« Back to Results
AEA Poster Session
Sunday, Jan. 3, 2021
7:00 AM - 6:00 PM (EST)
Monday, Jan. 4, 2021
7:00 AM - 6:00 PM (EST)
Tuesday, Jan. 5, 2021
7:00 AM - 3:00 PM (EST)
American Economic Association
"No Man Is an Island": An Empirical Study on Team Formation and Performance (L2)
Many organizations rely on decentralized arrangements where employees choose their projects and teams. Most of the empirical literature on working collaborations instead focuses on teams that are exogenously formed. I develop a structural entry model with heterogeneous strategic interactions where agents decide whether to join a project. The decision depends on who else may potentially join the project, the project quality, as well as other individual and project characteristics. In turn, this decision affects the probability of project completion. I estimate the model using a novel dataset from an important scientific collaboration. I find that agents' decisions to select into projects highly depend on the pool of teammates and the size of the team whereas projects' quality is of lesser importance. Heterogeneity in agents' characteristics explains this selection, which needs to be accounted for to obtain unbiased estimates of teams' performance. With a counterfactual experiment, I show that moving from a decentralized
to a centralized arrangement leads to fewer completed projects.
'Pay-Later' vs. 'Pay-As-You-Go': Experimental Evidence on Present-Biased Overconsumption and the Importance of Timing (D9, Q4)
When consuming goods provided by public utilities, such as telecommunication, water, gas or electricity, the predominant payment scheme is pay-later billing. This paper identifies one potential consequence of pay-later schemes, present-biased overconsumption of the respective good, and tests the effectiveness of pay-as-you-go schemes in reducing consumption. Specifically, I run a lab experiment which mimics an energy consumption choice and randomizes the timing of when consumption costs are paid: Either immediately ('pay-as-you-go') or one-week after consumption ('pay-later'). Results show that pay-as-you-go billing significantly decreases consumption, and in particular wasteful consumption. As the design controls for contaminating effects, these results can be solely attributed to present-biased discounting under the pay-later scheme. These results imply that pay-as-you-go schemes will be welfare improving both from agent's own perspective and from a social perspective if externalities are involved. In contrast, classic price-based policies will need correctives to account for present bias arising under pay-later schemes.
A 'Bad Beta, Good Beta' Anatomy of Currency Risk Premiums and Trading Strategies (F3, G1)
We test a two-beta currency pricing model that features betas with risk-premium news and real-rate news of the currency market. Unconditionally, beta with currency market risk-premium news is bad because of a significantly positive price of risk of 2.52% per year; beta with global real-rate news is good due to a negative price of risk. The price of risk-premium beta risk is counter-cyclical, while the price of the real-rate beta risk is pro-cyclical. Most prevailing currency trading strategies either have excessive bad beta or too little good beta, failing to deliver abnormal performance. Our empirical results can be delivered by a no- arbitrage model with precautionary savings and a pricing kernel characterized by two separate global shocks.
A Pay Scale of Their Own: Gender Differences in Performance Pay (M5, J1)
Using data from the online platform Glassdoor, we find that women earn 21% less and are 6.3 percentage points less likely to receive performance pay than men within the same employer and occupation. This substantial gender gap contributes to persistent disparities in total income and is directly related to the under-representation of women in performance-paying jobs. The gender gap worsens as women build their careers and is not driven by differences in latent ability and motivation, as proxied by school pedigree and major, or income growth. Instead, we find that women search less for, apply less to, and are less likely to be employed in performance-pay-intensive jobs, leading to shallower income growth since women earn and work for performance pay less often. These differences in search activity and representation cannot be explained by gender preferences for avoiding income risk, competition, high-paying jobs, managerial roles, or poor match quality.
Aid Against Trees? Evidence from a Community-Driven Development Program in the Philippines (Q3, O1)
Community-driven development (CDD) programs are becoming integral components in the development portfolios of major international development agencies, and are further being positioned as a parallel strategy to the sustainable development goals relating to climate change mitigation and adaptation. As this dual positioning occurs, little is known about the environmental effects stemming from such programs, especially in terms of deforestation. Using satellite-generated forest coverage data, this paper aims to address this gap in the literature by examining the impact of CDD programs on deforestation. More specifically, I apply a regression discontinuity design (RDD) and a randomized control trial (RCT) to a large-scale CDD program over two different time periods in the Philippines. Eligible municipalities in the RDD period experienced an average of 236 percent more deforestation and treated municipalities in the RCT period experienced an average of 265 percent more deforestation than the control. I then explore heterogeneous effects on a detailed dataset of the subprojects to show that the greatest impact on deforestation arose from infrastructure subprojects, which include trails, bridges and roads, followed by support, education and health facility subprojects. As international development agencies continue to invest heavily in CDD programs, more focus should be placed on the sustainability of such programs and on how CDD programs can be more in line with forest conservation policy.
Air Pollution Quotas and the Dynamics of Internal Skilled Migration in Chinese Cities (Q5, J6)
This paper examines the role of a sulphur dioxide (SO2) emissions quota introduced as part of China’s 11th Five-Year Plan on internal movements of high-skilled labour across Chinese prefecture cities. Using data on migration flows calculated through changes in Hukou status, our results suggest that a 1% increase in the SO2 emission reduction quota on average leads to a 0.15% increase in the high-skilled net outmigration. A possible mechanism could be that SO2 quota decreases the employment share in polluting industries at regulated cities in the short term, thereby resulting in a skilled net outmigration flows. However, this net outmigration trend alleviates in the long term due to improved air quality. Our findings contribute to a broader understanding regarding the effects of environmental policies on internal labour migration and labour force dynamics.
An Early Warning System for Tail Financial Risks (G2, C5)
This paper formulates an Early Warning System (EWS) for tail financial risks based on real-time multi-period forecast combinations of Value-at-Risk (VaR) and Expected Shortfalls (ES) of portfolio returns of non-financial firms and banks. Forecast combinations include baseline (VaR,ES) forecasts conditional on a domestic risk factor, as well as stress (sVaR,sES) forecasts conditional on CoVaRs of the risk factor, thereby integrating stress testing into forecasting. Using monthly data of the G-7 economies for the period 1975:01-2018:12, the proposed EWS delivers significant out-of-sample tail financial risk forecasts and reliable vulnerability signals up to a 12-month forecasting horizon, with stress forecasts in the combination improving forecasting ability prior to periods of severe financial stress.
Are Nonvoters Fence-sitters? (D7, H0)
In this paper, we compare 18- and 19-year-olds to 20- and 21-year-olds in twelve U.S. interim election years; the former is ineligible to vote in the presidential election two years before while the latter is eligible. Using the voting eligibility as an instrumental variable, we find that nonvoters are 23.5 percent more likely to affiliate to the same party with president than voters. Three placebo tests show that this finding is not driven by the age difference. Instead, we contend that cognitive dissonance is the main cause. Nonvoters, especially for those who would cast a vote for losers if they were eligible, tend to change their attitude after election to go along with most people.
Are Women Really Better Borrowers in Microfinance? Evidence from Matrilineal and Patrilineal Societies in India (C9, D9)
While the universal policy of gender targeting in microfinance stems, inter alia, from the conventional wisdom that women are better credit risks, there has been little research on the underlying reasons for it. This study examines whether social context and norms lead to gender differences in behavior among borrowers by conducting microfinance field experiments in comparable matrilineal and patrilineal societies in India. I observe a reversal of gender effect on loan default across the two societies. I find that women have a significantly lower default in the patrilineal society but significantly higher default in the matrilineal society compared to their male counterparts. I also find that matrilineal women are more likely to invest in risky projects than women in a patrilineal society. Moreover, they are likely to default strategically not only more than patrilineal women but even more than patrilineal men. My results suggest that gender difference among microfinance borrowers is driven by the differences in social context and gender norms across the matrilineal and patrilineal societies. My results indicate that policymakers should take into consideration the heterogeneity across societies and the social context in which a policy is implemented to be able to design better-targeted policies.
Artificial-Intelligence Assisted Decision Making: A Statistical Framework (C8, J2)
This paper proposes a statistical framework in which artificial intelligence can assist human decision making. Using observational data we benchmark the performance of each decision maker against the machine predictions, and replace decision makers whose information process quality is dominated by machine predictions based on the proposed criteria. The statistical frameworks that we proposed are applicable based on both Bayesian principles and frequentist principles of hypothesis testing and confidence set formation. Our theoretical discussion is illustrated by an example of birth defect detection, using a large data set of pregnancy outcomes and doctor diagnosis from the Pre-Pregnancy Checkups of reproductive age couples that are provided by the Chinese Ministry of Health. Based on doctor’s diagnosis, we find doctors, especially those who are from rural areas, can be replaced by the machine learning prediction. Statistically, the overall quality of our algorithm on a testable data set outperforms the diagnoses made only by doctors, with higher true positive rate and lower false positive rate. Our example also informs that decision making with artificial intelligence is more beneficial to poor areas relative to developed places.
Automation Adoption and Financial Regulation: Evidence from Stock Trading Firms and Workers (G2, G0)
How did trading automation impact broker-dealer firms and workers? While electronic trading platforms have been available for decades, widespread adoption of automated trading mostly occurred after the 2007 major market redesign promoted by the US Securities and Exchange Commission. With the intent of lowering access costs to stock markets, the policy fostered speed-driven competition between exchanges and trading firms. By leveraging several regulatory records to construct a rich linked employer-employee panel of equities traders, I study how employment, profits, and market structure were affected by higher returns to technology upgrading. Using variation in availability of local IT stock in investment firms, I find that automation eliminated 100 trading jobs on average during 2007-2009 for each additional computer per worker existing before SEC’s Regulation National Market System became effective. Through a series of tests, I show that these results are unlikely to be driven by the Great Recession or the rise in online brokerage services.
Bank Capital Requirements and Asset Prices: Evidence from the Swiss Real Estate Market (G2, R3)
In this paper, we empirically study the effects of bank capital requirements on asset prices. In particular, we analyze the 2013 globally first activation of the Basel III countercyclical capital buffer (CCyB) proposed by the Swiss National Bank. Since the Swiss sectoral implementation of the CCyB applies to residential mortgage lending only, we investigate the consequences of the intervention for the Swiss real estate market. As an identification strategy, we exploit the heterogeneity in treatment intensity across cantons, using data on the composition of mortgage supply in each canton. Our results suggest that the CCyB’s effectiveness in stabilizing asset prices crucially depends on the market’s underlying financing structure. We find that, being to a large extent financed by less mortgage-specialized universal or wealth management oriented banks, cantons with a more overheated real estate market were less affected by the intervention. Additionally, we show that cantons with higher exposure to the CCyB treatment experienced mitigated price growth for single-family houses. However, the effect is not observed for condominiums, which are less dependent on mortgage loans and are more considerably financed by “deep-pocketed” institutional investors seeking positive yields. The underlying channel is supported by our auxiliary finding that more CCyB-affected banks relatively reduced their mortgage lending activity. Our work raises important policy implications by shedding light on the intended and unintended effects of a novel macroprudential tool. For instance, in the presence of heterogeneous developments of real estate prices across regions, the CCyB requirements could be calibrated accordingly.
Bank Funding Costs and Solvency (G1, G2)
This paper investigates the relationship between bank funding costs and solvency for a large sample of euro area banks using two proprietary ECB datasets for both wholesale funding costs and deposit rates at the level of individual debt securities. In particular, the paper studies the relationship between bank solvency, on the one hand, and senior bond yields, term deposit rates and overnight deposit rates, on the other using a rigorous panel set-up. We find a significant negative relationship between bank solvency and the different types of funding costs and show that this relationship is non-linear, namely convex, for senior bond yields and term deposit rates. We also identify a positive realistic solvency threshold beyond which the effect of an increase in solvency on senior bond yields becomes positive. We also find that senior bond yields are more sensitive to a change in solvency than deposit rates. Among the deposit rates, the interest rates of the overnight deposits are the least sensitive. Banks' asset quality, profitability and liquidity seem to play only a minor role in determining the funding costs, while the ECB monetary policy stance, sovereign risk and financial market uncertainty appear to be material drivers.
Bank Loan Announcement Effects-Evidence from a Comprehensive 8-K Sample (G2, G3)
Using a comprehensive sample of over 10,000 bank loan announcements, we find results that differ from the findings of Maskara and Mullineaux (JFE 2011) and also that of Fields et al (JMCB 2006), which indicated that announcement effect of bank loans on borrower stocks disappeared as of late. We find bank loan announcements still have significant impact on borrowing firms' equity prices in our large sample, and our results are in-line with the findings of Billet et al (JF 1995), which was disputed by subsequent papers. Furthermore, we find that firms with lower abnormal spreads relative to the KMV-Merton default-risk model have higher announcements returns. We also document that, although information leakage (in terms of the run-up of borrowers' stock price prior to announcements) was quite significant in earlier sample periods, in recent periods there is much less information leakage prior to 8-K announcements of bank loans, and at least in this aspect the Dodd-Frank Act can be deemed as quite effective.
Barriers to Labor Migration for the Rural Poor: An RCT Experiment in Bangladesh (J6, O1)
Vocational training programs aimed at rapidly growing sectors have the potential to reduce skill gaps as well as improve income and employment possibilities. Such programs have often been unsuccessful because they are not driven by the industry demand and market linkages. Also, they do not tackle other barriers faced by trainees, such as savings and credit constraints, and the uncertainty of migration. We introduced a training ``plus'' program for the apparel sector jobs offered to the poor, rural households in northern Bangladesh, where we relaxed some of these constraints in a rigorous Randomized Control Trial (RCT) setting. Analyzing the program uptake demonstrated an interesting heterogeneity, where gender-specific social barriers, as well as risk and time preferences, play influential roles. Data from the follow-up surveys--six and eighteen months after the intervention--showed a statistically significant, persistent, and a large effect of the training program on employment when combined with an apprenticeship [on-the-job training (OJT)] or stipend component. We also found substantial income and remittance impacts, especially during the time of a seasonal shock, as well as a reduction in income poverty, both for the stipend and OJT groups. However, the rural household estimates--twelve months after the intervention--showed no impact on consumption poverty in the origin households.
Bond, Currencies and Expectational Errors (F3, G1)
We propose a model in which sticky expectations concerning short-
term interest rates generate joint predictability patterns in bond and
currency markets. Using our calibrated model, we quantify the effect
of this channel and find that it largely explains why short rates and
yield spreads predict bond and currency returns. The model creates the
downward sloping term structure of carry trade returns documented
by Lustig et al. (2019), difficult to replicate in a rational expectations
framework. Consistent with the model, we find that variables that
predict bond and currency returns also predict survey-based expecta-
tional errors concerning interest and FX rates. The model explains why
monetary policy induces drift patterns in bond and currency markets
and predicts that long-term rates are a better gauge of market’s short
rate expectations than previously thought.
Bunching Evidence of Cognitive Bias Caused by Eco-Labeling - The Case of Japan's Top Runner Program (Q5, L7)
In this paper, the effects of eco-labeling used in The METI's Top Runner program are investigated. The METI's Top Runner program is a program that attempts to improve the energy efficiency of durable goods, such as air-conditioner, refrigerator, and Television, in which the criteria is used that the energy efficiency of the most efficient durable goods in the previous period must be exceeded on average by each manufacturer's durable goods in the current period. The program used eco-labels to indicate whether the criteria had been met. Achieved products were labeled in green, while unachieved appliances were labeled in yellow.
Illustrating air conditioners case, we find clear evidence of bunching around the standard of criteria. Thus we prove the effects of color of the label causes the consumer behavior and attribute of the durable goods, because the standard is randomly determined and the standard can be met by a weighted average of each manufacturer's sales volume.
The technology is made public to some extent, when choosing the top runner criteria. Thus the manufacturer choose whether to produce the durable goods that achieves the standard at a certain cost, or to use existing technology. As it is, the information induction by eco-labeling cause all the manufacturer choose to achieves the standard.
Can a Rise of Intangible Capital Explain an Increase in Markups? (E2, D2)
This paper contributes to the current discussion on the reasons behand an increase in markups observed in several countries. Using a heterogenous-firm model, I show how the uncertainty and scalability properties of intangible capital imply that firms that succeed in their intangible capital investment can charge high markups relative to other firms. Sweden is one of the most intangibles-intensive economies in the world and I use data on all Swedish firms to study the empirical relationship between intangible capital and markups. I find that markups are positively related to intangible capital at the firm and industry level. Aggregate markups in Sweden have been low and stable over the past two decades. This finding provides evidence against the rise of intangible capital as the sole explanatory factor behind the rise in markups observed in other countries during the same time period.
Can Credit Default Swaps Improve Employee Treatment? (G3)
Employees are concerned about human capital risk when there is an increase in default risk for credit default swap (CDS) firms. We find that CDSs improve employee treatment ex ante, including employee compensation and employee welfare. The increase in employee welfare is mainly derived from firms' proactive cash profit-sharing programs. The results are robust to the endogeneity of CDS introduction. The positive effect of CDSs on employee treatment increases with employees' expected exposure to unemployment risk and employees' bargaining power. These findings suggest that credit derivatives can have real effects on employees by intensifying their concerns on human capital risk.
Can Estimated Risk and Time Preferences Explain Real-life Financial Choices? (D9, D1)
We combine experimentally elicited preferences with administrative micro data to study actual financial decision-making. Firstly, we estimate risk and (present-biased) time preferences in a real-life context, with horizons up to 10 years, for a large group of pension fund participants. We estimate a present-bias factor of 0.88, an annual discount rate of 3.91% and a CRRA utility curvature of 0.97. Secondly, using a life-cycle framework, we show that the individually estimated preferences explain actual retirement decisions up to 83% of our sample. Freedom of choice creates annual welfare gains up to 4.8%, but realized welfare gains are lower or even negative.
Can Food Voucher Assistance Program Enhance the Variety of Food Consumption? Evidence from a Field Experiment (D9, I3)
The Korean government is designing a food voucher assistance program (FVAP) to improve food consumption and nutrition intake of low-income households and promote consumption of agricultural products. As the Dietary Guidelines for Koreans emphasizes the importance of eating a variety of foods including whole grains, fruits and vegetables and dairy products, the task to design FVAP should pursue not only the quantitative improvement of food consumption (i.e., total food expenditure) but also the quality of food consumption that is often represented by the variety of purchased foods (i.e., Berry or Entropy Index). Prior to finalizing the FVAP design, a field experiment was conducted in two regions (rural versus urban) for three months in 2018 for approximately 1,200 household recipients (treatment group) and 400 non-recipients (control group) who were asked to report the details of all food shopping. Using data from the housekeeping book, this study investigated whether the introduction of FVAP improves the variety of purchased foods measured by Berry and Entropy Index. Furthermore, this study identified the factors associated with the variety of household food consumption by estimating a set of Tobit models, aiming for the most effective FVAP design. This field experiment study is expected to provide a priceless opportunity to assess each policy alternatives considered (i.e., cash versus voucher support; paper versus electronic voucher; limited versus unlimited supermarket access; restricted versus unrestricted food categories that are allowed to purchase using voucher).
Can Large Trade Shocks Cause Crises? The Case of the Finnish-Soviet Trade Collapse (F4, E3)
We study the macroeconomic consequences of a major trade disruption using the example of the Finnish-Soviet trade collapse in 1991. This is a rare case of a well-identified large trade shock in a developed economy. We find that the shock significantly affected Finnish output. While the direct trade effect was moderate, the shock was endogenously amplified by tightening credit conditions. Even so, the trade collapse was insufficient to generate an all-out crisis, and accounts only for a part of the Finnish Great Depression (1990−1993). We show that the financial system was both a trigger and an independent source of shocks throughout the depression.
Can Stock Market Bubbles in China Be Predicted? (G4, C1)
This paper presents an analysis of bubble prediction on China’s stock market based on the Shanghai Composite Index (SSE) and the Shenzhen component index (SZSE). The method that will be employed to define bubble periods is the multiple bubble testing that was generated by Phillips et al. (2015), which is based on sup augmented Dickey-Fuller (ADF) test. Through this method, the historical stock price data could be labeled as bubble periods or non-bubble periods. Then the classification machine learning models are used to predicate the upcoming bubble periods on Chinses stock market based on complexity, sentiment and structure variables. The complexity variables include autocorrelation, Hurst exponent, sample entropy, approximate entropy and Lyapunov exponent. The goal of this research is to accurately predict China’s stock market bubbles in order to help investors or policy makers to foresee the upcoming risks and could make better trading strategies or policy. Meanwhile, this paper also aims to contribute on the discussion between rational bubble and market efficiency.
Can Technology Solve the Principal-Agent Problem? Evidence from China’s War on Air Pollution (Q5, O3)
We examine the introduction of automatic air pollution monitoring, which is a central feature of China’s “war on pollution.” Exploiting 654 regression discontinuity designs based on city-level variation in the day that monitoring was automated, we find that reported PM10 concentrations increased by 37% immediately post–automation and was sustained. City-level variation in manipulation is negatively correlated with income per capita and positively correlated with true pre-automation PM10 concentrations. Further, automation’s introduction increased online searches for face masks and air filters, suggesting that the biased and imperfect pre-automation information imposed welfare costs by leading to suboptimal purchases of protective goods.
Casualties of Trade Wars (F1, F5)
Although trade wars have existed throughout modern history, there is little empirical evidence as to how countries choose which industries to target for retaliatory tariffs. We approach this question by studying the retaliatory measures applied by numerous countries to President Trump’s Section 232 steel and aluminum tariffs. We contrast this to the EU and Canadian retaliation against the US’s Continued Dumping and Subsidy Offset Act (CDSOA) of 2000. President Trump’s tariffs were broadly maligned, and its commensurate retaliation avoided the WTO dispute settlement process. In contrast, the CDSOA was opposed by President Clinton and the retaliatory measure that followed were WTO-approved. Our empirical results indicate that retaliation to President Trump’s tariffs were more likely to be applied to politically sensitive industries, including products made in voting districts of House and Senate leaders as well as larger industries. We also find evidence of heterogeneity across countries in their retaliatory behavior, with the EU displaying an increased tendency to target politically sensitive industries, including those located in Presidential swing states. Our results suggest these effects were absent in retaliation against the CDSOA. We also find evidence that the EU and Canada were more likely to target industries in which they had more market power, and thus were able to extract more welfare from the US as predicted by terms-of-trade models of trade protection. Finally, industries benefiting from the original policies that led to retaliation were also more likely to face sanctions.
Causal Effect of Local Social Capital on Political Violence and Its Relational Dynamics: Evidence from Africa (Z1, N4)
This paper has two aims: 1) to identify the causal effect of social capital on internal violence in Africa; and 2) to find the relational mechanism that mediates the adverse effect of the local social capital on political stability. To isolate the exogenous variation of social capital, this paper uses the mode of production of an ethnic group in the precolonial period determined by the proportion of common resource–rivers and lakes–of its historical homeland. As common-pool resource requires collective management that can lead to the emergence of solidarity within the group, it exogenously extends the level of individual trust from the family to the extrafamilial community. Combining the value survey data covering 69,890 individuals in 24 African countries from 2004 to 2017 with the data of satellite-measured water bodies of the homeland of the respondent’s ancestor, this article finds that a 1 standard deviation increase in the trust in local chief raises casualties of the battles against the state militants by 0.85 standard deviations. However, higher trust in local chief reduces non-battle violence against civilians and it exerts no discernible effect on non-state conflicts. Moreover, examining within-country variation, this paper uncovers that the broadcasting of state power, estimated by the regional density of strategic roads, proportionately alleviates the offensive nature of local social capital. Averaging above the 50th percentile of the road density, the causal effect of local social capital is transformed to be inclusive to the central state authority, having the parallel pacifying effects to both within and external entities. The evidence not only highlights the detrimental repercussions of local social capital on the regime stability in Africa, but also reveals the relational nature of local social capital whose effect is largely determined by its connectiveness with the state in which the local community is embedded.
Climate Risk Perceptions and Demand for Flood Insurance (R2, G5)
Using detailed micro-level data, we show that individuals' beliefs about climate change influence their choice and level of flood insurance coverage. Our empirical strategy exploits the heterogeneous impact of widening partisan polarization on climate change beliefs after the 2016 general election. We find that, in areas where flood insurance is not mandatory, a one-standard-deviation drop in the fraction of adults who believe global warming is happening leads to a 26% drop in the demand for flood insurance. In areas where flood insurance is mandatory, a similar drop in beliefs is associated with a lower propensity to carry voluntary content coverage and a higher likelihood of choosing the maximum deductible amount. As a secondary test, we exploit the flood insurance premium increases due to the Biggert-Waters Flood Insurance Reform Act of 2012. We show that homeowners who do not believe global warming is happening were more likely to terminate mandatory flood insurance coverage by prepaying mortgages.
Coherent Preferences and Asset Prices in Stock Market (C2, G4)
We study individual coherent preferences underlying asset prices and propose a set of explicit models for nonlinear V-shaped price pressure utility in a new framework. Coherent preferences are consistent interactive choices between momentum trading and reversal trading in stock market where market dynamic equilibrium exists. We find that: 1) coherent preferences generate nonlinear V-shaped price pressure and market dynamic equilibrium whereas beliefs contribute to discrepancy between market equilibrium prices and fundamental prices; and 2) subject individuals can display either disposition effect or inverse disposition effect. Our study suggests a better asset price model with trading volume distribution in finance.
Community Advocacy Forums and Public Service Delivery – Impact, and the Role of Information, Deliberation, and Administrative Placement (O2, I1)
To improve governance and public service delivery, the Government of Uganda organizes community forums – popularly known as barazas – where citizens receive information from government officials, and get the opportunity to directly engage with them. We run a cluster randomize control trial to assess the impact of the baraza intervention on a range of outcomes related to agriculture, health, education, and infrastructure. Using a factorial design, we further test the relative importance of the two main components of the intervention – information provision and citizen engagement. Furthermore, we compare the effectiveness of barazas organized at the district level to the effectiveness of barazas organized at the sub-county level, as the administrative placement of the barazas is a key determinant of the cost-effectiveness of this policy intervention.
Competition, Non-Patented Innovation, and Firm Value (G3, L1)
This paper studies how competition impacts non-patented corporate innovation and firm value by exploiting adoptions of state anti-plug molding laws – laws that prohibit “unscrupulous” reverse engineering by competitors – and their subsequent invalidation by the U.S. Supreme Court. Firms decrease patenting activity following the laws’ adoptions while also showing increasing investment spending, profitability, and value. Value gains are larger for firms at greater risk of imitation, and that are more innovative. After the laws are overturned, firms reinitiate patenting whereas prior investment spending, profitability, and value gains dissipate. These results suggest that more intense product market competition disincentivizes value-enhancing corporate innovation.
Computers with Internet Access and Wage Disparities across Regions: Evidence from China (J3, L8)
A long-standing question involves whether technological developments necessarily enhance people’s welfare or lead to relative inequality. If a dilemma exists between economic growth and disparity, a resolution to this puzzle would be worthwhile. Further, the widespread utilization of computers has created a technological revolution, as computers—and Internet access in particular—can help workers conduct complex tasks and decrease the costs for routine work. Specifically, the use of information communication technology has been found to promote workplace productivity and increase relative wages for skilled labor. Consequently, any wage disparity may be larger between skilled and unskilled labor due to their differences in both education and the ability to use advanced technology. Our research question focuses on whether the use of computers with Internet access augments the wage disparity across regions, as indicated with statistical data from provinces, cities, and counties in China. We use this data to empirically investigate the relationships between the different degrees of adoption of computers with Internet access, human capital, and the wage gaps across regions. We find that the use of computers with Internet access is associated with regional wage disparities and GDP growth in general. However, we also observe a U-shaped relationship between technology use and wage inequality. We further test whether educational endowments are a solution to this inequality, or a catalyst to enlarge the income gap for workers; the results hold with multiple model modifications. We discuss policy implications based on the empirical results.
Conceptual Differences Between Decision Utility And Experienced Utility: A Theory for Jevons' Wish? (D9, C0)
In the 19th century Jevons wished to capture the quantity of feeling. My theory of experienced utility aims to satisfy such wish. I define experienced utility as a function which represents an individual's hedonic experience from an activity over time. My theory is descriptive and relies on three assumptions:
I) Finite number of activities;
II) Choice of a single activity; and
III) Rate of change of experienced utility is proportional to difference between the experienced utility and the other experienced utilities up to a positive coefficient function.
The theory is presented by discussing primal, existential and functional differences between decision utility and experienced utility. Primal differences are in terms of their primitives. Experienced utility is a more basic primitive than preference relation. Existential differences are in terms of the assumptions required for the existence of each concept of utility. Functional differences are in terms of functional forms for each concept of utility. My theory is fundamentally different from the one by Kahneman et al. (1997). They present a normative theory of total experienced utility extending decision utility whereas I present a descriptive theory of instant experienced utility independent of decision utility.
Because my theory does not require rationality or include preferences I view my assumptions to be weaker than those required for the existence of decision utility. I prove the existence of a unique family of experienced utilities which are expressed explicitly, take real values and are linearly independent. A concept associated with each experienced utility is what I have termed as non-experienced utilities. Non-experienced utilities are experienced utilities from activities which an individual has not chosen while spending time on the chosen activity. If used to analyze individual choice, my theory has the potential to explain both time allocation and the sequence of activities. A current area of research is analyzing how non-experienced utilities determine the switch time from an activity to another.
Conditional Cash Transfers and High School Attainment: Evidence from a Large-Scale Program in the Dominican Republic (I2, I3)
Over the past two decades, conditional cash transfer (CCT) programs have been implemented extensively across the world, reaching over sixty countries, and have become one of the most popular social programs in developing regions. In Latin America, roughly one of every four individuals in 17 countries have received cash transfers. While there is an abundance of literature on the short-term effects of CCT programs on educational and non-educational outcomes, much less is known about their longer term effects in part due to the difficulty of following individuals for prolonged periods of time. This paper examines the impact of a large-scale CCT program, PROSOLI, on high school attainment in the Dominican Republic from 2005 to 2017. We implement a quasi-experimental approach combining extensive administrative, household and educational records from program and nonprogram participants across the country. We find that exposure to school transfers is, on average, associated with a 6-8 percentage points higher probability of completing high school relative to nonparticipation. The effect of the program seems larger on urban areas and increases with the amount of program exposure, while we do not observe major differences between male and female participants. The results are robust to alternative estimation methods and the use of different samples. Considering that one in two male students in the public school system who finish elementary school do not complete high school in the country and the returns to high school attainment are 26% (9%) higher than only completing primary school (not completing high school), the subsequent program effects on earnings are non-negligible. The study contributes to the still relatively scarce literature analyzing longer term schooling effects of CCT programs and to the discussion in the Dominican Republic about the lasting effects of this large-scale social program, which has benefited more than a quarter of the population.
Consumption Tax Reform and the Real Economy: Evidence from India's Adoption of a Value-Added Tax (H3, H2)
We study the impact of a consumption tax reform on firm capital and productivity by examining India's replacement of the sales tax with a value-added tax (VAT). Unlike the sales tax, the VAT allowed firms to offset their tax liability with VAT paid on capital inputs, effectively reducing the tax-related cost of capital. Exploiting the staggered adoption of the tax reform across Indian states, we show that VAT adoption increased firm capital by 3%. The effects are driven by financially-constrained firms -- an important source of heterogeneity in a developing country context. We also document a corresponding improvement in the productivity of financially-constrained firms. Our findings thus suggest that beyond revenue generation, consumption tax reforms can have the additional effect of stimulating investment and productivity in resource-constrained environments.
Corporate Board Diversity and Disruptive Innovation (G3, O3)
This study examines the effects of board diversity on disruptive innovation at the firm, and the underlying mechanisms for these effects. Prior research suggests that diverse corporate leadership, including the board of directors, is critical for firm value creation. However, it remains unclear how board diversity influences firm development of disruptive technology, which is a cornerstone for the creation of shareholder and stakeholder value, especially in the long run. Using an instrument approach, we identified the causal relationship between board diversity and firm innovation. Specifically, board diversity in demographics and cognitive characteristics can increase the quantity, impact, disruptiveness, and value of firm innovation. We probe resource-, governance-, and risk-based mechanisms to explain the baseline results. We document evidence that a diverse director board spurs disruptive innovation by encouraging inter-firm technological collaborations, appointing skilled upper-echelon managers, and increasing firm risk tolerance to pursue R&D investments.
Corporate Leverage Ratio Adjustment under Cash Flow-Based Debt Covenants (G3, G2)
Debt covenants attempt to solve the agency problem between shareholders and bondholders, but it is unclear how they impact a firm’s speed of adjustment towards its target capital structure. While covenants constrain firms in their ability to change their leverage, they may serve as a strong incentive to actively manage the capital structure. Recent research suggests that the presence of covenants slows down firms in their adjustment. However, in the absence of an appropriate measure of covenant slack, these findings are based on simple covenant counts or the probability of violation one period ahead. We argue that both methods fall short. Firstly, simple counts neglect how constrained a firm is by the prescribed financial ratios and the rich nature of covenant types. Secondly, while violations can be costly, lenders frequently waive them, and thus the violation probability may not be management’s most pressing concern. To overcome these shortfalls, we introduce a new measure for covenant slack. Our non-parametric method makes different covenant types comparable and relates a firm’s slack to that of other firms and other time periods. Our findings challenge existing research by showing that the mere presence of covenants increases the speed of adjustment. Moreover, using our proposed new measure, we show that firms with the least slack adjust significantly faster towards their target.
Corruption as Collateral (O1, E4)
We propose corruption can substitute for conventional collateral in enforcing financial commitments
when institutions are poor. A theoretical framework with agency frictions is built, in which
corruptive relations with government officials keep firms committed to loan payments. Based on
this framework, we hypothesize the anti-corruption investigation destroys the commitment mechanism
so that firms default and, most importantly, firms default strategically as long as they can
substitute corruption with other collateral. We investigate regional data and firm-level data from
China, and find powerful evidence supporting our hypotheses.
Cost of United States New Protectionism for Mexico (F0, E0)
We investigate the short-run impact of new wave of U.S. protectionism on Mexico's economy. First, we use event studies approach to see the impact of changes in trade policies on Mexico's imports and exports for the period of 2016-2019. Second, we apply Zoutman et al. (2018) approach to estimate the trade elasticities for Mexico using the six-digit level of the Harmonized System product codes. Third, we modify a small open economy DSGE model developed by Christiano et al. (2011) to study the effects of U.S. protectionism on Mexico's economy and the impacts of potential changes of trade policies on Mexico's economy through different bilateral trade shocks. The estimated elasticities of Mexico's imports demand and exports supply are -1.08 and 0.44 respectively. Our result shows that in the short run, the imposition of 10 percent tariffs on Mexico's exports potentially may reduce the Mexico's GDP to more than 4 percent, and 10 percent retaliatory tariffs on Mexico's imported consumption goods may reduce consumption by less than 2 percent and increase CPI inflation by less than 5 percent.
