The full text on this page is automatically extracted from the file linked above and may contain errors and inconsistencies.
____________ Review____________ Vol. 67 1 No. 8 O ctober 1985 5 Ilvo Views of the Effects of Government Budget Deficits in the 1980s 17 The Status of Farm Lenders: An Assessment of Eighth District and National Trends The Review is published 10 times p e r year by the Research and Public In form ation D epartm ent o f the Federal Reserve Bank o f St. Louis. Single-copy subscriptions are available to the public fr e e o f charge. Mail requests f o r subscriptions, back issues, o r address changes to: Research and Public In form ation Departm ent, Federal Reserve Bank o f St. Louis, P.O. Bop 442, St. Louis, M issouri 63166. The views expressed are those o f the individual authors and do not necessarily reflect official positions o f the Federal Reserve Rank o f St. Louis o r the Federal Reserve System. A rticles herein may be reprinted provided the source is credited. Please provide the Rank's Research and Public Inform ation D epartm ent with a copy o f reprinted material. Federal Reserve B ank of St. Louis Review O ctober 1985 In This Issue . . . Leonall C. A n dersen 1924-1985 It is w ith a deep sense of loss that we note the passing of our friend an d colleague Leonall "Andy" Andersen. A ndy ’s contributions to the econom ics profession, m any of w h ich were published in this Review, were n um erous an d significant. The focus of A ndy’s research after joining the Federal Reserve Bank of St. Louis in 1962 was on the im pact of m onetaiy an d fiscal actions on econom ic activity. Am ong his best-known contributions was the article, "M onetaiy and Fiscal Actions: A Test of Their Relative Im portance in Econom ic Stabilization,” w hich, co-authored w ith Je n y Jordan, appeared in this Review in 1968. Andy held num erous positions w hile at the St. Louis Fed, in clu d in g director of research and, at the tim e of his retirement in 1978, econom ic advisor. Upon retiring from the bank, Andy held an endow ed chair in m oney a n d banking at the University of Florida until 1981, w hen he became professor of econom ics at Gustavus A dolphus College in St. Peter, Minnesota, a position he held until his death on October 27, 1985. ★ ★ ★ The effects of ballooning federal budget deficits on the econom y have been a m ajor concern for some time. In the first article in this Review, "Two Views o f the Effects of Government Budget Deficits in the 1980s,” Jo h n A. Tatom explains the predicted effects of deficits based on the conventional analysis an d a com peting theoretical approach. The conventional analysis, according to Tatom, e m p h a sizes that deficit-increasing fiscal policies initially result in increased de m an d for goods an d services an d a reduced supply of national saving. These two effects lead to increases in national output, em ploym ent, prices an d interest rates. The dow n side of such policies is a decline in private saving an d investment in business structures, equipm ent an d inventory, an d housing. Since the sharp increase in deficits in 1981-82, econom ic developments have been sharply at odds with these predictions. Tatom explains that the alternative set of hypotheses em phasizes the substitut ability between government and private expenditures an d the im portance of perm anent instead of m easured incom e in private spending decisions. These hypotheses indicate that deficit-increasing fiscal actions do not raise d e m an d for goods an d services or reduce the supply of national saving. Therefore, they do not affect output, em ploym ent, prices or interest rates. Fiscal deficits arising from government expenditures directly “crowd-out” private spending, especially in vestment, w hile those arising from tax reductions have no direct effect on national saving, investment or interest rates. Com paring the changes in the shares of private saving a n d investment in GNP in 1980 an d 1984, Tatom indicates that both have risen sharply, instead of declining as the conventional view w o uld predict. The author explains that investment incentives provided in the 1981 tax act played a central role in raising the private saving rate and the share of real business investment in real GNP to 3 In This Issue Digitized 4for FRASER record levels in 1984. These increases are more broadly consistent w ith the alternative hypotheses than w ith the conventional view. ★ ★ ★ The plight of the nation's farmers has received a great deal of attention this year. M u ch of it has focused on the deteriorating financial condition of farmers overburdened w ith debt an d faced w ith falling land values a n d com m o d ity prices. In the second article of this Review, Michael T. Belongia a n d Kenneth C. Carrara examine the performance of the m ajor lenders to the farm sector to gain some perspective about the sector’s financial condition. These financial institutions are the Farm Credit System, agricultural banks a n d the Farmers Hom e A dm inistra tion (FmHA). Belongia an d Carrara first offer a brief description of these three agricultural lenders. They suggest that institutions that sharply increased their lending to agriculture during the 1970s an d early 1980s, w hen inflation, foreign d e m an d for U.S. farm products an d real com m odity prices were increasing or were expected to increase rapidly, should be experiencing the greatest deterioration in portfolio quality. The available data indicate that the performance o f agricultural loans, as measured by delinquencies an d loan losses, has worsened at all three lenders since 1982. In terms of profitability, however, agricultural banks have suffered less severe dow nturns th an Farm Credit System lenders. The Farmers Home A d m in is tration (FmHA), w h ich is a governmental agency a n d therefore does not report profitability data, also has exhibited greater forbearance on delinquent farm loans. Although nearly 25 percent of their farm ow nership loans were delinquent in 1984, only 0.22 percent of all FmHA farm loans were written off. Two Mews of the Effects of Government Budget Deficits in the 1980s John A. Tatom _l_ EDERAL budget deficits in the United States have become a m ajor concern since they rose to nearly $200 billion in fiscal 1983. In the absence of new policy efforts, the deficit is projected to continue at $200 to $250 billion per year for the rest of this decade. Deficits, according to m ost p o pu lar analyses, raise aggregate d e m an d for goods, services and credit, w hich boosts output, em ploym ent, prices an d interest rates an d reduces private investment.' This article examines the em pirical and theoretical basis of this m ainstream view. It also presents an alternative set of hypotheses, w hich indicates that fiscal policy actions are largely an d directly offset by private spending changes, rendering the aggregate dem and an d interest rate channels of influence in significant. T H E CON VEN TION AL ANALYSIS AND R E C E N T E X P E R IE N C E Conventional w isdom holds that recent a n d pro spective U.S. budget deficits have significantly raised John A. Tatom is an assistant vice president at the Federal Reserve Bank of St. Louis. Michael L. Durbin provided research assistance. An earlier version of this paper was presented at the Spring Conference of the Financial Research Foundation of Canada on April 16, 1985, in Ontario, Canada. 'Most introductory textbooks emphasize the boost to aggregate demand, interest rates and prices arising from “ expansionary” fiscal policy actions. See, for example, the macroeconomics sections of Dolan (1983), McConnell (1984), or Samuelson and Nordhaus (1985). These texts also discuss some of the theoretical reserva tions about these channels of influence raised below. Note that the hypothesized reduction in investment does not exceed the initial rise in aggregate demand for goods and services that arises from deficitincreasing fiscal actions. interest rates a n d have prom oted the crowding out of investment. But this view is based on the conventional deficit/aggregate-demand hypothesis that also holds that an expanded deficit should increase both output and the price level. The latter conclusions became center stage in 1980-81 w h e n the Reagan econom ic program was debated. Their em pirical validity, w hich remains largely unquestioned, was strongly rejected after mid-1981 w hen, w ith the deficit expanding, in flation plu m m e te d from double-digit levels an d the econom y entered the longest an d m ost severe reces sion since the 1930s. Interest rate developments were also at odds with the conventional view. Chart 1 shows the total govern m ent deficit as a percent of GNP a n d the AAA b on d yield since 1950. The surge to historically high interest rates occurred well before the 1981-82 surge in the deficit.2T h e recent rise in the deficit occurs from the third quarter of 1981 to the fourth quarter of 1982, w hen the AAA bond yield declined from about 15 percent to 12 percent. Then, in 1983-84, the deficit declined sharply relative to GNP, but the AAA bond yield rose. The principal difficulty in finding a positive relation ship between deficits an d interest rates arises from the fact that both the budget deficit an d interest rates move cyclically an d in opposite directions. Hence, it is not surprising, especially for short-term interest rates, that empirical studies often turn u p supposedly sig 2Similarly, the appreciation of the dollar precedes the deficit surge. The steady upward appreciation of the trade-weighted exchange rate for the U.S. dollar began at the end of 1979 (when interest rates soared) and was not noticeably affected by the 1981-82 deficit surge. 5 FEDERAL RESERVE BANK OF ST. LOUIS OCTOBER 1985 A a a Bond Yield and Total Governm ent Deficit as a Percent of GNP P e rc e it 15 1950 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 1985 S h a d e d a re a s re p re s e n t pe rio ds o f business recessions. Latest d a ta p lo tte d : 2 n d q u a rte r nificant negative statistical relationships between in terest rates and deficits. W hen one uses deficits con structed on a high-employment basis — that is, w ith systematic cyclical influences removed — there still is no evidence of a positive relationship between deficits and either short- or long-term interest rates over the period 1955-83.' 3See Tatom (1984). Efforts to control for future inflation expectations to capture real interest rate changes do not affect the observed absence of a deficit effect on interest rates. Also, some analysts conjecture that the debt/GNP ratio positively influences the interest rate. Regressions of quarterly changes in the AAA bond yield or three-month Treasury bill rates on changes in the ratio of net federal debt to GNP, controlling for changes in the capacity utilization rate and the inflation rate one quarter ahead, yield a negative but insig nificant relationship for the debt ratio over the period 1/1955 to III/ 1984. The independence of interest rates from the deficit has been observed by Evans (1985). Also see Feldstein and Eckstein (1970), Sargent (1973), and the recent Treasury study (1984). Plosser (1982) details many of the theoretical and econometric difficulties of previous tests of the interest rate/deficit hypothesis. Digitized 6 for FRASER Has Recent Crowding Out Reduced Investment? Chart 2 shows private dom estic investment as a share of GNP. Investment has declined a n d risen cycli cally since 1980, but these swings obscure the strength of investment over the past four years. W he n the capacity utilization rate is relatively low, the dem and for new capital can be met more easily by the re em ploym ent of existing capital instead of investment in new facilities. Thus, the share of investment in GNP an d the capacity utilization rate tend to move in tan dem or to be positively correlated. In 1984, the capac ity utilization rate was well below its 1979 level, w hen the prior peak investment ratio was achieved. Never theless, the share of private dom estic investment in GNP in 1984 virtually m atched this peak level. Even plant an d equipm ent (nonresidential fixed) investment has been quite high by historical stand ards, despite the recessions in 1980-82. W hen nom in al FEDERAL RESERVE BANK OF ST. LOUIS OCTOBER 1985 C h a rt 2 Nonresidential Fixed Investment as a Percent of GNP Perceat P erceit Mi 1950 S I -- 14 52 53 54 55 56 57 51 59 60 S h ad ed are as re p re s e n t p e r io d s o ( bu s in e s s