Academic journal article Federal Reserve Bank of New York Economic Policy Review

Do "Too-Big-to-Fail" Banks Take on More Risk?

Academic journal article Federal Reserve Bank of New York Economic Policy Review

Do "Too-Big-to-Fail" Banks Take on More Risk?

Article excerpt

* Large or complex banks might have a greater appetite for risk if they expect future rescues.

* Using data for more than 200 banks in 45 countries, the authors find higher levels of impaired loans after an increase in government support, as measured by Fitch Ratings' support rating floors (SRFs).

* A one-notch rise in the SRF increases an average bank's impaired loan ratio by roughly 8 percent; the authors show similar effects on net charge-offs and for U.S. banks only.

* The authors also show that riskier banks are more likely to take advantage of potential government support.

* The findings suggest that banks classified by rating agencies as more likely to receive government support engage in more risk taking.


In 1984, U.S. regulators made the unprecedented move of insuring all of Continental Illinois's liabilities. The Comptroller of the Currency indicated during the hearings after Continental's resolution that regulators would not allow the eleven largest banks in the Unites States to fail. Ever since, there have been many concerns with banks deemed "too big to fail." (1)

These concerns derive from the belief that the too-big-to-fail status gives large banks a competitive edge and incentives to take on additional risk. If investors believe the largest banks are too big to fail, they will be willing to offer them funding at a discount. Together with expectations of rescues, this discount gives the too-big-to-fail banks incentives to engage in riskier activities. This, in turn, could drive the smaller banks that compete with them to take on further risks, exacerbating the negative effects of having too-big-to-fail banks in the financial system.

The debate around too-big-to-fail banks has given rise to a large literature. Part of this literature attempts to determine whether bank investors, including depositors, believe the largest banks are too big to fail. Some studies seek to answer this question by investigating spreads on bank bonds (Flannery and Sorescu 1996; Sironi 2003; Morgan and Stiroh 2005; Anginer and Warburton 2010; Balasubramnian and Cyree 2011; Santos, forthcoming). Other studies consider spreads on bank credit default swap contracts (Demirguc-Kunt and Huizinga 2013; Li, Qu, and Zhang 2011), bank stock returns (Correa et al. 2012), and deposit costs (Baker and McArthur 2009). Yet others focus on the premiums that banks pay in mergers and acquisitions (Brewer and Jagtiani 2007; Molyneux, Schaeck, and Zhou 2011).

Another part of that literature investigates whether too-big-to-fail banks behave differently by looking at balance-sheet data (Gropp, Hakenes, and Schnabel 2011), syndicated loans (Gadanecz, Tsatsaronis, and Altunbas 2012, and bank z-scores (Brandao Marques, Correa, and Sapriza 2013), among other measures.

Our paper is closer to the latter studies in that we are also interested in finding out whether the too-big-to-fail status affects bank behavior. Specifically, we study whether banks that rating agencies classify as likely to receive government support increase their risk-taking.

An important novelty of our paper is the way we measure the likelihood of a bank receiving government support. Previous studies, including Haldane (2010), Lindh and Schich (2012), and Hau, Langfield, and Marques-Ibanez (2013), attempt to infer support from the difference between Moody's all-in credit ratings (long-term bank deposit ratings, which capture a bank's ability to repay its deposit obligations and include external support) and Moody's stand-alone ratings (bank financial strength ratings, which exclude external support). The difference between Moody's all-in credit and stand-alone ratings is commonly known as a ratings "uplift" Using uplifts, however, presents two potential issues. First, a change in uplift may arise from movement in either of the two underlying ratings, with completely different implications. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed


An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.