Why Aren't Bank Lending?: Credit Growth & Regulatory Micromanagement
Credit creation continues to lag the levels necessary to support economic growth. In the first quarter of 2011, the total amount of credit market borrowing in the United States was $912 billion (annualized), or about 6% of GDP. Of that total, the Federal Reserve supplied $995.3 billion (line 34, monetary authority), on net. This means that the Fed, through QE2, supplied 109% of all lending in the U.S. economy in the first quarter. Another way of looking at it is that the total credit market would have contracted by about $80 billion in the quarter without the Federal Reserve asset purchases. This would have been just the fifth quarter of credit contraction on record (the first four came in 2009).
Since 1980, credit growth (as a percent of GDP) has been positively correlated with nominal GDP growth. A 1% increase in credit has been associated with a 0.45% increase in the size of the economy. During the rapid economic growth of 1983-85, lending expanded at an average rate of more than 20% of GDP (nominal GDP grew by over 11% during this period). Credit growth also averaged more than 20% of GDP from the fourth quarter of 1997 to the first quarter of 2000, when the economy grew rapidly as corporations borrowed heavily to invest in new plant, equipment, and communications technologies. Less fondly remembered will be the record 30% growth in credit market borrowing averaged between 2005 and 2007, when the borrowing came largely from the household sector, which capitalized on liberal credit policies to convert home equity into consumption expenditures.
The lending excesses of 2005-07 directly impact current economic conditions in two ways. First, because the excesses of the period helped to contribute to the global financial crisis, today’s unemployment rate is partly a reflection of the dislocation caused by the credit losses attributable to bad lending. Secondly, the excesses of the period have made credit less available for high-quality borrowers today by increasing the regulatory scrutiny applied to prospective loans. Some of the decline in lending is surely attributable to “debt overhang,” where embedded losses on the balance sheet cause a bank to choose to conserve equity and fund fewer loans than it would otherwise. But it seems clear that part of the problem stems from regulators seeking to overcompensate for past failings by creating what local bankers describe as an “oppressive examination environment." Banks contend that regulators are pursuing five specific examples of over-regulation that are impeding economic growth:
- Requiring write-downs of performing loans based on the value of collateral;
- Classifying performing loans as in non-accrual status;
- Disregarding third party appraisals of collateral used to secure loans;
- Criticizing certain funding sources; and
- Increasing regulatory requirements beyond those specified in law.
Taken together, this oppressive environment has caused lenders to refuse to fund loans in certain circumstances irrespective of the loan’s economic prospects. The judgment of bankers is being supplanted by the judgment of regulators. Given the quality of the lending decisions made in 2005-07, some may argue that this is precisely what’s necessary now. The problem is that the easiest way to avoid making a bad loan is to refuse to approve any loans.
In frequentist (as opposed to Bayesian) statistics, hypothesis tests can be subject to Type I and Type II errors. In the banking context, the null hypothesis is that a prospective loan is bad. A Type I error would be failing to reject a bad loan; a Type II error would be wrongly rejecting a good loan by mistakenly classifying it as bad. Although analysts often focus on the “credit cycle” where credit conditions seem to oscillate predictably from lax in good times to stringent in bad, George Soros identifies the “regulatory cycle” (p. 86) as being every bit as impactful on lending conditions. This means that regulators try to make up for yesterday’s Type I errors by committing as many Type II errors in the subsequent period as possible.
Whether banks’ credit conditions are stringent today or not, it’s clear that the five specific examples of regulatory overreach all serve to erode bank lending capacity and make new loans less economic. Forcing banks to write down the value of loans based on collateral values causes bank capital to fall for no good reason. Imagine a bank loaned $7 million to a developer to build a $10 million office building [a 70 loan-to-value (LTV) ratio]. If the value of the office building falls by 30% to $7 million, the economic value of the loan has, in fact, been eroded because the bank would have certainly made a 100 LTV loan on different terms than a 70 LTV loan. But requiring this change to be reflected in bank capital even when no default event has occurred – or is likely to occur – is precisely the logic of mark-to-market accounting, which requires banks’ trading portfolios to be revalued daily.
These rules contribute to the pro-cyclicality of lending because mark-to-market gains in asset prices create more bank capital and lead to more lending, while declines in asset prices destroy capital and make credit more scarce. While some pro-cyclicality is the inevitable consequence of transparency, regulatory rules generally seek to mitigate the fair value accounting impact on bank capital. As explained previously by e21, between 13% and 55% of the major banks’ assets were fair valued, with the rest held at cost. The new supervisory guidelines seek to expand the mandate to illiquid loans for which no secondary market exists. Ironically, current examination policy towards lending is actually worse than mark-to-market accounting because of its asymmetry: the value of the loan can only deteriorate and bank capital (and lending capacity) can only fall because the bank receives no credit if the value of the collateral were to rise.
The other examples of regulatory micromanagement similarly erode bank capital and exacerbate the “credit cycle.” In testimony before the House Financial Services Committee, entrepreneur and developer Mike Whalen explained his difficulty obtaining a loan for a construction project with the “strongest financial structure” in his company’s 33-year history. The fact that his loan would be classified as “commercial real estate” and “hospitality” scared lenders because of the heightened scrutiny applied to these loans. As the Fed explains in regulatory guidance, an individual loan can be classified as nonaccrual because of its relationship to “similar types of loans” that have gone bad. For example, the Fed explains that a regulator may require the bank to write down the value of a group of loans even when “the particular loans [in that group] that are uncollectible may not be identifiable.” So if the regulator requires all “hospitality” loans to be discounted 10%, then extending a $10 million loan to Mike Whalen would result in an immediate loss of $1 million to the bank irrespective of how strong the specific loan may actually be.
The economy is too weak to withstand this level of regulatory micromanagement. Regulators do not have enough information to make decisions about each loan in the portfolio so they necessarily generalize in ways that disadvantage high quality borrowers in troubled sectors. This limits growth. Worse, the discourse in Washington tends to focus on whether slow loan growth is attributable to “supply” or “demand,” as though all borrowers’ credit histories, business plans, and capital structures can be easily aggregated into a demand curve that could be interpreted on the national level. In a decentralized, capitalist economy, lending decisions must necessarily be made at the local level by bankers willing to risk their own capital to fund projects in the pursuit of profit. Regulations that inhibit this localized process retard economic recovery and make the economy more dependent on the Fed to supply credit and the Federal government stimulus to provide incremental demand.