Google books provides the extracts below from The Crisis in American Banking.
(emphasis mine) [my comment]
The Crisis in American Banking
By Lawrence White
Lawrence H. White
This volume offers six original essays keyed to the continuing crisis in the U.S. banking industry. Five were first presented at a small conference—more like a series of seminars—on "The Crisis in the Banking Industry," sponsored by NYU's Austrian Economics Program (which is directed by Israel Kirzner). The conference was organized by Mario Rizzo and myself, and was held at New York University on April 29, 1991. The papers generated lively discussion among the authors and other participants in the conference, and have been revised to reflect (or deflect) constructive criticism received there. A sixth paper, by Richard Salsman, was solicited soon after the conference. All six have been updated to reflect developments through September 1991.
The U.S. banking system, its regulation and deregulation, and its troubles, have been much in the news lately. Banking topics have been discussed by economists in a number of monographs and conference volumes. The rationale for adding the present volume to the discussion is the hope that its contributors provide fresh perspectives by viewing the banking scene from unusual angles. In particular, several authors draw ideas from modern Austrian economics or from public choice theory that have seldom been applied to explaining contemporary banking problems.
A pervasive theme of the contributions here is that the U.S. banking crisis is fundamentally linked to the political regulation of banking. (Popular alternative explanations point to supposedly excessive competition or deregulation in banking, or to a supposed decline in the ethical standards of bankers in the 1980s.) Taken together, the chapters below (1) indicate that government regulatory, macroeconomic, and fiscal policies have seriously impaired the health of the banking industry; (2) contribute to explaining how rent-seeking, ideology, and the historical accretion of regulations have given banking policy its current unfortunate form; and (3) consider the long-term prospects for reform of banking regulation, and for the banking industry itself in light of the current and foreseeable regulatory environment.
In the first chapter 1 attempt to provide an overview of the U.S. banking Industry's troubles and the FDIC's insolvency, and to trace them to regulatory and macroeconomic policies. The secularly shrinking profits brought to banking by techological change, as discussed in this volume by George Kaufman, explain why the banking industry should contract, but not why its contraction should be punctuated by a crisis involving high levels of loan losses and bank failures. I suggest that cyclical losses in bank lending, especially severe today in the area of real estate lending, represent correlated errors induced by unpredictable monetary policy. Regulation (particulary restrictions on geographic and product-line diversification) and deposit insurance explain why U.S. banks, as the thrifts did earlier, have reacted to a downturn with increased risk taking. Increased risk taking has led to dramatic exits in the fashion of the Bank of New England, rather than simple shrinkage or redeployment of capital. Policy reform should seek to remove the distortions that promote excessive risk taking.
Roger Garrison addresses the impact of government budget deficits on the economy in general and on banking in particular. He argues that unusually large borrowing, to cover today's deficit, forces market participants to guess about tomorrow's policy for servicing or repaying the debt, and to guess about how other market participants will view the situation and respond. To what extent will a continued debt burden absorb domestic saving, causing high real interest rates and crowding out domestic investment? To what extent will it absorb funds from abroad, keeping real interest rates at their normal level but weakening demand for export goods? To what extent will taxes be raised, when, and on what? To what extent will the debt be monetized, causing inflation? Deriving surprising illumination from standard national-income accounting, Garrison shows that some combination of these (jointly exhaustive) repercussions must follow a deficit.
Empirical studies may show no strong or regular connection between deficits and any one of these repercussions, but Garrison points out that it would be fallacious to conclude that deficits have no repercussions or are harmless. The lack of a predictable mix of repercussions in fact means that market participants face added uncertainty. This excess uncertainty hinders decision making in the banking sector—portfolios will prove to have been misallocated when guesses about answers to the above questions prove to have been mistaken, or rates of return will be reduced by greater hedging—and in the rest of the economy.
Thomas Havrilesky examines the role of private interests (the S&L; and big-bank lobbies) in shaping banking legislation between 1985 and 1987, a period in which thrift industry problems incubated, hi provocative language he argues that banking- and thrift-industry policy, like politics generally, is about "rent-seeking" or redistributing income. Critical changes in the regulatory rules governing S&Ls; in the 1980s (an increase in the deposit insurance ceiling, regulatory forbearance to close insolvent thrifts, relaxed accounting standards) appear to reflect the capture of legislators and regulators by the S&L; lobby. Studying the statistical relationship between a Congressperson's contributions from S&L; or big-bank lobbies (Political Action Committees, or PACs) and his or her voting or bill sponsorship, Havrilesky finds evidence that each lobby was influential. Congress- people who sponsored pro-big-bank bills received a bigger share of their PAC money from big-bank PACs; those who voted as the S&L; lobby preferred likewise received greater proportionate S&L; PAC contributions.
Popular accounts of the thrift-industry fiasco, and of the FDIC's current difficulties, have blamed fraud and mismanagement among bankers, citing anecdotal evidence of renegades like Charles Keating. Richard Saisman's essay provides a useful historical perspective by showing that the same sort of charges were made in previous U.S. banking crises: the "wildcat banking" episodes of the antebellum period; the money panics of the National Banking era, especially the Panic of 1907; the banking collapse of the Great Depression; and the S&L; crisis of the 1980s. In each case bankers were blamed for ills that Salsman—drawing on important "revisionist" work by monetary theorists and historians in the last twenty years—indicates should in fact be traced to government interference in banking. Legal restrictions, in the later episodes combined with central banking and deposit insurance, have created climates in which unsafe banking practices can persist and become institutionalized. He concludes that systemic bad b anking is a symptom of bad policy, rather than an independent cause of crisis.
Why then have bankers been made scapegoats? Saisman cites a number of cases in which federal legislators, agency heads, and commissioners have led the movement to blame bankers, rather than government policies, for the banking system's failings. Selfinterest provides an obvious motive. A question that remains to be answered is why the popular press have not been more discerning.