recessions. Latest d a ta p lo tte d : 2 n d q u a rte r nonresidential fixed investment and GNP are adjusted by their respective deflators, in order to measure real investment as a share of real GNP, the recent strength of plant an d equipm ent spending relative to real GNP represents a postwar peak performance. This share is show n in chart 3. Note that, even at the depths of the previous two recessions, real plant an d equipm ent purchases were about as large a share of real GNP as the 11 percent attained at the peaks of previous invest m ent boom s in 1966 and 1969.4The conventional argu ment, that investment has been unusually weak due to the higher real rates of interest, is not obviously im p or tant in explaining recent investment experience. “The reason for the greater real strength is that the relative price for new plant and equipment declined sharply since 1980. The nonresi dential fixed investment deflator declined 14.2 percent relative to the GNP deflator from 1980 to 1984. T H E M A C R O E C O N O M IC E FF E C T S O F FISC A L P O L IC Y A closer look at the theoretical m echanism underly ing conventional analyses of the deficit reveals some of the potential shortcom ings of these analyses. In the textbook view of the effects of fiscal policy on the economy, increased government deficits expand ag gregate dem and, spending, o u tp u t a n d em ploym ent, regardless of w hether larger deficits arise from in creases in purchases, transfer paym ents or reductions in taxes. So-called balanced-budget increases in trans fer payments, in w h ic h a rise in transfer paym ents is m atched by a rise in taxes, leave aggregate dem and unchanged (ignoring distribution effects), while taxfinanced increases in government purchases raise ag gregate dem and. Such conventional analyses also take into account crowding out — reductions in private spending that 7 FEDERAL RESERVE BANK OF ST. LOUIS OCTOBER 1985 C h a rt 3 Private Domestic Investment as a Percent of GNP 1 9 5 0 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 8 4 1985 L a te s t d a ta p lo tte d : 2 n d q u a r te r occur due to fiscal policy changes. Increases in the deficit, so the argum ent goes, result in increased com petition in credit markets, thus b id din g u p interest rates. Also, if fiscal policy actions raise aggregate de m and, the increased com petition in the market for goods a n d sendees bids u p the general level of prices. For both reasons, real private spending is reduced, or crowded out. Households reduce their current real consum ption expenditures a n d increase saving; firms reduce real investment spending in response to a higher interest rate. Fiscal Policy Crowding Out: Saving, Investment and the Interest Rate The im portant link between fiscal policy, aggregate dem and a n d interest rates an d the concept of crow d ing out of private expenditures can be illustrated in the market for saving. In figure 1, the de m an d for saving is taken to be the dem and for funds to finance Digitized 8 for FRASER investment. O ther things that influence investment rem aining the same, the d e m an d for investment or for saving to finance it, is inversely related to the interest rate. The supply of saving consists of private saving — household disposable incom e less desired co n s u m p tion expenditures — a n d government saving — the excess of tax receipts over government expenditures, or the budget surplus. In figure 1, the national saving schedule is draw n as upward-sloping, indicating that, given income, households reduce consum ptio n ex penditures a n d save more at higher interest rates. In equilibrium at p oint A, the interest rate equates the supply and d em an d for national saving at interest rate In the conventional analysis, fiscal policy actions affect national saving, investment an d the interest rate by (1) directly changing the budget surplus or govern m ent saving, and/or (2) altering private saving. Such changes shift the national saving schedule. Given GNP FEDERAL RESERVE BANK OF ST. LOUIS F ig u r e 1 National Saving, Investment and the Rate of Interest OCTOBER 1985 change. In this case, however, the rise in aggregate dem and is the governm ent’s, w hile before it was the policy-induced change in private co n sum ptio n ex penditures. As before, however, interest rates w ill tend to rise, increasing private saving a n d reducing con sum ption an d investment expenditures. Tax-financed changes in government purchases, on the other hand, reduce private saving, given the inter est rate an d GNP. The higher tax reduces disposable incom e and therefore both con sum ptio n expendi tures an d private saving. The reduction in private saving is less than the tax increase, because private expenditures on goods an d services also decline. Since the government deficit does not change with such a fiscal action, the decline in national saving equals the reduction in private saving. The reduction in national or private saving again indicates a rise in aggregate d e m an d for goods an d services. Taxfinanced changes in transfer payments have no effect on aggregate de m an d for goods a n d services or the national saving schedule in figure 1, since the govern m ent deficit an d disposable incom e rem ain u n changed. Thus, private an d total spending on goods an d services an d private an d national saving are unaffected. an d interest rates, a fall in taxes or a rise in transfer payments (financed by borrowing) adds to disposable income, increasing both private consum ptio n ex penditures an d private saving. Since part of the tax cut or transfer paym ent is spent for consum ption, the rise in private saving is less than the deficit increase. Thus, national saving declines. Such a decline also indicates that desired aggregate d em and for goods and services has risen so that it m ust exceed the given level of GNP; the aggregate d em and increase equals the reduction in national saving. W ith no change in GNP, the interest rate m ust rise to equate national saving an d investment. In re sponse to the higher interest rate, investment is crowded out, or declines, but some of the initial short fall in national saving is elim inated since individuals also increase private saving. A rise in government purchases