The full text on this page is automatically extracted from the file linked above and may contain errors and inconsistencies.
Ft E July/August 1994 Vol. 76, No. 4 II Exp lan ation s for the In creased R iskiness of Banks in the 19 8 0 s Trade Betw een the United States and Eastern Europe The New Stru ctu re o f the Housing Fin an ce System The Inflation Tax and the M arginal W elfare Cost in a W orld o f C urren cy and Deposits THE FEDERAL A RESERVE BANK of ST. LOUIS R E V I E W President T h o m a s C. M e lz e r Review is published six times per year by the Research Department of the Federal Reserve Bank of St. Louis. Single-copy subscriptions are available free of charge. Send requests for subscriptions, back issues or address Director of Research W illia m G. D e w a ld Associate Director of Research C le tu s C. C o u gh lin Research Coordinator and Review Editor W illia m T. G avin changes to: Federal Resen/e Bank of St. Louis Public Information Department P.O. Box 442 St. Louis, Missouri 63166 The views expressed are those of the individual authors and do not neces sarily reflect official positions of the Federal Reserve Bank of St. Louis, the Federal Reserve System, or the Board of Governors. Articles may be reprinted Banking R. A lto n G ilb e rt D avid C. W h e e lo c k International C h ris to p h e r J . N e e ly M ic h a e l R. P ak ko P a tric ia S. P o llard or reproduced only if the source is credited. Please provide the Public Information Department with a copy of the reprinted materials. All nonproprietary and nonconfidential data and programs for the articles published in Review are available to our readers. This information can be Macroeconomics R ich ard G. A n d erson J a m e s B. B u llard M ic h a e l J . D u e k e r J o se p h A. R itte r Jo h n A. T ato m D a n ie l L. T h o rn to n P e te r Yoo obtained from three sources: 1. FRED (Federal Reserve Economic D ata), an electron ic bul letin board service. You can access FRED by dialing 314-621-1824 through a modem-equipped personal computer. Parameters should be set to: no parity, word length = 8 bits, 1 stop bit, and the fastest baud rate the modem supports (up to 14,400 bps). Information will be in directory 11 under the file name ST. LOUIS REVIEW DATA. For a free brochure Regional M ic h e lle A. C la rk K e vin L. K lie s e n A d am M . Z a re ts k y on FRED, please call 314-444-8809. 2. The Federal Reserve Bank of St. Louis. You can request data and programs on either disk or hard copy by writing to: Research Director of Editorial Services D a n ie l P. B ren nan Managing Editor C h a rle s B. H e n d erso n Graphic Designer B ria n D. E b e rt Department, Federal Reserve Bank of St. Louis, Post Office Box 442, St. Louis, MO 63166. Please include the author, title, issue date and page numbers with your request. 3. Intra-university Consortium for P o litical and Social Research (ICPSR). Member institutions can request data through the CDNet Order facility. Nonmembers should write to: ICPSR, Institute for Social Research, P.O. Box 1248, Ann Arbor, Michigan 48106, or call 313-763-5010. 1 Federal Reserve Bank of St. Louis Review July/August 1994 In This Issue... Explanations for the Increased Riskiness of Banks in the 1980s Sangkyun Park In the 1980s, the number of bank failures increased sharply, and banks in general experienced increasing problem loans and dwindling capital. The main explanations for the deterioration of bank asset quality include increased incentives for risk-taking by bank stockholders, desperate risktaking by bank managers to increase profits, and unexpected economic shocks. Sangkyun Park reviews the logic of the three explanations and examines their empirical significance. He finds that deliberate risk-taking by both stockholders and managers was consistent with the behavior of a sizable proportion, though not the majority, of banks. He also concludes that economic shocks were significant, but do not negate the effects of deliberate risk-taking. 25 Trade Between the United States and Eastern Europe Patricia S. Pollard Most of the trade between the United States and Eastern Europe since the end of World War II has been very small. W hile the United States has maintained high tariffs on imports from most Eastern European countries— and restricted its own exports to them as well, particularly high technology—Eastern Europe has maintained trade restrictions on imports from the United States. W ith the collapse of the Soviet system, however, Eastern Europe has begun to re-direct trade to the West as it initiates both political and economic reforms. Patricia S. Pollard describes the recent changes in trade flows between the United States and the three Eastern European countries that have made the greatest progress in adopting market reforms: the Czech and Slovak Federal Republic (CSFR), Hungary and Poland. She concludes that increased trade between the East and the West benefits not only the United States’ economy, but is also directly linked to Eastern Europe’s efforts to establish and maintain political stability. 47 The New Structure of the Housing Finance System John C. W eicher The Financial Institutions Reform, Recovery and Enforcement Act of 1989 (FIRREA) marked both the culmination of a 20-year period of changes in the U.S. housing finance system and the beginning of a period of legislative reforms intended to forestall any recurrence of the savings-and-loan industry debacle that forced the enactment of FIRREA. JULY/AUGUST 1994 2 John C. Weicher describes the process of change in the housing finance system and its implications for the ability of the system to fulfill the pur poses for which it was established. The new system is dominated by two large national institutions, chartered by the federal government and serving the secondary mortgage market, instead of thousands of small, local lenders who both make mortgage loans and hold them in portfolios. Legislation of the last five years has imposed higher capital requirements on nearly every institution involved with housing finance and changed the regulatory structure for every private institution in the system. In addition, these changes have created potential conflicts with the stated public objectives of the system: access to home mortgage funds for areas and groups that are “underserved” and support for the mortgage market during economic downturns. 67 The Inflation Tax and the Marginal Welfare Cost in a World of Currency and Deposits Alvin L. Marty Since inflation is the rate at which the purchasing power of money declines, it is a tax on real money balances. Alvin L. Marty provides an analysis that determines the optimal rate of inflation as an efficient tax rate. He explains the government’s revenue and the total cost associated with inflation and how each is affected by the level of the inflation rate. An efficient tax system minimizes the total cost of collecting a given tax revenue. For inflation to be chosen as a component of an efficient tax system, however, the inflation rate must be set so that the marginal cost per dollar of government revenue from inflation is equal to the marginal cost per dollar of other sources of revenue. Marty shows how the analysis of such an optimal rate of inflation is altered when there are two components of money—currency and deposits— one produced directly by the government, or its central bank, and the other provided by a competitive banking system. FEDERAL RESERVE BANK OF ST. LOUIS 3 Sangkyun Park Sangkyun Park is a senior economist at the Federal Reserve Bank of St. Louis. Jonathan Ahlbrecht provided research assistance. Explanations fo r the Increased Riskiness of Banks in the 1980s X N T E R E S T IN BANKING MATTERS surged in the 1980s, when the U.S. banking system experienced considerable difficulties after several decades of stability. During the decade, the number of bank failures increased sharply, and banks in general experienced increasing problem loans and dwindling capital. Most banks recov ered their financial strength in the early 1990s thanks to improved economic conditions and low short-term interest rates that increased interest margins. With the recovery of the banking sector, public interest in bank failures is fading, but many questions remain unanswered. To effectively prevent repetition of the banking turmoil, it is important to study the fundamental causes of the sudden deterioration of the financial health of the banking system in the 1980s. Without recognizing the causes, future banking policies intended to improve the safety and soundness of the banking sector might produce more unintended effects than intended ones. Numerous studies have proposed explanations for the deterioration of bank asset quality, but empirical evidence is sketchy. This study explores theoretical explanations for the financial problems of commercial banks in the 1980s and examines their empirical consistency. I focus on commer cial banks because many previous studies have examined the financial problems of savings and loan associations (S&Ls). Three theoretical possibilities for the increased riskiness of banks in the 1980s are considered: (1) increased incen tives for risk-taking by bank stockholders; (2) desperate attempts of bank managers to increase profits by assuming additional risk; and (3) unexpected economic shocks. To evaluate the empirical significance of the three hypotheses, the empirical section examines the effects of capital adequacy and earnings on the risk-taking behavior of banks and also looks at the relationship between regional economic conditions and bank performance. Capital ratios and earnings may be related to the risk-taking incentives of stockholders and managers, respec tively. Regional economic conditions should largely explain bank performance if unexpected economic shocks are the main reason for the deterioration of bank asset quality. For selected years of the 1980s, I divide banks into several groups based on year-end capital ratios and earnings on assets. I then compare year-to-year changes in various risk measures across groups. Although deliberate risk-taking by stockholders and managers does not appear to apply to the majority of banks, it is found to be consistent with the behavior of a sizable proportion of banks. This result holds even after controlling for the effects of local economic conditions. The next section describes the economic and institutional developments that are related to JULY/AUGUST 1994 4 the financial troubles in banking during the 1980s. The following section explores theoretical expla nations for increased risk-taking and existing empirical evidence. Empirical results of this study follow. Lastly, the article’s findings are summarized. DEVELOPM ENTS IN BANKING The United States enjoyed stable banking during the four decades following the establishment of the Federal Deposit Insurance Corporation (FDIC) in 1934. The relatively high number of bank failures between 1934 and 1942, about 44 per year, may be regarded as an aftermath of the financial crisis of 1933 and the following years of depressed economic activity.1 Between 1943 and 1974, only 121 banks failed. Bank failures began increasing in the second half of the 1970s and became notable in the second half of the 1980s. Between 1985 and 1990, the number of bank failures averaged 169 per year. To understand the dramatic increase in bank failures, we need to look at the economic and institutional developments that are relevant to the banking business. Econom ic Developments The banking industry encountered numerous unfavorable shocks in the 1970s and 1980s. These included sudden increases in interest rates, the collapse of many real estate investment trusts (REITs), the Latin American debt crises and sharp declines in real estate values. High inflation in the 1970s and the early 1980s raised interest rates. Unexpected increases in interest rates usually lower bank interest margins because the average maturity of bank liabilities is generally shorter than that of their assets. The mismatch of matu rities means that lending rates adjust more slowly than funding rates. Increases in interest rates, hence, tend to lower bank profits. REITs, which channel investors’ money into the real estate, mortgage and construction markets, grew rapidly in the early 1970s. Banks were the major supplier of funds during the rapid expan sion and often served as REIT advisors, whose functions included proposing investment projects and overseeing daily operations.2 A slump in the construction and real estate industries in the 1 The bank failure numbers do not include uninsured banks. See FDIC (1991). 2 See Sinkey (1979) for a detailed discussion of REITs. 3 See Cline (1984) for a detailed discussion of sovereign debt crises. FEDERAL RESERVE BANK OF ST. LOUIS m id-1970s decreased the asset value of many REITs and the banks that extended credit to them. After the oil shock of 1974, oil-exporting coun tries enjoyed large trade surpluses. U.S. banks took an active role in channeling the surpluses to developing countries. As a result, loans by U.S. banks to developing nations increased rapidly during the 1970s to exceed $100 billion in the early 1980s, when sovereign debt problems emerged.3 Large developing-country debtors, including Brazil, Mexico and Argentina, failed to meet their debt obligations as they were strained by worldwide recession, high interest rates and the second oil shock in 1979. Their debt-servicing difficulties lowered the value of the loans and, hence, the value of lending banks’ capital. A real estate boom in the 1980s resulted in sharp increases in property prices and over building of commercial properties in many parts of the United States. Banks financed the boom by expanding loans rapidly. Toward the end of the decade, the value of commercial properties plunged as vacancy rates surged. The price of residential properties also dropped in some regions, the Northeast and California in particular, where income growth substantially lagged behind increases in housing prices. Declining real estate prices placed many banks in financial trouble by increasing delinquency rates and lowering collateral values. Institutional Developments The Banking Act of 1933 transformed a rela tively competitive banking system into a highly protected one. Before its enactment, the main barrier to competition was the prohibition of interstate branching. The Banking Act of 1933, however, insulated the banking business by pre venting banks from engaging in other businesses, including the securities business, and vice versa. It also prohibited payment of interest on demand deposits and authorized the Federal Reserve Board to limit interest rates on time and savings deposits at member banks. Most notably, the act created the FDIC. The above measures relieved banks from com petitive pressure. The continued prohibition of interstate branching preserved some geographic 5 monopoly power. The separation of banking from other industries bolstered monopoly power by deterring other businesses from making inroads into banking. With interest rate ceilings, banks were unable to bid up interest rates to attract deposits. Government-backed deposit insurance made it unnecessary for banks to prove their soundness to depositors. Protected from compe tition, banks enjoyed relatively stable market shares and profits. Legislative and institutional developments in the 1970s and 1980s relaxed regulation and permitted greater competition. Many states relaxed their restrictions on bank ownership by out-of-state holding companies. This develop ment lowered geographic entry barriers, despite the prohibition of interstate branching.4 The Depository Institutions Deregulation and Monetary Control Act of 1980 (DIDMCA), which phased out interest ceilings and permitted all depository institutions to offer NOW accounts, allowed greater competition for deposits among banks. This act, in combination with the GarnSt. Germain Depository Institutions Act of 1982, also increased competition between banks and S&Ls by expanding the realm of the S&L business. Institutional developments also added compet itive pressure on banks. The 1980s witnessed rapid growth of financial products that could substitute for banking products.5 For example, many large corporations with established credit records found it attractive to borrow directly by issuing commercial paper. Money market mutual funds began offering convenient features of bank deposits such as checking privileges, and purchases and redemptions without fees. The emergence of competing products offered by nonbank institu tions eroded the profitability of banks. In sum, the banking industry experienced unfavorable economic shocks and increased competition in the 1970s and 1980s. Slumps in real estate markets and sovereign debt crises impaired the financial health of many banks. Legislative changes lowered geographic entry barriers and restored the mechanism of price competition in the banking sector. In addition, increasing sophistication of financial markets enabled nonbank institutions to circumvent industry barriers. TH EO RETICAL EXPLANATIONS AND EXISTIN G EVIDENCE Numerous studies have proposed explanations for the deterioration of bank asset quality during the 1980s, which can be broadly classified into three groups. First, “moral hazard” explanations argue that the changed economic and institutional environment in the 1980s increased the incen tives of bank stockholders to take risk. A second explanation is increased risk-taking incentives of managers, rather than stockholders. A third possibility is that the quality of bank assets deteriorated mainly because of unexpected external events rather than deliberate risk-taking. The three explanations are generally consis tent with developments in the banking sector. Convincing conclusions require detailed empirical examination. Unfortunately, the difficulty of obtaining adequate risk measures has discouraged empirical examination. Furthermore, it is difficult to determine causality because the hypotheses are interrelated with one another. For example, a negative economic event that decreases bank capital may increase the incentive to take risk, and risk-taking may make banks more vulnerable to economic shocks. Moral Hazard Stockholders im plicitly hold a put option, the right to sell their stocks at a prespecified price. With limited liability, stockholders of a corporation can walk away without incurring further losses when the net worth of the corpora tion falls below zero. Escaping from a firm with a negative net worth is eco n o m ica lly equivalent to selling the firm for the price of zero. Stock holders can thus increase their expected wealth at the expense of debtholders by increasing the variance of the return from assets, that is, by taking more risk. With a larger variance, it is more likely that the return from assets will turn out to be very high. A larger variance also increases the possibility of an extremely low return and, hence, a substantially negative net worth. Limited liability, however, protects stockholders 4 According to unpublished data compiled by the Board of Governors, bank assets held by out-of-state bank holding companies totaled about $470 billion as of June 30,1990, which accounted for about 16 percent of total bank assets. 5 Wheelock (1993) discusses the increasing competition faced by banks both in lending and funding markets. JULY/AUGUST 1994 6 from incurring additional losses once net worth has fallen to zero. In other words, while improved upward potential of asset portfolios clearly ben efits stockholders, downside risk mainly harms debtholders. Thus, limited liability gives stock holders an incentive to take higher risk than they otherwise would. In general, market forces prevent stockholders from taking advantage of the put option value. Debtholders demand interest rates high enough to compensate for higher default risk when a corporation holds a risky portfolio. In addition, bond covenants and needs to refinance short-term debt restrict the portfolio selection of corporations. Thus, it is difficult for stockholders to exploit debtholders because the cost of debt increases with the riskiness of a corporation. The effec tiveness of this market mechanism is limited in the banking sector, however, because of government-backed deposit insurance. In the 1980s, most deposits were insured either explicitly or im plicitly.6 Thus, most depositors, who were the major debtholders of banks, were indifferent about the riskiness of individual banks and, hence, did not demand higher interest rates to riskier banks. Furthermore, the FDIC charged a fixed-rate insurance premium until the end of 1992. Therefore, banks enjoyed risk-insensitive funding costs and had greater incentives to take risk than they would have if deposits were not insured.7 Because of the risk-taking incentives created by deposit insurance, banking authorities need to lim it the portfolio selection of banks.8 In other words, government action must substi tute for market forces to prevent banks from taking excessive risk. Proponents of the moral hazard view argue that banks had increased incentives to take risk in the 1980s for two main reasons: losses that impaired capital and reduced charter values due to greater competition.9 Poorly capitalized banks have a greater incentive to take risk. With smaller capital, 6 Studies on market discipline find that the presence of large deposits exceeding the FDIC insurance limit of $100,000 was not an effective source of discipline on banks. Gilbert (1990), who surveys the market discipline literature, points out that uninsured depositors and holders of subordinated debt rarely absorbed losses of failed banks. In most cases, failed banks were merged with other banks, and the acquir ers assumed all liabilities of the failed banks. 7 Merton (1977) makes theoretical arguments based on the option-pricing model. 8 Calomiris (1989) compares deposit insurance systems in U.S. history and concludes that a necessary condition for the success of deposit insurance systems is effective enforcements of regulations. Digitized for FEDERAL FRASER RESERVE BANK OF ST. LOUIS it is more likely that losses will be born ultimately by debtholders. Stockholders have less exposure to losses when capital is low and, hence, are less concerned about probable losses resulting from risk-taking. In addition to tangible capital, firms have charter values, which may be defined as the economic value deriving from the opportunity to do business in the future. If a firm fails, it loses its charter value. Thus, firms with large charter values may refrain from taking risk and inject more tangible capital to avoid failure. Effective restrictions on competition raise charter values. When there is less competition, the opportunity to do business in the future is more valuable because firms can expect larger profits. Increased competition in the banking sector in the 1980s reduced the charter value of banks and, hence, increased their incentives to take risk.10 Empirical tests of the moral hazard hypothesis present mixed results. Gunther and Robinson (1990) examine the behavior of insured commer cial banks in the Dallas and Houston metropolitan areas between 1983 and 1984 and find a negative relationship between capital growth and changes in loan-to-asset ratios. They interpret this result as a negative relationship between capital adequacy and risk-taking. Other studies examine S&L data.11 Barth and others (1986) find that delay in closing insolvent S&Ls increased the resolution costs to the Federal Savings and Loan Insurance Corpora tion between 1981 and 1985. They suggest that desperate risk-taking by insolvent institutions was largely responsible for the increased costs. McKenzie and others (1992) compare returns of thrifts in 1987 and 1988 on traditional and nontraditional assets allowed by the DIDMCA and the Garn-St. Germain Act. They find that returns were generally lower on nontraditional assets, particularly on those held by capital-deficient institutions, than on traditional assets. They conclude that low-capital thrifts undertook 9 See Marcus (1984) and Keeley (1990). Relaxed portfolio restrictions are frequently cited as a main cause of the increased risk-taking by S&Ls. The increased risk-taking opportunities are relevant but not equally important to banks. The DIDMCA of 1980 and the Garn-St. Germain Act of 1982 substantially relaxed restrictions on S&L assets but did not notably loosen constraints on bank assets. 10 Keeley (1990) finds a relationship between risk-taking by banks and declining charter values during the 1980s. 11 The risk-taking behavior of S&Ls may be similar to that of banks in many respects. Both banks and S&Ls benefited from deposit insurance and underwent economic shocks and regulatory reform in the 1970s and 1980s. 7 projects with low net present values to increase the variance of the return and, hence, took more active advantage of risk-taking opportunities opened by legislative changes. While the studies discussed above support moral hazard theories, some others fail to find evidence of moral hazard. According to Benston and Carhill (1992), investment in nontraditional assets was a major cause of thrift failures that occurred between 1985 and 1991, but that investment was undertaken primarily by initially solvent thrifts. Thus, the shortage of capital was a result, rather than a cause, of increased risk. Gilbert (1991), who studies the behavior of under capitalized banks between 1985 and 1989, shows that banks reduced their assets substantially while undercapitalized. This finding does not support a negative relationship between the capital ratio and risk-taking. In addition, recent studies on the credit crunch find a positive relationship between the capital ratio and loan growth in the early 1990s.’2 Incentives o f M anagers Moral hazard theories assume that managers maximize stockholder wealth. Some recent studies, however, suggest that the incentives of managers may differ from those of outside shareholders.13 Managers of failed banks may have difficulty finding comparable positions. Thus, bank failure reduces expected future earnings of the bank’s managers. In this regard, bank managers are similar to stockholders with a large charter value who tend to refrain from taking risk. Furthermore, managers, whose com pensation packages are predetermined in many cases, may not benefit as much from risk-taking as stockholders.14 In addition, it is difficult for stockholders to monitor managers due to the 12 For example, Johnson (1991), Bernanke and Lown (1991) and Peek and Rosengren (1992). Since tightened capital requirements implemented in the early 1990s can partly explain the contraction by capital-deficient banks, the credit crunch does not necessarily contradict moral hazard. Nevertheless, their findings are still suggestive of the magni tude of moral hazard problems. problem of analyzing the quality of existing assets and investment opportunities. Many studies recognize that stockholders are not as well-informed as managers about the earnings prospects of firms.15 These arguments suggest that managers may not act in the best interest of stockholders. In general, managers may prefer not to take as much risk as stockholders want because they face a larger cost and a smaller benefit from addi tional risk.16 Many managers, however, may adopt extremely risky strategies under certain conditions, namely, when a large proportion of managers are incompetent and profits are declining.17 The banking sector appears to have satisfied these conditions in the 1980s. Incompetent managers, who cannot effectively control costs and make wise investment choices, may need to pursue more risky strategies to keep their jobs longer. Stockholders want managers to be competent and loyal to them. When stock holders are not accurately informed about the future earnings prospects of their banks, they may rely on some easily identifiable measures to judge the quality of management. Earnings records may reflect the quality of a bank’s man agement. If a bank takes high risk as demanded by stockholders, it will occasionally suffer low earnings resulting from unlucky outcomes. Managers are not to blame in that case. Stock holders may fire managers, however, if earnings turn out to be consistently lower than the industry average. Of course, consistently below-average earnings may be a result of bad management. Since incompetent managers cannot do as well as competent ones on average, a conservative strategy by incompetent managers w ill consis tently result in below-average earnings. One way they can occasionally have above-average earnings not seem to apply to the banking industry of the 1980s, when banks suffered low profits. 17 Gorton and Rosen (1992) present a game-theoretic model showing that banks on average pursue risky strategies in these conditions. 13 See Saunders and others (1990), Allen and Cebenoyan (1991) and Gorton and Rosen (1992). 14 Houston and James (1993) fail to find any evidence that the structure of management compensation in banking serves to reward managers for exploiting risk-taking opportunities. 16 See Myers and Majluf (1984), Miller and Rock (1985) and MacKie-Mason (1990). 16 Jensen (1986) argues that managers with large free cash flow may expand their firms beyond the optimal size to increase their power and perquisites. This argument does JULY/AUGUST 1994 8 is to take on high-variance projects. Occasional high earnings can confuse stockholders, making judgements about the quality of management difficult. Thus, taking risk may be a rational strategy for incompetent managers. Intuitively speaking, one has to take chances and hope for good luck if he or she cannot rely on ability. In a tightly regulated industry, incompetent managers may be able to survive without taking excessive risk because stockholders do not observe much difference between competent and incompetent managers.18 Even incompetent ones can generate decent profits when there is little competition. Moreover, it is difficult for competent managers to excel if competition is limited by regulation. Considering that the banking business was tightly regulated until the early 1980s, the proportion of incompetent man agers may have been higher in the banking sector than other industries. Deregulation increased competition, and the banking industry experi enced declining market shares and profitability in the 1980s. When tight cost control and hard search for profit opportunities are called for, disparities between competent and incompetent managers become clear because managerial ability plays a significant role. The banking develop ments in the 1980s may have induced some incompetent bank managers to take excessive risks in a desperate attempt to preserve their jobs. Thus, it is possible that a change in managers’ incentives may have been largely responsible for the increased riskiness of banks. Evidence of risk-taking driven by the incentives of incompetent managers is indirect and limited. Gorton and Rosen (1992) show that “entrenched” mangers made more risky loans between 1984 and 1990 and interpret the result as evidence that excessive risk-taking was driven largely by managers’ incentives.19 Allen and Cebenoyan (1991) find that the most powerful managers actively acquired other banks between 1980 and 1987 and that those acquisitions were not valueenhancing.20 Although the focus of their study is not risk-taking by banks, this result is consistent with that of Gorton and Rosen. 18 Flood (1993) also argues that incompetent managers can survive indefinitely in a protected industry. 19 By their definition, entrenched managers are the mangers with enough shares to be protected from outside shareholders’ pressure but not enough shares to have the same objective as that of outside shareholders. 20 The power of managers increases with managerial stake and decreases with concentration of outsiders’ shareholdings. Digitized forFEDERAL FRASER RESERVE BANK OF ST. LOUIS U nexpected Shocks Given that banks experienced many external events that impaired their financial structure in the 1970s and 1980s, deliberate risk-taking is not required to explain the increased riskiness of banks. Banks may not have realized that certain categories of loans were particularly risky until many borrowers became unable to repay. Con sidering that geographic and portfolio restrictions limit the ability of banks to diversify, the financial strength of banks may greatly depend on the quality of major-category loans. Before the establishment of the FDIC, banking distress was often triggered by the collapse of a major industry or a stock market crash. In the 19th century, it was not unusual for banks to finance the expansion of a booming sector such as cotton or railroads, only to experience finan cial difficulties when that sector went bust.21 In those cases, deterioration in the asset quality of banks could hardly be attributed to increased risk-taking incentives resulting from institutional changes because the banking sector was governed largely by market forces at that time. Thus, it is possible that realization of highly unlikely outcomes or judgement errors were largely responsible for the financial problems of banks in the 1980s. Banks became extremely unlucky, or they unknowingly, rather than deliberately, increased the variance and lowered the expected value of their asset portfolios because of inability to assess the soundness of investment opportu nities. For example, banks might have expected that the real estate boom of the 1980s would continue for a long time and were surprised by the bust. McKenzie and others (1992) and Emmons (1993) find that the condition of local economies significantly contributed to the earnings and failures of financial institutions. These results imply that external shocks had significant effects on the performance of banks, but does not disprove the role of deliberate risk-taking. Risk-taking generally makes banks more vulner able to external shocks. 