Counterpoint Theory and the Novel Function Able to Classify All Continuous Solutions to the Nash Bargaining Problem (C7, D5)
This paper is the first to solve the Nash bargaining problem of how two firms optimally share a surplus they jointly create in repeated bargaining rounds by defining the single mathematical function able to classify all continuous solutions to the Nash bargaining problem consistent with Nash’s axioms of Pareto optimality, symmetry, invariance with respect to affine transformations of utility, and independence of irrelevant alternatives, and the alternative axiom of monotonicity. Here, a unique equilibrium point is determined, representative of an infinite number of points in R3 that cannot be reached by any two vector combinations. This unique solution is defined by paired point theory where the unique equilibrium point solution represents a pair of equilibrium points and corresponds to a pair of paired equilibrium points in R2. This unique equilibrium solution is modeled using a continuous bi-continuous function f: JR3 of J, a manifold open mobius band, on which each point on the band (which is an R2 projective plane) uniquely represents a pair of equilibrium points and corresponds to a pair of paired equilibrium points. It is useful in explaining the unique equilibrium solution of balanced vertically integrated firms which integrate both forward and backward to avoid the economic hold-up problem of incomplete markets.
Credit, Income and Inequality (G2, O1)
Analyzing unique data on credit granted to individuals who are majority owners of small firms,
we detail how a bank’s credit decisions affect the future income of accepted versus rejected loan
applicants. The bank’s cutoff rule, which is based on the applicants’ credit scores, creates a sharp
discontinuity in the decision to grant loans. We show that application acceptance increases
recipients’ income five years later by more than 10% compared to denied applicants. This effect
is mostly driven by the use of borrowed funds to undertake investments and is more pronounced
when positive soft information contributes to loan approval. We then quantify unequal income
developments following the bank’s credit decisions across regions and time. We document that
credit approval has a stronger effect on applicants’ income in low-income areas and during a crisis
period, when individuals are more credit constrained.
Cross-Border M&A and the Exchange Rate: Evidence from Switzerland (F3, G3)
I exploit the natural experiment induced by the 2015 Swiss National Bank repeal of the minimum exchange rate of 1.20 CHF per Euro to empirically test Blonigen’s (1997) theory. I find evidence that a sudden, sizeable, and persistent appreciation of the local currency is associated with reduced cross-border mergers and acquisitions activity involving local targets relative to comparable countries. I complement my study by applying the data-driven procedure devised by Abadie, Diamond, and Hainmueller (2010), known as the Synthetic Control Method, to construct a synthetic Switzerland. My findings are supported by the discrepancy between the actual number of cross-border deals involving domestic targets and its synthetic counterfactual, representing how the number of cross-border M&As would have evolved in Switzerland in the absence of the exchange rate intervention. In addition, in line with Blonigen’s (1997) model, I find a larger effect for high-technology firms.
CROSS-STATES HETEROGENEITIES IN THE TRANSMISSION OF THE U.S. NARRATIVE TAX CHANGES (R1, H0)
This paper investigates the assumption of homogeneous effects of federal tax changes across the U.S. states and identifies where and why that assumption may not be valid. More specifically, what determines the transmission mechanism of tax shocks at the state level? How vital are states' fiscal structures, financial conditions, labor market rigidities, and industry mix? Do these economic and structural characteristics drive the transmission mechanism of the tax changes at the state level at different horizons? This study employs a panel factor-augmented vector autoregression (FAVAR) technique to answer these issues. The model identifies tax shocks using sign restrictions with Uhlig's penalty function. The findings show state economies respond homogeneously in terms of employment and price levels; however, do react heterogeneously in terms of real GDP and personal income, and, in most states, these reactions are statistically significant. The heterogeneity in the effects of tax cuts is significantly related to the state's fiscal structure, manufacturing and financial composition, and the labor market's rigidity. The cross-state regression analysis shows that states with higher tax elasticity, higher personal income tax, strict labor market rigidity, and policy uncertainties are relatively less responsive to federal taxes. In contrast, the magnitude of the response in real GDP, personal income, and employment to tax cuts is relatively higher in states with a larger share of finance, manufacturing, lower tax burdens, and flexible credit markets.
Cyclical Drivers of Euro Area Consumption: What Can We Learn from Durable Goods? (E2, E3)
We study the cyclical dynamics of consumption in the euro area (EA) and the large EA countries by distinguishing durable from nondurable expenditures. We adopt a theoretical partial equilibrium framework to justify the identification strategy of our empirical model, a time-varying parameter structural vector autoregression (TVP-SVAR). Following the main insight from the theoretical model, that liquidity constraints induce important interactions between durables and nondurables, we distinguish durable-specific demand and supply shocks, while taking into account monetary and credit conditions. Our main findings are: (i) durables react faster and more strongly than nondurables after monetary shocks in the euro area and in the largest EA countries, a confirmation of an outcome commonly reported for the US; (ii) there is a large degree of cross-country heterogeneity in how different factors (including durable-specific ones) explain consumption; (iii) the strength of spillovers from durable to nondurable consumption, as predicted by theory, is empirically correlated with how much households across countries are likely to be liquidity constrained.
Dark Banking? Banks and Illicit Financial Flows (G2, K4)
Do banks enable crime? Does regulation insulate finance from criminal activity? I address these questions using evidence from the drug trade in Mexico, finding that local drug cartel activity causes an increase in bank deposits. Accordingly, branch networks grow in affected areas; this growth is not driven by increased lending opportunities. After the election of a “law-and-order” government, these effects dissipate, with liquidity flowing into branches of U.S. banks along the border. I interpret this as evidence that “finance follows crime” in weak institutional environments, and that, absent transnational policy coordination, regulatory arbitrage via cross-border liquidity flows undermines banking regulation.
Deep-Rooted Cultures and Gradual Diversification: How to Shape Housing, Credits, and Consumption (E5, R0)
In terms of uncertainty avoidance and indulgence, cultures influence dynamics across housing, credits, and consumption. The analysis in this study is restricted to the Eurozone, eliminating differential monetary policy impacts. We evidence contrary relationships between housing construction and private credits, comparing uncertainty avoidance panels that indicate credit approval determinants to risk aggressiveness and wealth effect in the low-level panel. Recursive relationships among house prices, construction, and credits in the low-indulgence panel indicate higher homeownership. Migration increases cultural diversification that leads to short-term disequilibrium, interactions between construction and credits, and wealth effect. Diversification fosters financing channels other than credits and homeownership.
Default Costs and Fiscal Limits in a Small Open Economy (F4, E6)
We revisit the link between default costs and the risk of default on government debt. We show that with high default costs two equilibria may co-exists. In the `good' equilibrium households' investment in foreign assets is low, revenue from tax collection is high and the probability of default is low. In the `bad' equilibrium households hedge against losses associated with default by investing large proportion of their income in foreign assets. This reduces output, decreases revenues from tax collection and raises the probability of default. We conclude that high default costs may tighten fiscal limits and increase the probability of default. This result is at odds with the notion that the higher the default costs, the lower the probability of default. At the same time, it is in line with the literature asserting that financial liberalization may raise risks of sovereign default.
Differential Effects of Affirmative Action for College Admissions – Evidence from a Selective Law School in Brazil (H4, J4)
Affirmative action (AA) policies continue to be a controversial solution for leveling the playing field for college admissions. We use one of the first quota policies implemented in Brazil, at Rio de Janeiro's State University (UERJ), to separately investigate its long-term effects for AA applicants who benefited from the policy and non-AA applicants displaced by it. We focus on applications to the undergraduate law major at UERJ for three reasons. First, UERJ's application process allows us to identify applicants to either AA or non-AA slots, and, among them, those who were offered admissions. Second, this is a highly selective undergraduate program. A 30-40 point (out of 100) difference in the cutoff scores between AA and non-AA shows that AA applicants were subjected to a much lower bar for admissions. Third, a high-stakes post-college exam (lawyers' licensing process) enables tracking applicants into the law career after college. In addition, we combine government data, including employment information (RAIS), firm ownership, and graduate degrees, along with online scraped data for the lawyer licensing exam, internship applications, and college graduation. For applications between 2006-2011, we are able to track about 87\% of AA and 79\% of non-AA applicants around each group-specific cutoff across outcomes. Our results suggest that, for beneficiaries, this AA quota policy increases the probability of graduating from college from 41\% to about 80\%, becoming a certified lawyer from 31\% to about 70\% , and being employed as such from 13\% to about 30\%. We find that applicants displaced by the policy do not appear to be negatively impacted, possibly because they could've been admitted to other quality universities. We estimate that non-AA applicants both slightly above (admitted) and below (displaced) the cutoff have a 71\% chance of graduating from college, about 70\% of becoming a licensed lawyer, and 25\% of being employed as such. We interpret the net effect of this policy to be positive, giving opportunities for those who don't typically have it without any significant direct impacts on others.
Differential Impact of New Public Housing Announcement on the Property Prices in Rich and Poor Neighbourhood (D8, R1)
This study investigates the differential effect of the announcement of the locations of new public housing complexes on the property prices in expensive and poor suburbs. Our novel approach employed the difference-in-differences method in conjunction with hedonic quality adjustment facilitated by (1) a single unanticipated government announcement of new public housing in multiple locations in Canberra, the capital of Australia, and (2) the fact revealed by the property price data of actual transactions that these locations include both more-expensive and less-expensive suburbs than their respective adjacent suburbs. Our investigation reveals that, the announcement of new public housing has a negative effect on the property prices in expensive host suburbs but not on the less-expensive ones. Our findings can make important contributions to policy decision-making in finding the right location of public housing.
Discretionary Trading: Evidence from the FIFA World Cup (G1, G4)
This paper rigorously tests the predictions of adverse selection models with discretionary trading. This is achieved by exploiting intra-day discretionary trading surrounding FIFA World Cup football matches that occur during trading hours. The extraordinary volatility and price discovery dynamics that occur on match days conform to the predictions of adverse selection models. Price volatility is increased by up to 0.223 standard deviations prior to matches. Volatility is reduced by 1.301 standard deviations when matches extend into extra-time. Consistent with the Back and Pedersen (1998) model, there is limited evidence of systematic patterns in transaction costs on match days.
Disentangling the effects of multidimensional monetary policy on inflation and inflation expectations in the euro area (E5, C5)
The European Central Bank (ECB) has adopted a mixture of conventional and unconventional tools in order to achieve its mandate of price stability in a low-inflation, low-interest-rate environment. This paper contributes to the existing literature by providing a taxonomy of the ECB's policy toolkit and by evaluating its implications on price stability and the anchoring of inflation expectations. Developing a novel high-frequency identification scheme for a large Bayesian Vector Autoregression, I find evidence that forecasters revise their long-term expectations upwards in response to quantitative easing and forward guidance shocks. Consequently, inflation increases and remains significant for over a year after the shock, which stresses the crucial role of expectations for the transmission of monetary policy.
Dividing Lines: Racial Segregation between Local Governments in U.S. Metropolitan Areas (H7, J7)
A large literature in economics studies the roots of urban segregation as a result of endogenous sorting of households across local housing markets. This theory assumes that neighborhoods have fixed amenities that are valued differently by households, and that segregation is the result of racial differences in preferences over amenities such as public school quality and policing services. But there is little empirical work quantifying the role of the political fragmentation of cities and local jurisdictional boundaries in shaping segregation. Political fragmentation may be an important driver of today’s segregation, as an emerging literature documents governments’ explicit role in creating racial segregation in US cities during the early 20th century (Rothstein 2017, Aaronson et al. 2020).
In this study, we comprehensively quantify how much today’s urban racial segregation and inequality is attributable to local political fragmentation and jurisdictional boundaries - municipalities and school districts. We begin with a descriptive analysis of the correlation of racial inequality in economic outcomes and between-jurisdiction segregation in metropolitan areas. We show that between-jurisdiction segregation explains a substantial share of the variation in racial inequity across cities, a larger share than that explained by within-jurisdiction segregation. These patterns establish our motivation for a deeper investigation of the impact of jurisdictional boundaries on urban segregation.
Next, we combine granular GIS data on local government jurisdictions with 2010 census block data on population by race. This novel dataset identifies residential blocks located near boundaries, enabling useful comparisons between blocks on either side of a local government boundary, similar to a regression discontinuity design. We use this dataset to estimate boundary discontinuities in racial composition across all adjacent local government pairs in US metropolitan areas (N > 20,000). These estimates identify places where the racial makeup of residents "jumps" as soon as one crosses a boundary, allowing us to make statistical statements about the most racially unequal “dividing lines” in the country. We find considerable cross-sectional variation in the extent to which city segregation is explained by local discontinuities across particular political boundaries. In some cities, particular dividing lines seem to be key drivers of segregation, akin to “racial borders”, raising important policy considerations.
To understand whether boundaries have a causal impact on segregation today, we estimate the effect of local jurisdiction changes on the racial composition of census blocks across the three most recent census waves (1990, 2000, 2010). We first show that there have been substantial changes to school district and municipal jurisdictions across this time period, especially in some states like Tennessee and Missouri. We infer the causal effect of jurisdictional change on racial sorting by comparing blocks that were reassigned to jurisdictions with greater (or lower) minority resident populations, to blocks in the same neighborhood that remained in the original jurisdiction. In essence, we implement a flexible selection on observables research design, relying on an assumption that within small geographic neighborhoods the likelihood of reassignment is uniform across all residential blocks, once we control for block composition and population prior to reassignment.
Do Counter-Stereotypical Female Role Models Impact Women's Occupational Choices? (J0, J1)
Despite the ``grand convergence'' in the gender gap in education and labor market outcomes over the last 30 years, women are still underrepresented in lucrative and competitive professions, such as STEM and business, which in turn perpetuates the gender pay gap. Among other reasons, women might be avoiding male-dominated fields because of a lack of appropriate female role models that would otherwise nudge them into these fields. In this study, to the best of our knowledge, we are the first to create a systematic measure of counter-stereotypical female role models based on a long time series of public opinion surveys and study its impact on occupational choices and labor market outcomes for women in the US.
Based on the Gallup survey question ``What woman do you admire the MOST?'' we create a measure of Counter-stereotypical female role models from 1951 to 2014. It is based survey respondents indicating that they most admire women who are politicians, writers, journalists, businesswomen, astronauts, or activists. On the contrary, stereotypical female role models include women who are famous for their husbands or sons, and women who work in gender stereotypical occupations such as actress, singer, or nurse.
We find that women admiring counter-stereotypical female role models are more likely to work full time. They are also more likely to enter male dominated industries and occupations. Further, their occupations are characterized by abstract tasks and they are more likely to work in managerial positions. We also find that the gender pay gap is lower in states where counter-stereotypical female role models are admired.
To establish causality, we use a quasi-natural experiment and examine a negative shock to counter-stereotypical female role models when Hilary Clinton lost the democratic primary in 2008.
Do Expert Panelists Herd? Evidence from FDA Committees (D0, I1)
We develop a structural model to address the question whether, and to what extent, expert panelists engage in herd behavior when voting on important policy questions. Our data comes from FDA advisory committees voting on questions concerning the approval of new drug applications. We utilize a change in the voting procedure from sequential to simultaneous voting to identify herding. Estimates suggest that around half of the panelists are willing to vote against their private assessment if votes from previous experts indicate otherwise and, on average, 9 percent of the sequential votes are actual herd-votes. Temporary committee members are more prone to herding than regular (standing) members. We find that simultaneous voting improves information aggregation given our estimates.
Do Financial Incentives Matter for Maternal Healthcare? Evidence from Programs in India (I1, I0)
This paper evaluates the impact of financial incentives provided by Janani Suraksha Yojna (JSY), a conditional cash transfer program, on maternal health behavior and child mortality in India. JSY provides cash assistance for delivery at public facilities; it alters women's choices by changing the relative prices of different delivery options. Using a difference-in-differences approach, I exploit JSY eligibility variations across individuals and states and the program's timing to estimate its impact. I find that JSY significantly increases public facility deliveries. The increase comes from shifts away from both home births and private facility delivery. Besides, the decline in private facilities' use is considerably larger than the decline in home births. I nd a modest effect of the program on child mortality, explained by the small decrease in home births. I also estimate the impact of JSY eligibility on women's pregnancy timing using a discrete-time hazard model. I find that the program reduces teen pregnancies. I further estimate the impact of JSY on the use of ante- and post-natal services, and its heterogeneous impact on women by education, wealth, and social group. Policymakers could integrate the program with financial incentives for other supporting healthcare services to improve health outcomes further.
Do Poor Countries Gain from Rich Countries’ Tariff War? (F1)
President Trump’s willingness to use tariffs to “make America great” vis-à-vis America’s major trade partners has made trade war a reality, especially with China. This literature is concerned almost uniquely about the benefits and harms of trade war for the two belligerent countries typically in two-country models. It is the primary objective of this paper to shift focus from the literature and examine the effect of bilateral trade war between rich and large countries on bystander countries, especially developing countries, which are usually too small to influence the terms of trade and dependent on the vagaries of rich countries’ trade policy. We use a three-country Ricardo-Matsuyama (2000) model to study the effects of a tariff war between a high-income and middle-income country who can both trade freely with a low-income country. Furthermore, the consequence of Nash equilibrium tariff wars (with non-negligible endogenous tariff rates) are also analyzed numerically. Our main results are as follows, First, the larger country tends to win the trade war but when the belligerent countries are about the same size both lose the war. We also find that the low-income country is made better off as the medium-income country gets larger. Most interestingly, there exist a range of population values in which both the high-income and middle-income countries lose the tariff war, yet the low-income country is made better off by the tariff war. Lastly, we study exogenous changes in immigration policies to analyze the effects on welfare. When the middle-income country is relatively small compared with the high-income country, trade war makes the low-income country worse off. We find that in such case out-migration of labor exacerbates the harm done to those remaining in the low-income country. The opposite is true when the middle-income country is relatively large.
Do Potential Future Health Shocks Keep Older Americans from Using Their Housing Equity? (J1, R2)
Housing constitutes a significant share of assets held by retired Americans as most own a home. However, there is little evidence that retirees use housing equity as a source of income. Most homeowners remain in their home throughout retirement and very few take-up reverse mortgages. This contradicts the predictions of the Life-Cycle Hypothesis (LCH) which suggests that households save during their working years and draw down those savings in retirement.
As homeowners age, it becomes more difficult to borrow money. Therefore, they may choose to engage in precautionary savings using their home and sell it to cover unexpected medical bills. In this paper, I explore how this option might preclude homeowners from using the equity in their home to increase consumption in retirement.
I use the Health and Retirement Study to model and calibrate an economy that consists of overlapping generations of heterogeneous agents who must make decisions in each period: whether to rent or buy, what size house or apartment to inhabit, and how much to spend on consumption in each period. In late retirement, households have a chance of receiving a negative health shock where they incur expenses that they are forced to pay for through either their income or accumulated assets, including the home. I compare this to an economy where there are no health shocks and find that homeownership rates decrease by as much as 13-percentage points.
I find that if retires had a health insurance policy that covered all out-of-pocket medical expenses, then even with a potential health shock, there is a 13-percentage point decrease in homeownership rates. This suggests that some households do use their house as precautionary savings and that if Medicare was expanded to cover these medical bills, then more homeowners would act in accordance with the predictions of the LCH.
Does Building Highways Reduce Traffic Congestion? (R4, D5)
In a seminal study, Duranton and Turner (2011) find evidence that points to the existence of the fundamental law of highway congestion in the US. They build a causal model using an instrumental variable (IV) approach that yields an estimate of 1.03 for the elasticity of vehicle miles traveled (VMT) to the stock of interstate highways in US metropolitan areas. The result means that government efforts to alleviate traffic congestion by expanding capacity are likely to fail — any increase in the stock of highways is accompanied by a commensurate increase in VMT, leaving travel times unaffected. In this article, we explore the impact of unobserved heterogeneity on the fundamental law. We begin by using a simple partial equilibrium model to demonstrate how metropolitan statistical areas (MSAs) that are identical in most respects but have different initial congestion levels respond differently to added capacity due to individual differences. These differences in MSAs gives rise to heterogeneity in the elasticity of VMT to capacity. We derive conditions under which the elasticity decreases with the initial congestion level. We then revisit the empirical analysis in Duranton and Turner (2011) using the instrumental variable quantile regression (IV-QR) model due to Chernozhukov and Hansen (2005, 2006, 2008). The IV-QR model allows us to incorporate variation in the elasticity due to the presence of unobserved differences across MSAs. Moreover, it allows us to evaluate the impact of changes in the stock of interstate highways on the entire conditional distribution of VMT, not just the impact on the conditional mean as in Duranton and Turner (2011). The IVQR estimates show that as predicted by the simple partial equilibrium model, the elasticity declines as one goes up the quantile ladder, being more than one at the lower quantiles and less than one at the higher quantiles. The median IV-QR estimate being close to one. The IV-QR model implies that among observationally identical cities, expanding road capacity can lower the number of cities experiencing severe congestion, although the mean or median congestion levels are likely to remain constant. We also estimate the impact of increased road capacity on the unconditional distribution of VMT using the generalized quantile regression (GQR) model due to Powell (Forthcoming). The GQR estimates mirror the IV-QR estimates, but their conclusions are starker at the upper quantiles: building highways have no statistically significant impact on VMT at the highest quantiles. The GQR results imply that building roads can lower the total number of cities having different observed characteristics experiencing severe congestion levels. We further explore the mechanisms that drive the empirical findings by running simulations using a spatial general equilibrium model with an extensive road network calibrated to the Greater Los Angeles (LA) Region. We find that the elasticity of VMT to capacity in LA is 0.321, and the elasticity decreases consistently with the initial congestion level.
Does Mortgage Regulation Affect the Supply and Demand for Alternative Home Financing? (G2, G5)
We study the impact of the LTV regulation on supply and demand for unregulated credit that is used for home acquisition. A combination of including only personal mortgages in the numerator of the LTV ratio, and a large variation in pre-existing unregulated building-level mortgage, provides an exogenous variation in unregulated credit amount available for home financing. We find that to achieve 1 dollar reduction in equity downpayment, a household is willing to pay about 1.8 dollars more in unregulated credit after the LTV cap is introduced. This increase is not substituted by a decrease in other consumer loans, and is rather complemented by a further increase in the latter in localities with more constrained homebuyers. We observe absence of our findings in Norway, where regulation was applied to all mortgages. Our results have direct implications for the design and types of credit used in borrower-based macroprudential instruments.
Does Ownership Change Intensify Competition? Evidence from the Texas Lodging Industry (L2, D2)
This study investigates how ownership changes increase/decrease competition in the Texas lodging industry. Firms with different ownership structures have been found to show different behaviors, intensifying competition in local markets. Ownership changes related to the ability to influence the intensity of competition may be the essential determinant of market structures. However, none of the existing studies analyzes the relationship. Therefore, this research will aim to provide insights to fill the gap in the entry and exit literature.
Does Paid Family Leave Save Infant Lives? Evidence from California (I1, I3)
One goal of the paid family leave program in the U.S. is to help working parents balance their careers and family responsibilities and hence improve the well-being of their infants. A large body of literature evaluates the effects of California’s Paid Family Leave program (CA-PFL) on early childhood outcomes, but most studies have been based on the analyses of surviving infants. If the CA-PFL reduces infant deaths, then such analyses would understate the program’s true effects. Using the linked birth and infant death data in the U.S. with a difference-in-differences framework, I find that the implementation of the CA-PFL reduced the post-neonatal mortality rate by 0.135 (per 1,000 live births), or it saved approximately 339 infant lives in California from 2004 to 2008. The effects were driven by death from internal causes and there were larger effects for boys than girls. These results are stable across a variety of robustness checks and no evidence suggests that these estimates result from the endogeneity of policy, simultaneously shocks, and changes in fertility.
Does Retirement Increase Stock Market Participation? Evidence from a Fuzzy Regression Discontinuity Design (G5, D1)
Existing studies suggests that the lower information cost facilitates the stock market participation. In this paper, we examine whether retirement, a direct way to lower information cost by relaxing time constraints, increases the stock market participation. We use U.S. Health Retirement Study (HRS) survey data and exploit the exogenous variation in social security benefit rule that shifts individuals’ retirement likelihood who are above the early entitlement age. We find that individuals above the early entitlement age 62 are more likely to retire than those below age 62 for low and median wealth group but not high wealth group. In addition, we find that that it does increase the stock market participation for the median wealth group but not the low wealth group. We further verify that the increased stock market participation due to retirement is through lower information cost by relax time constraint.
Does Test-Based Teacher Recruitment Work in the Developing World? Experimental Evidence from Ecuador (I2, J4)
Since 2007, the Ecuadorian government has required teacher candidates to pass national skill and content knowledge tests before they can participate in merit-based selection competitions for tenured positions at public schools so as to raise teacher quality. We evaluate the impact of this policy by linking administrative teacher records to data from a unique experimental study where almost 15.000 kindergarten children were randomly assigned to their teachers in the 2012-2013 school year in Ecuador. For our estimation strategy, we prove a successful random assignment of these children to teachers tenured through the new recruitment policy, which allows us to overcome the potential bias caused by the matching of teachers to students. Our results show that kindergarten students randomly assigned to teachers who passed mandatory entry test and won merit-based competitions, have significantly higher end-of-year test scores of at least a 0.105 standard deviation in language and a 0.085 standard deviation in math. These effects persist even after controlling for teacher education, experience, cognitive ability, personality traits and classroom practices.
Dollar borrowing by non-financial firms and the real effects of US monetary policy abroad (F3, E5)
I provide firm-level estimates of the real effects of US monetary policy on investment in 36 countries. The key identification idea is that firms, that roll over US Dollar debt shortly after FOMC meetings, are more exposed than firms that do not. Reductions in business investment after US monetary tightening are largest in countries with pegged or managed exchange rates (non-floaters) but also significant in floaters. The stronger spillovers to investment in non-floaters arise from a relatively stronger response by firms with high leverage. Exchange rate fluctuations contribute to the spillover heterogeneity because they amplify the financing spillover channel for non-floaters but dampen it in floaters. A simple framework of endogenous currency choice rationalizes my findings. Exchange rate pegs lead to higher financial vulnerability because they allow smaller and less productive firms to borrow in foreign currency, a conjecture which I confirm in the data.
Effects of External Assumptions on Forecast Errors in the Euro Area (C5, E3)
This paper analyzes the role of professional forecaster's misconceptions about ex-ante technical assumptions in explaining the heterogeneity of predictions for main economic aggregates and differences across individual forecast errors. Evaluating the panel provided by the European Central Bank's Survey of Professional Forecasters for the euro area, we test to which extent disagreement for key macroeconomic indicators - GDP growth, inflation and unemployment rate - can be explained by the heterogeneity in assumptions for the oil price, the exchange rate and the interest rate. We document that oil price disagreement matters in particular for inflation disagreement, while interest rate disagreement is closely associated with the dispersion of output growth and unemployment rate forecasts. Building upon this finding, we investigate whether individual forecast errors are related to individual assumption errors. In line with the evidence for disagreement, oil price errors are most closely related to inflation errors, whereas interest rate errors matter more for forecast errors for GDP growth and the unemployment rate. Our results indicate that survey participants can improve the accuracy of their predictions by reducing assumption errors. We contribute to the literature that seeks to better understand the expectation formation process of economic agents.
Election Day Shocks and Political Preferences (D7, D0)
This paper examines whether and how the suspension of elections in the Dominican Republic affected voter turnout, political preferences, and attitudes toward democracy. Through the use of polling data leading up to the suspension, we approximate the counterfactual scenario of expected turnout and likely results had the elections taken place without interruption. We rely on results obtained from the special elections celebrated a month after to examine the impact of said rare event. Though we find generalized preferences favoring the opposition party remain unchanged, the suspension resulted in a significant improvement of opposition performance. Social media data and newspaper coverage provide evidence of increased levels of collective action following the failed election.
Employment Effects of Climate Policy and Clean Energy: Evidence from China (Q0, Q4)
Climate change affects the global energy structure. China's climate policy has greatly promoted the development of the clean energy sector, which has had an uncertain impact on the economy and the labor market. From 2008 to 2019, China ’s share of clean energy consumption increased from 11.8% to 23.4%. At the same time, workers in traditional energy sectors were unemployed, and the supply of high-tech labors in clean energy were insufficient. First, we summarize the relationship between climate policy, clean energy sector development and employment, and study the climate and energy policies of national and local governments separately, depict China's energy development framework dealing with climate change. In addition, based on data from 2008 to 2018 in China, we use the dynamic panel data model to test the impact of clean energy on employment.
We found that the development of clean energy is mainly driven by the top-level design of the central government, local government implementation of supervision, and market mechanisms. Increasing clean energy production, applying environmentally friendly energy technologies, and developing green transportation are the three main ways in which clean energy responds to climate change. According to empirical research results, the clean energy still has a significant positive impact on employment. Meanwhile, clean energy impacts labor market by the employment creation or employment destruction effects. On the one hand, The clean energy sector has developed many new types of jobs, driving up the employment rate. However, it has partially replaced, on the other hand, the employment of traditional energy sectors, causing involuntary unemployment. Due to different regional economic foundations and industrial structures, the employment effects of clean energy have regional differences. Industrial policies to address climate change need to consider employment rates and employment equity.
Employment Protection and Firm Provided Training in Dual Labour Markets (J4, M5)
In this paper we leverage a labour market reform (Fornero Law) which reduced firing restrictions for open-ended contracts in the case of firms with more than 15 employees in Italy. The results from a Difference in Regression Discontinuities design demonstrate that after the reform, the number of trained workers increased in firms just above the threshold by approximately 1.5 additional workers. We show that this effect can be explained by the reduction in worker turnover and a higher use of permanent contracts. Our study highlights the potentially adverse effects of EPL on training in dual labour markets.
Energy Efficient Technology and Vintage Capital in Chinese Industry (Q4, E2)
By incorporating energy-saving through technology-embodied investment, disembodied and embodied technical changes into a dynamic stochastic general equilibrium (DSGE) model with heterogeneous investment, this paper identifies avenues through which firms adjust to rising energy prices. Using Chinese firm-level data from 1997-2004, we estimate a set of stylized facts regarding how firms of various ownership types respond to energy price changes. We then use these stylized facts to recover the key parameters in the DSGE model through indirect inference. The results show that within Chinese industry, in response to rising energy prices, state-owned enterprises, domestic non-state enterprises, and foreign-funded enterprises employ significantly different means to achieve their energy efficiency. Such differences can be substantially explained by government policy affecting energy pricing and the cost of investment finance across firms of different ownership types.
Environmental Change and State Capacity: Two Millennium of Data from China (N5, O1)
How does state capacity and society in general respond to changes in natural environment? The Yellow
River gave birth to the Chinese civilization, but it also has a violent nature and brought wide‐spread
misery to the lower Yellow River region. Most famously, the Yellow River changed its lower course at
least six times in the past two millenniums. Utilizing the randomness of river course shifts, this paper
examines the long‐run economic and political impact of environmental change. Evidence shows that
being along the Yellow River significantly decreased the population density in a region; but it rebounded
quickly to its pre‐existing level within a century after the river shifted away. The impact on state
capacity, however, followed a different dynamic. In the short run, state capacity expanded in the
affected region in response to greater need of collaborated water management, but shrank in the long
run and demonstrated certain level of history‐dependence long after the river shifted away.
Environmental Quality Perception, Reality, and Policy: Evidence from an Air Quality Mobile Device Application (Q5)
This study investigates how perceived air quality affecting human behaviors and preferences for associated policies. I developed a mobile device application to collect perceived air quality data, which can be contrasted with real-time scientific measures, as well as location and phone usage data from a year-long field experiment involving 10,275 individuals. The results first show that income is negatively associated with perceived air quality. I find that more engagement with the air quality information provided through the app significantly decreases the discrepancy between objective and subjective air quality, and such effect increases with income but decreases with age. More accurate assessment of the air quality further leads to more defensive behavior, such as reducing time spent outdoor when the air pollution level is unhealthy for sensitive groups or worse. Preference for using nuclear energy, which has been widely campaigned as one of the primary short-term solutions for air quality degradation, were surveyed before respondents installing the application and one year after the installation. I find that better assessment of the air quality decreases the likelihood of supporting nuclear energy. In 2018, a nationwide referendum was approved to repeal the planned end of nuclear power stations in the study region. This result highlights the problem that misperceived environmental quality may lead to suboptimal public policy decisions.
Estimated Policy Rules for Capital Controls (F3, E5)
This paper borrows the tradition of estimating policy reaction functions from monetary policy literature to ask whether capital controls respond to macroprudential or mercantilist motivations. I explore this question using a novel, weekly dataset on capital control actions in 21 emerging economies from 2001 to 2015. I introduce a new proxy for mercantilist motivations: the weighted appreciation of an emerging-market currency against its top five trade competitors. This proxy Granger causes future net initiations of non-tariff barriers in most countries. Emerging-markets systematically respond to both mercantilist and macroprudential motivations. Policy-makers respond to trade competitiveness concerns by using both instruments—inflow tightening and outflow easing. They use only inflow tightening in response to macroprudential concerns. Policy is acyclical to foreign debt; however, high levels of this debt reduces countercyclicality to mercantilist concerns. Higher exchange rate pass-through to export prices, and having an inflation targeting regime with non-freely floating exchange rates, increase responsiveness to mercantilist concerns.
Estimating the Armington Elasticity: The Importance of Data Choice and Publication Bias (D1, F1)
A key parameter in international economics is the elasticity of substitution between domestic and foreign goods, also called the Armington elasticity. Yet estimates vary widely. We collect 3,524 reported estimates of the elasticity, construct 34 variables that reflect the context in which researchers obtain their estimates, and examine what drives the heterogeneity in results. To account for inherent model uncertainty, we employ Bayesian and frequentist model averaging. We present the first application of newly developed non-linear techniques to correct for publication bias. Our main results are threefold. First, there is publication bias against small and statistically insignificant elasticities. Second, differences in results are best explained by differences in data: aggregation, frequency, size, dimension. Third, the mean elasticity implied by the literature after correcting for both publication bias and potential misspecifications is 3.
Estimating the Effect of Conflict on Agricultural Activity in the Central African Republic with Remotely Sensed Data (Q1, Q3)
Ongoing armed conflict poses major challenges to the economic recovery and development of the Central African Republic. Although the conflict occurs in agricultural areas, little evidence exists on the magnitude of the effect of conflict on agricultural activity and the channels. We construct a geo-referenced monthly panel (2000-2018) during which numerous conflict events occurred. Exploiting the prevalent practice of burning fields as a measure of active land under preparation for cultivation, we find that the presence of a conflict event 12 months prior lowers fire presence during land preparation by 9% and lowers biomass by 8% during the sowing/growing season - particularly in areas where maize, millet and cassava are highly suitable. While the nature of this effect differs across crops that are driven largely by differences in production processes particularly during land preparation, these decreases are suggestive of the abandoning of farm lands due to conflict.
Estimating the Impact of Weather on CBOT Corn Futures Prices using Machine Learning (C6, Q1)
Weather is one of the most important factor affecting corn production in the US. This component is also very unpredictable and sometimes can be significantly different from its historical pattern. A favorable weather over the crop growing season induces high production and similarly, a bad weather brings down the production. The prevailing weather condition in the season causes a certain level of production expectation in that year. That expectation influences futures prices in the market. In short, weather condition affects corn futures prices. This paper applies machine learning (ML) methods on weather and soil data to predict corn futures prices at Chicago Board of Trade (CBOT). The input variables, which include county level hourly weather data since 1980 and county level soil data of all corn producing states in the US, have spatial as well as temporal dimensions.
