Reframing Financial Regulation

December 14, 2016

The financial system is a critical engine that keeps the economy vibrant. Serving consumers faced with unanticipated expenses, investors planning their futures, and small businesses looking to expand, it creates economic opportunities for all participants in society. In recent years, however, worries about too-big-to-fail and too-small-to-borrow have led many to question whether financial markets are working the way they should.

Designing a regulatory framework that ensures the resilience of the financial system and supports economic growth and individual prosperity has proven elusive to policymakers and interested observers. Every several years, policymakers scramble to re-regulate the financial system, particularly after a financial crisis has occurred. Too often, new “this time we really mean it” regulations are simply added onto existing ones, which makes for an unwieldy regulatory framework. Real reform requires reimagining the framework from the ground up, a task this book undertakes.

Reframing Financial Regulation: Enhancing Stability and Protecting Consumers, edited by Hester Peirce and Benjamin Klutsey, brings together a diverse set of authors to provide alternative ways to regulate different aspects of the financial system. The chapters embody approaches that rely less on centralized, top-down regulations and more on market discipline and oversight. The book, which reflects a wide variety of viewpoints and approaches, seeks to initiate a lively conversation about how a thoughtfully regulated, market-based financial system can facilitate risk sharing, efficiently provide access to capital, and enable households to save for the future.

About the Editors

Hester Peirce is a senior research fellow and the director of the Financial Markets Working Group at the Mercatus Center at George Mason University. Before joining the Mercatus Center, she served as a senior counsel to the Republican staff on the Senate Committee on Banking, Housing, and Urban Affairs. Before that, she served as counsel to Commissioner Paul S. Atkins at the US Securities and Exchange Commission and as a staff attorney in the Division of Investment Management at the US Securities and Exchange Commission. She earned her BA in economics from Case Western Reserve University and her JD from Yale Law School.

Benjamin Klutsey is the program manager for the Financial Markets Working Group and the Program on Monetary Policy at the Mercatus Center at George Mason University. Klutsey was previously with the Institute of International Finance, where he served as policy assistant analyzing international financial regulations related to liquidity risk management and risk governance. He received his MA in international commerce and policy from George Mason University and his BA in government and philosophy from Lawrence University.

Table of Contents

Click each chapter title to download the PDF.

Introduction: Market-Based Financial Regulation
Hester Peirce and Benjamin Klutsey

Part 1: Regulating Loss Absorption

Chapter 1: Risk-Based Capital Rules
Arnold Kling | Mercatus Center at George Mason University

Chapter 2: On Simpler, Higher Capital Requirements
Stephen Matteo Miller | Mercatus Center at George Mason University

Chapter 3: Alternatives to the Federal Deposit Insurance Corporation
Thomas L. Hogan | Johnson Center for Political Economy, Troy University
Kristine Johnson | Mercatus Center MA Fellow at George Mason University (Alumna)

Part 2: Addressing Failure

Chapter 4: Title II of Dodd-Frank
Peter J. Wallison | American Enterprise Institute

Chapter 5: The Rise of Bail-Ins and the Quest for Credible Laissez-Faire Banking
Garett Jones | George Mason University

Part 3: Regulating Securities and Derivatives Markets

Chapter 6: US Broker-Dealer Regulation
Hon. Daniel M. Gallagher | Patomak Global Partners, LLC, and Former Commissioner, US Securities and Exchange Commission

Chapter 7: Reconsidering the Dodd-Frank Swaps Trading Regulatory Framework
Hon. J. Christopher Giancarlo | Commissioner, US Commodity Futures Trading Commission

Chapter 8: Rethinking the Swaps Clearing Mandate
Hester Peirce and Vera Soliman | Mercatus Center at George Mason University

Chapter 9: The Past and Future of Exchanges as Regulators
Edward Stringham | Trinity College

Chapter 10: Using the Market to Manage Proprietary Algorithmic Trading
Holly A. Bell | University of Alaska Anchorage

Chapter 11: Offering and Disclosure Reform
David R. Burton | The Heritage Foundation

Part 4: Regulating Consumer Finance

Chapter 12: Market-Reinforcing versus Market-Replacing Consumer Finance Regulation 
Todd J. Zywicki | Antonin Scalia Law School, George Mason University

Chapter 13: Examining Arguments Made by Interest Rate Cap Advocates
Thomas W. Miller Jr. | Mississippi State University
Harold A. Black | University of Tennessee, Knoxville (Emeritus)

Part 5: Facilitating Innovation

Chapter 14: Regulating Bitcoin—On What Grounds?
William J. Luther | Kenyon College

Chapter 15: Financial Technology 
Houman B. Shadab | Center for Business and Financial Law, New York Law School

Chapter 16: Ending the Specter of a Federal Corporate Law 
J. W. Verret | Antonin Scalia Law School, George Mason University

Part 6: Improving the Regulatory Process

Chapter 17: Is Regulatory Impact Analysis of Financial Regulations Possible?
Jerry Ellig and Vera Soliman | Mercatus Center at George Mason University

Alternatives to the Federal Deposit Insurance Corporation

January, 2016

Federal deposit insurance creates moral hazard that encourages risky banking practices and sets the stage for bank failures and financial crises. Alternatives to our current scheme include the creation of a more stable, privatized deposit insurance system, and the strengthening of market discipline through the lowering of mandated coverage levels, or doing away with deposit insurance requirements altogether.