21 Park (1993) examines the economic environment that led to banking panics. 9 EM PIRICAL STUDY To test the relevance of each hypothesis for explaining bank risk in the 1980s, we need to empirically examine the implications of each hypothesis using a consistent sample and com parable methodology. The application of rigorous econometric techniques to the risk-taking behavior of banks involves many problems, such as deter mining appropriate measures of risk and specifying a reasonable functional form. To avoid these difficulties, this study divides banks into several groups based on year-end capital ratios and earnings on assets for selected years, and com pares changes in various risk measures such as loan growth, the proportions of risky loans and funding strategies during the following year across groups. By looking at the behavior of banks in following years, we can better deter mine causality. It is difficult to infer causality from contemporaneous relationships because risk measures interact with capital ratios and earnings. For example, a low capital ratio can be either a cause or a result of increased risk. Aggressive risk-taking by poorly capitalized banks may be interpreted as evidence of moral hazard; stockholders more actively took advan tage of deposit insurance when they did not have much of their own wealth to lose. The association of low earnings with more risk-taking can be an indication of desperate profit-seeking by incompetent managers. Incompetent managers, who are more likely to have low earnings, may have adopted risky strategies to obscure their incompetence with occasional high earnings. The hypothesis of unexpected economic shocks may be supported by insignificance of capital adequacy and earnings along with significance of the condition of local economies in explaining the performance of banks. Data This analysis uses the Consolidated Reports of Condition and Income (Call Reports) data. The sample consists of domestically chartered, FDICinsured commercial banks. The sample excludes banks that were less than five years old at the 22 Potential biases resulting from the exclusion of merged banks will be discussed below. 23 For example, the act authorizes federal banking regulators to limit the growth of an institution and to remove any person causing harm to an insured financial institution. time of financial statements, and banks that were involved in mergers and acquisitions during the year analyzed. Relatively new banks may behave unusually. For example, they may expand rapidly despite low profits to cultivate a customer base. Mergers and acquisitions can dramatically change the financial characteristics of banks involved.22 In these cases, risk-taking may not have much to do with changed financial characteristics. The time span covered by this analysis is 1984 to 1988. Changes in risk measures during 1985, 1986, 1987 and 1988 are examined based on capital ratios and earnings on assets at the end of 1984, 1985, 1986 and 1987, respectively. The deregulation of the early 1980s, which under mined the charter values of banks and opened more risk-taking opportunities for them, might have led to active risk-taking in the mid-1980s. Legislative changes in the late 1980s discouraged banks from taking risk. The Federal Reserve Board introduced risk-based capital guidelines in 1988 based on the 1987 Basle Accord, and the Financial Institutions Reform, Recovery and Enforcement Act of 1989 tightened bank and thrift regulation.23 The FDIC Improvement Act of 1991 added more provisions designed to prevent risk-taking, such as risk-based deposit insurance premiums, prompt corrective action on undercapitalized banks, and more frequent bank examinations. Thus, risk-taking is likely to have been most active during the years examined by this study. Capital Ratios Banks are divided into four groups based on the year-end ratio of capital to assets: Group C l—less than 5 percent; Group C2—greater than or equal to 5 percent but less than 7 percent; Group C3—greater than or equal to 7 percent but less than 10 percent; and Group C4—greater than or equal to 10 percent.24 According to moral hazard theories, the first group has the strongest incentive to take risk and the incentive decreases with the capital ratio. The incentive, however, may not translate into actual risk-taking because of regulatory constraints. The capital ratios of since the stock market data are not available for most small banks. The book value may differ from the market value. The difference, however, may introduce merely a random noise rather than a systematic bias. 24 This analysis uses the book value of equity capital, consist ing of common stock, perpetual preferred stock, surplus and retained earnings. The market value cannot be obtained, JULY/AUGUST 1994 10 most banks in the first group fall short of the required minimum.25 Thus, banks in the first group may have been subjected to tight supervi sion by regulators and unable to increase risk. Most aggressive risk-taking by the second group, therefore, is also consistent with the moral hazard theory. Risk measures used here include: the rate of loan growth; the change in the ratio of commercial real estate loans (loans secured by construction and land development, multifamily residential properties and nonfarm, nonresidential properties) to total loans; the change in the ratio of loans to insiders (executive officers and principal share holders) to total loans; the change in the ratio of large time deposits (time deposits of $100,000 or more) to total assets; and the interest rate on large time deposits.26 Since sound investment opportunities are limited, rapid loan growth generally increases the risk of banks. Commercial real estate loans are regarded as relatively risky loans, and some banks may apply lower lending standards to insiders. Thus, the riskiness of loan portfolios may increase with the proportions of commercial real estate loans and loans to insiders. Increases in the share of large time deposits and high interest rates on large time deposits may indicate the banks’ desire to expand rapidly and to take risk. Bidding up interest rates on large time deposits is a fast way to raise funds for rapid expansion because large time deposits are more sensitive to interest rates than other deposits. In comparing risk-taking behavior across the four groups, this analysis focuses mainly on the distribution of percentile ranks of each risk measure. The use of percentile ranks, instead of actual values, makes the analysis concise. Examination of the distribution, rather than summary statistics, enables more comprehensive analyses of bank behavior. Furthermore, summary statistics of ratios or growth rates can be seriously contaminated by outliers. This analysis some what sacrifices analytical rigor in that it does not rely on formal statistical tests. The lack of formal statistical tests may, however, be partly compen 25 The minimum ratio of primary capital to total assets was set at 5.5 percent in 1985 and remained in effect until the end of 1990. Primary capital includes loan-loss provisions. Risktaking incentives may be more closely related to stockhold ers’ equity than primary capital. 26 The interest rate is estimated by dividing interest expense on time certificates of deposit of $100,000 or more by the average amount of the same deposit outstanding during the year. Digitized forFEDERAL FRASER RESERVE BANK OF ST. LOUIS sated by the fact that the entire population of commercial banks is examined over an extensive time period. Table 1 presents summary statistics and a con densed distribution of the percentile rank of loan growth, in which higher percentiles are associated with higher rates of loan growth. The table shows the lowest median rate of loan growth for Group C l banks in all four years, but fails to show any apparent pattern among the remaining three groups. Although Group C2 generally had high median rates of loan growth, the group with the highest median differs across years. Thus, com parison of median loan growth rates fails to present clear evidence of moral hazard. The distribution of loan growth is more inter esting. The table reports the percentage of banks in each group belonging in a certain percentile rank group of the entire population. For example, 28.6 percent of Group C l banks in 1984 had loan growth rates ranking them in the 0-10 per centile in 1985. Thus, the proportion of Group C l banks belonging in the 0-10 percentile group was almost three times as large as that of the entire population of banks. If the distribution of a group is roughly equal to the distribution of the entire population, about 10 percent of the group should belong in the 0-10 percentile group, another 10 percent in the 10-20 percentile group, about 30 percent in the 20-50 percentile group, and so on. Many Group C l and Group C2 banks grew very slowly (0-10 percentile ranks), while many other banks in the same groups grew very fast (90-100 percentile ranks). In contrast, well-capitalized banks belonged mostly in the middle percentile groups (20-50 percentile ranks and 50-80 per centile ranks). In 1987 and 1988, the proportion of Group C l banks belonging in the 90-100 per cent group decreased markedly, but remained larger than the proportion in the 80-90 per centile group. A possibility is that regulators more successfully prevented undercapitalized banks from expanding loans in those years, but that some undercapitalized banks continued to 11 Table 2 Changes in the Ratio of Commercial Real Estate Loans to Total Loans Table 1 Rates of Loan Growth 1984-1985 Percentile 0-10 10-20 C1 C2 28.6% 11.9 10.3% 8.0 C3 8.6% 10.0 1984-1985 C4 10.7% Percentile C1 C2 13.5% C3 12.5% 9.2% C4 12.3 0-10 9.1 10.7 10.5 8.1 8.5% 20-50 19.8 22.9 32.6 33.7 10-20 50-80 16.7 31.9 31.0 26.8 20-50 20.7 24.2 30.4 37.0 80-90 10.7 12.7 9.5 7.7 50-80 29.0 28.4 30.9 30.2 90-100 12.3 14.0 8.4 8.5 80-90 11.6 11.7 9.4 9.1 Number of banks 318 3,190 6,224 2,620 90-100 16.0 12.5 9.6 7.1 Mean 0.0074 0.0832 0.0577 0.0548 Number of banks 318 3,190 6,224 2,620 -0.0104 0.0782 0.0434 0.0249 Mean 0.0129 0.0092 0.0073 0.0049 Median 0.0052 0.0031 0.0011 0.0000 Percentile C1 Median 1985-1986 Percentile C2 C1 11.6% C3 8.4% 1985-1986 C4 7.9% 0-10 34.5% 10-20 12.6 9.6 9.6 11.0 0-10 20-50 20.3 24.5 31.6 34.5 50-80 16.8 29.4 31.4 80-90 4.7 11.4 10.3 90-100 11.1 13.4 Number of banks 380 3,119 Mean -0.0276 0.0598 0.0449 Median -0.0777 0.0484 0.0299 Percentile 10-20 C1 C2 34.8% 16.5 12.6% 9.6 C3 C4 10.5% 10.1% 10-20 7.9 9.7 9.9 10.9 29.5 20-50 25.5 23.1 29.9 39.2 8.4 50-80 25.7 30.6 31.1 27.4 8.8 8.7 80-90 14.5 12.7 9.4 7.4 6,050 2,541 90-100 15.5 13.2 9.6 6.2 0.0608 Number of banks 380 3,119 6,050 2,541 0.0179 Mean 0.0215 0.0176 0.0122 0.0065 Median 0.0113 0.0102 0.0047 0.0000 Percentile C1 1986-1987 0-10 C2 10.8% C3 7.2% 9.4 1986-1987 C4 7.1% 9.0% C2 C3 9.4% C4 10.4 0-10 11.1% 11.4% 10.7 10.2 10.2 9.2 9.2% 20-50 22.3 27.5 31.0 32.9 10-20 50-80 12.6 27.6 32.7 31.0 20-50 25.5 25.1 29.7 38.1 80-90 6.2 11.2 9.6 10.3 50-80 24.6 29.2 31.4 29.2 90-100 7.6 11.5 10.1 8.4 80-90 11.4 11.4 10.2 7.4 Number of banks 552 3,048 5,597 2,370 90-100 16.7 12.8 9.1 6.9 Mean -0.0305 0.0648 0.0757 0.3530 Number of banks 552 3,048 5,597 2,370 Median -0.0616 0.0523 0.0579 0.0492 Mean 0.0170 0.0128 0.0092 0.0048 Median 0.0050 0.0054 0.0034 0.0000 Percentile C1 1987-1988 Percentile 0-10 C1 38.0% C2 11.4% C3 7.5% 1987-1988 C4 8.1% 10-20 15.7 10.1 9.4 10.0 20-50 20.4 27.6 31.6 30.8 0-10 16.3% C2 C3 10.7% 9.9% C4 8.2% 10-20 10.4 8.6 10.3 10.7 20-50 22.6 25.6 30.1 35.7 50-80 15.0 29.5 31.4 30.6 80-90 3.8 9.9 10.7 9.8 50-80 26.7 31.1 30.0 29.5 10.6 12.2 9.8 8.0 90-100 7.1 11.5 9.3 10.7 80-90 2,550 5,557 2,526 Number of banks 521 Mean -0.0170 Median -0.0356 90-100 13.4 11.8 9.8 8.0 0.0811 0.0874 0.1532 Number of banks 521 2,550 5,557 2,526 0.0766 0.0774 0.0779 Mean 0.0061 0.0089 0.0055 0.0045 Median 0.0007 0.0035 0.0002 0.0000 Notes: Group C1: Capital Ratio < 0.05; Group C2: 0.05 < Capital Ratio < 0.07; Group C3: 0.07 < Capital Ratio <0.10; and Group C4: Capital Ratio >0.10. JULY/AUGUST 1994 12 grow extremely fast when they could circumvent supervision. The rapid loan growth of many poorly capitalized banks may have been moti vated by moral hazard. deposit at risky banks (supply). Nevertheless, the pattern of changes in the ratio of large deposits appears to be consistent with those of loan growth and shifts in loan portfolios. Tables 2 and 3 show changes in the ratio of commercial real estate loans to total loans, and changes in the ratio of loans to insiders to total loans. The changes in the ratio of commercial real estate loans display patterns similar to the rates of loan growth. W hile the ratio fell sharply for many Group C l and Group C2 banks, it increased significantly for many others. Group C3 and Group C4 banks were concentrated in the middle percentile groups. The ratio for Group C l banks in the highest percentile, how ever, did not decrease substantially in 1987 and 1988, unlike the rate of loan growth. In addi tion, the median change was inversely related to the capital ratio. These results may reflect the difficulty of regulating the riskiness of loan portfolios, compared with restricting the rate of loan expansion. In the case of loans to insiders, only Group C l banks were distributed heavily toward both tails. It appears that only those banks facing the possibility of imminent failure relied on insider loans as a means of increasing risk. The median fails to show a clear pattern. In sum, the average behavior of poorly capital ized banks was not notably different from that of well-capitalized banks. A more interesting finding is that a higher proportion of low-capital banks adopted highly risky strategies. This effect might have been more pronounced if the behavior of banks that failed during the year were included in the sample. Failed banks, most of which had been poorly capitalized, might have pursued extremely risky strategies if they were given the opportunity.27 In general, large time deposits appear to have grown slower at poorly capitalized banks. Table 4 shows smaller changes in the ratio of large time deposits to total assets for many Group C l and Group C2 banks. A sizable proportion of group C2 banks, however, showed relatively fast increases in the ratio. Furthermore, although a relatively small proportion of Group C l banks belonged in the 90-100 percentile groups, the proportion was consistently higher than that belonging in the 80-90 percentile group. Poorly capitalized banks generally paid higher interest rates on large time deposits, but the interest rates were not distributed heavily toward the tails (Table 5). It is complicated to interpret the behavior of large time deposits, which are not fully insured, and the interest rates on those deposits because demand and supply factors interact. In other words, the two variables reflect both the desire of low-capital banks to grow (demand) and the willingness of investors to 27 Failed banks include FDIC-assisted mergers. Another source of potential bias is voluntary mergers. The proportion of poorly capitalized banks (Groups C1 and C2), however, was not significantly higher than the proportion of well-capi talized banks (Groups C3 and C4) among acquired banks. Digitized forFEDERAL FRASER RESERVE BANK OF ST. LOUIS Earnings on Assets This section compares four groups of banks, representing each quartile of earnings on assets; Group E l and Group E4 represent the lowest and highest quartiles. Managers of Group E l banks, who were incompetent or had bad luck, may have had the strongest incentives to increase risk because they needed high earnings in the next period to preserve their jobs. Thus, the theory based on the managers’ incentives pre dicts the most aggressive risk-taking by Group E l banks and the most conservative strategies by Group E4 banks. The same risk measures used in the previous section are analyzed. Table 6 shows a positive relationship between earnings on assets and the rate of loan growth. The median growth rate w as g enerally high er for banks with higher earnings. This result suggests that the availability of profitable investment opportunities was the main determinant for loan growth; banks with high earnings, which might have more profitable lending opportunities, grew faster. A notable pattern in the table, however, is that the propor tion of Group E l banks belonging in the 90-100 percentile group consistently exceeded that belonging in the 80-90 percentile group. It is thus possible that the pattern of loan growth reflected the desperate attempts of some bank managers to increase profits. The pattern of changes in loan portfolios for low-earnings banks is similar to that for 13 Table 4 Changes in the Ratio of Large Time Deposits to Assets Table 3 Changes in the Ratio of Loans to Insiders to Total Loans 1984-1985 1984-1985 Percentile 0-10 10-20 C2 C1 16.0% 10.0% 8.0 9.9 C3 C4 Percentile C2 C1 03 8.4% C4 6.1% 10.7% 0-10 27.7% 14.4% 10.0 10.4 10-20 13.5 12.0 10.2 6.8 28.7 20-50 26.4 28.4 29.7 33.3 9.4% 20-50 26.2 30.1 30.7 50-80 29.8 31.4 30.0 28.2 50-80 16.4 26.5 31.6 32.0 80-90 7.6 9.1 10.4 10.6 80-90 6.9 8.9 10.1 11.5 90-100 12.4 9.6 9.5 11.4 90-100 9.1 9.8 10.1 10.2 2,758 5,379 2,015 Number of banks 318 3,190 6,224 2,620 Number of banks 275 Mean -0.00027 0.00010 -0.00006 0.00088 Mean -0.0184 -0.0038 0.0018 0.0049 Median -0.00001 -0.00001 -0.00001 0.00000 Median -0.0128 -0.0024 0.0008 0.0019 Percentile C1 C4 Percentile C1 1985-1986 1985-1986 C2 C3 C2 C3 C4 0-10 29.5% 12.6% 9.1% 6.1% 9.9 10-20 11.8 12.9 9.6 7.0 29.5 29.0 20-50 26.6 29.8 31.2 28.2 9.2% 15.8% 10.2% 10-20 9.7 8.3 10.9 20-50 23.5 32.4 0-10 10.7% 50-80 27.8 31.7 30.4 27.0 50-80 17.9 25.2 30.2 37.2 80-90 10.6 8.5 9.9 12.3 80-90 6.3 8.6 10.1 11.9 90-100 12.6 8.9 10.0 11.1 90-100 7.9 11.1 9.8 9.6 Number of banks 341 2,665 5,276 1,952 Number of banks 380 3,119 6,050 2,541 Mean 0.00042 -0.00060 Median 0.00000 -0.00002 Percentile C1 Mean -0.0256 -0.0058 -0.0039 0.0006 0.00000 Median -0.0177 -0.0061 -0.0031 -0.0002 C4 Percentile C1 9.0% 0-10 22.6% 12.4% 8.7% 6.9% 0.00049 -0.00003 -0.00001 1986-1987 1986-1987 C2 C3 C2 C3 C4 0-10 14.7% 9.8% 10.0% 10-20 10.6 8.5 10.2 11.3 10-20 10.7 10.4 9.9 9.4 20-50 23.8 30.7 30.1 30.3 20-50 26.2 24.4 30.5 37.2 50-80 30.9 33.7 29.2 26.7 50-80 20.8 29.0 31.1 31.0 80-90 8.9 8.7 10.5 10.7 80-90 6.9 11.1 10.3 8.6 90-100 11.0 8.5 9.9 11.9 90-100 12.9 12.6 9.5 7.0 Number of banks 462 2,565 4,883 1,836 Number of banks 552 3,048 5,597 2,370 Mean -0.00046 -0.00045 Mean -0.0027 0.0030 0.0038 0.0029 Median -0.00002 -0.00002 Median -0.0017 0.0033 0.0024 0.0004 0.00017 0.00080 -0.00004 -0.00006 1987-1988 1987-1988 Percentile C1 C2 Percentile C1 C2 C3 C4 29.0% 12.8% 8.4% 6.5% 10.1 10.3 10-20 13.1 12.0 9.8 7.7 27.8 31.2 32.0 20-50 23.7 26.6 29.9 35.1 31.0 33.7 29.9 26.0 50-80 18.7 27.3 31.5 31.9 9.0 10.0 9.8 10.5 80-90 7.1 10.5 10.3 9.6 13.3 9.9 9.5 10.7 90-100 8.3 10.7 10.2 Number of banks 521 2,550 5,557 Mean -0.0130 0.0022 0.0052 0.0066 Median -0.0062 0.0029 0.0044 0.0040 9.8% 10-20 13.6 8.8 20-50 19.5 50-80 80-90 2,202 Number of banks 435 Mean -0.00133 0.00044 Median -0.00005 -0.00004 C4 0-10 13.6% 90-100 03 10.4% 0-10 9.6% 4,858 -0.00021 1,944 0.00001 -0.00015 -0.00025 9.3 2,526 JULY/AUGUST 1994 14 Table 5 Interest Rates on Large Time Deposits Table 6 Rates of Loan Growth 1984-1985 1985 Percentile C1 C2 0-10 7.7% 7.9% 10.1% 12.8% 0-10 21.4% 6.7% 5.3% 10-20 8.0 8.6 10.3 11.3 10-20 15.5 8.1 7.7 8.8 20-50 27.6 34.4 29.6 25.6 20-50 30.5 29.7 28.7 31.1 50-80 32.8 32.1 30.0 26.9 50-80 19.2 34.0 35.2 31.7 80-90 12.2 9.1 10.0 11.0 80-90 6.2 10.9 12.1 10.8 90-100 11.6 8.0 10.0 12.4 90-100 7.3 10.7 11.1 10.9 Number of banks 311 3,143 6,014 2,360 Number of banks 3,056 3,082 3,104 3,113 Mean 0.0882 0.0871 0.0871 0.0868 Mean Median 0.0876 0.0861 0.0863 0.0864 Median Percentile C1 C4 Percentile C3 C4 Percentile E1 0-10 7.4% 8.5% E3 E4 6.8% 0.0059 0.0770 0.0835 0.0824 -0.0148 0.0613 0.0721 0.0565 1985-1986 1986 C2 E2 C3 10.1% E1 E2 E3 E4 13.5% 0-10 23.8% 7.3% 4.7% 4.7% 17.4 9.0 6.8 7.2 7.9 9.0 10.2 11.3 10-20 20-50 23.5 33.9 29.8 26.6 20-50 32.0 31.9 27.9 28.3 50-80 35.9 31.8 29.7 27.6 50-80 16.4 31.4 35.6 36.1 9.9 80-90 4.6 10.0 13.1 12.1 5.8 10.3 12.0 11.6 2,945 3,036 3,057 3,062 10-20 80-90 10.4 9.3 90-100 13.7 8.5 Number of banks 365 3,067 Mean 0.0747 Median 0.0745 0.0736 0.0725 10.4 10.0 11.2 90-100 5,846 2,333 Number of banks 0.0733 0.0725 0.0726 Mean -0.0178 0.0541 0.0781 0.0819 0.0723 Median -0.0591 0.0354 0.0615 0.0611 1986-1987 1987 Percentile C1 C2 0-10 9.7% 7.6% 10-20 9.8 8.1 C3 10.2% 9.9 C4 12.8% 12.8 20-50 24.1 28.5 31.2 30.4 50-80 34.8 35.8 28.9 23.6 80-90 9.8 10.5 9.8 9.9 90-100 11.8 9.5 Number of banks 518 Mean 0.0650 Median 2,968 0.0658 0.0654 0.0654 9.9 5,407 0.0643 0.0643 10.5 2,180 0.0634 0.0633 0-10 C1 C2 8.6% 7.1% C3 10.0% C4 13.3% 10-20 10.0 7.4 9.5 13.9 20-50 25.8 26.7 30.7 32.7 50-80 80-90 32.6 12.6 36.2 12.1 30.4 9.7 22.0 7.9 90-100 10.6 10.5 9.6 10.3 Number of banks 501 2,492 5,419 2,357 Mean 0.0700 Median 0.0713 0.0711 0.0717 E1 E2 E3 E4 4.1% 0-10 25.6% 7.2% 4.3% 10-20 18.1 9.1 7.3 6.2 20-50 30.7 33.4 29.0 26.8 50-80 15.1 31.0 35.3 37.5 80-90 4.3 9.5 12.7 13.0 90-100 6.2 9.7 11.4 12.3 2,722 2,924 2,956 2,975 Number of banks Mean Median 0.2278 0.0720 0.0911 0.1143 -0.0375 0.0521 0.0767 0.0863 1987-1988 1988 Percentile Percentile 0.0698 0.0703 Digitized forFEDERAL FRASER RESERVE BANK OF ST. LOUIS 0.0684 0.0685 Percentile E1 E2 E3 E4 0-10 25.4% 8.0% 3.8% 4.1% 10-20 16.0 9.7 7.9 6.9 20-50 28.2 32.7 30.8 28.2 50-80 16.8 29.6 35.3 36.9 80-90 4.8 10.2 12.2 12.4 90-100 8.7 9.6 10.0 11.5 2,603 2,845 2,838 2,878 Number of banks Mean 0.0379 0.0852 0.1045 0.1498 Median 0.0011 0.0727 0.0916 0.1016 Notes: Each group represents a quartile of earnings on assets. During the four years, the cutoff points were about 0.005 between Groups E1 and E2, 0.011 between Groups E2 and E3, and 0.016 between Groups E3 and E4. 15 low-capital banks but somewhat less pronounced. While many Group E l banks reduced the ratios of commercial real estate loans and loans to insiders to total loans, a large proportion of banks in the same group rapidly increased the portfolio shares of the risky loans (Tables 7 and 8). No apparent pattern is found for other groups. These findings suggest that a higher proportion of lowearnings banks pursued highly risky strategies. The ratio of large time deposits to assets increased faster for banks with higher earnings (Table 9). Yet a sizable proportion of Group E l banks increased the ratio of large time deposits very fast (90-100 percentile ranks). In addition, a relatively high proportion of Group E l banks paid the highest interest rates on large time deposits (Table 10). Thus, many Group E l banks appear to have bid up interest rates to attract large time deposits. Examination of the effects of earnings on risk-taking presents findings similar to the ones derived from the relationship between capital adequacy and risk-taking behavior. Although the majority of banks with low earnings were comparatively conservative, many other banks with low earnings adopted extremely risky strategies. As in the case of low-capital banks, this result could have been strengthened if the behavior of banks that failed during the year were taken into consideration.28 The divergent risk-taking behavior among low-earnings banks, however, is not as pronounced as that among low-capital banks. Thus, although this analysis supports the hypothesis that low earnings induced some banks to increase risk aggressively, evidence is somewhat weaker than that for moral hazard. In addition, the true motivation for risktaking is not clear for some low-earnings banks because many low-earnings banks also had low capital.29 This problem will be addressed below. 28 The majority of failed banks belonged in Group E1. Among acquired banks, the proportion of Group E1 banks was the highest in three of the four years. The proportion, however, was not overwhelming. Conditions o f Local Economies If external shocks are the main factor affecting the condition of banks, the strength of local economies should largely explain behavior and performance differences across banks. Table 11 shows the results of two sets of regressions that examine the effects of employment growth in home states on earnings on assets and loan growth. The employment growth, which reflects the con dition of local economies, is found to positively affect earnings and loan growth. Although they support the significance of local economic con ditions, these results do not disprove the roles of other factors. We may better assess the roles of stockholders’ and managers’ incentives by examining whether the behavior of banks that cannot be explained by the condition of local economies is systemati cally related to capital ratios or earnings on assets of the previous period. Thus, the residual rates of loan growth are examined.30 Higher demand for bank loans generally follows improving eco nomic conditions. Thus, the residuals reflect the loan growth that is unexplained by demand effects. Systematic differences in the residual loan growth across groups may be viewed as consequences of risk-taking in previous periods. Tables 12 and 13 report the distribution of the residual rates of loan growth. The distribution of the residuals is similar to that of loan growth observed in Tables 1 and 6. Group C2 banks are distributed heavily toward both tails. For Group C l and E l banks, the percentage of banks belonging in the highest percentile group is not high, but consistently higher than that in the second highest percentile group. Thus, after considering the condition of local economies, systematic differences still exist in loan growth across the groups classified based on capital ratios and earnings on assets. Accordingly, the recognition of large loan loss provisions by banks with lowquality loans. For this reason, the residual changes in earn ings are not discussed in detail. 29 The correlation coefficient between capital ratios and earn ings on assets was 0.393 in 1984, 0.326 in 1985, 0.335 in 1986, and 0.432 in 1987. The correlation may become even higher if the market value, instead of book value, of capital is used. Banks with high-quality assets are more likely to have high earnings and high market values relative to book values. 30 The residual changes in earnings for low-capital and lowearnings banks are found to be distributed heavily toward both tails. It is possible to interpret large variances in earn ings changes as a result of risk-taking in previous periods. The large variances, however, can result from delayed JULY/AUGUST 1994 16 Table 7 Changes in the Ratio of Real Estate Loans to Total Loans Table 8 Changes in the Ratio of Loans to Insiders to Total Loans 1984-1985 Percentile 0-10 E1 E2 12.4% 9.3% 1984-1985 E3 8.8% E4 Percentile 9.5% 0-10 E1 E2 13.1% 8.9% E3 7.6% E4 10.5% 10-20 10.0 10.9 10.6 8.5 10-20 10.2 9.2 9.7 10.9 20-50 28.2 29.2 31.6 31.0 20-50 25.0 31.4 32.7 30.8 50-80 28.9 30.6 29.5 31.2 50-80 27.9 31.7 31.9 28.3 80-90 9.8 10.7 9.9 9.6 80-90 11.9 9.4 8.9 9.8 90-100 10.7 9.5 9.5 10.2 90-100 11.9 9.2 9.2 9.7 Number of banks 3,056 3,082 3,104 3,113 Number of banks 2,641 2,694 2,630 2,465 Mean 0.0065 0.0078 0.0075 0.0079 Mean 0.00024 -0.00019 0.00058 -0.00001 Median 0.0010 0.0016 0.0009 0.0018 Median 0.00003 0.00000 -0.00001 -0.00007 1985-1986 Percentile 0-10 E1 E2 10.0% 9.8% 1985-1986 E3 9.7% E4 10.5% Percentile 0-10 E1 E2 13.3% 9.4% E3 8.6% E4 8.7% 10-20 9.6 10.4 9.7 10.2 10-20 9.1 10.6 9.1 11.3 20-50 30.2 30.2 29.4 30.0 20-50 26.3 30.2 31.9 31.6 50-80 29.2 29.8 30.1 30.8 50-80 29.6 29.9 31.8 28.8 80-90 10.2 10.0 10.7 9.1 80-90 10.3 10.0 9.7 10.0 90-100 10.9 9.7 10.1 9.4 90-100 11.5 10.0 9.0 9.6 Number of banks 2,945 3,036 Number of banks 2,497 2,626 2,654 2,462 3,057 3,062 Mean 0.0135 0.0122 0.0130 0.0122 Mean Median 0.0048 0.0046 0.0055 0.0043 Median -0.00070 0.00043 0.00001 -0.00001 1986-1987 Percentile E1 E2 0.00029 0.00000 -0.00006 1986-1987 E4 Percentile 9.0% 9.0% 0-10 9.7 10-20 9.9 E3 0-10 12.2% 10-20 10.9 9.8 9.6 20-50 28.8 29.9 30.3 30.8 20-50 5 -8 0 28.3 31.3 30.6 29.7 50-80 80-90 10.0% 0.00037 E1 E2 12.5% 10.0% E3 8.6% E4 9.0% 10.0 10.4 9.6 25.2 29.5 31.8 33.1 28.1 31.0 30.9 29.7 9.0 8.7 9.4 11.4 10.4 80-90 11.6 10.2 9.4 90-100 10.9 9.6 9.1 10.4 90-100 12.6 9.2 8.8 9.6 Number of banks 2,722 2,924 2,956 2,975 Number of banks 2,263 2,500 2,536 2,457 Mean 0.0081 0.0096 0.0102 0.0106 Mean -0.00014 -0.00005 Median 0.0018 0.0032 0.0036 0.0032 Median 0.00003 -0.00002 Percentile E1 E4 Percentile E1 8.7% 1987-1988 E2 0.00024 0.00034 -0.00006 -0.00008 1987-1988 E3 E2 E3 E4 0-10 13.3% 10.0% 8.2% 0-10 13.0% 10-20 10.2 10.2 9.8 9.8 10-20 11.2 10.3 9.6 9.0 20-50 28.9 30.4 30.7 29.9 20-50 23.5 30.2 32.8 32.8 50-80 26.5 30.1 32.3 30.8 50-80 28.0 28.2 31.9 31.8 80-90 11.0 11.4 9.1 8.6 90-100 13.4 10.2 8.1 Number of banks 2,196 2,429 9.4 9.9 9.5 11.1 90-100 11.7 9.3 9.4 9.7 Number of banks 2,603 2,845 2,838 2,878 80-90 Mean 0.0046 0.0056 0.0068 0.0071 Mean Median 0.0000 0.0000 0.0009 0.0013 Median Digitized forFEDERAL FRASER RESERVE BANK OF ST. LOUIS 9.8% 8.3% 2,462 9.2% 8.6 2,360 0.00021 -0.00001 -0.00014 -0.00031 -0.00005 -0.00014 -0.00015 -0.00015 17 Table 10 Interest Rates on Large Time Deposits Table 9 Changes in the Ratio of Large Time Deposits to Assets 1984-1985 1984-1985 E1 Percentile E2 16.4% 0-10 8.7% E3 7.8% Percentile E1 E2 E4 0-10 9.4% 8.8% E3 7.2% 10-20 8.9 10.5 10.4% 9.9 E4 11.4% 10.7 10-20 12.8 9.5 8.9 8.9 20-50 27.8 32.7 30.0 29.5 20-50 29.6 30.1 30.7 29.8 50-80 30.9 29.4 31.1 28.6 50-80 24.7 31.2 32.3 31.5 80-90 10.8 10.0 9.9 9.3 80-90 7.5 10.4 10.9 11.2 90-100 12.2 8.5 8.7 10.6 2,928 2,991 3,013 2,899 90-100 Number of banks 9.1 10.1 9.5 11.3 Number of banks 3,056 3,082 3,104 3,113 Mean 0.0886 0.0867 0.0864 0.0865 Median 0.0873 0.0859 0.0862 0.0860 Mean -0.0064 0.0017 0.0022 0.0043 Median -0.0038 0.0009 0.0011 0.0021 1985-1986 1985-1986 E1 Percentile E2 0-10 17.6% 10-20 11.8 8.8% E3 E4 Percentile E1 0-10 9.5% E2 E3 10.1% 9.2% E4 11.2% 7.3% 6.6% 10-20 9.8 9.9 10.1 10.2 11.2 9.3 7.7 20-50 28.0 31.0 31.1 29.8 29.8 30.4 30.4 29.3 10.0 20-50 28.5 32.4 31.3 27.9 50-80 50-80 25.0 28.8 31.7 34.2 80-90 10.4 9.6 10.0 80-90 8.3 9.5 10.3 11.9 90-100 12.5 8.9 9.2 9.5 90-100 8.8 9.2 10.1 11.9 Number of banks 2,804 2,948 2,972 2,896 2,945 3,036 3,057 3,062 Mean 0.0745 0.0728 0.0731 0.0728 Median 0.0732 0.0723 0.0725 0.0723 Percentile E1 Number of banks Mean -0.0123 -0.0044 -0.0021 0.0020 Median -0.0078 -0.0045 -0.0022 -0.0001 1986-1987 1986-1987 E1 Percentile E2 E3 E4 0-10 17.2% 10.5% 6.3% 6.6% 10-20 12.6 10.3 8.9 8.4 20-50 30.7 29.1 29.7 30.6 50-80 23.1 29.4 33.4 33.4 80-90 6.8 10.1 11.2 11.7 90-100 9.7 10.5 10.6 9.2 2,722 2,924 2,956 Number of banks 2,975 Mean -0.0035 0.0031 0.0068 0.0052 Median -0.0010 0.0020 0.0042 0.0039 0-10 E1 0-10 19.2% 13.5 10-20 E2 8.9% 10.4 E3 6.6% 8.3 E4 6.2% 8.1 27.4 31.7 30.0 30.6 50-80 22.4 29.4 34.7 32.7 7.5 10.8 11.8 9.9 9.8 9.7 10.6 2,603 2,845 2,838 2,878 90-100 Number of banks 9.6 Mean Median -0.0036 -0.0004 0.0039 0.0033 9.2% E4 10.5% 10-20 11.3 9.9 9.5 9.4 29.5 30.1 29.8 30.7 50-80 29.6 30.3 30.3 29.8 80-90 9.0 9.5 10.8 10.6 90-100 10.5 10.2 10.4 Number of banks 2,557 2,797 2,863 9.0 2,866 Mean 0.0644 0.0644 0.0649 0.0641 Median 0.0643 0.0645 0.0649 0.0645 Percentile E1 1987-1988 20-50 80-90 E3 10.1% 20-50 1987-1988 Percentile E2 10.2% 0.0068 0.0059 0.0079 0.0057 E3 E4 10.5% 9.4% 10-20 10.8 10.6 8.9 9.8 20-50 30.5 30.8 30.0 28.8 50-80 29.8 28.6 31.6 30.1 80-90 9.7 9.3 10.2 10.7 90-100 9.8 10.1 10.0 10.0 2,487 2,742 2,763 2,786 0-10 Number of banks 9.4% E2 10.6% Mean 0.0697 0.0696 0.0702 0.0696 Median 0.0702 0.0700 0.0707 0.0705 JULY/AUGUST 1994 18 finding that local economic conditions were an important explanation of the performance of banks does not deny the roles of the risk-taking incentives of stockholders and managers. Ch aracteristics o f Risk- Takers We may better understand the motives for risk-taking by looking at other characteristics of the low-capital and low-earnings banks that took high risks. This section examines more closely the Groups C l, C2 and E l banks that belonged in the 90-100 percentile group in loan growth (defined as “risky banks”). The positive correlation between capital ratios and earnings mentioned above suggests the pos sibility that the risky low-capital (Groups C l and C2) banks may also be low-earnings banks (Group E l). Then, the above analyses cannot distinguish between stockholders’ and managers’ incentives. Table 14 classifies risky low-capital banks based on earnings and risky low-earnings banks based on capital ratios. In general, the percentage of risky low-capital banks belonging in Group E l was no higher than that of the entire population of banks. Compared with the