Introducing a Series on Large and Complex Banks
This post is the first in a series of thirteen Liberty Street Economics posts on Large and Complex Banks. For more on this topic, see this special issue of the Economic Policy Review.
The chorus of criticism levied against mega-banks has, in some cases, outrun the research needed to back the criticism. To help the research catch up with the rhetoric, financial economists here at the New York Fed have engaged in a systematic study of the economics of large and complex banks and their resolution in the event of failure. The result of those efforts is a collection of eleven papers, each of which was subject to review (internal and external). The papers are now online in our Economic Policy Review. Today, we begin a two-week series of posts that present the key findings of each paper. Here, I’ll give a taste of each and some of the essential points delivered by them.
Before proceeding, the whole team would like to thank Darrell Duffie, Mark Flannery, Scott Frame, Joe Hughes, Jack Reidhill, David Skeel, and Phil Strahan for their help in refereeing the papers. Their detailed critiques, swiftly rendered, greatly improved our product. Mark Flannery deserves special thanks for refereeing multiple papers. Naturally, the referees aren’t to blame if you don’t agree with the papers.
Now to the content. While contributors took a wide-angle approach to the issue, the papers fit, without too much shoving, into one of three topics: size (and costs and benefits thereof), complexity (sources and measurement), and resolution. These three topics are, of course, closely related; given the size and complexity of a bank, whether it’s too big or complex to let fail depends on the resolution “technology” (DeYoung et al. 2001) available to regulators. The more efficient the technology, the larger and more complex the bank can be without its failure having systemic consequences. I discuss each paper in the order in which its associated blog post appears.
Bank Size, a Double-Edged Sword
Until recently, having mega-banks seemed like an unmitigated bad; they create systemic risk and there was little convincing evidence of economies of scale beyond a relatively small size. However, just in the last five years several papers have found scale benefits even for trillion-dollar banks. The first paper in the volume, “Do Big Banks Have Lower Operating Costs?” by Anna Kovner, James Vickery, and Lily Zhou, contributes to that recent literature by showing that bank holding company (BHC) expense ratios (noninterest expense/revenue) are declining in bank size. In a back-of-the-envelope calculation, the authors estimate that limiting BHC assets to 4 percent of GDP, as has been advocated, would increase noninterest expense for the industry by $2 billion to $4 billion per quarter. Breaking up mega-banks is not a free lunch.
The other edge of the sword, of course, is the potential funding advantages and moral hazard associated with being perceived as too big and complex to fail (TBTF). A paper by João Santos, “Evidence from the Bond Market on Banks’ ‘Too-Big-to-Fail’ Subsidy,” adds to the growing literature that tries to quantify the TBTF funding advantage, but Santos adds a twist; he tests whether all very large firms, including nonfinancial firms, enjoy a funding advantage. He finds that, in fact, the very largest (top-five) nonbank firms also enjoy a funding advantage, but for very large banks it’s significantly larger, suggesting there’s a TBTF funding advantage that’s unique to mega-banks.
Along with a funding advantage, being perceived as TBTF may create moral hazard. While it’s almost universally presumed that TBTF banks take excessive risk, recent research challenges that presumption; if the TBTF subsidy increases mega-banks’ franchise value, they may play it safe to conserve that value. In “Do ‘Too-Big-to-Fail’ Banks Take On More Risk?” Gara Afonso, João Santos, and James Traina test the moral hazard hypothesis using Fitch’s government support ratings as a proxy for TBTF status (a support rating reflects a rating agency’s views on the likelihood of government assistance for a systemically important bank). They find that a one-notch increase in support ratings is associated with an 8 percent (relative to average) increase in the impaired loan ratio, consistent with the traditional moral hazard story.
The takeaway from these three papers is that bank size has benefits and costs: The upside is the potential for economies of scale and lower operating costs; the downside is the TBTF problem and the attendant funding advantages and moral hazard.
Samuel Antill, David Hou, and Asani Sarkar take a different approach in a paper on size, “Components of U.S. Financial Sector Growth, 1950-2013,” as they document trends in the size of the financial sector as a whole over the last half-century. Among many interesting findings, the authors observe that the traditional banking sector (as opposed to the shadow banking sector) actually shrank by some measures, with growth in the banking sector mostly accounted for by large banks.
Measuring Bank Complexity
It’s been said that all science begins with measurement, and measurement is where our contributions on bank complexity make the greatest strides. We don’t definitively answer the question of why mega-banks are so complex, much less whether they’re overly complex, but we do take steps toward quantifying bank complexity by delineating its different dimensions, measuring them, and documenting their trends.