Walker Todd and Gerald P. O'Driscoll, Jr., provide further evidence of the destabilizing effects of government deposit guarantees. They emphasize that explicit deposit insurance is only one part of the "safety net" whose historical growth they document; implicit guarantees are an additional and crucial part. They warn that government supervisory agencies will not rein in excessive risk taking by banks (being inherently prone to err on the side of wishful thinking) until it is too late. Political pressures will then be brought to bear on the supervisors (the FDIC and the Fed) to bail out failing institutions, as long as deposit guarantees of any amount are in effect. They find evidence of these pressures not only in the much- discussed doctrine that some banks are "too big to fail" but also in the abuse of the Fed's discount window for bank rescues, a feature of the current system that has scarcely been mentioned elsewhere. The central bank discount window, under classical lender of last resort doctrine, is supposed to provide only liquidity support to the banking system, not capital support to individual insolvent institutions. Walker and O'Driscoll note that Fed loans to banks declared insolvent are repaid out of the FDIC's Bank Insurance Fund, and the BIF is ultimately replenished with taxpayer money, so the current system "converts unsound banking policy use of the discount window to keep insolvent banks afloat into unsound fiscal policy." They conclude that to achieve stability the entire safety net needs to be reformed, not only deposit insurance.
Todd and O'Driscoll argue for a comprehensive set of reforms that would eliminate federal deposit guarantees. At the very least, they favor a bank closure policy that exposes depositors and shareholders, but not taxpayers, to losses. Observing that the deposit guarantee system will soon be transferring wealth from the average citizen (through taxation to recapitalize the FIJIC) to the wealthy citizen (who has parked savings in insured deposits), they reasonably suggest that having an FDIC makes the average citizen worse off. Risk-free savings vehicles paying competitive rates are already available in the form of Treasury bills and savings bonds.
Looking beyond the current crisis, George Kaufman surveys the secular trends that are shrinking the banking industry in comparison to other financial service providers. He finds that these trends stem partly from technological changes (such as advances in telecommunications and computerization), and partly from bad policy decisions (such as mispriced deposit insurance, forbearance to close insolvent institutions, and geographic and product-line restrictions). Shrinkage of the banking industry due to loss of comparative advantage is efficient and nothing to mourn. But shrinkage due to ill-conceived public policy is not efficient. Kaufman calls for correctly pricing deposit insurance, and for removing restrictions on the geographic and product-line powers of banks. He warns against softening the balance-sheet standards for banks, which some voices have urged as a way to combat a supposed "credit crunch." With broader powers but without subsidies, efficient U.S. banks should be able to compete on a level playing field. The success of foreign banks and nondepository financial firms in recent years shows that lesser deposit guarantees (and correspondingly higher capital ratios) do not preclude growth.
As this preface is written, the state of the U.S. commercial banking industry and the FDIC continues to suggest disturbing parallels to the state of the savings and loan industry and the FSLIC a decade earlier. With the BIF's balances down to $2 billion in late 1991 (less than the amount needed to close the Bank of New England earlier in the year), the FDIC appeared to be putting off closing banks that were insolvent (on a market-value accounting basis, i.e., when counting their assets at market value rather than book value). After predicting earlier in the year that 180 to 230 institutions would be seized in 1991, the FDIC reduced its estimate in December to only 137. In pleading for "recapitalization" of the BIF, FDIC chairman William Taylor acknowledged in so many words that lack of funds was hindering the agency from closing unsound banks. Such a policy of forbearance carries the danger of duplicating in banking the second phase of the thrift crisis, that is, of creating a new cohort of "zombie" institutions rationally pursuing risky strategies at the expense of future taxpayers who pick up the tab for losses by government deposit insurance agencies.
In November 1991 the FDIC's Bank Insurance Fund was "recapitalized" by omnibus banking legislation granting the agency an additional $70 billion in borrowing authority. The FDIC Improvement Act requires the FDIC to repay any borrowings from its asset sales or insurance premiums. It remains to be seen whether the agency will be able to repay, or whether taxpayers will be presented with the tab at a later date. The legislation includes a number of deposit insurance reforms: changes in accounting and examination rules; a schedule of restrictions to be placed on undercapitalized instutitions; a mandate for the FDIC to impose risk-based Insurance premiums by 1994; and a requirement that the agency resolve failures by the method generating the least cost to the FDIC, even if that means exposing uninsured depositors to losses, beginning 1995. Proposals for structural reform of banking, to eliminate geographic and product-line restrictions, were excluded from the legislation.
With its new access to funds, the FDIC is expected to begin working off a backlog of insolvent but not-yet-seized institutions. The agency officially expects to close 200 to 239 banks, with total book assets in the neighborhood of $100 billion, in 1992 alone. Private analysts estimate $50 billion in losses to the FDIC from the closure of some 150 sick savings banks in New York and New England, in addition to losses from closure of ailing commercial banks. As the situation unfolds, observers who currently fear a replay of the $150 billion FSLIC bailout may find that they were overly pessimistic—or overly optimistic. Either way, the policy regime that allowed both the earlier and later problems to develop does not seem to be on the verge of any dramatic change. The reluctance of Congress to enact real reforms means that the critical analyses and reform proposals in this volume, against the wishes of their authors, will remain relevant for some time to come.
Why Is the U.S. Banking Industry in Trouble? Business Cycles, Loan Losses, and Deposit Insurance
Lawrence H. White
We learned from the U.S. thrift-industry debacle that congress peopie and regulators have incentives to mask and deny the size of insolvencies among deposit-taking institutions when they first arise. Rather than promptly resolve the widespread insolvencies that existed among thrifts in 1981, the authorities chose to revise the regulatory accounting rules, to practice "forbearance," and to gamble that economically insolvent thrifts might climb back into the black (Eisenbeis 1990, 19—20). As it turned out, the cost of resolving the problem grew, to the point where taxpayers have been saddled with an enormous expense in covering the thrift deposit guarantees made by the late FSLIC. Estimates of the expense, beginning at $10 billion in 1986, have been revised upward steadily to more than $150 billion (as of 1991, excluding Interest).