also affects national saving. In contrast to a tax cut or a rise in transfer payments, a rise in government purchases does not change disposable income, so consum ptio n expendi tures and private saving rem ain unchanged. But the rise in purchases raises the budget deficit or reduces government saving. National saving falls by exactly the change in aggregate d e m an d for goods a n d services, as was the case above for the tax or transfer paym ent In summary, the initial effects of fiscal policy actions on private an d national saving are the critical counter parts of any initial change in aggregate d e m an d for goods an d services; both indicate the extent of upw ard pressure on interest rates. The analysis here illustrates the im portance of both of these initial shifts. It also indicates w hy crow ding out tends to occur. In the conventional analysis, however, crow ding out is gen erally presum ed not to be complete." C row ding In The growth in aggregate d e m an d associated w ith reductions in national saving can raise or “crowd in ” GNP. W hen GNP rises, disposable incom e, co n s u m p tion expenditures an d private saving rise; the initial reduction of private an d national saving is offset bv 5The Council of Economic Advisers (1985), pp. 70-77, suggests that economic theory and evidence support “ complete” crowding out, where the total real demand for goods and services is unaffected by fiscal policy actions. Whether this crowding out, primarily of invest ment, arises through interest-rate or price-level crowding out or direct substitution of public for private expenditures is not indicated. The Congressional Budget Office (1985) also provides a detailed discussion of the effects of deficits. A recent review by Brunner (1984) provides the best recent discussion of the theoretical issues associated with the macroeconomic theory of fiscal policy. Also, see Carlson and Spencer (1975). 9 FEDERAL RESERVE BANK OF ST. LOUIS increases in both as GNP increases. The full adjust m ent of GNP, however, w ith interest rates constant, cannot raise national saving back to its initial level, so the interest rate increase and crow ding out w ill still occur. Note, however, that GNP cannot increase, just as interest rates cannot rise, unless the initial reduc tions in national saving occur. Some Reservations: The Permanent Income Hypothesis and Ex Ante Crowding Out An alternative set of hypotheses about the effects of fiscal policy actions on the economy, sometimes called classical or Ricardian, em phasizes two theoreti cal considerations called the perm anent incom e hy pothesis a n d e?c ante crow ding out. According to this view, consum ptio n expenditures are a function of permanent income; consequently, variations in saving (and saving relative to GNP) have a large cyclical c om ponent.6 The perm anent incom e hypothesis also en tails a government budget constraint, w h ich indicates that the present value of current an d future govern m ent expenditures m ust equal the present value of current a n d future taxes. This constraint im plies that the m ethod of financing government expenditures is irrelevant; that is, w hether current expenditures are financed through taxation or borrowing (future taxes w ith an equivalent present value) has no influence on the econom y/ Thus, changes in taxes are offset by equal changes in private saving, an d national saving is unaffected. The second consideration is that government ex penditures are, to some degree, substitutes for private expenditures.8 For example, an increase in govern m ent expenditures for school lunches may reduce private consum ptio n expenditures on such goods; increased public expenditures for transportation ser OCTOBER 1985 vices m ay reduce private d e m an d for such investment goods; increased transfer paym ents provide assis tance that m ay substitute for private saving a n d invest ment. To the extent that such substitution occurs, growth in government purchases crowds out private purchases w ith no net effect on econom ic activity; such growth in government purchases results in off setting reductions in private expenditures in clud in g investment. Similarly, growth in transfer payments can affect the m ix of desired private spending. No excess de m an d for national saving occurs, n o r is ag gregate d e m an d for goods a n d services altered; thus, GNP a n d interest rates are not affected by fiscal policy. The em phasis in this view of fiscal policy is on e?c ante crowding out, in w h ich fiscal policy actions are largely offset by direct private sector responses.11An increase in government purchases does not have to affect the interest rate; either national saving could remain unaffected by governm ent purchases, as these substitute for private consum ptio n, or investment de m a n d cou ld be reduced equally, as government p u r chases substitute for investment purchases. Similarly, national saving an d private investment can be re duced due to increased transfer payments. Thus, ag gregate dem and, interest rates a n d the price level may not be affected by fiscal actions '" If e?c ante crow ding out leads to private expenditure changes that fully offset fiscal policy actions, then the effects of fiscal actions on the private a n d national saving w ill not be the same as in the conventional analysis. One fundam ental difference is that a rise in taxes w ill reduce private saving by an equal am ount. Thus, a tax hike will result in an equal reduction in private saving, leaving national saving un changed.11 This im plies that the effects of government expendi tures on national saving are the same w hether they are tax- or bond-financed. Another m ajor difference is that a rise in govern- 6Textbook analyses typically distinguish between permanent and temporary changes in fiscal actions, based on the permanent in come hypothesis. Temporary changes in taxes or transfer payments are generally regarded to have little effect on private spending or national saving since such changes do not alter perceptions of permanent income or wealth. A type of temporary, or at least transitory, change in the budget arises from the “ cyclical deficit.” When unemployment rises due to a cyclical fall in income, tax receipts decline and federal expenditures, especially transfer pay ments for unemployment insurance, rise. As a result, the budget deficit rises. T h is consideration has come to be called the Ricardian Equivalence Theorem. It is developed by Barro (1974, 1979) and has received strong support from Plosser (1982), Aschauer (1985), Tanner (1979) and Kormendi (1983). See also Kochin (1974). 