There are several variables but not all are important and can cause unnecessary noise in the model. Principal Component Analysis (PCA) is used for feature selection (identifying relevant variables). Fully Connected Neural Network (FCNN) is employed on the selected inputs to forecast daily percentage change in futures prices. Our findings suggest that more specialized and focused training of the data incorporating spatial attributes using Convolutional Neural Network (CNN) delivers better result. The study demonstrates that price change is more responsive to the unanticipated part of weather variables i.e. weather surprise. Weather surprise is defined as the difference between the current weather variable and the preceding four years average of that variable for that month. Finally, we utilize Long Short-Term Memory (LSTM) along with CNN to additionally include the temporal characteristics of the input data. The model helps traders making informed decisions in the market to earn huge profits.
Excessive Firm Turnover in the Shadow of Unemployment (E2, E6)
This paper studies how pecuniary self-employment affects business cycle dynamics, macroeconomic efficiency, and the outcomes of structural reforms. I employ a two-sector dynamic general equilibrium model with endogenous producer entry. One sector (the "hiring sector") is populated by monopolistically competitive firms that employ workers subject to search-and-matching frictions in order to produce output. The other sector consists of self-employment firms that use the output of the first sector as input. Self-employment is introduced as a possible occupational choice for the unemployed. This feature relates firm creation more directly to the state of the labor market and to workers’ opportunity costs. Consistent with the U.S. data, the model shows that self-employment represents 7.4% of employment and is procyclical. The procyclicality of self-employment arises because positive productivity shocks in the hiring sector cause profits for the self-employed to rise strongly enough that additional unemployed workers are drawn into self-employment, despite tighter labor market conditions and a competing incentive to seek traditional employment. This dispels the common misconception that all labor market rigidities increase self-employment; on the contrary, economies with no significant unemployment benefits or a very weak bargaining power could still show high rates of self-employment. Novel sources of inefficiency exist since neither workers nor firms internalize the consequences of self-employment. As a result, the number of firms is more volatile and welfare costs of business cycles are higher in the presence of self-employment dynamics. Furthermore, I show that reforms facilitating entry are more effective when the self-employed are relatively less productive or have greater monopoly power.
Exploring the Impact of Economic Integration Agreements through Extreme Bounds Analysis (F1, C5)
We provide an empirical strategy guided by the data to estimate the dynamics and effects of Economic Integration Agreements (EIAs) on trade flows. The strategy uses Extreme Bounds Analysis (EBA) to guide the choice of lags and leads in the effects without researchers' discretion involved. We show that arbitrarily selected year intervals and starting year can result in non-robust estimates of transitional dynamics of the effects of EIAs on trade flows. The empirical strategy follows two steps: EBA firstly sifts lags and leads of EIAs robustly related to trade flows from candidates, then these are included in the gravity equation to estimate the effects of EIAs on trade. We find that various lags and leads are robustly and positively related to trade flows, and the lag and lead structure depends on the level of integration. Our results show that EIAs have a long-term effect of 64% on trade flows. Under the richer lag and lead structure, deep-integration agreements beyond the level of free trade agreements have a much higher impact on trade flows than free trade agreements do (112% versus 33%). The estimates of effects of EIAs obtained from EBA-based estimation have a smaller contemporaneous effect and larger phased-in effects compared to previous studies relying on the subjective choices of year intervals while similar results are observed with the decomposed EIAs.
Fading Legacies: Human Capital in the Aftermath of the Partitions of Poland (N3, I2)
This paper studies the role of institutions for the longevity of historical legacies in human capital. The Partitions of Poland (1772-1918) represent a large-scale natural experiment that subjected Poland to three different sets of educational institutions. Poland’s independence after WWI, in turn, rapidly harmonized the institutions of education. To study the evolution of human capital under these different institutional regimes, I construct a large, unique dataset on schooling and education in the Polish territories from 1911 to 1961. Using a spatial RD design, I find substantial differences in primary enrollment between the partitions prior to WWI. However, within two decades of Polish independence, enrollment becomes universal in all former partitions, with a particular strong growth in female access to schooling. This is accompanied by a high intergenerational mobility in education that equalizes literacy and educational attainment between the former partitions. Educational institutions hence drive both substantial divergence and convergence in human capital.
Family Time Allocations over the Last Half Century (D1, J1)
Over the last half century, married American women doubled their labor supply and halved the time they spend on household chores. By contrast, married American men reduced their labor supply and increased their involvement in chores. Both increased the time they devote to children but men, especially the most educated, did so by far more. What explains these dramatic changes in family time allocations? I develop a dynamic collective model that incorporates family time use and features financial and human capital, rich earnings dynamics, and endogenous divorce. The model quantifies the role of (i) the wage and education gender gaps; (ii) fertility and the cost to raise children; (iii) labor market experience; (iv) technical change in the household; and (v) features of the marriage market such as the expansion of divorce in the early 1970s. Preliminary results suggest that -at least- the gender wage gap and the divorce option are important for the observed time allocation dynamics. The latter implied an improvement in intra-household bargaining power among women born in the 1950s and a transfer of non-market work to their husbands.
Fed Tails: FOMC Announcements and Stock Market Uncertainty (E5, G1)
Uncertainty around FOMC announcements builds up days ahead of the meeting and fully resolves once the policy decision is announced. Disentangling tail uncertainty shows that the perception of bad economic states is the primary driver of this pattern, albeit policy operations are meant to be stabilizing. Investors are afraid of the revelation of bad states and are willing to pay a hedging premium of approx. 9% per meeting. FOMC announcements are special as uncertainty around other macroeconomic news releases is not driven by tail uncertainty. Not only does tail uncertainty predict pre-announcement stock market returns but also changes in the fed fund target rate for horizons up to one year. Our results indicate that policy makers closely monitor downside uncertainty and use this information as part of their decision-making process.
Fetal Origins of Covid-19 Mortality. Evidence from Peru (I1, O1)
In this study, we exploit the Cholera Epidemic in Peru in the early 1990s as a quasi-natural experiment to explore whether prenatal circumstances increase the risk of Covid-19 mortality. We find that a one-standard deviation increase in the incidence of Cholera during the first trimester in-utero increases the likelihood of working-age women to die of Covid-19 by 21 percent. As potential mediators we find a significant effect on BMI, obesity rates and high blood pressure, as well as on self-employment. In-utero infection with Cholera can result in nutritional deprivation, moreover, the epidemic represented an income and stress shock for many mothers, hence some, or a combination, of these factors could have prompted the results.
Firm Characteristics and the Gender Wage Gap: Evidence from the China Employer Employee Survey (J3, J7)
This article analyzes a novel employer-employee survey of manufacturing firms in China to
present new findings on the gender wage gap in China. The unexplained gender wage gap remains at 17 log points after adding a rich set of human capital, job, and firm characteristics. Ownership sector and firm characteristics related to market power play an important role in explaining gender differences in wage
determination. The gender wage gap varies across ownership sector, with the largest gaps in the Domestic Private Sector, and the weakest gaps in the Collective-owned sector. Rent elasticity estimates show that gender gaps are largest among the lowest-educated, but disappear among the highly educated and within high-skill firms. Variables related to female labor supply are significant in explaining both gender wage gaps and gender promotion gaps.
Firm Dynamics and Economic Development with Corruption and Financial Frictions (O1, O4)
We build a firm dynamics model with corruption to study its impact on firm entry and exit, capital accumulation, and innovation. The effect of corruption depends on the degree of financial frictions and the stage of economic development. In the model, corruption serves as an endogenous entry barrier that reduces firm churning and protects the incumbent firms, allowing them to accumulate capital more quickly and grow out of financial constraints. Corruption can therefore have a positive effect when economic growth relies mainly on capital accumulation. However, as the economy develops, corruption can lead to increasing productivity losses when capital becomes abundant and technological progress is the main driver of growth. In addition, more corruption at the early stage could lead to a highly skewed distribution of firms later on, making it easier for asset-rich incumbent firms to bribe the government officials and pre- vent successful innovators from entering the market. We test the predictions of our theory using the Chinese firm-level data from 1998 to 2007. Our theory also has implications for the optimal anti-corruption policy over the development process.
Firm Subsidies and Resource Misallocation (E6, E2)
Governments in developing and advanced economies subsidize firms in a discretionary fashion. Do such policies mitigate or exacerbate the misallocation of resources across firms? I analyze a typical EU policy using novel data on applicants and recipients of capital subsidies in Greek manufacturing. In my framework, firms face existing distortions that subsidies can correct or exacerbate. The actual policy exacerbates misallocation, decreasing aggregate total factor productivity (TFP) by 0.15%. The policy's potential effects are large as reallocating subsidies among firms can increase TFP by 2% or decrease it by 3%. The actual policy's effect is small because firms facing high distortions are as likely to receive a subsidy as those facing low distortions.
Fiscal Tools to Reduce Transition Costs of Climate Change Mitigation (Q5, H2)
How much the transition out of greenhouse gas emissions will cost to the economy? Current available estimates differ widely. This reflects different methodologies and assumptions adopted by different studies, combined with the inherent uncertainty related to forecasting future greenhouse gas emissions and temperature increases. This paper takes a different approach. It assumes as given widely-used scenarios for future carbon emissions and temperature increases (as the Paris agreement paths for example), and then backs out the combinations of fiscal tools (especially carbon price measures and tax incentives for green investments) that can minimize the transition costs. To this end the paper extends the work done in Catalano et al. (2019) by using a global overlapping generations model in the spirit of Kotlikoff et al. (2019) combined with a climate module. The results show that the estimated economic costs of the carbon transition vary significantly depending on the fiscal tools used to reduce greenhouse gas emissions.
Flirting with Disasters: Do Firms Financially Plan Ahead for Disasters? (G3, Q5)
There are two types of disasters: natural (Acts of God) and technological (human-caused) disasters. I investigate whether and under which conditions firms are hoarding precautionary cash holdings to address natural disaster risk, technological disaster risk or both.
The empirical analysis requires me to introduce a novel multidimensional risk measure for each type of disaster as early warning sign for possible future disaster strikes. Using these measures, I provide evidence that firms do not trade-off between these two types of disasters in determining their cash policy. Firms prioritize the preparedness of possible natural disaster strikes above possible technological accidents. The natural disaster related precautionary cash holdings hoarding policy is a long-term policy option that is funded by using external financing and focused on a few disaster types such as wildfires and landslides. Firms address only technological disaster risk by precautionary hoarding cash holdings when they are less internal financially constrained or in smaller countries by surface area.
Foreign Direct Investment Commitments in East Asia (F2, G3)
Do firms carry out planned investments? As firms rarely systematically disclose data on planned activities, we explore this question using country-level data on approved and utilized bilateral foreign direct investment (FDI) from high-income countries to four middle-income countries in East Asia. The ratio of utilized FDI to approved FDI, or the commitment ratio, captures the degree to which planned investments are implemented. We find that the commitment ratio is higher in China and Indonesia than in the Philippines and Thailand. The commitment ratio is higher when the host country has greater financial openness, larger population, and a more stable government. It is also higher when the bilateral exchange rate is more volatile. However, the commitment ratio is not affected by trade or investment agreements, past trade or investment experiences, political shocks, or cultural frictions.
Gender and Social Networks on Bank Boards (G2, J1)
We examine the effect of the social networks of bank directors on board gender diversity and compensation using a unique, newly compiled dataset over the 1999-2018 period. We find that within-board social networks are extensive, but there are significant differences in the size and gender composition of social networks of male vs female bank directors. We also find that same-gender networks play an important role in determining the gender composition of bank boards. Finally, we show that those connected to male directors receive higher compensation, but we find no evidence that connections to female directors are influential in determining pay and bonuses.
Gender Gaps in Political Careers: Evidence from Competitive Elections (J1, D7)
This paper investigates the impact of voter support on the representation of women in the political profession. The empirical analysis exploits two-stage elections in the United States and Italy to hold the selection of candidates constant. In two-stage elections, candidates are admitted to the second round of voting based on the outcome of the first round. I find that among candidates who marginally qualify for the final round, women are 20 percent less likely than men to be elected to the US House of Representatives and 40 percent less likely to be elected mayor in Italian municipalities. Using a difference-in-discontinuities design, I then show that the gender gap in the probability of being elected has long-lasting effects on career trajectories. Women are substantially less likely than men to win future elections and to climb the political hierarchy. My findings suggest that one of the reasons that few women reach the top in politics is that female candidates face hurdles at the beginning of their careers.
Gender Norms and Labor-Supply Expectations: Experimental Evidence from Adolescents (J2, J1)
Gender gaps in labor-market outcomes often exacerbate with the arrival of the first child.
We investigate how highlighting existing gender norms affects labor-supply expectations in
a sample of 2,000 German adolescents. At baseline, the majority of girls expects to work 20
hours or less per week when having a young child, and expects their partners to work 30
hours or more. We implement randomized treatments that (i) increase the salience of the
existing traditional norm towards mothers, and (ii) correct misperceptions about the norm’s
content. The treatments significantly reduce girls’ self-expected labor supply and increase
the expected within-family gender gap. In a second experiment, we randomly highlight
another, more gender-egalitarian, norm towards shared household responsibilities and show
that this attenuates the expected within-family gender gap.
Gender Wage Gaps in STEM Disciplines (J7, J4)
This study examines the academic gender wage gap in STEM and non-STEM disciplines at a public research university. The data used in this study are particularly useful because faculty are observed at four academic years (2008-09, 2010-11, 2015-16, 2019-20), allowing the construction of a panel data set which permits control of omitted productivity differences among faculty members.
Estimating earnings regressions for female and male faculty members, we compare the representation of women in STEM departments using two alternative definitions of STEM fields, one more inclusive. Controlling for years of experience, research productivity, and field salary differentials, we perform Oaxaca decompositions of the mean male-female wage gaps in STEM and non-STEM departments (Oaxaca and Ransom, 2002). We also conduct Mata and Machado quantile decompositions of gender gaps at the five quantiles to estimate potential discrimination over the wage distribution (Mata and Machado, 2005).
Our findings indicate that women are significantly underrepresented in STEM departments, but that comparable to national data, the underrepresentation has decreased over time. Our analyses indicate that a statistically significant gender gap in monthly salary in STEM departments that has decreased over time but has not disappeared. Further, the gender gap in STEM departments is significantly larger than that observed for non-STEM departments. Finally, despite finding positive effects of for women at the top end of the salary discrimination, we find there is potential discrimination at the low end of the salary distribution among faculty members working in STEM departments. This suggests that highly productive female academics working in STEM departments are well rewarded by the competitive market for academics, but female academics are apparently not paid on par with their White male peers at the lower end of the salary distribution.
Global Food Waste across the Income Spectrum: Implications for Food Prices, Production and Resource Use (Q1, Q3)
There are few examples in the existing literature that address the quantitative linkages between food waste, food security, and environmental sustainability, on a global scale. Here we develop a new panel database on household food waste at the national level based on the Energy Balance equation, including adjustments for changes in body weight over time. We use this to characterize the non-linear relationship between per capita income and the share of food availability wasted. By incorporating this relationship into a global partial equilibrium model of the agricultural sector (SIMPLE), we develop future trajectories of household food waste. We find that the emerging economies, particularly China and South Asia, are likely to play a key role in determining global food waste at mid-century. We also present several counterfactual scenarios that shed light on the implications for environmental and food security of limiting future growth in food waste. We find that the global impacts of these alternative pathways are greatly enhanced in the context of a more open international trade regime.
Government Debt, Dividend Growth, and Stock Returns (G1, G0)
This paper documents that the increase in public debt can lead to higher dividend payout to shareholders. It suggests that public debt can be a strong cash flow predictor which can help us to better predict future stock returns. Specifically, the higher public debt-to-GDP ratio can predict both higher dividend growth and higher stock returns, and the predicted changes are in the same magnitudes. The finding is consistent with Lettau and Ludvigson's (2005) argument that there exists a common component among stock returns and dividend growth. We argue that i) the existence of a common component can resolve the US asset pricing puzzle as emphasized by Cochrane (2007, 2011) that the dividend-price ratio can only predict discount rates but not cash flows; ii) the public debt can drive the co-movement among returns and dividend growth and capture this common component; iii) the strong cash flow predictability of the public debt-to-GDP ratio can not be consumed by the popular consumption-to-wealth ratio (cay) and many other macroeconomic states variables; iv) future stocks returns can be better out-of-sample predicted after controlling for public debt. The empirical evidence documented in the US aggregate market can also be extended to the US cross-section and the international markets, especially for the advanced financial markets, which help to explain the weak cash flow predictability recently documented by Rangvid et al. (2014) and Maio et al. (2015). To rationalize the finding, we propose a production-based asset pricing model incorporating cash-retention friction on the corporate sector. The model can produce testable predictions that the increase in public debt moves both dividend payment and the cost of capital in the same direction, resulting in the capture of the common component.
Grades, gender and early-career returns to academic performance (J7, I2)
Grade Point Average (GPA) is one of only a few costly signals that a university graduate can send to employers to distinguish themselves from other graduates. This study examined the extent to which the association between academic achievement and early-career incomes are different for male and female undergraduate degree recipients.
Undergraduate alumni records for a large state university system were matched to the state’s unemployment insurance database from 2004-2014 to determine wage data and industry of employment. Records were also matched to National Student Clearinghouse data that indicate alumni enrollment in a graduate program. I examine several specifications of income for graduates 7-12 months after graduation, when degree attainment and academic achievement is most likely to be a primary signal to employers, and estimate the returns to academic achievement in several industries.
Findings indicate that, on average, GPA is a meaningful signal to employers and graduates with a higher GPA earn an income premium, though there is significant heterogeneity across industries. Industries with higher average incomes also tend to have higher returns to GPA. I do not find evidence that men and women see differential returns to academic achievement within industries.
Grandfathers and Grandsons: Social Security Expansion and Child Health in China (H5, H2)
We examine the multigenerational impacts of a nationwide social pension program in China. The New Rural Pension Scheme (NRPS) was rolled out county-by-county since 2009 and enrollees age over 60 in rural areas are eligible to receive at least 70 CNY non-contributory monthly pension. We take ad- vantage of a panel data and use household age eligibility change as instrument for household pension receipt to identify the impacts of NRPS. Our findings indicate a substantial positive significant effect of the presence of pensioners on grandchildren’s short-term nutritional status measured by BMI-for-age z socres, while the impacts on their long-term nutrition status measured by height-for- age z scores are insignificant. We find a novel gender pattern that the impact are likely to be driven by boys subsample and only grandfather pensioner’s im- pacts are significant. Children living with NRPS eligible grandfathers are more likely to be overweight/obese. We examine the mechanism and find NRPS sub- stantially increases household income. Grandparents’ time allocation on child care might change moderately after being eligible for NRPS.
Green Asset Pricing (G1, Q5)
Climate change is one of the biggest economic challenges of our time. Given the scale of the problem, the question of whether a carbon tax should be introduced is hotly debated in policy circles. This paper studies the optimal design of a carbon tax when environmental factors, such as air carbon dioxide emissions (CO2), directly affect agents' marginal utility of consumption. Our first result is that the optimal tax is determined by the shadow price of CO2 emissions. We then use asset pricing theory to estimate this implicit price in the data and find that the optimal tax is pro-cyclical. It is therefore optimal to use the carbon tax to \cool down" the economy during periods of booms and to stimulate it in recessions. The optimal policy not only generates large welfare gains, it also reduces risk premiums and raises the average risk-free real rate. The effect of the tax on asset prices and welfare critically depends on the emission abatement technology.
Green Investments: Institutional Investors and Illiquid Assets (G0, Q2)
In the context of green investments through stock markets, the aim of this paper is to analyze to what extent illiquidity can affect delegated portfolio choice, and hence allocations in low-carbon assets. In this paper, we study optimal asset allocation of an open-end equity fund manager, who invests in a benchmark index and in an illiquid alternative risky asset. Illiquidity of the alternative asset is reflected through proportional transaction and execution costs incurred during trading operations. On the other hand, trading on the benchmark incurs no transaction costs. In our paper, the benchmark index and the alternative asset represents respectively brown and green assets. Fund flows to relative past performance are assumed to be positive and convex; investors reward manager’s positive relative past performance with substantial inflows, while redemptions resulting from poor past performance are not so harsh. Such reward structure implicitly incentivize fund managers to take more risks in order to beat the benchmark portfolio. Our model is an extension of Merton's (1971) continuous time model, where we consider a portfolio manager whose objective is maximize her expected value of her constant relative risk aversion (CRRA) utility of the terminal value of her fund’s assets under management. Preliminary results show that under transaction costs, a manager with high relative risk aversion, who is mostly concerned with her relative performance, will tend to allocate her assets close to those of the benchmark portfolio. Hence, due to the nature of green assets which can be less liquid than the already established brown assets, institutional investors concerned with maximizing shareholders’ value might not be fully inclined to invest in low-carbon capital.
Groupthink: An Experimental Study of Group Decision-Making (C9, D9)
Do we make better decisions in a group? Does group decision-making suffer from groupthink? Traditionally, the psychology literature tends to focus on within-group communication rather than the group itself; the economics literature emphasizes incentives and strategic interactions such as free riding and coordination. In this paper, I conduct an innovative laboratory experiment to investigate the causal effect of the group itself compared with the solo situation and identify the group effect free from free-riding incentives. First, subjects self-evaluate their ability in the solo setting and the ability of their group in the group setting. Second, they make a series of investment decisions according to their beliefs about their abilities, as the expected returns to investment positively correlate with the ability of the subject (in the solo setting) and with the ability of the group (in the group setting). More importantly, one’s own expected ability coincides the expected ability of one’s group by my experimental design, making solo and group settings directly comparable. The experiments provide strong evidence that group settings induce overinvestment. When subjects are in groups, they 1) reveal a stronger belief in their own ability in making profitable investment decisions by 36 percent, and 2) choose to invest more often (by 12 to 17 percent), lowering their final payoff. The group effect is significantly stronger for those who are randomly assigned to the group setting prior to the solo setting. Strikingly, the group effect without communication is as strong as in treatments with communication. I argue that when self-evaluation is involved, the group itself has a negative impact. Unlike what is suggested by groupthink theory, the negative group effect can occur even in the absence of within-group communication.
Happiness Makes Workers More Productive: Evidence from Large-Scaled Experiments (J2, D9)
There is an increasing interest among enterprises in investing in the happiness of their employees. However, the empirical evidence of the causal relationship between happiness and productivity is limited (Oswald et al., 2015; Bellet et al., 2020). Therefore, we conducted two different styles of large-scaled experiments which exogenously provide the variation in the level of happiness among employees of enterprises and civil servants in Japan (n=6,201) to test the causal relationship.
The first experiment is a Randomized Controlled Trial (RCT) showing a comedy clip to the treatment group while showing a control clip of moving shapes to the control group to test if the raised happiness induced by the comedy clip makes participants more productive. The second experiment exploits exogenous real-life negative shocks on happiness (death or serious illness of a partner) within a past year to test if the lowered happiness caused by the shocks lowers productivity. The productivity of each participant is measured by the number of correct answers of timed mathematical additions that participants solve for monetary incentives after watching a comedy/control clip.
As to the first experiment, the intervention of the comedy clip to raise happiness was overall not successful but successful only for those who live in Tokyo probably because of regional differences in the sense of humor. Therefore, we further study participants from Tokyo and find that those who are in the treatment group solve more additions correctly than those who are in the control group. As a result of the second experiment, we find that those who experienced the negative shocks on happiness within the past year show lower happiness levels as well as lower productivity compared to those who didn’t. Both experiment results support the causal relationship of happiness raising the productivity of workers.
Hate Is Too Great a Burden to Bear – The Effect of Hate Crime on Refugees’ Mental Health (I1, J6)
Against the background of increasing violence against immigrants and refugees, we estimate the effect of hate crime on refugees' mental health in Germany. For this purpose, we combine two innovative data-sets: administrative records on xenophobic crime against refugee shelters by the Federal Criminal Office and the IAB-BAMF-SOEP Survey of Refugees. Considering that refugees may not influence the timing of an attack nor the interview date, we use a regression discontinuity design to identify a local causal effect. Our results show that hate crime has a substantial negative effect on several mental health indicators. Further, we find suggestive evidence that country-specific human capital is an important mediator in our analysis, which emphasizes that our results have important policy implications.
Healthcare Consumption Disparities: New Evidence from the ACA Preventive Care Provision (D9, I1)
This study investigates whether a key provision of the 2010 U.S. Affordable Care Act (ACA) mandating private health insurers to cover preventive care without cost-sharing reduces the racial and ethnic disparities in health services consumption for uterine cancer survivors. Exploiting the nationally representative Medical Expenditure Panel Survey (MEPS) from 2008 to 2017 and using the difference-in-difference modeling, the empirical estimates indicate that the ACA has increased consumption of both preventive and curative services among Black and Hispanic survivors. Moreover, the ACA is associated with reductions in racial and ethnic disparities in prescription drug spending. The findings have implications for improving racial ethnic minority female population health outcomes, workplace productivity and earnings, and reducing early retirements.
Heterogeneity in Corporate Debt Structures and the Transmission of Monetary Policy (E5, G2)
We study how differences in the aggregate structure of corporate debt affect the transmission of monetary policy in a panel of euro area countries. Consistent with the bank lending view of transmission, we find the cost of bank loans to rise relative to the cost of corporate bonds in response to a standard monetary policy tightening shock. The strength of this effect depends on the financing structure prevailing prior to the shock. In economies with a high initial share of bond finance, the cost of credit rises by less as firms resort to bonds as a `spare tire' to compensate for the loan supply contraction and a smaller portion of firm credit is remunerated at the loan rate. In economies with a low share of bond finance, firms face a more limited scope to replace loans with bonds and the rise in the cost of credit is reinforced by a shift in the composition of credit demand towards bank loans. As a consequence, a higher share of bond finance goes along with a weaker transmission of standard monetary policy shocks to real activity. By contrast, the transmission of monetary policy shocks to longer-term yields tends to strengthen with the share of bond finance in the economy.
Heterogeneous Farmers’ Technology Adoption Decisions: Good on Average Is Not Good Enough (Q1, D9)
In spite of the importance of agriculture sector and persistently low agricultural productivity, smallholders in Sub-Saharan Africa are unenthusiastic about a seemingly profitable modern technology — fertilizer, and at the same time are keen on a seemingly unbeneficial traditional technology — intercropping. This paper aims to understand the rationale behind farmers’ decisions about agricultural technology adoption and explain that adoption puzzle. I construct a farmer’s decision-making model that takes into account both the expected value and the variance of a farmer’s profit. This model features the farmer's production function with multiple technology choices, heterogeneous returns, selection bias, and heterogeneous variances for each technology. Using a Tanzania panel dataset, I find that the expected returns of adopting the same technology vary significantly across farmers. Furthermore, adopting fertilizer significantly increases expected yields for farmers who adopt it every year, yet the higher expected returns are accompanied by larger variances. On the other hand, adopting intercropping does not increase the expected returns but significantly decreases the variance of yields. Farmers' technology adoption decisions are influenced positively by the expected value of profits and negatively by the variance of profits. These empirical results explain the low adoption rates of an intensively promoted higher-average-return technology such as fertilizer, and justify the high adoption rates of a seemingly unprofitable technology such as intercropping.
High Demand for Redistribution but Low Individual Giving: Evidence from Small and Large Groups (D6, C9)
We attempt to bridge the gap between the evidence on social preferences from laboratory experiments and their expression in the field. We report on the results of an online experiment of 1200 people, studying redistribution and giving decisions in small (four people) and large (200 people) groups. We document that preferences for giving and preferences for redistribution are distinctly different, and that this difference is increasing with group-size. This implies that demand for taxation among the rich cannot be substituted by charitable giving. Also, taxes can be a tool to help overcome coordination problems, insuring that the rich’s demand for redistribution is implemented.
Hiring Practices and Discrimination (J7, J2)
In this paper, I address two research questions. First, which job hiring designs curb discriminatory practices and result in better hires? Second, what kind of hiring committees should carry out such hiring designs? Exploiting the introduction of blind exams, the quasi-random assignment of recruitment committee members to selection processes, and plausibly exogenous variation of hiring designs within the same occupation or employer, I leverage rich information on applicants and committee members in Brazil’s public hiring to open the black box of hiring decisions. Public sector job selection processes in Brazil comprise a combination of one or several blind exams, resume analysis, oral evaluations, and interviews, where committee members individually score each candidate on every process step.I construct a large-scale applicant-reviewer panel by scraping over 35 million text documents containing information on federal, state, and local public sector hiring from official government gazettes. These documents contain highly dimensional unstructured text detailing the procedures of each hiring process, including rules, names of applicants and hiring committee members, wages, job description and requirements, and individual scores in each selection round. Because job selection rounds are contained among confounding information and are not linked over time, I develop an NLP algorithm that exploits common syntactic structure from legal writing in official documents. The algorithm filters out relevant information and match names of applicants and committee members to resumes, race, gender, age, and professional connections from other publicly available sources.
How Ambitious Can the Israeli Green Deal Be? (Q4, E1)
Israeli policy makers are considering carbon reduction targets for 2050. The goal of this study is to provide a comprehensive, economy-wide analysis of the alternative pathways for energy-related carbon emissions reduction in Israel. An integrated bottom-up, top-down modeling exercise, based on an original MESSAGEix_IL-MACRO framework, was performed over the Israeli energy system to assess the cost-effectiveness of greenhouse gas (GHG) emissions reduction options—firstly with renewable energy and transport electrification targets and secondly by imposing a carbon tax. The results show that, by the adoption of such a policy or a more ambitious policy (with a higher carbon tax), energy-related GHG emissions could be reduced by about 60% to 90% respectively, by 2050 relative to the reference year of 2005, with only a minor impact on the growth of the national GDP. Decarbonization of the Israeli economy will necessarily be based on increasing the electrification of transport and industry and on generating power from renewable energy resources (mainly solar). The unique challenge for Israeli policy makers is a population growth rate that is unprecedented in the developed world. The infrastructure should be developed rapidly to keep the growth of the standard of living intact. This challenge also presents the opportunity for a quick transition to a cleaner economy. The modeling tool and its outcomes can provide valuable insights for the design of clean energy policies that permit the fostering of sustainability targets. This methodology results in various scenarios that may help decision makers to understand the options available to them to accomplish the ambitious goals and targets they may set.
How Does Parental Out-Migration Affect Left-Behind Children's Education? (J6, I2)
In this paper, I investigate how parental out-migration affects the schooling outcomes of left-behind children in rural China. Unlike previous works which almost exclusively focus on the net effect of migration, I analyze three important causal mechanisms — parental absence, child’s study time, and investment in the child — simultaneously via a mediation analysis, disentangling the total effect of migration into mechanism-specific effects which are informative for policymakers. The analysis can be justified by the equilibrium solution of a theoretical two-agent model. The identification strategy is based on the rank condition for structural equation models to handle the endogeneity and Heckman selection model to correct for nonrandom missing. Using survey data on rural households from nine provinces, I find that the effects through parental absence and investment are both significantly negative with large sizes, while the effect through child’s study time is insignificant with a negligible size. The surprising negative effect through investment is mainly driven by reduced nutrition investment by de facto custodians, who may not have compatible incentives to allocate the remittances to the child. Through a refined subgroup analysis, I find that girls are suffering ten times more from the underinvestment than boys, revealing a shocking gender inequality in rural China. The findings suggest that policies that compensate for underinvestment, especially for girls, tend to be more effective in mitigating the negative effect of migration than other types of policies.
How Does Ride-Hailing Service Hit Household Vehicle Ownership? Evidence from National Microdata (R4, D1)
The ride-hailing service is trending up in the most recent decade. Like private driving and public transit, it becomes a major mode of commute in urban areas. The growing popularity of the ride-hailing service changes the landscape of the mobility market. Using data from the 2017 National Household Travel Survey, we examine the impacts of using ride-hailing services on the vehicle ownership for households from 43 metropolitan areas across the United States. Particularly, we estimate an ordered probit model of household vehicle ownership with endogenous treatments of using ride-hailing services in an instrumental variable approach. We find that highly frequent users are 2.21 and 3.44 percentage points more likely to possess no vehicle and one vehicle than regular users, respectively, while their probabilities of possessing two vehicles and three or more vehicles are 1.99 and 3.66 percentage points less. Also, our results show that the probabilities of possessing different numbers of vehicles do not vary significantly across respondents who use ride-hailing services no more than twice a week and that highly frequent users are more willing to reduce their vehicle holdings in contrast to others. Additionally, extrapolating our results from representative respondents to the population in the sampled areas, we find if all regular users convert to highly frequent users, their average vehicle holdings would reduce by 8.61 percent and the total decrease is approximately up to 190,000 vehicles, accounting for 1 percent of new vehicle sales in 2017.
Human Capital Portability and Careers of M&A Advisors (J4, J6)
We quantify the importance of firm-specific human capital in explaining workers' career choices. We develop a model that allows workers to accumulate both portable and non-portable human capital through their work experience and learn about their match quality with current employers over time. We also allow bankers to choose between firms that offer different levels of portability and production efficiency.
The model is estimated to match banker career data in the M&A advisory industry, which is populated by bulge bracket and boutique firms. Our estimation suggests that bankers in boutique firms accumulate less portable human capital but enjoy higher efficiency. Such a trade-off explains why bankers are more likely to choose bulge bracket banks at the start of their careers but increasingly migrate to boutique banks when they become more seasoned. We also gauge the extent to which non-portable human capital affects labor allocation and shapes industry structure.
Identifying Aggregate Shocks with Micro-Level Heterogeneity: Financial Shocks and Investment Fluctuation (E2, G3)
This paper identifies the aggregate financial shocks and quantifies their effects on business investment based on an estimated DSGE model with firm-level heterogeneity. On average, financial shocks contribute only 1.1% of the variation in U.S. public firms' aggregate investment. The negligible aggregate relevance of financial shocks mainly results from the interaction between firm-level heterogeneity and general equilibrium effects. Following a contractionary financial shock, financially constrained firms are directly forced to cut investment, which dampens the aggregate investment demand and lowers the capital good price. The lower capital good price motivates the financially-unconstrained firms to invest more, which largely cancels out the financial shock's direct effect in aggregation. If the firm-level heterogeneity is removed, the implied relevance of financial shocks to aggregate investment will be 50 times larger. This sharp difference indicates that representative firm models could overstate the relevance of financial shocks in driving the business cycle fluctuation and highlights the importance of micro-level heterogeneity in identifying the aggregate shocks.
Identifying Consumer Preferences from User-Generated Content on Amazon.com by Leveraging Machine Learning (D8, M3)
Inexperienced consumers may have high uncertainty about experience goods that require technical knowledge and skills to operate effectively; therefore, experienced consumers’ prior reviews can be useful for inexperienced one. However, one-sided review systems (e.g., Amazon.com) only provide the opportunity for consumers to write a review as a buyer and contain no feedback from the seller’s side, so the information displayed about individual buyers is limited. Therefore, this study analyzes consumers’ digital footprints (DFs) for programmable thermostats (home energy management devices) to identify and predict unobserved consumer preferences, using a dataset of 141 million Amazon reviews including consumer reviews and product-specific information. In addition, this study identifies consumers’ sentiment toward product content dimensions (PCDs) extracted from review text by applying topic modeling and domain expert annotations, while excluding questionable reviews (posted by “suspicious one-time reviewers” and “always-the-same rating reviewers”).
