The full text on this page is automatically extracted from the file linked above and may contain errors and inconsistencies.
August IS, 1988 • Conclusion Less regulation, lower interest rates, and improved economic conditions during the last five years helped most thrifts improve earnings as compared to those in the early 1980s. Using Ohio data, we found earnings at the best thrifts improved dramatically, while losses at the worst thrifts were generally reduced, until 1987 when losses accelerated. Losses stemmed from a range of factors, including poor asset quality, mispricing, and operational inefficiencies. The worst thrifts held a larger share of higherrisk loans such as commercial mortgages, but they earned lower yields and fees on loans, particularly on mortgage loans. The worst thrifts were much larger than the best thrifts and they made a larger share of non mortgage loans and were involved more with business and credit-card lending. Although higher-yielding business and consumer installment loans augmented overall loan yields, income was Significantly lower at the worst thrifts. While increasing differences in both revenues and expenses caused the thrift earnings gap to widen, cost differences played a key role. The worst thrifts were hurt by higher funding costs and operating expenses, but the heaviest blow came from sharply increasing nonoperating expenses attribtited to losses from selling assets States Gener.alAccounting Office, Briefing Report to the Chairman, Committee on Banking, Housing, and Urban Affairs,United States Senate, May 1987. and writing off loans. 3. The number of FSLICthrifts oper.ating in Ohio varied from 218 to 247 between 1983 and 1987, but only 192 of those instiruuons oper.ated in each of the five years. With larger losses in 1987, the worst thrifts produced a negative net worth. Their chances of recovery don't look good, especially since they were unable to capitalize on favorable mortgage-market conditions during the last few years-particularly in 1985 and 1986 when some thrifts did exceptionally well. Paul R. Watro is an economist at the Federal Reserve Bank of Cleveland. The author would like to thank Mark S. Sniderman and fames B. Thomson for their comments, and would like to thank Daniel j. Martin for research assistance. The views stated herein are those of the author and not necessarily those of the Federal Reserve Bank of Cleveland or of the Board of Governors of the Federal Reserve System. • Footnotes 1. An insolvent institution is one that has negative net worth based on gener.ally accepted accounting principles (GAAP). According to the Gener.al Accounting Office, the FSLICis technically insolvent with $13 billion more in liabilities than in assets. eCONOMIC COMMeNTaRY Federal Reserve Bank of Cleveland 4. Even in the early 1980s, Ohio's wor.st thrifts gener.ally did not attempt to grow out of the interest-rate problem as did some thrifts around the country, particularly in California and Texas. 5. Asset and loan yields are influenced by risk. Riskier loans should yield higher revenue because they impose additional expenses on lenders related to loan delinquencies and defaults. Is the Thrift Performance Gap Widening? Evidence from Ohio 6. Advances are much easier to acquire than are loans from Feder.alReserve Banks because discount-window borrowing is restricted to liquidity needs such as replacing deposit outflows. In contrast, advances can be used to make new loans, to purchase other assets, as well as to replace deposit outflows. Advances also vary in maturities up to sever.alyears. Moreover, the Feder.al Home Loan Bank Board has allowed the FSLICto act as guar.antor for advances when troubled thrifts did not have adequate collater.al. by Paul R Watro The savings and loan (thrift) industry reported a record loss of $6.8 billion in 1987, followed by a $3.8 billion loss for the first quarter in 1988. These losses are misleading, however, because they do not reflect the good performance of many thrifts. 7. Other borrowing comes largely from reverse repurchase agreements and mortgage-backed bonds. For example, over two-thirds of the nation's more than 3,100 thrifts were profitable last year with total earnings of $6.6 billion. Unprofitable thrifts, however, lost $13.4 billion in 1987, dragging down the industry's overall performance. 8. See William McGuire and Angelo DiMarzio, "The Profitability Impact of SFAS91,"Fifth District Review, Feder.al Home Loan Bank of Cincinnati, March 1988, pp. 18-21. 