8Bailey (1971) discusses at length the theoretical possibilities that fiscal actions directly influence private sector behavior. Digitized 10 for FRASER 9There are exceptions to the conclusion that fiscal actions do not affect aggregate demand. See Hall (1980) and Barro (1981) for discussions of the real output effects of temporary increases in government purchases, especially defense expenditures, even in a Ricardian world. 10The absence of effects of fiscal actions on GNP has been a feature of reduced-form estimates like the St. Louis equation for some time. See Hafer (1982) and the references there for recent analyses. Permanent adverse effects of government expenditures on invest ment are found in Carlson (1982). Also see the references in footnote 1. "In the conventional view, a rise in taxes initially reduces disposable income by an equal amount and results in a fractional reduction in private saving. The fraction, called the marginal propensity to save, is generally regarded to be relatively small, on the order of 20 to 30 percent. FEDERAL RESERVE BANK OF ST. LOUIS OCTOBER 1985 Table 1 The Share of Government Budget Components in GNP: 1980 and 1984 1980 1984 33.0% 34.3% 1.3% 20.4 12.6 20.4 13.9 0.0 1.3 Total government receipts 31.8 31.0 -0 .8 Total government surplus - 1 .2 -3 .4 -2 .2 Total government expenditures Purchases of goods and services Transfer payments Change m ent purchases w ill reduce private consum ptio n or raise saving, if such purchases are a substitute for private consum ptio n expenditures. Similarly, a rise in transfer payments w ill reduce private saving an d in vestment, if such paym ents are substitutes for saving. This can occur independently of distribution effects that in either view can yield a reduction in private saving. Finally, in the classical view, the effects of government expenditures on national saving can be associated w ith equal shifts in investment dem and that reflect the extent to w h ich government expendi tures an d investment are substitutes. Recent Fiscal Policy Developments and Saving and Investment A com parison of the im plications of the two views above can be facilitated by a look at the experience in the 1980s. Table 1 shows the principal com ponents of the total government surplus as a share of GNP in 1980, before the ballooning of the federal deficit, and 1984, the latest year available. From 1980 to 1984, the deficit w idened from 1.2 percent to 3.4 percent of GNP. The share of government purchases was unchanged, while the share of transfer paym ents rose. The rise in the deficit was accounted for prim arily by a rise in transfer payments and, to a smaller extent, by a decline in taxes.'3 These changes are explained to only a small "Several recent studies have examined the effects of fiscal actions on personal consumption expenditures in tests of ex ante crowding out. See Aschauer (1985), Feldstein (1982) and Kormendi (1983). These tests allow for direct substitution of government purchases for private consumption expenditures and transfer payments for private saving; they do not address the extent to which government expenditures directly affect private investment expenditures. ,3The decline in the share of government receipts in GNP matches the decline in the share of corporate profit tax liability in GNP. extent by relative differences in the cyclical perfor m ance of the econom y in 1980 an d 1984. The average unem ploym ent rate of 7.5 percent of the civilian labor force in 1984 was only slightly higher th an the 7.1 percent in 1980. W hen unem ploy m ent is higher, gov ernm ent transfer paym ents (especially unem plo y m ent com pensation) are higher, and, due to cyclical losses in income, tax paym ents are lower than they w o uld be otherwise. In the conventional analysis, the effect of the changes in the fiscal stance of the government sector show n in table 1 on saving is to raise the private saving rate (PSR) bv a fraction — on the order of about 25 percent — of the increased deficit (2.2 percent) or roughly 0.5 to 0.6 percentage points. Since the ex pected rise in the PSR is smaller than the rise in the deficit, the national saving rate (NSR) w o uld be ex pected to fall by the difference, about 1.6 to 1.7 per centage points. Associated w ith this shift in the na tional saving rate is an increase in the share of GNP allocated for co n sum ptio n expenditures a n d an ex cess dem an d for funds to finance investment. In the conventional view, aggregate d e m a n d sh o uld have risen, im proving the cyclical performance of the econom y a n d raising prices, a n d interest rates should have risen; the latter, of course, should have lowered investment. In the classical view, part of the increased deficit arose from the reduction in receipts as a share of GNP; this part is expected to be largely offset by a rise in the PSR, leaving the national saving rate unchanged. The rem ainder of the rise in the deficit, the rise in the share of transfer payments, w o uld be expected to reduce private saving a n d investment to the extent that households view transfer paym ents as substitutes for such avenues of w ealth accum ulation. Thus, the PSR a n d NSR could be expected to decline by some frac tion of the 1.3 percentage-point rise in transfer pay m ents. As a net result of these two forces, the PSR should rise by u p to 0.8 percentage points, a n d the NSR should decline slightly. Interest rates an d the cyclical com ponents of real GNP an d em ploym ent should be unchanged. Com paring 1984 w ith 1980, two central differences in expectations emerge between the conventional and classical views. These differences concern interest rates an d the cyclical perform ance of the economy. The cyclical performance of the econom y was slightly worse in 1984 than in 1980. Interest rates were gener ally higher in 1984 th an in 1980, despite a decline in inflation. For example, in 