Economists describing how regulators competed for “customers” by promising to be laxer in supervision coined two of the most telling phrases to come out of the S&L debacle: “competition in laxity” and “race to the bottom.” The novel aspect is that economists endorsed these pejorative terms because the race was toward greater deregulation. In the early 1980s, economists knew that regulation was the problem, so anything that reduced regulation was desirable. Richard Pratt shared this mindset when President Reagan appointed him Bank Board chairman in 1981.
The Bank Board was at the bottom of the federal financial regulatory heap before Pratt’s deregulation and desupervision. Jim Ring Adams (1990, 40) aptly described it as “the doormat” of federal regulators. I describe the problems that the board had with examination, supervision, and enforcement briefly, but they were among the most important contributors to the debacle, and a similar problem with SEC resources is one of the most important causes of the ongoing financial scandals at the time I write. Criminologists call this a “system capacity” problem (Calavita, Pontell, and Tillman 1997, 136). The regulatory and criminal justice systems lacked the resources (and often the will) to stop the control frauds.
The first problem was institutional structure. Agencies need to integrate examination and supervision, and the banking agencies did so. The Bank Board separated them in the worst possible manner. Examiners and supervisors worked for different bosses and different employers! The examiners were federal employees; the supervisors were Federal Home Loan Bank (FHLB) employees. Member S&Ls owned the twelve FHLBs. This posed an obvious potential conflict of interest. The FHLBs were not subject to federal limits on staffing or salary. We paid supervisory agents far more than examiners. The life of an examiner was constant travel and frustration; the life of a supervisor was cushy. The examiners had little authority. Only supervisors could recommend actions or issue directives. Naturally, the two groups often disagreed and were antagonistic.
The first common boss for examiners and supervisors was the Bank Board chairman, so no one else could resolve disputes. The Bank Board called its top supervisor the director of the Office of Examinations and Supervision (OES), but supervisory agents reported to the “principal supervisory agent” (PSA)—the president of each FHLB—not to the OES. The PSAs reported to the chairman of the Bank Board, not to the OES director. Each FHLB was a separate duchy with substantial political power through its industry membership. The structure violated every rule of proper management and proved disastrous.
Second, examiners used state-of-the-art techniques—from the 1930s. As late as 1986, examiners drafted each report in pencil. It took an average of two months to type a report.
Third, the industry hated the concept of examiners’ and supervisors’ exercising judgment, which is how banking regulators act (see Appendix B). S&L regulators could only enforce rules (Strunk and Case 1988). If an S&L was doing something unsafe, it could do so with impunity unless it violated a rule. If an S&L was acting sensibly but violating a rule, then supervisors would order it to stop. The enforcement branch reinforced this tendency. It would not take action absent a violation of an express rule (NCFIRRE 1993a, 50–51).
Fourth, the Bank Board virtually never made criminal referrals when it found fraud, and the Justice Department rarely prosecuted. The Bank Board had no formal criminal referral system. The attorney general, Edwin Meese, exacerbated the critical shortage of white-collar prosecutors by transferring many of them to prosecute pornography. (Meese acted to please the nation’s leading antiporn activist—Charles Keating.)
Fifth, the Bank Board got all of its money from industry assessments, but was subject to federal restrictions on how many examiners it could hire and how much it could pay them. Worse, its banking regulator “competitors” were exempt from many of these limits. Good examiners could make a lot more money by joining the banking agencies—the starting difference in annual salary was about $3,000; it grew to over $10,000 for senior analysts (Strunk and Case 1988, 141). Bank examiners had greater authority and prestige (and computers instead of pencils). There were exceptions, but the system ensured that Bank Board examiners generally would be low in quality.
Pratt faced an impossible situation. Virtually every S&L was insolvent on a market-value basis by 1981.1 By mid-1982, the industry was insolvent by roughly $150 billion (NCFIRRE 1993a, 1). The FSLIC fund had only $6 billion in reserves, so it was hopelessly insolvent. The Reagan administration refused to admit that the industry was insolvent, refused to give the FSLIC any additional money to close failed S&Ls, and ordered Pratt not to use the FSLIC’s statutory right to borrow even the paltry sum of $750 million from the treasury. Pratt’s orders were to cover up the S&L crisis.
The cover-up was particularly critical to the administration in 1981. Ronald Reagan’s campaign promises were to cut taxes, increase defense spending, and balance the budget. Those three promises, of course, were inconsistent, as his budget director, David Stockman, would later admit.2 The administration knew that if the public realized that the budget deficit was really $150 billion larger than reported, the resulting outcry could have prevented passage of the large tax cuts that the Economic Recovery Tax Act of 1981 (generally called the 1981 Tax Act) provided for.
The industry supported the cover-up because it didn’t want to report that it was insolvent. Pratt supported the cover-up because Bank Board officials shared the same nightmare, a national run on S&Ls. Pratt did not cause the interest rate crisis, but many would blame him if the system failed on his watch. Pratt made sure this did not happen. Congress supported the cover-up because the alternative was to cut popular social programs.
The cover-up optimized the industry for control fraud in several ways. The most direct contribution was abusive accounting. The Bank Board’s regulatory accounting principles (RAP) trumped generally accepted accounting principles (GAAP) for regulatory reporting purposes. Pratt developed “creative regulatory accounting principles”—the acronym said it all! I discussed the worst of these, loan loss deferral, in the introduction. The creative RAP provisions were the cherry on the sundae of accounting insanity. Two GAAP provisions composed the sundae. The largest accounting abuse came from GAAP’s failure to recognize market-value losses caused by interest rate changes. GAAP did not recognize the $150 billion loss in market value caused by interest rate increases.3
The other huge GAAP abuse was “goodwill accounting.” A word of encouragement: you will be able to understand it, you will be amazed at the scam, and you will know why policy makers must understand such scams. You will also be joining an elite group, for few understood it. In other books you can read that goodwill accounting was abusive, but not about how the scam worked. I describe in detail only the two accounting frauds central to the debacle; goodwill is the first.