In “Evolution in Bank Complexity,” Nicola Cetorelli, James McAndrews, and James Traina study the acquisition patterns of banks and other financial firms over the last thirty years or so to document trends in organizational structure by considering the number and types of subsidiaries owned by a given bank or BHC. They find that in the early part of the sample, banks were mostly buying other banks, but in the later part, banks were increasingly buying nonbank financial firms, such as asset managers and broker dealers. The authors note that as banks became less bank-centric, they became more organizationally complex, which makes them harder to resolve and supervise. The study hypothesizes that bank complexity increased in response to an evolving intermediation technology centered around securitization.
In “Measures of Complexity of Global Banks,” Nicola Cetorelli and Linda Goldberg measure different dimensions of complexity of our seemingly most complicated banking firms: global banks. Using Bankscope data, they count the number of affiliates (organizational complexity), the number of nonbank affiliates (business complexity or scope), and number of countries each entity operates in (geographic complexity). Among the many findings Cetorelli and Goldberg uncover is the surprising fact that size and complexity don’t necessarily go hand-in-hand; while organizational complexity is highly correlated with bank size, business and global complexity are much less so. Said differently, “large and complex” is not redundant.
Adam Kirk, James McAndrews, Parinitha Sastry, and Phillip Weed next investigate the complexities of dealer banks’ “internal” financing obtained via rehypothecation of collateral. In “Matching Collateral Supply and Financing Demands in Dealer Banks,” they study three activities where dealer banks can lower funding costs by repledging collateral: matched-book transactions, internalization (for example, where short positions offset margined long positions), and collateral received/posted in connection with derivative transactions. Among other findings, the authors show that “collateral efficiency” (the share of collateral that’s rehypothecated) fell sharply during the crisis and has remained low since. They also find evidence of a type of economy of scale among dealer banks; those with larger collateral pools exhibit higher collateral efficiency, presumably because a larger collateral pool contains securities that match more customer demands.
Large Bank Failure and Resolution
In “Bank Resolution Concepts, Tradeoffs, and Changes in Practices,” Phoebe White and Tanju Yorulmazer provide a useful primer on large bank failure resolution. They catalogue the costs of large bank failures, including the external costs arising from contagion and fire-sale externalities that may make large bank failures a public policy concern. White and Yorulmazer argue that private resolution, either mergers or purchase and assumption, are preferred resolution methods because they may not entail costs to taxpayers and they preserve any franchise of the failed bank. Unfortunately, as the authors note, if a systemic crisis is under way, healthy buyers may not be available, so regulators face a tradeoff between liquidation costs and public bail-outs.
Michael Fleming and Asani Sarkar then provide a detailed post-mortem of the Lehman bankruptcy—one of the largest and most complex in history—with a focus on the settlement of Lehman’s OTC derivative book. Their paper, “The Failure Resolution of Lehman Brothers,” finds that creditors recovered 21 percent of allowable claims, a below-average amount for comparable firms given the state of the economy and credit markets. Fleming and Sarkar attribute the lower payout in part to poor bankruptcy planning by Lehman. Liquidity provided by the Fed, they argue, mitigated losses for creditors. The paper will help inform the debate over the efficacy of traditional bankruptcy (Chapter 11) for large financial firms.
The last two papers in the volume provide rationales for bail-in, where the claims of creditors of the parent company are converted to equity in resolution.
In “Why Bail-in? And How!” Joseph Sommer argues that priority implicit in bail-in, with the creditors of the parent junior to the creditors of the subsidiaries, is efficient because the subsidiaries tend to issue financial liabilities that provide liquidity or risk shifting, extra services that the claims against the parent don’t provide. By providing loss absorption that protects those financial liabilities, bail-in helps preserve the value of those services.
In “What Makes Large Bank Failures So Messy and What to Do about It?” James McAndrews, João Santos, Tanju Yorulmazer, and I also argue for bail-in as a useful part of the resolution “technology,” and we provide a rationale for bail-in-able debt versus equity. The key analytical issue is how to model the behaviors of the resolution authority and those who can run the bank. We show numerically and algebraically that if the resolution authority waits until equity is depleted before closing a failing bank, requiring x in long-term debt and x in equity is better, in terms of preventing messy bank failures, than requiring 2x in equity. In the latter case, financial liability holders anticipate that there will be nothing to protect them from losses in resolution, so they’ll run. In the former case, financial liability holders are less likely to run because they realize when the bank is closed (when losses approach x), there will still be x in long-term debt to provide loss absorption; the bail-in-able debt provides a form of “equity in resolution.”
All of the authors and I hope you find these papers and the associated blog posts interesting and informative.
The views expressed in this post are those of the author and do not necessarily reflect the position of the Federal Reserve Bank of New York or the Federal Reserve System. Any errors or omissions are the responsibility of the author.
Donald P. Morgan is an assistant vice president at the Federal Reserve Bank of New York.