(FORTUNE Magazine) – IS AN S&L-SIZE DISASTER brewing among America's commercial banks? With banks already failing at the highest rate since the Great Depression, both the Congressional Budget Office and the General Accounting Office have recently issued dire predictions. If the economy slows further, they say, a key line of defense -- the Federal Deposit Insurance Corp., whose $13.2 billion insurance fund pays depositors when banks fail -- could quickly run out of money. Says Dan Brumbaugh, a former research scholar at Stanford University, one of the few experts to predict the S&L debacle: ''Make no mistake about it. We are in the incipient stages of the next crisis.'' So far, however, the warning lights are flashing only yellow. Banking's worst problems are concentrated in ten states (see map on following page). Compared with the thrifts, banks have far more capital, a much more solid earnings base, and considerably better management. If the U.S. avoids a recession and if policymakers learn the right lessons from the S&L mess -- admittedly no small assumptions -- the tab that's ultimately handed to taxpayers because of bank failures could be held to a small fraction of the $150 billion thrift bailout bill. No one denies that the industry faces hard times. ''The U.S. banking system is under stress,'' says FDIC Chairman William Seidman. Three hundred fifty- four banks have reported annual losses every consecutive year since 1986. Among big money center banks, profits are so poor that some have been eating into capital just to cover dividend payments. Write-offs as a percent of loans have risen from 0.84% four years ago to 1.16% -- worrisome by historical standards. As for liquidity, banks depend more than ever on brokered deposits -- an expensive and fickle source of funds, gathered together by stockbrokers and shunted off to whoever pays the highest yields. (Many now-defunct S&Ls were into the same game.) The unhappy result: Since 1987, U.S. banks have been failing at the rate of 200 per year, vs. just ten a year in 1980. To make depositors whole, the FDIC's insurance fund shrunk by $5.1 billion in 1988 and 1989. In July, three months after saying that the FDIC would not lose money a third straight year, Chairman Seidman was forced to acknowledge that the agency would be out another $2 billion in 1990. Since then 20 more banks have gone bust, notably the National Bank of Washington, whose failure could by itself cost the FDIC more than $500 million. SMALL WONDER that in the last year the average price of the 50 largest publicly traded commercial banks has dropped more than 40%. Frets James Sidell, president of UST Corp., holding company for a profitable and well- capitalized New England bank: ''This industry has one foot on a banana peel and the other in the grave.'' Recently Carole Berger, a respected analyst at the brokerage C.J. Lawrence, issued a report on high-risk lending at those same banks, with this headline: THE PLAGUE SPREADS. ANALYSTS FEAR EPIDEMIC. Bubonic banking has been incubating for more than a decade. It intensified in 1980, when the government removed longstanding ceilings on the interest rates banks and thrifts could pay depositors. With the demise of Regulation Q, as it was known, these institutions' cost of money soared as they were forced to bid for deposits against newly arrived money market funds. That change, in turn, spawned what Marc Perkins, chairman of the Tampa investment bank Perkins Smith Inc., calls ''the compulsion to lend'' among bankers. Says Perkins: ''The way it used to work was that toward the end of a credit cycle, the wise old bank president would go downstairs to his loan ( department and say, 'Fellas, don't make any more loans. Just buy T-bills, and instead of earning 4%, we will earn 2%.' But after Reg Q went the way of the dodo bird, the loan officer told the bank president, 'We can't do that because the yield on T-bills is below what we're paying for deposits.' So the president says, 'Well, forget I was here. You better keep making loans.' '' Other ailments with fancy-sounding labels fueled this compulsion by infecting earnings. ''Disintermediation'' -- epitomized by the shift among big corporate borrowers from bank loans to the commercial paper market -- began in the 1970s. But it exploded in the 1980s. So did ''securitization'' -- the packaging and sale, mainly by investment banks, of loans that were once the exclusive preserve of banks, such as mortgages, car loans, and credit card receivables. While banks could still originate these loans, securitization eliminated profits gained from holding them and earning interest over many years. WITH THEIR EARNINGS declining, banks have responded much as the S&Ls did, by making riskier loans for which they can collect higher fees and interest. In particular they have poured money into commercial real estate and highly leveraged companies. Last year for the first time ever in the U.S., real estate loans alone exceeded the banks' commercial and industrial lending. Unfortunately, what bankers are also discovering -- just as they did in the early 1980s after their disastrous lending spree in the Third World -- is that in an economic downturn, such loans go bad far more quickly and in far greater numbers than the more conventional variety. So far banking's woes haven't directly cost U.S. taxpayers a penny. The money that flows into the insurance fund comes from premiums paid by the 13,000 FDIC member banks. Despite the warnings from the General Accounting Office and the Congressional Budget Office, Seidman of the FDIC believes that can continue. ''We don't see failures sufficient to imperil our fund,'' he says. ''But that obviously depends on the economic environment.'' And what is Seidman's best guess about that environment? ''We're assuming that it remains about what it is now.'' That's also still FORTUNE's assumption (see Fortune Forecast). But to find out what might happen if the economy worsens we enlisted the help of Ferguson & Co., a well-regarded bank consulting firm in Washington, D.C. It built a model that predicts how much equity capital U.S. banks are likely to lose + under two scenarios -- one in which the economy merely slows further from its current sluggish pace and another in which the U.S. slides into a severe recession. The results suggest that the folks at the GAO and CBO are right to worry.