2. See, "Thrift Industry: Forbear.ance for Troubled Institutions 1982-1986," United Federal Reserve Bank of Cleveland Research Department p.o. Box 6387 A major reason for the large losses is that regulators continue to allow many insolvent thrifts to remain open because the federal insurance agencythe Federal Savings and Loan Insurance Corporation (FSLIC)-does not have adequate resources to close them.' As a result, over 500 insolvent thrifts continued operations in 1987, hurting the industry in general. BULK RATE U.S. Postage Paid Cleveland, OH Permit No. 385 Cleveland, OH 44101 Record losses in the unprofitable segment of the industry have sparked growing concern about what is being called the "thrift crisis." The crisis, which began in the early 1980s, is not new but the underlying problem causing it has changed. Today's crisis stems from poor-quality assets, whereas in the early 1980s it was due to high interest rates.' Material may be reprinted provided that the source is credited. Please send copies of reprinted materials to the editor. Address Correction Requested: Please send corrected mailing label to the Federal Reserve Bank of Cleveland, Research Department, P.O Box 6387, Cleveland, OH 44101 Since the end of the recession in 1982, interest rates have fallen sub- ISSN 0428- I 276 stantially, the economy has recovered, and the thrift industry has been deregulated. These developments gave thrifts promising opportunities to counteract interest-rate problems and to improve their precarious financial condition. Thrifts have reacted differently to the regulatory and economic changes. Some placed greater emphasis on growth and were less cautious about risk-taking and expenses. Others were more risk-averse and focused on improving profit margins through cost containment and traditional mortgage lending. • Performance Gap To illustrate contrasting thrift performance, we examined the best and worst thrifts that operated continuously in Ohio between 1983 and 1987.3 Performance was measured by the average return on assets. Based on five years of earnings, we classifledthe thrifts that ranked in the top 20 percent as the best performers and those in the bottom 20 percent as the worst performers. Each group in our study contained 38 thrifts. The worst performers had much larger assets and experienced less growth than the best performers. Average assets at the worst thrifts were $360 million, nearly twice as large as assets at the best thrifts. Over the last five years, asset growth at the worst thrifts averaged 4 percent annually, compared to 15 percent for the best thrtfts.s In examining thrift performance, the author discusses the differences in the earnings of the best and worst thrifts in Ohio and shows that the healthy thrifts are getting better and the sick thrifts are getting worse. Chart 1 shows that the earnings gap between the best and worst thrifts has widened substantially since the early 1980s. The expanding disparity is due not only to the dramatic improvement of the top performers, but also to the continued losses incurred by the bottom segment of the industry. In fact, the poor performers suffered their largest losses in 1987 when they recorded average losses amounting to 1 percent of assets. At year-end 1987, nearly one-half of the 38 poor performers were insolvent and their average net worth as a percentage of assets was -0.4 percent, down from 1.7 percent at year-end 1983. By separating earnings into two components-interest and noninterest -one can see in charts 2 and 3 that the increasing earnings disparity is largely attributed to differences in nonlnterest margins rather than to interest margin differences. The breakdown also allows us to link balance sheet differences to revenue and expense differences between the best and worst thrifts. • Net Interest Margin The net interest margin is the difference between interest income and interest expense as a percentage of average assets. This measure indicates how well interest- earning liabilities are being used to generate interest income. Chart 2 shows that net interest margins for thrifts in general have been increasing but that improvement has been greater at the best thrifts. Interest rates have strong effects on thrift profitability. Although thrifts have reduced the holdings of longterm fixed mortgages by substituting adjustable- rate mortgages, their portfolios still consist largely of longerterm assets that are being funded by shorter- term liabilities, such as moneymarket deposit accounts and traditional savings accounts. Consequently, when interest rates fall, thrifts generally improve earnings because the cost of funds usually declines by more than interest income. For example, net interest margins improved sharply between 1983 and 1986, when interest rates fell steadily from mid- 1984 to 1986. Low and falling interest rates also benefit thrifts by increasing the demand for houses and mortgages because lower borrowing costs enhance