This paper obtains three main results: first, I find that the factors that affect consumer ratings are: (a) user’ DFs (e.g., average rating across all categories), (b) reviewers’ attitudes toward eight product content dimensions (smart connectivity, easiness, energy saving, functionality, support, price value, privacy, and the Amazon’s service quality effect), and (c) other prior reviewers DFs (e.g., length of the review summary). Second, extreme gradient boosting (XGBoost) is found to obtain the highest performance for predicting the ratings of potential consumers before they make a purchase or write a review. Third, a convolutional neural network (CNN) on top of Bidirectional Encoder Representations from Transformers (BERT) embedding shows the highest performance for classifying consumers’ sentiment toward a specific PCD.
The two main contributions of this study are that it shows: (1) how to use user-generated-content about different products in different categories and at different times to identify unobserved consumer preferences; and (2) a detailed, step-by-step way to combine structured and unstructured data to predict individual heterogeneous consumer preferences and classify their sentiments. Overall, this approach developed in this paper is applicable, scalable, and interpretable for distinguishing important drivers of consumer reviews for different goods in a specific industry and can be used by industry to design customer-oriented marketing strategies.
Identifying Preference Shocks: Earthquakes, Impatience, and Household Savings (E2, D8)
The trauma of experiencing a natural disaster can have a devastating impact on individuals, even if they are not economically devastated. We exploit the panel structure of a nationally representative survey in Italy to explore plausible mechanisms linking traumatic experience to individual attitudes towards time. Combining the household survey with city level impact data of the 2009 L’Aquila Earthquake, we identify households who felt the shake but had no substantial changes in their economic circumstances. We elicit their patience using a survey measure about their willingness to pay for the immediate realization of a future payment. Our difference-in-differences estimates show that affected individuals become more impatient compared to similar individuals who did not feel the shake. Consistent with this finding, the same households increase consumption and de- crease their savings. We also show that temporary trauma can have a lasting impact. People save less from disposable income for several years after the earthquake.
Identifying SVARs from Sparse Narrative Instruments: Dynamic Effects of United States Macroprudential Policies (C3, E5)
We study the identification of policy shocks in Bayesian proxy VARs for the case that the instrument consists of sparse qualitative observations indicating the signs of certain shocks. We propose two identification schemes,
i.e. linear discriminant analysis and a non-parametric sign concordance criterion. Monte Carlo simulations suggest that these provide more accurate confidence bounds than standard proxy VARs and are more efficient than local projections.
Our application to U.S. macroprudential policies finds persistent effects of capital requirements and mortgage underwriting standards on credit volumes and house prices together with moderate effects on GDP and inflation.
Immigration and Worker-Firm Matching (J6, J2)
The matching process between firms and workers is an important mechanism in determining the distribution of wages. In a labor market characterised by large dispersion of workers' productivity and worker-firm complementarity, high quality firms have strong incentives to screen for and match with high quality of workers generating what is called positive assortative matching (PAM). Immigration in a local labor market, by increasing the variance of workers abilities, may drive stronger PAM between firms and workers. Using French matched employer-employee data over the period 1995-2005 we document that supply-driven increases in the local immigrant population share increased firm-workers PAM. We then show that this association is consistent with causality, is quantitatively significant, and leads to higher average wages and higher profits, but also higher wage dispersion. We also show that the increased degree of positive assortative matching is mainly reached by high-productive firms "attracting" higher quality workers;
Immigration Policy and Mental Health of the Hispanic Population in the United States (I1, D7)
The deterioration of public mental health in the US has become more prominent in recent years. The disparities in mental health among ethnicities have also widened and raised concerns in the public health sector. It can be attributed to various economic and social factors. One factor that has been neglected and is becoming a catalyst to the disparities is immigration policy. Among various immigration policies, Employment Eligibility Verification (E-Verify) mandate is the most widely adopted nowadays. It requires employers to verify the employment eligibility of new employees. While E-Verify does not target a specific ethnic group, there are concerns that the mandate may enlarge the mental health gap between the Hispanic and other populations because the Hispanic population may be disproportionally affected by E-Verify.
This study aims to examine the impact of E-Verify mandate on the mental health of different ethnicities and investigate if Hispanics are particularly affected by E-Verify. The study uses the Behavioral Risk Factor Surveillance data as well as the multi-year and multi-state E-Verify adoption information in a Difference-in-Differences model, and finds that the adoption of E-Verify increases Hispanics’ “mental health not good days” by 1.7 days and significantly increases their probability of having severe mental health issues, particularly for those who are unemployed. In addition to the negative mental health impact on undocumented immigrants, there is also a spillover effect on the general Hispanic population. However, E-Verify does not affect the white and black populations, thus enlarging the mental health disparities across ethnicities. The adverse mental health repercussions of stringent immigration policies on the Hispanic population indicate how immigration policies could bring disproportional impact on a specific population, and this research calls for effective interventions to address the unintended consequences of immigration policies.
Impact of Longer Maternity Leave on Maternal Mental Health and Wellbeing: Evidence from Chile (I1, J2)
I study the causal effect of longer maternity leave on postpartum depression and other measures of
mental health. I take advantage of exogenous variation in paid maternity leave introduced by a 2011
policy change in Chile, which increased the paid postnatal leave period from 12 to 24 weeks. Using a
difference-in-difference instrumental variable approach, I find that one extra week of paid maternity
leave increases the probability of being diagnosed with postpartum depression by 0.3 percentage
points. However, I find no evidence of an increase in self-reported maternal stress level during the
first 18 months after delivery using the Parental Stress Index (PSI) or an increase in longer term
depressive symptoms measured by the CES-D scale. A likely mechanism driving the increased
diagnosis is that longer paid maternity leave increases the likelihood of diagnosis given a fixed level of
depressive symptoms. I show that longer leave increases labor market attachment of mothers and
thereby affects insurance coverage and doctor visits. By remaining employed, women are not only less
likely to be uninsured (1.2 percentage points), but also they get access to better health insurance (7
percentage points). The availability of better health insurance combined with more time to go to the
doctor, increase doctor visits during the first year after delivery, especially mental health visits (1.2
percentage points). Taken together, the evidence suggests that longer maternity leave increases
postpartum depression diagnosis mainly by increasing diagnosis as opposed to worsening mental
health. This is important because diagnosis is an important prerequisite to get treatment.
Impact of the Change in Payments on the Actual and Perceived Behaviors of Medical Care Providers (I1, I3)
Prior literature established the link between a person aging out of a parent’s insurance coverage at age nineteen and a signiﬁcant decrease in insurance coverage of those nineteen-year-old young adults. Using the regression discontinuity framework, this paper furthers that research by establishing that although there was no change in the total income received by the medical care providers treating young adults who have aged-out of their parent’s insurance, there was a signiﬁcant change in the amounts received from various sources that comprise the total payment. I examine the impact of the change in the providers’ payments by the source on the providers’ behavior (supply-side) and on the patients’ perception of the providers’ behavior (demand-side), using a 14-year sample of unmarried young adults from the Medical Expenditure Panel Survey (MEPS). I ﬁnd that although there is a statistically signiﬁcant change in the sources of the total payments received by medical care providers from patients crossing the age of nineteen thresholds, medical care providers do not change their actual treatment decisions. However, the patients do perceive a statistically signiﬁcant negative change in the behavior of their medical care providers.
Impacts of the Digital Economy on Price Stickiness and Monetary Non-neutrality: Evidence from China (E3, E6)
In recent years, the rapid development of the digital economy has profound impacts on society, but meanwhile, short-term fluctuations in macroeconomics and finance have become more frequent. Therefore, it is worthwhile to study impacts of the digital economy on price adjustment and macroeconomic dynamics. This paper measures and compares online and offline price stickiness with unique micro big data in China, and employs the empirical evidence to calibrate the heterogeneous multi-sector general equilibrium price adjustment model. Specifically, our online data contains prices from more than 100 websites covering the whole basket of Chinese CPI with over 19 million price records, including 8 divisions, 46 groups, and 262 classes. The offline data is from the price monitoring center of the National Development and Reform Commission in China, which includes 126 types of food, daily industrial consumer goods, and services. We find that the weighted average monthly offline price change frequency (size) is 13.92% (16.99%), compared with the online price change frequency (size) of 47.18% (13.53%). With online markets accounting for about 20% in China, the effectiveness of monetary policy is estimated to be 53% of the pure offline market without intermediate input, which is further reduced to 42% considering the intermediate input. If the online market share reaches 100% in the future (under an extreme condition), the effectiveness of monetary policy is only 7% to 11% compared to the pure offline market. This paper shows that impacts of the digital economy on the transmission of monetary policy should not be ignored. In the future, central banks should pay more attention to high-frequency online inflation indicators, attach more importance to structural monetary policy, and accelerate the transformation of monetary policy in the digital economy era.
Incentives, Search Engines, and the Elicitation of Subjective Beliefs: Evidence from Representative Online Survey Experiments (C8, D8)
A large literature studies subjective beliefs about economic facts using unincentivized survey
questions. We devise randomized experiments in a representative online survey to investigate
whether incentivizing belief accuracy affects stated beliefs about average earnings by
professional degree and average public school spending. Incentive provision does not impact
earnings beliefs, but improves school-spending beliefs. Response spikes suggest that the latter
effect likely reflects increased online-search activity. Consistently, an experiment that just
encourages search-engine usage produces very similar results. Another experiment provides no
evidence of experimenter-demand effects. Overall, results suggest a trade-off between
increased respondent effort and the risk of inducing online-search activity when incentivizing
beliefs in online surveys.
Individualism, Human Capital Formation, and Labor Market Outcomes – International Evidence from an Adult Skill Assessment (J2, P5)
Previous research shows that family background is a highly relevant, maybe even the most relevant, determinant of human capital formation. While existing studies have mostly considered parental education, income, and wealth as those characteristics relevant for human capital development, much less attention has been devoted to another factor that is also transmitted from parents to children: culture. We investigate the role of individualism – as a main cultural trait emphasizing personal freedom and achievement – for human capital formation and labor-market success. To do so, we use internationally comparable data on cognitive skills and labor-market outcomes across 33 countries.
We combine several complementary empirical approaches, each based on a different source of variation in individualism: First, we apply an epidemiological approach, which relies on the idea that culture is persistent and that migrants take their original cultural toolkit with them when they migrate. We compare first- and second-generation migrants from different cultural backgrounds who move to the same destination country, controlling for differences in educational and labor-market institutions across destination countries. We also go beyond the classic epidemiological approach by constructing a novel individualism measure at the personal level. This allows us to more rigorously account for migrants’ origin-country factors and to even extend the within-country analysis to natives.
Consistently across the three approaches, a one-standard-deviation increase in individualism is associated with about 0.3 standard deviations higher numeracy skills. Individualists also have larger skill gains over time, are less likely to be unemployed, earn higher wages, invest more in on-the-job training, and choose more research-oriented occupations. These results are in line with the substantial social rewards for innovation, uniqueness, and personal achievement (rather than loyalty and in-group cooperation) in individualist cultures. We also find that individualism is more important for adult-life outcomes than other cultural dimensions, in particular, long-term orientation.
Inform Me When It Matters: Cost Salience, Energy Consumption, and Efficiency Investments (Q4, D9)
Effective attention to information may play a prominent role in consumer choice for energy-intensive services and it may simply be a function of receiving timely information when consumption takes place. This paper investigates whether and why the timing of utility bills leads to salience bias in heat energy consumption. In Germany, the 12-month billing period varies across buildings with a significant share of buildings receiving bills during the summer months, when the salience of heating costs is absent or low. I exploit this large-scale natural experiment in utility billing cycles at the building level to identify the salience effect of costs on energy consumption and the underlying heterogeneity in the average treatment effect. I find new evidence for consumer inattention to energy costs: consumers that are billed for heating during off-winter months demand more heat energy annually. Results suggest that households are paying attention to their heating costs in the first three months of the 12-month billing period. As a result, bills immediately before the winter heating season are most effective, allowing ample opportunity to adjust consumption. I show that salience bias in consumption is persistent and pervasive – affecting households in all regions and building/technology type. Engaging energy users with salient bills, not necessarily more frequent, has the potential to reduce energy consumption in the residential sector significantly. This paper further examines whether enduring differences in consumer inattention to energy costs had a long-run impact on thermal efficiency investments by building owners – with implications for the energy-efficiency gap.
Information Avoidance and Internet Privacy (O3, K0)
There is a widespread intuition that people are inconsistent about pro- tecting their privacy. This paper presents an experiment that demonstrates that people engage in information avoidance when making privacy decisions. People who are will- ing to pay nearly an hour’s worth of wages for privacy are also willing to give away their data for small money bonuses if given a chance to avoid seeing the privacy con- sequences. Placebo tests confirm that the same behavior does not occur when people make decisions between two money bonuses. The paper also presents evidence on how this pattern changed during the Cambridge Analytica scandal.
Information Choice and Shock Transmission (G0, D8)
During the 2007-2008 financial crisis, countries that were relatively more exposed to the crisis epicenter, the United States, were among the least affected. This counters the intuition that the impact of a shock increases with exposure to it, and raises the question of the mechanism through which the impact of shocks can decrease with exposure. I propose a model in which decision-makers learn about the risk factors they are exposed to, but have limited capacity to process information. I find that decision-makers optimally choose to learn more about the risk factors they are more exposed to, and this informational advantage mitigates the impact of shocks by enabling them to take better investment decisions. Relative to an exogenous information benchmark, the endogenous information model I propose predicts that shocks to risk factors that decision-makers are relatively more exposed to are attenuated, while shocks to risk factors that decision-makers are relatively less exposed to are amplified.
Information Illusion: Different Amounts of Information and Stock Price Estimates (D8, C9)
This study analyzes investors’ perception of placebic information and its impact on stock price estimates. We initiate a questionnaire-based stock price forecast competition among 196 undergraduate students in business administration. We show that placebic information increases the perceived amount of relevant information. Individual participants’ characteristics, such as gender, financial knowledge or overconfidence, do not affect these findings. Placebic information does not alter participants’ stock price estimates and their accuracy, but it has an impact on individual expectations about the stock price forecast competition itself. The findings indicate that placebic information leads to information illusion. As reaction to the illusion, less overconfident investors decrease their expectations with regard to payoff and chances to win a prize in the competition. More overconfident participants do not show the latter behavior. Our findings provide implications for practitioners and researchers alike. Since the participants in our study serve as a proxy for economically educated young adults who are likely to invest in stocks in the future, both regulators and policy makers should consider that placebic information can significantly impact investors’ perception and, therefore, regulation on information that is provided to retail investors should focus on relevant and avoid irrelevant information. Researchers should be aware that placebic information asymmetrically influences expectations of participants in experiments who show different levels of overconfidence.
Insights on Student Loan Debt from Linked Administrative Records (I2, D1)
American families carry more than $1.5 trillion in student loan debt. This debt provided many with the opportunity to pursue higher education, but remains for others a large, potentially crippling, financial burden. In this report, we explore how people of different socioeconomic groups are managing their student debt. We do this by linking administrative banking data, credit bureau records, and public records on race and ethnicity to create a unique data asset that includes the income, demographics, debt balances, and student loan payments of 301,583 individuals. In general, we find that borrowers of socioeconomic groups tend to manage student loans quite differently, often relying heavily on others—children, parents, and About the Institute spouses—in order to manage their debt. In particular, we find that while the median borrower is not unduly burdened by their debt, a significant minority of lower-income and younger borrowers are heavily burdened, required to make payments that constitute more than 10 percent of their take-home income. We also find that almost 40 percent of those involved in student debt repayment are making payments on other people’s loans, with 27 percent of those involved holding no student debt whatsoever. These outside helpers play a key role in helping borrowers make progress on their loan. Nevertheless, we find that low-income and older borrowers are more likely to be several months behind on their payments, and 7 percent of all borrowers not in deferral are on track to never pay off their loans. These dynamics of repayment put Black borrowers at a disadvantage, who, relative to White borrowers, have lower incomes and higher debt balances and are 4 times as likely to have no payments made against their loans, partly due to the fact that they are less likely to receive repayment help. This debt provided many with the opportunity to pursue higher education with commensurate income keeping debt burdens at reasonable rates. For others, student loan debt remains a large financial burden relative to income. In this report, we explore how people of different socioeconomic groups are managing their student debt.
Instruction Time and Student Achievement: The Moderating Role of Teacher Qualifications (I2, C2)
Recent evidence suggests a positive effect of the quantity of instruction on student achievement. In this paper, I focus on the interaction between the quantity and the quality of instruction. Using international TIMSS data, I exploit within-student between-subject variation. I find that on average, an additional hour of instruction time leads to an increase of 0.03 standard deviations in students’ test scores across all countries. Importantly, these effects of instruction time are significantly larger for students with better qualified teachers, resulting in an increase in test scores of 0.04 to 0.05 standard deviations. While on average, instruction time has no significant effect in developing countries, it increases test scores by 0.02 standard deviations when taught by a high-qualified teacher also in developing countries.
International Spillovers of New Monetary Policy (E5, E4)
We study spillovers of conventional and new monetary policies of a large economy to a small open economy (SOE). Building on Sims and Wu (Evaluating Central Banks' Tool Kit: Past, Present and Future," 2020, forthcoming in the Journal of Monetary Economics), we employ a medium-scale New Keynesian model that features all the major types of new monetary policies and the conventional monetary policy in a unified framework. We extend their model to an open economy setting. We use our model as a measurement device to
quantify the spillovers and study the economic mechanisms behind them. In our quantitative application, Canada is the SOE and the US is the large economy. Our results show similar spillover effects of conventional and new monetary policies on GDP of the SOE. However, the effects on various components of GDP (consumption, investment and net exports) differ by policy.
We also simulate counterfactual monetary policy scenarios for the US and Canada around the Great Recession of 2008. Three main conclusions emerge from these simulations:
(1) if the Fed had not engaged in quantitative easing (QE), the US recession in the wake of the 2008 financial crisis would have been deeper but Canada would have had better economic outcomes; (2) there are diminishing returns to QE in terms of its effects on both the US and Canadian real variables; and (3) had the Bank of Canada followed the Fed and engaged in QE of its own during the Great Recession, the real economic outcomes would have been better for Canada.
Investment Plans, Uncertainty, and Misallocation (E2, O1)
We use data on firms' expectations and planned capital expenditures to show planned investment (i) is partially flexible to real-time shocks, and (ii) is a strong predictor of actual investment, with higher statistical importance than expected sales. To explain these facts, we develop an investment model with endogenous learning and partially flexible investment plans. Our calibrated model shows managers actively use both strategies, but prefer better information over ex-post adjustments. Moreover, our results suggest that capital misallocation from uncertainty is much smaller than in a standard firm dynamics model. Finally, our model predicts counter-cyclical uncertainty via endogenous fluctuations in returns to learning.
Judicial Capacity Increases Firm Growth Through Credit Access: Evidence from Clogged Courts of India (O4, K4)
How do judicial institutions, such as sub-national courts, impact economic growth? I examine the effect of trial court capacity on local firms' performance by exploiting quasi-random variation in judge vacancies and mapping trial records for a third of such courts in India with court-level performance measures, bank lending, and firm outcomes. I find that reducing judge vacancy increases local firms' labor use, production, and profitability through improved access to bank credit arising from better enforcement of debt contracts. Addressing judge vacancy would generate orders of magnitude larger benefit relative to the cost.
Judicial Independence and Development: Evidence from Pakistan (O1, K4)
This paper provides plausibly causal evidence that Presidential appointment of judges considerably impacts judicial independence and decision quality in Pakistan. We find that when the judge selection procedure changed from Presidential appointment to appointment by peer judges, rulings in favor of the government decreased significantly and the quality of judicial decisions improved. The age structure of judges at the time of the reform and the mandatory retirement age law provide us with an exogenous source of variation in the implementation of the reform. We test for and provide evidence against potential threats to identification and alternative explanations for our findings. The analysis of mechanisms reveals that our results are explained by rulings in politically salient cases and by “patronage” judges who hold political office prior to their appointments. According to our estimates, judicial appointment by peer judges prevents land expropriations worth 0.14 percent of GDP every year.
Labor Market Effects of Occupational Licensing Exams in Spanish: Evidence from Cosmetology (J6)
We use the introduction of written occupational licensing exams in Spanish to examine the effect of occupational licensing on the labor market outcomes of cosmetologists. Historically, written licensing exams for cosmetology were only available in English in all 50 states plus Washington, D.C.; however, the licensing exams began being offered in Spanish in some states in the 1990s and early 2000s. We use a multi-period difference-in-differences estimator to exploit the varying availability of licensing exams in Spanish across states and time to assess the effects of a state’s introduction of exams in Spanish on employment and earnings of both Hispanics and non-Hispanics. Using data from the Annual Social and Economic Supplement (ASEC) of the Current Population Survey (CPS), Decennial Census, and American Community Survey (ACS), our initial findings show that the introduction of state cosmetology licensing exams in Spanish result in an increase in the likelihood that a foreign-born Hispanic is a cosmetologist, with the largest effect for individuals with limited English language proficiency. Offering the exam in Spanish also appears to reduce the employment of native-born cosmetologists and reduce the usual number of hours worked per week for non-Hispanic, foreign-born cosmetologists, which indicate that foreign-born, Hispanic cosmetologists may serve as a substitute for other cosmetologists. Earnings results are inconclusive. Overall, our findings suggest that expanding the available languages for licensing exams increases access to the cosmetology profession for immigrants.
Large Dimensional Latent Factor Modeling with Missing Observations and Applications to Causal Inference (C1)
This paper develops the inferential theory for latent factor models estimated from large dimensional panel data with missing observations. We estimate a latent factor model by applying principal component analysis to an adjusted covariance matrix estimated from partially observed panel data. We derive the asymptotic distribution for the estimated factors, loadings and the imputed values under a general approximate factor model. The key application is to estimate counterfactual outcomes in causal inference from panel data. The unobserved control group is modeled as missing values, which are inferred from the latent factor model. The inferential theory for the imputed values allows us to test for individual treatment effects at any time. We apply our method to portfolio investment strategies and find that around 15% of their average returns are significantly reduced by the academic publication of these strategies.
Large Macroeconomic shocks during the Pandemic: a DSGE analysis (E3, C3)
This paper estimates a DSGE model including the Covid-19 episode. To let the data speak about the diverse types of macroeconomic shocks caused by the pandemic, we build a Two-Sector One (Two)-Agent model and estimate it through Bayesian methods. The resulting medium-scale New Keynesian model includes the standard real and nominal frictions used in the empirical literature and allows for heterogeneous pandemic exposure across sectors. On account of the magnitude of the involved shocks, we solve the model nonlinearly. To conduct inference on the resulting non-linear non-Gaussian system, we employ a version of the Cubature Kalman filter suited to handle the large shocks and use the Sequential Monte Carlo sampler to obtain parameters draws from the posterior distribution. Large shocks pose questions about the way of modelling the interdependence between them, so we use a flexible specification that allows to distinguish between all possible combinations of disaster shocks. Results show that the pandemic-induced economic downturn can be reconciled with a combination of demand and supply factors.
Large-Scale School Meal Programs and Student Health: Evidence from Rural China (I1, I3)
Malnutrition is still one of the most serious challenges faced by several developing countries. In 2019, around 149 million children were stunted (low height for age), which not only serves as a marker for unrealized physical and cognitive potential but further signals likely difficulties in school learning and failure to accumulate sufficient human capital. In recent decades, many countries have implemented national school feeding programs (SFP) with the aim to improve children’s health and nutrition, and SFPs have become one of the most important social protection programs worldwide. This paper evaluates the impact of the Nutrition Improvement Program (NIP) on student health in rural China. The program, which started in November 2011, targets poverty-stricken areas and participating students receive school meal subsidies equivalent to 800 yuan (US$130) per year or 7.6% of the average per capita disposable rural income. We use data from multiple rounds of the China Health and Nutrition Survey (CHNS) from 2004 to 2015 and implement a quasi-experimental approach exploiting cross-county variations in NIP implementation. We find that program participation is, on average, associated with a higher height-for-age z-score (HAZ) in the order of 0.2-0.4 standard deviations, although the effect does not translate into a lower stunting rate. The effect seems larger on students with an initial higher HAZ, and on girls than boys. We do not find significant effects on the Body Mass Index BMI-for-age (BMIz) and weight-for-age (WAZ) z-scores. The results are robust to alternative estimation methods and the use of different samples. Our findings suggest that the program is partially improving students’ health in rural areas, at least on HAZ over the first few years after implementation, but more support is needed, maybe through more intensive and targeted health and nutrition interventions.
Leadership in Scholarship: Editors' Influence on the Profession's Narrative (A1, O3)
Academic journals disseminate new knowledge and therefore can influence the direction and composition of ongoing research by choosing what to publish. We study the influence of editors and coeditors of the American Economic Review (AER) on the topic structure of papers published in the AER between 1976 and 2013 using a textual analysis of manuscripts. We compare AER's topic structure to that of the other top general interest journals. The appointment of new AER editors, while accompanied by a minor co-movement of AER topics towards topics of editors' post-appointment publications, is not an indicator of editors' personal taste in topics, but rather indicates the desire of those who appoint editors to premediate trends in other Top 5 journals.
Legal Systems Comparison: Firms' Strategies in Patent Litigation (K0, G0)
We build a real options model to examine the impact of legal systems on firms' strategies in patent litigation, with the presence of capital market frictions. We show that the English rule ("loser pays") and the American rule ("both litigants pay their own cost") interact with firms' financial constraints differently, leading to the English rule shift the effective bargaining power from the operating patent owner ("incumbent") to the alleged infringer ("challenger"). We find that compared to the American rule, (1) the royalty rate in an ex-post settlement (i.e., settlement after the filing of a lawsuit) is lower under the English rule; (2) the negative effect of product market volatility on settlement likelihood is more significant under the English rule; (3) the litigation thresholds are more sensitive to a key product market characteristics ("gain-to-loss ratio") under the English rule. Furthermore, we find that the winning probability of the incumbent lowers the settlement likelihood under the American rule, but it raises the settlement likelihood under the English rule. This paper takes a first step in understanding how legal systems affect corporate innovation, and it generates new policy implications and testable implications regarding IP litigation with a financing consideration.
Lending Standards and Output Growth (G0, E7)
What drives macro-financial vulnerabilities? Inspired by Minsky-Kindleberger narratives, one prominent view emphasizes that lending standards repeatedly deteriorate in good times, creating exposure to widespread reassessments of risk. Another emphasizes that leverage amplifies negative shocks. This paper constructs panel data on lending standards and uses it to show that Minsky-Kindleberger dynamics interact with leverage. Standards erode with improving economic performance, but do not always co-move with aggregate leverage. The combination of deteriorating standards and leverage--above and beyond leverage alone--signals poor subsequent macroeconomic performance. Inconsistent with models incorporating rational expectations, this poor subsequent performance is systematically reflected in forecast errors.
Limits of Stress-test based Bank Regulation: Cues from the Covid-19 Crisis (G2, G2)
This paper studies state-contingent policies in the context of bank capital regulation. Unlike a one-size-fits-all policy, state-contingent regulation can be aligned to individual banks’ riskiness. However, when riskiness is not perfectly observable, as with supervisory reviews or stress-tests that provide noisy indicators of risk, state-contingent policy can be misdirected, and lead to excessive (insufficient) regulation of a less (more) risky bank. This can diminish banks’ ex-ante incentives to improve their risk-return profile, and even reduce welfare relative to a state-independent policy. We show that with information frictions, state-contingent policies must have limits that depend on the accuracy of supervisory reviews.
Liquidity Constraints and Buffer Stocks Savings: Theory and Experimental Evidence (C9, D9)
We present an experimental assessment of a three-period saving/consumption model (which is difficult using observational data). One model where the liquidity in the second period is constrained (and, thus, borrowing is not possible) is contrasted with the unconstrained model. We also control the variance of the stochastic income, resulting in a 2x2 experimental design. We test the models' comparative statics and find--- in contrast to the models' prediction---that the liquidity constraint does not increase savings in the first period. Remarkably, we cannot reject all other hypotheses regarding the comparative statics. In further analyses using threshold regressions and a RDD, we find that debt aversion can rationalize this result. We also examine the role of sociodemographic characteristics on saving behavior and find that subjects who achieved a higher score on a cognitive reflection task deviate less from optimal savings.
Liquidity Support and Distress Resilience in Bank-Affiliated Mutual Funds (G0, G2)
We study whether the stability of mutual funds and the propensity of a run among investors depend on the ownership structure of fund management companies. Analyzing a large portion of the European open-ended mutual fund industry, we find that flows of funds run by banks or by firms that belong to the same financial group as a bank are less volatile and less sensitive to bad past performance. This enables bank-affiliated funds to better weather distress and to hold lower precautionary cash buffers in comparison with their unaffiliated peers. We explain this finding by showing that banks provide liquidity support to distressed affiliated funds increasing their stakes in those funds that are experiencing large outflows. This in turn diminishes the severity of strategic complementarities in investors' redemptions. We find that liquidity support and other benefits of bank affiliation are particularly strong if the parent bank is more liquid and better capitalised. Analyzing the aftermath of two exogenous shocks to financial markets -- linked to the Brexit referendum and to political uncertainty in the aftermath of 2018 political elections in Italy -- we also gather evidence that distress in the banking system spills over to the mutual fund sector via ownership links. Our research highlights substantial dependencies between the banking system and the asset management industry and identifies an important channel via which financial stability risks depend on the organizational structure of the financial sector.
Liquidity Traps in a World Economy (F3, F4)
This paper studies a New Keynesian model of a two-country world with a zero lower bound (ZLB) constraint for nominal interest rates. A floating exchange rate regime is assumed. The presence of the ZLB generates multiple equilibria. The two countries can experience recurrent liquidity traps induced by the self-fulfilling expectation that future inflation will be low. These “expectations-driven” liquidity traps can be synchronized or unsynchronized across countries. In an expectations-driven liquidity trap, the domestic and international transmission of persistent shocks to productivity and government purchases differs markedly from shock transmission in a “fundamentals-driven” liquidity trap.
Author web page: http://www.robertkollmann.com
Loan Insurance, Market Liquidity, and Lending Standards (G2, G1)
We examine loan insurance when lenders can screen at origination, learn loan quality over time, and can sell loans in secondary markets. Loan insurance reduces lending standards but improves market liquidity. Lenders with worse screening ability insure, which commits them to not exploiting future private information about loan quality and improves the quality of uninsured loans traded. This externality implies insufficient insurance. A regulator achieves constrained efficiency by (i) guaranteeing a minimum price of uninsured loans to eliminate a welfare-dominated illiquid equilibrium; and (ii) subsidizing loan insurance in the liquid equilibrium. Our results can inform the design of government-sponsored mortgage guarantees.
Local Environmental Governance and the Agglomeration of Dirty Industries (R3, Q5)
We investigate how environmental policies circumscribe the location decisions of polluting firms. Using a barbell model of demand asymmetry, we analyze how environmental regulations impact the location and production of firms’ servicing geographically disjointed markets. Our analysis shows that regulatory stratification impacts environmental policy and market outcomes. Specifically, we analyze how the purview of a regulatory agency (local, industrial, national) influences emission policy. We find that nationally set environmental emission tax alter firms’ locations, i.e. higher emission taxes sway “dirtier” (“cleaner”) firms to disperse (agglomerate). However, industry-specific taxes incentivize dispersion only in the case of products with high transportation costs. We also examine how environmental policies influence firms' incentives to consolidate. Industrial consolidation, under an industry-specific emission tax, encourages the spatial dispersion of polluting plants. Plant dispersion improves welfare due to transportation savings and higher levels of production. Yet, these production increases generate greater emissions relative to the competitive scenario. These findings run counter to previous studies. We find that spatial competition of polluting firms can be detrimental due to the concentration of emitted pollutants. Therefore, governments should prefer industrial consolidation for dirty industries with high transportation costs. Our analysis contributes to the regional, spatial, and environmental economics literature in several significant respects. The use of emission taxes and permits has steadily increased in recent years. As of 2018, dozens of countries have proposed or implemented taxes or permits for the emission of sulfur dioxide, nitrogen oxides, and carbon. Importantly, we show how the board application of these policies may impact the concentration of local production. Our findings also provide important policy insights for the hierarchical structure of environmental agreements. The crafting of environmental policy must consider the purview of regulatory agency.
Local Property Taxation and the Rental Housing Market (H2, R3)
This paper investigates the effect of changes in the property tax multiplier on rental housing prices in Germany. According to the German tax system, the property tax is paid by the landlord. However, it can be legally shifted towards the tenants, which raises the question of the economic incidence of the tax burden. The empirical analysis is based on a geo-referenced dataset provided by ImmobilienScout24, the leading online broker for real estate in Germany, supplemented with detailed information on the municipality level. The data contains comprehensive information on advertised apartments for the period 2008 to 2015. The results show that in the short-run an increase in the tax burden is mainly borne by the landlord. The results differ for urban and rural municipalities: In urban municipalities, the landlord is able to shift most of the increased tax incidence towards the tenant, while the tax cannot be fully shifted in rural municipalities. This can partly be explained by different demand elasticities for urban and rural housing, with the demand for housing in cities being less elastic compared to rural areas.
Long-Term Climate Forecasts (C3, Q5)
Climate is a long-term issue and therefore climate forecasts should be long-run forecasts. Using the realized quantile methodology introduced in Gadea and Gonzalo (JoE, 2020) where quantiles are time series objects, this paper proposes a simple method to produce long-term temperature density forecasts from observational data. They complement the projections obtained by physical climate models mainly focused on the mean temperature. These averages usually conceal wide spatial disparities that among other distributional characteristics are captured by our density forecast.
Our proposed method consists of running an out of sample forecast accuracy model competition (using the Giacomini and White test, 2006) and combining the forecasts of the resulting Pareto-superior models to eliminate the forecast model dependency. Analyzing the observational data from global cross-section stations CRU 1880-2018 (similar results are obtained with central England daily temperatures CET 1772-2018) we obtain three sets of important results: (i) In 10-25 years the mean global temperature will be above the 2C degrees upper bound set by the PCA and by the end of the 21st century the increment will be of 3.5C-4C degrees above pre-industrial levels; (ii) this increase is larger in the lower quantiles (e.g. q05 will go from the 0.07C pre-industrial level to 2.07C in 25 years and to 4.06C in 100 years) than in the upper quantiles (e.g. q95 from 25.6C to 27.05C in 25 years and to 27.66C in 100 years) producing a decrease in the variance of the temperature distribution and leading to more serious consequences than the standard increase in the mean, and (iii) there is an acceleration process reflected in a larger temperature increase when more recent time intervals (e.g. from 1950 to 2018) are used to produce the density forecasts.
Long-Term Contracts and Efficiency in the Liquefied Natural Gas Industry (L1, Q4)
Long-term contracts facilitate efficient levels of investment by sellers faced with the risk
of ex-post holdup. Contractual rigidities, however, reduce the ability of firms to respond
flexibly to demand uncertainty. This paper provides the first empirical analysis of the trade-off
between hold-up risk and contract rigidity, focusing on the liquefied natural gas (LNG) industry,
where long-term contracts account for over 70% of trade. I structurally estimate a model of
contracting, investment and spot trade that incorporates hold-up risk and contractual rigidities,
using a rich dataset on LNG contracts, investment, trade flows, and spot prices. LNG buyers are
estimated to have substantial bargaining power, implying that sellers face considerable hold-up
risk when making investment decisions. In the short-run, allocative efficiency would improve
through reduced use of long-term contracts. However, investment decreases by 35% in the
absence of long-run contracting, suggesting that long-term contracts play a significant role in
combatting holdup at the cost of short-term allocative efficiency.