The Industry's and the FDIC's Troubles Are Large
In light of this experience, a sense of déjŕ vu accompanied news reports, beginning in late 1990, that the Federal Deposit Insurance Corporation, the agency that now guarantees both thrift and bank deposits, would soon run out of money without taxpayer assistance. The FDIC's Bank insurance Fund shrank as its annual disbursements in resolving bank failures, whose numbers swelled tremendously the last decade (see figure 1.1), exceeded its income from deposit insurance premiums. Beginning 1988 with $18.3 billion, the BIF lost $5.1 billion in 1988—89, and another $6.8 billion in 1990 alone, leaving its balance at only $6.4 billion at year-end 1990. The most recent FDIC projections imply net losses for 199 1—92 of $19 to $38 billion.'
As they did with the FSLIC, the authorities have persistently underestimated the FDIC's problems. Early in 1990 FDIC officials projected that the agency would break even for the year. By midyear they projected 1990 losses of $2 billion. In September 1990 FDIC chairman William Seidman raised the estimate to $3 billion. In retrospect the FDIC first found that it had actually lost $4.8 billion In 1990 (the BIF's balances had declined to $8.4 billion at year-end 1990 from $13.2 billion at year-end 1989); later it revised the year- end 1990 balance to $6.4 billion, implying 1990 losses of $6.8 billion. 2
The federal Office of Management and Budget projected in September 1990 that the FDIC would need infusions of $22.5 billion over the next five years to remain afloat. In October 1990, Seidman professed not to see the FDIC fund in danger for the foreseeable future, maintaining that the fund would remain in the black through the end of 1991. As 1991 began Seidman was asking Congress for a $10 billion loan to provide a margin of safety, though he had previously indicated that the fund could use a $25 billion infusion. The Congressional Budget Office projected in January 1991 that the BIF would develop a $2.8 billion deficit by mid-1992, but that (with increases in premiums) it would return to solvency by mid-1994. In March 1991 the Bush administration proposed that Congress provide the FDIC with $25 billion in borrowing authority. Within a week Seidman testified that under a "pessimistic economic scenario" the fund might have to borrow $30 to $35 billion, and the administration's request for FDIC borrowing authority was increased to $70 billion, just in case. As of June 1991, the FDTC's "baseline projection" gave the BIF a year-end 1991 balance of $3.2 billion, while its "pessimistic scenario" projected a year-end balance of $1.7 billion. By the fall of 1991 it was evident that even the "pessimistic scenario" was overly optimistic. Seidman was projecting that the fund would be insolvent at year's end, and reported that it had already borrowed $2.9 billion from the U.S. Treasury.
As in the thrift meltdown, the projections of private-sector experts have been more pessimistic ex ante and have proven to be more accurate ex post. Bank consultant David Cates projected in October 1990, before the recession had been officially declared underway, that in a recession banks could lose $86.3 billion in equity, 41% of the industry's cushion, and taxpayers could face a bill as high as $40 billion to cover FDIC losses. Cates's credibility was enhanced by his projecting that, under such a scenario, the Bank of New England would fail. The Bank of New England failed in January 1991, amidst depositor runs, with losses expected to make it (at $2.5 billion) the third most expensive resolution in the FDIC's history. Lowell Bryan, a bankwatcher at McKinsey and Company, similarly projected in December 1990 that the FDIC would need injections of $20—40 billion over the next few years. A study by the independent bank rating service Veribanc found that the FOIC was already insolvent at year-end 1990, facing resolution costs for then- insolvent banks (estimated at $11.6 billion) in excess of BIF balances. Edward Kane estimated in mid-1991 that the BIF was already $40 billion In the hole as of the end of 1990, if one recognized as a BIF liability the negative net worth that would appear under mark- to-market accounting for insured banks' assets.5
An authoritative study of the FDIC's condition appeared in a report by James R. Barth, R. Dan Brumbaugh, and Robert E. Litan, dated December 1990, commissioned by the Financial Institutions Subcommittee of the House Banking Committee. Barth, Brumbaugh, and Litan (1990, 6) concluded that the BIF at the end of 1990 appeared to be where the FSLIC was in the mid-1980s, "without sufficient resources to pay for its expected caseload of failed depositories." The authors rehearsed a number of scenarios, varying in the assumed severity of the incipient recession and the degree of forbearance to be exercised toward larger banks.
They estimated (Barth, Brumbaugh, and Litan, 93, 103) that even under a "mild recession" the FDIC's bank resolution costs for 1991— 93 would run between $31 and $43 billion, exhausting its expected resources of $28—31 billion (consisting of start-of-period reserves of $9 billion plus premium and interest income of $19—22 billion).
A major source of concern is that larger banks have begun to appear on the FDIC's list of "problem banks." The list numbered 975 banks at the midpoint of 1991, slightly fewer than in the immediately preceding years, but the aggregate assets of problem banks had increased. Likewise, although the number of bank failures In the first half of 1991 (fifty-seven) was fewer than in the first half of the previous year (ninety-nine), the average asset size of failed banks was much much larger: $475 million versus only $65 million (FDIC 1991,2,4).
In fact some of the very largest U.S. banks are teetering. The Economist commented in December 1990: "Nobody knows just how much rubbish EU.S.1 banks have on their books, or how many loans might become rubbish if a recession deepens. Among the banks that fail may be prominent money-centres."6 Barth, Brumbaugh, and Litan (1990, 13) commented that as of the end of 1990 "most" of the nation's largest banks were "on—or conceivably over—the edge of insolvency.... (Many of these banks not only currently have weak balance sheets by any reasonable standard, but they also are highly exposed to additional deterioration in their capital positions from their significant involvement in high-risk lending." They reported (Barth, Brumbaugh, and Litan 1990, 50) that six of the top twenty- five bank holding companies had "high risk loans" (loans for highly leveraged transactions [HLTI, medium and long-term LDC l oans, and commercial real estate loans) in excess of four times their "adjusted tangible common equity" (tangible common equity plus allowance for loan losses less 1% of all performing loans). That is, a 25% fall in the value of such a bank's high-risk loan portfolio, together with a normal 1% loss rate on other loans, would wipe out its capital, leaving it insolvent.
FDIC call reports show that the large banks (those with more than $10 billion in assets) as a size class have the weakest loan portfolios. Across most categories of loans, the large banks have the highest percentages of loans past due or noncurrent, and the highest percentage of loan charge-offs (FDIC 1991, 3). In recent years the class of large banks has had the highest percentage losing money. In the first half of 1991, 11.2% of all banks (1,361 of 12,150) lost money, while 19.6% of large banks (9 of 46) did so (FDIC 1991, 5).