Ferguson's model assigns a set of values to four key factors that reduce bank capital. For example, in the ''slowdown'' scenario, it assumes 30% of loans already on banks' books as nonperforming loans would go uncollected, whereas in the ''recession'' case, it assumes 50% must be written off. Such losses are then partly offset by the banks' compensating earnings power and various accounting complexities, such as the tax credits generated by the losses. In the less dire scenario Ferguson assumes that the average bank earns a return equal to 0.75% of its assets, while the severe scenario lowers that return to 0.50%. All these assumptions are drawn from close analysis of earlier banking disasters. Other banking experts consulted by FORTUNE agree that as a stress test, Ferguson's model is reasonable. It is also conservatively based on the most recent publicly available FDIC data. Because these date back to March 31, they understate the recent deterioration in many banks' real estate loan portfolios, particularly in hard-hit regions such as New England. So what's the potential damage? Under the slowdown scenario, the entire U.S. banking system suffers a net loss of some $19.5 billion in equity capital. David Cates, Ferguson's chairman and the model's creator, figures that would translate into a $5 billion hit to the FDIC's insurance fund. A loss that big, in turn, would give rise to further trouble. ''Once the fund slips into single digits ((below $10 billion)),'' says Cates, ''it risks becoming effectively insolvent. That's because as it dwindles, big uninsured depositors ((the FDIC only insures accounts up to $100,000)) will become alarmed and move their money from weaker banks into either strong banks or T-bills.'' Such a flight from weak banks could trigger failures and a corresponding drain on the fund. To forestall such a run, Cates estimates that taxpayers might be asked to pony up $5 billion to replenish the FDIC's coffers. In the recession scenario, the banks' equity capital shrinks by $86.3 billion -- a far more worrisome 41% cut in the system's capital cushion. But healthy banks would absorb part of the capital loss, Cates figures, and the FDIC could tap them for higher premiums. Taxpayers, he calculates, would be left facing a maximum bill of $40 billion. That's some $100 billion less than even the most conservative estimates of the potential cost of the S&L bailout -- good news if you're the type who insists on always looking at the half-full side of the glass. There's also comfort, though of a distinctly chilly sort, in Cates's finding that most big banks with the heaviest exposure to high-risk loans would survive a really nasty recession (see box). STILL, A POTENTIAL taxpayer liability of $40 billion is hardly chicken feed. And compared with the thrifts, banks play a far more critical role in the economy. That's why it's essential that legislators and regulators shore up their safety and soundness with greater skill and alacrity than they have displayed in the S&L debacle. At the moment Congress is eagerly awaiting a Treasury Department study on bank reform that's meant to provide a foundation for new legislation from the Bush Administration. But that study is not due until December. In the meantime, here are FORTUNE's reform proposals, drawn from extensive interviews with industry experts: -- Intervene earlier. Banking regulators lack the kind of clearly defined early warning policy that the SEC uses to police Wall Street. As a result they often step in after banks are too far gone to save. Lowell Bryan, the top banking consultant at McKinsey & Co., would like the government to draw up a well-defined system of green, yellow, and red light signals that could help both banks and regulators determine when an institution is veering off course. Here's an example. Richard Randall, a vice president of the Federal Reserve Bank of Boston, points out that as early as 1986, some medium-size New England banks had made loans to builders that amounted to more than 20% of their entire portfolios. In one extreme case, a bank's construction loans amounted to 70% of total loans -- a clear signal to call in the cops. -- Improve bank accounting. Today banks mask the true health of their balance sheets in a number of ways. For one thing, they can carry loans on their balance sheets at historical value -- even if their value later deteriorates. This practice creates a huge incentive for lenders to roll the dice. To remove that incentive, banks should be forced to do what regulators do only after they move in to liquidate -- mark loans and investments to market wherever possible. Where no liquid market exists for an asset, as in the case | of nonperforming loans, then some proxy for what