1980, the consum er price index rose 13.5 percent, w hile rising only 4.3 percent 11 OCTOBER 1985 FEDERAL RESERVE BANK OF ST. LOUIS C h a rt 4 Governm ent Surplus, N ational Saving and Private Saving as a Percent of GNP NOTE: D a ta a re n a tio n a l in c o m e an d p ro d u c t account ba sis. |_1 The ra tio o f gross p r iv a te s o v in g to GNP. [2 P riv a te s a vin g plus the g o v e rn m e n t s u rp lu s as a p e rce n t o f GNP. S h ad ed a re a s re p re s e n t p e r io d s o f business recessions. Latest d a ta p lo tte d : 2 n d q u a r te r in 1984. The average Aaa b o n d weld, however, aver aged 11.94 percent in 1980, w hile averaging 12.71 per cent in 1984. Thus, cyclical developments are more consistent w ith the classical view, but interest rate developments, considering these two years, are more consistent w ith the conventional w isdom .14 Of course, other factors that influence interest rates an d cyclical performance are not likely to have re m ained the same, an d more careful control for these factors is necessaiy to discrim inate between the hy potheses. Some insight into the im portance of these other factors can be gained by exam ining the other im plications of these hypotheses. 14ln fact, interest rate movements have not followed the deficit over the whole period, only in the two years indicated. This apparent contradiction arises from the fact that most of the increase in the deficit occurred in 1982 when interest rates had been declining and continued to decline, while the rise in interest rates over this period occurred in 1980 and 1981, and the decline in inflation occurred in 1981. Digitized 12for FRASER W hile the im plications of fiscal developments in these two years for the saving rates are sim ilar in the two views, it is useful to examine w hat h app e n e d to these rates. Chart 4 shows the private saving rate, government surplus share a n d national saving rate from 1950 to the end of 1984.'’ The PSR has been fairly constant com pared w ith the NSR. For example, from 1/ 1950 to IV/1984, the PSR averaged 16.5 percent, exhib ited no trend a n d h ad a standard deviation of only 1.02 percentage points.,c Such behavior, however, m ay 15ln the national income and product accounts, private saving in cludes both personal and business saving. Private saving plus government saving equals national saving. Government saving is the budget surplus of the federal, state and local governments. Gross private domestic investment equals national saving plus net foreign saving. 16This near constancy has been formulated as "Denison’s Law” which indicates that private saving is proportional to high-employment or trend GNP but is cyclical, rising in expansions and declining during recessions. Denison's Law is developed in Denison (1958), Hick man (1966) and David and Scadding (1974). The latter indicate that Denison’s Law provides strong support for Friedman’s (1957) permanent-income theory of consumption. FEDERAL RESERVE BANK OF ST. LOUIS OCTOBER 1985 Table 2 Shares of Components of Gross Private Saving in GNP: 1980 and 1984 Gross private saving1 Personal saving Business saving: Undistributed corporate profit2 Corporate capital consumption allowance2 Noncorporate capital consumption allowance with capital consumption adjustment Addendum: Corporate capital consumption adjustment 1980 1984 16.5% 18.4% Change 1.9% 4.2 4.3 12.4 1.2 (1.8) 6.8 (6.2) 14.2 3.2 (1.6) 6.7 (8.2) 4.3 4.3 0.0 - 0 .6 1.5 + 2.1 0.1 1.8 2.0 (-0 .2 ) (- 0 .1 ) ( + 2.0) 'Figures do not add due to rounding. 2The top entry for each of the components of corporate business saving includes the capital consumption adjustment. The figure in parentheses excludes the capital consumption adjustment. obscure the conflicting effects of various influences on the PSR. The NSR appears to be strongly cyclical, declining sharply in recessions. This pattern m ust arise from cyclical movements in government saving since the PSR does not appear to be cyclical. Cyclical differences may not have exerted a strong influence in com paring 1984 to 1980 performance, however. From 1980 to 1984, the PSR rose sharply from an average of 16.5 percent to 18.4 percent. Based on the conventional analysis, this rise is sharply higher than that expected. Similarly, the national saving rate fell from 15.4 percent to 15.0 percent, m uch smaller than the decline expected from the conventional analysis, but it may also be smaller than that expected from the classical view. The counterpart of strong saving, domestic invest ment, has been even stronger since 1981. In the na tional incom e a n d product accounts, gross saving equals gross investment, except for a m in o r statistical discrepancy. Gross private domestic investment as a share of GNP rose 2.1 percentage points to 17.4 percent in 1984, despite the 0.4 percentage-point fall in the national saving rate. This difference is accounted for by the inflow of net foreign saving, the largest share of w hich was due to domestic firms channeling their own funds from investment abroad into domestic investment. In 1980, U.S. assets abroad rose $96.3 bil lion, but this pace of investment plum m eted to $20.9 billion in 1984. The pace of foreign investment in the United States increased slightly over the period. For eign assets in the United States rose $98.8 billion in 1984, up slightly from the $84.7 billion pace in 1980. As a result, net foreign investment fell from an $11.6 billion outflow in 1980 to a net inflow of $77.9 billion in 1984. The recent behavior of investment suggests a strong candidate for the significant om itted factor account ing for the strength of dom estic saving and the rise in interest rates. This factor, the investment incentives in the 1981 tax act, accounts for the relative strength of investment, despite the higher level of interest rates in 1984 than in 1980. More direct evidence of these effects can be seen in the d o m in an t com ponent of saving in the United States, business saving. Table 2 provides a sum m ary of com ponents of pri vate saving in 1980 an d 1984. The rise in the private saving rate was virtually all due to an increase in the business saving rate. The latter, in turn, arose almost completely because of an increase in the corporate capital