Marketplace reflections of troubles at the large money-center banks are not hard to find. In 1980 Moody's Investors Services rated the debt of fourteen major banks "AAA"; today it gives only one U.S. bank that rating (Salsman 1990, 71). Moody's now rates a fifth of Chase Manhattan's debt even below investment grade (Byron 1990, 16). Stock traders' expectations of likely future difficulties in banking are reflected in low share prices (relative to current reported earnings) for all banks, but especially for large banks (Barth, Brumbaugh, and Litan 1990, 15). While healthier banks trade at slightly above book value, money center banks have been trading well beLow (Salsman 1990, 72).
In addition to the asset-quality problems at large banks, Barth, Brumbaugh, and Litan (1990, 51) pointed out that the Bank Insurance Fund is also threatened by renewed troubles at savings banks. They noted that BIF-insured savings banks in the aggregate lost $670 million in 1989. The latest available figures show the situation worsening. The number of "problem" savings banks, which stood at thirty-one in mid-1990, had risen to fifty-eight a year later. Savings banks in the aggregate dropped a staggering $2.5 billion in 1990, and lost another $662 million in the first half of 1991, exceeding losses In the first half of 1990. In New England, where most savings banks are located, two-thirds of the twenty-three largest institutions were unprofitable in the first half of 1991 (FDIC 1991, 7).
With the FDIC running out of cash, there is a great danger that the agency is neglecting to close insolvent banks, just as the FSLIC neglected insolvent thrifts for years. "Zombie" institutions (economically "dead" but still operating) may be afoot, piling up obligations that will eventually be laid at the doorstep of taxpayers. Barth, Brumbaugh, and Litan (1990, 81) note that during the years 1980—85, with fewer annual failures, the typical failed bank was resolved about fifteen months after it first appeared on the FDICs problem list. By 1987—89, amidst two hundred failures per year, the typical failed bank was not resolved for 21—28 months after first being listed.
The Immediate Source of Trouble Is Bad Loans
U.S. banks are failing or troubled primarily because so many of the risky loans they made in the 1980s are in default. In the four quarters ending September 30, 1990, the banks' net charge-offs for bad loans were $30.5 billion, the largest doflar amount for any four- quarter period ever, and a record high percentage of assets. Chargeoffs in the first half of 1991 continued at roughly the same high rate. Despite the charge-offs, the proportion of troubled loans on bank portfolios continued to rise. At the end of September 1990, the total of noncurrent loans and leases plus "other real estate owned" (foreclosed mortgage property) was $89.6 billion, up by $14.2 billion (19%) from a year earlier ($75.4 billion), and at a record level as a share (2.65%) of total assets (FDIC 1990, 1—3). By the midpoint of 1991, the total ($107.9) and the share (3.19%) had both risen even higher.
Despite the taking of historically large loan-loss provisions in 1989 and 1990, another round of large provisions was needed in 1991, and still another round was expected to be needed in 1992. Loan-loss reserves were down to 65 cents per dollar of noncurrent loans as of 1991:2, down from 73 cents as of 1990:3, and down from 83 cents a year before that. Losses from commercial real estate loans, LDC loans, and HLT loans were expected to increase in the recession, especially among the larger northeastern banks (FDIC 1990. 2).
The nature of the banking industry's bad loans has been widely reported. Southwestern banks, making up the majority of failed banks in the last few years, suffered big real-estate and energy loan losses in the late 1980s. The big money-center banks took sizable write-downs of LDC loans in 1987, such that all large banks posted negative returns for the year, and another milder round of write- downs in 1989. In both years, loss provisions on overseas loans more than accounted for the total negative income recorded; domestic business was profitable (Duca and McLaughlin 1990, 483). In 1989 through 1991, New England banks took large losses on real estate loans. Depressed real estate markets are pushing the value of much of the collateral held by banks below the value of the loans carried on many banks' books (Barth, Brumbaugh, and Litan 1990, 47), so that more loan defaults are expected. Bank inventories of foreclosed real estate were still accumulating as of mid-1991 (FDIC 1991, 2).
The Causes of Loan Losses Are Cyclical, Secular, and Regulatory
How can we explain the profile shown by figure 1.1, an extraordinary growth in the annual number of bank failures after 1981? Commercial bank profitability has trended downward since 1970 (FRBNY 1986), a trend that has continued in recent years and is consistent with banking firms exiting from the market. But the gradual secular trend in bank profitability cannot plausibly explain the dramatic shift in the trend of bank failures, or why it occurred when it did. The onset of a sharp recession in the second half of 1981 undoubtedly helped swell the number of failures in 1982 and 1983. But we clearly cannot explain the pattern solely by reference to the business cycle. The number of failures was much less In previous recessions, and failures continued to climb even after the 1982—83 recession gave way to a period of sustained expansion.
A recession swells the number of bank failures for obvious asset quality reasons. With unemployment up, more household loans go bad. As recession took hold in 1990, delinquency rates in home-improvement loans and revolving credit reached their highest levels in ten years (Farrell 1991, 29). More importantly, with corporate bankruptcies, business loans go bad. In the 1982—83 recession, energy and agricultural business loans especially went into default. in the most recent downswing, commercial property loans have been the most conspicuous source of losses and were the principal reason that the Bank of New England failed.
It would be myopic, however, to treat the recessionary phase of the cycle as the ultimate source or the exogenous cause of asset- quality problems. The loans that go bad were typically made years earlier during the expansion. Hundreds of banks had asset-quality problems well before the 1990—91 recession officially began, and bad assets brought on the 1988—89 wave of Texas bank failures in advance of the national recession. Viewing the upsurge in loan losses ex post, we s ee a "cluster of error" in bank lending: overexpansion in certain loan categories (LDC loans, commercial real estate, HLT loans), or overoptimism regarding their repayment prospects. 8 At the end of 1990:3, real estate assets (real estate loans plus mortgage-backed securities plus foreclosed properties) comprised 30.6% of total commercial bank assets, up from only 18.8% at end of 1984 (FDIC 1990, 1—2). Overbuilding in commercial real estate was evident from office vacancy rates, which approached 20% in many cities (Mandel 1991, 30). Analysts at the Federal Reserve Board (Duca and McLaughlin 1990, 487) noted that "concerns about the quality of real estate loans appear strongest in areas in which land prices had risen sharply in previous years."