It all starts with a simple, logical assumption drawn from economics: the best proof of market value is what an arm’s-length buyer pays for an asset. An arm’s-length buyer is an independent buyer acting in his own interests. (When economists assume “rationality,” they err if they fail to take into account what’s rational for a fraud.) Goodwill accounting among 1980s S&Ls was overwhelmingly fraudulent. Pratt’s priorities, because the FSLIC had only trivial amounts of money relative to the scale of the industry’s insolvency, were to avoid spending FSLIC funds to resolve failed S&Ls and to cover up the insolvency of the industry and the FSLIC. That meant that the FSLIC rarely used the normal means of resolving failures, i.e., paying a healthy firm to acquire the failed S&L. Instead, Pratt induced roughly 300 buyers to acquire failed S&Ls without any FSLIC assistance. Pratt called these “resolutions” and took credit for developing innovative techniques that reduced the average cost of resolving such failures by about 75 percent.
White-collar criminologists’ mantra is “if it sounds too good to be true, it probably is.” The obvious question is why entities knowingly took on net liabilities without FSLIC assistance. (A firm whose debts exceed its assets is insolvent; it is a net liability.) Accountants’ answer was “goodwill.” A firm can have greater value than the sum of its tangible assets less its debts. McDonalds is an example. It is worth far more than what it could sell its physical assets for, less its debts. It has a reputation for safety and cleanliness and is famous worldwide. This favorable reputation has great value, and we call that value “goodwill.” Accounting literature, however, calls it a “general, unidentified intangible” (FASB Statement 72), and I will explain later why that phrase is important to the S&L scam. The “intangible” part just means it isn’t a physical asset. The words “general” and “unidentified” indicate that the goodwill isn’t attributable to any specific, identifiable physical asset, such as the golden arches.
The concept of goodwill and the assumption of rationality are both reasonable propositions. Together, however, in the context of the mass insolvency of the S&L industry, goodwill created insane financial results. It optimized the S&L environment for control fraud. It helped cover up the mass insolvency of the industry. It allowed Pratt to claim that he had resolved failures at minimal cost and had contained the crisis, which allowed him to resign in triumph and begin a lucrative career at Merrill Lynch, trading mortgage products with the industry.
Here’s how the assumption of rationality and the concept of goodwill produced insanity. When you purchased an S&L through a merger, the assets and liabilities of the S&L you were buying were “marked-to-market.”4 As a practical matter, that meant that the S&L’s mortgage assets would lose roughly 20 percent of their value.5 Note that this result stems from the first GAAP accounting abuse that I discussed, the failure to recognize market-value losses caused by interest rate changes. Most S&Ls were insolvent on a market-value basis in 1981 by roughly 20 percent of their reported GAAP assets, so my example is realistic. This brings us to the fundamental balance sheet equation: assets –liabilities = capital. A typical acquired S&L might have reported under GAAP that it had $200 million in assets and $205 million in liabilities. Its GAAP insolvency was $5 million.
Here’s how the mark-to-market valuation transforms the situation. On a market-value basis, the S&L’s assets are worth 20 percent less than on a GAAP basis: $160 million, not $200 million. The market value of the liabilities is the same as their GAAP value, $205 million. You might think that this demonstrates that the S&L being acquired was insolvent by $45 million on a market-value basis, but if you think so, you have forgotten rationality and goodwill. It would be irrational knowingly and voluntarily to buy an S&L that was insolvent by $45 million without getting at least $45 million in financial assistance from the FSLIC. But buyers got zero FSLIC assistance. The deals were done knowingly; the mark-to-market prior to completing the deal ensured that. The deals were voluntary. The FSLIC had no leverage with which to extort buyers. If the deal was done knowingly and voluntarily, then it was an arm’s-length deal, and that made it the best possible evidence of the true market value of the S&L being purchased. The logic was inescapable: the S&L being acquired must not really be insolvent. It must have enormous goodwill value that accountants could not value directly in the mark-to-market. Indeed, in this example it had to have a minimum value of $45 million because if it had any lesser value, the S&L would be a net liability and it would be irrational to purchase it. Accountants recognized this value by creating a $45 million goodwill asset on the acquirer’s books.
Note how circular and irrefutable this chain of logic is: there is no need (indeed, no way) for the auditor to check whether the S&L being acquired really has any goodwill at all, much less $45 million of it. There is no need because the arm’s-length nature of the deal makes it the best evidence of market value; the auditor has no superior process. It is also impossible for the auditor to check because “general, unidentified intangible” is a fancy way to say “ghost.” The accounting jargon means “we don’t know where to look for it, and even if we did, it wouldn’t matter because it can’t be seen or measured.”
Stepping back from the circular arguments, however, allows one to take the criminology perspective: this is too good to be true. Five hundred S&Ls that are deeply insolvent on a market-value basis aren’t really insolvent on a market-value basis because they all turn out to have enormous amounts of goodwill? Then there is the odd way in which goodwill tracks insolvency. If one purchased the S&L a year later, when the mark-to-market showed it was insolvent by $60 million instead of $45 million, the accountants would put $60 million of goodwill on the books. The more insolvent the S&L being acquired, the greater its goodwill. That was, to say the least, illogical.
There was, in fact, no goodwill at the vast majority of failed S&Ls. Accountants did not consider what the source of the enormous goodwill could be. It couldn’t be deposit insurance or even the broad asset powers granted by states with the greatest degree of deregulation. One could start a new S&L that would be solvent and would have deposit insurance and the same asset powers. Everyone doing the deals knew that the goodwill was fictitious, but it was in their interest to pretend it was real, so they did.
Why did buyers do these deals? Some of the deals were honest. For example, a large S&L would buy a much smaller S&L that was its major competitor for deposits in a metropolitan area. The large S&L would then have the market power to pay less for deposits and charge slightly more for home loans. Or, a very large S&L would buy a smaller S&L that had a good branch network in a part of a state where the large S&L had no presence and wanted to expand. In both cases, there would be real intangible value, but it would be identifiable; in the second example, it was attributable to the branch network.