Repealing Banking Regulations Is the Best Path to Financial Stability

Monday, January 5, 2015

Many people blame the recent financial crisis on a lack of regulation and fraud in the financial system. However, the Federal Reserve appears to have been a significant contributor to the crisis in terms of both its poor monetary policy and faulty regulation of the financial system.

In contrast to the current approach of regulators attempting to outsmart the bankers, increasing competition by reducing the restrictions on banking activities and ending the special protections granted to privileged financial institutions will help increase financial intermediation and improve monetary stability. The best path to financial stability is to free the banking system from the current burden of inefficient and costly regulations.

Free and unregulated banking systems provide the financial intermediation consumers want while simultaneously reducing banks’ incentives to take risk. These benefits can be captured in spite of – indeed, because of – a lack of government intervention in the financial system. In a free banking system, banks are allowed to issue their own banknotes. These notes act as liabilities against each bank’s assets. Anyone who holds the banknote can redeem it or “cash it in” for a particular asset specified in advance, such as a regular U.S. dollar or possibly for a commodity such as gold.

Continue reading

American Manufacturing Is Not in Decline

Wednesday, December 24, 2014

While visiting a steel plant in Indiana in October, President Obama described the recent gains in American manufacturing employment as a reversal from previous decades when "everybody said American manufacturing is going downhill." But has manufacturing really been going downhill? Not exactly.

News outlets including CNN, NPR and The Washington Post have also lamented the decline of U.S. manufacturing. But what do they mean by a "decline" in manufacturing?

The U.S. manufacturing sector produces more stuff almost every year than it did the year before. Prior to the 2008 recession, we manufactured more stuff in the United States than had ever been manufactured by any country, ever, in the history of the world. Is that decline?

After a slight dip from the recent recession, the United States is once again at an all-time high. We are on pace to manufacture more stuff in 2014 than we have in any year in all of U.S. history. Is that decline?

Most Americans, including the president, have a misperception of decline because manufacturing employment has fallen in recent decades. Although it is true that the number of jobs in manufacturing has gone down, there are two reasons this change should not be thought of as "decline."

First, while the number of U.S. manufacturing jobs has fallen by almost 7.5 million jobssince its peak in 1979, this loss is modest compared to the jobs created over that time. The U.S. economy creates and loses almost 5 million jobs every month. The number of non-manufacturing jobs created in the last few months alone is more than the number of U.S. manufacturing jobs lost in the past 35 years.

Second, American workers have moved out of manufacturing and into industries where they are most needed. Yet despite having fewer workers, the manufacturing sector still keeps producing more stuff every year. Producing more stuff with less work is not decline.

Think about it: If "decline" means working less and making more, then I want to be in decline. We should all want to be in decline. We should want every industry to be in decline!

Another reason for the talk of "decline" is that in the last few years, China (which has four times the population of the United States) has finally started manufacturing more stuff than we do. But despite being the second largest manufacturer in the world rather than number one, we still produce more stuff today than we ever have before. That is hardly decline.

Many Americans worry that foreigners are "stealing our jobs." As the president said, they're "looking to bring jobs back from China." But the jobs lost to China are mostly low-skilled, low-paying jobs. We should be glad those jobs are moving overseas as long as we keep creating higher-paying jobs here at home. That's not decline.

Plus, as China gets richer, its people buy more stuff from the United States. Much of the equipment used by China's manufacturing sector is made in the U.S.A. Making good high-quality products in the United States while the cheap stuff is made abroad is not a sign of decline.

No, Mr. President, American manufacturing is not in decline.

Bernanke Policies at Fed Broke His Own Rules Of Central Banking

Friday, February 7, 2014

As Ben Bernanke's tenure at the Federal Reserve drew to a close, many wondered how history will remember this controversial chairman. Although he is widely respected as an academic, economists are divided regarding Bernanke's actions during the 2008 financial crisis.

Some think the Fed did everything by the book, while others contend its bank bailouts were unprecedented, possibly bordering on illegal. Former Fed Chairman Paul Volcker, for example, said the Fed's actions "extend to the very edge of its lawful and implied powers, transcending certain long-embedded principles and practices."

Did Bernanke save us from another Great Depression, or did he sell out to Wall Street by bailing out the big banks?

One way to evaluate his performance is to ask what course of action most economists recommend in a crisis and see if Bernanke followed their advice. In fact, Bernanke himself has proposed exactly that. He claims to have followed the standard playbook for financial crises: the rules for a classical lender of last resort set forth by 19th-century economist Walter Bagehot.