composition of the population, a higher proportion of risky lowearnings banks belonged in Group C l or C2. The proportion, however, was less than 50 percent in three of the four years. Thus, the overlap between risky low-capital and low-earnings banks suggests a more significant effect of moral hazard, but does not appear to be large enough to invalidate inde pendent roles of stockholders’ and managers’ incentives. Some researchers argue that the “too-big-to-fail” policy increased risk-taking incentives for large banks because the policy further reduced the expected loss to stockholders in the event of insolvency and weakened market discipline by uninsured depositors.31 Table 15 shows the size distribution of risky banks with low capital or low earnings. The percentage of the risky lowcapital banks with $5 billion or more in total assets (Size 3 or Size 4) was no higher in general than that of the population, and it was lower for risky low-earnings banks. Thus, this analysis does not support the too-big-to-fail hypothesis. This result, of course, does not disprove the hypothesis either. It may be difficult for large banks, which already have significant market shares, to expand rapidly. Thus, the selection 31 See Mishkin (1992) and Boyd and Gertler (1994). Digitized for FEDERAL FRASER RESERVE BANK OF ST. LOUIS Table 11 Regression Results (dependent variable: change in earnings on assets) Intercept 1985 1986 1987 1988 -0.0031 -0.0038 -0.0004 -0.0013 (-12.44) State* 0.0789 (8.49) Adjusted R-Square 0.0057 (-21.64) 0.1260 (16.17) 0.0211 (-1.69) 0.0532 (-2.19) 0.0731 (6.29) (3.89) 0.0033 0.0013 (dependent variable: rate of loan growth) Intercept 1985 1986 1987 1988 0.0083 0.0195 0.0222 0.0252 (2.04) State* 2.5195 (16.34) Adjusted R-Square 0.0211 (4.44) 3.1992 (16.51) 0.0219 (0.27) 4.7001 (0.62) 2.2534 (1.67) (1.78) 0.0002 0.0002 * State = Rate of employment growth in the home state. Notes: Numbers in parenthesis are f-values. The statistical significance of regressions fluctuates widely across years mainly because of extreme values of dependent variables. of the risk measure may be responsible for the result. Nevertheless, it indicates that large banks do not account for a large proportion of those driven by risk-taking incentives. Table 16 classifies risky banks based on holding com p an y status in an attem p t to observe the rela tionship between the ownership structure and risk-taking. Compared with the composition of the entire population, risky banks were more likely to be owned by multi-bank holding com panies, less likely to be owned by one-bank holding companies, and about equally likely to be independent. One possibility is that the ownership share of managers is greatest for inde pendent banks and smallest for banks owned by multi-bank holding companies. Under this assumption, the above finding suggests that banks controlled by stockholders took more risk, and, hence, supports the moral hazard view. Another possibility is that stockholders of multi-bank holding companies are least able to monitor managers of member banks because of the complicated organizational struct ure. In 19 Table 13 Residual Rates of Loan Growth Table 12 Residual Rates of Loan Growth 1984-1985 1984-1985 Percentile C2 C1 C3 04 Percentile E1 E2 E3 E4 0-10 20.8% 12.1 10-20 15.4 8.9 7.6 8.3 32.1 32.4 20-50 30.0 30.0 29.4 30.6 30.3 31.5 27.9 50-80 19.7 32.6 34.7 33.0 12.6 9.7 7.6 80-90 6.4 11.4 12.1 10.0 10.7 13.2 8.6 9.2 90-100 7.8 10.4 10.8 11.0 3,190 6,224 2,620 3,056 3,082 3,104 3,113 0-10 28.6% 10-20 13.2 20-50 10.8% 8.3% 10.7% 8.6 9.7 22.6 24.6 50-80 14.4 80-90 10.4 90-100 Number of banks 6.8% 5.3% 7.2% Number of banks 318 Mean -0.0638 0.0137 -0.0021 -0.0038 Mean -0.0520 0.0123 0.0201 0.0189 Median -0.0841 0.0061 -0.0152 -0.0301 Median -0.0717 -0.0009 0.0061 -0.0057 Percentile C1 04 Percentile 8.8% 0-10 22.6% 8.2% 5.0% 4.7% 10-20 15.3 9.8 7.2 7.8 1985-1986 C2 11.5% 1985-1986 C3 8.4% E1 E2 E3 E4 0-10 30.8% 10-20 14.7 9.1 9.9 10.6 20-50 20.5 26.8 30.9 33.1 20-50 30.1 30.1 30.0 29.8 50-80 17.2 28.5 31.6 30.0 50-80 19.7 31.6 33.7 34.6 80-90 7.1 11.3 10.2 8.3 80-90 5.1 10.6 11.8 12.2 90-100 9.7 12.8 8.9 9.2 90-100 7.0 9.7 12.2 11.0 Number of banks 380 3,119 6,050 2,541 2,945 3,036 3,057 3,062 Mean -0.0673 0.0051 -0.0038 0.0135 Median -0.0951 -0.0101 -0.0180 -0.0264 Number of banks Mean -0.0501 0.0017 0.0200 0.0266 Median -0.0897 -0.0146 0.0033 0.0039 1986-1987 Percentile 0-10 C1 C2 27.9% 11.3% 1986-1987 C3 8.0% 04 Percentile 8.8% 0-10 E1 E2 21.1% E3 8.1% 6.0% E4 5.7% 10-20 13.9 10.3 9.5 9.8 10-20 14.0 10.8 8.2 7.3 20-50 25.8 28.1 31.1 31.0 20-50 26.8 31.1 31.0 30.8 50-80 18.2 29.8 31.5 29.8 50-80 22.3 29.8 34.6 32.7 80-90 6.7 10.1 10.2 10.3 80-90 6.6 10.2 10.6 12.3 90-100 7.6 10.6 9.8 10.3 90-100 9.1 10.0 9.7 11.1 Number of banks 552 3,048 5,597 2,370 2,722 2,924 2,956 2,975 Mean -0.1272 -0.0614 -0.0518 0.2315 Median -0.1466 -0.0759 -0.0731 -0.0761 Median Percentile C1 04 Percentile 7.9% Number of banks Mean 0.1395 -0.0554 -0.0481 -0.0254 -0.1269 -0.0776 -0.0625 -0.0581 1987-1988 C2 1987-1988 03 E1 E2 E3 E4 0-10 35.1% 0-10 24.7% 10-20 18.6 9.5 9.4 10.2 10-20 16.1 10.1 7.5 6.8 20-50 18.8 27.3 32.1 30.6 20-50 27.4 32.4 31.1 28.9 50-80 15.2 29.5 31.2 30.9 50-80 17.8 29.7 34.9 36.4 80-90 4.8 9.8 10.7 9.7 80-90 5.2 10.4 12.2 11.8 90-100 7.5 12.0 9.1 10.6 90-100 8.8 9.3 9.9 11.8 Number of banks 521 2,550 5,557 2,526 Number of banks 2,603 2,845 2,838 2,878 Mean -0.1052 -0.0138 -0.0096 0.0577 Median -0.1284 -0.0206 -0.0212 -0.0194 12.0% 7.6% 8.0% 4.3% 4.2% Mean -0.0531 -0.0114 0.0069 0.0525 Median -0.0883 -0.0250 -0.0071 0.0029 JULY/AUGUST 1994 20 Table 14 Earnings on Assets and Capital Ratios of Risky Banks Table 15 Size Distribution of Risky Banks 1984 1984 Population Population Low capital E1 E2 E3 E4 Total 3,064 3,083 3,106 3,113 12,366 (24.8%) (24.9%) (25.1%) (25.2%) 107 178 135 (27.8%) 66 (22.0%) (36.6%) 2,951 3,039 3,063 3,068 (24.3%) (25.1%) (25.3%) (25.3%) 486 (13.6%) Low capital Size 1 Size 2 Size 3 Size 4 Total 11,736 576 45 9 12,366 (94.9%) (4.7%) 408 (84.0%) Low earnings 76 (15.6%) 204 19 (8.5%) (91.5%) 1985 Population Low capital 75 (16.3%) 154 (33.5%) 155 (33.7%) 76 Population (16.5%) Low capital 1986 Population Low capital 2,734 2,928 2,960 2,982 (25.2%) (25.5%) (25.7%) 85 (21.7%) 115 (29.4%) 122 (31.2%) 69 11,604 Low capital 2,613 2,847 2,838 2,882 (25.5%) (25.4%) (25.8%) 104 (31.4%) 87 (26.3%) Low earnings 59 81 (24.5%) 11,180 331 (17.8%) Low capital Low earnings C2 C3 C4 Total 3,193 6,229 2,621 12,363 12 (5.4%) 95 (42.6%) (50.4%) (21.2%) 90 (40.4%) 26 223 Population (11.7%) 1985 Low earnings 383 3,126 6,056 2,546 (25.8%) (50.0%) (21.0%) 17 58 (33.7%) 68 (39.5%) 29 12,111 Low earnings 5,606 2,376 (4.8%) (26.3%) (48.4%) (20.5%) 62 (36.9%) Low earnings 55 (32.7%) 28 2,554 5,559 2,529 (22.9%) (49.8%) (22.6%) 78 (34.5%) (19.6%) 154 18 (0.4%) 4 (0.9%) 0 (0.0%) (0.1%) 0 460 (0.0%) 0 172 (89.5%) (10.5%) (0.0%) Size 1 Size 2 Size 3 Size 4 Total 10,935 609 51 9 11,604 (94.2%) (5.2%) 341 48 (12.3%) 161 7 (4.2%) (0.4%) 2 (0.5%) 0 (0.0%) (0.1%) 0 391 (0.0%) 0 168 (0.0%) Size 1 Size 2 Size 3 Size 4 Total 10,489 629 54 8 11,180 (93.8%) (5.6%) 293 Low earnings 36 (10.9%) 214 12 (5.3%) (0.5%) 2 (0.6%) 0 (0.0%) (0.1%) 0 331 (0.0%) 0 226 (0.0%) 11,594 168 Size 2 - between $300 million and $5 billion in total assets. Size 3 - $5 billion or more in total assets, excluding the 10 largest banks. Size 4 - the 10 largest banks. Note: The numbers in parenthesis are percentages of the Total column. 528 26 90 (16.7%) (4.7%) (11.5%) (5.0%) 366 172 1987 Population (94.5%) Size 1 - less than $300 million in total assets. 3,055 23 12,121 (16.9%) 557 (13.7%) 10 (94.7%) 1986 Population 47 (88.5%) (3.2%) (9.9%) Total 611 1987 Low capital Population Size 4 11,453 (95.8%) 320 (25.8%) 223 Size 3 (87.2%) Low earnings C1 (2.6%) 0 (0.0%) 1986 1984 Population 486 391 Population (23.4%) 0 (0.0%) 0 (0.0%) Size 2 (17.6%) 1987 Population (0.4%) (0.1%) Size 1 (79.6%) (23.6%) 2 1985 12,121 460 (0.4%) 41 81 (35.8%) 11,170 226 (18.1%) Note: The numbers in parenthesis are percentages of the Total column. Digitized forFEDERAL FRASER RESERVE BANK OF ST. LOUIS 21 Table 16 Holding Company Structure of Risky Banks Table 17 Charter Types of Risky Banks 1984 1984 Independent One bank Multi-bank Total 4,811 4,567 2,988 12,366 Population (38.9%) 134 Low capital (27.6%) Low earnings (36.9%) 150 (30.9%) 69 (30.9%) 63 (28.3%) National State Member 3,789 830 (30.6%) (6.7%) Population Undercapitalized (24.2%) 202 Low capital 91 (5.1%) 186 28 (38.3%) 223 Low earnings (40.8%) 2 (56.4%) 486 (41.6%) 22 (5.8%) 66 12 (29.6%) 1985 Independent Population 4,215 (34.8%) Low capital 138 (30.0%) Low earnings One bank 4,596 (37.9%) 138 (30.0%) 66 (38.4%) 52 (30.2%) Multi-bank 3,310 12,121 National State Member 3,733 832 (30.8%) (6.9%) Population 184 460 Undercapitalized (40.0%) 54 18 (42.9%) 172 5 (11.9%) 189 Low capital (31.4%) 32 (41.1%) Low earnings 1986 Independent One bank Multi-bank Total 3,703 4,445 3,456 11,604 (31.9%) (38.3%) (29.8%) Population Low capital 100 (25.6%) Low earnings 109 (27.9%) 53 (31.5%) 48 (28.6%) 182 168 Undercapitalized (39.9%) 19 (11.0%) National State Member 3,531 825 (30.4%) (7.1%) 19 1 (45.2%) Low capital 1987 Independent Population Low capital One bank Multi-bank Total 11,180 3,478 4,450 3,252 (31.1%) (39.8%) (29.1%) 101 (30.5%) Low earnings 90 (39.8%) 93 137 (28.1%) 49 (41.4%) (21.7%) (38.5%) 87 (62.6%) 15 39 (38.5%) 272 486 (56.0%) 145 223 (65.0%) Nonmember Total 7,556 12,121 (62.3%) 19 42 (45.2%) 239 460 (52.0%) 105 172 (61.0%) 1986 Population 67 (7.0%) 48 (27.9%) 391 (46.5%) Total 7,747 12,366 1985 Total (27.3%) (5.4%) Nonmember (2.4%) 146 31 (37.3%) Low earnings (7.9%) 53 12 (7.1%) (31.5%) 331 Nonmember Total 7,248 11,604 (62.5%) 22 42 (52.4%) 214 391 (54.7%) 103 168 (61.3%) 1987 226 Population Undercapitalized National State Member 3,401 825 (30.4%) (7.4%) 14 (37.8%) Independent - banks not owned by a holding company. One bank - banks owned by a holding company that owns one bank. Low capital Multi-bank - banks owned by a holding company that owns more than one bank. Low earnings 104 (31.4%) 57 (25.2%) 2 (5.4%) 24 (7.3%) 17 (7.5%) Nonmember Total 6,954 11,180 (62.2%) 37 21 (56.8%) 203 331 (61.3%) 152 226 (67.3%) Note: The numbers in parenthesis are percentages of the Total column. National - national banks. State member - state banks that are members of the Federal Reserve System. Nonmember - state banks that are not members of the Federal Reserve System. Note: The numbers in parenthesis are percentages of the Total column. JULY/AUGUST 1994 22 this case, managerial incentives better explain aggressive risk-taking by banks owned by m ulti bank holding companies. It will require more extensive study to draw a clear conclusion. Unfortunately, the scarcity of data on the owner ship structure of the majority of banks makes the task difficult. sistent with the behavior of a subset of banks. Aggressive risk-taking may not have been a preva lent phenomenon. The finding that deliberate risk-taking was confined to a subset of banks, however, does not rule out the possibility that increased risk-taking was an important cause of a large number of bank failures in the 1980s. The enforcement of regulation may also affect risk-taking. Table 17 reports the number of risky banks for each type of charter. A notable pattern is that the proportions of risky undercapitalized (Group C l) and low-capital banks (Groups C l and C2) were relatively high among national banks throughout the four years examined by this study. No apparent pattern is found for risky low-earnings banks. The condition of local economies is found to have been important for explaining the performance of banks. This finding implies an important role for external economic events but does not disprove the role of deliberate risk-taking. Adverse economic conditions may have been largely responsible for the overall deterioration of the financial health of the banking industry during the 1980s. Yet delib erate risk-taking may have played a significant role in many bank failure cases. It appears that all of the three factors discussed by this article contributed to the banking problems of the 1980s. SUM M ARY There are numerous explanations for the deteri oration of the asset quality of banks during the 1980s. Moral hazard theories explain the problem on the basis of increased risk-taking incentives of bank stockholders, which arise from limited lia bility and the existence of deposit insurance. Deregulation in the early 1980s, which increased both intra- and interindustry competition, increased stockholders’ incentives to take risk by reducing the charter values of banks. Another possible explanation for increased riskiness in banking is desperate profit-seeking by bank managers to preserve their jobs. Increased competition and dwindling profit opportunities sharply lowered profits of the banks managed by incompetent managers. To make profits acceptable to stock holders, incompetent managers needed to increase risks and hope for a good outcome. W hile these two explanations blame deliberate risk-taking, it is also possible that unexpected external events impaired the financial structure of banks. In other words, although banks did not change their behavior, a sequence of adverse economic events such as the collapse of real estate markets under mined the financial strength of banks. This study indicates the presence of ex ante risk-taking by both bank stockholders and man agers, but evidence is stronger for moral hazard. A main finding is that risk measures for banks with low capital or earnings are distributed heavily toward both tails. In other words, while many banks with low capital or earnings refrained from taking risk, many other banks with similar char acteristics adopted highly risky strategies. These findings suggest that moral hazard and desperate profit seeking by incompetent managers are con Digitized for FEDERAL FRASER RESERVE BANK OF ST. LOUIS REFEREN CES Allen, Linda, and A. Sinan Cebenoyan. “Bank Acquisitions and Ownership Structure: Theory and Evidence,” Journal of Banking and Finance (April 1991), pp. 425-48. Barth, James R., R. Dan Brumbaugh, Jr., and Daniel Sauerhaft. “Failure Costs of Government-Regulated Financial Firms: The Case of Thrift Institutions,” Federal Home Loan Bank Board Research Working Paper No. 123 (October 1986). Benston, George J., and Mike Carhill. “FSLIC Forbearance and the Thrift Debacle,” Proceedings o f the 28th Annual Conference on Bank Structure and Competition (Federal Reserve Bank of Chicago, 1992), pp. 121-50. Bernanke, Ben S., and Cara S. Lown. ‘The Credit Crunch,” Brookings Papers on Economic Activity (1991, No. 2), pp. 205-47. Boyd, John H., and Mark Gertler. “The Role of Large Banks in the Recent U.S. Banking Crisis,” Federal Reserve Bank of Minneapolis Quarterly Review (winter 1994), pp. 2-21. Calomiris, Charles W. “Deposit Insurance: Lessons from the Record,” Federal Reserve Bank of Chicago Economic Perspectives (May/June 1989), pp. 10-30. Cline, William R. International Debt: Systemic Risk and Policy Response. MIT Press, 1984. Emmons, William R. “Increased Risk-Taking versus Local Economic Conditions as Causes of Bank Failures,” Proceedings of the 29th Annual Conference on Bank Structure and Competition (Federal Reserve Bank of Chicago, 1993), pp. 189-209. Federal Deposit Insurance Corporation. Annual Report (1991). Flood, Mark D. “Deposit Insurance: Problems and Solutions,” this Review (January/February 1993), pp. 28-34. Gilbert, R. Alton. “Market Discipline of Bank Risk: Theory and Evidence,” this Review (January/February 1990), pp. 3-18. _____ . “Supervision of Undercapitalized Banks: Is There a Case for Change?” this Review (May/June 1991), pp. 16-30. Gorton, Gary, and Richard Rosen. “Corporate Control, Portfolio Choice, and the Decline of Banking,” Federal Reserve Board Finance and Economic Discussion Series No. 215, Division of Monetary Affairs (December 1992). 23 Gunther, Jeffery W., and Kenneth J. Robinson. “Empirically Assessing the Role of Moral Hazard in Increasing the Risk Exposure of Texas Banks,” Financial Industry Studies Research Paper No. 4-90, Federal Reserve Bank of Dallas (October 1990). Houston, Joel, and Christopher James. “Bank Compensation, Turnover and Risk-Taking,” MidAmerica Institute Research Paper (January 1993). Jensen, Michael C. “Agency Costs of Free Cash Flow, Corporate Finance, and Takeovers,” The American Economic Review (May 1986), pp. 323-9. Johnson, Ronald. ‘T he Bank Credit Crumble,” Federal Reserve Bank of New York Quarterly Review (summer 1991), pp. 40-51. Keeley, Michael C. “Deposit Insurance, Risk, and Market Power in Banking,” The American Economic Review (December 1990), pp. 1183-200. MacKie-Mason, Jeffrey K. “Do Firms Care Who Provides Their Financing?” in Glenn R. Hubbard ed., Asymmetric Information, Corporate Finance, and Investment. University of Chicago Press, 1990, pp. 63-103. Marcus, Alan J. “Deregulation and Bank Financial Policy,” Journal of Banking and Finance (December 1984), pp. 557-65. McKenzie, Joseph A., Rebel A. Cole, and Richard A. Brown. “Moral Hazard, Portfolio Allocation, and Asset Returns for Thrift Institutions,” Journal o f Financial Services Research (April 1992), pp. 315-39. Merton, Robert C. “An Analytic Derivation of the Cost of Deposit Insurance and Loan Guarantees,” Journal of Banking and Finance (June 1977) pp. 3-11. Mishkin, Frederic S. “An Evaluation of the Treasury Plan for Banking Reform,” Journal o f Economic Perspectives (winter 1992), pp. 133-53. Miller, Merton H., and Kevin Rock. “Dividend Policy under Asymmetric Information," Journal o f Finance (September 1985), pp. 1031-51. Myers, Stewart C., and Nicholas S. Majluf. “Corporate Financing and Investment Decisions When Firms Have Information That Investors Do Not Have,” Journal of Financial Economics (June 1984), pp. 187-221. Park, Sangkyun. “Banking Panics in U.S. History: Causes and Policy Responses,” in George Kaufman, ed., Research in Financial Sen/ices: Private and Public Policy. JAI Press, 1993. Peek, Joe, and Eric S. Rosengren. “The Capital Crunch in New England,” Federal Reserve Bank of Boston New England Economic Review (May/June 1992), pp. 21 -31. Saunders, Anthony, Elizabeth Strock, and Nickolaos G. Travlos. “Ownership Structure, Deregulation, and Bank Risk Taking,” Journal o f Finance (June 1990), pp. 643-54. Sinkey, Jr., Joseph F. Problem and Failed Institutions in the Commercial Banking Industry. JAI Press, 1979. Wheelock, David C. “Is the Banking Industry in Decline? Recent Trends and Future Prospects from a Historical Perspective,” this Review (September/October 1993), pp. 3-22. JULY/AUGUST 1994 25 Patricia S. Pollard Patricia S. Pollard is an economist at the Federal Reserve Bank o f St. Louis. Richard D. Taylor provided research assistance. Trade Between the United States and Eastern Europe HROUGHOUT MOST OF THE post-World War II era, trade between the United States and Eastern Europe was minuscule. The United States maintained high tariff barriers on imports from most Eastern European countries and also restricted its own exports to these countries. In particular, the United States prohibited the export to these countries of high-technology goods related to national security interests. Eastern Europe also maintained various trade restrictions on imports from the United States. Most Eastern European trade was controlled by the state and conducted within the Council for Mutual Economic Assistance (CMEA), the trade organization of the Soviet bloc countries. With the disintegration of the Soviet system and the collapse of the CMEA trading bloc, Eastern European countries began to re-orient their trade to the West. As these countries undertook political and economic reforms, the United States reduced its tariff restrictions on their products. Consequently, trade between the United States and Eastern Europe has expanded substantially since 1988. This paper examines the growth and pattern of trade between the United In January 1993, the Czech and Slovak Federal Republic split into two independent countries: the Czech Republic and the Slovak Republic. With the exception of total export and import data, the data used in this paper end before the split occurred. States and the three Eastern European countries which have made the greatest progress in adopt ing market reforms: the Czech and Slovak Federal Republic (CSFR), Hungary and Poland.1 Studies have shown that the U.S. economy is likely to be one of the principal beneficiaries of economic liberalization in Eastern Europe.2 U.S. exports to, and investment in, the region should increase as the restructuring of the economies of Eastern Europe results in an increase in demand for capital goods and technology, and opens new markets for U.S. products. Such gains will be limited, however, if the Eastern European countries reverse the pattern of opening their markets and raise protectionist barriers against products from the United States. Despite the initial steps taken to reduce trade barriers on Eastern European products, the United States maintains quantitative restrictions and other forms of protectionism on many prod ucts from Eastern Europe. Most significantly, the United States maintains a high degree of protection against the importation of textiles and apparel, chemicals, steel and agricultural products from Eastern Europe. These goods See, for example, Wang and Winters (1992). JULY/AUGUST 1994 26 Table 1 Growth in U.S. Trade 1988-931 U.S. imports ($ millions) U.S. exports (S millions) 1988 1993 Growth 1988 1993 Growth CSFR2 $87.6 $341.5 289.9% $55.1 $300.1 444.2% Hungary 293.9 400.5 36.3 77.5 433.9 459.7 Poland 375.6 454.0 20.2 303.7 916.5 201.8 Combined CSFR, Hungary, Poland 759.0 1,196.0 57.6 436.3 1,650.5 278.3 441,282.4 580,054.4 31.6 322,718.3 464,767.2 44.0 World 1 Based on nominal dollar values. 2 The 1993 data for the CSFR were calculated by combining the data for the Czech Republic and the Slovak Republic. SOURCE: U.S. Department of Commerce, Bureau of the Census are produced by the sectors in which Eastern Europe is most competitive. The possibility exists for an increase in protectionism in Eastern Europe as these countries have become increasingly frustrated by the lack of progress in securing access to U.S. as well as other Western markets for their products. How the problems stemming from these trade barriers are handled will be an important determinant of future trade flows between the United States and the CSFR, Hungary and Poland. This paper describes the recent changes in these trade flows and examines the restrictions facing Eastern Europe in its trade with the United States. The structure of the paper is as follows. Section two provides an overview of trade between the United States and the CSFR, Hungary, and Poland. The causes of the recent growth in trade between the United States and Eastern Europe are examined in section three. The product composition of this trade is dis cussed in section four. Section five examines U.S. restrictions on the products in which the CSFR, Hungary and Poland have their greatest comparative advantage. The conclusions are presented in section six. 3 The European Union (Belgium, Denmark, France, Germany, Greece, Ireland, Italy, Luxembourg, the Netherlands, Portugal, Spain and the United Kingdom) remains the major trading partner for these three Eastern European countries, accounting for more than half of their exports and imports. 4 Exports from the United States to the three countries com bined grew at a faster rate than exports from the European Union (208 percent compared to 180 percent) between 1988 and 1992 (the latest year for which data for the European FEDERAL RESERVE BANK OF ST. LOUIS OVERVIEW OF TRADE Trade with the CSFR, Hungary and Poland has always comprised a low percentage of the total international trade of the United States. Neither U.S. exports to these countries nor imports from any of the three constitute more than 1 percent of total U.S. exports or imports. From the perspective of the CSFR, Hungary and Poland, however, trade with the United States constitutes a larger share of the international trade of each country.3 Despite its relatively small size, there has been a substantial expansion in trade between the United States and the CSFR, Hungary and Poland following the disintegration of the Soviet bloc. In dollar terms, U.S. imports from the three grew by 58 percent between 1988 and 1993 while U.S. exports to these three countries grew by 278 percent (see Table 1, and Figures 1 and 2).4 In comparison, total U.S. imports increased by 32 percent between 1988 and 1993 whereas total U.S. exports rose by 44 percent. U.S. exports to the CSFR, Hungary and Poland have grown much faster than imports from these countries. Consequently, in 1988 the United Union were available). U.S. imports from the three Eastern European countries, however, grew more slowly than European Union imports over the same time period (27 per cent compared to 133 percent). 27 Figure 1 Total Annual U.S. Imports from Eastern Europe Millions of dollars 500 450 400 350 300 250 200 - 150 100 - 50 0 - 1988 ' 1989 ' 1990 ' 1991 ' 1992 ' 1993 SOURCE: U.S. Department of Commerce, Bureau of the Census Figure 2 Total Annual U.S. Exports to Eastern Europe Millions of dollars SOURCE: U.S. Department of Commerce, Bureau of the Census JULY/AUGUST 1994 28 The Jackson-Vanik Amendment to the Trade Act of 1974 The Trade Expansion Act of 1951 suspended the most-favored-nation (MFN) status of any country deemed to be “under the control of the world Communist movement.” Between 1951 and 1974, the MFN status of Poland and Yugoslavia was restored. The U.S. Trade Act of 1974 reasserted the determination of the United States to withhold MFN status from Communist countries. Section 402 of the Trade Act states that MFN status could not be granted to any nonmarket economy country which did not already have such status. The Act established two exceptions, however, through which MFN status would be granted: 1) the President reported to Congress that the country did not restrict the emigration of its citizens; and 2) the President waived com pli ance with the freedom of emigration require ment. The Jackson-Vanik amendment estab lished the freedom-of-emigration requirement. This requirement could only be waived if the President determined that doing so would “promote the objective of the amendment.”1 MFN status still could not be granted until the United States had concluded a bilateral trade agreement with the country in question. Furthermore, as is standard, the granting of MFN status was to be done on a reciprocal basis. Both the waiver of the Jackson-Vanik amendment and the trade agreement had to be approved by a congressional resolution. Waivers were granted for a one-year period, ending July 3, but could be renewed each year by the President. Following the renewal of the waiver, Congress was given 30 days to pass a resolution eliminating the waiver. Between 1974 and 1988, no country was certified to have met the Jackson-Vanik requirement. Under a waiver of this amend ment, however, three countries were granted MFN status: Romania in 1975, Hungary in 1978, and China in 1980. In October 1989, Hungary was accorded permanent MFN status following the passage of a freedom-of-emigration law. The Czech and Slovak Federal Republic was granted waiver status in November 1990, and permanent MFN status in October 1991 in accordance with the Jackson-Vanik amendment. Although such status was referred to as “permanent,” the Jackson-Vanik amendment required that the President report to Congress semiannually on a country’s compliance with the emigration criteria. Countries which were not subject to the Jackson-Vanik requirements (such as Poland) had unconditional MFN status. In 1992, President Bush asked Congress to remove Hungary and the CSFR from the Jackson-Vanik restrictions, and these countries were granted unconditional MFN status. 1 United States International Trade Commission (June-July 1990, p. 7). States ran a trade deficit with all three, but by 1993 the deficit had turned to a trade surplus with Hungary and Poland. CAUSES OF GROWTH IN U .S .EA STERN EUROPEAN TRADE The recent growth in U.S. trade with Eastern Europe is due in part to a general increase in trade between the former nonmarket economy countries and the West. The collapse of intraDigitized forFEDERAL FRASER RESERVE BANK OF ST. LOUIS CMEA trade due to a movement to market pricing and the settlement of accounts in convertible currencies rather than the system of official exchange rates tied to the Soviet ruble, also played a part in re-orienting trade to the West. As these countries sought to modernize their production processes, they looked to the West as a source of capital and technology. Exports to the West provided a source of foreign currency with which to purchase these capital goods and were also a source of economic growth following 29 the shrinkage of the domestic market resulting from the collapse of the old economic system.5 REDUCTION IN U .S. TRADE B A RRIER S FACING EA STERN EUROPE The growth in trade between the United States and the CSFR, Hungary and Poland also reflects the reduction of trade barriers by all parties. In the initial euphoria following the collapse of the Soviet bloc, the United States pledged to open its markets to Eastern European products in order to aid these countries in developing a market system. One of the first steps the United States took to encourage reform in Eastern Europe was to grant most-favored nation (MFN) status to these coun tries, leading to a substantial reduction in the tariff rates on imports from Eastern Europe.6 Poland was originally granted MFN status in 1960, but this status was suspended in January 1981, following the imposition of martial law. It was not until November 1989 that Poland regained its MFN standing. Hungary was granted MFN status in 1978 under the waiver provision to the Jackson-Vanik Amendment to the 1974 Trade Act (see shaded insert on the opposite page for more on this amendment). The CSFR was the last of the three countries to gain MFN status, in November 1990. The importance of MFN status can be illustrated with reference to the textile and apparel industry. The MFN tariff rates on U.S. imports of textile and apparel range from 20 per cent to 35 percent. In contrast, non-MFN tariff rates range from 50 percent to over 100 percent.7 Additional changes in the U.S. treatment of imports from these countries have occurred as well. In November 1989, Hungary was designated as a “beneficiary developing country,” making it eligible for tariff reductions granted under the 5 Although output data for the former nonmarket economies are not totally reliable, estimates by the International Monetary Fund (1993) indicate that between 1988 and 1992, the economy of the CSFR shrank by 23 percent, the Hungarian economy shrank by 21 percent, and the Polish economy shrank by 16 percent. generalized system of preferences (GSP).8 In January 1990, Poland was deemed eligible for the GSP and in April 1991 the CSFR was deemed eligible. As part of the Trade Enhancement Initiative for Central and Eastern Europe (TEI) undertaken by the Bush administration in 1991, the United States expanded the list of products for which tariff reductions are granted to GSP countries to include products proposed by the CSFR, Hungary and Poland. The United States also concluded bilateral trade agreements with each country, increasing their import quotas on textiles and apparel.9 REDUCTIONS IN EASTERN EUROPEAN TRADE BA R R IER S FACING THE UNITED STATES The Eastern European countries also sharply reduced their barriers to imports. In the CSFR, most quantitative restrictions on imports were eliminated or converted into tariffs.10 The average unweighted tariff rate was 5 percent until January 1992, when the CSFR requested and received GATT approval to raise its average tariff rate to 6 percent." Hungary has an average unweighted tariff rate of 13 percent on imports in addition to a 2 percent customs clearance fee, while the average tariff on imports in Poland is 13.6 percent. Both countries have also eliminated most quan titative restrictions, although Hungary does maintain quotas on some consumer goods, while Poland limits imports of some alcoholic beverages.12 In comparison, the United States maintains a 6.8 percent average tariff rate on imports, while the tariff rate for the European Union and Japan is 6.5 percent.13 All of these entities also maintain nontariff barriers. Furthermore, the tariff rates in the CSFR, Hungary and Poland are lower than most coun tries at a comparable stage of development.14 8 The Generalized System of Preferences is a program whereby developed countries grant preferential tariff rates to products from developing countries. It is allowed under GATT rules as an exception to the MFN principle. The United States first granted preferences as part of the Trade Reform Act of 1974. As noted in the text, not all products are covered under the GSP. 6 MFN status guarantees that the tariffs imposed on a coun try’s products will be no higher than those imposed on the imports of any other country. MFN tariff rates have been reduced substantially through successive trade rounds held under the General Agreement on Tariffs and Trade (GATT). For the period covered in this paper, the average MFN tariff rate on manufactured goods imported into the United States was 4 percent. In contrast, non-MFN tariff rates are set by the Smoot-Hawley Tariff Act of 1930. 