Long-Term Impacts of Medical Education Reform on Local Children: Evidence from the Area Health Education Center Program (I1, I2)
This paper examines the long-term impacts of medical education expansion on local children’s health and education outcomes. By utilizing the federal Area Health Education Center (AHEC) Program, which was launched in 1972 and gradually expands to most states till recent years, this paper suggests that promoting medical education and training to local communities significantly improves children’s long-run health status and health behaviors. Using an event study empirical strategy with the controls of individual characteristics, family background, and regional time trends, the evidence shows that children with greater childhood exposure to this medical education reform are less likely to be overweight, report better health status, less depressed, and consume fewer alcohol drinks and cigarettes than the cohorts without any childhood exposure to the AHEC program from the same county. Furthermore, opening a local AHEC also increases local children’s higher-level educational attainment. While children from low-income and minority families benefit more from the opening of the local health education center, the impacts of the AHEC program still focus on the urban geographically.
Low Rates and Financial Stability (E4, G2)
I develop a recursive banking model based on Diamond and Dybvig (1983) and use it to study the effect of interest-rate shocks on financial stability. The key channel is through the franchise value of deposits, defined as the future net interest margins that the bank expects to make on its deposits. In the model, banks need a sufficient franchise value of deposits to be solvent. Due to the existence of a lower bound on deposit rates, low rates reduce the franchise value of deposits and may lead to financial instability. I characterise an effective lower bound on the interest rate, i.e. the lowest possible interest rate compatible with a stable banking system. The key finding of the paper is that this is sharply increasing with the persistence of the interest-rate shock.The model calibrated to the Euro Area implies that the financial system can withstand an interest rate of−28% temporarily. However, if the shock were permanent, interest rates below 1% would be enough to generate a banking crisis.
Lower-for-Longer under Endogenous Technology Growth (E5, E3)
When modeling total factor productivity endogenously in an otherwise standard DSGE setup, aggregate demand influences, in contrast to the conventional view, technology growth and hence the supply side of the economy. Money is non-neutral also in the long-run as monetary policy affects the technology stock via investment in R&D and technology adoption and thus the long-term trend level. We show that due to hysteresis effects in total factor productivity the true ZLB-induced losses are larger than commonly assessed. Standard Taylor rules lead to a premature tightening of monetary policy and permanent output losses. A hysteresis-augmented Taylor rules reverts output to its initial trend and via the inherent lower-for-longer property fosters inflation at the effective lower bound. We study price level targeting, average inflation targeting and temporary price level targeting strategies and show that they support the alignment of inflation with target during an ELB episode. In addition, they can significantly reduce the hysteresis-induced long-run output losses by alleviating the shortfall in technology-improving investments and thus total factor productivity.
Machine Learning for Zombie Hunting. Firms' Failures and Financial Constraints. (G3, C4)
In this contribution, we exploit machine learning techniques to predict the risk of failure of firms. Then, we propose an empirical definition of zombies as firms that persist in a status of high risk, beyond the highest decile, after which we observe that the chances to transit to lower risk are minimal. We implement a Bayesian Additive Regression Tree with Missing Incorporated in Attributes (BART-MIA), which is specifically useful in our setting as we provide evidence that patterns of undisclosed accounts correlate with firms’ failures. After training our algorithm on 304,906 firms active in Italy in 2008-2017, we show how it outperforms proxy models like the Z-scores and the Distance-to-Default, traditional econometric methods, and other widely used machine learning techniques. We document that zombies are on average 21% less productive, 76% smaller, and they increase in times of financial crisis. In general, we argue that our application helps design evidence-based policies in presence of market failures, for example optimal bankruptcy laws. We believe our framework can help inform the design of support programs for highly distressed firms, for example after the recent pandemic crisis.
Macro-Financial Interactions in a Changing World (F4, C5)
We measure the time-varying strength of macro-financial linkages within and across the US and euro area economies by employing a large set of information for each region. In doing so, we rely on factor models with drifting parameters where real and financial cycles are extracted, and shocks are identified via sign and exclusion restrictions. The main results show that the euro area is disproportionately more sensitive to shocks in the US macroeconomy and financial sector, resulting in an asymmetric cross-border spillover pattern between the two economies. Moreover, while macro-financial interactions have steadily increased in the euro area since the late 1980s, they have oscillated in the US, exhibiting very long cycles of macro-financial interdependence.
Macroeconomic and Financial Market Analyses and Predictions with Deep Learning (B4, C1)
Since Hinton, Osindero, and Teh (2006) developed the fast learning algorithm, deep learning has been powerful tools that have recently achieved impressive performance on a wide spectrum of industries as well as in academia. For the macroeconomic and financial variables, however, more elaborate approaches need to be taken due to the unique latent features of them. In this regard, we propose novel approaches to apply deep learning to the predictions of time series variables in those fields. Specifically, we suggest ensembles of neural networks and Bayesian learnings for estimating the posterior distributions of the forecasting outcomes as the out-of-sample forecasts. The examples are provided with the predictions of monthly Korea’s nominal exports and daily Korean won-US dollar exchange rates.
Make IT Work: the Labor Market Effects of Information Technology Retraining in the Netherlands (J6)
The aim of this paper is to evaluate the effects of an active labor market program (ALMP) for higher educated workers in the Netherlands. The one-year program is characterized by six months of full-time IT retraining followed by a six-month internship. To address potential selection effects into the retraining program, we apply matching methods to find a suitable control group. The matching variables include ability and personality scores. We merge our data with register data from Statistics Netherlands to estimate the effects of participating on earnings per month and working days per month. The results show significant lock-in effects during the program, lasting up to five months after program start. From six months after the program start onwards, we find significant positive effects on earnings and working days. These positive effects remain significant until the end of the 36-month evaluation period. A conservative cost-benefit analysis based on the effects on earnings shows an annualized internal rate of return of 2.52%, which is low compared to the returns to education in the Netherlands or around 8%. We conclude that IT retraining has positive effects on the labor market outcomes of the participants yet relatively low returns.
Male Migration & Changing Roles for Women in Agriculture in Rural India (J2, Q1)
Migration is widely prevalent across the world, and rural outmigration is a dominant stream of migration in developing countries, including India, and has important implications for agriculture. I study the effects of internal migration in rural India in the context of agricultural households.
By reducing the availability of family-labor to employ on the farms, migration could lead to labor re-allocation among the members left behind in order to substitute for the lost labor; on the other hand, the remittances could be reinvested in agriculture, increasing farm-labor and/or production. Migration could also result in the household modifying its involvement with agriculture – reducing cultivation area, changing cultivation-patterns, or maybe also move out of agriculture. Migration’s effect on labor-leisure choices among the left-behind members and the household’s decisions pertaining to agriculture are inter-twined. Particularly, if the migrant had participated in agricultural decision-making, his/her departure leads to a change in the decision-maker & nature of decisions taken, and subsequently, in the household’s agricultural outcomes.
Migration tends to be male-dominant, including in rural India. Women tend to be left behind, and take over agricultural operations, leading to Feminization of agriculture. This could reflect as more women toiling for longer hours on the farm – feminization-of-farm-labor. Women could also transition into new roles as farm managers – feminization-of-farm-management. Female-farm-managers tend to have lesser control over assets, know-how and credit. They may have previously been unexposed to the rigors of farm management, and face discriminatory norms impeding their managerial actions.
Using panel data, adopting a Difference-in-Difference approach combined with matching methods, I examine complex inter-relationships between migration, feminization and agriculture. I find that male-migration leads to feminization-of-farm-management, and female-managed farms report lower agricultural profits, suggesting the presence of challenges for female farm-managers to handle.
Manager Uncertainty and Cross-Sectional Stock Returns (G1, D8)
This paper evidences the explanatory power of managers’ uncertainty for cross-sectional stock returns. I introduce a novel measure of the degree of managers’ uncertain beliefs about future states: manager uncertainty (MU), defined as the count of the word “uncertainty” over the sum of the count of the word “uncertainty” and the count of the word “risk” in filings and conference calls. I find that manager’s level of uncertainty reveals valuation information about real options and thereby has significantly negative explanatory power for cross-sectional stock returns. Beyond existing market-based uncertainty measures, the manager uncertainty measure has incremental pricing power by capturing information frictions between managers’ reported uncertainty and investors’ perception of uncertainty. Moreover, a short-long portfolio sorted by manager uncertainty has a significantly positive premium and cannot be spanned by existing factor models. An application on COVID-19 uncertainty shows consistent results.
Market Liberalization, Dairy Intake and Adolescent Height: Results from the Chinese Health and Nutrition Study, 1991 – 2009 (D1, F6)
We explore the link between market liberalization and height-by-age among a panel of adolescent participants of the China Health and Nutrition Survey from 1991 to 2009, and the role of dairy consumption in moderating this relationship. China witnessed rapid (though internally uneven) economic growth in the late 1990s and early 2000s after the country’s accession to the World Trade Organization and the affiliated liberalization of many internal markets. We find that adolescents from provinces that experienced more intensive market liberalization were 1.2 cm taller than those from provinces experiencing less liberalization. We identify increased dairy consumption as a credible causal pathway between market liberalization and height, which is occasionally used as a proxy for net cumulative nutrition. Specifically, using a difference in difference model, we find panelists living in provinces with greater market liberalization consumed 10 grams more of dairy products per person per day than those in other provinces. While the life-long net nutritional impact of increased dairy consumption during childhood is the subject of ongoing research, increased childhood dairy consumption has been shown to promote greater attained height and increased growth velocity, particularly in regions with compromised overall diet quality (Willett and Ludwig 2020). Before market liberalization in the late 1990s, dairy products were primarily marketed through home delivery networks operated by local milk producers (Fuller et al 2006). Liberalization eased this geographic mismatch between dairy production and consumption and stimulated a substantial increase in available market channels. The study provides the first estimates of the role of market liberalization on Chinese dairy consumption, and should stimulate robust discussion of the role of market liberalization on population nutrition and the numerous causal pathways in addition to dairy consumption that may link liberalization and nutrition indicators such as height.
Measurement Error, Validation Data, and Program Evaluation (O1, Q1)
Measurement in applied economics involves trade-offs. Many outcomes of interest, like total output by a firm or farmer, are costly to measure objectively. In practice, these outcomes are often measured using survey data. When designing a study or field experiment, researchers typically work under financial or time constraints and must choose how to collect the best data possible given these constraints. What if the treatment in an experiment affects not only an outcome of interest, but also the measurement of that outcome?
In this paper, we explore the implications of a particular form of measurement error: differential mis-reporting by treated and control participants in a randomized experiment. We document the presence of this form of measurement error in self-reported cultivated acreage by farmers in a recent field experiment in western Kenya (Deutschmann et al 2019). We discuss some possible mechanisms for this error and demonstrate how the interpretation of the results of the experiment would change in the absence of GPS measures of cultivated acreage. We demonstrate the effectiveness of a validation data method adapted from Buonaccorsi and Tosteson (1993) and Carroll et al (2006) to correct for this error. We present Monte Carlo evidence that this method performs well, using both real data from field experiments and fully simulated data with a range of measurement error structures. Finally, we discuss the implications for experimental design and provide guidance for how researchers can account for this type of measurement error in practice.
Measuring Capital-Labor Substitution: The Importance of Method Choices and Publication Bias (E2, D2)
We show that the large elasticity of substitution between capital and labor estimated in the literature on average, 0.9, can be explained by three factors: publication bias, use of aggregated data, and omission of the first-order condition for capital. The mean elasticity conditional on the absence of publication bias, disaggregated data, and inclusion of information from the first-order condition for capital is 0.3. To obtain this result, we collect 3,186 estimates of the elasticity reported in 121 studies, codify 71 variables that reflect the context in which researchers produce their estimates, and address model uncertainty by Bayesian and frequentist model averaging. We employ nonlinear techniques to correct for publication bias, which is responsible for at least half of the overall reduction in the mean elasticity from 0.9 to 0.3. Our findings also suggest that a failure to normalize the production function leads to a substantial upward bias in the estimated elasticity. The weight of evidence accumulated in the empirical literature emphatically rejects the Cobb-Douglas specification.
Measuring Inequality Using Geospatial Data (O1, E0)
The main challenge in studying economic inequality is limited data availability. This is particularly problematic in developing countries, which are more prone to poverty and inequality. Furthermore, data quality is a challenge even in developed countries as quality standards of statistical offices differ substantially between countries and within countries over time. Finally, measurement of income inequality is hampered by tax evasion and shadow economy transactions that reduce the reliability of personal income data, which is the main input to traditional measures of income inequality. In this paper, we attempt to address these issues by constructing a new database of measures on economic inequality for a balanced global sample of 234 countries and territories, covering the period from 1992 to 2013. We use worldwide geospatial satellite data on nighttime lights emission as a proxy for economic prosperity and match them with data on geo-located population counts to construct Gini-coefficients. Key methodological innovations include the use of alternative levels of data aggregation, and a parsimonious calibration of the lights-prosperity relationship to match traditional inequality measures based on income data. Indeed, we obtain a measure that is significantly correlated with cross country variation in income inequality. However, we also find that our lights-based measures suggest generally higher inequality than traditional measures. The latter finding indicates that lights-based measures capture multidimensional aspects of inequality, such as shadow economies and wealth, that are likely missing in income-based measures. Finally, we provide illustrative applications of our new database, including the analysis of the effects of out-of-pocket health expenditure, epidemics, and financial liberalization on economic inequality. Particularly, we compare results obtained with income- versus our light-based Gini-coefficients, which also offer a substantial gain in terms of number of observations. We find that our measures lead to contrasting results mainly along country-specific dimension of inequality dynamics.
Measuring United States Manufacturing Services Trade Using United States Customs Records: A Proposed Methodology (F1, F6)
To align with international standards for compiling balance of payments statistics, the value of goods provided for processing without a change in ownership should be excluded from exports and imports of goods; instead the processing fee should be included in exports and imports of manufacturing services. These guidelines are based on the country that owns the goods for processing (the lead firm), regardless of the country that sends the intermediate goods or that receives the finished goods. It seems reasonable to assume, however, that most of the imports of manufacturing services involve roundtrip trade between the lead firm’s home country and a partner country. This paper explores the feasibility of using U.S. customs records to estimate imports of manufacturing services on roundtrip trade involving the United States. The paper presents a profiling approach to identify manufacturing services imports by using a set of profiling criteria to identify some of the firms and their roundtrip trades where intermediate inputs are exported to a country, processed, and imported back into the United States, for the year 2012. The paper uses the Census Bureau’s Longitudinal Firm Trade Transactions Database (LFTTD) that compiles, by firm, transaction-level data on trade in goods collected by the U.S. Customs and Border Protection, in conjunction with microdata from surveys of multinational enterprises conducted by Bureau of Economic Analysis (BEA) and technical coefficients from the BEA input-output accounts. With the objective of constructing a conservative lower-bound estimate, this study focuses only on the roundtrip trades between entities linked by a majority-owned related party relationship.
Mechanisms for a No-Regret Agent: Beyond the Common Prior (D8, C4)
A rich class of mechanism design problems can be understood as incomplete-information games between a principal who commits to a policy and an agent who responds, with payoffs determined by an unknown state of the world. Traditionally, these models require strong and often-impractical assumptions about beliefs (a common prior over the state). In this paper, we dispense with the common prior. Instead, we consider a repeated interaction where both the principal and the agent may learn over time from the state history. We reformulate mechanism design as a reinforcement learning problem and develop mechanisms that attain natural benchmarks without any assumptions on the state-generating process. Our results make use of novel behavioral assumptions for the agent -- based on counterfactual calibration -- that capture the spirit of rationality without relying on beliefs.
Mediation and Costly Evidence (C7, D8)
This paper studies the efficient mediation procedure where disputants are asymmetrically informed and hard evidence can be acquired probabilistically at a cost. A mediator commits ex-ante to a mediation plan that generates stochastic messages for the uninformed party, based on the informed party's reports, whether and which costly evidence to acquire. The model encompasses both facilitative mediation where mediator only transmits information, and evaluative mediation where mediator bases recommendation on evidence, thus has to acquire information.
The efficient mediation plan features two threshold where the lower threshold determines whether an evidence should be acquired, and the higher threshold determines whether the two parties can settle if evaluation turns out inconclusive. We show that (i) facilitation is always involved in efficient mediation, (ii) evaluation is required by efficient mediation if and only if efficiency demands all cases to be settled, and (iii) in an efficient mediation plan, weak cases are settled by facilitative mediation, strong cases are settled by evaluative mediation if required, and settlement for stronger case hinges on more precise yet risky evidence. Furthermore, who bears the burden of proof is irrelevant for efficiency.
While facilitation is the exclusive focus of previous literature, our findings suggest that evaluation is equally important for efficiency. Our findings highlight mediation default, which bears implications for the design of online platform, dispute resolution, ratings, and international relations.
Medical Decision Making Under Uncertainty (I1, D9)
The focus of this paper is to examine how individual differences in risk preference, tolerance of uncertainty and regret attitude could have effects on the medical decision making by looking at anaesthetist’s decision to operate under different risky scenarios. A survey is designed to elicit participant’s preference which includes three case-vignettes and three preference measures. Two models of single-stage and dual-stage were proposed in the analysis. We found that medical experts with stronger feelings of regret and stress from uncertainty report lower likelihood to operate, hence psychological traits have very different effects in the two-stage decision-making mechanism. We further characterized the decisions into two types: decision with absolute certainty in which one chooses to operate without doubt and decision with relative uncertainty where one operates with doubts. Participants are found less likely to operate with absolute certainty under medium-risk scenario than under the low-risk scenario and no association is found between the three effects and the decision making. In contrast, participants are less likely to operate with relative uncertainty under the high-risk scenario compared with low-risk scenario and regret and risk attitude plays a significant role in the process. Additional results show that females are less likely to choose to operate in both stages.
Menu Costs and Information Rigidity: Evidence from the Consumption Tax Hike in Japan (E3, E6)
When nominal interest rates hit the zero lower bound, conventional monetary policy will not work, meaning that new policies to stimulate the economy are needed. To solve this problem, Feldstein (2002) and Correia et al. (2013) propose that the government should raise consumption taxes to generate inflation in consumer prices and affect consumers' intertemporal decisions. Their unconventional proposal relies on the crucial assumption regarding firms' price-setting behavior. Specifically, they assume that tax-excluded prices (rather than tax-included prices) are sticky, so that consumption tax hikes are immediately passed through to prices.
To test this assumption, this study examines firms' price-setting behavior after Japan's consumption tax hike in April 2014. The main findings of this study can be summarized as follows. First, prices became less sticky after the tax hike. Specifically, the probability of price changes was higher after the tax hike than in the previous year, whereas the size of price changes was smaller. This finding suggests that firms paid menu costs (i.e., attached new price tags) when passing through the tax hike to their prices. Second, however, more than half of prices were not adjusted on a tax-excluded basis after the tax hike. This finding suggests that although firms attached new price tags, they failed to adjust their prices, likely due to the cost of information gathering.
To test the existence of information gathering costs, this study focuses on the price-setting behavior of multi-product firms. As Pasten and Schoenle (2016) argue, multi-product firms can use information on common shocks for repricing all their products, implying that they face economies of scope in information gathering. This study compares the impact of economies of scope on price adjustment after the tax hike and in the previous year to show that economies of scope arise mainly due to the cost of information gathering.
Migrant Remittances: Mixed Motives and the Impact on Household Expenditures (D1, R2)
Remittance helps maintain the bond between migrants and their households of origin. A huge collection of literature investigates why migrants remit and whether household expenditure is affected by the inflow of remittances. Existing studies primarily identify two theories of migrants' motivation for remittance sending: remittance as altruism and remittance as insurance. This study builds on the literature and identifies an additional motive: identity norm. Rather than being mutually exclusive, these motives are actually compatible with each other. To analyze all these potential motivations, I first develop a stylized theoretical framework that generates testable predictions. Then, using a panel dataset pooled from three waves of the China Laborforce Dynamic Survey (CLDS), I test whether the average amount of remittance from each migrant responds to household demographic structure, the experience of income shocks in the household, and the influence from peers via their remittance sending behaviors. Empirical findings echo predictions from the theoretical model, and provide supporting evidence for remittance sending under a mixture of different motives. Finally, I find a significant increase in household expenditures on farm inputs and housing following the receipt of remittances from migrants, which suggests that the inflow of remittance encourages both productive and consumptive expenditures for migrant-sending households in rural China.
Minimum Wage, Market Imperfection, and Production Efficiency: Panel Event Study in Vietnam (O1, L1)
Industries in developing countries involve a host of small businesses which compete with foreign or state-owned firms. This paper examines how an unexpected spike of minimum wages disproportionately affected productivity of domestic and foreign firms under the mixed market structure using Vietnam's enterprise census. While foreign firms were shielded from the policy shock, production of domestic firms has become smaller and informal. A panel event study estimation finds that small domestic firms constrained by the minimum wage regulation significantly reduced employment (-14%), profit margins (-1 percentage point) and TFP (-2%) over three years. Unproductive capital-labor adjustments contributed to the productivity loss. Moreover, a triple-difference regression indicates that small firms faced significantly large negative minimum wage effect when foreign or state-owned firms held strong price setting power in the same market.
Minimum-Wage Policy Implications in Higher Education (I2, J3)
In North America, approximately 50% of minimum-wage workers are between ages 17-29. Within this group, half are students. Therefore, it is natural to wonder whether minimum-wage policies affect post-secondary education (PSE) decisions. Using Canadian longitudinal data, we find that high minimum wages reduce university enrolment but stimulate enrolment in community colleges.
The effect on university enrolment is driven by high-school students from lower socio-economic backgrounds, who are less likely to attend university if the minimum wage is high upon graduation from high school. These students are typically less likely to enrol in university and are even less likely to do so when the minimum wage increases. The minimum wage strengthens the link between parental background and children educational attainment, reducing intergenerational mobility.
On the contrary, the positive effect on community-college enrolment is driven by older students. As the minimum wage increases, students already enrolled in community college are less likely to drop out, and workers who separate from their job are more likely to return to community college to acquire additional skills. These results are consistent with the hypothesis that the minimum wage creates unemployment and increases competition in the labour market, which persuades individuals to return to school and acquire new skills.
While the literature has extensively analysed the disemployment effects of the minimum wage, little is known about the effects on education. This paper shows that minimum-wage policies could have important effects on individuals' lifetime income by influencing schooling decisions.
We focus on Canada for two main reasons. First, the frequency of minimum-wage changes is high. Canadian provinces set their minimum wages autonomously generating great variation across provinces and time. Further, among OECD countries, Canada has the highest percentage of individuals with a tertiary degree, with university and community college being equally popular choices.
Mismeasuring TFP and the myth of productivity shocks (E3, E2)
Changes in measured total factor productivity (TFP) are correlated with changes is real GDP. However, this may be due to measurement error that is correlated with business cycle or true cycles in unobservable actual productivity. Using a labor productivity series based on manufacturing workers from 1899 to the present, I am able to address this issue of measurement error which results in an acyclical productivity series, while measured TFP remains strongly cyclical. A simulated real business cycle model for the Great Depression with measured TFP shocks matches changes in Real GDP data during the Depression well, but my alternative productivity series sees no decline in GDP (and even a counterfactual increase in the early 1930s). These findings cast doubt on the relevance of productivity shocks in explaining business cycles.
Mitigating Credit Rationing: The Role of Development Banks in Loan Syndication (G2, O1)
Abstract: Using a novel dataset of development finance institutions (DFIs) worldwide, this paper examines the effect of the DFI participation upon the terms of syndicated loans. DFIs are government-supported financial institutions with an official mandate to pursue public policy objectives. The industrial policy view holds that DFIs provide long-term loans to investment projects that would not otherwise be funded by profit-driven commercial banks, whereas the political view contends that DFIs provides subsidized loans to firms crowding out private credits. But few has used loan-level data on a global scale to test which view holds water. To fill the gap, we test these arguments by conducting a study with a sample of 50,542 syndicated non-sovereign loans made to 11,058 borrowers in 76 countries during the period 1996-2016. To correct for the potential self-selection of DFI participation in loan syndication, we employ an endogenous treatment model. We also control for a large set of loan features, borrower characteristics, and country-specific macroeconomic and institutional variables. The results show that the DFI participation on average is associated with higher loan spreads, longer maturity, and smaller loan amount. Overall, our findings suggest that DFIs tend to provide high-cost credits to risky borrowers mitigating the credit rationing problem. We then further breakdown DFIs into four groups, namely, multilateral development banks as well as national development banks from high-income countries, middle-income countries and low-income countries. We examine the impact of the participation by different types of DFIs on loan terms and investigate the conditions upon which the effect at the aggregate level holds.
Modeling and Forecasting Serially Dependent Yield Curves (E4, C5)
Considering that yield curves usually are serially dependent, this paper proposes a new method to estimate and forecast yield curves based on factors driving serial dependence of yield curves. Gathering information at different lags of yield curves, the dimensionality and the lag order of yield curves are jointly determined. Applying this method to monthly U.S. government bond yields from January 1985 through December 2009, I find that the dynamic structure of yield curves reduces to a vector process lying in a 3-dimensional space, with 1-month lag information. Yield curve residuals from this new model over time exhibit zero mean and less autocorrelation. Moreover, this new model’s 1-month ahead forecasts outperform
those of all competitors including the dynamic Nelson–Siegel and random walk forecasts at all maturities.
Monetary Policy and the Mortgage Market (E5)
When a central bank changes the interest rate, it affects many households directly through their mortgage interest payments. If these households are constrained in their spending, this channel can have real and direct effects on aggregate demand. However, this channel is absent in standard frameworks of monetary policy. In such frameworks, changes in the policy rate affect consumption demand only via a forward-looking Euler equation. To quantify the mortgage interest rate channel, I build a heterogeneous-agent life-cycle model with housing and long-term mortgage contracts. The illiquid nature of housing gives rise to wealthy hand-to-mouth households, and the existence of mortgage financing allows for households to be both relatively poor and have high exposures to changes in the interest rate. I find that the aggregate response of consumption to a real interest rate shock is highly dependent on the type of mortgage contracts available and the possibility to refinance. In an economy with fixed-payment long-term mortgages, the response of consumption is 50 percent higher due to changes in mortgage interest rates and the endogenous response in house prices. However, in an economy with adjustable-rate mortgages, the consumption response is more than six times as large as compared to when fixed-rate mortgages are used. Hence, a detailed understanding of the contract structures in the mortgage market is an important input into the analysis of monetary policy.
Monetary Policy Financial Transmission and Treasury Liquidity Premia (E5, E4)
We identify and quantify the macroeconomic dynamic effects of well-identified monetary policy interest rate shocks on the yield curve due to changes in Treasuries liquidity premia. When the Fed raises interest rates, the spread between less-liquid assets and Treasuries of the same maturity and risk increases, as the liquidity value of holding Treasuries increases when the aggregate amount of banks’ customer deposits decreases. The longer the maturity, the smaller - but still significant - increase in the spread, as longer-term Treasuries are less liquid and more heavily discounted when interest rates rise. Monetary policy thus affects real interest rates through changes in liquidity premia across the yield curve.
The monetarist literature has emphasized the role of deposit fluctuations as proxies for various substitution effects of monetary policy when many asset prices matter for aggregate demand (Nelson, JME 2003). This paper identifies and quantifies the macroeconomic dynamic effects of substitutions between short- and long-term Treasuries, due to changes in liquidity premia, which occur because of the bank customer deposits' response to monetary policy interest rate movements. We thus identify and quantify monetary policy effects on the yield curve through relative liquidity premia.
Nagel (QJE 2016) documents, using static regressions, an observed positive relationship between the short-term Treasury liquidity premium and the Fed funds rate. When the Fed funds rate increases, the spread between non-liquid short-term assets and T-bills increases. Drechsler, Savov and Schnabl (QJE 2017) present a monetary policy transmission channel through the banking system, or deposits channel, which can explain Nagel’s empirical findings. The spread increases as the liquidity value of holding short-term T-bills increases when the aggregate amount of banks’ customer deposits decreases.
We use a macro SVAR model to quantify the effects of well-identified monetary policy interest rate shocks on the yield curve due to changes in liquidity premia. We show that the monetary effects on the yield curve via liquidity premia vary across maturities. When the Fed raises interest rates, the liquidity premium of longer-term Treasuries significantly increases, but by less than that of shorter T-bills, as they are less liquid and more heavily discounted when rates rise.
This points to a substitution from long- to short-term Treasuries, leading to a relative steepening of the slope of the yield curve, not related to policy expectations or risk premium, but reflecting liquidity premia reactions to monetary policy. When measured with Treasury yields, monetary policy interest rate actions thus have a proportionally larger effect on long-term rates than on short-term rates, as the liquidity premium of T-bills increases by more than the liquidity premium of longer-term Treasuries when the policy interest rate increases.
Our results can shed light on the expectation hypothesis, as monetary policy affects the term structure through liquidity premia. Moreover, Treasury liquidity premia fluctuations have recently been used to understand various issues like real equilibrium interest rate movements (Del Negro, Giannone, Giannoni and Tambalotti, 2018) or exchange rate forecasting (Engle and Wu, 2019). Our results contribute to quantifying the effect of monetary policy on those fluctuations.
Monetary Policy Surprises and Inflation Expectation Dispersion (E5, E3)
This paper documents that inflation expectation dispersion increases in response to monetary policy surprises in the United States. Relying on daily data of federal funds rate forecasts and inflation expectations at the analyst level from major financial institutions, we calculate monetary policy surprises of individual analysts as the unexpected changes in the federal funds rate during the two and a half days before the Fed meetings. We then assess the impact of these surprises on the dispersion of inflation expectations, which is based on the same analysts' inflation projections produced during the two and a half days following the Fed meetings. With an identification strategy that hinges on this tight window around the Fed meetings, we find that surprises in federal funds rate decisions generate dispersion of inflation expectations at short horizons and have no effect at longer horizons. To rationalize these results, we propose a partial equilibrium model with rational expectations and sticky information. We show that, when we allow the degree of information rigidity to depend on the realization of firm-specific shocks, the theoretical results are qualitatively consistent and quantitatively close to the empirical ones.
Monetary Policy under Data Uncertainty: Interest-Rate Smoothing from a Cross-Country Perspective (E5, D8)
Cross-country estimates of Taylor rules suggest that higher data uncertainty is associated with a more inertial behavior of interest rates. Data uncertainty is measured by the volatility of differences between real-time data and their revisions. Using a simple structural model with Kalman ﬁlter learning, I replicate the cross-country pattern of the inertial behavior. More inertial behavior results not because central banks gradually adjust interest rates in the face of data uncertainty, but because the central banks’ inference about the true data is correlated with past interest rates. Thus, I endogenize inertial behavior of interest rates as resulting in part from the learning process.
Monetary Policy with Endogenous Money Supply (E4, E5)
We develop a modified cash-in-advance model where we add asset market to goods and money market. In our model, money supply changes with households' borrowing decisions. Moreover, the model has two separate price systems; asset and general price systems. As a result, in our steady state analysis, while an expansionary monetary policy can tame asset prices, its effect on general price level is the same regardless of asset demand.
Monetary-Fiscal Policy Games and Uncertainty Shocks: The Role of Institutions (C7, H3)
The paper examines how rules and institutions as well as the monetary-fiscal coordination setup impact welfare outcomes of a reform during uncertainty shocks. We define uncertainty shocks as sudden events that create ambiguity about future course of economic policies chosen by policy makers as well as the possible responses of economic agents to the new policies. We examine uncertainty within a framework of interaction between several institutional factors: (1) the scheme of coordination between monetary and fiscal authorities, (2) credibility and reputation of authorities and (3) the degree of dependence on discretionary measures as opposed to commitments. Through a game theoretic model, we show that: first, a reform implemented under fiscal dominance accompanied with dependence on discretionary measures result in worst welfare outcomes that are magnified when uncertainty parameters are involved. Second, in the case of central bank independence and immunity to fiscal concerns, reforms signal fiscal discipline and produce desired outcomes. Third, proper fiscal rules and commitments under a scheme of benevolent and coordinating authorities are second best during uncertainty shocks and relatively lessen undesired outcomes due to higher credibility and lower time inconsistency. The game results are proposed for empirical examination on the Egyptian economy; a MENA-developing economy that has been subject to repeated reform attempts and several uncertainty shocks. In Egypt, there appears to be a continued setup of fiscal dominance paralleled with lacking efficient institutions and a high reliance on politically-motivated discretionary interventions that become apparent during political cycles and spells of political and economic uncertainty. We examine the above game through calibrating with Egyptian data during the period 1990-2020 using Non-Linear Structural Vector Autoregression. Counterfactual simulations are proposed to examine possible welfare outcomes under alternative schemes of monetary-fiscal setups as well as different scenarios of dependence on policy rules as opposed to discretionary measures.
Moving to Better Health Care? Estimating the Causal Impact of Medicaid Expansion on Homelessness (I1, H5)
Recent homelessness figures in the United States show a puzzling trend of aggregate decline but lopsided changes across states. This study examines the puzzle through healthcare access and provides the first causal evidence of the Medicaid expansion’s impact on homeless adults’ location. Using the state and county-level data on the homeless population from 2010-2017, the estimates from a difference-in-differences model show a significant 10.3 % post-expansion increase in homeless individuals per capita in states that adopted Medicaid expansion. Furthermore, utilizing the difference in homeless individuals’ coverage status vis-a-vis homeless people in families, estimates from a triple difference (DDD) model also confirm the post-expansion increase in homelessness in expansion states. This study contributes further by uniquely utilizing county-level data to provide subsample analysis on metropolitan counties and counties located at state borders. Results from the state, county, and border-county-discontinuity design reveal the evidence of homeless individuals’ migration from non-expansion to expansion states. Two mechanisms explain the migration process: post-expansion coverage eligibility of previously uninsured homeless individuals and the increased ability of homeless service providers in expansion states in offering healthcare and housing-related services. This paper concludes by measuring the state spending on Medicaid to demonstrate the implications of these findings on state welfare policymaking and fiscal expenditure.
Multi-Product Firms and Misallocations (D2, L1)
Shocks and distortions could affect the aggregate economy via the multi-product channel (Bernard, Redding, and Schott, 2010, BRS henceforth; Miniti and Turino, 2013; Jaef, 2018). This paper finds that (i) compared to the U.S., multi-products firms are fewer and smaller in China; (ii) firms with higher Hsieh and Klenow(2009, HK henceforth) taxes are less likely to be multi-product producers, controlling for the size effect. While the empirics confirm the existance of the product channel, our quantitative analysis using a discrete multi-product choice model with endogenous firm entry and exit suggests that this channel generates a small misallocation. This is because marginal firms that drop products due to distortions are of medium productivities. The output loss from the product extensive margin is small once the granularity of the firm size distribution is matched in the model. Most welfare loss come from the firm extensive margin and the product intensive margin.