The bulk of the goodwill mergers, however, were accounting scams. The buyers weren’t irrational; they were taking advantage of an accounting abuse with the encouragement of the Bank Board and the blessing of a Big 8 audit firm. There are two keys to understanding why it was rational to merge despite fictitious goodwill. First the buyer was normally an insolvent S&L. Second, goodwill accounting was so perverse that the more insolvent the S&L acquired, the more “profit” reported.
The owner of an insolvent S&L and the owner of a healthy one had very different incentives when it came to making acquisitions. It was rational for an insolvent S&L to buy, without FSLIC assistance, another insolvent S&L. The insolvent buyer had no downside: limited liability meant that once the S&L was insolvent, the creditors bore any new losses. The owner of the insolvent S&L was no worse off if the merger made the S&L insolvent by an additional $45 million (as in my first example).
Goodwill mergers guaranteed that fraudulent, insolvent buyers won a trifecta even when the goodwill was bogus. First, buying an insolvent S&L was an elegant way for a control fraud to optimize the S&L as a fraud vehicle. Life is full of trade-offs, even for frauds. Control frauds normally have to trade off several factors. The ideal fraud vehicle would be a large company: there is more to steal and the prestige is greater. The larger an S&L’s assets in the early 1980s, however, the greater its insolvency. Control frauds do not want to report that they are insolvent: a regulator can close them down or restrict their operations. A goodwill merger was perfect because it gave one control of a huge S&L and “eliminated” the insolvency of the purchased S&L. Under honest accounting methods, merging with a deeply insolvent S&L without FSLIC assistance should hurt profitability: the acquirer takes on more liabilities than assets, and so it should lose money.
That takes us to the second leg of the trifecta. I was serious about the claim that the more insolvent the S&L acquired, the higher the reported income. Goodwill mergers created fictitious profits in three ways. The principal means was “gains trading.” Remember that the problem in the early 1980s was that S&Ls had lent most of their mortgage money in the 1970s at much lower interest rates and that the fixed-rate mortgages had thirty-year maturities. When interest rates go up, the value of long-term fixed-rate debt instruments (mortgages, bonds, treasury bills) goes down.
The S&L industry had roughly $750 billion in assets during the worst of the interest-rate crisis. Those assets were overwhelmingly long-term (typically thirty-year) fixed-rate mortgages. Fixed-rate assets do not earn higher rates of interest when market interest rates rise. As a result, they can lose a great deal of their market value when rates rise (no one wants to buy a mortgage if it is only earning 10 percent when he can buy a recently issued mortgage and earn 20 percent interest). By mid-1982, the S&L industry had lost about $150 billion in the market value of its mortgages. That works out to a 20 percent loss of total asset values. I will use that percentage loss in my hypothetical examples to provide a realistic explanation of why a “goodwill” merger could produce tremendous, albeit fictitious, profits.
For simplicity, assume the same insolvent S&L example I have been using. We buy an S&L that has $200 million (book value) in mortgages that the S&L lent in 1977 at an 8 percent interest rate. On a market-value basis, however, they are only worth $160 million because the market interest rate for a comparable mortgage is now 16 percent. The key is that we create a new book value when we acquire these mortgages through the merger. Their book value becomes $160 million. The $205 million in liabilities at the S&L we are buying are very short-term deposits. Short-term deposits do not change materially in value when interest rates change, so their book value is unchanged by the merger accounting.
Now assume that interest rates begin to fall after we buy the S&L. One year later the market interest rate for a comparable mortgage is 12 percent. Remember: interest rates and the market values of mortgages go in opposite directions. Interest rates have fallen by 25 percent since the merger, and the mortgages we acquired in the merger have increased in market value to $180 million. We sell the mortgages for $180 million and book a $20 million “gain on sale.”
There were four remarkable things about this “gains trading” scam that made it one of the most perfect frauds of all time. First, one books an enormous profit through a deal that actually locks in an enormous loss. In my example, the acquirer assumes $205 million in liabilities to do this deal. The liabilities are real. The acquirer has just sold every asset acquired in the merger for $180 million. The sale locked in a $25 million loss. The merger has been disastrous, but one reports record profits! And these record profits are derived from GAAP, not creative regulatory accounting principles. Consider the policy implications of this. If one held the mortgages and if interest rates had continued to fall until the market value of the mortgages came back to $200 million, one might have survived. If one sold at $180 million, then it is irrelevant whether rates continue to fall. The other implication is that the acquirer knows that the profit is fictitious and that failure is certain, which maximizes the perverse incentives to engage in reactive control fraud.
Second, the Internal Revenue Service (IRS) treats this transaction for tax purposes as a loss. The IRS says that if one started with assets that had a book value of $200 million and sold them for $180 million, there is a $20 million loss for tax purposes that can be used to offset tax liability on GAAP profits. This is the second way in which goodwill mergers increased net income.6
Third, one could maximize this fictitious income only through a merger. Here’s a simple way to understand the point. Assume the buyer was an S&L with assets and liabilities identical to those of the seller. (That is not a bizarre assumption. Most S&L acquirers were other S&Ls, and virtually every S&L was insolvent on a market-value basis during the peak merger years.) The important thing to understand is that only the seller’s mortgages are marked-to-market by the merger. Again, we assume that market interest rates for comparable mortgages are 16 percent at the time of the merger and fall one year later to 12 percent. One could sell only the acquired mortgages for a gain because only they got a new (lower) book value through the mark-to-market. The buyer’s mortgages have market, but not book, values identical to those of the mortgages acquired through the merger. The book value of the buyer’s mortgages is still $200 million. If we sell them one year after the merger for their market price of $180 million, we have to book a $20 million loss under GAAP.
The S&L league seriously proposed that the entire industry mark its assets to market and create $150 billion in goodwill so that S&Ls could engage in gains trading without finding a merger partner. Even the administration thought this was beyond the pale. One can now see why S&Ls were desperate to acquire other S&Ls.