home affordability. During the 1983·86 period, wider rate spreads and greater mortgage lending contributed to sharply higher earnings at the best thrifts and to a reduction in losses at the worst thrifts. On the other hand, thrifts are affected adversely by high and rising interest rates. After significant improvement, interest margins increased only slightly during 1987 because interest rates rose during part of the year. Interest income, constituting the bulk of thrift income, comes from loans and investments. As a percentage of assets, interest income in 1987 was 81 basis points lower at the worst thrifts than at the best thrifts. This difference has increased more than four times over the last three years. The worst thrifts avoided even lower interest income by expanding into higher-yielding consumer installment loans and business lending. By yearend 1987, nonmortgage loans, primarily in the form of consumer loans, accounted for 8 percent of outstanding loans at the worst thrifts, cornpared to less than 4 percent for the best thrifts. Not only did the worst thrifts extend more consumer and business loans, but they also earned higher yields on such loans, probably because they were more involved with credit-card and business lending. Average deposit rates paid by the worst thrifts were between 1 basis point and 16 basis points higher than those paid by the best thrifts over the last five years. Since the worst thrifts held a Significantly larger share of lower- paying transaction and traditional savings accounts, one can deduce that they either paid higher rates on sirnilar time deposits or held more higherpaying, longer-term time deposits. The worst thrifts also held a smaller share of deposits over $100,000. All deposit balances over $100,000 are not federally insured and typically require high premiums from institutions having financial difficulties. The worst thrifts also took greater risk by making more commercial mortgage loans. Nonresidential mortgages accounted for 12 percent to 17 percent of the loans held by the worst thrifts, compared to 6 percent to 11 percent of the loans held by the best thrifts over the last five years. Despite making a larger volume of riskier loans, average loan yields were significantly lower at the worst thrifts, particularly for mortgages.' the worst thrifts increased their reliance on borrowed funds to support assets. In ]987, borrowed funds accounted for about 15 percent of the liabilities of the worst thrifts, up from less than 12 percent of the] 983 total. The majority of borrowed funds come from the district Federal Home Loan Bank. Such loans are called advances and are heavily collateralized, like loans made to depository institutions by Federal Reserve Banks.s Apparently, the worst thrifts did not adequately take credit risk into account when they priced loans. Although figures on delinquent and non performing loans for individual thrifts are not publicly available, one available indicator of past loan quality is the volume of foreclosed and Advances often constitute a low-cost, subsidized loan to problem thrifts that because of their precariOUS financial condition typically must pay higher rates for loans from alternative private sources.' Although borrowings may involve smaller transaction costs than generating and maintaining retail repossessed property held by thrifts. Since 1983, such properties have more than doubled at the worst thrifts and now account for more than one percent of their assets. This percentage is nearly four times the amount held by the best thrifts. In addition to earning lower interest revenue, the worst thrifts also incurred Significantly higher interest expense paid to creditors and depositors. This expense difference was between 53 and 64 basis points higher than at the best thrifts, and could be due to several factors. For example, the worst thrifts held relatively more liabilities to assets, had greater reliance on borrowed funds, and paid higher deposit and borrowing rates. CHART 1 EARNINGS GAP Percent 1.8 r--------------..., ~ 1.2 II) ~ 0.6 ~ § ~ O.O~'---------------------------_; ~ ·0.6 .1.2 ~-----" 1982 •••••.. -~-1983 •.....•. __ _'_ __ 1985 1984 1986 .