13 USITC (August 1991, p. 6). 7 Erzan and Holmes (1992, p. 4). 14 Rodrik (1992, p. 2). 9 Quotas set a limit on the quantity of a product which a coun try can sell to another country. 10 OECD (1991, p. 84). 11 Green (February 6, 1992). 12 Rodrik (1992, pp. 3-4). JULY/AUGUST 1994 30 Tariff rates in the CSFR, Hungary and Poland also tend to be lowest on capital goods and raw materials, the major U.S. export products to these countries. In contrast, as discussed below, trade restrictions in the United States are highest on the goods for which the three Eastern European countries have a comparative advantage. Table 2 Major Product Composition of U.S. Exports, by End-Use Category (percent of total) CSFR PRODUCT COMPOSITION OF TRADE More than half of all U.S. merchandise exports to the CSFR and Hungary, and slightly less than one-half of U.S. exports to Poland, are capital goods (Table 2). Although capital goods were one of the largest categories of U.S. exports to the CSFR, Hungary and Poland in both 1988 and 1992, there was a clear shift during this period from industrial supplies and materials to capital goods. Put simply, there was an increase in the demand for capital due to industrial restructuring. Another factor contributing to the shift toward imports of capital goods is the easing of the Coordinating Committee on Multilateral Export Controls (COCOM) restrictions. COCOM was created in 1949 to control the exportation to the Soviet bloc countries of products and technology which could be used for military purposes.15 The importance of the relaxation of COCOM restrictions is highlighted by the growth in U.S. exports of computers, semiconductors and telecommunications equipment—high-technology industries, relying heavily on research and development conducted by highly skilled workers. Such exports grew from 4.9 percent to 20.3 percent of total exports to the CSFR, from 4.4 percent to 12.4 percent for Hungary, and from 1.0 percent to 10.9 percent of total U.S. exports to Poland. In contrast, the CSFR, Hungary and Poland are countries whose productive resources are characterized by relatively large amounts of semiskilled labor, and all suffer from a lack of up-to-date capital. These factors, in combination with their relatively low-wage rates, point to production cost advantages in products requiring large amounts of semiskilled labor. The product composition of U.S. imports from the CSFR, Hungary and Poland does fit this pattern (Table 3). In 1992, consumer goods, particularly apparel 15 COCOM was disbanded on April 1, 1994. The members of COCOM were Australia, Belgium, Canada, Denmark, France, Germany, Greece, Italy, Japan, Luxembourg, the Netherlands, Norway, Portugal, Spain, Turkey, the United Kingdom and the United States. Digitized forFEDERAL FRASER RESERVE BANK OF ST. LOUIS Hungary Poland 1988 1992 1988 1992 1988 1992 Food & beverages 0.8 3.9 0.7 2.1 38.1 15.4 Industrial supplies 44.2 4.3 37.8 8.9 24.3 7.9 Capital goods 46.2 73.8 40.3 61.5 15.2 45.8 2.7 11.4 7.0 1.3 2.7 Automotive 0 Consumer goods 5.8 12.0 7.9 17.0 7.8 13.0 Exports, n.e.c. 3.1 3.3 1.9 3.4 13.4 15.2 and footwear, accounted for the largest category of U.S. imports from each country. The CSFR and Poland have increased their exports of capital goods to the United States, although these goods are not high-technology products. For the CSFR and Poland, nearly all capital goods exported to the United States are nonelectrical machinery and parts. Within this group, industrial and agricultural machinery, and machine tools are the most important. A more formal way to analyze the exports of a country is to calculate an index of relative com parative advantage (RCA). This index is calcu lated as follows: where X n are exports of commodity n; i is the country of origin; j is the country of destination; and - j is the rest of the world (all countries excluding country j ) . Equation 1 indicates that the relative comparative advantage of country i in any good n depends on the share of that good in country i ’s exports to country j relative to the share of good n in the rest of the world’s exports to country;. In general, if this ratio is greater than 1, then country i has a comparative advan tage in producing that product relative to the rest of the world.16 16 For more details on this index and its use in determining the comparative advantages of Eastern European countries, see Murrell (1990) and USITC (1991). 31 categories each of the three Eastern European countries has the greatest relative comparative advantage, and also to look at changes in each country’s comparative advantage as each has ini tiated the transition to a market economy.18 Table 3 Major Product Composition of U.S. Imports (percent of total) CSFR Hungary 1988 1992 1988 1992 Poland 1988 1992 Food & beverages 5.6 4.1 20.5 17.0 35.9 15.0 Industrial supplies 21.6 23.2 20.4 17.2 26.9 29.3 Capital goods 16.6 22.2 6.3 13.5 10.5 18.3 Automotive Consumer goods Exports, n.e.c. 5.0 3.8 13.9 14.5 0.8 1.4 49.3 43.1 38.2 36.8 24.7 33.9 2.0 3.6 0.8 1.0 1.2 2.2 Table 4 shows the relative comparative advantage indexes for each of the three Eastern European countries, by principal end-use category of exports in 1988 and 1992, based on U.S. Bureau of the Census data. Appendix tables provide the RCAs for each country using five-digit, end-use cate gories in each year from 1988 to 1992. CHANGES IN RELATIVE COMPARATIVE ADVANTAGE Two major developments occurred between 1988 and 1992 that may have affected the rela tive comparative advantages of the Eastern European countries. The first was the progress made in moving from a command system of pro duction to a market-oriented one. Producer and consumer prices were decontrolled, government subsidies to industry were reduced or in many cases eliminated, and privatization programs were implemented. These measures should eventually result in more efficient production leading to an index of comparative advantage more directly related to market forces. For example, in 1992 the CSFR exported $242 million of merchandise to the United States, with shipments of footwear accounting for $14 m illion of this total. In contrast, world merchan dise exports to the United States totalled $532 billion in 1992, and footwear exports accounted for $7 billion of this total. Thus, whereas footwear comprised nearly 6 percent of the merchandise exports of the Czech and Slovak Federal Republic to the United States, it accounted for slightly more than 1 percent of world exports to the United States. Since the share of footwear in the CSFR’s exports to the United States was larger than the share of footwear in world exports to the United States, the CSFR is said to have a relative comparative advantage in this product (RCAfootwear> l) .17 If the share of footwear in the CSFR’s exports to the United States had been smaller than the share of footwear in world exports to the United States, the CSFR would have a relative comparative disadvantage in this product (RCAfoo,wear< l). The second development was the easing of trade restrictions by the United States. As noted above, only Hungary enjoyed MFN status in its trade with the United States in 1988, and none of the three countries was considered eligible for GSP status.19 By 1992, all three had both MFN and GSP status, as well as increased quotas for textiles and apparel. The relaxation of these restrictions should allow the computed relative comparative advantage to more accurately reflect the true comparative advantage of each country. Using the index of relative comparative advan tage, it is possible to determine in which product Despite these two developments, the evidence presented in Table 4 and the Appendix does not 17 More precisely, the index of relative comparative advantage for footwear from the CSFR is 14,270,978 7,294,287,012 ------------------:----------------------------= .05895 242,077,791 532,017,422,033 „ .01371 = 4.3 18 Another standard method used is revealed comparative advantage, which is calculated by x; -/w; x; +m; ’ (1990) and Collins and Rodrik (1991) to calculate the com parative advantages of the Eastern European countries in trade with the West prior to economic liberalization. Both of these studies calculate the index of comparative advantage only for major product categories, but their results are similar to the results based on 1988 data used in this paper. 19 The United States did not extend GSP benefits to the Soviet bloc countries. the difference between country i ’s exports of good n and its imports of good n divided by the sum of country i ’s exports and imports of good n. This index was used by Fieleke JULY/AUGUST 1994 32 Table 4 Relative Comparative Advantage Indexes: By Principal End-Use Category CSFR Hungary Poland Category 1988 1992 1988 1992 1988 1992 0 1.00 0.78 3.64 3.26 6.41 2.87 Agricultural 00 1.41 1.07 5.14 4.51 8.44 3.49 Nonagricultural 01 0.05 0.05 0.18 0.10 1.70 1.28 1 0.81 0.89 0.76 0.66 1.01 1.13 Energy products 10 0.00 0.00 0.02 0.02 0.00 0.04 Paper & paper products 11 0.04 0.01 0.01 0.00 0.00 0.00 4.82 Foods, feeds & beverages Industrial supplies & materials 12000 & 121 4.81 7.45 5.48 1.55 3.74 Chemicals, excluding medicinals 125 0.40 1.27 0.97 1.16 0.63 1.31 Building materials, except metals 13 0.52 0.20 0.02 0.00 0.08 0.39 Metals 14 2.19 0.73 0.64 0.39 1.89 0.75 Textile supplies & related materials Metallic products 15 0.81 0.80 2.50 4.07 2.54 4.75 Nonmetallic minerals & nonmetallic prods. 16 0.73 0.48 0.53 1.07 0.23 0.14 2 0.72 0.88 0.27 0.54 0.46 0.72 Capital goods except automotive Electric generating machinery, electric 20 0.31 0.07 0.31 1.29 0.12 0.79 Nonelectric incl. parts & attachments 21 0.93 1.11 0.30 0.50 0.48 0.74 Transportation equipment, except auto. 22 0.00 0.06 0.01 0.05 0.68 0.51 Automotive vehicles, parts & engines 3 0.25 0.22 0.70 0.84 0.04 0.08 Consumer goods (nonfood), except auto. 4 2.25 1.87 1.75 1.59 1.13 1.46 Consumer nondurables, manufactured 40 2.68 1.84 2.72 2.29 1.49 1.62 Consumer durables, manufactured 41 1.59 1.44 1.08 1.06 0.92 1.45 Unmanufactured consumer goods 42 4.68 6.17 0.01 0.02 0.12 0.08 50 0.68 1.09 0.26 0.31 0.43 0.65 apparatus & parts Exports not elsewhere counted show major shifts in relative comparative advan tage between 1988 and 1992. In general, the product categories in which a country exhibited a relative comparative advantage (RCA>1) in 1988 are the same as those in 1992, and vice versa.20 Furthermore, Hungary, which had made the most progress toward reforming its economy at the start of 1988 and faced the lowest tariffs on its exports to the United States of the three countries, had no fewer shifts in its relative comparative advantage (movements from RCA>1 to RCA<1, and vice versa) between 1988 and 1992 than the CSFR or Poland. The lack of major shifts in relative comparative advantage is not surprising given the years needed to restructure the production of the formerly command-based economies. Such restructuring could change the pattern of comparative advan tage of the Eastern European countries. There is some evidence, however, that the estimates given in this paper may be close to reflecting the com parative advantage of each country which will prevail after the transition to a market-based system is completed within the CSFR, Hungary and Poland. The product composition of each country’s trade with the West was different from 20 Another method to determine changes in relative compara tive advantage is to compute a rank correlation coefficient for each country. A rank correlation coefficient of 1 indicates no change in the ordering of industries by the RCA index between 1988 and 1992, zero indicates no relationship between the 1988 ordering and the 1992 ordering, and -1 indicates a complete reversal in the ordering. The rank correlation coefficients are .74 for the CSFR, .77 for Hungary, and .63 for Poland, using the five-digit product categories. Digitized forFEDERAL FRASER RESERVE BANK OF ST. LOUIS 33 its trade with the other CMEA members. As noted by Collins and Rodrik (1991), Eastern Europe’s trade with the West was less distorted than its intra-CMEA trade and, thus, more likely to be reflective of its comparative advantage.21 With respect to intra-CMEA trade, the products traded and their prices were set through bilateral government agreements.22 In contrast, products exported to Western countries and the prices of these products were subject to international competition. An OECD study of Hungary found that even prior to 1989, Hungary based its trade with the West on its comparative advantage.23 There is little evidence that the CSFR, Hungary or Poland have redirected their intra-CMEA sales to the West.24 Relative Comparative Advantage in 1992 The one-digit, end-use categories show that the CSFR, Hungary and Poland exhibited a relative comparative advantage in consumer goods in 1992. The latter two countries also had a com parative advantage in foods, feeds and beverages. On a more disaggregated basis, all three coun tries in 1992 exhibited a relative comparative advantage in agricultural goods, textile supplies & related materials, chemicals (excluding medicinals), and manufactured consumer goods. Among consumer goods, all three countries had a relative comparative advantage in some type of textile apparel & household goods made from textiles, and footwear.25 In addition, Hungary and Poland had a relative comparative advantage in m etallic products. In summary, these three Eastern European countries exhibit a relative comparative advantage in agricultural products, chemicals, textiles, apparel and footwear, with the exception of the CSFR in metallic products.26 This pattern of comparative advantage fits the typical pattern for developing countries.27 21 Collins and Rodrik (1991, p. 50). 22 OECD (1992, p. 83). 23 OECD (1993, p. 91). 24 See, for example, Rodrik (1992, p. 18) and OECD (1993, pp. 91-3). 25 See the Appendix for details. 26 These findings are supported by changes in prices relative to the overall producer price indexes in Hungary and Poland between 1989 and 1991. In conjunction with the liberaliza tion of prices, reductions in subsidies, and the progress made in making their currencies convertible, Hungary and Poland both experienced a decline in the prices of textiles, If these sectors do indeed represent the compar ative advantage of the CSFR, Hungary and Poland, one would expect to see further increases in the export to the United States of these products. Furthermore, as these countries become more adept at marketing and supplying goods for export, trade in these products should increase. In actuality, however, the potential for increased exports to the United States of the products in which the three have a comparative advantage is limited by the fact that these goods fall into the “sensitive sectors” categorization. These are products typically produced by sectors in decline and are highly protected from interna tional competition. U .S. TRADE RESTRICTIONS ON EA STERN EURO PEAN PRODUCTS The initial emphasis in the United States on opening its markets to Eastern European products has given way to protectionist sentiments as the CSFR, Hungary and Poland have shown that they can compete successfully with certain Western industries and, as a consequence, imports to the United States from these countries have expanded. As noted above, many sectors in which Eastern Europe is most competitive are highly protected in the United States. For example, the textile, apparel and footwear industries enjoy the highest level of tariff protection of all U.S. manufacturing industries. Tariff rates for textiles and apparel average 18 percent ad valorem, while tariffs on certain footwear products range as high as 40 percent.28 The U.S. dairy industry also enjoys a high level of tariff protection, with rates ranging from 10 percent to 25 percent.29 Furthermore, none of these products are eligible for GSP tariff reductions. In addition to the tariff barriers facing Eastern Europe on the products in which it has a com parative advantage, most of these products are subject to nontariff barriers. For example, the clothing, leather and metal products relative to their producer price indexes. Hungary also saw a drop in the relative price of food, while Poland experienced a fall in the relative price of chemicals. The decline in the relative prices of these products was due primarily to the availability of lower cost inputs and an increase in production efficiency. See the Organization for Economic Co-operation and Development (1993) for the details of the relative price changes. 27 Bank for International Settlements (1993, p.70) and Collins (1991, p. 223). 28 For a further discussion of U.S. tariff protection, see Finger (1992) and Ray (1991). 29 U.S. Harmonized Tariff Schedule 1993. JULY/AUGUST 1994 34 United States maintains quotas on textiles, apparel and some agricultural products, most notably dairy products, a category in which the CSFR, Hungary and Poland have a relative com parative advantage. In addition, quotas on steel exports to the United States from these countries, as well as from many others, were in place during most of the period covered in this paper. Although many of the quotas which apply to the CSFR, Hungary and Poland remain underutilized, they still may act as an effective restraint on trade. Quotas are not applied by major product category but to specific products. Thus, rather than setting a limit on the importation of wool clothing, the United States places limits on specific types of wool clothing— for example, m en’s and boy’s wool suit-coats. The quota limit on a specific product may be so low that it is not profitable to export such a small quantity. Furthermore, the quota agreement may require that exports be spread out over each year, further limiting the profitability of trade. For example, the agreement limiting the export of Polish steel to the United States prohibited Poland from exporting more than 60 percent of its yearly quota allotment in any two consecutive quarters.30 The 1992 textile agreement between the United States and the CSFR requires the latter to space its exports to the United States “within each category evenly throughout each agreement period.”31 The United States has restricted the flow of certain goods simply by threatening to place limits on their importation. For example, the most recent textile agreement between the United States and Poland lists products for which no quotas are set, but for which the United States reserves the right to “consult” with the govern ment of Poland to restrain the trade of these products if the United States believes such products are causing “market disruption” or the “risk of market disruption.” The agreement even sets limits on imports of these products while consultations are in progress.32 Another method 30 USITC (March 1990, p. 11-8). 31 United States Department of State (July 21, 1992, p. 4). 32 See U.S. Department of State (January 28, 1992, pp. 8-10). 33 Dumping is the practice of selling a product in foreign mar kets at a lower price than in the home market, or at price below the cost of production. For an analysis of the use of anti-dumping regulations as a protectionist device, see Coughlin (1991). 34 The United States is not alone in restricting access to prod ucts from Eastern Europe. The European Union, despite concluding association agreements with the CSFR, Hungary RESERVE BANK OF ST. LOUIS FEDERAL used to restrict exports of sensitive products to the United States is anti-dumping regulations.33 Tariff barriers, nontariff barriers and anti-dumping measures have all been used to restrict the flow of goods from the CSFR, Hungary and Poland to the United States. Although some relief has been granted to Eastern Europe in recent years, restric tions still prevail on the products in which these countries have a relative comparative advantage.34 The importance of these restrictions for each sector are discussed below. Steel The granting of MFN status to the CSFR and Poland reduced the tariffs they faced on steel exports to the United States from an average of 20-25 percent to an average of less than 5 percent.35 However, quota restrictions and anti-dumping measures act as limits on steel products. Until March 1992, steel from the CSFR, Hungary and Poland was subject to quantitative restrictions. These limits were raised in 1989 and 1991, and, as part of the Trade Enhancement Initiative for Central and Eastern Europe, the United States committed itself to adjusting the ceilings further, either through increased flexibility in the admin istration of the quotas, or increasing the actual quotas.36 This commitment, however, was never acted upon. At the end of March 1992, the United States allowed all of the quantitative restrictions on steel to elapse. The end of quantitative restrictions on U.S. steel imports did not open the U.S. market to steel products from Eastern Europe. Within two months of the elimination of quotas, the U.S. steel industry accused every significant foreign supplier of steel to the United States of dumping their product in the U.S. market. In the summer of 1993, the USITC ruled that steel producers in 19 countries had dumped their products in the U.S. market, and that this action had resulted in injury or the potential for injury to the U.S. steel industry. In accordance with this finding, the and Poland, still maintains restrictions on many products, most notably, agricultural, chemicals, iron and steel, textiles and apparel, and footwear. 35 See USITC (November 1991, p. 117). 36 As noted in the text, not only did the United States place lim its on each type of steel product but even in the amount which could be imported during subperiods of the year. 35 USITC imposed duties ranging from 18 percent to 109 percent of the value of the steel product on imports from the dumping countries. Dumping duties on Polish exports to the United States of carbon steel plate were levied at 62 percent, effectively eliminating Polish exports of this product to the U.S. market. Although no charges of dumping were filed against the CSFR and Hungary, the size and extent of the dumping duties is likely to limit the growth of steel exports to the United States from all countries.37 Prior to the anti-dumping case, the CSFR, Hungary and Poland had begun programs to make their “steel enterprises more market-oriented, costconscious and perhaps more export oriented.”38 Use of anti-dumping measures by the United States is an indication to these countries that even if they follow the prescriptions of the West and develop an efficient industry, they still may be denied access to U.S. markets. Textiles and A pparel U.S. imports of textiles and apparel from the CSFR, Hungary and Poland are covered by quotas in accordance with the Multifiber Arrangement.39 Before the granting of MFN status to the CSFR and Poland, quota utilization rates (which indi cate how close a country comes to meeting the quota on a particular product) for these countries were very low, because the tariffs acted as an effective barrier to trade.40 Even though MFN tariffs are high, utilization rates have increased since the granting of MFN status. Utilization rates rose even as the United States has negotiated new textile and apparel agreements with the CSFR, Hungary and Poland which have increased these quotas. Under the Trade Enhancement Initiative for Central and Eastern Europe, the United States pledged to take steps to increase its imports of textiles and apparel from Eastern Europe. In accordance with this initiative, the United States raised the quotas on some imports from the CSFR, Hungary and Poland. The United States also promised to consider setting quotas for more broadly defined product categories which would allow the countries more flexibility in meeting the quotas.41 Chemicals Tariffs on industrial chemicals and fertilizers average only 2 percent ad valorem and, thus, since the granting of MFN privileges to all three countries, they do not represent a significant barrier to trade. According to the Organization for Economic Co-operation and Development (OECD), the main obstacle to the growth of Eastern European chem ical exports has been the use of anti-dumping measures by the West.42 For example, the U.S. chem ical industry filed dumping charges in 1992 against the imports of sulfanilic acid from Hungary.43 The USITC found preliminary evidence that the Hungarian producers were dumping this product in the United States and causing harm to the U.S. chem ical industry. Temporary duties equal to 58 percent of the value of Hungarian shipments of sulfanilic acid were assessed. These duties were rescinded when, in its final decision in February 1993, the USITC ruled that there was not sufficient evidence that these imports were injuring the domestic industry. Agriculture Agricultural exports from the CSFR, Hungary and Poland are affected both by U.S. agricultural subsidies and nontariff barriers. According to the USITC, the only nontariff barriers in agriculture that significantly affect the CSFR, Hungary and 37 The imposition of duties which block certain producers from the U.S. market does not necessarily lead to an increase in imports from the “nondumping” producers. The U.S. indus try is free to file charges of dumping against foreign competi tors at any time. Thus, the finding of dumping may act as a deterrent to other producers to expand their exports to the United States. 40 As noted in the text above, non-MFN tariffs in textiles range from 50 percent to 100 percent, while MFN tariffs range from 20 percent to 35 percent. 38 USITC (November 1991, p. 117). 41 See USITC (November 1991, p. 46). 39 The Multifiber Arrangement (MFA) refers to the bilaterally negotiated quota restrictions on textiles and apparel, which are placed by developed countries on imports from develop ing countries. The MFA is negotiated under the auspices of the GATT committee on textiles. See Hamilton (1990). If Congress approves the GATT Uruguay Round of mulitilateral trade agreements, the MFA will be phased out over a 42 OECD (1992, p. 92). 10-year period beginning in July 1995. Quota restrictions on textiles and apparel are then to be replaced by GATTnegotiated tariffs. 43 Sulfanilic acid is a gray-white to white crystalline solid. Its main uses are in the production of synthetic dyes that in turn are used in foods, drugs and cosmetics, and in the produc tion of optical brightening agents. Sulfanilic acid is also used in concrete additives. (USITC, February 1993.) JULY/AUGUST 1994 36 Poland axe the quantitative restrictions on cheese imports.44 Most cheese products are covered by quotas and those which are not face high tariff barriers. Furthermore, as noted above, cheese products are not eligible for GSP treatment. As part of the TEI, the United States committed itself to increasing the access of cheese products from these countries into the U.S. market. None theless, no progress has been made on this pro posal. For example, in 1991 Hungary petitioned the United States to allow GSP benefits for the importation of goya cheese, one of the few cheeses for which importation into the United States is not limited by quotas. Imports, however, are restricted by a 25 percent tariff. Hungary provided 25 percent of the total U.S. imports of goya cheese in 1990. Although no goya cheese is produced in the United States, the U.S. dairy industry opposed the extension of GSP benefits to goya cheese, arguing that this product was a substitute for domestically produced, hard, Italian-type cheeses.45 Because of this opposition, the United States refused Hungary’s request to add goya cheese to the list of GSP-eligible products. CONCLUSION Foreign trade is vitally important for the CSFR, Hungary and Poland to facilitate the re-structuring of their economies. These countries are depen dent upon exports to ensure a supply of foreign currency to finance capital purchases (reducing the pressures to incur foreign debt), and to pro mote economic growth, which in turn is critical to their political stability. The governments of the CSFR, Hungary and Poland have made great progress over the past few years in reforming their economies. The role of the state has been reduced substantially through the deregulation of prices, the privatization of industries, and the adoption of legislation aimed at fostering the market system. Furthermore, all of these countries have substantially liberalized their trading environments by eliminating quotas, harmonizing tariffs, and permitting the convert ibility of their currencies. Officials in these coun tries cite the continuation of Western trade barriers as one of the primary hindrances to their successful transition to market democracies.46 44 USITC (April 1992, p. 18). 45 See USITC (March 1992) for details. 46 Burke (January 19, 1994). Digitized forFEDERAL FRASER RESERVE BANK OF ST. LOUIS The United States’ economic growth has benefitted from the reforms undertaken by the Eastern European countries. Most notably, U.S. exports to these countries have expanded substantially. Despite these gains, the United States continues to restrict access to its markets to goods produced in Eastern Europe. As shown in this article, the products in which the CSFR, Hungary and Poland have the greatest comparative advantage are pre cisely those in which the United States maintains the greatest restrictions on trade. Reducing the trade barriers to these products will spur economic growth in Eastern Europe, and is an important step the West can take to ensure that the coun tries of Eastern Europe continue along the path of reform. REFERENCES Bank for International Settlements. 63rd Annual Report, June 1993. Burke, Justin. “Central European Countries Cite Obstacles to Economic Integration,” The Christian Science Monitor, January 19, 1994, p. 9. Collins, Susan M., and Dani Rodrik. Eastern Europe and the Soviet Union in the World Economy. Institute for International Economics, May 1991. Collins, Susan M. “Policy Watch: U.S. Economic Policy Toward the Soviet Union and Eastern Europe,” The Journal o f Economic Perspectives (fall 1991), pp. 219-27. Coughlin, Cletus C. “U.S. Trade-Remedy Laws: Do They Facilitate or Hinder Free Trade?” this Review (July/August 1991), pp. 3-18. Erzan, Refik, and Christopher Holmes. “The Restrictiveness of the Multi-Fibre Arrangement on Eastern European Trade," Policy Research Working Papers, WPS 860, The World Bank (February 1992). Fieleke, Norman S. “Commerce With the Newly Liberalizing Countries: Promised Land, Quicksand, or What?” New England Economic Review (May/June 1990), pp. 19-33. Finger, Michael J. ‘Trade Policies in the United States,” National Trade Policies, Handbook o f Comparative Economic Policies, vol. 2. Greenwood Press, 1992, pp. 79-108. Green, Paula L. “U.S. Monitors EC Ties With Eastern Europe,” Journal of Commerce (February 6, 1992). Hamilton, Carl B., editor. Textiles Trade and the Developing Countries: Eliminating the Multi-Fibre Arrangement in the 1990s. The World Bank, 1990. International Monetary Fund. World Economic Outlook (May 1993). Murrell, Peter. The Nature of Socialist Economies: Lessons from Eastern European Foreign Trade. Princeton University Press (1990). 