Legislative Proposals Regarding Bank Examination Practices
by Simon Johnson, Peterson Institute for International Economics
Testimony submitted to the House Committee on Financial Services, Subcommittee on Financial Institutions and Consumer Credit, hearing on "Legislative Proposals Regarding Bank Examination Practices"
July 8, 2011
H.R. 1723 "The Common Sense Economic Recovery Act of 2011"
H.R. 1723 "The Common Sense Economic Recovery Act of 2011" proposes to create a legislative constraint on the ability of bank examiners to determine whether a loan is "accrual" or expected to be repaid in full. The intent is to effectively lower capital requirements for banks that have impaired assets, i.e., to allow them to have less equity funding relative to total assets than would otherwise be the case. This would not be a good idea.
In a free market system—without government guarantees—financial institutions fund themselves with a relatively large amount of equity (30 percent is not uncommon), because they need a strong buffer against losses. Investors are not willing to fund a bank that is prone to collapse.
But in a system with deposit insurance, the downside risk for one class of investors—retail depositors—is limited or zero. This encourages banks to fund themselves with more debt relative to equity, which means little capital relative to total assets. This increases the upside payoff to equity—i.e., for a given return on assets, the return on equity is higher when things go well. But it also increases the downside returns on equity, creating more volatility and a higher probability that the bank will become insolvent.
The FDIC and other regulators are charged with ensuring that banks maintain enough capital—i.e., sufficiently high buffers against losses—so they remain solvent. In the case of insolvency, there are typically costs to the FDIC's Deposit Insurance Fund. There can also be significant taxpayer costs, both direct and indirect, from bank failures.
To see the fiscal impact of the finance-induced recession, look at changes in the CBO's baseline projections over time. In January 2008, the CBO projected that total government debt in private hands—the best measure of what the government owes—would fall to $5.1 trillion by 2018 (23 percent of GDP). As of January 2010, the CBO projected that over the next eight years debt will rise to $13.7 trillion (over 65 percent of GDP)—a difference of $8.6 trillion.
Most of this fiscal impact is not due to the Troubled Assets Relief Program—and definitely not due to the part of that program that injected capital into failing banks. Of the change in CBO baseline, 57 percent is due to decreased tax revenues resulting from the financial crisis and recession; 17 percent is due to increases in discretionary spending, some of it the stimulus package necessitated by the financial crisis (and because the "automatic stabilizers" in the United States are relatively weak); and another 14 percent is due to increased interest payments on the debt—because we now have more debt.2
In effect, a financial system with dangerously low capital levels—hence prone to major collapses—creates a nontransparent contingent liability for the federal budget in the United States.
The last economic cycle—culminating in the financial crisis of 2008—demonstrated that allowing banks and other financial institutions to reduce their capital levels is dangerous not just for those institutions but also for the system as a whole.3 Individual banks never take into account the spillover effects caused by low equity capital funding levels—and these negative externalities can be very large.
Repeated boom-bust cycles have also shown that it is very dangerous to allow banks to reclassify their loans as performing when management knows that the loan is not likely to be paid in full.
The rules in this area are clear and well-established.4 In some ways, the bill duplicates what the FDIC already does. For example, under current practices, after six months of performance on a modified loan, the loan can be restored to accrual status.
But classifying loans as "accrual" when they are either not receiving the full interest due or when the institution knows there will be a problem is not a good idea. This will make the bank's capital position look better than it really is and detract from the credibility of their financial statements.
It will also make it harder for the FDIC to implement prompt and appropriate corrective action. Most likely, this measure will increase costs to the Deposit Insurance Fund in the long run. This increases the potential future liability of all banks that have to contribute to this fund.
In addition, modified loans are treated the same way as other loans in the bill—so that if a borrower comes to the bank knowing that he or she will not be able to make the next payment, those loans could still be treated as accrual loans because—under the bill—they are still technically current even though the bank knows that future payments are uncertain. It would be contrary to reasonable safety and soundness criteria for a loan to be treated as accrual if the bank does not expect full repayment of the remaining contractual principal and interest.
The following provisions could be usefully added to the bill, with regard to when a loan can be treated as accrual:
This bill would also affect the "shared national credit" ("SNC") program, which requires examiners assign the same classification to all loans to the same large borrower. If a borrower has defaulted on a loan from one bank, then a current loan from another bank should be classified no better than the loan from the first bank. If one loan is nonaccrual then the other loan must be nonaccrual. This is simply sensible accounting.
H.R. 2056 "A bill to instruct the FDIC Inspector General to study the impact of insured depository institution failures"
H.R. 2056 calls for a study of the impact of bank failures. This would be helpful, for example in shedding light on how supposedly "uniform" supervisory enforcement actions have actually been applied across the big four Office of the Comptroller of the Currency (OCC) banks relative to thousands of community banks.
In all likelihood the problem is not that the FDIC is closing too many banks; it is that the FDIC is closing too few banks. Arguably, there are dozens of "zombie banks" that the FDIC has failed to close—even though they are in worse shape than many that it has closed. Conditions of these zombie banks continue to deteriorate, increasing the ultimate cost to the Deposit Insurance Fund.
As far as can be determined by outside observers, the FDIC appears to prefer closing banks for which it has bidders because the Deposit Insurance Fund has limited resources and the FDIC wishes to avoid borrowing from the Treasury.
It is time to draw on experience from the early 1990s and discuss creating another Resolution Trust Corporation to take over the assets of troubled banks, with the charge of rapidly returning them to the private sector.
Of course, such an entity would also be useful should the "extend and pretend" game for one of the big four banks also come to an end.
Simon Johnson is a senior fellow at the Peterson Institute for International Economics and a professor at MIT.
1. The definitive arguments on this issue are presented in detail by Anat Admati, Peter DeMarzo, Martin Hellwig, and Paul Pfleiderer, "Fallacies, Irrelevant Facts, and Myths in the Discussion of Capital Regulation: Why Bank Equity is Not Expensive."
2. See also the May 2010 edition of the IMF's cross-country fiscal monitor [pdf] for comparable data from other industrialized countries. The box on debt dynamics shows that mostly these are due to the recession; fiscal stimulus only accounts for one-tenth of the increase in debt in advanced G-20 countries. Table 4 in that report compares support by the government for the financial sector across leading countries; the United States provided more capital injection (as a percent of GDP) but lower guarantees relative to Europe.
3. Capital levels were reduced by various kinds of financial innovations—such as moving assets supposedly "off balance sheet"—as well as by lobbying regulators and legislators. Arguably the single most disastrous decision was to allow investment banks to operate with significant more debt relative to their equity. See Simon Johnson and James Kwak, 13 Bankers, Chapter 5.
4. See, for example, "Uniform Retail Credit Classification and Account Management Policy," Federal Register 64.27 (February 10, 1999), issued by the Federal Financial Institutions Examination Council.