The fourth bit of elegance comes from an arcane point whose importance I promised to explain. Gains trading would not have been very attractive if one had had to reduce the goodwill asset figure when selling mortgages. A reduction in goodwill would have caused a dollar-for-dollar reduction in capital, and would soon have led to recognition of the GAAP insolvency of the S&L. It seems obvious that selling the assets obtained in the merger must also reduce any goodwill. But here is where the words “general” and “unidentified” proved so useful to the scams. Because the goodwill was not associated with any tangible acquired asset, it was not written off, even if every tangible asset acquired in the merger was sold.
In addition to gains trading and the IRS treatment of the gain as a loss for tax purposes, goodwill mergers created fictitious income through a device so arcane that perhaps one accountant in a thousand knew about it. Here’s the quick and dirty version. Accountants actually created two new accounts when there was a goodwill merger. In addition to goodwill they created an account called “discount.” Why they did this is not important to this discussion.7 For S&Ls in the early 1980s, discount and goodwill were nearly identical in size. A percentage of goodwill had to be recognized every year as an expense. Pratt and the buyers, of course, wanted to minimize expense recognition in order to inflate net income. They found a blunt but effective way to do so. S&Ls had to recognize only 2.5 percent of goodwill a year as expense.8 In my example of a $45 million goodwill figure, that would mean a bit over $1 million a year. S&Ls recognized a portion of discount as income every year. (Note that neither the expense nor the income represents cash flows.) The rate at which discount was recognized as income was significantly greater than the rate at which goodwill was recognized as expense.9 Given my first point, the nearly identical size of goodwill and discount, this meant that the more insolvent the S&L acquired, the more “income” it produced.10 The increased income from discount could be three times the increased expense from goodwill.11
Years later, brilliant lawyers produced an unexpected, fourth source of income from these goodwill scams. In 1989, Congress finally began cleaning up the S&L rot through the Financial Institution Reform, Recovery, and Enforcement Act (FIRREA). One of the abuses it ended was fictitious goodwill. The beneficiaries of that fictitious goodwill hired lawyers who managed to convince the Supreme Court that the Congress was taking something of value from these scam artists and that the Constitution requires the taxpayers to compensate them for the fictitious goodwill. We may have to pay $20 billion to the least deserving claimants in the history of takings litigation!
I have now explained two benefits of mergers to control frauds. Both the ability to control a large S&L without having to recognize its insolvency and the multiple sources of fictitious income are relevant to the third leg of the trifecta. The mergers protected the buyers from regulators and supercharged the S&Ls as control-fraud vehicles.
I explained that most acquirers were themselves insolvent S&Ls. Again, I want to emphasize that some of the goodwill mergers were legitimate acquisitions designed to strengthen branch networks. They should not be tarred with what I am about to explain. The goodwill mergers gave the acquirers de facto immunity from regulatory controls for several related reasons. The most obvious stems from my discussion of the fictitious income the mergers produced. Both the buyer and the S&L it was about to acquire would, in 1979–1982, have reported losing money. The merger would occur, and, miraculously, the combined entity would almost immediately be profitable—extremely profitable. It is very difficult to take supervisory action against a firm that is profitable.
The profit turnaround was “too good to be true,” but Pratt hailed it as proof that his strategy for rescuing failed S&Ls was not only much cheaper than other solutions, but was also transforming the industry by attracting entrepreneurs whose superior management produced profits in awful economic times. Pratt knew better, as shown by his testimony before the National Commission on Financial Institution Reform, Recovery and Enforcement (NCFIRRE 1993c, 12).
Under GAAP, as they were applied in the early 1980s, two institutions with massive negative earnings could merge, and the combined entity could show positive income without the operations of either institution changing (NCFIRRE 1993a, 38).
The acquiring CEOs responded with becoming modesty or bravado, depending on their personality, and raked in the rewards (raises, bonuses, and perks) that their superior skills and the responsibility of running a much larger S&L entitled them to. The values of stock S&Ls (depositors owned many S&Ls in “mutual” form) surged after mergers in response to the “profits.”
The remarkable profits of the goodwill mergers caused nearly everyone to view the CEOs as stars. From the Bank Board perspective, the acquirer was not simply a star manager—he was someone who deserved the agency’s gratitude for resolving a failure at no cost to the FSLIC. The Bank Board wanted to believe that the profits were real. Pratt proclaimed that the profits proved his policies were correct. It would take a very brave or a very stupid examiner to say that the goodwill, income, and “successes” of the mergers were equally fictitious.12
The fact that senior field supervisors often recruited the acquirers in the goodwill mergers was a particular problem. The supervisor would recommend approval of the merger, vouching for the character of the principals. The merger produced the inevitable surge in “profits.” The supervisor would write a glowing letter praising the new management, and, in turn, would receive a bonus and a letter in her file praising the merger. The supervisor would then ask the CEO to come to dog and pony shows where she would try to interest other acquirers. She would introduce the CEO as her prime example of how superior management produced great profits, and she would regale the group with odes to the CEO’s brilliance. The CEO would tell his fellow real estate developers that we had met lots of jerky regulators, but his friend Mary was great.
Regulators are human. We are grateful to those who help us, particularly in times of greatest need. We are sensitive to the criticism that we are too negative; we like to say positive things. We deal constantly with the industry and make friends. We are reluctant to see our friends as crooks, and we know how embarrassing it would be if the CEO we recruited and praised turned out to be a fraud. We are subject to cognitive dissonance.
The combination of these factors meant that Bank Board supervisors were very unlikely to expose the goodwill accounting scams. Two other things compounded this problem. Very few people, even within the Bank Board, understood how the scam worked. I don’t want to overstate this point—many people were skeptical that goodwill was real—but only a handful knew how goodwill and mark-to-market virtually guaranteed substantial fictitious profits if the insolvency of the acquired S&L was large relative to the size of the acquirer. This problem became even worse once Pratt and his senior staff left the agency, for the folks who remained were even less likely to understand.13
The acquirers, however, did not rely solely on these human weaknesses. They had lawyers, and they typically got forbearance provisions that said that the Bank Board could not take supervisory action against them even if they were in violation of the agency’s pathetically weak net-worth requirements if the failure was due to acquiring the failed S&L. This meant that it was difficult to crack down on S&Ls involved in goodwill mergers even if they were highly unprofitable.