This pattern—that recession is the sharpest where the expansion had previously been most vigorous—is consistent with "monetary maIinvestment' theories of the business cycle, a class of theories that includes the work of the Austrian school in the 1920s and 1930s and that of Robert E. Lucas, Jr., in the 1970s and 1980s. As Lucas (1981, 9) has commented, this work insists on "the necessity of viewing (recurrent business cyclesi as mistakes." The theoretical problem is then "to rationalize these mistakes as intelligent responses to movements in nominal 'signals' of movements in the underlying 'real' variables we care about and want to react to." That is, a monetary malivestment theory traces the clustered business failures and unemployment of the recession phase to decisions (retrospectively inappropriate) made by labor and capital owners during the expansion phase, and appeals to monetary disturbances to explain why these decisions appeared sensible at the time they were made.
Rational-expectations work in monetary business cycle theory emphasizes that unanticipated monetary expansion generates unexpectedly high nominal demand for outputs. The Austrian theorists emphasized that new money, injected into the loanable funds market, reduces real interest rates in the short run. Both effects misleadingly signal businesses that their real profitability has increased, and so spur the unsustainable real expansion that constitutes the boom period. Real interest rates (measured by the annualized nominal interest rate on three-month T-bills minus the contemporaneous annual inflation rate) In the United States fell during the 1970s, and were actually negative from 1974 through 1980 (Kohn 1991, 729). Merely holding inventories appeared to be profitable. With disinflation after 1980, real Interest rates rose sharply, and nominal demand no longer outran expectations.
Richard Saisman (1990, 25—28) has offered the interesting hypothesis that U.S. banks may have been directly (as well as indirectly, via the business cycle) weakened by expansionary monetary policy. The Fed rapidly expanded the monetary base, leading in textbook fashion to the multiple expansion of bank deposits and loans. Meanwhile, Saisman argues, banks were not able to raise or Internally generate enough capital to keep pace, so that their capitali asset ratios fell. This argument implies that movements in the banking industry's capital/asset ratio should be predominantly associated with movements in the denominator (assets), with the numeratar (capital) remaining relatively stable. At least in the 1980s, however, industry aggregates do not show this pattern. Figure 1.2 plots year- to-year compound growth rates (differences in natural log levels) for capital, assets, and the capital-asset ratio. It shows that capital growth has been at least as volatile as asset growth.
Various regions of the United States have taken turns being worst hit with bank asset-quality problems during recent quarters. but all except the Midwest and Central regions have been seriously hit. Southwestern banks' portfolios have not yet recovered from problems that came to a head in 1988 and 1989. At midyear 1991 the "troubled real estate asset rate" as measured by the FDIC ("noncurrent real restate loans plus other real estate owned as a percent of total real estate loans plus OREO") still exceeded 10% in Arizona, Texas, Oklahoma, and Louisiana and exceeded 8% in New Mexico and Colorado. Northeastern banks have had the biggest recent problems with real-estate loans. Troubled real estate asset rates exceeded 10% in Rhode Island, Massachusetts, Connecticut, New Hampshire, New Jersey, New York, and the District of Columbia and exceeded 8% in Maryland and Virginia. No other state in the nation had a rate exceeding 8%, and only four (Maine, Vermont, Pennsylvania, and Florida) had rates exceeding 6%. Nationally, 11.2% of banks lost money in the first half of 1991, but 25.7% in FDIC's Northeast region did so. The biggest increases in loan loss provisions were taken during that period by banks in the Northeast and in the West. The largest increase in noncurrent loans was recorded by banks in the West. where 19.2% of banks lost money.
Secularly Shrinking Profitability
For the aggregate of federally insured commercial banks, the ratio of net income to assets has declined from 0.68 for 1980—85 (the average of annual figures) to 0.53 for 1986—90:2 (Barth, Brumbaugh, and Litan 1990, 121).° At the lower tail of the earnings distribution, where net income can be negative, cumulative losses can deplete capital and cause insolvency. Some 354 banks have reported losses every year from 1986 to 1989 (Fromson 1990, 119).
Much has been written in recent years about the erosion of the profitability of commercial banking under the impact of competition from nonbank intermediaries and from securities markets. Securitization is estimated to have reduced the spread on residential mortgages by fifty to one hundred basis points (Barth, Brumbaugh, and Litan 1990, 116, citing Rosenthal and Ocampo 1988). Increasing numbers of corporations, especially those with better bond ratings than the money-center banks, find it cheaper to issue commercial paper to investors directly than to borrow from banks at traditional spreads. With deposit interest rate ceilings being lifted after 1980, there has also been greater price competition for deposits among banks and between banks and thrifts.
In light of this we should note the apparently contradictory view that observed spreads between deposit and loan interest rates have not shrunk. A Federal Reserve Bank of New York staff study of Recent Trends in Commercial Bank Profitability (1986, 16) declared that "despite all the structural changes relating to interest rates and interest rate competition of recent years, there has been no visible impact on the net interest margins of the banks." It is true that, abstracting from Loan loss provisions, industry-wide net income has been stable as a share of assets (Duca and McLaughlin 1990, 477). Net interest income actually shows a slight upward trend over the 1980s, from 3.03% of assets in both 1980 and 1981, to 3.40% in both 1989 and the first half of 1990 (Barth, Brumbaugh, and Litan 1990, 121).