Bagehot's rules are considered the most effective guidelines for central bankers in times of panic. In his famous book "Lombard Street," Bagehot proposed a course of action for the Bank of England in response to banking crises.

Bagehot was not an advocate of central banking, but given the role taken on by the Bank of England as a central banker, he believed that the bank should act as the lender of last resort in times of panic. Bagehot had four rules for a lender of last resort:

1. The central bank should lend freely to solvent banks. As long as they are backed by sound collateral, there should be no limit on the loan amounts.

2.The central bank should only provide last-resort loans at a high rate of interest. This penalty rate serves as a self-selection mechanism so only the banks that are truly in need of funds seek them.

3. The central bank should only lend to illiquid but fundamentally solvent institutions. During a crisis, the central banker is under pressure because many banks are short on liquidity. However, the central bank should only make loans that it expects to be repaid in the future.

4.The central bank should announce its policies before any crisis takes place. This creates an expectation that the central bank will help to stabilize the banking system in future financial crises.

It is clear that the Fed's bailouts deviated significantly from Bagehot's recommendations. The Fed did not charge a high penalty rate of interest on its loans to at-risk banks. To the contrary, the Fed lowered its lending rates during the crisis.

Nor did the Fed deal only in good securities. Rather, it freely accepted risky assets such as mortgage-backed securities in order to get these toxic assets off banks' balance sheets and replace them with safe, liquid assets.

Did the Fed announce its policy in advance? Hardly. It bailed out Bear Stearns, then let Lehman Brothers fail, and then bailed out everyone else.

Perhaps the most important deviation from Bagehot's rules is the Fed's bailout of insolvent financial institutions.

The Fed made three rounds of "emergency" loans to American International Group before taking over almost 80% of the firm's equity. Citigroup received its third bailout just before the firm went bankrupt, after its stock price fell from 57 before the crisis to less than 1 in 2009. After coercing Bank of America to buy a failing Merrill Lynch, the Fed had to write guarantees on Merrill's assets in order to keep BofA afloat.

If you believe Bernanke, this was all just part of the master plan.

Were the Fed's methods successful? It depends on the exact definition of success. Judging by the yardstick of averting another Great Depression, the Fed's actions were clearly successful. However, by Bernanke's own metric of comparing the Fed's actions to Bagehot's rules, they surely were not. Bernanke failed by the standard he himself selected.

Perhaps Bernanke will be remembered as neither a sellout nor a savior — but merely as a politician, saying one thing and doing another. And as any politician would say, he had only our best interests at heart.

The Failure of Risk-Based Capital Regulation

January 31, 2013

Capital requirements are a primary component of US banking regulation. Since 1991, the Federal Reserve has used a “risk-based” method of capital regulation that attempts to account for the riskiness of various types of bank assets. However, evidence shows that this system has increased, rather than decreased, risk in the US banking system.

In this policy brief, we explain the fundamentals of risk-based capital (RBC) regulation and discuss some potential shortcomings of this system. We propose that the Fed end its use of RBC regulation and return to the use of simple capital ratios as measures of bank risk.

BANK CAPITAL REGULATION

Bank equity, or “capital,” functions as a cushion against unexpected losses in the value of bank assets. Equity, or capital, is the value of the bank to its investors, which is calculated as the current value of the bank’s assets minus its liabilities. The greater a bank’s capital, the greater the loss in asset value it can absorb before becoming insolvent. From 1980 to 1991, bank regulators required banks to maintain some minimum level of capital as a percentage of total assets. This standard did not differentiate between different types of bank assets in terms of risk.

In 1991, the Fed introduced a new system of RBC regulations based on the international agreement known as the Basel Accords[1]. These standards were intended to “enhance the resiliency and stability of the banking and financial system.”[2] RBC standards assign a risk weight to each type of bank asset so that banks with greater quantities of risk face a higher capital requirement. The RBC ratio, the Fed’s primary metric for measuring bank risk, is calculated as bank capital divided by risk-weighted assets (RWA). Since RWA is in the denominator of this formula, holding a greater number of risky assets causes the RBC ratio to fall, while holding fewer risky assets causes the RBC ratio to rise.

As often happens with consumer protection laws, the actual effects of RBC regulation have been the opposite of those intended. Rather than limiting banks’ risk, these rules give banks an incentive to acquire risky assets that are not rated properly by the regulators. The system also increases systemic risk by encouraging all banks to hold the same types of assets, thereby reducing diversification and increasing fragility in the banking system. Proponents of RBC regulation claim that these costs are outweighed by the Fed’s improved ability to identify risky banks. However, several studies which will be discussed below have shown that the RBC ratio is actually inferior to the standard capital ratio as a predictor of bank risk. 