The goodwill mergers were central to Pratt’s cover-up. Every merger transmuted net liabilities into fictitious assets. This is accounting alchemy. Every dollar of goodwill made a dollar of insolvency disappear. By the end of 1983, the industry had over $20 billion in goodwill. The goodwill mergers also produced, as I explained, large amounts of fictitious income. This effect was so large that the industry reported a net profit in 1993, but would have reported a net loss if not for the “income” produced by the mergers (NCFIRRE 1993a, 39). The goodwill mergers let Pratt declare victory and leave.
Some books on the debacle have concentrated on lambasting Pratt for his creative RAP. That criticism is misdirected. GAAP created far more consequential accounting abuses than creative RAP. That fact, however, is no defense of Pratt’s creative RAP. The last thing that an industry engaged in rampant GAAP-sanctioned accounting abuses needed was additional abuse. The provisions were indefensible. Control frauds used several of Pratt’s creative RAP provisions, but they were not critical to the success of the frauds. I do not discuss them other than to recall that loan loss deferral damaged Pratt’s purported strategy of waiting for a fall in interest rates.
I have explained that according to economic theory, deregulating an industry that is massively and pervasively insolvent and has deposit insurance guarantees disaster. Pratt designed the Bank Board’s deregulation, and the deregulatory Garn–St Germain Act of 1982. Pratt and Roger Mehle, an assistant secretary of the treasury, drafted the legislation.
Pratt’s deregulation was doomed from the outset because he used the worst possible model of deregulation, Texas, to guide his efforts. He gave Texas an award in recognition of serving as his model. Texas-chartered S&Ls caused by far the worst losses during the debacle. Pratt chose Texas as his model for a logical reason. Texas S&Ls were more deregulated than those in any other state, and they reported smaller losses during 1979–1981 than almost any others. Again, stupid regulation was a major cause of the interest rate crisis, and almost everyone assumed that deregulation was the answer to the crisis. Pratt ignored the fact that dubious accounting drove the superior results in Texas.
Pratt’s first major deregulation was a good one. The interest rate crisis had discredited the members of Congress who had prevented the Bank Board from approving adjustable-rate mortgages (ARMs), so Pratt was able to adopt a rule allowing ARMs.
The second major avenue of deregulation was the removal of controls on how much interest S&Ls and banks could pay depositors (“Reg Q”). This occurred in complex legislation and rulemaking, and the details are not important to this book. This deregulation was harmful—it was critical to making S&Ls the ideal Ponzi scheme—but it was also essential. Unless the nation was prepared to extend interest rate controls to money market funds (which was impossible politically and would have been very bad economics), S&Ls and banks had to be allowed to pay competitive rates of interest, or else depositors would have removed most of their deposits and transferred them to money market funds that were paying three times the interest rate permitted under Reg Q. The administration, not Pratt, led the charge to repeal Reg Q.
Another act that critics often blame incorrectly for the resulting S&L debacle was Congress’s raising the deposit insurance limit from $40,000 to $100,000. In fact, this played no meaningful role in the debacle. The way that Congress increased the limit was wrong, and it showed the extreme clout of the S&L league, but the higher limit had no substantive effect. The thing to understand is that the limit was $40,000 at each S&L and that there were roughly 3,000 S&Ls. Under the lower limit, one could take $120 million to a deposit broker (Merrill Lynch was the largest) and have them deposit $40,000 into each of these 3,000 S&Ls, and every dollar deposited would be fully insured. There are only a handful of entities that might wish to deposit more than $120 million. And I haven’t mentioned banks and credit unions, each with the same insurance limit. Altogether, that made about 20,000 insured depositories, which meant one could deposit about $800 million in insured funds under the old limit. In short, the old limit imposed no meaningful restraint, so the new limit’s only impact was a tiny reduction in transaction costs, because Merrill Lynch’s computers no longer had to divide a $80,000 deposit into two $40,000 deposits to attain full insurance coverage.
Other than approving ARMs, every act of deregulation that Pratt undertook contributed to the debacle. Pratt was a whirlwind who deregulated a broad range of activities. I discuss the seven areas most responsible for producing the wave of control frauds. The first two related directly to what CEOs could do. First, Pratt made it possible for one man to own an S&L. Pratt eliminated the Bank Board rules that prevented an individual from owning more than 15 percent of the stock and that required there be at least 400 shareholders. Second, Pratt relaxed conflict-of-interest rules that restricted officers and directors from using their positions for personal gain.
Pratt’s deregulation spurred massive growth in several ways. The third key facet of deregulation was Pratt’s further weakening of the net-worth requirement. He did so in two ways. The obvious change was reducing the net-worth requirement to 3 percent of total liabilities. That is a ludicrously low level of capital. A typical manufacturing company in the 1970s would have had fifteen times as much capital. An S&L with 3 percent capital is a few bad loans away from failure. The less obvious change came from all the income and assets attributable to abusive GAAP and creative RAP. An S&L that was deeply insolvent could report under both GAAP and RAP that it exceeded its required net worth. Pratt’s fourth key measure spurred growth by ending restrictions on deposit brokers.
We need to view these deregulatory steps in conjunction with prior acts of deregulation that Pratt inherited but did not change. He inherited two insane rules that encouraged massive growth. One was five-year averaging. An S&L’s capital requirement could be far less than the nominal requirement because it could, for example, meet the 3 percent requirement by showing that its capital represented 3 percent of its average liabilities over the last five years. The next footnote provides an example that illustrates the point; but the perverse bottom line was that the faster you grew, the lower your effective percentage capital requirement.14 By lowering the nominal requirement to 3 percent and continuing five-year averaging, Pratt knew that he was essentially eliminating the capital requirement, and that came on top of the pervasive accounting abuses.