This only means, however, that in an accounting sense it is loan loss provisions, and not declining spreads as they are measured ex ante, that account for the decline in return on assets. Increasing loan loss provisions (typically a belated response to increasing default rates) reveal ex post that spreads actually have declined for loans of a given risk class. Loan Loss rates have been rising more or less steadily since 1962 (Barth, Brumbaugh, and Litan 1990, 117, 119). Lower-quality loans have been booked at ex ante spreads that used to be re served for higher-quality Loans. Barth, Brumbaugh, and Litan (1990, 117) argue that the movement of blue-chip borrowers to the commercial paper market, making banks unable to place loans of the traditional sort, helps explain why banks have taken on the additional risk that is evident from rising loan losses. The loss of traditional borrowers does explain why commercial and industrial (C&I;) loans fell from 21% of bank assets down to 19% over the course of the 1980s, to be replaced by loans with higher default risk. But it does not explain why banks, as Barth, Brumbaugh, and Litan (1990, 128) elsewhere document, chose to take on more portfolio risk over the course of the 1980s by reducing the share of cash and investment securities in their portfolios, or why they increased the share of real estate loans by more percentage points than the share of C&I; loans declined. It does not explain why "within the real estate loan category, banks shifted toward the riskiest borrowers," namely construction and development loans and commercial mortgages.
Most importantly, the shrinkage of margins on traditional loans does not by itself explain why banks have underpriced loans to their new borrowers. That is, it does not explain why they collected exante spreads too small to preserve income net of loan losses except in those unlikely states of the world (which did not obtain) in which the new loan portfolios had no higher a default rate than the old. To explain this mistake we need either to explain why banks would have failed to perceive that default rates were likely to be higher, or we need to explain why they became more willing to take on portfolio risk.
The Roles of Regulation and Deposit Insurance
The roots of the thrift industry crisis in regulation and deposit insurance are now well understood (Kane 1985, 1989; Brumbaugh 1988; Benston and Kaufman 1986). A number of long-standing legal restrictions, particularly restrictions against product-line diversification, made (and still make) thrifts weaker and more vulnerable to adverse shocks than they would otherwise be. Such restrictions alone cannot explain why failures exploded in the 1980s, for they did not suddenly become more binding. But they help to explain why the adverse cyclical and secular factors discussed above brought down as many thrifts as they did.
In the most general terms, the surge of thrift failures in 1981—82 can be attributed to interest-rate risk.' By funding Long-term fixed- rate mortgages with short-term deposits, thrifts were implicitly betting heavily against a large rise in interest rates. They lost the bet. The average explicit interest cost of savings deposits in FSLIC insured institutions rose from 6.6% in 1978 to 11.2% in 1982 (Kane 1989, 12—13, table 1—2). Although interest rates on new mortgages also rose, the thrifts' assets consisted largely of conventional fixed-rate mortgage loans made in the 1960s and '70s, paying between 6% and 9%. Borrowing at 11% to fund old mortgages paying 6 to 9%, hundreds of thrifts soon found their equity consumed by negative income flows.
The thrift crisis continued (or a second thrift crisis arose), despite the fall in interest rates after 1982, because the regulators' failure to close literally hundreds of insolvent thrifts created an army of institutional "zombies," economically dead but not yet buried, that rationally gambled for "resurrection." In general terms, unclosed thrifts substituted credit risk for interest-rate risk. Long-odds gambling in the form of high-risk lending offered the owners of insolvent thrifts their best hope of getting back into the black, making their shares worth something again. The downside risk fell entirely on the FSLIC: the owners of thrifts with zero net worth had literally nothing to lose. Cole, McKenzie, and White (1990) provide econometric evidence that thrift failures in the late 1980s were swelled by risk- taking strategies begun earlier in the decade, strategies motivated by low net worth and forbearance. Failed thrifts had riskier portfolios than nonfailed thrifts, namely, higher concentrations of nonresidential mortgages, land loans, and real estate investinents. "Moral hazaid" behavior is indicated by the lower failure probabilities for institutions whose ownership structures gave their managers less to gain from risk taking: mutual institutions compared to stock institutions, publicly traded stock Institutions compared to closely held stock institutions. Failure probabilities were higher for banks with higher managerial expenses. [this sounds familiar]
Competition from zombie thrifts, who bid deposit rates up and loan rates down, made survival more difficult for still-solvent thrifts. By 1986 the average explicit earnings spread on new mortgage loans was smaller than it had been since 1972, and the spread net of average thrift operating expenses had reached a historic low (Kane 1989, 12—13, table 1—2). The result was a mushrooming number of economically insolvent institutions (Kane 1989, 26, table 2—1) and an accumulation of red ink that finally exceeded the FSLIC's resources by hundreds of billions of dollars,
Because commercial banks were carrying less of a mismatch between the repricing frequencies of their assets and liabilities than were thrifts, the banks were less victimized by the run-up in the level of interest rates in 1979—82. But banks with large maturity mismatches (borrowing short to lend long) did suffer large losses because of frequently inverted yield curves (short rates above long rates) during the period (FRBNY 1986, 117—21). The yield curve returned to its normal slope in 1983. The number of problem banks grew steadily up to 1986, however, with the rising failure rate among commercial and industrial business and increasing loan losses. The secular trend toward declining spreads continued to exert itself.
The increasing number of banks approaching the brink of insolvency has meant, as it meant in the case of thrifts who reached or crossed the brink a few years earlier, an increased attraction to financial "gambling." Booking high-risk loans without premia sufficient to cover probable defaults is a way to maximize the expected value of the bank to its owners, given that the owners can pass losses In excess of equity on to the FDIC. Deposit insurance, in other words, has created in commercial banking the same two "moral hazard" problems that have been much discussed in connection with the second wave of the thrift industry crisis.
1. Insured depositors (de jure or de facto) do not discipline weak institutions by demanding higher deposit rates or by moving funds to stronger institutions. Riskier banks can essentially sell a lower- quality product at the same price because customers are fully covered against product failure. Without risk-sensitive insurance premiums as a substitute for depositor discipline, deposit insurance subsidizes risk taking. A bank's expected return on assets can be increased by taking on a riski er (higher-variance) loan portfolio, but its cost of funds does not rise even though the bank is more likely to fail. Maximizing expected net payoff therefore pushes the bank to a riskier position on the risk-return frontier than would be taken by a bank whose uninsured depositors demanded compensation for an increased risk of default on their claims, or by a bank whose insurer priced its premiums to reflect insolvency risk. Banks have accordingly chosen lower capital/asset ratios as deposit guarantees have grown in scope and in implicit value.