HOW RBC REGULATION WORKS

RBC standards divide each bank’s assets into four categories, based on their levels of risk. Each category is assigned a different “risk weight.” The assets that regulators perceive as safe, such as cash and government bonds, receive a 0 percent risk weight. Slightly riskier assets, such as securities issued by government agencies, are assigned a 20 percent risk weight. Mortgages, considered riskier still, are weighted in the 50 percent category. All other assets, such as corporate bonds and commercial loans, receive a 100 percent risk weight. The weighted sum of all four asset categories is the bank’s RWA. The RBC ratio is calculated as the bank’s capital divided by its RWA.

The RBC ratio is intended to provide a better estimate of bank risk than the simple capital ratio does. Since the RBC ratio is bank capital divided by RWA, banks holding more risky assets must maintain higher levels of capital to maintain the same RBC ratio. When RBC regulations were implemented in 1991, banks were required to maintain a minimum RBC ratio of 8 percent, up from the previous minimum non-RBC capital ratio of 6 percent.[3] Since regulators expected RBC regulations to increase banks’ levels of capital, they allowed for a period of “adjustment by banks who need time to build up to those levels.”[4] In retrospect, however, it is not clear that replacing the unweighted capital ratio with the RBC ratio should have encouraged banks to hold more capital or reduce their holdings of risky assets.

Consider the simple example of a bank with assets, liabilities, and equity, as shown in figure 1. The bank’s capital ratio is its equity of $60 billion divided by total assets of $1 trillion, yielding a ratio of 6 percent. In 1990, this bank would have met the minimum level of 6 percent capital required of all banks. But what would happen in 1992, when the bank is required to maintain an RBC ratio of 8 percent? Let us assume that the bank’s cash assets receive a 0 percent risk weight, safe assets receive a 20 percent risk weight, and risky assets receive a 100 percent risk weight. This bank’s total RWA are calculated as follows:

RWA­=­(100­×­0.0)­+­(500­×­0.2)­+­(400­×­1.0)­=­$500 ­billion.

With capital of $60 billion divided by total RWA of $500 billion, the bank’s RBC ratio would be 12 percent. This is above the required minimum of 8 percent, so the bank could reduce its RBC while still meeting the minimum required level—either by converting its cash holdings to holdings of risky assets, or by increasing its liabilities without increasing capital. Thus, moving from the standard capital regulation to the RBC regulation would increase, rather than decrease, the bank’s level of risk.

Let us consider another case that might have similar results. RBC regulation relies on the ability of bank regulators to properly assess the relative riskiness of a wide and varied array of bank assets. If regulators over- or underestimate the risk of a particular type of asset, those assets will be assigned an improper risk weight. Banks can then profit from “regulatory arbitrage” by the buying or selling of the misrated asset. Mortgage-backed securities (MBS), for example, were thought in the early 1990s to be relatively safe but are now considered to be very risky and are widely viewed as a major cause of the recent financial crisis. With the adoption of RBC regulation in 1991, however, MBS were assigned low risk weights of 20 percent or even 0 percent. Banks were therefore able to increase their profits by acquiring these high-risk, high-return assets while simultaneously reducing their RBC ratios.

To see how this worked, consider the change from the balance sheet in figure 1 to that of figure 2. Suppose the bank in figure 1 sells $250 billion worth of corporate bonds that are assigned a risk weight of 100 percent. The bank uses the funds to buy $250 billion in MBS. Using the same calculations as in the previous example, with MBS receiving a 20 percent risk weight, we find that the bank’s capital ratio is unchanged but that its RBC ratio has increased to 20 percent. Because MBS were misrated by the regulator, the bank is able to increase its RBC ratio while also increasing its holdings of risky assets. As described in the next section, this is what occurred in the US banking system during the 1990s and the first few years of the 21st century. In a similar situation, the incentive for banks to hold “safe” government bonds caused the Greek government’s debt crisis to destabilize the financial system of the entire European Union.

Banks are required to fulfill two other capital requirements in addition to the regular RBC ratio. These requirements both pertain to “Tier 1 capital,” which includes common equity, some preferred equity, and interest in subsidiaries less goodwill.[5] First, banks must maintain Tier 1 capital equal to or greater than 4 percent of RWA. Second, they must have Tier 1 capital equal to or greater than 4 percent of total assets, a requirement similar to the standard capital ratio used before 1991. Despite these failsafes, the Fed’s RBC regulations may still fall victim to the scenarios outlined in the examples above. Considering the conflicting incentives inherent in the system, it is impossible to accurately predict the effects of RBC regulation.[6] Studies of banking regulation since the adoption of the Basel Accords may be helpful in judging whether RBC regulation has reduced risk in the banking system. 