consum ptio n adjustm ent as a share of GNP. This figure corrects reported profits an d capital con sum ption (depreciation) allowances for the un d e r statement or overstatement of true econom ic depreci ation, in cluding losses from the use of historical rather 13 FEDERAL RESERVE BANK OF ST. LOUIS OCTOBER 1985 Table 3 The Federal Budget as a Share of GNP 1990' 1980 1981 1982 1983 1984 22.4% 20.4 2.0 23.2% 21.1 2.0 24.3% 20.6 3.7 25.3% 19.5 5.8 24.0% 19.2 4.8 Current Services Administration Share in GNP of: Expenditures Receipts Deficit 23.5% 19.5 3.9 20.9% 19.5 1.4 'Unified Budget Estimates from Budget of the United States Government 1986. than replacement cost in com p uting depreciation al lowances. The sharp change in this adjustm ent reflects the slow dow n in inflation from 1980 to 1984, reducing the extent of underdepreciation due to his torical cost accounting; more im portant, the change in the adjustm ent reflects the acceleration of deprecia tion allowed by the 1981 tax act. The latter is indicated by the large ju m p in the size of the corporate capital consum ption allowance (without capital co n sum p tion adjustm ent) relative to GNP. This ju m p accounts for the reported rise in the share of undistributed corporate profits (with adjustm ent) despite the lack of im provem ent in the cyclical performance of the economy. Thus, other things have not been equal in the deter m ination of saving and investment. Tax cuts arising from accelerated depreciation have added substan tially to the private saving rate a n d m ade possible the cash flow to finance the deficit ind uce d bv such a loss in government revenue, w ithout interest rate changes. But the new incentives also in d uced a substantial rise in the share of investment in GNP, especially in the share of plant and equipm ent investment and a redi rection of investment by U.S. firms from abroad. Not surprisingly then, yields on m ost private assets rose sharply from 1980 to 1984. The changes in saving a n d investment rates from 1980 to 1984 conform more closely to the expectations of the classical view th an to those of the conventional analysis, especially w h e n the investment incentives of the 1981 tax act are taken into account. In the absence of more detailed statistical analysis, however, the data do not yield decisive evidence supporting either view to the exclusion of the other. The strength of invest m ent — arising from im proved incentives, despite nearly unchanged cyclical performance of the econ om y an d a sharply higher real rate of interest — has 14 been associated w ith a substantially smaller decline in the national saving rate an d a m u c h larger rise in the private saving rate than that suggested by the conven tional view, however. The Implications o f Federal Deficits For Crowding Out in the 1980s Table 3 shows the growth in federal expenditures as a share of GNP from 1980 to 1984 a n d unified budget estimates for 1990. The latter are constructed assum ing either no further policy changes or the im p le m e n tation of adm inistration proposals. In the absence of policy changes, expenditures are higher in each year than in 1980, resulting in an im plicit crow ding out of investment.17W hile expenditures a n d deficits peak as a share of GNP in 1983, the declines to 1990 are small .18 The e* ante crowding-out view suggests that tax changes have no effect on national saving, b u t that changes in government expenditures reduce invest m ent to the extent that such expenditures lower na tional saving.151Increases in government expenditures ,7The Congressional Budget Office (1985) discusses the effect of such deficits on the ratio of federal debt to GNP, including the view that it is the level of the debt relative to GNP rather than the deficit that affects interest rates. Their current services estimate of this ratio rises to near 50 percent of GNP in 1990, roughly its level in 1959. The view that the comparable decline in this ratio from 1959 to 1974 reduced interest rates is noticeably absent from contemporary or earlier studies. Also see footnote 3 above. 18The growth in the government budget deficit from 1980 to 1982 was cyclical in nature and would not have raised interest rates in any case. Investment demand is typically more strongly cyclical than budget deficits so that, even if the conventional view were correct, interest rates would not have risen due to cyclical deficit increases. Barro (1983) and Tatom (1984) detail the cyclical deficits since 1980. 19Since gross domestic investment equals national saving plus net foreign saving, the fiscal effects on saving must be mirrored in similar changes in investment, other things equal. FEDERAL RESERVE BANK OF ST. LOUIS have little effect on interest rates or GNP, in this view, although they do change the mix of GNP and, d e p en d ing on how they are financed, alter the mix of national saving.20 In the absence of policy changes to reduce the share of government expenditures in the n a tio n ’s output, crowding out w ill rem ain a serious concern. The ad m inistration has proposed cutting the share of federal expenditures in GNP by 1990. This proposal focuses on reductions in government purchases.21 Such a p o l icy w o uld boost capital form ation an d econom ic growth by raising private a n d national saving rates. According to the classical view, however, this may have little effect on interest rates. This view indicates that deficit reduction efforts that focus on raising taxes will have no short-term im pact on econom ic perfor mance, but w ill instead sim ply reduce private saving by a corresponding am ount.” SU M M ARY Popular analyses of recent a n d prospective U.S. gov ernm ent deficits suggest that deficits have raised o ut put, prices and interest rates a n d crowded out private investment. The im plication of this view is that future budget cuts, in the short run, w ill retard the growth of aggregate dem and but will lower interest rates, lead ing to a strengthening of private investment and longrun growth. There are reasons to question the relevance an d the “ The link between deficits and the price level depends on whether increased deficits raise aggregate demand and on the extent to which deficits are accommodated by monetary