2. In combination with forbearance, deposit guarantees enable and encourage insolvent banks to gamble for resurrection from economic insolvency. An insolvent institution can bid for funds with little risk premium to try to grow back into the black. Its owners have everything to gain and, with FDIC absorbing the downside risk, nothing to lose. With enough forbearance they can even operate a Ponzi scheme, using new deposits to pay the interest on old deposits [This is what has happened during the last twenty years] (Kaufman 1988, 574).
The U.S. system of deposit guarantees thus serves as a background condition explaining banks' risk-preference behavior: it encourages banks close to insolvency to take on greater risks. This "moral hazard" problem likely intensified in the 1980s because the effective coverage of deposit guarantees was extended, and because the guarantees were increasingly mispriced. The Depository Institutions Deregulation and Monetary Control Act of 1980 raised explicit deposit coverage to $100,000 per account, from $40,000. This made it cheaper for depositors to get full coverage for large amounts by spreading funds among banks, with or without the help of a broker, and correspondingly made it easier for risk-prone banks to acquire funds (Kaufman 1988, 574).
Perhaps more importantly, the FDIC extended de facto full coverage to all liability holders of large banks. The FDIC increasingly resolved bank failures by "purchase and assumption," arranging takeovers that fully protected uninsured liability holders from loss, as in the Franklin National Bank case in 1974. There still remained the threat that uninsured depositors might take losses if a bank had to be liquidated because it was in such bad shape that no purchaser could be found. The FDIC removed even that threat in the 1984 Continental illinois case, when the FDIC itself effectively purchased (nationalized) the insolvent bank (O'Driscoll 1988, 666). The FDIC enunciated the so-called too big to fall doctrine, under which even uninsured liability holders would be protected from any loss. Consistent with the extension of de facto guarantees, Short and Robinson (1990, 14—15) note that while high-risk banks normally had to pay a premium rate to attract large uninsured CDs in the mid-1970s, studies of more recent data do not consistently show risk premia on CDs or subordinated debt. Finally, because banks' exposure to interest-rate risk increased with the increased volatility of interest rates in the 1980s, but their deposit insurance premiums did not, the risk subsidy implicit in FDIC guarantees increased (Benston and Kaufman 1986, 62).
Evidence of increased bank gambling can be found in the changing composition of bank portfolios. Cash and investment securities declined to only 27% of assets in 1990 from 36% in 1980. Loans rose to 61% of assets from 54%, real estate loans to 23% from 15%. Within real estate, as noted above, banks shifted toward riskier borrowers, namely construction and development loans and commercial mortgages (Barth, Bmmbaugh, and Litan 1990, 128). In 1989, bank real estate loans exceeded C&l; loans for the first time ever (Fromson 1990, 120). Further evidence of increased bank gambling can be seen in the rising cost of resolving failed institutions: 20.3% of deposits in 1989 failures, double the 10.2% figure for 1985 failures [TODAY, THE COST OF RESOLVING FAILED INSTITUTIONS IS 100%. Failing banks have no assets left in them] (Barth, Brumbaugh, and Litan 1990, 29, table 4). Higher resolution costs mean either that authorities were slower to close banks after their economic net worth crossed into the negative region, or that the banks' net worth fell more rapidly [With the cost now 100%, the authorities are more desperate than ever to hold the system together]. Increased reliance on high-risk "double-or-nothing" lending strategies can have both effects. It can increase the discrepancy between the economic value of assets and their value according to regulatory accounting principles, so that authorities are slower to recognize negative net worth. And it can make asset values fall more rapidly because returns on high- risk assets are more sensitive to changes in the state of the world.
Just as competition from zombie thrifts impaired the profitability of solvent thrifts, competition from zombie banks has weakened solvent banks. These spillover effects are ironic, because the ostensible purpose of deposit insurance was to prevent spillover effects (namely the spread of runs) from unsound to sound banks. Short and Robinson (1990, 7—8) find that Insolvent (but still open) Texas banks bid up the deposit rates paid by other Texas banks, and may have increased the number and size of insolvencies. They also point out that the FDIC policy of resolving institutions with assistance, absorbing bad assets to keep them open, puts unassisted institutions (who have to swallow their own bad assets) at a competitive disadvantage. Conversely, others have noted that the shrinkage of the thrift industry, in part due to RTC closures, has helped strengthen commercial banks by giving them retail deposits that are on average a cheaper source of funds than brokered deposits (Duca and McLaughlin 1990, 488).
A banker quoted anonymously in the New York Fed study Recent Trends in Commercial Bank Profitability (FRBNY 1986, 43) explained clearly the incentive of an unprofitable bank to gamble for recovery;
One banker said that traditional corporate banking faced two alternatives, both of which are "routes to going out of business." One is just to say "no" to underpriced risky deals. The other is to take the risks, the alternative most organizations are driven to by the need to occupy an existing staff of loans officers and supporting personnel. This latter alternative, he said. simply results in going out of business "more dramatically," especially in a disinflationary period when the Inflation that temporarily hid the risks is no longer there to mask them.
Our current deposit guarantee system allows a bank to take the risks without a correspondingly greater cost of funds, despite the increased likelihood that it will exit the business "dramatically." We have seen all too many dramatic exits in recent years. Taxpayers have discovered that they are the financial "angels" obliged to cover the costs of what threatens to be a very expensive show.
A minimal goal for banking policy would be to give bankers the proper incentive to choose the less dramatic route to going out of business. A bank should be led to retire gracefully as Its profitability declines, rather than to run up a bill for other banks, or taxpayers, in the course of fighting the inevitable. If market forces dictate that the banking industry as a whole is to shrink, let It not consume others' wealth in the process. Let it shrink quietly and promptly, so that financial resources can be reallocated with minimum waste to what promise to be more valuable uses.
The absence of incentives to gamble for resurrection can be seen in historical banking systems with unlimited liability for bank shareholders. In such systems unprofitable banks would voluntarily wind up their affairs without waiting for insolvency. Bank owners had no incentive to pursue double-or-nothing strategies even as the bank's net worth became negative, because further losses in net worth continued to fall entirely on the shareholders, rather than on depositors or on a deposit guarantee agency. It may be that unlimited shareholder liability, or even extended liability, is not generally the optimal risk-sharing arrangement between shareholders and depositors. That is a question financial markets can resolve in the absence of subsidies and legal restrictions. It is not obvious that extended liability is incompatible with tradable shares, though for obvious reasons shareholders whose own exposure depends on the wealth of their coshareholders might want shares to carry covenants regarding ownership qualifications.