STUDIES OF RBC REGULATION

With the second round of RBC regulation in 2007, the Fed claimed that “the advanced approaches of Basel II are a significant improvement” that would enhance “bank safety and soundness and overall financial stability.”[7] However, RBC regulation has done just the opposite, according to recent evidence thoroughly documented in the book What Caused the Financial Crisis, edited by Jeffrey Friedman.[8] The first chapter, written by Friedman, provides evidence that the misrating of MBS did in fact cause banks to increase their holdings of MBS. Chapters by Viral Acharya and Matthew Richardson[9] and by Juliusz Jablecki and Mateusz Machaj[10] demonstrate how the securitization and acquisition of MBS created systemic risk in the banking sector.

Supporters of the Basel system propose that, despite these shortcomings, the RBC ratio can improve the accuracy of bank regulation by identifying particularly risky banks. Indeed, a 1991 study by Robert Avery and Allen Berger finds that RBC regulations were superior to pre-1991 capital regulations as predictors of bank risk and performance.[11] However, recent studies directly comparing the capital and RBC ratios tend to find that the standard capital ratio is a better indicator of risk than the RBC ratio. Arturo Estrella, Sangkyun Park, and Stavros Peristiani find that “the risk-weighted ratio does not consistently outperform the simpler ratios, particularly with short horizons.”[12] Asli Demirgüҫ-Kunt, Enrica Detragiache, and Ouarda Merrouche find that the capital ratio performs better than the RBC ratio as a predictor of bank stock returns, which, since the financial crisis, have been strongly related to bank risk.[13]

Our own study, “Evaluating Risk-Based Capital Regulation,” reconsiders the evidence provided by Avery and Berger.[14] Using more recent data, we find that the standard capital ratio is significantly better than the RBC ratio as an indicator of bank risk and performance and that using both ratios simultaneously does not produce better results. Taken in conjunction with the other available evidence, our findings indicate that RBC regulations lead to more risk-taking by individual banks, and more overall risk in the banking system, without improving the effectiveness of the Fed’s capital regulations. 

CONCLUSIONS

Since 1991, the Federal Reserve has employed a risk-based measure of bank capital as its primary tool for regulating bank risk. However, RBC regulations are easily exploited and susceptible to regulatory arbitrage. Evidence indicates that such regulations have increased individual bank risk as well as systemic risk in the banking system. In addition, RBC regulations do not appear to improve the Fed’s identification of risky banks, even when used in conjunction with the standard capital ratio. On these grounds, we propose that the Fed end its use of RBC regulation and return to the capital ratio as the primary basis for bank regulation.

ENDNOTES

1. Basel Committee on Banking Supervision, “International Convergence of Capital Measurement and Capital Standards,” Bank for International Settlements (July 1998).

2. Randall S. Kroszner, “Statement by Governor Randall S. Kroszner,” press release, Board of Governors of the Federal Reserve System, November 2, 2007, http://www.federalreserve.gov/newsevents/press/bcreg/krosznerbasel20071102.htm.

3. Total capital typically includes common and preferred equity, subordinated debt, and interest in consolidated subsidiaries less goodwill. See Robert B. Avery and Allen N. Berger, “Risk-Based Capital and Deposit Insurance Reform,” Journal of Banking & Finance 15 (1991): 851–53.

4. Basel Committee, “International Convergence,” 14.

5. See Avery and Berger, “Risk-Based Capital,” 852–53.

6. “This literature produces highly mixed predictions, however, regarding the effects of capital regulation on asset risk and overall safety and soundness for the banking system as a whole.” David VanHoose,  “Theories of Bank Behavior Under Capital Regulation,” Journal of Banking & Finance 31 (2007): 3680.

7. Randall S. Kroszner, “Statement by Governor Randall S. Kroszner,” press release, Board of Governors of the Federal Reserve System, November 2, 2007, http://www.federalreserve.gov/newsevents/press/bcreg/krosznerbasel20071102.htm.

8. Jeffrey Friedman, ed., What Caused the Financial Crisis (Philadelphia: University of Pennsylvania Press, 2011). For a summary, see Thomas L. Hogan’s “Book Review of Friedman, Jeffrey, ed., What Caused the Financial Crisis,” Reason Papers 34, no. 2 (October 2012): 222–28.

9. Viral V. Acharya and Matthew Richardson, “How Securitization Concentrated Risk in the Financial Sector,” in What Caused the Financial Crisis, ed. Jeffrey Friedman (Philadelphia: University of Pennsylvania Press, 2011), 183–99.

10. Juliusz Jablecki and Mateusz Machaj, “A Regulated Meltdown: The Basel Rules and Banks’ Leverage,” in What Caused the Financial Crisis, ed. Jeffrey Friedman (Philadelphia: University of Pennsylvania Press, 2011), 200–27.

11. Avery and Berger, “Risk-Based Capital.”

12. Arturo Estrella, Sangkyun Park, and Stavros Peristiani, “Capital Ratios and Credit Ratings as Predictors of Bank Failures,” FRBNY Economic Policy Review (July 2000): 33–52.

13. Asli Demirgüҫ-Kunt, Enrica Detragiache, and Ouarda Merrouche, “Bank Capital: Lessons from the Financial Crisis” (IMF Working Paper No. 10/286, Washington, DC, 2010), http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1751399.