Multiple Change Points Detection in High-Dimensional Data (K0, C1)
Regression analysis is widely used in economics for understanding the nature of relationships between variables. However, it is naïve to assume that the parameters are dynamically stable. The relationship between the predictor and the response variables may differ over time and space. The threshold point—where the parameters change—is called “change-point”. The detection of change points becomes particularly challenging in high dimensional data, i.e., when the parameters outnumber the observations. Although socioeconomic and marketing data are often high-dimensional, change point methods for high-dimensional models have received less attention in the literature. Many studies consider multiple change-points via grid search which can be computationally expensive, are unable to detect no-change cases, and are not extendable to high dimensional data. The current study extends the algorithm of Kaul et al. (2019) to detect multiple change-points via sequential binary segmentation based on L1/L0 regularization—where a change-inducing variable may switch the regression parameters in a dynamic setting. The model is able to detect zero to multiple change points of the parameters, and are applicable to high-dimensional data. We further compare the performance of our model in the detection of multiple change points with existing methods that use arbitrary segmentation and grid search. Simulation results indicate that our model is more stable, computationally efficient and has a lower average bias. For the empirical application, we predict community crime rates with socioeconomic predictors in the United States. We find that population, social security, and rental density respectively induce one, two and three change points in parameters. Thus, policies aimed at crime reduction should undertake separate measures across communities based on the estimated thresholds.
Mystery of Increasing Gender Imbalance of Tertiary Education in China: Causation and Influence (J1, O1)
In the past two decades, there was a mystery of increasing gender imbalance of tertiary education in China. Specifically, the more recently and in higher degree education, there was a much higher female students ratio, while considering the birth ratio of boys was higher than the one of girls. In this paper we combined the Becker Classic Discount Tradeoff and Risk Tradeoff based on the students’ heterogeneities into one mechanism to describe how the students would make the decision between entering job market and applying for further education, which is the engine. The students would consider the basic cost-return problem in Discount Tradeoff, and evaluate the potential risk to both two choices based on their heterogeneities including gender, capital, and grades etc. in Risk Tradeoff. We found that with three inputs from male’s wealth requirement of Chinese marriage custom, job market gender bias, and the increasing average salary of mid-educated jobs since entering WTO in 2001, the engine would release a result which is in accordance with the present gender imbalance situation. We conducted a three year junior undergraduate students survey on their choices between applying for further degree education and entering into job market, testified the deductions of our analysis. Finally we discussed two macro-level growth impacts from such a gender imbalance structure continuing in the long run, including: changing labor structure as a result of higher average education level of females than that of males would postpone average age of marriage and childbearing, and a huge waste of human capital accumulation due to many educated females have to abandon partial career to raise children while it is also beneficial to the education of next generation.
Oligopolistic Investment, Markups and Asset-Pricing Puzzles (D4, G1)
We analyze a multi-consumption good general equilibrium production-based asset pricing model with an oligopolistic sector that follows subgame perfect pricing, production, and capital investment strategies. The model is calibrated with U.S. aggregate and industry data from 456 manufacturing industries. The oligopoly model provides a better fit to product and asset markets' data compared to the benchmark competitive industry. In particular, under the "classical" assumptions of time-additive power utility and Markov shock structure, and assuming reasonable risk aversion, the model generates relatively high industry and aggregate equity premia and their volatilities, as well as Sharpe ratios. The model also fits well the volatility of industry investment and the cyclicality of price-cost markups. We find support for theoretical predictions on the link between industry competition and product and asset market outcomes.
Operating Leverage and Price Dispersion in the Airline Industry (G3, L9)
Over the past few decades, the U.S. economy has experienced a number of fundamental changes in its market structure including a sharp increase in market concentration and a steady rise of mark-up ratios (De Loecker et al., 2020). Since superstar firms can take advantage of substantial fixed cost investment to create barriers against start-up firms, they can expand market share and solidify market power. However, while high fixed cost structure or operating leverage (OL) can help firms gain competitive advantage and accelerate profitability during favourable times, it limits the firm’s flexibility to scale down operation when facing adverse demand shock. Despite the importance of OL on firm performance and stability during economic cycles, studies on its relationship with corporate policies remain scant.
In this paper, we examine how OL affects corporate pricing strategy in the airline industry. We observe dramatic changes in OL over time and across airlines. These significant movement and dispersion of the leverage enable us to isolate the intra-firm effect of OL on price dispersion, therefore alleviating the omitted variable problem caused by the heterogeneity of airlines attributes.
Based on a multi-product heterogeneous firm model, we predict that firms with high fixed costs discriminate prices more than firms with low fixed costs, in an effort to boost sales and improve profitability. Employing detailed pricing and financial statement data, we offer empirical evidence in support of this hypothesis. We also show that the price distribution of high fixed cost firms skews to the left, leading to lower average ticket prices. Market competition and financial leverage amplify the effect of operating leverage on price dispersion. We exploit an exogenous oil price shock to establish causality.
Operation Allied Force: Unintended Consequences of the NATO 1999 Bombing on Children’s Outcomes (I1, J1)
The goal of this paper is to estimate the causal effect of the NATO’s Operation Allied Force bombing of Serbia in 1999 on children’s outcomes. We examine in utero effect on children both in terms of short-term outcomes (birthweight), as well as long-term outcomes (grades of 15-year-old pupils). We use two different data sets for our analysis and we complement these with primary data on bombing. First, we use the population of birth record from the Statistical Office of the Republic of Serbia (SoRS). Second, we use the final examination dataset from the Serbian Ministry of Education (SMoE), which covers the whole population of pupils finishing primary school and contains data on pupil's grades. Our main identification strategy compares outcomes of children who were in utero during the NATO bombing (born between July 10 and October 15, 1999) (treatment group) with children born in the months June to October, 1998 (control group 1) and children born in the months June to October 2000 (control group 2). Results show that affected children had lower birthweight and lower grades in Serbian language and mathematics; no effect is found for behaviour. There are no differences in results between boys and girls. The negative effect of being in utero during bombing is largely driven by children in Belgrade. The paper contributes to the literature on short- and long-term effects of conflicts on future generations by shedding light on a conflict which has not been previously studied in the literature. One policy implication could be that governments need to intervene and design policies to alleviate the negative in utero effects on children in the aftermath of large-scale disasters.
Optimal Default Retirement Saving Policies: Theory and Evidence from OregonSaves (H3, G5)
I study the optimal default savings rate in automatic enrollment retirement saving plans. If individuals tend to procrastinate to make an active decision, the optimal default rate should be high to encourage people to opt out of the default. If individuals tend to actively undersave, the optimal default rate should be low to encourage people to stay at the default. Using exogenous increases in the default savings rate in OregonSaves, the first state-sponsored auto-enrollment plan in U.S., I identify individual adherence to the default rate. Using survey data from OregonSaves-eligible workers, I estimate the degree of undersaving if workers actively switch to a non-default rate. Combining individual-level administrative data with survey data, I suggest that the optimal default savings rate 7%.
Optimism Gone Bad? Persistent Effects of Traumatic Economic Experiences on Households’ Investment Decisions (E7, G5)
The long-enduring flourishing stock prices since 2010 have not increased stock market participation of German households. This is a puzzling observation since, according to Malmendier and Nagel (2011), households (i) use their life-time information on asset returns to make financial investment decisions and (ii) near events matter more for their current investment decisions than the farther ones.
In this paper, I show that the traumatic crash of the Deutsche Telekom shares in 2001 can explain the persistent non-participation of German households in the stock market. This event is unique due to its emotional character: the company focused on extensive brand advertisement with huge media coverage and rapidly established the reputation of their shares as the “people’s share”. Almost 2 million retail investors bought shares of Deutsche Telekom, where most of them have never invested in stocks before. Therefore, the disappointment of German households was huge as they experienced the fall of the shares in 2001, which was followed by corruption scandals of the company.
For my analysis, I closely follow the method of Malmendier and Nagel (2011) and investigate whether German households anticipate life-time experiences of asset returns for their investment decisions. To analyze the investment behavior of German households starting from the 1980s, I combine three German household survey data sets, which are the SOEP, SAVE study, and the PHF. In addition, I construct a variable that measures the exposure of German households to the Telekom event. Since the Telekom shares were highly represented in the media, I apply a sentiment analysis using financial newspapers and TV-coverage of the company.
Output-Inflation Trade-offs and the Optimal Inflation Rate (E5)
In staggered price models, a non-CES aggregator of differentiated goods generates empirically plausible short- and long-run tradeoffs between output and inflation: lower trend inflation flattens the Phillips curve and decreases steady-state output by increasing markups. We show that the aggregator reduces both the steady-state welfare cost of higher trend inflation and the inflation-related weight in a model-based welfare function for higher trend inflation. Consequently, optimal trend inflation is moderately positive even without considering the zero lower bound on nominal interest rates. Moreover, the welfare difference between 2% and 4% inflation targets is much smaller than in the CES aggregator case.
Paying over the Odds at the End of the Fiscal Year (H5, L2)
Governments often tend to increase their spending at the end of the fiscal year partially because they cannot roll over money from one year to the next. This is often considered a problem for policymakers due to the possibility of some of this spending being wasteful. This, however, represents an opportunity for firms to sell their services and goods at inflated prices. While there is some anecdotal survey evidence of civil servants in the United Kingdom, no paper has so far tested this problem empirically. We use a novel dataset composed of all Ukrainian government procurement auctions since 2015 to show that government suppliers raise their prices to profit from the governments' heightened willingness to spend, which creates substantial waste in government spending.
Peer effects in Fertility and Son Preference of China (J1, R2)
The increasingly unbalanced sex ratio in China and associated social challenges have been widely documented, though few studies have rigorously investigated the role that peer effects have played in this unbalanced sex ratio. This paper fills this gap by focusing on peer effects in the decision to have a second child, and to have a son. To identify peer effects, we separate out contextual and correlated effects that are known to hamper empirical studies on peer effects. The data we use comes from the 2016 data of China Family Panel Studies, and is a ten-year cohort of women aged 45-54 by 2016; we use a structural discrete choice model to estimate the peer effects. We find that peer choices significantly influence the probability that a family has a second child, but not the probability of having a son. Instead, having a son is largely driven by contextual effects, and in particular, by the education level of one's peer group. Our findings indicate that recent fertility incentives such as the two-child policy may generate spillover effects that encourage more families to have a second child.
Population Dynamics and Optimal Family Policies (E1, J1)
Many governments in the developed world are actively pursing policies to encourage aggregate fertility in the face of population aging. To study the optimal design of large-scale family policies, this paper builds a general equilibrium overlapping-generations model with heterogeneous agent, endogenous fertility and human capital investments. We find that (1) education policies and family policies go hand-in-hand, (2) family policies that are cost-effective in the short-run could be much more costly in the long-run, and (3) optimal policies achieving replacement fertility combines subsidized childcare and increased public education expenditures. We also propose the use of Reproduction Possibility Frontier (RPF) to evaluate the aggregate trade-off between output and fertility as well as its distributional consequences.
Portfolio instability and socially responsible investment: experiments with financial professionals and students (G4, G1)
Socially responsible investment (SRI) is regarded as a potential instrument to prevent behavioral biases and contribute to portfolio persistent growth. Moreover, SRI could be more resilient to ﬁnancial crisis and economic events. The stabilizing property of SRI has been overlooked by the literature on portfolio management, despite the fact that more stable portfolios are less costly in terms of human resources and incur lower transaction costs. We assess the eﬃciency and stability of socially responsible portfolios from the perspective of behavioral ﬁnance. Based on the data collected from incentivized experiments with 153 ﬁnancial professionals and 233 students, we compare a baseline treatment to a ranking treatment in which participants received feedback about their average investment in SRI assets. We found a signiﬁcantly positive inﬂuence of SRI on the level of portfolio stability and less need for portfolio rebalancing. Portfolios with a majority of SRI shares exhibited more stability in both treatments compared to conventional portfolios. Moreover, we observed that in the ranking treatment subjects invested more in SRI assets than in the baseline. In addition, the experiment revealed the convergence of professionals' and students' behavioral patterns.
Positive Future Thinking Predicts Stock Price Crash Risk: Evidence from S&P 1500, 1994-2015 (G4, G1)
In this study, I address the question of why stock markets are vulnerable to crashes (e.g., Hong & Stein 1999) from social psychological perspectives. I propose that positive thinking about the future of corporate elites predicts higher stock price crash risk. Research has shown that positive future thinking is associated with economic downturn (e.g., Sevincer et al. 2014) and I consider such psychological mechanism can be an invisible force behind stock price crashes too. I choose three domains of the mostly used and widely recognized big data sources for corporate America – 1) a large and structured set of corporate annual reports (‘10-K’), 2) a business intelligence dataset for financials and market information of 49 industries (‘Compustat’), and 3) the largest and comprehensive proprietary databases in stock market research (‘CRSP’) – to analyze the relationship between ‘positive future thinking’ and ‘stock price crash’. A crash risk is an unusually large, negative, and abrupt movement in stock prices that occurs without a correspondingly large event (see Hong & Stein 1999). The financial economists mainly have used three indicators of stock price crash risks and I also employ this convention too: (1) negative skewness in stock returns; (2) “down-to-up volatility” measure for the ratio of the negative and positive firm-specific weekly returns; and (3) firm-specific weekly returns exceeding three standard deviations below its mean value to generate a frequency of 0.1% in the normal distribution. I find that the results are consistent across different indicators of crash risks and time periods (i.e., from 1994 to 2015), and alternative econometric methods (i.e., industry effect, winzorization at 1 percent, and statistically non-significant reverse causality model specifications). In sum, I show that, at least from 1994 to 2015, it is detrimental to the financial health of corporate America, based on psychological research.
Predicting Authoritarian Crackdowns: A Machine Learning Approach (C6, N4)
We have developed a quantitative indicator to predict if and when a series of protests in China, such as the one that began in Hong Kong in 2019, will be met with a Tiananmen-like crackdown. The indicator takes as input protest-related articles published in the People's Daily---the official newspaper of the Communist Party of China. We use a set of machine learning techniques to detect the buildup in the articles of negative propaganda against the protesters, and the method generates a daily mapping between the current date in the Hong Kong protest timeline and the "as if" date in the Tiananmen protest timeline. We call this counterfactual date the Policy Change Index for Crackdown (PCI-Crackdown) for the 2019 Hong Kong protests, showing how close in time it is to the June 4, 1989, crackdown in Tiananmen Square.
Predicting Bank Distress in the UK with Machine Learning (C3, G2)
Using novel data and machine learning techniques, we develop an early warning system for bank distress. The input data comes from confidential regulatory returns, and our measure of distress is derived from supervisory assessments of bank risk in the UK from 2006 through to 2012. We contribute to a nascent academic literature utilising new methodologies to anticipate negative firm outcomes, comparing and contrasting classic regression techniques with modern machine learning approaches that are better able to capture complex non-linearities and interactions. We find the random forest algorithm outperforms other models on a host of performance metrics, including when varying the relative cost of false negatives (missing actual cases of distress) and false positives (wrongly predicting distress) for discrete decision thresholds. To improve the transparency of the random forest and examine the drivers of distress, we utilise state of the art machine learning interpretability and inferential techniques -- Shapley values and regression.
Preregistration and Incentives (D8)
Preregistering study designs is broadly supported as improving scientific credibility but criticized for limiting the scope of what can be learned. The paper investigates this tradeoff in a model where a researcher conducts a study and aims to convince an evaluator that the results are worth publishing. When both begin equally informed, the aim to publish is closely aligned with producing informative research, leaving preregistration redundant. When better informed, the researcher can credibly signal confidence by committing to a hypothesis. Thus, whether preregistration should be the norm in a field depends on how critically private information plays in designing studies.
Pretextual Traffic Stops and Racial Disparities in Their Use (K4, J1)
A moving-violation traffic stop is pretextual when it is motivated by suspicion of an unrelated crime. Despite widely expressed concerns that they infringe on civil liberties and enable discrimination against minority motorists; evidence on the use, frequency, and nature of pretextual stops is mostly anecdotal. Using nearly a decade's worth of traffic citation data from Louisville, KY, I find evidence suggesting that pretextual stops predicated on a particular moving violation---failure to signal---were relatively frequent. Compared to stops involving other similarly common moving violations, where arrest rates range from 0.01 to 0.09, stops involving failure-to-signal yield an arrest rate of 0.42. More importantly, pretext to stop a vehicle requires only one traffic violation. In stops involving failure-to-signal, the arrest rate is 0.52 when no other traffic violations are cited, and the presence of other traffic violations yields a 55% relative decrease in the probability of arrest. Relative to conventional traffic stops, black and Hispanic motorists account for a disproportionate share of likely pretextual stops. Yet, within likely pretextual stops, they are arrested at a significantly lower rate than other motorists. Following departmental adoption of body-worn cameras I find that the pretextual-stop arrest rate increases 33-34%, and the racial disparity in pretextual-stop arrest rate becomes much smaller and statistically insignificant. These findings are consistent with racially-biased use of pretextual stops prior to body-cam introduction, and more judicious and race-neutral use thereafter.
Preventing Child Maltreatment: Beneficial Side Effects of Public Childcare Provision (J1, I3)
We investigate the impact of childcare provision on cases of child maltreatment. For identification, we exploit a governmental reform introducing mandatory early childcare in Germany that generated large temporal and spatial variation in childcare coverage at the county-level. Using high-quality administrative data covering all reported child maltreatment cases in a county per year, our results show that maltreatment cases decline by 1.8% if a county increases childcare slots by one percentage point. This figure suggests that the expansion avoided approximately 20,000 maltreatment cases in our observation period of 2002-2015 compared with a scenario of no childcare expansion. As a mechanism, we find that the strongest reduction occurs in households in which a male partner or husband is present while childcare expansion has no effect in single-parent households. Because child maltreatment leads to enormous societal costs, we provide evidence that the provision of universal public childcare can prevent some of these costs.
Price Discounting at United States Land Grant Universities: A Supply-Demand Analysis (I2, L3)
We examine whether the 1862 land grant universities that do better in the U.S. News and World Report’s (USNWR) rankings really have the ability to charge higher tuition and offer less financial aid than institutions that do less well in the rankings. Developing a demand-supply framework to deduce relevant hypotheses, and drawing relevant data from multiple sources from 2005-2017, we find that parameters estimated using a generalized linear model (GLM) suggest each one unit improvement in national ranking is associated with an increase in (a) inflation adjusted in-state sticker price by 0.33% to entering undergraduates, (b) inflation adjusted out-of-state sticker price by 0.35% to entering undergraduates, and (c) inflation adjusted financial aid per undergraduate student by 0.33%. In addition, each one unit improvement in the USNWR ranking score is associated with more increase in the inflation adjusted out-of-state sticker price relative to its in-state counterpart across the1862 land grant universities.
Pricing Like the Competition: Excessive Tax Pass-through and Retail Prices in the Mexican Soda Market (H2, L1)
I analyze price adjustments following Mexico’s 2014 tax on sugar-sweetened beverages. First, I show evidence of tax over-shifting: in response to a one-peso tax, retailers increased prices by 1.32 pesos. I find that local competition partially limits over-shifting. Second, when adjusting prices, stores use a “catch-up” strategy where the price of cheaper products increases more than the price of more expensive products. Third, I find evidence of uniform adjustments to the tax at the store-chain level targeting modal prices across chains. Together, these results suggest that retailers facing more competition do in fact lower prices that are initially set by a store's chain and contradict the hypothesis that taxes are smoothly passed into prices.
Product-Level Trade Elasticities: Worth Weighting For (F1, F6)
Trade elasticity is a crucial parameter in evaluating the welfare impacts of changes in trade frictions. Its value varies widely across products, however, which is especially important for developing countries' evaluation of such welfare impacts. We estimate, and make publicly-available, trade elasticities at the product level by exploiting the variation in bilateral tariffs for each product category for the universe of country pairs over the 2001 to 2016 period. Homogenous elasticities lead to the underestimation of the welfare impact of trade, in particular for developing economies, and all the more so for those with high import penetration in less-elastic sectors.
Protesting Austerity: The Effect of Reform Legislation on Riots and Strikes (H3, K0)
Demonstrations and strikes are often expressive in the sense they occur regardless of any bargaining process. I construct a new data set of high frequency observations of new legislation voted in the Greek parliament and the daily incidence of protests, strikes and riots. Focusing on an era of fiscal adjustment, I establish the causal effect of austerity measures voted in the parliament on the incidence and length of protests, strikes, and riots across Greece. The incidence of protest, riots, and strikes rises sharply in the days prior to the voting day of an austerity bill in the parliament, while it dies out slowly in the days following the vote. I find that around half of all protests, strikes, and riots associated with a particular austerity bill take place in the days following a parliamentary vote. Most of the strikes associated with austerity measures are organized by federation unions. Strikes by single-corporation unions are only 17 percent of the total strikes or 11 percent of total protests, indicating that expressive motivations alone explain the majority of the strikes and protests during a period of fiscal contraction. The results suggest that union members obtain expressive utility from striking and protesting, and that the protest choice maximizing their expressive utility is their actual protest choice. Last, I proxy the length of structural reform prescribed by an austerity bill by the number of pages in the bill. I find strong and positive association between the pervasiveness of an austerity bill--proxied by its length in pages-- and the incidence of protests and riots in the days leading up and immediately following its vote in the parliament.
Quantitative Easing versus Laissez Faire in a Low Interest Rate World (E5)
In a liquidity trap with the interest rate at its zero lower bound, the effectiveness of quantitative easing (QE) depends on the interest-rate sensitivity of the risk-averse firms’ expected net investment return relative to its “certainty equivalent”, and the “expected investment return-real interest rate” boundary bifurcates the slope of the aggregate demand curve into an upward sloping portion (low investment return) and a downward sloping portion (high investment return). While QE can help turn lower real interest rates into greater investment if the expected investment return is indeed lower than its rate-sensitive boundary value, QE could also backfire and trigger a deflationary spiral under the opposite condition. I show that when the expected investment return is sufficiently large relative to the prevailing real interest rate, it is laissez faire rather than QE that turns the disinflationary force into investment demand and therefore breaks the deflationary spiral. With a higher expected investment return, the bifurcation point of aggregate demand curve moves with lower inflation and higher output, and such a favorable trade-off opens more opportunity for laissez faire than for QE.
Racial Gaps in Financial Outcomes (J1, D1)
Even amidst strong macroeconomic conditions, families experience high levels of income volatility that have important implications for well-being. Families with limited liquid assets are dramatically less likely to smooth consumption in the face of income fluctuations, and it stands to reason that racial gaps in liquid assets could result in racial differences in consumption smoothing. In this report, the JPMorgan Chase Institute uses administrative banking data to study racial gaps in liquid assets, take-home income, and families’ consumption response to income volatility from the vantage point of a novel de-identified data source: administrative banking data paired with self-reported race information from voter registration files. We find large racial gaps in take-home income and liquid assets which persist across age, income, gender, and geographic segments. Additionally, we find racial differences in consumption smoothing. Compared to White families, Black and Hispanic families exhibit sharper drops in spending after involuntary job loss and larger increases in expenditures after the arrival of the tax refund. However, these racial differences in consumption smoothing are explained by racial gaps in liquid and financial asset buffers. Taken together, our results shed light on the distributional impacts and importance of efforts to reduce financial volatility and increase liquid assets for low-income families and address the structural factors that contribute to racial gaps in income and assets.
Racial Gaps in the Early Careers of Two Cohorts of American Men (J1, J7)
This paper studies how the early career racial gaps and the underlying driving forces have changed across two cohorts of young American men (NLSY--79 and NLSY--97). Tracking Black and white men from early adulthood to their mid-30s, I first document cross-cohort changes of labor market trajectories over the first eight years post-schooling. Exploiting the richness of individual and family characteristics in the NLSY data and using the restricted geocode files to construct neighborhood measures, I then use a semi-parametric decomposition to examine the explanatory power of 1) education and skills, 2) family background, 3) childhood neighborhood, and 4) school-to-work transition to the observed racial gaps in labor market outcomes, and how the pattern has changed across cohorts. I establish three main results. First, measured racial differences in education and skills, especially cognitive skills, are crucial in explaining racial labor market gaps in the younger cohort as well as in the older cohort, despite a broad convergence of racial skill gap across cohorts. Second, Black men initiated their careers with an inferior local labor market condition and a worse employment outcome. In the younger cohort, this Black disadvantage in transition explains an important share of the racial labor market gaps measured six to eight years later. Third, the explanatory power of childhood neighborhood is small or negligible after conditioning on racial differences in family background and individual skills.
Racial Segregation and Exclusionary Desire (R2, J1)
Massive wishful integration intervention could trigger white ight and thus exacerbate residential segregation rather than promote racial integration. Yet, there is little large-scale empirical evidence of tipping. This paper proposes a rened holistic estimation scheme to explore the breadth and nature of tipping. There are two major innovations in this identication strategy. First, the revealed integration preferences are elicited from a random coecient discrete choice model. Second, the restrictive assumption of common tipping point is relaxed, which allows an in-depth investigation of tipping among neighborhoods of dierent preferences.
This paper nds that, among households exhibiting moderate exclusionary desire, there is substantial evidence of tipping (around the minority ratio of 12%), which is also documented in Card, Mas & Rothstein (2008). The evidence of neighborhood tipping warrants caution, when designing racial integration policies.
Contrary to expectation, no evidence of tipping is found among neighborhoods with strong exclusionary desires. These neighborhoods are immune to tipping, because the black home seekers are tacitly kept out, which keeps the minority ratio below tipping points. Policymakers must be mindful of the concealed tipping points.
Reference point adaptation and air quality – Experimental evidence with anti-PM 2.5 facemasks from China (Q5, D9)
The formation of reference points has drawn increasing interest ever since the introduction of prospect theory. Given that most studies focus on tradable goods such as stocks, for which the prices are observable, few have focused on environmental goods. This paper attempts to fill this gap in the literature in this regard.
In our experiment, we divided the subjects into buyers and sellers and asked them to trade four PM 2.5 filters using the Becker–DeGroot–Marschak auction (BDM auction). We have two treatments in this experiment: a) the experience of seven weeks of heavy air pollution; and b) the receiving of information on the relationship between death rates and air pollution. The different bidding prices for the four PM 2.5 filters in these treatment groups make it possible to trace the adjustment of the reference points as a result of these treatments without having to know their precise values.
Our results show that, for buyers, the heavy air pollution drives them to fully downwardly adjust their reference points on air quality. For sellers, however, the reference points adaptation caused by heavy pollution is not a full adaptation. Moreover, the new information on the damage to health from air pollution causes buyers to upwardly adjust their reference points on air quality but does not significantly change the sellers’ reference points.
We show that, for both treatments, sellers are more reluctant to adjust their reference points on air quality than are buyers. Our results confirm the asymmetric reference point adaptation in that adaptation after a loss is harder than adaptation after a gain.
Relative Wealth Placement and Risk-Taking Behavior (D8, D9)
This study provides evidence that relative wealth placement, i.e., an individual's position in the wealth distribution relative to a peer group, substantially impacts risk-taking behavior.
We first develop predictions on risk-taking in a theoretical setup where utility does not only depend on consumption, but also on placement in the wealth distribution. In a labratory experiment, we present individuals with a simple investment decision with constant absolute wealth, but varying relative wealth placement. Individuals invest up to 50 percent more depending on their relative placement.
More specifically, we find that individuals placed at the bottom (top) in the relative wealth distribution exhibit more (less) risk-taking.
Revisiting the Optimal Inflation Rate with Downward Nominal Wage Rigidity: The Role of Heterogeneity (E5, E2)
In this paper, I study the optimal inflation rate in a sticky price economy in which workers are heterogeneous in labor productivity and wage changes are subject to asymmetric adjustment costs. The model calibrated to U.S. micro wage data implies downward nominal wage rigidity (DNWR). The optimal inflation rate is substantially higher than stated in the literature in the presence of worker heterogeneity. A key to understanding the result is that DNWR causes an inefficient cross-sectional allocation of labor as well as inefficient aggregate dynamics, enlarging the "grease the wheels" effect of inflation.
Right to Work Laws and Total Worker Compensation: Evidence from Synthetic Control (J5, J3)
There are currently 27 states with right-to-work (RTW) laws. However, identifying their effects is complicated by the simultaneity between passing RTW laws, economic conditions, and union sentiment. I address these issues by exploiting variation in timing of the RTW law adoption and using synthetic controls for a pooled sample of states. I estimate the effect of RTW laws on fringe benefits, union membership, and hourly wages in the private sector using a pooled synthetic controls methodology for the years 1990 to 2020. While there has been ample research on how RTW laws affect union membership and wages, there has been no analysis on how the laws may impact total worker compensation. This study extends prior research by estimating the impact of RTW laws on employer pension plan and health insurance participation in the private sector. I find that RTW laws decrease union membership, pension plans and hourly wages, but increase employment. I then conduct a heterogeneous analysis by industry and find that these effects are mostly concentrated in the construction industry. Since the effect varies by industry, policy makers should consider the industry composition of their state economy before adopting RTW laws.
Rising Cohabitation and Child Development (E6, E2)
Cohabitation rates of couples without children have steadily increased in the U.S. over the past 50 years. Yet, cohabitation rates of couples with small children have only increased for the less educated. What explains this differential rise in cohabitation rates by education and what are the implications for child investment and child outcomes? We show empirically that cohabiting women experience smaller childbirth penalties, work more in the labor market, and spend less time with their children as compared to married women. Subsequently, their children are less likely to obtain a college degree. To rationalize these facts, we build an overlapping generations model of marriage, cohabitation, and child development. Parents are altruistic towards their children and invest time and goods into their development. This, in turn, increases the probability that a child completes college. Couples can choose to separate in every period but married couples pay a divorce cost. Assets are split equally between spouses if couples were married prior to separation, but not if spouses previously cohabited. The model matches differences in hours worked, time, and money invested in children between married and cohabiting women. A comparison of the 1975 and 2015 steady states reveals that changes in the gender wage gap and the college premium are important drivers of the rise in cohabitation among less educated women with children over this period.
Road Pricing and Vehicle Usage: Evidence from the Gothenburg Congestion Charge (Q5, D6)
This paper uses household-level microdata to estimate changes in car-owning and driving decisions in response to the introduction of a congestion charge in the second largest Swedish city. Applying difference-in-differences methods, we find that the charge is associated with a modest reduction in car-owning rate and passenger-kilometers driven. During the first three years after implementing the Gothenburg congestion charge, the car-owning rate reduces by 0.82 percentage points. Furthermore, the average annual mileage decreases by about 130 kilometers per car-owning household, representing a one percent reduction in car usage. As wealthier households are more likely to own and drive cars, our preliminary results show that a congestion charge expands the gap in private passenger car usage between the rich and the poor households. Based on the estimate that the total amount of driving is lowered by 16 million kilometers annually, we also calculate the environmental and economic benefits arising from pollution and congestion avoided.
Robust Inference in Time-Varying Structural VAR Models: The DC-Cholesky Multivariate Stochastic Volatility Model (C1, E5)
This paper investigates how the ordering of variables affects properties of the time-varying covariance matrix in the Cholesky multivariate stochastic volatility model. It establishes that estimated covariance matrices, obtained under alternative orderings of variables, are systemically different when the data exhibits idiosyncratic volatility dynamics. Specifically, simulations show that estimated covariances and correlations become more different, the larger the ratio of individual volatility paths becomes. This paper shows that this property is important for empirical applications as alternative estimates on the evolution of U.S. systematic monetary policy and on inflation-gap persistence indicate that conclusions may hinge on a selected ordering of variables. The dynamic correlation Cholesky multivariate stochastic volatility model is proposed as a robust alternative.
Role of Inadequate Housing in the Spread of COVID-19 in the United States (R2, I1)
We use the national American Housing Survey data from 2017 (latest available) on housing adequacy index, per capita square feet and other physical neighborhood quality variables , such as lack of adequate plumbing, at the MSA level to examine whether MSAs that had inadequate housing also generally showed higher confirmed COVID-19 cases compared to MSAs with adequate housing. Daily COVID-19 confirmed data is obtained from the John Hopkins and CDC. Controlling for detail demographic and economic factors we also estimate whether the neighborhood lacking adequate housing showed higher confirmed COVID-19 cases post strict shelter at home is imposed at the national level. We will also use the staggering strict shelter-in-place, stay at home and 6 feet social distancing by states to see if there are any significant increase in the confirmed COVID-19 cases immediately after the local restrictions are in place in neighborhood with inadequate housing. Findings from this paper will shed light on the role of inadequate housing on a pandemic like COVID-19. This paper will also show whether there are possible negative effect of sheltering and strict stay at home for certain neighborhoods.
Screening Loss-Averse Consumers (D8, D9)
We study optimal pricing strategy of a monopolist who faces consumers that have heterogeneous private tastes, have reference-dependent preferences, and are subject to loss-aversion. There is asymmetric of information and monopolist does not observe the consumers’ valuations. Assum- ing that the monopolist can make consumers expect to buy the desired variety of the good, and that these expectations determine the consumers’ reference points, we obtain two main results. First, with expectation-based loss aversion, menu pricing is possible even if the single-crossing property is violated (high-valuation consumers do not have a larger marginal utility of quality than low-valuation consumers). Second, when firms face consumers with expectation-based loss aversion, menu pricing may become more desirable to the monopolist compared to selling only to high-valuation consumers.
Shared Ride-Hailing and Tip Payments: Evidence From Chicago (R4, D6)
Popular services such as Lyft Line and Uber Pool allow customers to authorize sharing their ride with a stranger. However, a portion of rides in which sharing was authorized end up as solo rides. I exploit the good-as-random assignment of these solo rides to identify the effect of solo rides on tip payments, both at the intensive and extensive margins. To deal with selection into treatment I condition on ride time and place, fare level, and a set of date-time fixed effects. Using ordinary least squares weighted by the propensity a given rides ends up shared, I find that when pooling-authorized rides end up unshared, drivers enjoy a tip premium of about 6.46%. I build upon previous research by examining the reasons for this gap, finding that shorter travel times for solo rides explain some but not all of the tip premium. Riders respond to less convenient shared rides by tipping lower and less frequently, even if the reduced convenience is beyond the driver's control. Results are robust across different fare levels and to tests for bias due to unobservable omitted variable bias.
Short-Run Equilibrium of International Trade under Heterogeneous Discrete Firms with Multiple Continuous Varieties (F1, C3)
This study demonstrates the impact of international trade on the lowering markups of multiple varieties produced by a discrete firm differentiated in productivity. Market liberalization promotes head-to-head competition among productive heterogeneous firms. Therefore, in this pro-competitive market, productive firms would survive by reducing markups and adjusting their range of products, while inefficient firms fail to survive. Despite the importance of this topic, most trade studies employing firm-level granularity and heterogeneity mute a change in markups in the short-run, while they identify properties in the long-run equilibrium in terms of prices, a range of varieties, and profits. With an assumption of fixed firm-level productivities across symmetric economies in the short-run, market liberalization introduces head-to-head competition. In this pro-competitive market, productive firms survive by pricing competition with exporters with symmetric productivities. In the quantitative analysis, the results show that the survivors in the liberalized market lower their markups and change the range of varieties in the short-run equilibrium.