••• 1987 SOURCE: Author's classification and Federal Home Loan Bank Board's semiannual and quarterly financial statements. As can be expected, deposits, and despite their subsidy relative to other borrowed funds, average borrowing rates have been higher than average deposit rates over the last three years, particularly at the worst thrifts. In summary, we found that increased differences in net interest margins contributed to the widening of the thrift performance gap. However, the noninterest margin played the dorninant role in the acceleration of earning differences between the best and worst thrifts. • Net Noninterest Margin The net noninterest margin is the difference between non interest income and noninterest expenses as a percentage of assets. This ratio reflects CHART 2 such as repossessed properties, securities, and loans. Noninterest expenses consist of all expenses involved with overall thrift operations such as employee salaries and benefits, marketing costs, insurance premiums, legal fees, as well as expenses of premises and other fixed assets. It also includes nonoperating expenses, such as the provision for writing- off loans and losses from selling assets. Chart 3 illustrates the importance of the net noninterest margin in explaining the expanding earnings gap and the deteriorating performance of the worst thrifts. NET INTEREST MARGIN Percent 3.0 !!1 the operational efficiency and marketing and pricing decisions of thrifts. Noninterest income includes fee income and profits from leasing office building operations, real estate investments, service corporation activities, plus income from selling assets II) '" '" 2.4 «S Since experiencing the effects of deregulation and high and volatile interest rates in the late 1970s and early 1980s, thrifts as well as commercial banks have placed increased II) OIl «S •... 1.8 ~ 1.2 c 0 c •... 0.6 ti •... 0.0 ::l ~ •... II) II) .S ti Z ·0.6 -1.2 1982 CHART 3 ~ 1983 1984 1985 1986 1987 SOURCE: Author's classification and Federal Home Loan Bank Board's semiannual and quarterly financial statements. emphasis on fee income. Thrifts generated fee income from originating mortgages, from servicing loans, from other loan charges such as prepayment fees, and from service charges on transaction accounts. Higher fees for originating and servicing loans might suggest that some thrifts are behaving more like mortgage banking firms, that is, they originate mortgages with the intent to sell them to investors. NET NONINTEREST MARGIN Percent O.Or------------------, Between 1982 and 1986, fee income accounted for most of the noninterest income earned by thrifts. The best thrifts earned more fee income, which, as a percentage of assets, rose by 50 basis points to a 79 basis- point level ~ f c o ~ ~ between 1982 and 1986, compared to a 30 basis-point gain and a 63 basispoint level at the worst thrifts. ~ II) E '§ c ti z -3.0 ••..•.. ---"---'--1982 1983 ••..•.. ---"-1984 SOURCE: Author's classification Bank Board's semiannual 1985 1986 ..•• 1987 and Federal Home Loan and quarterly financial statements. Refinancing was a Significant part of the relatively strong demand for mortgage loans over the last few years. However as mortgage rates increased in 1987, lending tapered off, causing loan-fee income to decline from 1986 levels. The best thrifts experienced an Ll-basis-point drop in fee income, which contributed to an overall 18-basis-point earnings decline in 1987. In contrast, fee income dropped by 7 basis points at the worst thrifts, which was an insignificant part of their 99-basis-point increase in last year's losses. Recent accounting changes on the treatment of loan- fee income will have adverse effects on future thrift earnings. Effective this year, regulators have required thrifts to recognize lending revenues and expenses over the life of mortgage loans rather than at the time loans are made. Although cash flow will not be affected, the accounting change will reduce reported earnings at thrifts over the next several years.s The negative impact, however, should have less effect on the worst thrifts since loan fees account for a smaller share of total income (5 percent), as cornpared to 6 percent for the best thrifts. Expenses were quite important in explaining the increasing disparity in the nonlnterest margin and overall earnings between the worst and best thrifts. The worst thrifts have been experiencing sharply rising, large nonoperating expenses that can be attributed to bad assets, to the need for cash, and to the rise in interest rates over much of last year. Nonoperating expenses accounted for 42 percent of the worst thrifts total noninterest cost in 1987, up from 27 percent in 1986 and 14 percent in 1984. Even the best thrifts were hurt significantly by a near doubling of nonoperating costs over the past year to 22 percent of their total noninterest cost. These expenses largely reflect the increasing severity of the asset- quality problem at thrifts. The worst thrifts incurred higher operating costs than their more fortunate counterparts. These differences were broad-based, with the worst thrifts paying higher salaries, office expenses, and marketing outlays. As a percentage of assets, operating expenses were 54 basis points higher at the worst thrifts than at the best