37 Organization for Economic Co-operation and Development. Economic Surveys: Czech and Slovak Federal Republic. OECD, December 1991. _____ . Reforming the Economies of Central and Eastern Europe. OECD, 1992. _____ . Economic Surveys: Hungary. OECD, 1993. Ray, Edward J. “Protection of Manufacturers in the United States,” in David Greenaway and others, eds., Global Protectionism. St. Martin’s Press, 1991, pp. 12-39. Rodrik, Dani. “Foreign Trade in Eastern Europe’s Transition: Early Results,” National Bureau of Economic Research Working Paper Series No. 4064 (May 1992). U.S. Department of State, Bureau of Economic and Business Affairs, Textiles Division. “United States and the Republic of Poland Sign New Textile Agreement” (Textiles Division, Public Release, January 28, 1992). _____ . “United States and the Czech and Slovak Federal Republic Amend Their Bilateral Textile Agreement” (Textiles Division, Public Release, July 21, 1992). United States International Trade Commission. U.S. Laws and U.S. and EC Trade Agreements Relating to Nonmarket Economies. USITC, March 1990. United States International Trade Commission, Office of Economics. “Legislation to Ensure Role for Congress in Extending MFN Status to Nonmarket Economy Countries,” International Economic Review (June-July 1990), pp. 7-8. _____ . “Liberalization of Foreign Trade in Czechoslovakia, Hungary, and Poland: Progress and Prospects,” International Economic Review (August 1991), pp. 5-7. United States International Trade Commission. Central and Eastern Europe: Export Competitiveness of Major Manufacturing and Services Sectors. USITC Publication 2446 (November 1991). _____ . President’s List of Articles Which May Be Designated Or Modified As Eligible Articles For Purposes of the U.S. Generalized System of Preferences. USITC Publication 2491 (March 1992). _____ . Trade Between the United States and China, the Former Soviet Union, Central and Eastern Europe, the Baltic Nations, and Other Selected Countries During 1991. USITC Publication 2503 (April 1992). _____ . Sulfanilic Acid from the Republic of Hungary and India. USITC Publication 2603 (February 1993). Wang, Z.K., and L. Alan Winters. ‘The Trading Potential of Eastern Europe,” Journal of Economic Integration (autumn 1992), pp. 113-36. JULY/AUGUST 1994 38 Appendix Index of Relative Comparative Advantage: CSFR 1991 1992 3.597 1.072 0.031 3.616 4.585 4.418 0.047 3.189 1.130 0.000 0.000 0.017 0.495 0.151 0.231 0.731 0.270 0.000 1.184 0.000 0.000 3.019 0.028 0.000 0.000 0.116 00190— Wine & related products 3.415 3.773 4.261 4.305 1.546 00200— Feedstuff and foodgrains 0.129 7.632 16.709 12.935 11.816 01000— Fish & shellfish 0.049 0.000 0.000 0.000 0.000 01010— Alcoholic beverages, except wine 0.010 0.944 0.077 0.038 0.126 01020— Other nonagricultural foods & food additives 0.133 0.595 0.885 0.468 0.530 10010— Fuel oil 0.000 0.004 0.000 0.000 0.000 10020— Other petroleum products 0.002 0.004 0.000 0.000 0.001 10030— Liquified petroleum gases 0.000 0.000 0.000 0.000 0.000 10100— Coal & other fuels, except gas 0.000 0.000 0.000 0.032 0.000 10300— Nuclear fuel materials & fuels 0.000 0.000 0.000 0.005 0.000 11000— Pulpwood and woodpulp 0.000 0.080 0.000 0.000 0.000 11100— Newsprint 0.038 2.309 0.000 0.000 0.000 11110— Paper & paper products, n.e.s. 0.080 0.000 0.000 0.017 0.041 12000— Cotton, wool & other natural fibers 0.000 3.923 0.000 0.000 0.000 12030— Hides & skins, & fur skins-raw 0.417 0.000 0.000 0.000 0.000 12060— Farming materials, including farm animals 0.000 0.031 0.125 0.410 0.366 12070— Other (tobacco, waxes, nonfood oils) 0.004 6.477 14.058 8.566 8.418 12100— Cotton cloth & fabrics, thread & cordage 3.268 4.971 2.902 1.812 3.902 12110— Wool, silk & other vegetable fabric 7.372 14.032 17.173 19.819 23.339 12135— Synthetic cloth & fabric, thread 0.528 5.618 5.218 9.048 10.218 12140— Other materials (hair, synthetics, etc.) 67.477 12.076 0.000 0.200 0.000 12150— Finished textile industrial supplies 11.145 3.257 1.468 1.555 0.776 0.000 0.012 0.341 0.000 0.932 12320— Other materials, except chemicals 1.637 0.070 0.040 0.068 0.000 12500— Plastic materials 0.100 0.278 0.276 0.091 0.090 12510— Fertilizers, pesticides, and insecticides 0.007 0.000 0.023 0.000 0.000 12530— Industrial inorganic chemicals 0.863 3.117 0.736 1.282 2.610 12540— Industrial organic chemicals 0.732 0.075 0.114 0.439 0.500 12550— Other chemicals (coloring agents, print inks, paint) 0.026 0.418 0.100 1.726 4.199 13000— Lumber & wood in the rough 0.074 0.428 0.040 0.000 0.005 13010— Plywood & veneers 0.000 0.000 0.015 0.000 0.000 0.225 Product 1988 1989 1990 00100— Meat, poultry & other edible animals 3.692 3.938 00110— Dairy products & eggs 3.215 3.239 00120— Fruits & preparations 0.046 0.013 00130— Vegetables & preparations 0.034 00160— Bakery & confectionery products 0.360 00170— Tea, spices & preparations 0.000 00180— Other (soft beverages, processed coffee, etc.) 12160— Leather & furs-unmanufactured 13020— Stone, sand, cement & lime 0.000 0.859 0.266 0.077 13100— Glass-plate, sheet, etc. (excluding automotive) 5.740 5.540 6.474 2.646 8.581 13110— Other-finished (shingles, molding, etc.) 1.676 0.280 0.302 0.140 0.089 13120— Nontextile floor & wall tiles and other covering 0.000 0.338 0.045 0.000 0.052 14000— Steelmaking & ferroalloying material 0.000 0.000 0.000 1.882 1.462 14100— Iron & steel mill products-semifinished 6.779 5.233 1.603 1.936 1.828 14200— Bauxite & aluminum 0.000 0.000 0.000 0.000 0.000 Digitized for FEDERAL FRASER RESERVE BANK OF ST. LOUIS 39 Product 1988 1989 1990 1991 1992 14220— Copper 0.000 0.000 0.000 0.000 0.013 14240— Nickel 0.736 0.000 0.000 0.000 0.010 14280— Other precious metals 0.000 0.000 0.000 0.000 0.000 14290— Miscellaneous nonferrous 0.050 0.000 0.000 0.550 0.346 15000— Iron and steel products, except advanced 1.911 2.256 0.664 3.417 1.730 15100— Iron and steel manufactured, advanced 0.044 0.008 0.014 0.285 0.834 15200— Finished metal shapes, except steel 0.006 0.004 0.042 0.042 0.098 16040— Sulfur & nonmetallic minerals 5.051 0.037 0.000 0.000 0.016 16050— Other (synthetic rubber, wood, cork, gums, etc.) 0.000 0.000 0.000 0.000 0.120 16120— Other (boxes, belting, glass, abrasives, etc.) 0.370 0.160 0.355 0.484 0.735 0.103 20000— Generators, transformers & accessories 0.977 0.019 0.000 0.074 20005— Electric apparatus & parts, n.e.c 0.008 0.049 0.002 0.065 0.061 21000— Drilling & oil field equipment 10.966 10.488 5.828 5.448 0.000 21010— Specialized mining & oil processing equipment 0.000 0.000 0.000 0.434 1.833 21030— Excavating, paving & construction 0.040 0.016 0.000 0.350 0.863 21040— Nonfarm tractors & parts 1.440 4.487 3.207 0.434 0.797 21100— Industrial engines, pumps, compressors & generators 0.000 0.002 0.081 0.115 0.093 21110— Food & tobacco processing machinery 0.000 0.000 0.000 0.118 0.483 21120— Machine tools, metal working 3.611 2.725 3.775 2.599 4.361 21130— Industrial textiles, sewing, & leather working machinery 2.023 3.872 3.845 6.689 6.092 21140— Woodworking, glass working, and plastic machinery 0.014 0.004 1.492 1.125 0.308 21150— Pulp & paper machinery 2.249 3.725 5.315 1.998 2.347 0.363 0.838 21160— Measuring, testing & control instruments 0.000 0.036 0.343 0.197 21170— Materials handling equipment 0.000 0.006 0.437 1.331 21180— Other industrial machinery 0.158 0.454 1.461 4.629 5.436 21190— Photo & other service industry machinery 0.076 0.138 0.019 0.221 0.179 13.782 14.761 22.738 11.487 9.580 21200— Agricultural machinery and equipment 21300— Computers 0.000 0.000 0.000 0.004 0.006 21301— Computer accessories, peripherals 0.001 0.105 0.025 0.003 0.062 21320— Semiconductors 0.000 0.000 0.000 0.014 0.009 21400—Telecommunications equipment 0.033 0.777 0.318 0.698 0.697 21500— Business machinery & equipment, except computers 0.283 0.121 0.183 0.070 0.050 21600— Laboratory, testing & control instruments 0.000 0.000 0.165 0.079 0.335 21610— Other scientific, medical & hospital equipment 0.000 0.000 0.045 0.016 0.085 22000— Civilian aircraft, complete - all 0.000 0.000 0.127 0.023 0.086 22010— Parts for civilian aircraft 0.002 0.006 0.000 0.070 0.161 22020— Engines for civilian aircraft 0.000 0.000 0.000 0.037 0.007 22220— Marine engines & parts 0.000 0.000 0.000 0.000 0.000 30000— Complete & assembled-new & used 0.001 0.000 0.000 0.000 0.000 30100— Complete & assembled 0.004 0.478 0.000 0.000 0.016 30200— Engines & engine parts 0.000 0.008 0.020 0.085 0.114 30220— Automotive tires & tubes 10.637 9.019 12.498 13.744 9.249 30230— Other parts & accessories 0.000 0.023 0.005 0.011 0.049 40000— Apparel & household goods-cotton 0.689 0.398 0.398 0.709 1.023 10.616 40010— Apparel & household goods-wool 20.882 13.846 14.072 20.096 40020— Apparel & household goods-other textiles 0.042 0.542 0.436 0.894 0.974 40030— Nontextile apparel & household goods 0.804 1.142 0.820 0.418 0.427 40040— Footwear of leather, rubber, or other materials 4.939 8.644 8.741 5.845 4.306 40050— Sporting & camping apparel and footwear & gear 9.108 0.942 0.392 2.958 2.179 40100— Medicinal, dental & pharmaceutical preparations 2.254 0.116 0.249 0.369 1.367 JULY/AUGUST 1994 40 P ro d u c t 1988 1989 1990 40110— Books, magazines & other printed material 3.030 5.387 3.712 1.521 1.906 40120—Toiletries & cosmetics 0.000 0.000 2.034 0.000 0.367 1992 1991 40140— Other products (notions, writing & art supplies) 1.579 1.194 1.538 1.601 1.598 41000— Furniture, household items, baskets 1.898 2.914 3.516 3.304 2.699 19.451 19.780 20.002 13.120 16.421 0.719 41010— Glassware and porcelain 41020— Cookware, chinaware, cutlery, house & garden wares 0.402 0.210 0.419 1.202 41030— Household & kitchen appliances 0.004 0.000 0.000 0.009 0.007 41040— Rugs & other textile floor covering 0.108 0.060 0.417 1.404 0.743 41050— Other (clocks, portable typewriters, other goods) 0.236 1.614 1.064 0.751 1.020 41100— Motorcycles & parts 2.011 1.968 1.052 2.383 1.825 41110— Pleasure boats & motors 0.000 0.000 0.064 0.011 0.000 41120— Toys, shooting & sporting goods & bicycles 1.734 0.499 0.535 0.676 1.337 41130— Photo & optical equipment 0.005 0.000 0.000 0.026 0.013 41140— Musical instruments & other recreational equipment 2.628 7.830 9.132 14.905 12.279 41210— Radios, phonographs, tape decks & other stereo 0.000 0.000 0.000 0.000 0.000 41220— Records, tapes & disks 0.580 2.122 6.494 2.147 1.460 41300— Numismatic coins 0.096 0.273 0.293 0.186 0.033 41310— Jewelry (watches, rings, etc.) 1.239 0.306 0.163 0.251 0.685 41320— Artwork, antiques, stamps and other collectibles 1.527 1.866 2.124 1.877 3.018 24.705 0.055 0.000 0.000 0.000 0.000 0.000 0.000 0.011 0.000 42000— Nursery stocks, cut flowers, Christmas trees 42100— Gem diamonds-uncut or unset 42110— Other gem stones-precious, semiprecious, & imitations 18.437 28.459 29.001 26.076 28.718 50000— Military aircraft & parts 0.000 0.000 0.000 0.072 0.223 50010— Other military equipment 0.000 1.168 2.261 2.980 3.917 50020— U.S. goods returned, & reimports 0.083 0.238 0.068 0.243 0.224 50030— Minimum value shipments 3.031 2.657 2.932 2.595 2.258 50040— Other (movies, miscellaneous imports & special transactions) 0.401 2.284 1.511 5.017 7.577 Index of Relative Comparative Advantage: Hungary P ro d u c t 00000— Green coffee 1988 1989 1990 1 9 91 1992 0.000 0.097 0.000 0.000 0.000 00100— Meat, poultry & other edible animals 14.591 9.405 12.703 13.159 8.179 00110— Dairy products & eggs 11.418 19.675 14.758 19.731 21.707 00120— Fruits & preparations 6.771 10.356 10.951 12.691 8.122 00130— Vegetables & preparations 4.861 9.420 4.692 3.826 2.697 00140— Nuts & preparations 0.096 0.069 0.000 0.000 0.050 00150— Food oils & oilseeds 0.057 0.000 0.418 0.000 0.614 00160— Bakery & confectionery products 0.144 0.814 1.713 1.816 1.361 00170— Tea, spices & preparations 4.940 6.259 5.211 5.039 3.535 00180— Other (soft beverages, processed coffee, etc.) 2.051 2.679 1.673 1.157 0.271 00190— Wine & related products 1.393 1.778 1.993 2.197 2.516 00200— Feedstuff and foodgrains 0.819 1.749 1.310 1.441 5.223 01000— Fish & shellfish 0.000 0.001 0.026 0.000 0.000 01010— Alcoholic beverages, except wine 0.016 0.044 0.066 0.104 0.150 01020— Other nonagricultural foods & food additives 2.170 1.416 1.667 1.439 1.268 10010— Fuel oil 0.000 0.000 0.000 0.001 0.000 10020— Other petroleum products 0.179 0.303 0.177 0.190 0.234 0.069 0.014 10300— Nuclear fuel materials & fuels RESERVE BANK OF ST. LOUIS FEDERAL 0.000 0.000 0.056 41 Product 1988 1989 1990 1991 1992 11000— Pulpwood and woodpulp 0.004 0.000 0.000 0.000 0.000 0.010 11110— Paper & paper products, n.e.s. 0.021 0.027 0.004 0.000 12000— Cotton, wool & other natural fibers 1.253 2.597 0.465 1.331 1.180 12030— Hides & skins, & fur skins-raw 0.159 0.067 0.099 0.026 0.000 11.475 25.183 5.298 0.403 3.101 5.543 3.044 12060— Farming materials, including farm animals 12070— Other (tobacco, waxes, nonfood oils) 1.093 2.792 3.089 12100— Cotton cloth & fabrics, thread & cordage 9.846 6.120 4.046 1.343 1.325 12110— Wool, silk & other vegetable fabric 6.299 8.118 4.086 4.559 3.813 12135— Synthetic cloth & fabric, thread & cordage 6.093 4.726 2.439 1.953 1.364 0.047 12140— Other materials (hair, synthetics, etc.) 2.417 2.933 5.336 0.758 12150— Finished textile industrial supplies 5.273 2.070 3.206 1.927 2.001 12160— Leather & furs-unmanufactured 0.143 1.307 0.611 2.222 0.000 12320— Other materials, except chemicals 0.293 0.029 0.740 1.006 1.140 12500— Plastic materials 0.049 0.541 2.453 1.372 1.371 12510— Fertilizers, pesticides, and insecticides 0.000 0.007 0.000 0.789 1.679 12530— Industrial inorganic chemicals 2.253 1.391 1.542 1.481 1.621 12540— Industrial organic chemicals 0.476 1.137 1.234 0.750 1.171 0.203 12550— Other chemicals (coloring agents, print inks, paint) 3.683 0.737 0.653 0.575 13000— Lumber & wood in the rough 0.000 0.000 0.000 0.001 0.001 13010— Plywood & veneers 0.000 0.013 0.009 0.000 0.000 13020— Stone, sand, cement & lime 0.095 0.006 0.000 0.000 0.000 13100— Glass-plate, sheet, etc. (excluding automotive) 0.033 0.154 0.000 0.128 0.033 0.009 13110— Other-finished (shingles, molding, wallboard, etc.) 0.000 0.001 0.002 0.004 13120— Nontextile floor & wall tiles and other covering 0.000 0.000 0.012 0.088 0.000 14000— Steelmaking & ferroalloying materials 0.046 0.000 0.003 0.000 0.000 14100— Iron & steel mill products-semifinished 1.917 1.957 2.083 1.162 0.988 14200— Bauxite & aluminum 0.154 0.010 0.000 0.086 0.000 14220— Copper 0.000 0.000 0.000 0.000 0.040 14240— Nickel 0.000 0.296 0.145 0.000 0.000 14270— Nonmonetary gold 0.032 0.029 0.034 0.007 0.000 0.000 14280— Other precious metals 0.000 0.000 0.000 0.000 14290— Miscellaneous nonferrous 0.063 0.042 0.069 0.645 1.095 15000— Iron and steel products, except advanced manufacturers 0.076 0.238 0.003 0.003 0.138 15100— Iron and steel manufacturers, advanced 0.148 0.011 0.017 0.026 0.055 15200— Finished metal shapes, except steel 6.805 7.929 6.608 5.552 9.806 16040— Sulfur & nonmetallic minerals 0.549 0.205 0.203 0.091 0.000 16050— Other (synthetic rubber, wood, cork, gums, resins, etc.) 0.022 0.000 0.065 0.136 0.066 16110— Audio & visual tapes & other media 0.000 0.008 0.001 0.000 0.000 16120— Other (boxes, belting, glass, abrasives, etc.) 1.040 1.426 3.016 2.535 1.666 1.268 20000— Generators, transformers & accessories 0.003 0.075 0.180 0.522 20005— Electric apparatus & parts, n.e.c 0.449 0.787 0.640 0.897 1.295 21000— Drilling & oil field equipment & platforms 0.467 0.000 0.000 0.000 0.000 21010— Specialized mining & oil processing equipment 0.000 0.000 0.000 0.000 0.024 21030— Excavating, paving & construction machinery 0.202 0.014 0.191 1.150 0.113 21040— Nonfarm tractors & parts 0.639 1.689 0.340 0.025 3.454 21100— Industrial engines, pumps, compressors & generators 0.054 0.174 0.325 0.610 0.635 21110— Food & tobacco processing machinery 0.089 0.087 0.079 0.009 0.006 21120— Machine tools, metal working, molding & rolling 0.183 1.046 0.531 0.949 0.292 21130— Industrial textiles, sewing, & leather working machinery 0.031 0.000 0.010 0.016 0.176 21140— Woodworking, glass working, & plastic & rubber machinery 0.029 0.012 0.119 0.152 3.631 JULY/AUGUST 1994 42 P ro d u c t 1988 1989 1990 1991 21150— Pulp & paper machinery 0.002 0.015 0.010 0.600 0.371 21160— Measuring, testing & control instruments 0.254 0.314 0.212 0.193 0.671 21170— Materials handling equipment 0.138 0.332 0.289 0.514 0.322 21180— Other industrial machinery 1.301 1.588 1.846 0.908 0.963 21190— Photo & other service industry machinery 0.225 0.290 0.176 0.144 0.122 21200— Agricultural machinery and equipment 3.852 7.856 7.170 8.139 10.255 21300— Computers 0.000 0.001 0.003 0.002 0.001 21301— Computer accessories, peripherals & parts 0.000 0.007 0.004 0.013 0.006 21320— Semiconductors 0.007 0.001 0.009 0.003 0.030 21400—Telecommunications equipment 0.191 0.014 0.012 0.026 0.014 21500— Business machinery & equipment, except computers 0.010 0.290 0.141 0.079 0.030 21600— Laboratory, testing & control instruments 0.013 0.078 0.060 0.432 1.204 21610— Other scientific, medical & hospital equipment 0.071 0.172 0.075 0.113 0.163 22000— Civilian aircraft, complete - all 0.009 0.003 0.009 0.004 0.011 1992 22010— Parts for civilian aircraft 0.000 0.005 0.019 0.033 0.010 22020— Engines for civilian aircraft 0.000 0.000 0.000 0.000 0.096 22100— Railway transportation equipment 0.086 0.000 0.174 0.512 0.010 22210— Other commercial vessels, new and used 0.000 0.000 0.000 0.698 0.000 22220— Marine engines & parts 0.000 0.000 0.000 0.000 0.011 30100— Trucks, buses, & special purpose vehicles 0.000 0.005 0.000 0.095 0.001 30110— Bodies & chassis for trucks & buses 0.001 2.774 15.454 29.076 9.517 30200— Engines & engine parts 0.026 0.002 0.004 0.001 0.001 30220— Automotive tires & tubes 5.801 5.468 4.798 1.667 0.643 30230— Other parts & accessories 2.261 2.303 2.417 2.601 2.921 40000— Apparel & household goods-cotton 1.245 1.688 1.715 1.646 1.972 21.349 40010— Apparel & household goods-wool 17.527 18.727 14.530 13.901 40020— Apparel & household goods-other textiles 2.414 2.043 1.512 1.472 1.920 40030— Nontextile apparel & household goods 1.054 1.365 1.937 0.293 0.333 40040— Footwear of leather, rubber, or other materials 4.334 2.658 4.401 3.076 2.342 40050— Sporting & camping apparel, footwear & gear 0.467 0.364 0.448 0.614 0.516 40100— Medicinal, dental & pharmaceutical preparations 5.698 5.677 6.617 4.566 3.644 40110— Books, magazines & other printed material 0.586 0.458 0.303 0.741 0.153 40120— Toiletries & cosmetics 0.059 0.039 0.016 0.172 0.060 40140— Other products (notions, writing & art supplies) 0.071 0.583 0.375 0.323 0.258 41000— Furniture, household items, baskets 2.028 2.181 1.558 1.816 0.929 41010— Glassware and porcelain 4.854 6.253 6.242 7.284 9.624 41020— Cookware, chinaware, cutlery, & other household goods 0.919 1.953 1.018 1.209 1.365 41030— Household & kitchen appliances 0.011 0.000 0.006 0.028 0.005 41040— Rugs & other textile floor covering 0.652 0.557 0.914 0.709 0.704 41050— Other (clocks, portable typewriters, other household goods) 6.246 2.651 2.073 2.936 1.817 41100— Motorcycles & parts 0.000 0.054 0.000 0.000 0.000 41110— Pleasure boats & motors 0.164 0.588 0.174 0.007 0.123 41120— Toys, shooting & sporting goods, & bicycles 0.100 0.086 0.137 0.965 1.794 41130— Photo & optical equipment 0.000 0.253 0.215 0.159 0.149 41140— Musical instruments & other recreational equipment 0.300 0.352 0.102 0.318 0.315 41200— Television receivers, vers & other video equipment 0.000 0.000 0.000 0.000 0.000 41210— Radios, phonographs, tape decks & other stereo 0.000 0.000 0.001 0.003 0.000 41220— Records, tapes & disks 2.405 1.617 2.066 1.810 1.391 41300— Numismatic coins 0.059 0.078 6.805 0.668 0.223 41310— Jewelry (watches, rings, etc.) 0.108 0.070 0.048 0.007 0.000 Digitized forFEDERAL FRASER RESERVE BANK OF ST. LOUIS 43 Product 1988 1989 41320— Artwork, antiques, stamps and other collectibles 0.664 42000— Nursery stocks, cut flowers, Christmas trees 0.000 42100— Gem diamonds-uncut or unset 0.017 0.000 42110— Other gem stones-precious, semiprecious & imitation 0.000 0.016 1992 1990 1991 0.695 1.063 4.373 1.429 0.003 0.055 0.205 0.246 0.000 0.000 0.000 0.004 0.000 0.000 50000— Military aircraft & parts 0.000 0.000 0.000 0.000 0.008 50010— Other military equipment 3.655 3.570 2.698 2.423 4.116 50020— U.S. goods returned, & reimports 0.106 0.090 0.070 0.224 0.129 50030— Minimum value shipments 0.511 0.364 0.380 0.401 0.378 50040— Other (movies, miscellaneous imports & special transactions) 0.603 0.577 3.002 0.050 1.433 Index of Relative Comparative Advantage: Poland Product 1988 1989 1990 1991 1992 00000— Green coffee 0.000 0.025 0.000 0.000 0.000 0.000 0.000 0.072 0.000 0.000 38.943 37.600 25.330 11.456 8.119 00110— Dairy products & eggs 6.980 13.598 7.710 11.218 21.245 00120— Fruits & preparations 1.227 2.234 2.184 4.453 3.546 00130— Vegetables & preparations 0.803 1.038 1.214 1.788 1.298 00150— Food oils & oilseeds 0.490 2.364 18.895 0.677 0.548 00020— Cane and beet sugar 00100— Meat, poultry & other edible animals 00160— Bakery & confectionery products 1.024 1.522 1.335 2.500 3.081 00170— Tea, spices & preparations 0.941 0.104 0.509 0.014 0.192 00180— Other (soft beverages, processed coffee, etc.) 0.905 0.957 1.090 0.570 2.039 00190— Wine & related products 0.096 0.122 0.074 0.193 0.296 00200— Feedstuff and foodgrains 7.577 0.153 0.024 4.586 4.428 01000— Fish & shellfish 2.261 2.057 4.079 3.731 1.579 01010— Alcoholic beverages, except wine 0.151 0.263 1.560 1.573 0.623 01020— Other nonagricultural foods & food additives 0.612 0.000 0.000 0.000 0.000 10010— Fuel oil 0.000 0.000 0.000 0.000 0.408 10020— Other petroleum products 0.000 0.000 0.000 0.000 0.000 10100— Coal & other fuels, except gas 0.000 36.870 0.000 0.000 0.139 10300— Nuclear fuel materials & fuels 0.000 0.000 0.000 0.023 0.000 11000— Pulpwood and woodpulp 0.000 0.000 0.000 0.000 0.000 11110— Paper & paper products, n.e.s. 0.002 0.047 0.002 0.000 0.001 12030— Hides & skins, & fur skins-raw 0.566 0.795 0.593 1.845 2.689 12050— Natural rubber & similar gums 0.000 0.000 0.000 0.000 0.000 9.954 2.174 2.867 1.913 2.326 14.229 8.507 9.420 10.847 14.617 12060— Farming materials, including farm animals 12070— Other (tobacco, waxes, nonfood oils) 12100— Cotton cloth & fabrics, thread & cordage 2.222 0.634 1.167 1.794 0.672 12.146 15.070 19.067 27.927 34.972 1.011 1.782 2.071 2.919 2.880 44.130 0.559 0.000 0.000 0.000 12150— Finished textile industrial supplies 0.001 0.676 0.051 0.010 0.000 12160— Leather & furs-unmanufactured 0.000 0.094 0.324 0.092 0.104 12110— Wool, silk & other vegetable fabric 12135— Synthetic cloth & fabric, thread & cordage 12140— Other materials (hair, synthetics, etc.) 12320— Other materials, except chemicals 0.028 2.298 0.000 0.002 0.016 12500— Plastic materials 0.016 0.016 0.034 0.015 0.148 12510— Fertilizers, pesticides, and insecticides 0.002 0.131 0.000 1.959 2.332 12530— Industrial inorganic chemicals 0.624 0.470 1.738 2.921 0.974 12540— Industrial organic chemicals 0.445 0.731 0.263 0.513 0.250 JULY/AUGUST 1994 44 Product 1988 12550— Other chemicals (coloring agents, print inks, paint) 2.636 13000— Lumber & wood in the rough 0.000 13010— Plywood & veneers 0.734 13020— Stone, sand, cement & lime 1989 1990 1991 1992 1.089 1.319 4.231 4.385 0.000 0.000 0.000 0.000 0.000 0.090 0.741 3.899 0.000 0.000 0.000 0.021 0.000 13100— Glass-plate, sheet, etc. (excluding automotive) 0.084 0.000 1.746 1.788 0.931 13110— Other-finished (shingles, molding, wallboard, etc.) 0.000 1.413 0.387 0.000 0.005 13120— Nontextile floor & wall tiles and other covering 0.000 0.002 0.000 0.000 0.000 14000— Steelmaking & ferroalloying materials 0.000 0.000 0.194 0.000 0.580 14100— Iron & steel mill products-semifinished 3.339 3.309 2.773 4.461 1.915 14200— Bauxite & aluminum 0.151 0.090 0.088 0.178 0.000 14220— Copper 6.709 13.997 0.000 0.000 0.018 14240— Nickel 0.550 0.000 0.637 0.000 0.000 14260—Zinc 6.233 0.017 1.277 0.020 1.756 14270— Nonmonetary gold 0.000 0.000 0.000 0.030 0.000 14280— Other precious metals 0.000 0.000 0.000 0.002 0.000 14290— Miscellaneous nonferrous 0.099 0.000 0.006 0.187 0.000 15000— Iron and steel products, except advanced manufacturers 3.122 2.625 2.380 2.249 1.404 15100— Iron and steel manufacturers, advanced 1.122 1.831 1.563 2.255 1.311 15200— Finished metal shapes, except steel 2.742 4.467 6.841 8.239 9.644 16040— Sulfur & nonmetallic minerals 0.164 0.054 0.026 0.004 0.018 16050— Other (synthetic rubber, wood, cork, gums, resins, etc.) 0.084 0.010 0.549 0.328 0.163 16110— Audio & visual tapes & other media 0.000 0.000 0.020 0.000 0.000 16120— Other (boxes, belting, glass, abrasives, etc.) 0.430 0.120 0.182 0.401 0.199 20000— Generators, transformers & accessories 0.375 0.647 0.527 1.216 0.443 20005— Electric apparatus & parts, n.e.c 0.008 0.276 0.433 0.544 0.892 21000— Drilling & oil field equipment & platforms 0.772 0.052 0.017 0.057 0.202 21010— Specialized mining & oil processing equipment 0.000 0.000 0.000 0.219 0.291 21030— Excavating, paving & construction machinery 0.284 0.191 0.168 0.078 0.520 34.255 15.577 22.543 27.693 12.064 21100— Industrial engines, pumps, compressors & generators 0.497 0.168 0.318 0.490 0.773 21110— Food & tobacco processing machinery 0.054 0.186 0.292 0.400 1.063 21120— Machine tools, metal working, molding & rolling 3.013 3.118 4.097 4.574 4.903 21130— Industrial textiles, sewing, & leather working machinery 0.011 0.103 0.024 0.074 0.000 21140— Woodworking, glass working, & plastic & rubber machinery 0.148 0.319 0.742 0.710 0.570 21150— Pulp & paper machinery 0.034 0.122 0.032 0.056 0.148 21160— Measuring, testing & control instruments 0.151 0.461 0.631 0.686 0.632 21170— Materials handling equipment 1.264 0.922 0.526 0.674 0.140 21180— Other industrial machinery 0.249 0.936 1.224 1.311 2.106 21190— Photo & other service industry machinery 0.757 0.687 0.841 0.709 0.739 21200— Agricultural machinery and equipment 2.403 4.637 7.728 7.220 10.719 21040— Nonfarm tractors & parts 21300— Computers 0.000 0.000 0.001 0.013 0.000 21301— Computer accessories, peripherals & parts 0.006 0.000 0.005 0.018 0.011 21320— Semiconductors 0.000 0.000 0.024 0.050 0.043 21400— Telecommunications equipment 0.010 0.010 0.069 0.028 0.010 21500— Business machines & equipment, except computers 0.025 0.142 0.093 0.189 0.076 0.342 21600— Laboratory, testing & control instruments 0.197 0.209 0.138 0.120 21610— Other scientific, medical & hospital equipment 0.095 0.103 0.037 0.083 0.226 22000— Civilian aircraft, complete - all 0.858 0.678 0.610 0.362 0.210 22010— Parts for civilian aircraft 0.334 0.154 0.146 0.223 0.008 22020— Engines for civilian aircraft 1.080 0.815 0.883 0.665 0.583 Digitized forFEDERAL FRASER RESERVE BANK OF ST. LOUIS 45 Product 1988 1989 1990 1991 1992 22100— Railway transportation equipment 0.000 0.107 0.753 2.995 4.603 22210— Other commercial vessels, new and used 0.000 0.000 0.198 0.000 0.000 22220— Marine engines & parts 0.000 0.007 0.000 0.021 0.000 0.000 30000— Passenger cars complete & assembled (new and used) 0.001 0.000 0.000 0.000 30100— Trucks, buses, & special purpose vehicles 0.349 0.357 0.376 0.335 0.214 30110— Bodies & chassis for trucks & buses 0.000 0.000 0.000 0.094 0.000 30200— Engines & engine parts 0.014 0.159 0.233 0.177 0.212 30220— Automotive tires & tubes 0.000 0.000 0.125 0.855 0.660 30230— Other parts & accessories 0.023 0.035 0.028 0.040 0.096 40000— Apparel & household goods-cotton 2.692 2.802 3.593 2.109 1.981 40010— Apparel & household goods-wool 7.677 7.618 7.535 8.635 14.474 40020— Apparel & household goods-other textiles 1.284 0.952 1.330 1.527 1.384 40030— Nontextile apparel & household goods 0.176 0.098 0.301 0.074 0.040 40040— Footwear of leather, rubber, or other materials 1.138 1.699 1.178 1.891 2.439 40050— Sporting & camping apparel, footwear & gear 0.571 0.253 0.151 0.187 0.456 40100— Medicinal, dental & pharmaceutical preparations 1.728 1.006 0.293 0.485 0.435 40110— Books, magazines & other printed material 0.122 0.118 0.150 0.059 0.174 40120—Toiletries & cosmetics 0.055 0.025 0.032 0.012 0.046 40140—Other products (notions, writing & art supplies) 0.010 0.014 0.069 0.094 0.078 41000— Furniture, household items, baskets 2.617 2.883 3.681 3.452 2.989 41010— Glassware and porcelain 5.992 14.285 19.087 25.244 28.039 41020— Cookware, chinaware, cutlery, & other household goods 0.745 0.489 0.720 0.437 0.330 41030— Household & kitchen appliances 0.120 0.116 0.553 1.337 1.084 41040— Rugs & other textile floor covering 0.295 0.053 0.170 0.135 0.166 41050— Other (clocks, portable typewriters, other household goods) 2.462 1.848 2.991 3.058 2.453 0.071 41100— Motorcycles & parts 0.000 0.002 0.000 0.000 41110— Pleasure boats & motors 0.048 0.076 0.297 0.000 0.009 41120— Toys, shooting & sporting goods, & bicycles 0.448 0.488 0.507 0.316 0.385 41130— Photo & optical equipment 0.000 0.044 0.052 0.043 0.052 41140— Musical instruments & other recreational equipment 0.015 0.007 0.071 0.407 0.160 41200—Television receivers, vers & other video equipment 0.000 0.000 0.000 0.000 0.029 41210— Radios, phonographs, tape decks & other stereo 0.000 0.000 0.007 0.007 0.014 41220— Records, tapes & disks 0.160 0.066 0.125 0.380 0.505 41300— Numismatic coins 6.560 1.878 0.112 0.807 0.975 41310—Jewelry (watches, rings, etc.) 0.014 0.034 0.092 0.112 0.144 41320— Artwork, antiques, stamps and other collectibles 0.084 0.674 0.234 0.309 0.365 42000— Nursery stocks, cut flowers, Christmas trees 1.269 0.000 0.061 0.042 0.380 42100— Gem diamonds-uncut or unset 0.000 0.000 0.024 0.000 0.000 42110— Other gem stones-precious, semiprecious & imitation 0.050 0.022 0.044 0.103 0.191 50000— Military aircraft & parts 0.004 0.000 0.163 0.221 1.420 50010— Other military equipment 0.105 0.000 0.130 0.311 0.373 50020— U.S. goods returned, & reimports 0.422 0.462 0.332 0.412 0.515 50030— Minimum value shipments 0.727 0.567 0.535 0.535 0.728 50040— Other (movies, miscellaneous imports & special transactions) 0.372 0.170 0.728 0.246 1.640 JULY/AUGUST 1994 47 John C. W eicher John C. Weicher is a senior fellow at the Hudson Institute and a visiting scholar at the Federal Reserve Bank o f St. Louis. Heidi L. Beyer provided research assistance. The New Structure of the Housing Finance System A, . FTER 25 YEARS OF ECONOMIC evolution and 15 years of political turmoil, the U.S. housing finance system has changed in fundamental ways, and the structure of the new system is becoming apparent. The system is still intended to allocate credit to housing and hold mortgage rates below their free-market level, but this subsidy is provided through different institutional arrangements. The dominant institutions are now extremely large government-sponsored enterprises (GSEs] which operate in the secondary mortgage market, issuing securities that are backed by mortgages and buying mortgages originated by other institutions. They have taken the place of small local savings and loan associations which make loans directly to homebuyers and hold the mortgages in their own portfolios. The cost of the subsidy falls on tax payers who bear the risk of failure by the GSEs. The public purpose of government intervention in the housing finance system has always been to promote homeownership, giving force to a social preference that derives from a widely held but rarely analyzed belief that homeowners are better citizens, because they “have a stake in society.” Subsidies are appropriate because families will not take this social benefit into account in deciding whether to buy a home. The system also has two subsidiary goals: (1) countercyclical support for housing production; and (2) geographic equity (as defined by public policy) in the mortgage market. The latter is more directly related to the purpose of promoting homeownership. Achieving these purposes is the responsibility mainly of privately owned institutions which are supposed to meet them while maximizing profit and avoiding direct cost to taxpayers. This is also true of the major housing finance agencies within the federal government; they do not normally receive funds from the U.S. Treasury. The private as well as public institutions operate under statutes which define their powers, limitations and privileges, and delineate what they can hold as assets and liabilities. To some extent, they compete against each other. The home mortgage market consists of some $3 trillion of household debt, nearly all of it held by private institutions, of which more than $1 trillion is explicitly or im plicitly guaranteed by the federal government (not counting deposit insurance). There is continuing tension between the public purposes of the system and the safety and soundness— and profits— of the privately owned institutions that predominate in it. This paper first describes the present structure of the housing finance system, contrasting it with the traditional system and explaining why the system has changed. It concludes with a discus sion of the major issues that will face public policy over the next few years. JULY/AUGUST 1994 48 THE CURRENT SYSTEM Mortgage and Housing Markets The mortgage market has traditionally been separate from other capital markets, because mortgages differ in key respects from other debt instruments. A home mortgage is a loan to an individual or couple for the purpose of buying a particular house. The amount the lender is willing to loan depends on the value of the prop erty. The mortgagor promises to repay the loan over time. If he or she fails to do so, thus default ing on the mortgage, the lender can foreclose, take title to the property, and sell it to someone else. The default risk (also termed credit risk) depends on changes in the value of the property and in the circumstances of the owner, such as job loss, divorce, or the death of a husband or wife. The extent of loss in the event of default depends on changes in property values. Real estate markets are local markets, so evaluating a particular piece of property requires local expertise. For these reasons, mortgages have traditionally been illiquid; investors have been willing to buy them only if they have the knowl edge required to evaluate them. The standard mortgage instrument—until the 1980s, almost the only mortgage instrument—is a fixed-rate, level-payment, long-term, self-amor tizing loan. The term usually runs for 30 years. Such mortgages carry prepayment risk as well as default risk. The term can be shortened only at the option of the borrower, by prepaying the loan. If market interest rates fall, as happened most recently during 1993-94, mortgagors are likely to prepay their loans and refinance their homes at a lower interest rate. If rates rise, mortgagors are unlikely to prepay unless they are moving, and lenders will find themselves earning belowmarket rates on their mortgage portfolios. Thus, lenders bear the risk of adverse movements in interest rates. This was a particular problem during the inflation of the late 1970s. As a result, mortgages with adjustable rates (ARMs) were authorized and became common in the early 1980s. A variety of other new instruments also have come into existence, such as balloon mort gages, which are not fully amortized over their term, and graduated payment mortgages, which carry fixed interest rates but have lower payments in the early years. The traditional standard mortgage remains the most common, although its popularity relative to ARMs has varied as interest rates have fluctuated. Fixed-rate mort gages are more popular among borrowers when Digitized for FEDERAL FRASER RESERVE BANK OF ST. LOUIS the general level of interest rates is low and ARMs are more popular when the level is high. The dominant position of the standard mort gage developed under the auspices of the federal government, specifically the Federal Housing Administration (FHA), which was created in the 1930s to insure home mortgages. Buyers pay a mortgage insurance premium to FHA, and in return FHA guarantees that lenders w ill receive payment of the outstanding principal balance on the mortgage in the event of default and foreclo sure. FHA is required by law to operate on an actuarially sound basis; its insurance premiums are supposed to cover its losses and operating costs. Once FHA demonstrated the viability and profitability of such mortgages, the “FHA mort gage” also became the norm for conventional mortgages (those not insured or guaranteed by a government agency). Primary Lenders The local nature of real estate markets has meant that mortgage lenders have traditionally been local institutions. The most important have been the savings and loan associations (S&Ls), both in terms of their share of the mort gage market and the share of mortgages in their portfolios. The S&Ls started as local specialized mortgage portfolio lenders, obtaining deposits from “small savers” within their locality and making mortgage loans there also. Until 1983 their lending areas were geographically limited by statutes and regulation. Typically, their deposits have been locally generated as well, though there has been no geographic limitation on liabilities. Mutual savings banks, concentrat ed in the Northeast, are similar to S&Ls as spe cialized mortgage portfolio lenders, but they developed independently and started with a dif ferent purpose: to provide a range of financial services to households. Savings banks and S&Ls together are usually termed “thrifts.” They have access to the national capital markets through the Federal Home Loan Bank System, a set of 12 regional Federal Home Loan Banks which they own. Chartered and regulated by the federal government, the Home Loan Banks are able to borrow at preferential rates in the capital mar kets. Commercial banks also hold a significant share of home mortgages, but mortgages com prise a minor fraction of their assets. Until 1989 they could not belong to the Home Loan Bank System; they now can if their mortgage holdings are large enough. 