14. Thomas L. Hogan, Neil Meredith, and Xuhao Pan, “Evaluating Risk-Based Capital Regulation,” (Mercatus Working Paper, Arlington, VA: Mercatus Center at George Mason University, January 2013).

Evaluating Risk-Based Capital Regulation

January 29, 2013

I. Introduction

The standard capital ratio of equity over assets has long been used as an important indicator of bank risk. Banks with more equity are less affected by asset depreciation than are other banks because a drop in the value of their assets affects only their equity and not their liabilities. In response to the savings and loan crisis of the 1980s, the Federal Reserve adopted risk-based capital (RBC) regulations in 1991 based on the Basel Committee on Bank Supervision (1988, hereafter “Basel Accords”) to improve the effectiveness of US banking regulation and to standardize the US banking system with other systems around the world. The RBC ratio measures equity as a percentage of risk-weighted assets (RWA); each category of assets is assigned a weight appropriate to its perceived level of risk. Banks with riskier assets must maintain more capital, while banks with safer assets require less capital. However, if this method of assessing risk is flawed, its use may increase, rather than decrease, systemic risk in the banking system.

Critics of the Basel system have pointed out several ways in which RBC regulation has increased risk in the banking system.[1] First, RBC can encourage risk-taking by individual banks, especially if regulators have not properly identified the riskiness of a particular class of assets. Jablecki (2009, 16) shows that the misrating of risky assets in the Basel Accords has encouraged US banks to adopt “regulatory capital arbitrage techniques, in particular securitization.” This problem cannot be resolved simply by reevaluating the assets’ risk weightings, because regulators themselves cannot be certain of all potential risks. Indeed, regulators seriously underestimated the riskiness of mortgage-backed securities (MBS), which were viewed in the 1980s as safe, low-risk assets but are now recognized as very high in risk. Second, risk- weighting systems can create systemic risk by encouraging many banks to invest heavily in the same class of assets. Friedman’s (2011) work explains how RBC regulations give banks an incentive to hold certain classes of risky assets, such as MBS and Greek government bonds, and that this approach has increased systemic risk in the United States and the European Union respectively.[2]

It is possible, however, that the benefits of RBC regulation might outweigh the potential costs. Some studies, such as Estrella, Park, and Peristiani’s (2000), have proposed that optimal banking regulation might utilize some combination of capital and RBC regulations. The Fed currently employs such a system (discussed further in the next section), which is based on the Basel Accords. If the RBC ratio is, in fact, an effective predictor of bank risk, then the RBC ratio might help regulators identify particularly risky banks, and this advantage might offset the disadvantage of increased systemic risk. However, if the RBC ratio does not improve the predictive power of the capital ratio, then RBC regulation may cause significant harm without providing any added benefit. Therefore, we must turn to the empirical question of whether or not the RBC ratio is better than the capital ratio as a predictor of bank performance.

Avery and Berger’s (1991) paper is among the first empirical studies to validate the use of RBC as a measure of bank performance, and it is widely cited in support of RBC regulation.[3] This study uses FDIC Call Report data to calculate the RWA of all US commercial banks from 1982 to 1989. The authors find that RWA is correlated with several indicators of bank performance, such as income, nonperforming loans, and bank failures.4 In addition to evaluating RWA as a predictor of bank risk, Avery and Berger (1991) examine the levels of capital required under the new and old regulatory capital standards. Banks whose levels of capital failed to meet the new standard were found to have significantly worse performance than banks whose levels of capital failed to meet the old standard, leading the authors to conclude that the new RBC standards provide a better indication of risk. However, there are shortcomings in their method of analysis. First, the RWAs calculated by Avery and Berger (1991) differ significantly from those reported by the banks themselves. We use instead the RWA values reported by banks in their FDIC Call Reports. Second, this approach compares specific policies rather than comparing the capital and RBC ratios as analytical tools. For example, comparing a 6% capital ratio with an 8% RBC ratio may only tell us that an 8% ratio is better than a 6% ratio, and not necessarily whether the capital ratio or the RBC ratio is a better indicator of performance. Third, the regression analysis of Avery and Berger (1991) examines the influence of RWA separately from capital requirements, whereas RBC regulations depend crucially on the interaction between capital and RWA. (For example, banks with higher RWA may still be considered safe if they are highly capitalized.)

We contribute to the debate on bank regulation by comparing the capital and RBC ratios reported by US commercial banks in their Call Reports as indicators of performance. Contrary to Avery and Berger (1991), we find evidence that the capital ratio is a better indicator of bank performance than the RBC ratio. The next section discusses our sample of bank-income and balance-sheet data, including some summary statistics. In section 3, we describe the analysis used by Avery and Berger (1991) and discuss some shortcomings of their approach. Section 4 proposes a more direct method of comparing capital and RBC ratios. Section 5 provides the results of our analysis, which indicate that the standard capital ratio is a significant predictor of the measures of bank performance used by Avery and Berger (1991). In contrast, the RBC ratio is not a significant predictor of performance, even when used in conjunction with the capital ratio. Section 6 concludes our study.