thrifts in 1987, up from a 29-basispoint difference in 1984. nonlnterest margins rather than to interest margin differences. The breakdown also allows us to link balance sheet differences to revenue and expense differences between the best and worst thrifts. • Net Interest Margin The net interest margin is the difference between interest income and interest expense as a percentage of average assets. This measure indicates how well interest- earning liabilities are being used to generate interest income. Chart 2 shows that net interest margins for thrifts in general have been increasing but that improvement has been greater at the best thrifts. Interest rates have strong effects on thrift profitability. Although thrifts have reduced the holdings of longterm fixed mortgages by substituting adjustable- rate mortgages, their portfolios still consist largely of longerterm assets that are being funded by shorter- term liabilities, such as moneymarket deposit accounts and traditional savings accounts. Consequently, when interest rates fall, thrifts generally improve earnings because the cost of funds usually declines by more than interest income. For example, net interest margins improved sharply between 1983 and 1986, when interest rates fell steadily from mid- 1984 to 1986. Low and falling interest rates also benefit thrifts by increasing the demand for houses and mortgages because lower borrowing costs enhance home affordability. During the 1983·86 period, wider rate spreads and greater mortgage lending contributed to sharply higher earnings at the best thrifts and to a reduction in losses at the worst thrifts. On the other hand, thrifts are affected adversely by high and rising interest rates. After significant improvement, interest margins increased only slightly during 1987 because interest rates rose during part of the year. Interest income, constituting the bulk of thrift income, comes from loans and investments. As a percentage of assets, interest income in 1987 was 81 basis points lower at the worst thrifts than at the best thrifts. This difference has increased more than four times over the last three years. The worst thrifts avoided even lower interest income by expanding into higher-yielding consumer installment loans and business lending. By yearend 1987, nonmortgage loans, primarily in the form of consumer loans, accounted for 8 percent of outstanding loans at the worst thrifts, cornpared to less than 4 percent for the best thrifts. Not only did the worst thrifts extend more consumer and business loans, but they also earned higher yields on such loans, probably because they were more involved with credit-card and business lending. Average deposit rates paid by the worst thrifts were between 1 basis point and 16 basis points higher than those paid by the best thrifts over the last five years. Since the worst thrifts held a Significantly larger share of lower- paying transaction and traditional savings accounts, one can deduce that they either paid higher rates on sirnilar time deposits or held more higherpaying, longer-term time deposits. The worst thrifts also held a smaller share of deposits over $100,000. All deposit balances over $100,000 are not federally insured and typically require high premiums from institutions having financial difficulties. The worst thrifts also took greater risk by making more commercial mortgage loans. Nonresidential mortgages accounted for 12 percent to 17 percent of the loans held by the worst thrifts, compared to 6 percent to 11 percent of the loans held by the best thrifts over the last five years. Despite making a larger volume of riskier loans, average loan yields were significantly lower at the worst thrifts, particularly for mortgages.' the worst thrifts increased their reliance on borrowed funds to support assets. In ]987, borrowed funds accounted for about 15 percent of the liabilities of the worst thrifts, up from less than 12 percent of the] 983 total. The majority of borrowed funds come from the district Federal Home Loan Bank. Such loans are called advances and are heavily collateralized, like loans made to depository institutions by Federal Reserve Banks.s Apparently, the worst thrifts did not adequately take credit risk into account when they priced loans. Although figures on delinquent and non performing loans for individual thrifts are not publicly available, one available indicator of past loan quality is the volume of foreclosed and Advances often constitute a low-cost, subsidized loan to problem thrifts that because of their precariOUS financial condition typically must pay higher rates for loans from alternative private sources.' Although borrowings may involve smaller transaction costs than generating and