Research Highlights Featured Chart

April 8, 2025

Still too big to fail?

How bank bailout expectations changed after the global financial crisis

Tyler Smith

Source: David Shankbone, CC BY 3.0

As the 2008 financial crisis threatened to collapse the global economy, governments around the world stepped in to rescue struggling banks deemed "too big to fail." This extraordinary intervention raised concerns that large financial institutions could expect taxpayer-funded bailouts in the future, potentially subsidizing their reckless behavior and shielding them from the costs.

In a paper in the American Economic Review, authors Antje Berndt, Darrell Duffie, and Yichao Zhu provide evidence that instead of increasing after the Global Financial Crisis, market expectations of a government bailout for globally systemically important banks (GSIBs) actually decreased.

They drew their results from a structural model of bank valuation, unlike previous studies of bank bailouts which used reduced-form methods. The authors calibrated their model using financial statements and market prices of debt, equity, and equity options for six major US GSIBs, including Bank of America, Citigroup, Goldman Sachs, JP Morgan Chase, Morgan Stanley, and Wells Fargo.

Figure 3 from the authors’ paper shows their estimates of the “no-bailout” probability—the probability that GSIBs would not be bailed out by the government if they became insolvent—between 2001 and 2023.

Figure 3 from Berndt et al. (2025)

The dark blue line represents the estimated no-bailout probability for each month, while the shaded light blue area indicates the range between the fifth and ninety-fifth percentiles of the bootstrap distribution of this probability, indicating the statistical confidence in the estimates. The gray vertical bands indicate recessions.

The chart depicts a dramatic and persistent increase in no-bailout probabilities after the 2008 global financial crisis. Prior to the crisis, markets implied only about a 5 percent chance that these massive banks would not be rescued if they approached insolvency. After the crisis, this probability jumped to around 10 percent, eventually climbing to as much as 20 percent. 

This shift in bailout expectations aligns with post-crisis regulatory reforms, including the Dodd–Frank Act's establishment of new insolvency resolution mechanisms aimed at forcing creditors, rather than taxpayers, to bear losses when large banks fail.

The authors’ model still implies a high probability, roughly 87 percent on average, that GSIBs will be bailed out. But their research provides some quantitative evidence that too-big-to-fail distortions in the US banking system have moved in the right direction.