49 S&Ls, mutual savings banks and commercial banks are primary lenders; they originate mort gages which they hold in their own portfolios. A large number of mortgages, however, are origi nated by mortgage bankers, for immediate sale in the secondary market to an investor who expects to hold them. Mortgage bankers make their money from fees for originating mortgages and often for servicing them, collecting the monthly payments and transmitting them to the investor. Mortgage banking developed as an important component of the housing finance system when FHA began to insure mortgages. FHA insurance and its uniform national under writing standards meant that specialized knowl edge of local housing markets was less important for investors. Mortgage bankers have since developed the skills necessary to originate conventional mortgages and now originate just under half of all home mortgages. Securitization and the GovernmentSponsored Enterprises The limitations of mortgages as investment vehicles led to the creation of mortgage-backed securities (MBSs) beginning in 1970. Mortgage securitization consists of combining a group of mortgages into a pool and selling shares in the pool to investors. This spreads the risk of default over a number of mortgages and allows investors to calculate the probability of default for the mortgages in the pool with more accuracy than for any individual mortgage. The earliest and simplest M BSs are known as pass-through secu rities; the servicer collects principal and interest payments and passes them through, without taxation, to the investor. The pass-through security does not reduce prepayment risk. More recent forms, the Collateralized Mortgage Obligation (CMO) and the Real Estate Mortgage Investment Conduit (REMIC), partition the principal cash flow from a pool of mortgages or MBSs into maturity class es, or tranches. Each investor receives a propor tionate share of all interest payments. Principal payments are allocated in full as they occur to each tranche in turn, starting with the shortestmaturity tranche. Tranches are separately priced and sold to investors with different time horizons. Investors in the longer-term tranches 1 The collateral for CMOs and REMICs can be either whole mortgages or MBSs. Thus, the ratio of CMOs and REMICs to all MBSs does not represent either their share of all MBSs or their share of all securitized mortgages. incur greater interest rate risk, but have some protection against prepayment. They are not, however, protected in the event of a protracted decline in interest rates; if mortgagors prepay in large numbers, the securities will be redeemed. This type of security was developed in 1983 and it now accounts for about half of all mortgages that are securitized.1 Securitization has broadened the mortgage market by creating instruments that appeal to investors without special knowledge of local housing markets. The payment streams are sim ilar to bonds, and the consequences of default and prepayment are minimized. Until the last few years, the market for MBSs has generally required a government guarantee on the mortgages, the securities, or both. Securitization was developed by a government agency, the Government National Mortgage Association (GNMA or Ginnie Mae). GNMA issued pass-through securities based on pools of FHA-insured mortgages, and added its own guarantee of timely payment of principal and interest to the FHA guarantee of principal pay ment in case of default. (GNMA also issues securities backed by pools of mortgages guaran teed by VA, originally the Veterans Administration and now the Department of Veterans Affairs, in a program created after World War II and modeled on FHA.) The other major MBS guarantors are the housing GSEs: the Federal National Mortgage Association (FNMA or Fannie Mae) and the Federal Home Loan Mortgage Corporation (FHLMC or Freddie Mac).2 Both also buy and hold mortgages in their own portfolios, financing them by issuing debt securities in the capital markets, and in fact they have the first- and fourth-largest mortgage port folios, respectively, of all mortgage lenders in the United States. They are secondary market agen cies; they do not originate mortgages but buy loans from primary lenders or mortgage bankers. FNMA and FHLMC are privately owned insti tutions with stockholders and private boards, but they are federally chartered corporations with a variety of special privileges, and the President appoints five members out of 18 to the board of each one. The most important of the privileges are exemption from state and local 2 GNMA can also be considered a GSE, but it is fully owned by the federal government and operates as an agency within HUD, as is FHA. JULY/AUGUST 1994 50 income taxes, exemption from Securities and Exchange Commission registration and state securities laws, the ability to borrow $2.25 bil lion from the U.S. Treasury in an emergency, and the fact that their debt securities are “quali fied investments” for regulated financial institu tions. Their securities are also issuable and payable through Federal Reserve Banks.3 These privileges give them “agency status” in the capital markets, a general perception that the government w ill stand behind them. This per ception is reinforced by the fact that both are very large financial institutions, “too big to fail.” Agency status allows them to borrow at relatively low rates and to issue or guarantee payment on securities based on pools of conventional mort gages. The market treats their guarantees of timely payment of principal and interest as equivalent to a government guarantee.4 The Role o f the F ed era l Government It should be clear that the federal government has a large role in the housing finance system. It insures some mortgages; it issues securities backed by pools of those mortgages; and it has chartered corporations which are believed in the capital markets to have an im plicit government guarantee behind their debt securities and their mortgage securities. In addition, it regulates and insures the deposits of primary lenders, and has chartered institutions which provide them with access to the capital markets. The federal gov ernment is generally credited with conducting three successful social experiments in the mort gage market: demonstrating the feasibility of long-term, self-amortizing loans; mortgage insur ance; and securitization. In all three cases, the private sector has successfully copied the feder al models. The FHA mortgage became the stan dard for conventional mortgages, and a private mortgage insurance industry has developed to insure them. The private sector now accounts for more business than the federal government in both instances. The demonstration that secu ritization is feasible has been followed by a sub 3 Most of these privileges date back in some form to the FNMA Charter Act of 1954 or its initial 1938 charter. In 1970 the same privileges were extended to FHLMC when it was created. 4 The securities may be issued directly by a GSE or alterna tively by a subsidiary of a private entity such as a Wall Street firm, with the GSE guaranteeing the timely payment of prin cipal and interest. 5 These statements are based on unpublished tabulations of FHA-insured loans between 1989 and 1992. The tabula- Digitized for FEDERAL FRASER RESERVE BANK OF ST. LOUIS stantial volume of private M BS activity only since about 1990, however, and private securi ties are still a small part of the total market. The Dividing Lines The federal government also demarcates the market segments of the various institutions by means of two statutory numerical concepts: the FHA ceiling and the conforming loan limit. The FHA ceiling is the maximum principal balance on a mortgage that FHA can insure. The ceiling is set in law in nominal dollars; since 1980 higher amounts have been allowed in areas with higher housing costs. The present ceiling is $67,500, or 95 percent of the area median home price if that is higher, up to a maximum of $151,725. The maximum is still less than 95 percent of the area median home price in a num ber of large markets, among them New York and the largest metropolitan areas on the West Coast, and it is raised every few years. FHA insurance is intended for the first-time homebuyer who can only afford a relatively small down payment, and who thus poses a greater risk of default to the lender. Most FHA buyers make a down payment of five percent or less.5 Below the ceiling, nearly all low down payment loans are insured by FHA and securi tized by GNMA. The conforming loan limit is the maximum principal amount of a mortgage that FNMA and FHLMC can buy. Before 1974, they were restricted to mortgages with principal amounts below the FHA ceiling. A higher limit was set by statute in that year. In 1977 the limit was set at 25 percent above the maximum mortgage amount for S&Ls, and both were raised in 1979. After the S&L maximum was abolished in 1980, the conforming loan limit was set by statute at its then-current value of $93,750, and indexed on the basis of the annual percentage change in the mean price of homes bought with conven tional mortgages. Since 1980, the limit has been about 37 percent above the mean price. In 1993 tions differ from published data because FHA allows part (before 1991, all) of the closing costs to be financed in the mortgage, as is discussed in Price Waterhouse (1990, pp. 18-19). The published data do not adjust the loan-to-value ratios to reflect financing of closing costs, and therefore show somewhat lower loan-to-value ratios. For example, Price Waterhouse (1990, p. 17), reports that in 1988-89 slightly less than half of FHA-insured mortgagors had loanto-value ratios above 95 percent. 51 the conforming loan limit was $203,150, while the mean price was $144,000.6 Because of their agency status and the perception that they are too big to fail, the GSEs can offer lower interest rates than the S&Ls and therefore dominate the market below the conforming loan limit. Estimates of their cost advantage are in the range of 20 to 35 basis points.7 The S&Ls remain as portfolio lenders above the conforming loan limit. Below the limit, they operate largely as mortgage bankers. They origi nate mortgages not for their own portfolios but for sale to the GSEs, although they may buy back the securities issued on a pool of the mortgages that they have sold. Regulations issued under the Financial Institutions Reform, Recovery and Enforcement Act of 1989 (FIRREA) gave the S&Ls an incentive to move away from portfolio lending by setting capital requirements only 40 percent as high against MBSs issued by the GSEs as against whole mortgages. Other thrifts and commercial banks have a similar role. The mortgage market, therefore, can be divided into the FHA/ GNMA submarket, for low down payment loans below the FHA ceiling; the GSE submarket, for most other loans below the con forming loan limit; and the “jumbo” submarket, occupied by S&Ls, other thrifts and commercial banks, for loans above the conforming loan limit. mortgages that were above the lim it when issued may be below it a few years later. Thus, the share of the market open to the GSEs is larger when measured in terms of all outstanding mortgages than it is when measured in terms of mortgages originated in the current year. Primary lenders do make loans below the con forming loan limit. Some are nonstandard loans which the GSEs do not choose to purchase, but mortgages above as well as below the limit are likely to be underwritten to the guidelines of the GSEs, to keep open the option of selling them in the secondary market. In addition, the conforming loan limit appar ently only adjusts in one direction. Declining house prices during 1993 resulted in a reduction of $6,050 in the calculated conforming loan limit for 1994, as reported by the Federal Housing Finance Board. FNMA and FHLMC announced that they would not lower the limit, because the 1980 statute referred only to “increases,” and not to “decreases” or “changes.”8 HUD Secretary Henry Cisneros, as GSE regulator, first challenged this action and then accepted it. THE GROWING DOMINANCE OF THE GSEs Actual market segmentation, however, is less clear-cut than the dollar demarcations suggest. The FHA ceiling only applies to FHA. The GSEs, the S&Ls and any other lender can originate mortgages below the ceiling, and some such loans are made. To some extent, the private mortgage insurers compete with FHA by offering insurance on loans with low down payments, though they do not insure a large share of these mortgages. The GSEs have been the thriving and expand ing institutions in the system. In 1992, the latest year for which full data are available, FHA and VA loans were about 10 percent of the total dol lar volume of all home mortgages issued, nonconforming loans about 20 percent, and conven tional conforming loans about 70 percent. The dominant role of FNMA and FHLMC in the con forming loan market is reflected by the fact that they securitized over half of these loans and added to their mortgage portfolios as well. Similarly, the conforming loan limit applies only to the GSEs. The S&Ls and other primary lenders can make loans below the limit. But the conforming loan limit is much less restrictive on the GSEs than the FHA ceiling is on FHA. Since house prices in most years rise at least modestly, The growth of the GSEs is shown in Table 1, which depicts the mortgage market in terms of the total dollar volume of loans outstanding at various dates. The GSEs now hold or securitize about 30 percent of the total, compared to about 7 percent in 1980. Since 1980 they have accounted 6 The annual adjustment is based on the percentage change in prices of homes sold during the last five business days in October. 8 The staff director of the Senate Housing Subcommittee as of 1980 has stated that the intent of the statute was that the limit should move in accord with house prices in both direc tions, but prices had risen so long and so much by 1980 that nobody remembered the possibility of a decrease when the bill was written. 7 See, for example, ICF (1990) and Hendershott and Shilling (1989). The former estimates a differential of 23 basis points as of 1987, the latter 30 to 35 basis points as of 1986. Both apply to loans that are at least 15 percent above the conforming loan limit and, therefore, unlikely to be sold in the secondary market when they are seasoned. These are apparently the most recent analyses. JULY/AUGUST 1994 52 Table 1 Single-Family Mortgage Debt Outstanding, 1968-92 (end-of-year values) Dollar Values (billions of current dollars) 1968 Portfolio lending FNMA portfolio FHLMC portfolio S&Ls Commercial banks Others* Subtotal 7 0 110 39 109 265 1980 52 4 411 160 1989 91 18 512 1992 124 31 375 452 229 856 336 529 1486 0 0 0 0 27 20 63 110 200 129 592 921 1380 0 0 0 0 0 265 0 14 92 4 110 965 220 266 358 77 921 2408 436 402 411 132 1380 2954 591 1573 Security holdings** S&Ls Commercial banks Others* Subtotal Securities issued** FNMA MBSs FHLMC PCs GNMA MBSs Private pools Subtotal Total Percent Shares 159 307 914 1968 1980 1989 1992 2.6 0.0 41.5 14.7 41.1 100.0 5.4 0.4 42.6 16.6 23.7 88.7 3.8 0.7 21.3 14.0 4.2 1.0 12.7 15.3 22.0 61.7 20.0 53.3 Security holdings** S&Ls Commercial banks Others* Subtotal 0.0 0.0 0.0 0.0 2.8 1.9 6.5 11.4 8.3 5.4 24.6 38.2 5.4 10.4 31.0 46.7 Securities issued** FNMA MBSs FHLMC PCs GNMA MBSs Private pools Subtotal 0.0 0.0 0.0 0.0 0.0 0.0 1.5 9.5 0.4 11.4 9.1 11.1 14.9 14.8 13.6 13.9 4.5 46.7 Portfolio lending FNMA portfolio FHLMC portfolio S&Ls Commercial banks Others* Subtotal 3.2 38.2 * Others include mutual savings banks, life insurance companies, finance companies, the Farmers Home Administration, the Federal Housing Administration, the Veterans Administration (Department of Veterans Affairs in 1989 and later), mortgage companies, real estate investment trusts, state and local credit agencies, state and local retirement funds, noninsured pension funds, credit unions, other U.S. government agencies, and individuals. ** Security holdings show the distribution of securities issued. Either can be added to portfolio lending data to obtain the totals; both cannot be added without double counting. Security holdings can be added to data on portfolio lending to show mortgage market activity of thrifts, banks and other institutions; securities issued can be added to data on portfolio lending to show mort gage market activity of the GSEs. SOURCES: Board of Governors; U.S. Department of Housing and Urban Development; Inside Mortgage Capital Markets; Inside Mortgage Securities; Savings and Loan Fact Book. for over 40 percent of the net increase; since 1989, over 70 percent. This is nearly the entire con ventional conforming loan market. The new interpretation of the conforming loan limit allows them to further increase their market share. FHA and VA insure or guarantee a gradually declining share of home mortgages, as Table 2 shows. It is not feasible to calculate noncon forming loans as a fraction of the outstanding stock of mortgages at any given time. The non conforming market has been stable at about 20 to 22 percent of all conventional mortgages origi nated in a given year, measured in terms of dollar volume, but there has been a fairly steady shrinkage when measured in terms of the number of mortgages, from 11.6 percent in 1984 to 6.4 per FEDERAL RESERVE BANK OF ST. LOUIS cent in 1993. Even the stable dollar share of annual originations implies a declining share of all mortgages outstanding, as the conforming loan limit rises from year to year. The housing finance system is an emerging duopoly, dominated by the two large GSEs. Other institutions are increasingly limited to segments of the market which are effectively barred to the GSEs by statute, and which are declining in importance. The dominant position of the GSEs is rein forced by their relationship to other market institutions. Thrifts and banks are both their competitors and their customers. They compete as portfolio lenders, but at the same time they sell 53 Table 2 Government-Guaranteed and Conventional Mortgages, 1968-92 1968 1980 1989 1992 51 34 94 102 180 265 770 965 283 157 1968 326 164 2464 2954 Dollar Values (billions of current dollars) FHA VA Conventional Total 19.2 12.8 67.9 9.7 10.6 79.8 11.8 6.5 81.7 11. 5. CO 00 Percent Shares FHA VA Conventional 2408 Source: Dept, of Housing & Urban Development, Survey of Mortgage Lending Activity. mortgages to the GSEs and buy mortgage securities from them, and also buy the debt securities that the GSEs use to finance their portfolios. THE TRADITIONAL SYSTEM: A COMPARISON This is very different from the housing finance system as it existed about 25 years ago. In 1968 it was still recognizably the New Deal system. It was dominated by the S&Ls, which gathered deposits locally, borrowed from the Home Loan Banks during recessions or when rates were high, and made long-term, fixed-rate loans (up to a maxi mum of $40,000) on homes located within 50 miles of their home offices (100 miles after 1964). The Federal Home Loan Bank Board (the Bank Board) regulated and supervised both the S&Ls and the Home Loan Banks, and insured the S&Ls through the Federal Savings and Loan Insurance Corporation (the FSLIC). FHA was losing business to the S&Ls, and at the same time taking on greater credit risk, because of the growing private mortgage insur ance industry. FHA had a single premium for 9 This paragraph is based on Kaserman (1977). 10 See Fredrikson (1971) and the data and literature therein cited on changes in regional differentials. Actual average mortgage rates varied by about 1 percentage point between the Northeast at the low end and the South and West at the high end, and may have risen slightly between 1940 and 1963. These rates are not risk-adjusted, but Fredrikson makes adjustments for loan-to-value ratio and term, and finds they have little effect on the regional differentials. all loans; on the “principle of cross-subsidization,” profits on the less risky loans were supposed to subsidize losses on those with higher loan-to-value ratios. The outcome was that FHA lost the better loans to the conventional market, because private mortgage insurers could charge a lower premium on the less risky loans.9 The only secondary market agency was FNMA, established in 1938 as a fully governmental agency and limited to FHA and VA mortgages. Its purpose was to smooth out the flow of mort gage credit, over time and between places. Securitization had not yet been invented; FNMA issued bonds and bought mortgages from primary lenders and mortgage bankers. It was supposed to be a dealer, selling as well as buying mortgages. It had begun to operate as a portfolio lender in the post-war period, however, buying VA mort gages in particular, until directed to liquidate its portfolio by the 1954 FNMA Charter Act. Its portfolio then fluctuated between $2 billion and $3 billion until 1965. At that point it again began to buy mortgages in large volume. Its portfolio reached $7 billion in 1968, less than 3 percent of the total market. This system was considered a success in terms of its policy objectives. Housing production reached unprecedented levels in the post-war period; the pre-war peak of 937,000 housing starts in 1925 was eclipsed in 1946 and indeed in all later years. There was also a remarkable increase in homeownership, from 44 percent of all households in 1940 to 55 percent in 1950 and 62 percent by 1960. Total home mortgage debt doubled in the first five years after the war and doubled again in the next five years. Not all the goals were met; regional differences in mort gage rates probably did not diminish, but this was a secondary concern.10 Contemporary econ omists were divided over whether the housing finance system was a major contributor to these outcomes, or whether the same results could have been reached some other way, but as a policy matter the system was credited with the successes that occurred.11 11 Grebler, Blank and Winnick (1956) argue that the changes in the housing finance system were important; Saulnier, Halcrow and Jacoby (1958) conclude they were not. JULY/AUGUST 1994 54 Several problems were inherent in this system, and by the late 1960s it was already starting to break down. The S&Ls were expected to incur interest rate risk routinely. They borrowed short and they had to lend long. If interest rates rose, their cost of funds would rise faster than the earnings on their long-term mortgage portfolios. Second, they operated under a kind of one-way Glass-Steagall Act. They had no protection from competition on either side of the balance sheet. They could issue only time deposits and had to specialize in home mortgages. Commercial banks could, however, also issue time deposits and make mortgage loans. Third, the geographic lending restrictions meant that S&Ls incurred credit risk from local as well as national eco nomic changes. This had already occurred on a large scale when the Florida land boom collapsed in the late 1920s.12 CHANGES IN THE SYSTEM All these potential problems became real ones after 1965 and forced changes in the system. The disintegration that began with the onset of inflation in the late 1960s has been described and analyzed by many economists.13 In this paper only a brief review of the process of change is necessary. Table 1 traces its course. It shows the importance of different institutions in the mortgage market in 1968, as the New Deal system was starting to unravel: in 1980, when policy makers were forced to recognize that the S&Ls could no longer function as portfolio lenders in an inflationary world; in 1989, when FIRREA w as p assed to ad d ress the losses of the S&Ls and the insolvency of the FSLIC; and in 1992, the latest year for which information is available. Inflation and the D ecline o f the S&Ls The S&Ls remained the dominant institutions in the mortgage market during the 1970s. They held over 40 percent of all home mortgages in 12 Between 1927 and 1929, 40 percent of the S&Ls in Florida, with almost half of the assets, went out of business, while S&Ls in the rest of the country were expanding. See Bodfish (1931) for these data. 13 For analyses of the problems of the S&Ls during the 1970s and 1980s, see Barth (1991) and White (1991); White was a member of the Federal Home Loan Bank Board and Barth was chief economist there during the late 1980s. Jones (1979) describes the policy process during the 1970s. The most extensive analysis of the secondary market is U.S. Department of Housing and Urban Development (1987). Weicher (1988) describes the changes in the system as a whole and the developing problems that led to the passage of FIRREA. Digitized for FEDERAL FRASER RESERVE BANK OF ST. LOUIS 1980 as they did in 1968, and they accounted for almost half of the net increase in mortgages outstanding over the period. But this was increasingly against their will. Once inflation began to accelerate, they could not finance their portfolios of fixed-rate long-term mortgages with short-term deposits, unless depositors would accept below-market rates. The small saver proved unwilling to subsidize the homebuyer, if alternative investments paying market rates were available. Money market mutual funds (MMMFs) were such an investment. Beginning in 1972 the S&Ls’ cost of funds began to rise relative to short-term Treasury rates.14 By 1980 the net income of the S&Ls as a whole was approaching zero, tangible net worth was start ing to fall, and the industry had a net worth in market value terms variously calculated to be between -8 and -19 percent of its assets.15 Major legislation was enacted in 1980 and 1982 that liberalized both the asset and liability sides of S&L balance sheets. The goal of saving them as institutions took precedence over the goal of promoting housing. They were allowed to make loans and direct investments outside of housing altogether, up to 40 percent of their assets. Between 1980 and 1989, they accounted for less than a quarter of the increase in out standing mortgages, and more than half of their growth took the form of security purchases rather than portfolio lending. By 1989 they held only about 30 percent of outstanding mortgages either in portfolio or as securities. Since the passage of FIRREA, closure of failed S&Ls has red u ced the total portfolio of the ind u stry by over $175 billion, and their share of the market is now under 20 percent. The Evolution o f the Secondary M arket The dates in Table 1 also represent stages in the evolution of the secondary market. Between 14 The Eleventh District Cost of Funds rose from 9 basis points below the three-month Treasury rate in 1972 to 61 basis points above it in 1987. (The Eleventh District Federal Home Loan Bank is located in San Francisco and the district includes the states of Arizona, California and Nevada. Its Cost of Funds, measuring the interest rate on deposits paid by S&Ls in the district, is one of the common indices for ARMs.) 15 See, for example, Brewer (1989), Brumbaugh (1988) and Kane (1985). Brewer’s estimate is the lowest, Kane’s the highest. Brumbaugh’s is -12.5 percent. 55 1968 and 1980, the secondary market took on its present institutional structure, securitization was invented, and the first mortgage securities won market acceptance. Two legislative changes— one in 1954 requir ing FNMA to buy mortgages on privately owned low-income housing projects subsidized by the government, and the other in 1967 treating the purchases as a federal budget outlay—resulted in splitting FNMA into two agencies in 1968 and changing its role.16 GNMA was created to take responsibility for the low-income mort gages, and FNMA went off-budget as a federally chartered corporation with agency status in the capital markets. In 1970, amid concerns about rising interest rates and a new “credit crunch” in the primary mortgage market as a result of Reg Q, policymakers responded by turning to the secondary market. In the Emergency Home Finance Act of 1970, FNMA was given authority to buy conventional mortgages as well as those guaranteed or insured by the federal government. At the same time, the S&Ls acquired their own federally chartered secondary market facility, the Federal Home Loan Mortgage Corporation (FHLMC). FHLMC was supposed to buy the S&Ls’ current conven tional mortgage portfolios to “free up” funds for new loans—in effect trying to shift the conse quences of monetary policy to other sectors of the economy. Like the Home Loan Banks, FHLMC was wholly owned by the S&Ls.17 Thus, where there had been one secondary market agency in 1967, there were three in 1970. They had different ownership and they acted in different ways. GNMA operated as a secondary market agency very much in the original intent of the New Deal system. It did not buy and sell mortgages, but it achieved the same result by issuing securities. 16 Under the 1967 federal budget reform, purchases of subsi dized mortgages were raising outlays on a dollar-for-dollar basis, even though part of the principal and interest on the mortgage would be paid to the government by the borrower. Subsequent budget reforms have changed this accounting practice. Under current law, the entire principal amount of a mortgage purchased or insured by the federal government is counted in the credit budget, but only the anticipated subsidy is included as a outlay in the administrative budget. It created the first M BSs in 1970; by 1980 it was securitizing virtually all new FHA and VA loans, and its M BSs accounted for almost 10 percent of all outstanding mortgages. FHLMC followed GNMA into the securities business on a much smaller scale and became primarily an issuer of MBSs backed by conventional mortgages. FNMA took a different route. During the 1970s, it turned itself into the largest conventional mortgage portfolio lender and thus, in effect, the largest S&L in the country, albeit with different sources of funds. Its experience paralleled that of the S&Ls. It did not foresee the inflation of the 1970s, so that its net worth also turned nega tive in the late 1970s: its 1980 value of -16 per cent was similar to the S&L industry.18 During the 1980s, securitization accounted for over half the growth in the total volume of mortgage credit. In 1981 both FNMA and FHLMC initiated mortgage swap programs, buying S&L portfolios and issuing pass-through securities on exactly the same mortgages in return. This brought FNMA into the business of issuing securities, rather belatedly. Since then, both its portfolio and its M BS volume have grown rapidly. Its outstanding M BSs are now 3.5 times the size of its portfolio, but it remains the largest portfolio lender. Besides issuing pass-throughs to S&Ls, FHLMC created the CMO in 1983 and expanded its securities business almost twentyfold. FIRREA marks a further stage in the evolution of the secondary market. It turned FHLMC into nearly a carbon copy of FNMA, giving it exactly the same kind of board of directors and a very similar charter. After FIRREA, the secondary market institutions assumed a dominant position in the mortgage market. Between 1968 and 1980, about 80 percent of the net increase in mortgages was held in portfolio and 20 percent was securi tized; since 1989 the proportions have been holdings of FHA and VA mortgages, and also that FNMA’s debt issuances would drive up interest rates and raise the cost of funds to the S&Ls and other primary mortgage lenders. Burns raised the issue of illiquidity and expressed concern that FHLMC would drive up interest rates on FHA and VA mortgages. See U.S. Department of Housing and Urban Development (1987). 