In the absence of government deposit guarantees, a caveat emplor policy prevails. With entry free into both limited-liability and extended-liability banking, depositors who choose limited-liability banks are choosing freely to expose themselves to default risk, in exchange for whatever comparative benefits limited-liability banks can offer them. The usual objections to such a policy are (1) that depositors would attempt to free-ride on one another's efforts to monitor the bank, so that too little monitoring would take place; and (2) that depositors would run on suspect banks, setting off contagious banking panics.
The first objection is not really specific to banking. Quality-assurance problems of this sort are generally handled by certification agencies. In banking, a private clearinghouse association has historically been the agency acting to certify the solvency and liquidity of its member banks, primarily because each member bank (who accepts the liabilities of its fellow members daily) has a strong interest in receiving such quality assurances (Timberlake 1984; White 1992, ch. 2).
The second objection is undercut by the historical evidence that a run on a suspect bank is not generally contagious. In the absence of legal restrictions that weaken banks in similar ways, bank failures do not occur in droves, and so depositors do not rationally infer from one bank's difficulties that all others are about to default. No contagions are recorded in Canadian or Scottish banking history. where banks were free to branch nationwide and to capitalize adequately. Even in the United States there is little evidence (outside the exceptional years of 1929—33) of runs spreading generally from insolvent to solvent banks (Kaufman 1988, 566—71; Schwartz 1988, 591—93).
Depositors fleeing suspect banks generally redeposit their funds in sound banks. Such movements were occurring in the early 1930s, and one suspects that the private interest of weak banks in opposing such a "flight into quality" explains why small banks enthusiastically supported the formation of the FDIC, while many large banks opposed it. If so, the same sort of interest today would oppose "coinsurance" proposals (limiting deposit guarantees to, say, 90% of deposits). Weak banks (a category that today includes many of the largest banks as well as the smallest U.S. banks) may fear that coinsurance might deliver on its advocates' claims: it might reimpose market discipline.
Proposals to limit deposit guarantees to "narrow banks" (Litan 1987) are a step toward the goal of an undistorted financial system provided that banks are free to issue explicitly and credibly unguaranteed accounts on whatever terms informed customers find agreeable. A number of options are open to banks to make their accounts run-resistant or even run-proof. Deposit contracts could contain notice-of-withdrawal clauses. Checking accounts could be linked to money-market mutual funds, equity rather than debt claims. Capital adequacy assurances, or even extended shareholder liability, could be offered. Some sort of private deposit insurance might be feasible.
Assuring that the best sorts of financial contracts win out on a level playing field requires eliminating the discriminatory practices in the current operation of the clearing and settlement system (e.g., the exclusion of money-market funds from direct use of the Fedwire, and the unpriced guarantee of interbank payments made by wire). Ideally the payments system would be entirely privatized.
The case for government deposit guarantees in a deregulated environment is not persuasive. Any government deposit guarantees that remain in this environment must at a minimum be self-financing. If the guarantee system cannot cover its costs, it is hard to defend its efficiency. If the deposits of banks (however narrow) are provided with government guarantees at rates subsidized by general taxation, there are inadequate incentives for savers to seek efficient alternative intermediary forms (such as mutual funds). In the context of proposals currently on the table, this means that if making the Bank Insurance Fund sell-financing by raising FDIC assessments on banks makes the banking industry shrink that much faster, so be it.
Bankers as Scapegoats for Government. Created Banking Crises in U.S. History
Richard M. Saisman
If there is anything more tragic than our current banking crisis, it is that the crisis is being blamed on the wrong group, on the bankers, instead of on the primary culprit, government intervention [exactly!!!]. The tragedy lies in falling to identify the fundamental cause of the problem, thereby ensuring its continuance. Bankers are not entirely innocent of wrongdoing in the present debacle, but to the extent that bankers have been irresponsible, it has been primarily government intervention that has encouraged them to be so. More widely, it is irresponsible government policy that has made the U.S. banking crises of the past century so frequent and seemingly so inevitable. Government has created these banking crises—sometimes inadvertently, at other times with full knowledge—by making it nearly impossible to practice prudent banking. Having done so, government has then pointed to bad banking practices as sufficient cause for still further interventions in the industry.
I. The Context of the Current Banking Crisis
The view that today's banking crisis is due primarily to the mismanagement and fraud of pri vate bankers underlies most popular accounts of the crisis.' Critics are Inclined to blame private bankers for banking instability because they wrongly believe that unregulated banking systems are inherently unstable and that regulation is required to restrain some natural tendency of private bankers to engage in mismanagement and fraud. Central banking is said to provide a restraining influence on the destabilizing urges of the private banking system, while free banking is seen as inherently prone to instability. The recent, burgeoning literature on free banking overthrows this conventional wisdom and defends free banking as an inherently stable system made unstable only by legal restrictions and central banking-related interventions. In this view, bad banking comes not from free markets but from perverse public policy. [absolutely]
Guided by erroneous assumptions about the nature of free banking and central banking, analysts of the current crisis typically stress the symptoms (bad banking practices), and overlook the underlying disease (government intervention), as the cause of our problems. For example, many commentators and bank regulators are satisfied to cite anecdotal evidence from the current banking crisis to draw the obvious conclusion that bankers like Charles Keating are incompetent and dishonest, and then to claim that these and similar cases represent the sum and substance of an explanation of the banking crisis. Such figures simply are not representative of the entire industry. While there is no denying that bankers such as Charles Keating exist, we can only understand the fundamental cause of our banking crisis by identifying the institutional arrangements that make such bankers possible. As I argue below, central banking and legal restrictions have institutionalized unsafe banking.
Today's banking crisis is only the latest in a long series of U.S. banking crises blamed on bankers but actually caused by government intervention.
My reaction: I am working on an entry putting everything together.