The combination of very low nominal capital requirements and five-year averaging meant that the fastest growing control frauds could grow by roughly $1 billion for every additional $1 million of “capital” they could report. This thousand-to-one leverage opportunity was one of keys to the astonishing growth that made the control frauds ideal vehicles for Ponzi schemes and imposed horrific damages on taxpayers. Every $1 million in fraudulent accounting income could put the taxpayers at risk by an additional $1 billion.
I will only mention the other similar act of capital depravity that Pratt inherited and then exploited instead of ending. “Twenty-year phase-in” provided that a newly created (de novo) S&L did not have to meet the nominal capital requirement. It only had to meet one-twentieth of the requirement by the end of year one, one-tenth of it by the end of year two, etc. De novos had no real capital requirement in their early years of operation. Five-year averaging and twenty-year phase-in were indefensible and disastrous. Pratt did not create them, but instead of ending them, he exploited them and made them worse by lowering the nominal capital requirement and encouraging fictitious accounting income.
The final three key acts of deregulation made construction lending the ideal Ponzi scheme. Pratt’s fifth step was allowing federally chartered S&Ls to place a much higher portion of their assets (40 percent) in construction lending. Sixth, Pratt weakened the “loan-to-one-borrower” (LTOB) requirement to permit S&Ls to loan 100 percent of their RAP net worth to a single borrower. Pratt had inflated RAP, even beyond the enormous inflation provided by GAAP, so an S&L loaning at its LTOB limit was loaning far more than the total of its real capital to a single borrower. That meant that the S&L was one bad loan away from insolvency. (Of course, this discussion assumes the S&L had real, positive net worth to begin with, which it usually did not in 1982.) Seventh, Pratt ended the requirement for a down payment on secured real estate loans. S&Ls could loan 100 percent of the collateral’s appraised value. If the loan defaulted (and default rates on speculative commercial real estate construction are high), the S&L was sure to suffer losses, even if the appraiser only slightly overvalued the project.
Pratt’s desupervision of the industry compounded the disaster his deregulation caused. Desupervision helped make the industry ideal for control fraud. First, and most disastrously, Pratt froze and then reduced the number of examiners. This was a terrible mistake, but Pratt was not alone in making it. President Reagan’s first act was to freeze new hires. The Office of Management and Budget (OMB) wanted the Bank Board to reduce its examiners and supervisors. President Reagan appointed Vice President Bush to head his financial deregulation task force. Bush recommended that financial regulators rely more on computer analyses of industry financial statements and cut both the frequency of examinations and the number of examiners. Martin Lowy (1991, 36) says that Pratt fought with the administration for new examiners and was denied them.
The OMB went so far as to threaten Pratt with criminal sanctions if he didn’t obey its spending restrictions. On another occasion, the OMB cut off FSLIC funds for liquidations (ibid.).
Deregulation requires increased supervision in an industry with deposit insurance. An NCFIRRE consultant interviewed Paul Allen Schott, an assistant general counsel for the Treasury Department in 1981–1985. Treasury and the OMB worked together throughout that time to try to prevent the Bank Board from hiring additional examiners. Schott explained:
When OMB [blocked] efforts to hire more examiners, the thought was deregulation meant not supervising the institutions. There was a misguided perception that supervision wasn’t needed. (NCFIRRE 1993b)
Examinations fell sharply from 1981 to 1983 as a result of the combination of deregulation, the cut in the number of examiners, the loss of experienced examiners, and the wave of control frauds. Deregulation meant examiners had to review much more complex assets, and control frauds meant that they could not rely on the S&L’s records or personnel to be truthful. Both conditions vastly increased the required examination time and examiners’ required expertise. Experienced examiners are far more efficient. The Bank Board conducted 3,171 exams in 1991, 2,800 in 1982, and 2,131 in 1983. At a time when greatly increasing the number of examinations was vital, desupervision led to one-third fewer exams.
Pratt inherited an undermanned staff paid so poorly relative to its counterparts in the banking agencies that examiner quality was poor. Pratt knew how desperately inadequate Bank Board examination, supervision, and enforcement were. He could have used the Garn–St Germain Act and his powers as chairman to fix these problems. Specifically, the Bank Board needed the statutory authority to pay competitive wages, and it needed a chairman who would dramatically improve Bank Board examination and supervision. The Bank Board also needed to make the organization of its examiners and supervisors rational and to adopt modern examination techniques; as I’ve written, its examiners were still writing out reports in pencil in 1986. Pratt, however, produced no meaningful improvements. He missed his critical opportunity against the control frauds. The administration actively made things worse.
The administration took extraordinary steps to prevent Pratt from closing insolvent S&Ls. The first act was the congressional testimony of Roger Mehle, the assistant secretary that Treasury Secretary Donald Regan chose to take the lead in the S&L crisis. Mehle’s testimony started with a fact: GAAP financial statements did not reflect the true market value of S&Ls. That was bad news, for their market values were deeply negative, but Mehle tried to make it good news by arguing that because GAAP was irrelevant, insolvency was irrelevant. He said the industry was “sound” (Black 1993a, 20). Market-value insolvency, however, is critical. If an S&L’s liabilities exceed its assets, the insurance fund suffers a loss; and if the liabilities of the insurance fund exceed its assets, the taxpayers bear the loss.
Mehle had an answer to that: we bear the loss only if we close the S&L, so we should not close failed S&Ls. Mehle argued that an S&L needed to be closed only if it could not raise the cash to pay its current obligations. He noted that deposit insurance meant that this should never occur because the insolvent S&L could grow and use new deposits to pay interest on old deposits. The administration encouraged insolvent S&Ls to engage in Ponzi schemes. Mehle was not on a frolic of his own; he testified for the administration.