18 U.S. Department of Housing and Urban Development (1987), based on Kane and Foster (1986). 17 Opposition to allowing FNMA to buy conventional mortgages was stated by Federal Reserve Chairman Martin in 1969 and opposition to creating FHLMC was stated by Federal Reserve Chairman Burns in 1970. Martin expressed con cern that FNMA’s conventional mortgage portfolio would be illiquid and, therefore, might ultimately displace FNMA’s JULY/AUGUST 1994 56 reversed. GSE portfolios also continued to grow as a share of the market. Did the System “Work” A fter 1965? Housing advocates opposed many of the policy changes for fear they would weaken the ability of the system to allocate credit to housing, regardless of the consequences for the financial system or the economy. This was the key concern preventing financial reform in the 1970s: If the S&Ls were allowed to diversify, who would “fill the gap” in the mortgage market? This concern proved to be unfounded. The secondary market agencies filled the gap, as Table 1 shows, and the volume of outstanding mortgages almost tripled between 1980 and 1989. A second concern was the cost of mortgage credit. This too has proved to be largely unfounded. The spreads between mortgage and bond rates may have widened during the 1970s and early 1980s, but by the late 1980s the conventional mortgage market may have become fully integrated with the capital markets.19 Here also, the growing role of the GSEs offset the declining presence of the S&Ls, holding down the mortgage rate in the conforming loan market. Regional differences in mortgages rates probably disappeared by the m id-1980s, as a result of securitization and deregulation. Available data on interregional flows of mortgage funds suggest that securitization resulted in transfers from the Northeast and Midwest to the South and West, where population and housing demand were growing, and, as already noted, in 1983 the S&Ls were allowed to make mortgage loans anywhere in the country.20 The developing S&L crisis in the 1980s also helped create a national mortgage market. One way in which the Bank Board 19 Hendershott and Van Order (1989) conclude that the interest rate on conventional, fixed-rate mortgages rose by about 100 basis points between the late 1970s and the mid-1980s, compared to the rate that would have prevailed in a perfect market; then it fell by about 50 basis points between 1986 and 1988 to the market rate. Other studies covering a shorter period and comparing the mortgage and bond rates also find that the conventional mortgage rate began rising in relative terms sometime in the 1970s, before deregulation of the S&Ls. See, for example, Kaufman (1981), Tuccillo, Van Order and Villani (1982) and Hendershott, Shilling and Villani (1983). The Hendershott and Van Order study ends in 1988, and there does not appear to be any more recent analysis of the spread; given the year-to-year fluctuations in the spread which they calculate, it would be desirable to see more recent data before concluding that the actual conventional conforming loan rate is the same as the rate in a perfectly competitive market. Cotterman (1994) notes that the spread between the MBS and Treasury rates fluctuated between 1984 and 1990, and was at its lowest level in 1988 and 1990. Digitized for FEDERAL FRASER RESERVE BANK OF ST. LOUIS handled failing S&Ls was to sell them for their franchise value to other S&Ls—at first in the same market, then in the same state, then in other states, as the number and severity of failures rose and the financial resources of the FSLIC were increasingly inadequate. The system had less success in achieving its other purposes. Since 1965 the homeownership rate has fluctuated in a fairly narrow range, between 62 and 66 percent of all households.21 There was a notable increase among young families during the 1970s, but this was simply a result of inflation. Young families bought homes as soon as they could because owning a home was the best inflation hedge available, especially as inflation pushed them into higher marginal tax brackets. In the 1980s disinflation and reductions in marginal tax rates caused their homeownership rate to drop quickly back to its 1970 level. Housing cycles, like economic cycles in general, became more pronounced. Record years of over 2 m illion housing starts annually in 1971-73 were followed by a postwar low in 1975; another year of 2 m illion in 1978 was followed by new lows in 1981 and 1982. The housing finance system could not have been expected to offset completely the effects of the oil shocks and other macroeconomic changes, but it is doubtful if it achieved its stated more modest objective of mitigating their impact on housing and shifting part of the consequences to other sectors. The S&Ls used advances from the Home Loan Banks to offset deposit outflows, and this may have had some effect. FNMA may also have mitigated the cycles to a lesser extent through the late 1970s, but it was not aggressively countercyclical, and it may have had no effect in the recessions of 1980 20 Rudolph, Zumpano and Karson (1982) find that interregional interest rate differences still existed in the mid-1970s, while Karson, Rudolph and Zumpano (1986) conclude that they did not exist by the mid-1980s. King and Andrukonis (1984) report that FHLMC securities generated a gross transfer of over $5 billion during their first decade. Information on net interregional flows of mortgage funds is not available, but the existence of substantial gross flows suggests that securi tization played a significant role in eliminating regional rate differentials. 21 The difference in the trend after 1960 may be partly attribut able to demographic changes, especially the increasing pro portion of households in categories in which homeownership is less common, such as single individuals and single parents. 57 and 1981-82.22 FNMA’s purchases of conven tional mortgages dropped by about one-third from 1979 to 1981, which does not suggest that it tried to act countercyclically. Thus, even during this period of institutional change and upheaval, the system continued to allocate credit to housing, albeit at somewhat higher mortgage rates, and a fully national mort gage market developed. But cyclical fluctuations in housing were severe and the homeownership rate stopped rising, raising the question of whether the system was still achieving its basic purpose. POLICY ISSUES: SAFETY AND SOUNDNESS VS. PUBLIC PURPOSE The housing finance system continues to evolve. Congress has enacted three major laws in the last five years, affecting every institution in the system, and may consider further legisla tion for the Federal Home Loan Bank System.23 Some important provisions of these laws have not yet been implemented; when they are, they may in turn provoke further changes. The laws have addressed two kinds of issues: policy mat ters—what purposes the system w ill serve and how it will achieve them; and regulatory mat ters—what powers different institutions will have and how they will be regulated. In the policy area, all three new laws represent a balancing of public purpose against “safety and soundness,” the implicit objective that the system not impose direct costs on taxpayers that must be met by legislated appropriations. In the wake of the S&Ls’ problems, there has been a much stronger emphasis on safety and soundness; new capital requirements have been imposed on most insti tutions within the system. But there are elements in each law that concern the public purposes, and there is some evidence that the pendulum may be swinging back toward a renewed emphasis on 22 See Grebler (1977) and Jaffee and Rosen (1979) for the earlier cycles, and Kaufman (1985) for the 1980-82 reces sions. Grebler analyzes both Home Loan Bank advances and FNMA purchases. 23 The three laws are: FIRREA, concerning the S&Ls, the Federal Home Loan Bank Board (which it abolished), FHLMC, and to a lesser extent the Federal Home Loan Banks; Title 5 of the National Affordable Housing Act of 1990, concerning FHA; and the Federal Housing Enterprises Financial Safety and Soundness Act of 1992, enacted as Title 13 of the Housing and Community Development Act of 1992, concerning FNMA and FHLMC. In addition, Congress in 1992 required five separate studies of the Federal Home Loan Bank System as the precursor to future legislation. Four of these studies appeared during 1993, and the fifth in April 1994. (The GSE legislation in 1992 was enacted after these purposes. At the same time, some provi sions of the new laws may make it more difficult to achieve them. Safety and Soundness The new laws raise capital standards and take account of risk differentials among assets for virtually all institutions within the housing finance system. FIRREA imposed more stringent capital requirements on the S&Ls. They must have 1.5 percent tangible capital relative to assets and must meet the same risk-based capital standards as national banks. The tangible net worth of the S&Ls as a whole was only 0.7 percent in 1989. Both FHA and the GSEs are required to hold more capital than they had when the laws were passed. The existing standards were not raised so much as they were changed conceptually. FHA had no specific capital standard, beyond the legislative requirement that it be actuarially sound, which was undefined. FNMA’s capital requirement, established in its Charter Act, was calculated as a debt-to-capital ratio. This meant that it was only required to hold capital against its portfolio. Because it did not need to issue debt to finance its MBSs, FNMA’s capital-to-asset ratio (including MBSs) was 1.1 percent in 1990. Prior to FIRREA, FHLMC had no statutory capital requirement; the Bank Board determined it as a policy matter. FIRREA gave FHLMC the same debt-to-capital standard as FNMA. With its larger proportion of M BSs, its capital-to-asset ratio was 0.8 percent in 1990. These are quite low levels of capital; had the GSEs been required to meet the risk-based capital standards set for the S&Ls in FIRREA, FNMA would have needed 2.5 times as much capital as it actually had in 1990, and FHLMC more than three times as much.24 Both the FHA and GSE standards are established by “stress tests.” In other words, the entity must Congress required and received nine separate reports from various federal government agencies.) 24 The calculations of these various capital ratios are reported in U.S. Department of Housing and Urban Development (1990a, 1990b). The published debt-to-capital ratios in the reports include only stockholders’ equity; on that basis, the ratios are 0.8 percent for FNMA and 0.6 percent for FHLMC. Regulatory capital is defined in the Charter Acts to include retained earnings and subordinated debt. Subordinated debt is not counted as equity under Generally Accepted Accounting Principles because it takes precedence during bankruptcy over ownership interests. The subordinated debt of the GSEs is due and payable in the event of bankruptcy or insolvency, which appears to limit the government’s ability to rely on it as capital. JULY/AUGUST 1994 58 have enough capital to survive a recession, with the amount determined in advance by econo metric analysis. Both are risk-adjusted capital standards; riskier loans are more likely to default and more capital is required against them. The FHA standard was set on the basis of an actuarial study by Price Waterhouse during 1989-90. Price Waterhouse recommended that FHA should at a minimum have a net worth of 1.25 percent against insurance in force, instead of the then-current level of about 1.0 percent. The purpose was to ensure that FHA would have positive net worth in the event of a typical post-war recession. This standard was enacted in 1990, to be effective in 1992, with a higher standard of 2.0 percent in the year 2000. To reach these targets, the insurance premium was raised by about 70 percent, in present value terms, and risk-related premiums were estab lished for the first time. Minimum equity requirements for FHA homebuyers were also raised, to reduce defaults and strengthen the insurance fund; this increase, however, was partly rolled back in 1992.25 The GSE stress test is based on their worst actual regional experience, which for both has been Texas in the mid-1980s. They are supposed to have enough capital to withstand such a recession if it occurred on the national level, and also to survive large (600 basis points) upward or downward changes in interest rates occurring within a period of one year and lasting for 10 years. Both mortgages held in portfolio and M BSs are included. The GSE capital standard includes more than the stress test. It also defines two lower levels of capital, in the form of ratios against mortgages held in portfolio and MBSs. The authority of the regulator varies depending on the GSE’s actual capital relative to the levels defined in the statute. The “minimum” capital level is 2.5 percent against mortgages held in portfolio and 0.45 per cent against M BSs, which matched FNMA’s 25 The 1990 legislation required FHA mortgagors to have at least 2.25 percent equity in their home when they bought it (1.25 percent for mortgages under $50,000). Previously, it was possible to buy a home with no real equity, because buyers were allowed to finance the closing costs in their mortgage. On loans below $50,000, the minimum down payment is 3 percent; on loans over $50,000, it is 3 percent of the first $25,000 and 5 percent over $25,000. Closing costs average 2 to 3 percent, ranging up to 6 percent in a few states. The down payment in effect paid the closing costs for a substantial number of FHA-insured homebuyers. For analysis of the relationship between defaults and initial FEDERAL RESERVE BANK OF ST. LOUIS actual capital position as of 1991, and was slightly more than FHLMC’s capital. (FHLMC was given 18 months to meet the minimum level without incurring any regulatory sanctions, and it now does.) The “critical” capital level is half of the minimum level; if capital falls below it, the regulator can immediately put the GSE into receivership. The GSEs’ risk-based capital standard is almost certainly less stringent than the standard for thrifts and banks, and the minimum standard clearly is lower. Depository institutions must have 4 percent capital against a mortgage in their portfolio, while the GSEs must only have 2.5 percent. If a mortgage is securitized by the GSEs and the security is held by a thrift or bank, total capital is still less: 2.05 percent, consisting of 1.6 percent for the depository institution to protect against interest rate risk and 0.45 percent for the GSE to protect against credit risk. The FHA and GSE standards were established in very different ways. In the case of FHA, an econometric analysis was conducted and the results were known before legislation was passed. The law set new parameters for FHA insurance partly on the basis of whether they would enable FHA to achieve the standard. In the case of the GSEs, the stress test is prescribed as much as possible in the statute, but it was not performed before the bill was passed. Instead, it was nego tiated between the Bush Administration and the GSEs and written into law by Congress without analysis of how much capital w ill be required to meet it. The test must be formally stated in reg ulations by November 1994, and does not become effective for another year. Capital standards for the Federal Home Loan Banks may be the subject of legislation in the near future. Their capital now consists only of the stock that S&Ls and other institutions have had to buy in order to be members of the Home Loan Bank System and to obtain advances. Members can withdraw from the System and sell loan-to-value ratios, showing a strong positive correlation, see Price Waterhouse (1990) and Hendershott and Schultz (1993), and the literature cited therein. For a more detailed discussion of the 1990 FHA legislation, see Weicher (1992). 59 their stock back to their Home Loan Bank, subject to an advance notice requirement of six months. Thus, it is problematic whether the capital would be available if individual Home Loan Banks began to incur losses. The members have an incentive, and a right, to withdraw their capital just when it is most needed. But the Home Loan Banks have so little capital because of public policy. They had $2 billion in retained earnings until FIRREA took that money to cover part of the cost of S&L failures. FIRREA also required them to contribute $300 m illion per year out of their future earnings. That is why the only capital they now have is the stock owned by the members. It is going to be difficult to build capital through new retained earnings. Even if Congress were to repeal the $300 m illion annual contribution, the S&Ls and probably the newer members are likely to prefer having this money passed through as dividends rather than remaining as retained earnings, which could be taken away again. In their studies of the Home Loan Banks, both the General Accounting Office and the Clinton Administration have argued that they need more capital, but neither has offered a specific proposal for raising it. Countercyclical Support fo r Housing Construction The emphasis on safety and soundness creates a clear potential conflict with the traditional countercyclical objective of the housing finance system. This is most obvious with respect to the GSEs, where it was explicitly discussed in eval uating the adequacy of their present capital. In its annual reports as regulator of FNMA and FHLMC prior to legislative consideration of a new capital standard, HUD used a Depression scenario to assess capital adequacy, based on one used by Moody’s to rate private mortgage insurers. HUD concluded that neither GSE could survive 10 years of the Depression scenario, but 26 These results are for 1991, as reported in U.S. Department of Housing and Urban Development (1992a, 1992b). Similar tests for 1990 show that FHLMC could survive six years and FNMA seven (U.S. Department of Housing and Urban Development, 1991b). The 1991 test is more detailed and sophisticated. The test is stringent: It includes a 10 percent decline in house prices for four straight years, for example. It is necessary to survive 10 years of the Moody’s Depression scenario to qualify for an AAA rating; very few financial institutions are rated AAA. The HUD stress test is not identical to the Moody’s scenario, although it is closely modeled on the scenario. Thus, it cannot be said that the GSEs would or would not receive an AAA rating from Moody’s, or conversely that any other financial institution would or would not survive 10 years of the HUD stress test. both could last six years.26 This analysis assumed that they continued to be active in the mortgage market during the Depression to the same extent as previously; FNMA and FHLMC could survive a full 10 years of the Depression scenario, if they immediately suspended opera tions at the beginning of the Depression. In response, the GSEs stated that they would in fact cease buying mortgages immediately.27 This, of course, raises the question of whether they could recognize the onset of a depression immediately (a delay of two years would be enough to cut the period of survival from 10 to six years for both GSEs). The 1992 legislation accepted the GSE position on at least an interim basis. The stress test must assume that the GSEs accept no new business until the General Accounting Office and Congressional Budget Office complete studies of the appropriate new business assumptions. The studies are due in November 1995. The GSEs have also said they would increase their guarantee fees if necessary to remain solvent, but it is not easy to raise prices during a recession. It should be possible for a GSE to buy mortgages in periods of high interest rates, earn a profit if rates decline, and perhaps moderate fluctuations in housing production in the process. The opportunity to profit from interest rate declines is limited, but not eliminated, by the prepayment option. On the other hand, countercyclical behavior may result in credit risk. If the quality distribution of loans offered in a recession is the same as during an expansion, then it is necessary to buy lower quality mortgages in order to be actively countercyclical. The potential conflict has been discussed in every downturn. The legislative changes probably heighten it. The GSEs still have Charter Act responsibilities to “provide ongoing support to the mortgage market,” but a capital standard that assumes they do not. An AAA rating is not likely, however. In 1991 Standard and Poor’s evaluated both GSEs at the request of the Treasury, rating FHLMC as A+ and FNMA as A- on the assumption that the GSEs did not have agency status, which in fact they do. 27 The reaction of the GSEs appears in U.S. Department of Housing and Urban Development (1991b). JULY/AUGUST 1994 60 At present these issues are hypothetical, but in a severe downturn they would become real. They would then attract policymakers’ attention and perhaps result in further changes. “U nderserved” A reas and Groups The traditional goal of making mortgage credit equally available across the country has taken new forms in FIRREA and the GSE legislation. The focus of concern has shifted from regions to communities, particularly low-income urban neighborhoods and those with predominantly minority residents, and also to individuals. Explicit subsidies through the housing finance system are being provided and proposed to meet this goal. FIRREA required the Home Loan Banks to subsidize low-income housing directly by allocating 20 percent of their profits to a new Affordable Housing Program, with a minimum annual amount starting at $50 m illion and increasing to $100 m illion by 1995. This is essentially another tax on the Home Loan Banks and through them the S&Ls. The Clinton Administration’s report on the Home Loan Bank System goes farther and proposes a specific mandate to facilitate mortgage lending to lowerincome families and targeted populations. This would be a new role for the Home Loan Banks. Support for low- and moderate-income housing has been one of the purposes of FNMA and FHLMC, as stated in their Charter Acts.28 This became the major Congressional concern in 1992, once the Administration and the GSEs agreed on a capital standard. The law sets two general goals that certain percentages (to be determined by the HUD Secretary) should be for units occupied by families with incomes below median, and for housing in central cities. Both goals include rental housing as well as homes.29 The law allows a number of acceptable reasons for not meeting any goal in a given year, and a multi-year regulatory process before any sanc tions could be imposed for falling short. 28 Both GSEs are required “to provide ongoing assistance to the secondary market for residential mortgages (including activities relating to mortgages on housing for low- and moderate-income families involving a reasonable economic return that may be less than the return earned on other activities)...” See 12 U.S.C. § 1716. 29 The law also sets a special affordable housing goal that 1 percent of each GSE’s purchases should be for housing affordable to low-income families or located in low-income neighborhoods, with allocated shares for single-family and multifamily housing. FEDERAL RESERVE BANK OF ST. LOUIS The law also sets transition goals. Each GSE is to have at least 30 percent of its purchases for housing occupied by families below median income by 1994, rising in two steps from their 1992 levels. This is about the share of the con ventional conforming loan market consisting of buyers in this income range. The GSEs have not come very close to this percentage in the past for single-family houses; both were below 25 percent in 1991. Nearly all apartments, however, are affordable by families of median income by the rules of thumb set forth in the legislation. Both GSEs met the 1993 targets established by HUD. Similar transition goals were established for central cities, and in this case neither GSE met the 1993 goal. Under HUD’s regulations they are now required to file housing plans to describe how they will meet them in the future. These housing goals pose another potential conflict between the safety and soundness of the housing finance system (and also the financial interests of the private and quasi-private housing finance institutions) and a public purpose which is becoming increasingly prominent. Both the GSEs and the Home Loan Banks have pointed out the conflict. The issue is more serious for the Home Loan Banks because they must fund the Affordable Housing Program. This require ment, like the $300 m illion contribution to the cost of S&L resolutions, has been used by the Home Loan Banks to justify making investments with the funds that they borrow on the capital markets as the demand for advances has fallen, and thus perhaps to undertake new types of risk. FNMA and FHLMC are required only to buy loans for moderate-income housing, not to pro vide subsidies, and they need not lower their underwriting standards. The general conclusion of the mortgage default literature is that default is largely a function of the loan-to-value ratio on the mortgage and not closely related to either the value of a home or the income of the buyer.30 But even if moderate-income housing turned out 30 See Hendershott and Schultz (1993) and the literature cited therein. Hendershott and Schultz do find that foreclosures on FHA-insured loans are inversely related to loan size, which they attribute to differential underwriting standards or house price appreciation rates. The latter explanation might indicate greater risk for small loans. 61 to be riskier, it would be possible for the GSEs to take more risk than the S&Ls, given their lower cost of funds. The charter acts and the legislation seem to expect that the GSEs will use their agency status to make loans to moderate-income buyers, without jeopardizing their safety and soundness. The Administration’s proposed lower-income mandate for the Home Loan Banks takes a similar view, stating that collateral requirements should not be relaxed to meet the mandate. REGULATORY ISSUES The Structure o f Regulation At the same time that capital standards were being raised for most entities within the system, all of the private institutions were given new regulators. The Bank Board was abolished and its duties parcelled out among several agencies. Some of the potential conflicts between public purpose and safety and soundness are reflected in the new regulatory structure. The 1992 GSE legislation divided authority between the Secretary of HUD and the Director of a new Office of Federal Housing Enterprise Oversight (OFHEO), which is formally part of HUD but is effectively independent of the Secretary. The Director reg ulates for safety and soundness; the Secretary establishes, monitors, and enforces the housing goals and regulates the GSEs in all areas other than safety and soundness. For example, the Secretary rather than the Director approves new programs, and the Secretary rather than the Director raised the issue of whether the GSEs had to reduce their maximum mortgage amount when the conforming loan limit declined last year. The effectiveness of this relationship has not yet been tested, since the Office is still in the formative stage and has not yet had to issue any of the regulations required by the law. Both the GAO and the Administration studies of the Home Loan Banks have recommended 31 White (1991) favors deregulation, but says it came 15 years too late and should have been accompanied by stronger safety and soundness regulation, and by risk-adjusted deposit insurance premiums. He reviews the literature showing that losses were positively related to use of the new powers. Barth (1991) favors deregulation but says that it contributed to the problems of some S&Ls. Kane (1989) argues that deregulation did not cause the problems of S&Ls and re-regulation would not solve them, but goes on to say that deregulation expanded opportunities for poorly man aged S&ls to fail, as well as allowing well-managed ones to rebuild their net worth. Rudolph (1989) analyzed the subse quent behavior of S&Ls that were insolvent in 1982 and found that traditional housing lenders were more likely to be that they be regulated by OFHEO. At present they are regulated by the Federal Housing Finance Board, created in FIRREA as an after thought for that sole purpose. The ostensible reason for abolishing the Bank Board was that it acted as an advocate for the industry it was sup posed to regulate. The risk that the regulated entities would capture the regulator would seem to be still greater for the Federal Housing Finance Board, with nothing to do but regulate the 12 Home Loan Banks, and for OFHEO, regu lating only the two GSEs. It is probably prefer able to have one regulator for all 14 institutions, rather than two, although FNMA and FHLMC do not favor sharing their regulator. Whether that proposal is adopted, it seems likely that the reg ulatory structure of the system will be revised again. Is T here A Future fo r Specialized Portfolio Lenders? Public policy has wrestled for two decades with the question of whether specialized mort gage portfolio lenders can exist. In the 1970s policymakers decided they could and tried to keep the S&Ls operating as they always had in the face of inflation and new competition. In the 1980s policy reversed itself, and S&Ls were given broad new asset powers. In the 1990s policy has reversed itself again. FIRREA adopted the premise that deregulation caused the S&L failures, a view not shared by most economists.31 It explicitly took away some of the powers granted in 1980-82 and required S&Ls to put a higher percentage of their assets in mortgages and housing investments to keep their tax advantages, although the latter restriction was somewhat relaxed in 1991.32 S&Ls are not forced to be spe cialized housing portfolio lenders, however; the new capital rules give them an incentive to move away from portfolio lending by requiring only 40 percent as much capital against MBSs issued by the GSEs as against whole mortgages. insolvent in 1982-83, but less likely in 1986, which suggests that among insolvent S&Ls, at least, those taking advantage of the new powers were less successful than those which “stayed in housing.” 