maintaining retail repossessed property held by thrifts. Since 1983, such properties have more than doubled at the worst thrifts and now account for more than one percent of their assets. This percentage is nearly four times the amount held by the best thrifts. In addition to earning lower interest revenue, the worst thrifts also incurred Significantly higher interest expense paid to creditors and depositors. This expense difference was between 53 and 64 basis points higher than at the best thrifts, and could be due to several factors. For example, the worst thrifts held relatively more liabilities to assets, had greater reliance on borrowed funds, and paid higher deposit and borrowing rates. CHART 1 EARNINGS GAP Percent 1.8 r--------------..., ~ 1.2 II) ~ 0.6 ~ § ~ O.O~'---------------------------_; ~ ·0.6 .1.2 ~-----" 1982 •••••.. -~-1983 •.....•. __ _'_ __ 1985 1984 1986 .••• 1987 SOURCE: Author's classification and Federal Home Loan Bank Board's semiannual and quarterly financial statements. As can be expected, deposits, and despite their subsidy relative to other borrowed funds, average borrowing rates have been higher than average deposit rates over the last three years, particularly at the worst thrifts. In summary, we found that increased differences in net interest margins contributed to the widening of the thrift performance gap. However, the noninterest margin played the dorninant role in the acceleration of earning differences between the best and worst thrifts. • Net Noninterest Margin The net noninterest margin is the difference between non interest income and noninterest expenses as a percentage of assets. This ratio reflects CHART 2 such as repossessed properties, securities, and loans. Noninterest expenses consist of all expenses involved with overall thrift operations such as employee salaries and benefits, marketing costs, insurance premiums, legal fees, as well as expenses of premises and other fixed assets. It also includes nonoperating expenses, such as the provision for writing- off loans and losses from selling assets. Chart 3 illustrates the importance of the net noninterest margin in explaining the expanding earnings gap and the deteriorating performance of the worst thrifts. NET INTEREST MARGIN Percent 3.0 !!1 the operational efficiency and marketing and pricing decisions of thrifts. Noninterest income includes fee income and profits from leasing office building operations, real estate investments, service corporation activities, plus income from selling assets II) '" '" 2.4 «S Since experiencing the effects of deregulation and high and volatile interest rates in the late 1970s and early 1980s, thrifts as well as commercial banks have placed increased II) OIl «S •... 1.8 ~ 1.2 c 0 c •... 0.6 ti •... 0.0 ::l ~ •... II) II) .S ti Z ·0.6 -1.2 1982 CHART 3 ~ 1983 1984 1985 1986 1987 SOURCE: Author's classification and Federal Home Loan Bank Board's semiannual and quarterly financial statements. emphasis on fee income. Thrifts generated fee income from originating mortgages, from servicing loans, from other loan charges such as prepayment fees, and from service charges on transaction accounts. Higher fees for originating and servicing loans might suggest that some thrifts are behaving more like mortgage banking firms, that is, they originate mortgages with the intent to sell them to investors. NET NONINTEREST MARGIN Percent O.Or------------------, Between 1982 and 1986, fee income accounted for most of the noninterest income earned by thrifts. The best thrifts earned more fee income, which, as a percentage of assets, rose by 50 basis points to a 79 basis- point level ~ f c o ~ ~ between 1982 and 1986, compared to a 30 basis-point gain and a 63 basispoint level at the worst thrifts. ~ II) E '§ c ti z -3.0 ••..•.. ---"---'--1982 1983 ••..•.. ---"-1984 SOURCE: Author's classification Bank Board's semiannual 1985 1986 ..•• 1987 and Federal Home Loan and quarterly financial statements. Refinancing was a Significant part of the relatively strong demand for mortgage loans over the last few years. However as mortgage rates increased in 1987, lending tapered off, causing loan-fee income to decline from 1986 levels. The best thrifts experienced an Ll-basis-point drop in fee income, which contributed to an overall 18-basis-point earnings decline in 1987. In contrast, fee income dropped by 7 basis points at the worst thrifts, which was an insignificant part of their 99-basis-point increase in last year's losses. Recent accounting changes on the treatment of loan- fee income will have adverse effects on future thrift earnings. Effective this year, regulators have required thrifts to recognize lending revenues and expenses over the life of mortgage loans rather than at