Mehle’s second act was even more amazing. While still the senior Treasury Department official with responsibility for S&Ls, he testified on behalf of an S&L suing the Bank Board. Pratt’s predecessor placed Telegraph Savings in receivership because it was insolvent. The owners filed suit challenging the receivership. Had the plaintiffs won, they doubtless would have sought damages for an unlawful takeover—from the Treasury Department! It was, therefore, bizarre that Treasury permitted Mehle to testify on the owners’ behalf against the agency. Mehle was a dream witness for the plaintiffs, testifying that it was arbitrary for the agency to close an S&L because it was insolvent. He said that the S&L industry had to be healthy because it was growing rapidly. He even testified that it was irrelevant whether an S&L was already insolvent and had growing losses because “It can solve it by borrowing” or bringing in additional deposits (Black 1993a, 20). (Bankrupt companies solve their problems when they borrow more and go more heavily into debt?) After he left the administration, Mehle helped form a “shadow financial regulatory committee” that gave similarly expert advice on how to regulate.
The courts found Mehle unpersuasive, but the administration succeeded in its real goal: ensuring that Pratt would not close enough S&Ls to reveal the cover-up. The numbers tell the story: Pratt spent just enough FSLIC money in 1981 and 1982 to report that FSLIC reserves increased very slightly to $6.15 billion and then to $6.3 billion (Kane 1989, 9). The FSLIC fund, of course, was massively insolvent on a market-value basis because the industry was massively insolvent. Government accounting standards required the FSLIC to recognize that liability and to report that it was insolvent. Needless to say, Pratt (like his predecessors) did not do so. Pratt should have reduced the FSLIC fund by resolving failed S&Ls. Pratt, however, wanted the FSLIC fund to grow in order to aid the cover-up, to project an image of strength that would reduce the risk of a run, and to provide emergency liquidity if there were a run.
Third, I have explained how Pratt welcomed and praised the entrepreneurs and how this attitude plus the fictitious profits produced by the goodwill mergers made it very unlikely that the agency would take timely action against the control frauds. And, in fact, the agency did not take effective action against any control fraud during Pratt’s tenure. Indeed, the agency took few enforcement actions. The goodwill mergers and the wave of new entrants that Pratt encouraged diverted critical supervisory resources into (non)resolutions at precisely the time they were desperately needed to counter the wave of control frauds.
Another term for “competition in laxity” was “the race to the bottom.” S&Ls could change freely from a federal to a state charter (the permission from the government to run an S&L) and still be insured by the FSLIC. The charter determined what the S&L could invest in. Texas led the race by deregulating in the 1970s, and California followed the lead. Many federally chartered S&Ls in those states converted to state charters. By granting federally chartered S&Ls greater investment powers, the Garn–St Germain Act and Pratt’s deregulation led many Texas and California charters to convert quickly to federal charters. Texas had the equivalent of a “most favored nation” clause in its charters that allowed Texas-chartered S&Ls to do whatever federally chartered ones could, so the rush to convert to federal charters was greatest in California. California responded to the Garn–St Germain Act with the Nolan Act (named after its sponsoring senator, the notably corrupt and soon-to-be-convicted Pat Nolan). The Nolan Act became effective on January 1, 1983. It won the race to the bottom by going directly to the bottom. A California-chartered S&L could invest 100 percent of its assets in anything (with the commissioner’s approval).
Despite Nolan’s corruption, this was not a conspiracy, but a bungled mess of epic proportions. No one was clever enough to design this disaster. The conversion of large numbers of California-chartered S&Ls into federal ones caused the state legislature and the industry to push for immediate adoption of the Nolan Act.15
A similar dynamic occurred in Texas. The result was that scores of federally chartered S&Ls converted to California and Texas charters. The second wave was the rush for de novo Texas and California charters—particularly California. The theory was that the state was supposed to be the primary regulator of state-chartered S&Ls. But neither Texas nor California had enough examiners and supervisors to handle their existing charters; the loss of assessment income gutted the ranks of both types of regulators. The surge of charter conversions overwhelmed them. It far outpaced the rate at which they could hire and train additional staff. In these circumstances, the last thing a rational state commissioner would have done is to encourage a wave of de novos: she would have had to turn her already hopelessly inadequate staff into application reviewers. The industry was suffering from mass insolvency and was undergoing unprecedented deregulation; it needed vastly more examination and supervisory resources to avoid catastrophe.
Naturally, Texas and California did the worst possible thing, at the worst possible time, in the worst possible manner. Their commissioners encouraged hundreds of de novo applicants for state charters, and they directed their overwhelmed staffs to expedite approval of the applications with little or no review for quality. They approved the applications of hundreds of real estate developers who were insolvent, had poor track records, and had severe conflicts of interest for de novo charters. Larry Taggart, commissioner of the California Department of Savings and Loans (CDSL), did not deny any de novo charters.
The second worst control fraud, Don Dixon’s Vernon Savings & Loan (aka “Vermin”), provided prostitutes to the Texas commissioner, Lin Bowman.16 Taggart was worse. He went on to work for the three most notorious control frauds and tried to get the administration to fire Gray. His judgment was so poor that he put the following in writing in his August 4, 1986, letter to Donald Regan, now the president’s chief of staff (and the man leading the administration’s effort to force Gray to resign):
[T]hese actions being done to the industry by the current chief regulator of the Federal Home Loan Bank Board are likely to have a very adverse impact on the ability of our Party to raise needed campaign funds in the upcoming elections. Many who have been very supportive of the Administration are involved with savings and loans associations which are either being closed by the FHLBB, or threatened with closure. (U.S. House Banking Committee, 1989, 3:630)
The wave of opportunistic control frauds entered the industry overwhelmingly through Texas and California charters in 1981–1984. The state commissioners were not simply ineffective; they were often the frauds’ allies. In the states with the greatest deregulation and the greatest need for supervision, examinations became rare and supervision farcical.