Simple Mechanisms for Non-linear Agents (D0)
We consider auction design for agents with non-linear preferences. We quantify the extent to which posted pricing approximately optimizes welfare and revenue for a single agent. We give a reduction framework that extends the approximation of multi-agent pricing-based mechanisms from linear utility to non-linear utility. This reduction framework is broadly applicable as Alaei et al.  have shown that mechanisms for linear agents can generally be interpreted as pricing-based mechanisms. We give example applications of the framework to oblivious posted pricing [e.g., Chawla et al., 2010], sequential posted pricing [e.g., Yan, 2011], and virtual surplus maximization [Myerson, 1981].
SNAP Participation and Labor Supply Decisions: A Dynamic Structural Modeling Approach (I3, J2)
The Supplemental Nutrition Assistance Program (SNAP, or Food Stamps) is an essential federal assistance program in the United States. It guarantees financially strapped households with supplemental (in-kind) income to purchase grocery food. In this manner, SNAP acts as an automatic stabilizer for the macroeconomy. Previous reduced form studies establish the negative impact of stricter work requirements on SNAP participation, but, present tenuous evidence of changes in SNAP policy on labor supply margins. Furthermore, for a sub population and we do find that SNAP has a near-linear relationship between labor income and benefit amounts: for each additional dollar of labor income, SNAP households lose $0.30 of benefits. We explore the dynamic interplay between participating in SNAP and labor supply decisions by designing a structural dynamic discrete choice model. The model utilizes changes in the SNAP policy rule regarding work and the corresponding eligibility requirements such as asset tests. We evaluate how these policy structures impact a SNAP participant’s decision to work or not and calculate the welfare of the baseline system with alternative policies to encourage SNAP participants to work. Our model is calibrated to the U.S. data using the NLSY’97 working age cohort measures the welfare in the baseline economy under the pre and post SNAP poly rule change in 2008. We then model and discuss how several proposed changes to SNAP (namely, changes in work requirement parameters and the earnings deduction ties to unemployment insurance) would impact labor supply decisions.
Social Networks and Corporate Social Responsibility (G3, D8)
I show that corporate social responsibility (CSR) spreads through the social networks of firms' directors. This result is obtained using a novel identification strategy exploiting the imperfect overlap between industry, geographic and social peers, a diff-in-diff relying on directors' deaths, and a regression discontinuity design based on CSR proposals. Social network effects are concentrated in firms pursuing product differentiation strategies for which CSR is more likely to add value, firms strategically positioned in the social network to acquire valuable information, and firms in which the incentives of managers and shareholders are aligned. This suggests that some firms aim to create value by using social networks as a market for information exchange on CSR. I find little evidence for alternative explanations such as social norms and agency problems.
Softening the Price War: An Information Design Approach (L1, D8)
This article generalizes the model of product differentiation by introducing an additional information disclosure process. A mass of consumers decide to purchase a unit of product from one of two costless manufacturers. Whether each product matches each consumer is independently identically distributed among all product- consumer combinations. Two manufacturers set their prices simultaneously. The market maker, whose objective is to maximize the total revenue of manufacturers, commits to a disclosure policy publicly before the beginning of price competition. This article shows that each information disclosure policy is one-to-one mapped to a consumer distribution in Hotelling model with linear transportation costs. Without information disclosure, the model is reduced to a Bertrand competition, which leads to a severe price war. Under some information disclosure policy, the model is reduced to a Hotelling competition with uniformly distributed consumers, which leads to a mild price war. This article finds out the optimal information disclosure policy that maximizing revenue of manufacturers. The maximized revenue is twice as many as equilibrium revenue in Hotelling competition with uniform distribution. This article also studies several extensions by relaxing some assumptions in benchmark model and shows the robustness of main results. This article implies that market maker can soften the war competition through two equivalent channels, public disclosure and customer subsidy.
Sowing the Seeds of Financial Imbalances: The Role of Macroeconomic Performance (C3, G2)
The seeds of financial imbalances are sown in times of buoyant economic growth. We study the link between macroeconomic performance and financial imbalances, focusing on the experience of the United States since the 1960s. We first follow a narrative approach to review historical episodes of significant financial imbalances and find that the onset of financial disturbances typically occurs when the economy is running hot. We then look for evidence of a statistical link between measures of macroeconomic conditions and financial imbalances. In our in-sample analysis, we find that strong economic growth is followed by a build-up of financial imbalances across all dimensions of the National Financial Conditions Index. In our out-of-sample analysis, we find that the link between strong economic performance and increases in non-financial leverage is particularly strong and robust. Using a structural VAR identified with narrative sign restrictions, we also demonstrate that business cycle shocks are important drivers of non-financial leverage.
Specialists or Generalists? Cross-Industry Job Mobility and Occupational Wages (J6, J3)
This paper investigates low cross-industry job mobility of low and middle-wage occupations, and
its consequence, higher industrial concentration of employment(HHI), as a new factor contributing to wage growth suppression in low and middle-wage american labor markets. I first present empirical evidence documenting the negative effect of HHI on workers in low and middle-wage occupations. To study the policy implications of these findings, I use a general equilibrium two sector on-the-job search model simulating the effect of cross-sectoral job mobility/skill transferability on industrial concentration of employment and on the distribution of wages and wage-productivity gaps across and within industries. The model confirms the empirical observation that an increase in between-sector job mobility rates for workers in the less productive sector not only increases wages in that sector but in all sectors. It additionally decreases both aggregate wage inequality and the wage-productivity gaps in all industries. Drawing on the model’s results, I assess the effect of Career and Technical Education(CTE) Programs—an educational policy encouraging the training of "career-oriented" skills transferable across industries—on the HHI index and wages. I find that a one standard deviation increase in CTE enrollment decreases
industrial concentration of employment, and ultimately increases wages by 8.3 percent. My results
suggest that policies fostering occupational skill transferability and thus cross-sectoral mobility may be welfare improving for workers in low wage occupations.
Spillover Effects of the United States and the Western Economic Sanctions against Russia on other Transition Economies (P2, F4)
A review of the literature on international sanctions suggests a growing need in the assessment of economic impact the sanctions may have on third, not involved in the conflict, countries. This study evaluates the medium-term economic impact of multilateral, imposed by Western economies, and unilateral, imposed by U.S., sanctions against the Russian Federation on transition countries of the former Soviet Union, and Central and Eastern Europe. To address the question two gravity models- bilateral trade and direct investment flows- are developed covering the period from 2014 to 2018. We construct the variables of sanctions against Russia and use the Poisson pseudo-maximum likelihood econometric technique to run the analysis. This paper demonstrates that sanctions against a large open economy may have negative medium-term spillover effects on third countries. We recommend to evaluate the economic risks that third economies bear from sanctions and to supply them with specific mechanisms of mitigation when imposing sanctions against target countries.
Staying Ahead of the Curve XR Futures Foresight & Unified Field Theory: Combining the Econometrics of Network, Spatial, and Panel Data, with XR Immersive Experience to See New Solutions to the Global Problem of Ocean Plastic Pollution (Q5, B5)
Solving global problems requires staying ahead of the curve. This paper presents a critique of current approaches using spatial, panel, and network data for economic, social, and political decision-making under risk, and develops an alternative model called XR futures foresight and unified field theory. Currently, a single theoretical framework (i.e. a unified field theory) does not exist that can account for the fundamental forces of nature and relativity which influence how global problems such as ocean plastic pollution, or global pandemics like COVID-19 are addressed. Here, we describe a unified field framework that permits us to use pairs of physical and virtual fields for futures foresight. Physical fields are defined by numeric values where there is a value for each point in space and time as captured in spatial, panel, and network data. Virtual fields are temporary changes around a point in space that can help us to understand relative changes when we change our frame of reference in space and time. Using the global problem of ocean plastic pollution as a proxy,we examine how network, spatial, and panel data can be paired with virtual data under unified field theory.We apply the theory to a proactive XR (virtual,augmented,and mixed reality)futures foresight approach which enables individuals and organizations to test how reflexive and prepared they are to address alternative futures.This research introduces an XR futures foresight framework which combines virtual, augmented, and mixed reality technology with spatial, panel, and network data.We find that pairs of physical and virtual fields allow individuals “not so much to see what nobody has yet seen, as to think what nobody has yet thought, concerning that which everybody sees”(2)and enables them to make more informed local agreements and pre-commitments to better address global problems today.
STEM Women and Gender Pay Gap in IT Career (J2, J3)
With the rapid advancement of high-tech industries, it is intriguing to understand whether the long-lasting gender pay gap issue still sustains in high-skilled jobs. Jobs in the information technology industry, conventionally perceived as male-dominant, now are more favorable for women under the technological and organizational change. Prior evidence shows that the pay gap may be alleviated when women can gain similar STEM skills as their male counterparts. Extended from this stream of literature, we adopt a large-scale archived behavioral data from an online recruitment platform to explore gender pay gap with STEM ability in the matched job application-position pairs. Our results show that in IT careers women tend to show lower expected wage than males. Having STEM ability can raise their self-confidence in expected wage. But gender disparity exists in the pay gap between employer and job applicant, which implies women need more wage bargaining than man . The increase in pay gap between the employer and the job applicant is only for STEM women . We further construct determinants of matched skills to explore heterogeneous effects of women’s self-confidence and under-representation on the gender pay gap. We find that a higher level of skill match leads to an increased female representation in IT career . The results are robust across different settings.
Student Loan Supply, Parental Saving and Portfolio Allocation (D1, I2)
I show that an expansion of student loan supply affects parents' saving decisions and portfolio allocation. By exploiting policy-induced variation on expected student aid, I find a sizable increase parental saving rate. The mechanism that drives this result is the positive effect of student aid on students' college enrollment. Consistent with this interpretation, I find a disproportionate increase in college enrollment for children of families affected by the reform. The positive saving response is largest among lower- and middle-income families, in areas with higher average college expenses and for parents with strong saving preferences. A placebo test validates that the effect is absent in families without children. Moreover, I show that affected parents shift the allocation of saving flows towards riskier assets.
Tax Avoidance through Cross-Border Mergers and Acquisitions (H2, F2)
We document a novel tax avoidance strategy: Cross-border, tax-haven mergers and acquisitions (M&A). Tax havens have $2.4 trillion in M&A deal value beyond what is predicted based on economic fundamentals. Cross-border, tax-haven M&As result in $24.7 billion in recurring annual tax avoidance, and cross-border, non-haven M&As result in an additional $31.3 billion in recurring annual tax avoidance. This is the first paper to document that tax havens affect real investment on a large scale, and not just capital flows on paper. Moreover, we create an algorithm, which is available to others, to derive the tax residence of any company given data on the firm's country of incorporation and headquarters.
Terrorist Pasts, Criminal Futures: the Evolution of Paramilitary Violence in Northern Ireland (R1)
After the UK voted to leave the EU in June 2016, the avoidance of a hard border on the island of Ireland quickly became an area of focus and sensitivity because of the border's historical controversy. The Troubles in Northern Ireland was a brutal ethno-nationalist conflict that erupted between 1968-1998, between the Protestant unionists, who desired the province to remain part of the UK, and the Roman Catholic nationalists, who wanted Northern Ireland to be reunited with the Republic of Ireland. Although Northern Ireland remains in peaceful times today, it is still a deeply divided and fractured society. with significant spatial inequality. Using spatially detailed crime data, this research project seeks to explore the evolution of past terrorism into the modern organized criminality that we see operating across Northern Ireland today.
The Consequences of Sorting for Understanding School Quality (J2, I2)
I study the sorting of students to school districts using new lottery data from an inter-district school choice program in Massachusetts. I find that moving to a more preferred school district increases student math scores by 0.19 standard deviations. The program also generates positive effects on coursework quality, high-school graduation, and college attendance. Motivated by these findings, I develop a rich model of treatment effect heterogeneity and estimate it using an empirical-Bayes-type procedure that leverages non-experimental data to increase precision in quasi-experimental designs. The estimator I propose is a weighted average of experimental and non-experimental variation, with the weights chosen according to the correlation of the heterogeneous effects across samples. I use the heterogeneous effects to examine Roy selection into the choice program. Students who would be negatively impacted by the program are both less likely to apply and, conditional on taking up an offer to enroll, are more likely to subsequently return to their home district. I find that this selection drives almost all of the program evaluation treatment effect identified with the lottery. The fact that families sort students to school districts according to potential benefit suggests that research relying on school choice lotteries to learn about differences in school quality may lack a broad claim to external validity.
The Deadly Connection between Hurricanes and Sinkholes: Analyzing Market Responses to Multiple Environmental Risks (Q5, R3)
Real estate purchase is a major transaction for many households in the USA and the price of real estate properties may or may not reflect a wide range of associated environmental risks. Earlier hedonic studies consider one or the other type of these risks and analyze how that affect the property values. But in this study, we consider both geological (sinkholes) and hydro-metrological risk (hurricanes) and analyze their impacts on property values. We are particularly interested in learning if the information revealed through one type of risk triggers the impact of another type of risk. With that objective, we focus on Lake County in Florida which have large number of identified sinkholes and have also experienced a recent hurricane. We investigate how the proximity of sinkholes affects the housing price using real estate sales data of almost 35000 single family homes in Lake county from 2014 to 2018. Given that the area has experienced significant damages during a recent hurricane (Irma in 2017) and the state of Florida has instituted a change in sinkhole insurance law (2016), it allows us to exploit these variations in a quasi-experimental setup and use difference in differences technique and regression discontinuity design to establish causal relationships. We find that houses that are near to known sinkhole locations experience a significant price discount, which is amplified following a hurricane event. We also analyze the effect of the new sinkhole insurance law on housing prices in Florida. We find that houses that are located close to known sinkhole locations face a subsequent price discount due to reduced sinkhole damage protection offered by the new insurance law. Thus, an institutional change has led to a similar effect as a natural hazard event for sinkhole proximate real estate properties in the housing market in Florida.
The Dollar and Corporate Borrowing Costs (F3, G2)
We show that an appreciation of the U.S. dollar raises syndicated loan spreads for U.S. borrowers, even for those with no trade exposure. We identify the effect of dollar movements using spread and loan amount adjustments during the syndication process. Using this high-frequency, within loan variation, we find that dollar movements affect loan terms and that these effects are concentrated in appreciations. A one standard deviation increase in the dollar index increases spreads by up to 16 basis spreads and reduces loan amounts and underpricing by up to 3.6 percent and 6 basis points, respectively. The dollar therefore transmits global shocks to corporate borrowing costs in the United States, suggesting that the dollar is a strong indicator for the global demand for risky assets.
The Dynamic Incidence of the Corporate Income Tax (E6, H2)
After Harberger published his influential paper in 1962, many authors have assessed empirically whether the incidence of the corporate income tax (CIT) falls on capital owners, consumers, or workers (Krzyzaniak and Musgrave, 1963; Gordon, 1967; Arulampalam et al., 2008; Fuest et al., 2018). Today, there is little agreement among economists about who bears the incidence of the CIT (Gruber, 2007; Harberger, 2008a,b; Wagner et al., 2018; Auerbach, 2018). The reason for the little convincing evidence is that the econometric models used in the literature ignore that the factors that motivate changes in corporate tax policy are sometimes correlated with other developments in the economy and disentangling those dynamic effects from exogenous policy changes requires tremendous effort.
Using annual information at the industry level for the United States, this paper estimates impulse response functions in a structural vector autoregression (SVAR), with residual-based moving block bootstrap (Jentsch and Lunsford, 2019; Merterns and Ravn, 2019), to study the consequences of exogenous changes in corporate tax policy. The identification of these exogenous events follows Romer and Romer (2009, 2010) by using narrative accounts from 1950 until The Tax Cut and Jobs Act of 2017. The results validate Harberger’s original predictions. That is, in the short-term, capital owners bear the full burden of the tax. Over time, however, capital owners can shift this burden either by raising consumers’ goods prices or decreasing workers’ wages. The magnitude of these effects depends on the degree of capital intensity as well as the access to international markets for the industry under consideration.
The Effect of Aging Out of WIC on Food Insecurity (I1, I3)
The Special Supplemental Nutrition Program for Women, Infants, and Children (WIC) and the National School Lunch Program (NSLP) are designed to increase food security and reduce hunger for children from low-income households. Since the cutoff age for WIC is five, and school enrollment is required for receiving free or reduced-price NSLP, some children from low-income households cannot receive both WIC and free or reduced-price NSLP. Using data from the current population survey, the partial identification method developed in this paper addresses the problems of self-selection into WIC and systematic underreporting of program participation. Due to this loophole in food assistance programs for children, aging out of WIC is found to increase child food insecurity by at least 1.1 percentage points. This result indicates that the prevalence of child food insecurity would decline by 15 percent if WIC extended its cutoff age until children enroll in kindergarten.
The Effect of COVID-19 on Price Dispersion in China’s Online Market (L1, O1)
In this paper, we investigate the effect of the shock of COVID-19 and the subsequent stay-at-home order on the online price dispersion in China. We employ a difference-in-differences framework by comparing the changes in price dispersion in a 3-month period around the date of lock-down order, with the corresponding changes around the same period in the lunar year 2019. Surprisingly, we find a decrease in product price dispersion during the epidemic outbreak, contrary to what the neo-Keynesian model predicts during a crisis, which naturally leads to an unexpected inflation shock. We propose that a specific feature of this epidemic – the nation-wide lock-down and the extended holiday – allows consumers to search extensively in online shopping, which reduces information cost and improves market efficiency. After taking into account such an “extra holiday effect”, we find the net effect of the epidemic on price dispersion to be positive, but the effect is more than offset by the large decrease during holidays. We provide further evidence by comparing products with different crisis-time demand elasticity (e.g., food versus clothes) and online search intensity, and find the “extra holiday effect” is most evident in products that are of inelastic demand and high search intensity. Incorporating the search intensity into the model absorbs the holiday effect, confirming our hypothesis that the holiday effect is exemplified through a high level of search activities. Price dispersion is widely considered as a measure of market efficiency and consumer welfare. China’s well-developed online market appears to be able to serve as a buffer and even an offsetting force during the COVID-19 crisis.
The Effectiveness of Monetary Policy Communication: Evidence from Textual Analysis of Daily Newspaper Articles (E5, G1)
The goal of this paper is to measure how effectively the Fed communicates monetary policy through FOMC minutes. Using sentiment analysis on daily newspaper articles, I measure the surprise in the daily news sentiment that is caused by FOMC minutes release. The surprise in the sentiment is calculated as the change in the news sentiment that is measured over three weeks prior to the release of the FOMC minutes and news sentiment with in one day after the release of the minutes. Using an event-study methodology, I test whether the sentiment surprise due to the release of the FOMC minutes affect the one-day returns in the S&P 500. The analysis is conducted in both the pre- and post- 2011 periods. In March of 2011, the Fed began holding a press conference on the day of the FOMC meeting. If the Fed effectively communicates its policy stance through the press conference, the new information content of the minutes release should diminish. I find that the magnitude and significance of the effect of sentiment changes due to the minutes release on the daily returns falls across the two periods. I am able to disaggregate the sentiment surprises into individual topics. I find that the Fed is more effective in communicating about monetary policy and economic growth than about inflation and labor markets.
The Eﬀects of College Admissions Probabilities on Major Choice (I0)
Field of study is a key determinant of lifetime earnings. In fact, major choice explains a significant portion of the persistent wage gaps across college graduates. In standard worldwide contexts in which applications consist of a joint college and major decision, students choose a major partly in response to their chance of being admitted. In this paper, I study how the likelihood of college admission aﬀects socioeconomic differences in choice of major. I explore changes in admissions probability induced by an affirmative action program adopted by a flagship university in Brazil. Admissions are a variation of the Boston Mechanism (not strategy-proof) and applicants must choose only one major before they take the entrance exams. The admissions rule adopted by universities in Brazil is pre-determined and based exclusively on the applicant’s ranking in a series of entrance exams. The affirmative action policy changes the admissions rule by reserving 40 percent of college seats for low-income applicants from public elementary and high schools. Using the quasi-experimental nature of the intervention, I estimate the eﬀects of changes in admissions probabilities on the socio-economic gap in major choice. I compare major choices between eligible and non-eligible populations before and after the affirmative action policy. I find that the policy reduced the gap in application to selective majors between low and high socioeconomic status groups by 60 percent. This effect is increasing in exam achievement levels. I find evidence of adverse gender effects, with female-male gap in application to male dominated majors worsening, especially among non-eligible applicants. My ﬁndings contribute to an emerging literature on the determinants of and socioeconomic differences in major choices. Encouraging individuals to apply to higher return majors may be an important channel through which affirmative action policies increase economic mobility.
The Global Effects of United States Monetary Policy on Equity and Bond Markets: A Spatial Panel Data Model Approach (E5, C3)
We use spatial panel data model analysis to study the international transmission of U.S. monetary policy shocks in the global equity and bond markets. Through this analysis, we decompose the overall effect of such a shock into 1) direct effects, 2) higher-order network effects transmitted through global economic networks, and 3) simultaneous effects transmitted between local equity and bond markets. Theoretically, our analysis of the transmission mechanism for the shocks relies on a network model with monetary policy spillovers. Empirically, we study asset price responses around the scheduled Federal Reserve announcements to demonstrate the significant roles of all three effects in the transmission of shocks.
The Huey Long Public Works Program in Louisiana and Fiscal Multipliers in the Great Depression (N1, E6)
What was the economic impact of government spending during the Great Depression? We revisit this seminal question using a unique state- and parish-level government spending dataset covering the ambitious public works program in Huey Long’s Depression-era Louisiana. Long was elected governor in 1928 on a populist platform and embarked on a large-scale spending program while similar, neighboring states did not. New Deal spending in Louisiana later lagged due to Long’s feud with President Roosevelt, but after Huey’s assassination in 1935, the Long faction made peace with the President. Federal New Deal spending in Louisiana surged. We exploit this temporal and state-level variation in government spending, along with variation in economic slack, to estimate fiscal multipliers during the Great Depression. Given the salience and significance of corruption for both the theory and empirics of fiscal multipliers in this episode, we propose a “corruption dismultiplier” and discuss potential estimation techniques.
The Impact of Education Reforms on Household Adult Welfare Outcomes in Ethiopia: The 1994 Free Primary Education (FPE) Reform (I2, D6)
This study examines the effect of free primary education reform on years of schooling and on various indicators of welfare in Ethiopia. Welfare is measured using multiple poverty indicators, including per adult equivalent consumption expenditure, relative deprivation in terms of consumption expenditure, and poverty gap. Using variation in individuals’ dates of birth at the time of the reform as a source of exogenous variation in education, cohorts of age 14 and younger in 1994, who were either in pre-school or primary school, are presumed to be exposed to the reform, whereas those above age 14 are presumed not to be exposed. I used both difference-in-differences (DID) and instrumental variable estimation strategies to estimate the impact of the reform on education, and the causal impact of education on adult welfare outcomes. Preliminary results show that the reform led to an increase in years of schooling of 0.80 years (without controls) and 0.77 years (with controls), and increased the welfare of individuals who were age 8 or younger in 1994. Therefore, in general the reform increased the education and welfare outcomes of individuals age 8 or younger in 1994, who were likely to be in preschool or in the first cycle of primary school when the reform started.
The Impact of Unconventional Oil and Gas Developments on Population Dynamics: A County-Level Analysis of Migration Inflow and Outflow (Q4, J2)
The purpose of this paper is to employ a difference in difference methodology to determine the direction and magnitude of population changes, wages and employment for counties with and without shale gas and tight oil in Ohio and Pennsylvania. Such an effort involves creating a control group of counties with similar employment by industry and comparing employment, population and wage trends in the control group to those counties in the treatment group that are witnessing oil and gas drilling. This study includes twenty-four counties in Ohio and Pennsylvania, twelve of which began drilling wells for shale around 2011 and twelve that did not. The research question is whether producing oil and gas via shale have an economically significant effect on the job market, specifically the number of employed individuals in these counties. The analysis incorporates migration inflow and outflow between counties in the control and treatment groups. The analysis is based on county-level data for number of wells in each county, shale gas and oil production, number of paid employees, annual payrolls, population change, and employment. The results of the analysis show that counties that have shale oil extraction observed a 3.6 – 6.5% increase in employment due to the new technology. Previous literature has mostly ignored population trends or used synthetic control groups to measure the economic impact of shale gas, however this paper incorporates actual migration flows, to account for the fact that people may move to a neighboring county with shale development for new job openings. Finally, the paper discusses main concerns for the environmental and social costs when considering further development of hydrofracking.
The Impacts of Open Access on Scientists, Inventors, and the Public (O3, O0)
The main goals of making scientific literature open access are twofold: 1) speed scientific discovery and 2) give access to the public who funded the research. In this paper, I use citations from articles, patents, and Wikipedia to examine whether open access achieves these goals by estimating whether innovators (scientists and inventors) or the general public increase their use of articles after those articles become freely available on PubMed Central, the largest repository of free full-text biomedical articles. Estimates suggest that innovators modestly increase their use of the typical article after it becomes freely available, but that the public substantially increases their use. Using the NIH's Public Access Policy (PAP) as an instrument for an article becoming freely available suggests that, in contrast to the modest effects for the average article, innovators substantially increase their use of complier articles -- those articles that are freely available only because the PAP requires them to be. This is possible if particular subsets of innovators only gain access to complier articles after implementation of the PAP.
The Ins and Outs of Employment: Labor Market Adjustments to Carbon Taxes (Q5, J0)
Environmental policies have swept the world for decades. These policies received increasing attention from the public partly because of potential job and wage losses. Are environmental policies and adverse labor market consequences connected?
This question has been argued at length and empirical evidence is mixed. While many studies document the significant labor market effects of environmental policies in regulated firms, some find the effects weak. If these effects are weak, why is the public so concerned with environmental lawmaking? If these policies create significant job and wage losses, what hinders the literature from identifying the significant unemployment and wage effects? What is missing between the public and prior literature?
To answer these questions, this paper provides an in-depth study of how an environmental policy shapes a labor market. To establish a causal relationship, I exploit a unique opportunity provided by the introduction of a revenue-neutral carbon tax policy in British Columbia. I identify the causal effects of BC's carbon tax on labor market outcomes with the coarsened exact matching method and the difference-in-differences approach.
This paper reveals the mechanisms explaining the differences in the dynamics between unemployment and wage effects. I find that the unemployment effect arrives without lags and decays quickly. The initial unemployment effect was significant because of the job-loss and job-finding effects. In other words, the tax made job losses increasingly common and made job-hunting harder. The unemployment effect quickly decayed because the job-loss effect was short-lived. Intuitively, the carbon tax increased the marginal cost of production, depressing labor demands. When firms decided to lay off workers, those workers were laid off without substantial lags. This mechanism partly explains why the job-loss effect happens only in the early stage of the policy and why the significant unemployment effect decays quickly.
The adverse wage effect comes with lags and grows gradually. The incumbent wage effect was negligible for years, reinforcing our understanding of nominal wage rigidity among incumbent workers. Since incumbent workers were the majority in employment, the initial average wage effect was negligible. Meanwhile, the average hiring wages plunged, and this effect lasted long. Average wages continued to decrease with a gradual increase in the proportion of new hires in employment. Wage adjustments operated through this slow process of labor turnovers, explaining why the average wage effect grew gradually.
These findings enhance our understanding of labor market adjustments. While it is commonplace to say that wage cuts are alternatives to layoffs, I find the opposite---layoffs mediate wage cuts. Moreover, these findings reveal that a complete understanding of the dynamics of unemployment and wage effects requires an explanation of employment flows---the inflow and outflow of employment.
Furthermore, the findings offer an explanation bridging between the public and prior literature. The documented significant unemployment and wage effects recognize the public concerns on potential job and wage losses created by environmental policies. Meanwhile, this paper explains why prior studies may fail to capture the significant labor market effects of environmental policies.
The Macro-financial Effects of International Bank Lending on Emerging Markets (F3, E5)
We provide novel empirical evidence on the effects of cross-border bank lending on emerging market economies' (EMEs) macro-financial conditions. We identify causal effects by leveraging the heterogeneity in the size distribution of bilateral cross-border bank lending to construct granular instrumental variables for aggregate cross-border bank lending to 22 EMEs. We find that cross-border bank credit causes higher domestic activity in EMEs, and looser financial conditions. Financial condition indices ease, nominal and real effective exchange rates appreciate, sovereign and corporate spreads narrow, domestic interest rates fall, and housing prices increase. Similarly, real domestic credit grows, real GDP expands, and imports rise. Effects are weaker for countries with relatively higher levels of capital inflow controls, supporting the view that these policy measures can be effective in dampening the vulnerabilities associated with external funding shocks.
The Macroeconomic Impact of Oil Industry Uncertainty: New Evidence from Millions of Financial Analyst Forecasts (E3, Q4)
We develop measures of oil industry uncertainty (OIU) using analyst forecasts drawn from a large firm-level dataset. Our oil industry uncertainty measure is related to future economic downturns, so that some versions of our measure may serve to forecast future downturns. An increase in oil uncertainty also has adverse effects on the US oil sector. The results are robust to conditioning on aggregate uncertainty. At the same time, oil industry uncertainty is related to increases in stock prices. This is in contrast to aggregate uncertainty, which has the opposite effect on stock prices. Our oil industry uncertainty measure is thus an independent influence on both the oil industry and on economic aggregates. We also look at the correlation between our baseline OIU shocks and other oil shocks identified in the literature namely oil supply shocks, economic activity shocks, oil specific demand shocks, and oil speculative demand shocks, and find that the impact of oil uncertainty is not due to it functioning as a transmission channel for first-moment oil shocks. However, we find that technical change specific to the oil industry, approximated by the stock of patents associated with the oil and gas industry, is significantly correlated with OIU, and weakly Granger-causes OIU. This suggests that our oil industry uncertainty measure may at least partly reflect uncertainty stemming from advances in oil industry technology.
The Monetary Policy Rescuer: A Long-Run Nominal Interest Rate Target (E5, E3)
Monetary policy in most developed nations is under pressure. With key policy rates either negative or close to zero, analysts are wondering what central banks would do to fight the next recession. This paper proposes that FOMC (and other central banks) can revive the effectiveness of monetary policy by declaring a commitment to an explicit long-run nominal policy rate target.
Our work suggests that one major reason of the diminishing effectiveness of monetary policy is the lack of an explicit long-run Federal Funds Rate (FFR) target. Due to the absence of a long-run FFR target, market participants are unable to gauge whether the current policy stance is accommodative or restrictive and by what magnitude? One benchmark is the equilibrium interest rate, or r-star. However, there are at least half a dozen different measures of the r-star which makes it very difficult to evaluate monetary policy accommodation.
We propose 4% as a long-run target for the nominal FFR. Some of the major benefits of our proposed framework include: helping market participants gauge whether the current stance is accommodative/restrictive; anchoring policy watchers’ expectations in the sense that analysts would expect the FFR to stay close to its target; reduce the risk that the FFR would hit and stay at the zero lower bound for an extended period of time; reduce changes in the Fed’s balance sheet by providing enough room to cut rates in the case of a slowdown/recession; and, with the inflation target rate set at 2%, ensure that the real FFR will be positive when the FOMC meets its interest rate and inflation targets.
Therefore, we suggest a long-run FFR target would boost the effectiveness of the monetary policy and central banks should entertain the idea of an explicit long-run policy rate target.
The Poor Tax: Redistributive Pressure and Labor Supply (O1, H3)
In developing countries, financial transfers within kin and social networks are ubiquitous. We test whether redistributive arrangements act as a tax on earnings, thereby distorting the incentive to work: workers reduce labor supply because they cannot retain the full benefits of their effort. We enable full-time piece rate factory workers in Côte d’Ivoire to deposit earnings increases into blocked private savings accounts over 9 months, lowering the marginal tax on increased effort. We find evidence for large labor supply distortions: the intervention increases output and earnings by 11%, implying an 26% Treatment on the Treated effect on worker earnings. When others in the workers’ kin network would learn of the existence of the savings accounts, take-up plummets from 60% to 14%—consistent with redistributive pressure. Our findings suggest that informal insurance, while potentially welfare improving, may generate substantial efficiency costs. This highlights a way in which underdevelopment can contribute to low productivity in developing countries.
The Price of Silence: Marriage Transfers and Women’s Attitude Toward Intimate Partner Violence (J1, Z1)
Violence against women is a major health problem, yet intimate partner violence continues to be widely justified around the world. Thus, targeting attitudes towards IPV is a powerful way to reduce violence and empower women. In Jordan, despite considerable public efforts, women's justification of IPV remains high and the media and public opinion point to the dower as the culprit.
In Jordan and throughout the Middle East, North Africa and Muslim regions of Asia, the husband must pay a dowry - a transfer of money - to the bride upon validation of the marriage. Although the prevalence of the dower and its magnitude are considerable, we continue to have very limited knowledge of its effects on women's well-being. Using the Jordanian Labour Market Survey 2016, I estimate how women's justification of IPV is affected by the amount of dower received at the time of marriage. I find a positive and significant association, robust to several alternative specifications, including an instrumental variable strategy. Since women must return the dower they received at marriage if they are seeking a divorce, these results are consistent with an intra-household model that predicts than a higher dower reduces women's outside options.
The Puzzling Politics of R&D: Political Connections and Innovation in Russia (O3, O1)
Technological progress is an important factor in economic development, yet it can be a destabilizing force, upending the existence balance of power in the economy. Such changes can be unwelcome to the government that would like to preserve the existing status quo. However, blocking technological development is hardly feasible. Instead of stifling innovation, the incumbent can have incentives to assure technological advantage of companies that belong to the set of individuals that are loyal to him by providing government grants. Unlike pork-barrel forms of favouritism that are often illustrated in the literature on the value of political connections, such support would require some efforts on the part of receiving company, so that it actually obtains technological advantage. Using trajectory balancing approach for all companies that applied for Russian program of R\&D support via government subsidy, I show that politically connected companies are more likely to obtain government R\&D grants. Furthermore, they are more likely to benefit from them than unconnected companies. The heterogeneity in the effect suggests that the role of political connections in technological development of companies goes beyond receiving government funds. In addition, given that all applicants had to be well-informed in order to win the grant, it suggests that the role of connections goes beyond obtaining relevant political information.
The Real Consequence of Failing Stress Test: Evidence from Mergers and Acquisitions (G2, G3)
We find that borrower firms of banks that failed U.S. stress tests subsequently conduct significantly less mergers and acquisitions (M&A). The effect is stronger for treatment firms with weaker corporate governance or treatment firms more susceptible to managerial agency problems such as empire building. Moreover, financial covenant usage in M&A-related bank loan contracts increases after bank stress test failure shocks. We further document a positive impact of bank failing stress test on M&A deal quality particularly when the acquirer needs to finance M&A via raising new bank loans. Taken together, these findings suggest that failure banks increase screening on financing corporate borrowers’ M&A projects. In line with treatment firms refraining from M&A activity that can hurt their shareholders, we also find these firms to subsequently improve their profitability. Our fresh empirical evidence on corporate borrowers’ M&A suggests a beneficial spillover effects of bank stress tests.
The Rise (and Fall) of Science Parks (R1, O3)
Science parks have gained the favor of many analysts and policy-makers (Katz and Bradley, 2013). Despite the importance of science parks in the real world (Link and Scott, 2003), it has attracted little attention in economic theory (Link and Scott, 2007).
We provide a full-fledged GE model where a science park may emerge as a decentralized equilibrium outcome. Our model includes: (i) a composite consumption good which produced by using an endogenous range of inputs; (ii) workers are hired to produce intermediate goods or to conduct R&D; (iii) the productivity of an intermediate firm depends on its level of R&D and inter-firm spillovers; and (iv) both workers and intermediate firms are spatially mobile and use land.