the time loans are made. Although cash flow will not be affected, the accounting change will reduce reported earnings at thrifts over the next several years.s The negative impact, however, should have less effect on the worst thrifts since loan fees account for a smaller share of total income (5 percent), as cornpared to 6 percent for the best thrifts. Expenses were quite important in explaining the increasing disparity in the nonlnterest margin and overall earnings between the worst and best thrifts. The worst thrifts have been experiencing sharply rising, large nonoperating expenses that can be attributed to bad assets, to the need for cash, and to the rise in interest rates over much of last year. Nonoperating expenses accounted for 42 percent of the worst thrifts total noninterest cost in 1987, up from 27 percent in 1986 and 14 percent in 1984. Even the best thrifts were hurt significantly by a near doubling of nonoperating costs over the past year to 22 percent of their total noninterest cost. These expenses largely reflect the increasing severity of the asset- quality problem at thrifts. The worst thrifts incurred higher operating costs than their more fortunate counterparts. These differences were broad-based, with the worst thrifts paying higher salaries, office expenses, and marketing outlays. As a percentage of assets, operating expenses were 54 basis points higher at the worst thrifts than at the best thrifts in 1987, up from a 29-basispoint difference in 1984. August IS, 1988 • Conclusion Less regulation, lower interest rates, and improved economic conditions during the last five years helped most thrifts improve earnings as compared to those in the early 1980s. Using Ohio data, we found earnings at the best thrifts improved dramatically, while losses at the worst thrifts were generally reduced, until 1987 when losses accelerated. Losses stemmed from a range of factors, including poor asset quality, mispricing, and operational inefficiencies. The worst thrifts held a larger share of higherrisk loans such as commercial mortgages, but they earned lower yields and fees on loans, particularly on mortgage loans. The worst thrifts were much larger than the best thrifts and they made a larger share of non mortgage loans and were involved more with business and credit-card lending. Although higher-yielding business and consumer installment loans augmented overall loan yields, income was Significantly lower at the worst thrifts. While increasing differences in both revenues and expenses caused the thrift earnings gap to widen, cost differences played a key role. The worst thrifts were hurt by higher funding costs and operating expenses, but the heaviest blow came from sharply increasing nonoperating expenses attribtited to losses from selling assets States Gener.alAccounting Office, Briefing Report to the Chairman, Committee on Banking, Housing, and Urban Affairs,United States Senate, May 1987. and writing off loans. 3. The number of FSLICthrifts oper.ating in Ohio varied from 218 to 247 between 1983 and 1987, but only 192 of those instiruuons oper.ated in each of the five years. With larger losses in 1987, the worst thrifts produced a negative net worth. Their chances of recovery don't look good, especially since they were unable to capitalize on favorable mortgage-market conditions during the last few years-particularly in 1985 and 1986 when some thrifts did exceptionally well. Paul R. Watro is an economist at the Federal Reserve Bank of Cleveland. The author would like to thank Mark S. Sniderman and fames B. Thomson for their comments, and would like to thank Daniel j. Martin for research assistance. The views stated herein are those of the author and not necessarily those of the Federal Reserve Bank of Cleveland or of the Board of Governors of the Federal Reserve System. • Footnotes 1. An insolvent institution is one that has negative net worth based on gener.ally accepted accounting principles (GAAP). According to the Gener.al Accounting Office, the FSLICis technically insolvent with $13 billion more in liabilities than in assets. eCONOMIC COMMeNTaRY Federal Reserve Bank of Cleveland 4. Even in the early 1980s, Ohio's wor.st thrifts gener.ally did not attempt to grow out of the interest-rate problem as did some thrifts around the country, particularly in California and Texas. 