In addition to cutting the marginal rate of taxation, the 1981 Tax Act created abusive tax shelters and encouraged investments in real estate that were driven by tax, rather than normal economic, considerations. The inevitable result was a bubble in real estate values, particularly in commercial real estate. The bubble helped S&L control frauds claim record profits and increased the ultimate losses to the taxpayers once the bubble burst. A recurrent myth is that the 1986 Tax Reform Act, which removed most of the abusive provisions of the 1981 act, caused real estate recessions and greater losses to the FSLIC. As the investigating commission explained in its report on the debacle:
Many observers blamed the 1986 Tax Act for the woes that befell the S&L industry, but it was the 1981 Act that created an unsustainable boom, and encouraged “over-building.” The 1986 law hastened the collapse in the Southwest, much of whose expansion had been based on expectations of continued inflation in property values. But had the 1986 act not been passed, over-building would have been even greater, and the eventual collapse in real-estate values deeper. (NCFIRRE 1993a, 55)
To an opportunistic control fraud, obtaining control more cheaply is the bronze standard, doing so for free wins a silver medal, and getting paid to take control takes the gold. Opportunistic control frauds lived up to the name I gave them: they were champions at getting something for nothing. I’ll explain the most common scams. (The scams were not necessarily mutually exclusive; some opportunists combined them.) The most common fraud mechanism was to have Hermann K. Beebe fund the purchase of an S&L. Beebe was a control fraud, a convicted felon running a Louisiana insurance company. He was an associate of the New Orleans mob. Beebe helped scores of control frauds acquire S&Ls and banks in the Southwest (Mayer 1990, 226). Beebe would loan the money to buy the S&L. The buyer would, in turn, cause his S&L to make far larger loans to Beebe’s straws. The straws, of course, would not repay the loans. Beebe won, the S&L owner won, and the taxpayers lost.
Similarly, Michael Milken, the “junk bond king” of Drexel Burnham Lambert, financed Charles Keating’s purchase of Lincoln Savings and David Paul’s purchase of CenTrust Savings (Black 1993c). Keating did not have to spend a penny of his own money to buy the S&L.
Another scam was to have the existing shareholders of the S&L direct it to fund the acquisition. I’ve explained that S&Ls were overwhelmingly insolvent in the early 1980s and that shareholders should receive nothing when the company is insolvent. They, however, do not want to receive nothing, and if they control the S&L (and have no ethical constraints), they may decide to have the S&L make a very large loan to a potential acquirer. The acquirer uses the loan proceeds to buy their stock (at a substantial premium above its true, i.e., nonexistent, market value). This gives the acquirer control of the S&L without having to use any of his own cash.
Pratt changed Bank Board rules to allow acquirers to contribute capital to the S&L in the form of real estate instead of cash. The acquirer had to present a real estate appraisal supporting the claimed value. I am confident that the reader can predict what happened, even though Pratt, the expert, did not: this change proved disastrous. In every case that I studied, the value assigned to the real estate was grossly inflated. David Paul, who, like Keating, was a Drexel “captive,” used this scam. Contributed capital was overwhelmingly fictitious. Again, the key is that every $1 million of fictitious capital allowed the control fraud to grow by up to $1 billion and increased the risk to the taxpayers by the same amount. The acquirers who won gold medals for fraud used an elegant variant of this scam. They would say that they wished to contribute $5 million in capital to “their” S&L. Unfortunately, the only way they could do so was to contribute a large real estate parcel. The parcel had an appraised value of $20 million. They would contribute the parcel to their S&L, and the S&L would pay them $15 million in cash. The parcel might, in fact, be worth as little as $2 million.
Outside professionals are like British journalists. There’s no need to bribe them. More precisely, you bribe them in a way such that no one considers it a bribe.
Here’s how it works with appraisers. The appraisal fee is larger for commercial real estate than for residential, and it is greater in absolute size for more expensive than for less expensive properties. As a result, both appraisers and S&L control frauds gained if the S&L did more commercial real estate lending on more expensive projects. The S&L loan officer calls the appraiser and asks him for a favor. The loan officer has to make a recommendation on a proposed $60 million loan in two weeks. Could the appraiser please give him a preliminary, oral estimate of value as soon as possible, before completing the written appraisal report? (Note that the loan officer has communicated the size of the loan to the appraiser.) The S&L can’t make the loan without violating Bank Board rules if the appraiser does not come back with a value of at least $60 million. The appraiser calls with preliminary estimate of value. If it is at least $60 million, the loan officer tells him to finalize the written appraisal, pays his full fee, and uses him in the future. If he comes back under $60 million the loan officer thanks him effusively, says that there is no need to complete the written appraisal given the inadequate value of the property, pays a reduced fee, and the S&L never uses the appraiser again. Functionally paying an appraiser a fee to value property (that is, say, really worth $35 million) at over $60 million is equivalent to a bribe. But it is a “perfect crime,” impossible to prosecute. Control frauds know that they only need a tiny group of appraisers to inflate property values; there is no need to suborn the entire profession. The thing that most people don’t understand is that this whole process can (and typically was) done in a way that a transcript of the conversation could appear on the front page of the local newspaper without embarrassing the appraiser or the loan officer.
Shopping for an accommodating auditor involves a similar process. The dynamics differ because control frauds virtually always use top-tier audit firms, whereas appraisers are often sole proprietors. Institutions matter a great deal, and control frauds are adept at finding the weak link in any institutional chain.
The law-and-economics movement claims that top-tier audit firms will never aid a control fraud because the financial loss resulting from a diminished reputation is so much greater than any possible gain from assisting the fraud (Easterbrook and Fischel 1991, 282). Judge Easter-brook has gone to the extreme of preventing a plaintiff from presenting a case charging an auditor with aiding a fraud, stating that it would be irrational for an audit firm to act as alleged, given its reputational interest (Prentice 2000, 136). As I noted earlier, whenever economists assume rationality but ignore fraud, the predictions are likely to fail. Every S&L control fraud got a clean opinion from a top-tier audit firm—usually they got several years worth of clean opinions—even though they were deeply insolvent and engaged in pervasive accounting fraud. All of the current control frauds got clean opinions from top-tier audit firms, even when they were deeply insolvent and engaged in massive accounting fraud. Control frauds are routinely able to find auditors to assist their frauds. Indeed, a prestigious audit firm is a control fraud’s most valuable ally.
The NCFIRRE report (1993a, 76) provides the bottom line:
The result was a sort of “Gresham’s law” in which the bad [accounting] professionals forced out the good.