that loan realistically should be earning -- the prime rate plus a few points, say -- should be used to discount its value. -- Raise deposit insurance premiums. Since its inception in 1933, deposit insurance has proved wonderfully effective in maintaining investor confidence. Its huge drawback, however, is that it enables bankers to make gamier loans than they would if the government were not there to bail out depositors. In 1980, the same year that Reg Q was axed, Congress made that drawback worse by raising the limit on insured accounts from $40,000 to $100,000. But current proposals to solve this problem by returning to the $40,000 ceiling or limiting insured accounts to one for each individual miss the point: The real exposure taxpayers face is the uninsured deposits of large corporations and pension funds. Under the ''too big to fail doctrine'' -- last seen at work in the 1984 bailout of Continental Illinois -- the Federal Reserve and other bank regulators have rightly served notice by word and deed that they would not allow large banks with vast numbers of such depositors to fail. The damage to the banking system from the stampede that would ensue as these depositors scrambled for safety would be too great. Says Charles Munger, the vice chairman of Berkshire Hathaway: ''You want banking insured. Bank panics are for the birds.'' Instead, Washington would do well to adopt a two-step reform. First, it should stitch up the existing safety net by raising the insurance premiums that banks pay. Seidman of the FDIC has already proposed boosting them from 12 cents for every $100 of insured deposits to 19.5 cents -- the maximum allowed by law. That would raise some $5 billion for the insurance fund. Congress should also follow Fed Chairman Alan Greenspan's recent advice and grant the FDIC new authority to go higher if necessary. Individual depositors, of course, will share the cost of higher premiums, either through higher fees for services or lower interest rates on deposits. They should. They've been getting too good a deal on deposit insurance for too long. -- Limit insurance to supersafe banks. Once banking's insurance safety net has been made more secure, it must then be cut back so that it covers only low- risk lending. McKinsey's Lowell Bryan calls this establishing ''core banks.'' To remain insured by the FDIC, such banks would abide by strict limits on how they operate. For example, they might be allowed to lend only to individuals, small businesses, and midsize companies and to invest only in government securities. At the same time Congress should create separate uninsured banks that would be free to engage in riskier activities, such as real estate or overseas lending. These banks would also be allowed to enter other long-forbidden financial service businesses like investment banking and insurance, as long as they met those industries' regulatory and disclosure standards. Such institutions would most resemble companies like American Express, General Electric Credit, and Ford Motor Credit. If they emulate this crowd's fine track record, these new banks should have little trouble attracting capital from investors. But under this arrangement, those seeking higher than average returns on deposits would be the ones exposed to any market risk -- not U.S. taxpayers. Finding a way to limit deposit insurance to a narrower version of banking ought to be the keystone of any comprehensive reform. But while this notion has been gaining support, it's still not clear that the White House and Congress will move swiftly enough to avert a crisis. Says Barry Weingast, a senior fellow at Stanford University's Hoover Institution and a close student of the politics of the S&L debacle: ''The issue is this: Are our politicians sufficiently chastened or embarrassed by the savings and loan crisis that they will do the responsible thing this time?'' There are grounds to worry that they aren't. For example, the recent reaction of some legislators from New England to that region's banking woes has been distressingly similar to the behavior of their Texas counterparts when the thrifts began to bleed in the mid-1980s. Regulatory forbearance, rather than rigor, is all too often the watchword. Most citizens, of course, can't be expected to follow the complex ins and outs of banking reform. Therein lies another problem. Observes Stanford's Weingast: ''The fundamental flaw we saw in the S&L crisis is that even when politicians understand a problem, unless they feel they can get credit from voters for solving it, they tend to let it unfold.'' The best way to avoid the fire next time will be if voters at least send Washington a clear message that they will neither forgive nor forget another S&L-style sleepwalk.