Continue Reading

Deposit Insurance Is Not Free

December 4, 2012

Economists are apt to point out that there is no such thing as a free lunch: Someone has to pick up the tab. Surprisingly, one common justification for government-provided deposit insurance maintains that it is a free lunch: No one has to pay for it. 

In this policy brief we review the theoretical argument for government-provided deposit insurance and compare it to the actual costs observed under the Federal Deposit Insurance Corporation (FDIC) in the United States. Contrary to theoretical arguments, deposit insurance is not free in practice. Moreover, we offer explanations for the increasing cost of government-provided deposit insurance observed over time[1]. We propose that economists and policymakers must consider the real costs when evaluating government-provided deposit insurance.

The Theoretical Argument for Government-Provided Deposit Insurance

In their influential article “Bank Runs, Deposit Insurance, and Liquidity,” Douglas W. Diamond and Philip H. Dybvig defend government-provided deposit insurance as a costless solution to the problem of bank runs[2]. The Diamond–Dybvig model provides a simple mathematical framework for analyzing the phenomenon of bank runs. Individuals deposit funds at a bank, which channels these funds to productive ventures. The bank agrees upfront to pay depositors a positive rate of return for each period their funds are left with the bank[3]. However, some fraction of depositors will find it necessary to withdraw their funds from the bank before the bank’s investments have generated returns. 

The bank pays a positive return on any early withdrawals, but its managers expect to maintain some capital investments to distribute to their patient depositors in the future. As long as the fraction of depositors showing up to withdraw their funds early is smaller than anticipated, the bank is able cover all commitments to depositors. If the fraction of early withdrawals exceeds some crucial level, however, the bank becomes insolvent since its commitments to depositors exceed the limited capital available to the bank before its productive ventures have been completed. If the bank is in danger of failing, even depositors who would prefer to leave their deposits in the bank until a future date will show up early to redeem their deposits, which causes a run on the bank. This is especially dangerous since only the expectation that the bank may fail can lead to a bank run. Thus, even safe, solvent banks are subject to bank runs.

Diamond and Dybvig proffer government-provided deposit insurance as a solution to the problem of bank runs. Unlike a private bank, a government can raise revenue at any time through the levy of taxes. If a bank run occurs, the government can tax those impatient depositors who withdrew their deposits early and use the money to ensure that patient depositors are paid. The fact that patient depositors will be paid even in the case of a run removes their incentive to run in the first place[4]. As a result, bank runs do not occur and the government need not pay any depositors. In the model, the mere commitment to pay depositors in the case of a bank run is sufficient to prevent bank runs. Hence, Diamond and Dybvig conclude that government-provided deposit insurance can be provided at no cost.

Thanks to Diamond and Dybvig, many economists have come to see government-provided deposit insurance as a free lunch. Their original article has garnered more than 1,000 citations in academic journals, is listed among the 25 most influential articles in economics, and forms the basis of discussion on deposit insurance at the undergraduate and graduate level[5]. There is only one problem with the narrative of costless government-provided deposit insurance: Government-provided deposit insurance in practice differs significantly from that proposed in theory.

The Cost of Government-Provided Deposit Insurance

In practice, government-provided deposit insurance is not a costless solution. The government must expend real resources to provide deposit insurance. Total expenses for the FDIC’s Deposit Insurance Fund have averaged $2.67 billion each year since its inception[6]. A total of $208.33 billion has been spent to date—with more than half (51.37 percent) of all expenses incurred in the last 10 years[7]. Figure 1 presents cumulative total expenses for the Deposit Insurance Fund, 1934–2011 (billion $2008).

Over 90 percent of the total cost of deposit insurance can be explained by two factors: (1) administrative and operating expenses and (2) net disbursements to depositors of failed or assisted banks[8]. Administrative and operating expenses total $32.15 billion (15.43 percent of total expenses). Net disbursements total $156.86 billion (75.29 percent of total expenses).

Administrative and operating expenses have been growing rapidly in recent years. From 1934 to 1974, administrative and operating expenses grew at an average annual rate of 1.24 percent, averaging just $0.12 billion per year over the period. Administrative and operating expenses grew at an average annual rate of 4.85 percent from 1974 to 2011. In 2011, annual administrative and operating expenses totaled $1.56 billion. Figure 2 shows annual administrative and operating expenses for the Deposit Insurance Fund, 1934–2011 (billion $2008).