5. Asset and loan yields are influenced by risk. Riskier loans should yield higher revenue because they impose additional expenses on lenders related to loan delinquencies and defaults. Is the Thrift Performance Gap Widening? Evidence from Ohio 6. Advances are much easier to acquire than are loans from Feder.alReserve Banks because discount-window borrowing is restricted to liquidity needs such as replacing deposit outflows. In contrast, advances can be used to make new loans, to purchase other assets, as well as to replace deposit outflows. Advances also vary in maturities up to sever.alyears. Moreover, the Feder.al Home Loan Bank Board has allowed the FSLICto act as guar.antor for advances when troubled thrifts did not have adequate collater.al. by Paul R Watro The savings and loan (thrift) industry reported a record loss of $6.8 billion in 1987, followed by a $3.8 billion loss for the first quarter in 1988. These losses are misleading, however, because they do not reflect the good performance of many thrifts. 7. Other borrowing comes largely from reverse repurchase agreements and mortgage-backed bonds. For example, over two-thirds of the nation's more than 3,100 thrifts were profitable last year with total earnings of $6.6 billion. Unprofitable thrifts, however, lost $13.4 billion in 1987, dragging down the industry's overall performance. 8. See William McGuire and Angelo DiMarzio, "The Profitability Impact of SFAS91,"Fifth District Review, Feder.al Home Loan Bank of Cincinnati, March 1988, pp. 18-21. 2. See, "Thrift Industry: Forbear.ance for Troubled Institutions 1982-1986," United Federal Reserve Bank of Cleveland Research Department p.o. Box 6387 A major reason for the large losses is that regulators continue to allow many insolvent thrifts to remain open because the federal insurance agencythe Federal Savings and Loan Insurance Corporation (FSLIC)-does not have adequate resources to close them.' As a result, over 500 insolvent thrifts continued operations in 1987, hurting the industry in general. BULK RATE U.S. Postage Paid Cleveland, OH Permit No. 385 Cleveland, OH 44101 Record losses in the unprofitable segment of the industry have sparked growing concern about what is being called the "thrift crisis." The crisis, which began in the early 1980s, is not new but the underlying problem causing it has changed. Today's crisis stems from poor-quality assets, whereas in the early 1980s it was due to high interest rates.' Material may be reprinted provided that the source is credited. Please send copies of reprinted materials to the editor. Address Correction Requested: Please send corrected mailing label to the Federal Reserve Bank of Cleveland, Research Department, P.O Box 6387, Cleveland, OH 44101 Since the end of the recession in 1982, interest rates have fallen sub- ISSN 0428- I 276 stantially, the economy has recovered, and the thrift industry has been deregulated. These developments gave thrifts promising opportunities to counteract interest-rate problems and to improve their precarious financial condition. Thrifts have reacted differently to the regulatory and economic changes. Some placed greater emphasis on growth and were less cautious about risk-taking and expenses. Others were more risk-averse and focused on improving profit margins through cost containment and traditional mortgage lending. • Performance Gap To illustrate contrasting thrift performance, we examined the best and worst thrifts that operated continuously in Ohio between 1983 and 1987.3 Performance was measured by the average return on assets. Based on five years of earnings, we classifledthe thrifts that ranked in the top 20 percent as the best performers and those in the bottom 20 percent as the worst performers. Each group in our study contained 38 thrifts. The worst performers had much larger assets and experienced less growth than the best performers. Average assets at the worst thrifts were $360 million, nearly twice as large as assets at the best thrifts. Over the last five years, asset growth at the worst thrifts averaged 4 percent annually, compared to 15 percent for the best thrtfts.s In examining thrift performance, the author discusses the differences in the earnings of the best and worst thrifts in Ohio and shows that the healthy thrifts are getting better and the sick thrifts are getting worse. Chart 1 shows that the earnings gap between the best and worst thrifts has widened substantially since the early 1980s. The expanding disparity is due not only to the dramatic improvement of the top performers, but also to the continued losses incurred by the bottom segment of the industry. In fact, the poor performers suffered their largest losses in 1987 when they recorded average losses amounting to 1 percent of assets. At year-end 1987, nearly one-half of the 38 poor performers were insolvent and their average net worth as a percentage of assets was -0.4 percent, down from 1.7 percent at year-end 1983. By separating earnings into two components-interest and noninterest -one can see in charts 2 and 3 that the increasing earnings disparity is largely attributed to differences in