32 White (1991) notes, however, that at the same time these provisions were enacted, many members of Congress were saying that S&Ls should be more like commercial banks. JULY/AUGUST 1994 62 Meanwhile, the GSEs are thriving in part because they are specialized lenders—very large S&Ls, in that sense— as well as secondary market agencies. Their portfolio operations are quite profitable: Almost three-quarters of FNMA’s revenue, and almost half of FHLMC’s, came from their portfolios in 1992. FHLMC has recog nized this since FIRREA gave it the same powers and incentives as FNMA; it is behaving similarly, expanding its portfolio by 70 percent and an nouncing a business objective of further rapid growth. FHLMC was already the fourth largest portfolio lender in the United States, although its portfolio always has been small relative to its securities volume. The specialized portfolio lender is alive and functioning to a greater extent than is generally recognized. Several factors contribute to the GSEs’ success as portfolio lenders. A number of recent studies conclude that the 20 to 35 basis point difference between the conforming market and the jumbo market is the margin between profit and loss for S&Ls, on average. The cost advantage of the GSEs is attributed to the lower capital require ments they face, the tax exemptions and smaller regulatory burdens granted them by their federal charters, the fact that they do not have to pay deposit insurance premiums or help to fund the resolution of failed S&Ls, and economies of scale.33 The first two factors are benefits conferred explicitly by act of Congress. Macroeconomic conditions may also con tribute to the GSEs’ recent success. If inflation is erratic and unanticipated, it is probably not possible to survive as a specialized portfolio lender, as FNMA and the S&Ls showed during the 1970s and early 1980s. If inflation is low and stable, it is apparently still possible to be a specialized portfolio lender, as the same institu tions showed when inflation receded after 1982 and both FNMA and a large number of S&Ls became profitable once again. FNMA especially was saved by disinflation. It gambled on a decline in interest rates (there was not much to lose from such a gamble, from the stockholders’ point of view), buying mortgages in large vol ume, and it benefitted greatly when rates fell. Its net worth became positive in 1985. It has tried to prevent a similar problem in the future. A HUD analysis in 1991 concluded that it has 33 See Cotterman (1994) and McNulty and Pearce (1994) for discussions of the literature. Goodman and Passmore (1992) calculate that the difference in capital requirements alone lowers the GSEs’ costs by about 35 basis points. FEDERAL RESERVE BANK OF ST. LOUIS effectively hedged against interest rate risk by changing the duration of its liabilities, even though most of its portfolio consists of fixed-rate mortgages. It is doubtful if S&Ls could do the same thing, given their higher costs. They may be able to prosper as portfolio lenders specializing to some extent in fixed-rate mortgages by accepting inter est rate risk, if inflation remains low and if the yield curve remains upward sloping. They may also survive as ARM lenders. The broader geographic authority of the GSEs may have been important in the 1980s. S&L fail ures since deregulation have been concentrated in states which suffered severe recessions, notably Texas but also other energy and farm states. The decline in domestic crude oil prices between 1982 and 1989 precipitated a regional recession in which commercial real estate val ues in the Southwest fell by one-third. (Over the same period, they rose nationally by about 10 percent.) Defaults on home mortgages and other real estate rose rapidly. The geographic restrictions on S&L portfolios meant that the bad investments and loans were held in the portfo lios of institutions in those states. One-fifth of the S&Ls which failed during the 1980s were located in Texas and they accounted for half of the losses. Deposit insurance turned the region al failures into a national problem. It is now possible for an S&L, like a GSE, to make or buy loans anywhere in the country. This may be significant. Both FNMA and FHA survived the Texas recession and other regional problems during the mid-1980s. Both sustained losses, but neither was driven to a negative net worth position. Their experience suggests that national portfolio lenders—in effect, national S&Ls—could be viable. But the cost advantages of the GSEs would remain. GSE Powers The competitive advantages conferred on the GSEs by agency status raise the question of their role as potential competitors in markets beyond their current activities. This has been a recurring issue for FNMA since it was privatized, when the Secretary of HUD was given new program approval authority. It has become an issue for FHLMC 63 since FIRREA. Typically, FNMA has sought legislative authority from Congress if the Secretary has denied or deferred approval. Usually, but not always, Congress has given approval. The most contentious instance is REMICs. The Tax Reform Act of 1986 created these securities and authorized FNMA and FHLMC to issue REMICs backed by conventional mortgages, subject to the approval of their regulators.34 Both agencies sought approval, over the strong objections of Wall Street securities issuers, who argued that they could enter this market if the agencies did not, but could not compete with them. The Federal Home Loan Bank Board originally decided not to allow FHLMC to issue REMICs, but HUD (under pressure from Congress) eventually allowed FNMA to issue them on a temporary basis, which became per manent in 1988, and FHLMC was allowed to issue them in 1988. FNMA has also sought to move beyond mortgage purchase and securitization, with less success. In 1985 it proposed to buy bonds collateralized by mortgages that would be issued by financing subsidiaries of financial institutions. In 1990 it requested approval of a program to buy debt obligations secured by conventional mortgages, or securities backed by such mortgages. The former proposal brought objections from invest ment bankers and the Senate Banking Committee chairman and ranking minority member, as being financing transactions rather than mortgage pur chases, and HUD did not approve it. The latter would have placed FNMA in competition with the Home Loan Banks by letting it issue advances against mortgage collateral, and HUD again denied approval.35 34 REMICs have tax advantages over CMOs and were expected to be a major innovation in mortgage securities. 35 HUD also denied approval on grounds that the program involved significant risks to all parties, including the federal government, and could adversely affect FNMA’s safety and soundness since it could affect FNMA’s needs for capital. The transactions would have allowed financial institutions to defer recognition of economic losses and encourage leveraging, possibly increasing the risk of bankruptcy by the institution. Denial of new programs on safety and soundness grounds has been relatively infrequent. FNMA has also undertaken or considered activities that are ancillary to its secondary market operations but that are already provided by private firms. In the m id-1980s, it raised the possibility of establishing a mortgage insurance subsidiary, in competition with the private mortgage insur ance industry. In 1985 it acquired a computer software firm with the intent of producing and selling a loan origination and servicing program.36 The GSEs have strongly opposed any regulatory limitation on their powers. This was a major issue in the 1992 legislation. Ultimately, the HUD Secretary retained the authority to deny approval for a new program if he or she determines that the program is “not in the public interest,” but the Secretary has to act within 45 days, and the law prescribes an appeals process which is heavily weighted toward the GSEs. The law also gives the GSEs broad authority to buy any kind of home mortgage, limiting the Secretary’s regulatory discretion.37 There is no reason to think that the GSEs will not attempt to extend their activities in the future, so the demarcation of authority between different institutions will probably continue to be a recurring issue. CONCLUSION One public policy issue which appears to be resolved is the desirability of a housing finance system to allocate credit to housing. This is a change from recent years. For about a decade beginning in the late 1970s, there was extensive public discussion about dismantling the system on the grounds that it was becoming too expen sive for society to continue allocating credit to housing, in two senses. Household investment in housing was growing rapidly in response to time, virtually the only conventional mortgages issued were standard mortgages. As new instruments were developed in the 1980s, FNMA claimed the authority to buy any kind of conventional first mortgage under the “Romney letter,” such as ARMs and balloon mortgages with different risk characteristics than the standard mortgage. Proposed HUD regulations issued in 1990 would have revoked this broad interpretation. FNMA objected to the regulations, and they were superseded by the 1992 legislation. 36 The regulatory issues raised by new activities between 1980 and 1985 are described in U.S. Department of Housing and Urban Development (1987), the proposal to purchase debt obligations secured by conventional mortgages in U.S. Department of Housing and Urban Development (1991a). 37 In 1970 FNMA sought and received regulatory approval from Secretary George Romney to begin a program of buying conventional first mortgages on single-family homes. At the JULY/AUGUST 1994 64 inflation in the late 1970s, with concomitant underinvestment in productive capacity, and the growing problems of the S&Ls were threatening to impose direct costs on taxpayers. Cotterman, Robert F. “The Effects of FHLMC’s and FNMA’s Mortgage Activities,” in U.S. Department of Housing and Urban Development, Report to Congress on the Federal Home Loan Bank System, vol. 2: Analytical Studies. U.S. Department of Housing and Urban Development, 1994. A number of proposals were offered to address these concerns. The Financial Reform Act con sidered by the House of Representatives in 1976 would have turned the S&Ls into commercial banks. The Reagan Administration offered budget proposals to levy fees on the GSEs, which could have been set at a level to cover the market value of agency status. There were several studies of privatizing one or both of them in the mid-1980s. Fredrikson, E. Bruce. ‘The Geographic Structure of Residential Mortgage Yields,” in Jack M. Guttentag, ed., Essays on Interest Rates, vol. II. Columbia University Press, 1971. But interest has faded since then. During con sideration of FHA reform, no one (including its competitors) made any suggestion to close FHA down or turn it over to the private sector. Most importantly, there was no real interest in priva tizing either GSE in the extended process of writing the 1992 legislation. The closest thing to it was an amendment offered by Rep. Jim Leach (R.-Iowa) on the floor of the House of Representatives, which would have set higher capital standards and eliminated some of the GSEs’ privileges. The amendment lost on a 298-119 vote. No comparable proposal was mentioned in the Senate. Five years after FIRREA and 15 years after public recognition that the New Deal housing finance system was no longer viable, the United States has a very different system. But the new system still has the same purposes as the old one, broadly speaking, and if anything the inherent conflicts b etw een the p u rp oses of th e system and the safety and soundness of the individual institutions in it are more sharply drawn. While the spectacular problems of the S&Ls have attracted most of the recent attention, their role has sharply diminished, and the important issues in the future are likely to involve the GSEs. REFERENCES Barth, James R. The Great Savings and Loan Debacle. American Enterprise Institute, 1991. Bodfish, Morton. History of Building and Loan in the United States. United States Building and Loan League, 1931. Brewer, Elijah, III. “Full-Blown Crisis, Half-Measure Cure,” Economic Perspectives (November/December 1989), pp. 2-17. Brumbaugh, R. Dan, Jr. Thrifts Under Siege: Restoring Order to American Banking. Ballinger, 1988. FEDERAL RESERVE BANK OF ST. LOUIS Goodman, John L., Jr., and S. Wayne Passmore. “Market Power and the Pricing of Mortgage Securitization.” Federal Reserve Board Finance and Economics Discussion Paper No. 197 (March 1992). Grebler, Leo. “An Assessment of the Performance of the Public Sector in the Residential Housing Market: 1955-1974,” in Robert M. Buckley, John A. Tuccillo, and Kevin E. Villani, eds., Capital Markets and the Housing Sector. Ballinger, 1977. _____ , David M. Blank, and Louis Winnick. Capital Formation in Residential Real Estate. Princeton University Press, 1956. Hendershott, Patric H., and William R. Schultz. “Equity and Nonequity Determinants of FHA Single-Family Mortgage Foreclosures in the 1980s.” Journal o f the American Real Estate and Urban Economics Association (winter 1993), pp. 405-30. _____ , and James D. Shilling. ‘The Impact of the Agencies on Conventional Fixed-Rate Mortgage Yields,” Journal o f Real Estate Finance and Economics (June, 1989), pp. 101-15. _____ , _____ , and Kevin E. Villani. “Measurement of the Spreads between Yields on Various Mortgage Contracts and Treasury Securities,” AREUEA Journal (winter 1983), pp. 476-89. _____ , and Robert Van Order. “Integration of Mortgage and Capital Markets and the Accumulation of Residential Capital,” Reqional Science & Urban Economics (May 1989), pp. 189-210. ICF, Incorporated. Effects of the Conforming Loan Limit on Mortgage Markets. U.S. Department of Housing and Urban Development, March 1990. Jaffee, Dwight M., and Kenneth T. Rosen. “Mortgage Credit Availability and Residential Construction," Brookings Papers on Economic Activity (1979, No. 2), pp. 333-76. Jones, Sidney L. The Development of Economic Policy: Financial Institution Reform. Graduate School of Business Administration, University of Michigan, 1979. Kane, Edward J. The Gathering Crisis in Federal Deposit Insurance. MIT Press, 1985. _____ . ‘The Looting of FSLIC: What Went Wrong?” Ohio State University College of Business Working Paper 89-3 (January 1989). _____ , and Chester Foster. “Valuing and Eliminating Subsidies Associated With Conjectural Government Guarantees of FNMA Liabilities,” College of Administrative Sciences Working Paper 86-71, Ohio State University (1986). Karson, Marvin J., Patricia M. Rudolph, and Leonard V. Zumpano. “Inter-Regional Differences in Conventional Mortgage Terms: A Test of the Efficiency of the Residential Mortgage Market,” paper presented at the American Real Estate and Urban Economics Association, 1986. 65 Kaserman, David L. “An Econometric Analysis of the Decline in Federal Mortgage Default Insurance,” in Robert M. Buckley, John A. Tuccillo, and Kevin E. Villani, eds., Capital Markets and the Housing Sector. Ballinger, 1977. Kaufman, George. “Impact of Deregulation on the Mortgage Market,” in Housing Finance in the Eighties: Issues and Options. Federal National Mortgage Association, 1981. Kaufman, Herbert. “FNMA and the Housing Cycle: Its Recent Contribution and Its Future Role in a Deregulated Environment,” in U.S. General Accounting Office, The Federal National Mortgage Association in a Changing Economic Environment. U.S. General Accounting Office, 1985. King, A. Thomas, and David Andrukonis. “Who Holds PCs?” Secondary Mortgage Markets (February 1984), pp. 12-17. McNulty, James E., and James Pearce. “An Economic Evaluation of Specialized Housing Lenders and the Qualified Thrift Lender Test: A Review of the Literature,” in U.S. Department of Housing and Urban Development, Report to Congress on the Federal Home Loan Bank System, vol. 2: Analytical Studies. U.S. Department of Housing and Urban Development, 1994. Morton, J.E. Urban Mortgage Lending: Comparative Markets and Experience. Princeton University Press, 1956. Price Waterhouse. An Actuarial Review o f the Federal Housing Administration’s Mutual Mortgage Insurance Fund. Price Waterhouse, 1990. Rudolph, Patricia M. “The Insolvent Thrifts of 1982: Where Are They Now?” AREUEA Journal (winter 1989), pp. 450-62. _____ , Leonard V. Zumpano, and Marvin J. Karson. “Mortgage Markets and Inter-Regional Differences in Conventional Mortgage Terms,” AREUEA Journal (spring 1982), pp. 94110 . Saulnier, R.J., Harold G. Halcrow and Neil H. Jacoby. Federal Lending and Loan Insurance. Princeton University Press, 1958. Tuccillo, John A., Robert Van Order, and Kevin E. Villani. “Homeownership Policies and Mortgage Markets, 1960 to 1980,” Housing Finance Review (January 1982), pp. 1-21. U.S. Department of Housing and Urban Development. 1991 Report to Congress on the Federal Home Loan Mortgage Corporation. HUD, 1992a. _____ . 1991 Report to Congress on the Federal National Mortgage Association. HUD, 1992b. _____ . 1990 Report to Congress on the Federal National Mortgage Association. HUD, 1991a. _____ . Capitalization Study o f the Federal National Mortgage Association and the Federal Home Loan Mortgage Corporation. HUD, 1991b. _____ . 1989 Report to Congress on the Federal Home Loan Mortgage Corporation. HUD, 1990a. _____ . 1988-89 Report to Congress on the Federal National Mortgage Association. HUD, 1990b. _____ . 1986 Report to Congress on the Federal National Mortgage Association. HUD, 1987. U.S. General Accounting Office. Federal Home Loan Bank System: Reforms Needed to Promote Its Safety, Soundness, and Effectiveness. U.S. General Accounting Office, GAO/GGD-94-38, December 1993. Weicher, John C. “The Future Structure of the Housing Finance System,” in William S. Haraf and Rose Marie Kushmeider, eds., Restructuring Banking and Financial Services in America. American Enterprise Institute, 1988. _____ . “FHA Reform: Balancing Public Purpose and Financial Soundness,” Journal of Real Estate Finance and Economics (March 1992), pp. 133-50. White, Lawrence J. The S & L Debacle. Oxford University Press, 1991. JULY/AUGUST 1994 Alvin L. Marty Alvin L. Marty is professor of economics and finance at the Center for Business and Government, Baruch College, City University o f New York. The author is indebted to Philip Cagan, Barry Ma and John Tatom for helpful comments and suggestions. Li Li provided research assistance. The paper was written while the author was a visiting scholar at the Federal Reserve Bank of St. Louis. The Inflation Tax and the Marginal Welfare Cost in a World of Currency and Deposits H ow HIGH IS THE OPTIMAL rate of inflation? The answer depends on the range of benefits and costs associated with inflation that are considered by the monetary authority in choosing the inflation rate. For example, if one considers the effects of inflation on the distribu tions of income and wealth, its interactions with the tax code or the transition cost of changing the expected rate of inflation, or if one adopts the alternative perspectives of different economic agents, the benefits and costs can be relatively large and difficult to assess. This article abstracts from transitory and largely avoidable aspects of inflation, and focuses instead on the fundamental public finance aspects of the monetary authority’s problem. In this case, the net benefits and costs are those associated with an inflation rate that is perfectly anticipated; the benefit of inflation that accrues to the monetary authority (typically the government) is the revenue from inflationary money creation. This benefit is analogous to the revenue arising from a specific tax on any other good or service. Inflation imposes a tax on money holdings because it is the rate at which individuals lose the purchasing power of a dollar. To lower the total cost of holding money, individuals change their holdings and their use of money when inflation rises. Their efforts to do so, however, reduce their total services from real money balances, thereby lowering individuals’ real income. This loss is the welfare cost of inflation. The optimal rate of inflation is found by com paring the marginal welfare cost of revenue from inflation with the marginal cost of alternative sources of revenue. An efficient system of tax collection minimizes the welfare cost of a given flow of tax revenue; this requires that the inflation rate must be chosen so that the marginal cost per dollar of revenue from inflation is the same as the marginal cost of alternative sources of revenue. In the analysis below, these concepts are devel oped for models involving a money stock made up of currency only, competitively priced bank deposits only, and a mix of both. The differences in each case clarify the analysis as well as provide some insight into the implications of the analysis for the optimal inflation rate. THE MARGINAL WELFARE COST OF REVENUE FROM MONEY CREATION: THE CURRENCY CASE Almost two decades ago, it was shown that a simple formula provides a method of calculating the additional welfare cost of collecting a dollar of revenue from money creation. This measure is the ratio of the marginal welfare cost of inflation JULY/AUGUST 1994 68 to the marginal revenue from a change in antici pated inflation.1 To derive this formula, assume the only money is currency and that the demand for real money balances depends only on the nominal rate of interest, holding other influ ences constant: (1) m = (p[i) The welfare loss, W, is (2) W = j(p (x )d x - i(p{i) and the marginal welfare loss from a rise in inflation, n, is reflected in the incremental loss from a rise in the nominal interest rate: (3) dW di Using the Phelps-Auernheimer (Phelps, 1973; Auernheimer, 1974) definition of the revenue, R, we have (4) R = (p{i)i, and the marginal revenue is r/R (5) — = (pU) + i(p'[i)- di Since the elasticity of demand for real balances is (6) JV; = - (p[i) ’ the marginal welfare cost per unit of revenue, the ratio of equations 3 and 5, is (7) dW dR 1 - N, Equation 7 is a variant of the well-known Ramsey tax rule (tax more heavily goods in inelastic demand) and assumes, as does the Ramsey rule, cross effects absent within the taxed sector. The formula is useful in answering the question: What rate of inflation (money rate of interest) would equalize the marginal welfare cost per dollar revenue accruing to inflation tax with an index of such costs due to other distortionary taxes? The analysis we are conducting is in the realm of balanced budget incidence. We raise the inflation tax on real balances until the marginal 1 The real rate of interest is held constant as the money rate of interest, /, varies in such an analysis (Marty, 1976). FEDERAL RESERVE BANK OF ST. LOUIS welfare cost per dollar of revenue is equal to an index of these per dollar distortions for other taxes. The increase in revenue is used by the government for exhaustive expenditures rather than rebated to consumers directly, or indirectly through reduced taxes. Another observation is in order. Although I have illustrated the use of these formulas by plugging in estimates of the marginal welfare cost, the main contribution of the paper lies in the provision of the formulas themselves. If these formulas pass muster, other empirical observations can be plugged in. Assume the demand for real cash balances fol lows the Cagan semi-log form M/P = A exp(-bi). Friedman (1971) uses three alternative values of b, 5, 10 or 20. Laidler (1986) cites .15 as the typical interest elasticity of demand for M l. If we assume the real rate is 1.5 percent, which equals the money rate at zero inflation, the value of b is .15/.015 or 10 percent. To err on the side of charity to inflationary finance, we use a value of 5 for b. Tower (1971) cites 10 percent as the upper limit to the index of the marginal welfare cost per dollar revenue for other distorting taxes. This estimate is considerably lower than those of Ballard, Shoven and Whally (1985), which range from 17 to 56. We assume 10 percent as the marginal welfare cost per dollar of revenue for other distortionary taxes. Using the Cagan function, we have in a currency-only world (8 ) dW dR ib 1 -ib ’ where dW/dR = 5i/(l-5i) = .1 and i* = .018. With the real rate = 0.015, the “optimal” inflation rate is approximately zero. Given the parameter values we have assumed, a very modest tax on real bal ances equalizes the marginal welfare cost per dollar of revenue to an index of distortions due to other taxes. THE MARGINAL WELFARE COST OF REVENUE: COMPETITIVELY PRICED DEPOSITS ONLY A variant of the above formula holds for a large number of competitive banks subject to a sterile legal reserve requirement, /, (Marty and Chaloupka, 1988). An individual bank would be forced by competition to pay (1 -f)i on its 69 deposits where i is the yield on its assets. The opportunity cost of holding deposits is then fi. In a deposit-only world, the welfare cost becomes (9) W = j(p [x )d x - if(p[if). The marginal change to welfare due to inflation is ai In this case, revenue is With the real rate equal to 1.5 percent, the “optimal” rate of inflation is 12.3 percent. THE MARGINAL WELFARE COST OF REVENUE FROM INFLATION: CUR RENCY AND COMPETITIVELY PRICED DEPOSITS We now show that the above analysis can be extended to a world of both currency and deposits. The demand function for each component is still referred to as <p, but they are potentially different and the different measure of cost, i or if, is used to indicate this. The counterpart measures are (11) R = fi(p{if). ' i \ ( if j(p (x )d x -i(p (i) + j(p [x )d x -if(p U f ) (14)W : The marginal increment to revenue as the interest rate changes, dR/di, is the bracketed term in equation 10. Since the elasticity of demand for deposits is (15 ) ^ (12) Nlf = - f i v V f ) (16) R = i(p[i) + if(p[if), the marginal welfare cost per dollar increment to revenue is (17) - j r = i<p'[i) + (p[i) + if(p '[if)f + (p [if) f Vo / V0 dW = - i (p ' [ i) - if( p ' [ if) f, di (pUf) flJ D (13) dW / di d R /d i Ny 1 -A L di and dW dR (18) (p[i) + icp'[i) + iffcp'[f) + f(p [if) ' Since The authorities in a bank-only world can set a money rate of interest equal to that in the currency-only world divided by the reciprocal of the reserve ratio, /. If the optimal money rate in the currency world were 10 percent, that rate can be set at 40 percent in a world of deposits (assuming the reserve ratio is 25 percent). Both the welfare loss and the tax revenue, however, are the same as in a currency-only world. Although the tax rate (the money rate of interest) is higher by the reciprocal of the reserve ratio, the tax base is reduced by the share of high powered money in the total money supply f[M/P). Assume initially that the demand for deposits has the same functional form as that for currency, that the marginal welfare costs per dollar of rev enue for other distortionary taxes is the same as in the world of currency, and that the reserve ratio is 13 percent (realistic for the United States). Since dW/dR = ifb/{l-ifb), we have .1 = i(.13)5/ [1 - j (.13)(5)] then i* = 13.8 percent, which is equal to the money rate in a world of currency, 1.8 per cent, divided by the reserve ratio, 13 percent. (19) Nif = (pdf) we obtain C Nr D — H +— fN if sat (20) dW dR — [ l - N ,) + — f [ l - N if M 1 MJ ,f where C is currency, <p(i), and D is deposits, ip(/i'). Once again, set the index of the marginal wel fare costs per dollar increment to revenue for other distorting taxes equal to 0.1. Let the reserve ratio be 13 percent and ratio of currency to the money supply be 30 percent (the ratio of bank deposits to money is then 70 percent). These figures correspond broadly to ratios in place in the United States for the early 1990s. Again set the semi-log slope of the Cagan func tion equal to 5. Then we have dW/dR = [(.3) (5i) + .7 (.65i) (-13)]/{(.3 - 1.5i) + .091 [1 - (,13)(5i)]} = 0.1. Then i* = 2.28 percent. It should be noted JULY/AUGUST 1994 70 that, although the formula is a weighted average of currency and deposits, the currency weight dominates the solution. W hile demand deposits are 70 percent of the money supply, the tax base is only the ratio of reserves to the money supply —that is, 9 percent. Given this low reserve ratio, currency commands dominate weight. The formula makes intuitive sense. If the rev enue ratio equals 100 percent, so that demand deposits pay no interest and, assuming for sim plicity, that the demand function (cp) for deposits is the same as that for currency, the formula reduces to (21) --- Nj + --- N: M ' M ' dW dR N, l-iV ; In effect, currency and deposits are of the same stuff. On the other hand, if the reserve ratio is zero, deposits produce neither seignorage nor a wel fare loss; we are, in effect, in a currency-only world because the monetary authority receives no revenue from deposits. In this case, the formula reduces to (2 2 ) dW dR 1 - N, This again makes intuitive sense since only currency is taxable. Variants of the above formulas can be derived. Consider, for example, a world in which an effective prohibition on the payment of interest on deposits is in effect. Then i i (23) W = | (p(x)dx - i(p{i)+ J (p(x)dx - i(p(i) 0 0 and (24) R = i(p{i) + if(p[i). It follows that (25) dW C D ---N; + --- N; M ' M ' dR — ( l- iV ,) + — /(1 -JV ,.) M ' M This is similar to the formula in equation 20, where deposits pay interest, but without the FEDERAL RESERVE BANK OF ST. LOUIS reserve ratio, /, in the second term of the numer ator. If the reserve ratio equals zero and there is no interest paid on deposits, bank deposits yield no government revenue, but a welfare loss accrues to both currency and deposits. If the reserve ratio equals 100 percent, the interest prohibition on deposits is unnecessary, but both currency and deposits incur a welfare loss and both provide seignorage. Once again, the formulas make intuitive sense. CONCLUSIONS AND COMMENTARY The above analysis has imposed the zero-profit condition that the return on interest-bearing assets is paid out in interest on deposits. This condition ignores the bank’s intermediation function, which has a necessary supply price. If the marginal costs of intermediation are constant, the interest paid on deposits is reduced by a given proportion. Since the tax base (reserves) is independent of intermediation costs, but deposits pay less interest, it follows that we have underestimated somewhat the marginal welfare costs and have erred on the side of overestimating the optimal rate of inflation. Although for purposes of exposition, the analy sis has in the main assumed that the demand schedule for deposits is the same as that for currency, all the formulas hold if the demand schedule for deposits differs from that for cur rency. All one needs to do is change the form of the function and plug in the relevant interest elasticities. The formulas are general and can be applied to economies with different indexes of marginal distortions and varying interest elasticities. A potential problem in using these formulas to predict dW/dR is that the ratio of currency to deposits may change with the rate of inflation. As an empirical matter, the currency-deposit ratio has remained remarkably stable in the United States since the period of financial de regulation in the late ’80s, when deposits began paying explicit interest. Moreover, a theoretical argument that the currency-deposit ratio is inde pendent of the money rate of interest has been made by Dwyer and Saving (1986). As we have seen the opportunity cost of holding currency is the rate of interest, i, and the opportunity cost of holding deposits is a fraction of the interest rate, if. Assuming the indifference curve between currency and deposits are homothetic, and that the ratio of these opportunity costs is the appro priate measure (by analogy with price theory) 71 determining the currency deposit ratio, this ratio is independent of the money rate of interest.2 Although these formulas have been used to assess dW/dR at hypothetical inflation rates, which requires predicting the currency deposit ratio, the formulas also can be used to calculate the ex post measure dW/dR at a prevailing money rate. All that is required is to observe the pre vailing currency-deposit ratio. Finally, some caveats are in order. The analysis deals with alternative positions of steady-state inflation. It does not handle the welfare costs of variable inflation—costs which may well be more significant than those associated with steady-state inflation. Moreover, our analysis has treated real balances as part of an optimal tax menu; this usual assumption is not without its critics (Lucas, 1986). REFERENCES Auemheimer, Leonardo. “The Honest Government’s Guide to the Revenue from the Creation of Money,” Journal of Political Economy (May/June 1974), pp. 598-606. Ballard, Charles L., John B. Shoven, and John Whalley. “General Equilibrium Computations of the Marginal Welfare 2 This reasoning, however, is not fully compelling. The ratio of the price of sowbellies to that of caviar has the dimensionali ty of sowbellies to caviar and is independent of proportionate changes which leave the relative price ratio unchanged. The ratio of the opportunity cost of currency to that of deposits is a dimensionless number and taking the ratio of the opportunity costs (by analogy with commodities) implies that one’s choice of currency and deposits is independent of the difference in their opportunity costs. Tatom (1979) makes an early attempt to determine the marginal welfare costs per dollar of revenue in a world of both currency and deposits. He takes the ratio of the oppor tunity costs combined with homothetic indifference curves as Costs of Taxes in the United States,” The American Economic Review (March 1985), pp. 128-38. Dwyer, Gerald P. Jr., and Thomas R. Saving. “Government Revenue from Money Creation with Government and Private Money,” Journal o f Monetary Economics (March 1986), pp. 239-49. Friedman, Milton. “Government Revenue from Inflation,” Journal o f Political Economy (July/August 1971), pp. 846-56. Lucas, Robert E., Jr. “Principles of Fiscal and Monetary Policy,” Journal o f Monetary Economics (January 1986), pp. 117-34. Marty, Alvin L. “A Note on the Welfare Cost of Money Creation,” Journal o f Monetary Economics (January 1976), pp. 121-4. _____ , and Frank J. Chaloupka. “Optimal Inflation Rates: A Generalization," Journal o f Money, Credit and Banking (February 1988), pp. 141-4. Phelps, Edmund S. “Inflation in the Theory of Public Finance,” Swedish Journal o f Economics (March 1973), pp. 67-82. Ramsey, Frank. P. “A Contribution to the Theory of Taxation,” Economic Journal (March 1927), pp. 47-61. Tatom, John A. “The Marginal Welfare Cost of the Revenue from Money Creation and the ‘Optimal’ Rate of Inflation," The Manchester School o f Economic and Social Studies (December 1979), pp. 359-68. Tower, Edward. “More on the Welfare Cost of Inflationary Finance,” Journal o f Money, Credit and Banking (November 1971), pp. 850-60. a compelling reason to treat the currency-deposits ratio as independent of the inflation rate. More importantly, Tatom does not build up his welfare costs from an explicit consider ation of the integral for currency and deposits separately, but conflates the two using a single integral running from zero to the money rate of interest. In fact, the integral for competi tively priced deposits should run from zero to if. Interestingly enough, Tatom’s analysis, although not general, applies to a world in which an effective prohibition on the payment of interest on deposits exists, and in which the form of the demand schedule is the same for currency and deposits. This is a special case of my analysis, and I am indebted to John Tatom for this reference and discussion. JULY/AUGUST 1994