Why are administrative and operating expenses on the rise? Part of the growth in administrative and operating expenses can be explained by increases in population and real per-capita wealth. Total domestic deposits held in FDIC-insured institutions increased from $643.66 billion in 1934 to $8,401.23 billion in 2011. Additionally, the maximum amount covered under the FDIC has increased over the period. Congress has raised the nominal maximum amount covered by FDIC insurance seven times since its inception. In 1934, the inflation-adjusted maximum amount covered under the FDIC was just $40,168; FDIC-insured deposits totaled $290.42 billion and roughly 45 percent of all domestic deposits held in insured banks were covered. In 2011, the maximum amount covered under the FDIC was $239,234; FDIC-insured deposits totaled $6,678.59 billion and around 79 percent of domestic deposits held in insured banks were covered. These figures suggest that, as the size of the fund increases, the costs of administering the fund similarly rise.

The biggest cost of government-provided deposit insurance is disbursements to the deposit holders of failed or assisted banks. When a bank fails, the FDIC pays its insured depositors out of the Deposit Insurance Fund and recovers any remaining assets. Net disbursements equal total disbursements less the value of assets recovered. Figure 3 shows net disbursements under the FDIC, 1934–2011 (billion $2008).

Net disbursements vary significantly from year to year[9]. Whereas annual net disbursements average $2.01 billion, the median is only $0.04 billion. Indeed, the bulk of disbursements (96 percent) are clustered in just two periods: the savings and loan crisis (1981–1994) and the financial crisis (2007–2011)[10]. During the savings and loan crisis, 1,464 banks closed at an average rate of 9.38 banks per month. Net disbursements averaged $4.72 billion per year over the period—ranging from a high of $12.59 billion in 1988 to a low of $0.26 billion in 1994. Each bank failure resulted in an average net disbursement of $0.04 billion. In the most recent financial crisis, 414 banks failed at an average rate of 8.63 per month. Net disbursements were significantly higher from 2007 to 2011. Roughly $22.31 billion was disbursed each year on net—more than $0.21 billion for each failed bank[11].

Conclusion

Contrary to the view put forward by Diamond and Dybvig, government-provided deposit insurance is not free. The reason is straightforward: Government-provided deposit insurance in practice differs significantly from that proposed in theory. First, the FDIC must expend real resources administering and operating the Deposit Insurance Fund. Second, the FDIC is committed to bailing out depositors whenever a bank goes bust regardless of whether the failure was the result of a bank run. Economists and policymakers would do well to keep these factors in mind when considering the costs and benefits of government-provided deposit insurance.

Endnotes

1.For a more extensive treatment of the topic, see Thomas L. Hogan and William J. Luther, “Implicit and Explicit Costs of Government-Provided Deposit Insurance” (working paper, Kenyon College, Gambier, OH, June 13, 2012), http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2083662.

2.Douglas W. Diamond and Philip H. Dybvig, “Bank Runs, Deposit Insurance, and Liquidity,” Journal of Political Economy 91 (1983): 401–19.

3.Technically, the bank as modeled by Diamond and Dybvig agrees to a prespecified rate to be paid to all depositors holding funds in the bank during the first period and divides the output from productive ventures (which takes exactly two periods to produce) among those still holding funds in the bank at the end of the second period, when the bank is dissolved.

4.As stated above, some fraction of depositors will experience circumstances precluding them from leaving their funds with the bank until all productive ventures are completed. Government-provided deposit insurance benefits these depositors as well. Since insurance prevents bank runs, they are sure to receive the full value of their deposits, which might not be true in the case of a run.

5.In a recent interview, Nobel laureate Thomas Sargent calls attention to the costless nature of government-provided deposit insurance in the Diamond-Dybvig model. See: Arthur Rolnick, “Interview with Thomas Sargent,” Federal Reserve Bank of Minneapolis The Region, (September 2010): 31.

6.The data presented throughout is available from Federal Deposit Insurance Corporation, FDIC 2011 Annual Report (Washington, DC: FDIC, 2011). All values are adjusted for inflation (100 = 2008).

7.We limit our analysis to the direct costs incurred by the FDIC. As a result, we omit indirect costs (for example, those stemming from moral hazard or the implicit subsidy on bank deposits relative to other investments).

8.The remaining costs are listed by the FDIC as “interest and other expenses.” Some expenses were also transferred from the Federal Savings and Loan Insurance Corporation Resolution Fund in the years 1989–1992. 

9.Some research suggests that while deposit insurance is capable of preventing bank runs, it may increase systemic risk in the banking sector, leading to less frequent but more severe financial crises See, for example, Asli Demirgüc-Kunt and Enrica Detragiache, “Does Deposit Insurance Increase Banking System Stability? An Empirical Investigation,” Journal of Monetary Economics 49 (2002): 1373–1406. If this is the case, FDIC does not merely incur the cost of disbursements to the deposit holders of failed banks but is also the cause of some of those failures.

10.Roughly 39 percent of all net disbursements occurred during the savings and loan crisis, compared with 57 percent in the financial crisis.

11.Although net disbursements were larger in the recent financial crisis than in the savings and loan crisis, the percentage of disbursements recovered was also higher.