What happens when a bank fails? In the decades prior to the financial crisis of 2007, there was an average of just five bank failures per year, a rarity that made it easy to overlook how important resolving failures can be to ensure the banking system’s stability. The crisis brought this question to the fore: bank failures spiked, reaching 157 failures in 2010.
Failed banks are taken over by the FDIC, which can either bail out, liquidate, or sell them. But since a rash of bank failures occurred in 1987, virtually no research has been conducted into what happens when banks fail. Booth School Associate Professor of Finance Gregor Matvos, with colleagues Amit Seru and Joao Granja of MIT, undertook the task of establishing basic facts about who buys failed banks and what forces determine how much the FDIC ends up paying as a result.
Matvos points out that bank failures are a special case. When private firms fail, the government doesn’t step in to help. But because banks play an outsized role in the economy and are difficult to establish, the human capital and the relationships developed within a bank are considered worth protecting with public resources. Selling failed banks maintains the bank’s benefits without putting taxpayers on the hook for a complete bailout. Between 2007 and 2013, 95 percent of failed banks were sold. Selling banks is, however, costly; the FDIC spent approximately $90 billion insuring deposits at failed banks during the crisis.
In previous decades, regulations against interstate banking prohibited banks from being sold across state lines. But despite a relaxation of regulations and the advancement of communications technologies which make interstate bank operations feasible, Matvos finds that bank sales continue to be extremely local. A third of failed banks are sold to a bank that has a branch in the same zip code, 54 percent go to banks within the same county, 84 percent stay within the state.
Matvos theorizes that most banks consider the value of a failed bank its soft information—its human capital and local relationships—which is difficult to take advantage of at a distance and dissuades distant potential buyers. Likewise, banks whose asset portfolios and specializations are similar to those of a failed bank are far more likely to buy because it’s easier to take advantage of economies of scale. For a bank whose primary business is commercial real estate, buying a bank that focused on consumer auto loans might not make much sense.
However, while the ideal buyer may be a local bank operating with a similar portfolio, in many scenarios a failed bank is the victim of a shock that has hurt all banks within the region. Better capitalized banks are more likely to buy failed banks, but a shock, like a substantial dip in real estate values, often leaves the local competitors best suited to buying the failed bank poorly capitalized. In these scenarios Matvos finds that the failed bank gets sold to a bank farther away and in a different line of business, less than ideal matching that frequently leads to higher costs.
Indeed, the costs incurred by the FDIC in selling a bank vary widely as a result of these competing forces. The average cost, at approximately 28 percent of the failed bank’s assets, is quite high and is positively associated with the distance at which the bank gets sold. Selling banks is thus not a miracle cure to resolving failed banks. It may spare taxpayers the full cost of bailing out a bank, but still requires substantial public resources. The fire sale nature of the sales often lead to unideal buyers and higher costs for the FDIC and taxpayers.