We identify three key rationales for a formation of a science park: (i) more localized knowledge spillovers, (ii) low commuting costs, and (iii) abundance of skilled labor. These are all typical features of high-tech industries that make a science park, which accommodates more intermediate firms and fosters research activities. This may explain why those firms are often clustered in science parks such as the Silicon Valley, the Hsinchu Science-Based Industrial Park in Taiwan or the Cambridge Science Park in the U.K. (Saxenian, 1994; Chen, 2008; Helmers, 2019) and why in the absence of localized knowledge spillovers, simple cluster policies are not sufficient for a science park (Duranton et al., 2010). Continual improvements in infrastructure may lead to the fragmentation of science parks as observed in the case of Silicon Valley (Saxenian, 1994).
Eventually, too large an inflow of migrants due to high wages will lead to their geographical fragmentation. Thus, science parks are more likely to be sustainable in urban environments that do not have too large a labor pool. This may explain why most successful science parks do not emerge in megacities.
The Saving Behavior of Heterogeneous Households and Credit Constraints: A Decomposition (D1, E2)
This paper sheds light on how saving decisions respond to credit constraints. More specifically, the paper is concerned with whether the households saving behavior responds to credit constraints to build wealth and relax liquidity constraints or accumulate funds for precautionary purposes. The paper attempts to understand to what extent credit-constrained households can build wealth when the presence of a liquidity trap characterizes the macro environment. This paper utilizes a probit model using cross-sectional data from the Survey of Consumer Finances (SCF) to examine the effect of credit constraints on constrained and discouraged households' savings behavior. By discouraged households, we refer to households that perceive a high probability of loan denials, while constrained households are those whose credit applications are denied by financial institutions. To do so, we classify the reasons for saving to develop a better understanding of whether saving plays a substantial role in households' wealth accumulation when saving is motivated primarily by uncertainty (precautionary saving), retirement, or investment needs. We also utilize a quantile regression to capture the effect of credit constraints on the different levels of household wealth. Indeed, credit-constrained households face difficulty saving to accumulate wealth; however, the effect of credit constraints on discouraged households fades out as households' wealth increases. The results of this study indicate that researchers should account for credit constraints when modeling household saving behavior.
The Signaling Role of Early Career Job Loss (J3, D8)
I examine the extent to which ability signaling explains long-term wage losses suffered by young workers who experience layoffs. Young workers are of particular interest because employers have limited information about their ability, so signaling theoretically plays a larger role in determining wages. In addition, young workers are unlikely to experience wage losses due to loss of industry-specific human capital or separation from high-quality job matches, which may explain long-term wage decreases among older workers. Using data from the National Longitudinal Survey of Youth 1997, I show that young workers of all ability levels initially experience similar wage losses following layoffs, but high relative ability workers fully recover within five years while low relative ability workers experience persistent wage losses. Consistent with traditional learning models, relative, not actual, ability affects wage trajectories. I illustrate a conceptual model of layoff signaling that varies by pre-layoff experience and can explain divergent wage trajectories across high and low relative ability workers. I test the model empirically and find that low relative ability workers’ inability to overcome negative layoff signals explains a substantial proportion of long-term wage losses among young workers.
The Skill Landscape of Germany: Evidence from Apprenticeship and Job Vacancy Data (J2, I2)
As a consequence of technological progress, jobs are subject to rapid changes in terms of their task content and the skills required to work in these jobs are changing with them. In this paper, we construct highly detailed measures of skill supply and skill demand in Germany using apprenticeship plans and online job vacancy data, respectively. This allows us to assess how the German apprenticeship system - which is highly praised internationally as a model for preparing young people for the age of automation (Economist, 2018) - is actually able to prepare its graduates adequately for the challenges of the modern labor market. We also investigate the economic consequences of the mismatch between skill supply and demand.
The German apprenticeship setting is unique in the sense that skill requirements of apprenticeships are codified in state-approved apprenticeship plans, which are standardized across the whole country. Thus, the same practical and theoretical skills are developed through a particular apprenticeship regardless of the training location in Germany. Extracting the skills conveyed through each apprenticeship by using text analysis techniques enables us to relate a precisely measured, highly detailed set of skills to each apprenticeship on the supply side. On the demand side, we depict employers' demand for skills using a rich data set containing over 550,000 web-crawled job vacancies from the federal employment agency and other job vacancy boards with in-depth information on job requirements in Germany. We can link apprenticeship plans and job vacancies at the occupational level, which allows us to assess the extent to which the German apprenticeship system satisfies employers' skill needs. Moreover, we study how mismatch between skill supply and skill demand affects labor market performance and wage inequality at the district level, using high-quality administrative data from the Sample of Integrated Employment Biographies (SIAB).
The Strategic Use of Corporate Philanthropy: Evidence from Bank Donations (G3, G2)
Using data on bank donations to nonprofit organizations, we examine the strategic nature of banks' charitable giving. We find that bank donation decisions are driven by local market competition and such donations subsequently lead to a higher local market share. We confirm our results by using two exogenous shocks: the application of antitrust laws in bank mergers and natural disasters. We further show that bank donations lead to increases in local mortgage originations and in the likelihood of entry into new markets through branch openings. Overall, our evidence suggests that banks participate in corporate philanthropy strategically to enhance performance.
The World Interest Rate (F3, F6)
The world interest rate is a key determinant of borrowing and lending behavior and plays a central role in most open-economy macro models. This paper uses a dynamic factor model to construct a world interest rate based on short run interest rates of over 70 countries. Compared to simple averages, such a process should better extract a common trend, rather than be influenced by idiosyncratic shocks. In a simple econometric model, the global rate can be dominated by whatever region has the most countries in the sample. With regional factors incorporated, the obtained global factor can reflect overall global influences. Consistent with recent evidence, we find the global rate is on a long run secular downward trend. Furthermore, capital account openness strongly affects both the co-movements of local-regional and local-world rates. Finally, we also make methodological suggestions for the practical use of dynamic factor models especially when the proper regional groupings are not obvious.
Trade Credit and Markups (F1, G2)
Trade credit is the most important form of short-term nance for firms. In 2019, U.S. non-financial firms had $4.5 trillion in trade credit outstanding, equaling 21 percent of U.S. GDP. This paper documents that firms with higher markups supply more trade credit, an effect that increases with the buyers' borrowing rate. We rationalize this finding in a model with positive markups and costly financial intermediation. In the model, reducing financial intermediation costs provides a strong rationale for the dominance of trade credit in firm-to-firm transactions. Using U.S. Compustat and detailed Chilean export data, we find strong support for the model.
Trade of the Customer Capital and Aggregate Welfare (F1, L1)
Using novel data sets that link production and sales of Chinese firms in the overseas market, we document a large number of firms outsources their exports to a firm within a narrowly defined homogeneous industry. Interpreting this as evidence of trading demand capital (e.g., accumulated customer base) in the foreign market where exporting firms often face a high barrier to access foreign customers, we build a model in which firms can borrow and lend their customer base via outsourcing. The model can reconcile the observed positive correlation between outsourcing and productivity dispersion. Our finding implies that how much frictions in the product market matters for the aggregate welfare crucially depends on the extent to which the customer capital can be traded.
Uncertainty, Imperfect Information, and Expectation Formation over the Firm’s Life Cycle (E2, D8)
Using a long panel dataset of Japanese firms that contains firm-level sales forecasts, we provide evidence on firm-level uncertainty and information frictions over its life cycle. We find that firms make non-negligible and positively correlated forecast errors. However, they make more precise forecasts and less correlated forecast errors, when they become more experienced in doing businesses. We then build a model with heterogeneous firms that gradually learn about their demand using noisy information. We quantify the learning channel and the real options channel (along the age dimension), through which a greater variance of informational noises adversely affects resource allocation and thus depresses productivity.
Understanding the Bimodality of the Export Intensity Distribution in Thailand (F1, D2)
Recent evidence finds that firm-level export intensity, defined as the ratio of exports to revenue, is bimodally distributed in at least 47 countries. In this paper, we investigate the determinants of the bimodality in a developing country by using Thailand's manufacturing firm-level census data covering the period between 2007-2017. Consistent with Melitz (2003), we do not find evidence that firm productivity can explain the variation in export intensity. We document that firms with the export intensity of at least 90 percent, so-called “pure exporters,” are relatively young, have foreign ownership, produce narrow product variety, and export to high-income countries.
Understanding the Changes in Labor Supply of Older Workers Across Cohorts in the United States (J0, E2)
During the past several decades, labor supply of older workers has increased dramatically in the United States. The goal of this research is to develop and estimate a dynamic life-cycle model that could explain the increase in labor supply across cohorts. To do so, I focus on two cohorts: the 1930s and 1950s ones. Using data from Panel Study of Income Dynamics (PSID), I document that the 1950s cohort, relative to the 1930s one, supplies more labor at the older ages, from ages 60 to 70, in both participation rates and working hours. In addition, the 1950s cohort faces a changed economic environment in terms of health dynamics, wages, spousal earnings and Social Security rules, compared to those faced by the 1930s cohort. To assess the contributions of these changed factors on the increase in labor supply of older workers across cohorts, I develop a quantitative dynamic life-cycle model of labor supply, retirement and savings decisions of men, and estimate the model using the Method of Simulated Moments to match the life-cycle profiles of labor participation, hours worked and savings for the 1930s cohort. I then assess their effects by replacing the changed factors faced by the 1950s cohort to the 1930s cohort’s estimated model. In a dynamic life-cycle model, all the behavior changes are jointly and quantitatively explained by a combination of changes in various labor supply determinants, with the changes in wages and Social Security earnings test that was eliminated beyond the Full Retirement Age from the 1930s to the 1950s cohorts being the dominant contributors.
Understanding Wage Growth: The Role of Coworkers (J0, J3)
We study a critical driver of wage growth: coworkers. Using linked employer-employee data for Italy, we explore coworkers’ effect on wage growth in two directions. First, using a novel estimation method and accounting for the endogenous sorting of workers into peer groups and firms, we estimate the impact of the average coworker’s quality on future wages. We find that a 10 percent rise in coworker’s quality increases one’s wage in the next year by 1.8 percent. The effect decreases gradually over time and becomes about 0.7 percent after five years. Second, we delve deeper into the channels that identify the peer effect. Using an event-study specification around mobility episodes, we study how the entry and leave of a high-quality or low-quality worker could affect both the mover and the coworkers' future wages. We find that hiring a high-quality worker and separating a low-quality worker are important drivers of wage growth for the incumbent workers. Movers experience an immediate gain when moving into high-quality peers. Knowledge spillover could play an important role in explaining the mechanisms behind our findings.
Unearthed: An Exploration of Shale Development on House and Income Inequality (R0, Q4)
Recent advances in shale development have produced both positive and negative outcomes for local communities, with higher employment and income known to be the most significant effects. Despite the stated importance of the distribution of economic gains among local populations in previous literature, adequate research on the shale boom's impact on inequality and affordability does not exist. I employ the difference-in-difference (DiD) method to study the unintended social consequences of the hydraulic fracturing boom in Oklahoma, the second-largest producer of oil and gas in the country, over the period of 2004-2017. Given the magnitude of the economic adjustment with the advent of hydraulic fracking in Oklahoma, it is a prime setting to study the impact of shale development on inequality and affordability. I find that shale counties experienced higher housing prices and lower affordability power compared to non-shale counties. However, the estimation fails to find any statistically significant effect on inequality. Oklahoma provides a unique setting because of the size of the fracking industry and the speed with which fracking grew. Nevertheless, the setting is not too unique: The results will be relevant for any locality
facing a natural resource boom.
Unequal and Unstable: Income Inequality and Bank Risk (G2)
We provide evidence that regions in the U.S. with higher income inequality tend to have a higher incidence of failed banks. However, not all banks are more risky, as reflected in a higher dispersion of bank risk. We show how a model based on risk-shifting incentives where banks channel insured deposits into subprime loans can account for both findings. In equilibrium, a competition to risk-shift emerges, leading to a subprime lending boom in which loans to high-risk borrowers carry negative NPVs. Some banks engage in risk-shifting by lending to high-risk subprime borrowers, while the rest specialize in lending to low-risk prime borrowers.
Universal Basic Income, Targeted Cash Transfers, and Progressive Taxation: Reducing Income Inequality in South Africa (H0, D6)
South Africa has one of the world’s most progressive tax systems, yet income inequality continues to be a major challenge for the country. Several fiscal policy initiatives have been implemented since the end of apartheid to reduce the high levels of inequality and poverty. Despite this, there has been no significant reduction in inequality in post-apartheid South Africa. Universal basic income (UBI) and better progressive taxation can be a new way to address the limited strength of fiscal policies in South Africa. In developing countries, however, data on income is limited for the vast majority of the population working in the informal sector - informal labor is about 86% in Africa (ILO, 2018). Additionally, inclusion in the formal tax system is low. This paper compares the magnitude by which UBI versus targeted cash transfer (TCT) funded by progressive taxation can reduce income inequality in South Africa. Empirically, I conduct a policy simulation exercise to analyze how additional revenue generated from tax progressivity can be used to finance UBI and TCT, and to what extent this can reduce income inequality. Results show that both UBI and TCT reduce income inequality by more than 30% when these policies are accompanied and financed through a progressive taxation; however, UBI performs better in reducing inequality than TCT.
Valuing Domestic Transport Infrastructure: A View from the Route Choice of Exporters (R1, F6)
A key input to quantitative evaluations of transport infrastructure projects is their impact on transport costs. This paper proposes a new method of estimating this impact relying on the widely accessible customs data: by using the route choice of exporters. We combine our method with a spatial equilibrium model to study the aggregate effects of the massive expressway construction in China between 1999 and 2010. We find that the construction brings 5.1% welfare gains, implying a net return to investment of 150%. Our analysis also produces some intermediate output of independent interest, for example, a time-varying IV for city-sector export.
Vertical Collusion and Bid Evaluation in Open Procurement Auctions (D4, D7)
Public procurement is a complex process and there are numerous opportunities for corruption at any of the stages. I formulate a new theoretical model of low-bid auctions which includes collusion between an auctioneer and a bidder at the bid evaluation stage. The model takes into account that collusion changes the behaviour of favoured bidders. It predicts that favoured firms bid less aggressively by increasing their bids by the amount of expected rent. As a result, this equilibrium response inflates the average prices of procurement. This method helps me to determine the welfare and efficiency loss due to bid tailoring. I also propose a method of model identification which is based on observed variation in the number of reported bids in bid evaluation reports. These properties represent potential contributions to the auction and corruption literatures.
For the model application, I have collected a large dataset containing over 2 million procurement purchases from the Russian Federation where abuse of the bid evaluation process is a known problem. The fixed-effects regression analysis suggests two effects: (a) the procurement prices are between 53.1 and 46.1% higher in the rigged auctions and (b) though the average number of bidders in such auctions is considerably higher the actual number of admitted and reported bidders are 20% lower. Then I show that these data patterns are not explained by a non-collusive bidding model and propose the new model with corruption. I estimate the model structurally to confirm that bidding behavior changes in line with the model. The preliminary estimates show that model with bid tailoring provides between 20-98.5% better fit than a standard English auction model in 21 studied markets.
Voluntary versus Mandatory Public Annuity Plans: A Unified Framework to Understand Their Pros and Cons (H5, G5)
Many emerging economies choose the fully-funded retirement financing system. One disadvantage of this system is that citizens have to bear more risks, including the longevity risk (the risk of outliving their resources when they live longer than anticipated). In this context, several societies, including Singapore and Hong Kong, have introduced the public annuity (PA) plans in recent years. We observe a major difference in existing PA plans: the voluntary public annuity with ceiling (VPAc) plan and the mandatory public annuity with flexibility (MPAf) plan.
We develop a simple model with asymmetric information on survival probability to study the pros and cons of voluntary versus mandatory PA plans. We find that the idea of restrictive flexibility provides a unified framework to understand different PA plans. Introducing the PA plan reduces the severity of adverse selection in the PA market, but further distorts the private annuity market.
The MPAf and VPAc plans have systematically different effects on the welfare of citizens with different survival probabilities. Whether the MPAf or VPAc plan is adopted, the healthy group loses but the intermediate health group benefits. On the other hand, the effect on the least healthy group depends on which PA plan is adopted. In particular, they lose under the MPAf plan because of the requirement tobuy the minimum mandated level, but benefit from the VPAc plan because the ceiling restriction does not affect their preferred amount of PA purchase. These results provide some guidance regarding which PA plan the government chooses.
Wage Determination and the Bite of Collective Contracts in Italy and Spain: Evidence from the Metalworking Industry (J5, J3)
In several OECD countries employer federations and unions fix skill-specific wage floors for all workers in an industry. One view of those “explicit” contracts argues that the prevailing wage structure reflects the labor market conditions back at the time when those contracts were bargained, with little space for renegotiation. An alternative view stresses that only workers close to the minima are affected by wage floors and that the wage structure reacts to current labor market conditions. We disentangle both models using a novel dataset that combines more than 1,000 signature dates and 15,000 wage floors set in the metalworking industry with labor market histories of metalworkers drawn from Social Security records in Italy and Spain. An increase in the contemporaneous local unemployment rate of 1 p.p. diminished contemporaneous mean wages by about 0.45 p.p. between 2005 and 2013 in both countries. Instead, a 1 p.p. higher unemployment rate back at the time of contract renewal reduced wages by 0.07 p.p., an impact driven by wages close to the negotiated wage floors. Even though the evidence for earlier periods is mixed in Italy, the results do not support the view that the wage structure reflects labor market conditions at the time of bargaining. The results support the hypothesis that (most) wages respond to local current unemployment, although the estimated elasticity falls short of the prediction of an off-the-shelf bargaining model.
Wage Risks, Child-Development, and Optimal Taxation of Families over Life Cycle (H2, H3)
Despite the "surprising" potential gain of age-dependent taxes found in the dynamic taxation literature (e.g. Weinzierl (2011) and Farhi and Werning (2013)), except for the social security part of the tax and transfer systems that include some age-dependency, age-dependent taxes have not been ever implemented in any country. Concerns over the robustness of the welfare gains to model assumptions and the distaste of societies to use tags based on age might be among the possible reasons.
In this paper, I study the potential gains of an alternative for age-dependent taxes. Marriage and child dependent taxes are very common in different countries. The high correlation of these characteristics with wages, labor supply elasticity of women, and tightness of household borrowing constraints make these forms of taxation a viable alternative for age-dependent taxes.
As my first contribution, I estimate a life-cycle family labor supply model with temporary and persistent wage shocks and changes in family structure and solve for the optimal marriage and child dependent tax schedule. I evaluate its welfare gains compare to the optimal age-dependent tax schedule. I also examine if there is substantial welfare gain from introducing additional age-dependency or capital taxation when the tax function is designed optimally using demography variables.
As my second contribution, I introduce the child-development process into the model and solve for the optimal taxation problem in this new environment. Insuring altruistic parents and unborn children against low investment on the human capital of in early ages, provides an additional motive for more progressive child-dependent taxes. I examine how the previous results on optimal tax schedule change when the child-development process is considered.
Welfare Costs of Occcupational Decline: Counterfactual Approach (J6, C5)
Both trade and technological changes constantly cause large structural shifts in labor markets and reallocation of workers between jobs and industries. McKinsey Global Institute (2017) predicts that by 2030 between 3 and 14 percent of global workforce will have to switch occupations due to automation. How much do workers loose or gain when their occupations get automated or outsourced?
The goal of this project is to measure welfare losses of workers' reallocation due to an occupational decline. I construct and estimate an extended multi-sector (Roy, 1951) model of occupational choice with multiple latent workers' characteristics. Workers' characteristics include both unobservable skills and preferences. Jobs differ in terms of skill contents and in terms of job amenities. As both workers and job characteristics are unobservable in my model, the identification is based solely on the data of occupational transitions and wages. Intuitively, workers are more likely to move between occupations which are close in terms of skill contents or amenities.
My preliminary results show that welfare costs of occupational decline depend on occupation and vary from 2% of welfare to around 20%. For most two-digit occupations losses lie in the range from 7 to 12%. The losses become higher if an occupational decline affects a larger group (one-digit occupation). My estimates exceed the measured earnings losses of workers experiencing an occupational decline (Edin et al, 2018).
The paper contributes to the literature on costs of workers' displacement by accounting for non-monetary costs. Rosen (1986) shows that fit between preferences and job amenities is an important component of worker’s utility. While several papers provide estimates of earnings losses of displaced workers, none of them specifically estimates welfare costs which depend not only on earnings but also on a fit between preferences and amenities.
Welfare Cuts and Well-Being: Evidence from Indiana in the Great Recession (I3, J2)
This project analyzes the short- and long-run effects of being unable to access a critical set of government transfer programs during the Great Recession. I focus on a unique natural experiment in Indiana, which in 2007 outsourced the automation of its welfare services (covering SNAP, TANF, and Medicaid) to the IBM Corporation. Instead of making it easier for individuals to apply or recertify for welfare, this effort led to a lack of personalization in the certification process and a zero-tolerance policy for errors on certification forms. The system was rolled out to two-thirds of Indiana’s counties before it was permanently halted at the end of 2009. Preliminary difference-in-differences estimates show that – one year after the automation effort was rolled out – SNAP, TANF, and Medicaid rolls fell by 15%, 20%, and 6% (respectively) in the exposed counties. For Medicaid and SNAP, these effects appear to be permanent – five years after the effort was disbanded, treated counties still had more than 5% fewer recipients than untreated counties. Smaller-population and higher-poverty counties experienced larger declines in welfare receipt. The individuals screened out were also more likely to have children, disabilities, and lower education levels. In ongoing work, I analyze the extent to which the outsourced automation effort undermined the ability of its welfare programs to insure recipients against economic risk. Using administrative data from various sources, I examine the broader effects of the automation effort on incomes, financial solvency (e.g., bankruptcy, evictions, foreclosures), health (e.g., hospital admissions, mortality), and other outcomes.
Welfare Effects of Reducing Meat Production in United States Agriculture (Q1, I3)
Modern food movement trends include reducing meat consumption for environmental
and health reasons. One animal product that receives some scrutiny is beef, as this
sector contributes 65% of the total animal agriculture emissions in the United States
This study provides monetary costs of welfare losses involved with an extreme production and diet shift in the U.S. population. Many efforts have gone into the benefits that would be gained from an overall diet shift away from meat products, but no studies have attempted to place economic costs using current meat prices and quantities, and price elasticities in order to place dollars that would need to be spent to achieve these benefits.
An equilibrium displacement model is used to construct a system that connects supply of meat products to consumption. Using a Cost-utility perspective, a value of cost per unit of GHGE reduced and health outcome achieved can be quantified using the literature for GHGE and health outcome estimates. Results show that it would cost consumers and producers $185,560,170,000 per year to achieve a 2.6% reduction in US GHGE, a 23% increase in available calories, and a loss avoidance of 306,220 health-related deaths, based on current consumer preferences for protein sources.
In the US, this would cost approximately $71 million dollars per 1% reduction in GHGE, $8 billion dollars per 1% extra calorie availability, and $6 million dollars per health-related death avoided, per year. The importance of this result lies in the comparison of the cost estimates to the cited benefits.
This study attempts to fill this information discrepancy with some estimates of costs of an extreme diet shift. Finally, the question can be asked, is the cost of a large-scale diet change worth the benefits gained?
Welfare Magnets and Internal Migration in China (H3, J6)
This study examines the causal effects of welfare benefits on internal migration decisions. Using a quasi-experimental migration reform across 283 Chinese cities from 2002 to 2015, combined with a difference-in-differences setup, I show that improved welfare benefits substantially increase migration. The observed impact is more pronounced for individuals such as the young, women and medium-low-skilled workers. It is relatively smaller in destinations exposed to larger positive demand shocks, suggesting that improved welfare benefits reduce migration costs. And it persists over the long term. All these findings confirm the existence of sizable welfare magnet effects.
What Could Possibly Go Wrong? Predictable Misallocation in Simple Debt Repayment Experiments (D1, D9)
How do individuals repay their debt? We use a series of lab and online experiments to answer this
question. In a simple debt repayment experiment we provide subjects with two credit cards with different
interest rates and levels of debt that are to be repaid. Even though this is arguably one of the simplest
financial decision imaginable from a traditional rational choice perspective, we find severe deviations
from optimal, i.e. interest minimizing, repayment decisions: There is significant misallocation measured
via the available money that is not repaid to the credit card with the highest interest rate. We show that a
large proportion of our subjects does not know how to solve repayment problems, and that they instead
use heuristics that are prone to various fallacies, which we predict and show in our experiments.
What Do We Learn from SARS-CoV-1 to SARS-CoV-2: Evidence from Global Stock Markets (G0, I1)
This paper studies global stock market reactions to COVID-19 outbreaks caused by the virus SARS-CoV-2. The stock markets in countries that suffered from 2003 SARS diseases caused by a similar virus (SARS-CoV-1) react more quickly and strongly to the first COVID-19 outbreak in Wuhan China during late January 2020. This pattern lasted for weeks until a number of severe outbreaks outside of China (e.g., South Korea and Italy) began in late February, and the stock markets in countries without SARS experiences started to tumble. We document an important underreaction to COVID-19 in countries without the early experience of similar crises.
What Firms Do: Gender Inequality in Linked Employer-Employee Data (J3, J7)
This paper investigates the contribution of firms to the gender gap in earnings on average, at different deciles of the earnings distribution, and over time and it sheds light on how the impact of firm pay policy comes about. Using a linked employer-employee dataset for Italy, covering the universe of workers in the private sector, we show that the gap in firm pay policy explains on average 30 per cent of the gender pay gap in the period 1995-2015. Sorting of women in low pay firms explains a larger fraction of the gender pay gap than differences in bargaining power, on average and at the bottom of the distribution, whereas the latter dominates at the top; in addition, differences in bargaining power have increased in importance over time. To explain sorting, we investigate whether women have a lower probability of moving towards firms with higher pay rates, and find that this is indeed the case, especially if these firms display high (unexplained) variance in pay. We also find evidence that the firm environment as captured by exogenous changes in the gender balance in leadership positions influences the gender gap in bargaining power, indicating that the latter is partly institution-driven.
When Firms Matter: Propagation of Firm Level Shocks through Production Networks (L1, E3)
Recent empirical evidence shows that firm-level shocks can affect aggregate fluctuations through the production network. The mechanism of this, however, has been less studied. In this paper, we show how i) switching rigidities in customer-supplier links, ii) production elasticities and iii) the distribution of intersectoral firm linkages can explain how firm-level shocks can propagate into aggregate economy-wide fluctuations. We show that when the distribution of intersectoral firm linkages is fat-tailed then idiosyncratic firm-level shocks are magnified to sector level shocks even if the number of firms becomes arbitrarily large in each sector. Further, we show that even if the downstream firm-linkages is thin-tailed, this amplification does not disappear. In other words, once a firm-level shock gets amplified as a sector-level shock, it is transmitted through the economy as a sector-level shock.
When in Rome: Lending to Small and Medium Enterprises by Foreign and Domestic Banks (G2, D8)
Conditional on a loan application filed by a small or medium enterprise ("SME"), we find that the existence of recent loans of that firm with private domestic banks increases the chance a loan will be granted by a foreign bank relative to a private domestic bank. On the other hand, recent loans extended by foreign banks or by domestic state-owned banks do not produce this differential effect. Furthermore, the aforementioned effect vanishes for large firms. These findings are consistent with a mechanism by which foreign banks overcome borrower informational asymmetries by relying on their domestic peers' recent behavior. Indeed, the higher ability of private domestic banks to access informationally opaque SMEs, dependent on soft information, makes recent loans with them a more valuable signal for foreign lenders who lack the same ability.
When Machine Learning Meets North Korean Text Data: Relationship between the North Korean Economy and Security Strategy (N4, O5)
Due to the lack of data, North Korea is a difficult object for analysis. The data publicly offered by the regime are unreliable and even these are scarce. On the other hand, survey of North Korean defectors living in South Korea, which is a common alternative source of data, is cost inefficient. The paper applied machine learning tools to the collected North Korean text data which are abundant and economical. In particular, the paper transforms text data, which is commonly considered as qualitative data, into quantitative data in its analysis. Through this method, the paper attempts to examine the political and economic relationship between the statements announced by the North Korean regime toward South Korea and foreign countries and the macroeconomic reality faced by the regime.
More specifically, the paper analyzes the correlation between North Korean economy and the regime’s statement toward South Korea and foreign countries under Kim Jong Un (January, 2009 to Present) (1) As the first step, 3786 articles from Korean Central News Agency (KCNA) related to the statements toward South Korea and foreign countries were transformed into quantitative data using the Word2Vec model. (2) Then, the tone of the North Korean media is put into a numerical scale through compiling monthly indices using transformed data. (3) Lastly, the quantified tone of the media is examined whether it had any correlation with either North Korea’s macro-economic variables or global market price of its major trading goods.
The paper has quantified the tone of North Korean media for the first time and has also used the most up-to-date machine learning technology. However, the paper underscores the need for further research through expanding the scope of texts and economic data as well as applying more recent natural language processing algorithm.
Which Entrepreneurs Are Financially Constrained? (G2, J6)
We study what type of entrepreneurs are affected by financial constraints by exploiting age-based discontinuities in the amount of funding available through a public program for unemployed workers. Our sample links ad- ministrative data on 2.1 million eligible workers to the firms they create, spanning a wide range of skills, sectors and outcomes. We find that access to funding increases the rate of entrepreneurship by 11%. The effect is stronger for entrepreneurs who incorporate their business, especially for those who were in the top decile of the wage distribution before unemployment. Among incorporated entrepreneurs, the effect is strongest in the ITC sector, followed by manufacturing. In terms of ex-post outcomes, we find that the effect is more pronounced for businesses in the upper half of the size, growth and profitability distributions. Our findings suggest that financial constraints hamper growth-oriented entrepreneurship.
Who Pays for the Iceberg Transport Costs? The Effect of Seasonal Natural Barriers on Trade and Regional Inequality (N7, R1)
We illustrate how trade, which is subject to time-variant transport costs, drives regional inequality. Using data on seasonal Atlantic iceberg drift and millions of ship locations, we observe exogenous variation in maritime transport routes between North America and Europe. The southward iceberg drift during summer forces vessels on more southern and thus longer trade routes. In our paper, we trace nineteenth and twentieth century cotton trade from farmers in the Southern United States through merchants in New York to textile manufacturers in the United Kingdom. This allows us to examine how the effect of fluctuating transport costs is distributed along the supply chain. Its vast geographic reach and status as one of the most traded commodities at the time makes cotton a particularly suitable and relevant application. By exploiting the exogenous and seasonally variant influence of icebergs on transport networks, we add new insights to the gravity-based market access literature, plagued by the endogeneity of man-made transport networks.
Why Do Central Banks Make Public Announcements of Open Market Operations? (E4, G1)
Central banks make public the results of open market operations (OMOs), which they use to adjust the liquidity available to the financial system so as to maintain the short-term borrowing rate in the range compatible with achieving their monetary policy objectives. This paper shows that such announcements are costly, because that they moderate the impact of changes in supply achieved through OMOs. Nevertheless, communication of OMOs is desirable because it improves the transparency of the funding market, which makes the price of liquidity -- a key input into economic decision-making -- more informative about underlying demand and supply.
Within-Firm Labor Heterogeneity and Firm Performance: Evidence from Employee Political Ideology Conflicts (J5, J2)
This paper explores the implication of within-firm labor heterogeneity for firm performance through the lens of employee political ideology. Using individual campaign donation information to capture political ideology, I find that political ideology conflicts, both those within employees and those between CEOs and employees, are negatively associated with firms’ future operating performance. This effect is stronger for firms whose employees are more geographically concentrated and more sophisticated. The reduced labor productivity and abnormal employee turnover are two plausible mechanisms through which employee political ideology conflicts hurt firm performance. To establish causality, I use an instrumental variable approach which relies on the exogenous variation in political ideology caused by local television station ownership changes.
Working for Too Little: The Choice to “Create” (J2, Z1)
The current work is one of a handful to present an individual’s unique motivation for creative expression as an additional element for utility function inclusion. We posit that creative expression should be particularly important for employment and job choice decisions, and, therefore, focus the empirical portion of this work on examining how individuals identify in “artistic” versus other types of professions. Specifically, we examine how job-identification varies based upon employment status, and how individuals choose to transition more or less readily from artistic professions relative to other professions and depending upon the business cycle. Our approach is unique in our focus on both gender and time period to allow for important differences in creative motivation. We employ various data sources to substantiate our analysis, however, the present version of this analysis focuses on the United States Current Population Survey Merged Outgoing Rotation group (CPS-MORG) files. Our results provide additional evidence for the need to consider individual creative expression when designing future analyses of utility, and particularly in relation to sectoral employment decisions.
Workload, Time Use and Efficiency (M5, J2)
Task juggling –working on multiple projects at the same time– is generally thought to be detrimental to productivity (Coviello et al. 2014). Yet fluctuations in workload often require workers to juggle multiple projects at once. This paper studies under what conditions task juggling is optimal and can help workers handle high workload.
We present a model of time allocation in a multi-task context with short-term learning within projects and within steps across projects. We find that in heterogeneous contexts, where there is more learning within projects than within the same step across projects, it is optimal to work sequentially, completing one project before starting the next. In contrast, in homogeneous contexts, in which learning within the same step across projects is relatively stronger, it is optimal to juggle tasks by working in batches, completing the same step across projects. Output increases with workload in either context, but while timeliness may decrease in heterogeneous contexts, quality and timeliness are expected to increase in the homogeneous contexts, as workload increases the efficiency of batch work.
We provide empirical evidence of the theoretical predictions using detailed workload, productivity, time and internet use data of insurance claims examiners across two departments that handle heterogeneous and homogeneous claims respectively, and who face plausibly exogenous variation in workload. We show evidence consistent with examiners working sequentially in the heterogeneous context and working in batches in the homogeneous context. A one standard deviation increase in workload increases output by 2% in the heterogeneous context and 6.1% in the homogeneous context, while tardiness decreases and quality increases in the latter context only.
You Do Not Know the Value of Water before the Well Runs Dry - the Impact of Sustainable Development Goals on Firm Value (G3, Q5)
The contribution to each of the 17 Sustainable Development Goals (SDGs) is the next generation of measures for the sustainability of business models and firms. In this context we are the first to study the impact of the firm’s SDGs performance on its value using unique data of SDGs aligned products and services from more than 5,800 global firms. Comparing those firms disclosing their SDGs performance to 25,800 non-disclosing firms reveals significant differences in firm characteristics. Therefore, we jointly estimate a SDGs disclosure-choice model and integrate the results in a firm-value model. First, the disclosurechoice model identifies the determinants of SDGs disclosure. Second, the firm-value model studies the impact of both, aggregated and single SDGs performance measures on Tobin’s Q. Our results reveal that a positive contribution to specific SDGs, e.g. “combating hunger”, “attaining gender equality”, and “optimizing material use” have a significant negative, whereas “ensuring health” and “mitigating climate change” have a significant positive impact on firm value. The results remain robust after controlling for firms’ ESG scores and the SDGs performance of countries. We therefore recommend investors, asset managers, and firms to include the firm’s SDGs performance into their investment decision to assess firm value precisely.