growth. The classi cal view indicates that increased deficits do not raise aggregate demand and, hence, cannot be inflationary. The second issue, however, whether deficits contribute to money stock growth and, hence, inflation, is not examined here. This link between the deficit and inflation is developed more fully in Hein (1981). See Hamburger and Zwick (1981) for an alternative view. 2'A detailed analysis of the unified budget proposals indicates that they focus on reductions in federal aid to state and local govern ments, agriculture and other purchases. These expenditures are principally either part of total government purchases directly, or they finance such purchases at the state and local government level. See Carlson (1985). “ The earlier discussion does not distinguish between the type of taxes. Thus, the effects discussed are for average relationships. One of the most important qualifications that this raises concerns business tax changes that change investment incentives. The 1981 improvements in tax incentives for investment certainly lowered taxes and raised the deficit and may, at unchanged interest rates, have left national saving unchanged, as the classical view suggests. But the increased investment demand played a major role in boost ing interest rates and thereby affected economic performance. Recent proposals to remove those incentives would reverse many of these effects on economic performance, even if the overall taxes and deficits are unchanged. OCTOBER 1985 accuracy of the conventional view. It provides an in consistent view of recent econom ic developments w ith inaccuracies ranging from the forecast of bo o m ing output, em ploym ent a n d inflation for 1981 and beyond, to the forecast of rising interest rates. In addition, the evidence here shows that the expected crowding out of investment has been offset by other factors, resulting in an investment boom since 1980. The alternative hypotheses examined here indicate that fiscal policy actions are largely an d directly offset by the private sector. Thus, tax changes are offset by adjustm ents to private saving, w ith n o direct effect on national saving or investment. This classical view of fiscal policy also em phasizes that increased govern m ent purchases are directly offset by reduced private expenditures (especially investment). According to the classical view, policy actions to reduce the deficit are not likely to affect interest rates an d m ay not affect the investment boom . For example, if deficit reduction entails sim ply raising taxes, private saving will fall by a like a m o un t a n d no additional investment will occur. To the extent that deficit reduc tion focuses on expenditures, however, investment will be strengthened, but w ithout the inducem ent of lower interest rates. The evidence from the recent experience suggests that the classical view is correct a n d indicates the im portance of business tax cuts in raising domestic saving, investment an d interest rates and reducing U.S. investment abroad. The evidence is not decisive as to w hich view more generally an d accurately depicts the effects of fiscal policy on the economy, however. But both views indicate that domestic investment and econom ic growth are im p ed e d by deficits arising from government expenditure growth, a n d that they are ultim ately im proved by restraint in such growth. Pro posals to deal w ith the deficit w ithout raising taxes focus largely on reducing government nondefense purchases. The successful im plem entation of these plans w o uld ultim ately raise private saving an d invest ment, alter the com position of national o utp ut and prom ote econom ic growth. REFERENCES Aschauer, David Alan. “ Fiscal Policy and Aggregate Demand,” American Economic Review (March 1985), pp. 117-27. Bailey, Martin J. National Income and the Price Level, 2nd ed. (McGraw-Hill, 1971). Barro, Robert J. “ Are Government Bonds Net Wealth?" Journal of Political Economy (November/December 1974), pp. 1095-117. 15 FEDERAL RESERVE BANK OF ST. LOUIS OCTOBER 1985 ________ _ “ On the Determination of the Public Debt,” Journal of Political Economy (October 1979), pp. 940-71. Friedman, Milton. A Theory of the Consumption Function (Princeton University Press, 1957). ________ . “ Output Effects of Government Purchases,” Journal of Political Economy (December 1981), pp. 1086-121. Hafer, R. W. “The Role of Fiscal Policy in the St. Louis Equation,” this Review (January 1982), pp. 17-22. ________ _ “ The Behavior of U.S. Deficits," University of Chicago (March 1983), unpublished. Hall, Robert E. "Labor Supply and Aggregate Fluctuations," Carnegie-Rochester Conference Series on Public Policy (Spring 1980), pp. 7-33. Brunner, Karl. “ Fiscal Policy in Macro Theory: A Survey and Evalu ation," in The Monetary vs. Fiscal Policy Debate, conference pro ceedings, October 12-13, 1984, Federal Reserve Bank of St. Louis, forthcoming. Hamburger, Michael J., and Burton Zwick. “ Deficits, Money and Inflation,” Journal of Monetary Economics (January 1981), pp. 141-50. Carlson, Keith M. “ Controlling Federal Outlays: Trends and Pro posals,” this Review (June/July 1985), pp. 5-11. Hein, Scott E. 3-10. ________ . “The Mix of Monetary and Fiscal Policies: Conventional Wisdom vs. Empirical Realities,” this Review (October 1982), pp. 7-21. Hickman, Bert G. "Investment Demand in the Sixties,” Reprint No. 120, The Brookings Institution, 1966. “ Deficits and Inflation,” this Review (March 1981), pp. Carlson, Keith M „ and Roger W. Spencer. “Crowding Out and Its Critics,” this Review (December 1975), pp. 2-17. Kochin, Levis. “ Are Future Taxes Discounted by Consumers? A Comment,” Journal of Money, Credit and Banking (August 1974), pp. 385-94. Congressional Budget Office. The Economic and Budget Outlook: Fiscal Years 1986-90,1985 Annual Report, Congress of the United States (GPO, February 1985). Kormendi, Roger C. “ Government Debt, Government Spending, and Private Sector Behavior,” American Economic Review (De cember 1983), pp. 994-1010. Council of Economic Advisers. (GPO, February 1985). Economic Report of the President David, Paul A., and John L. Scadding. “ Private Savings: Ultrara tionality, Aggregation, and Denison’s Law’,” Journal of Political Economy (March/April 1974), pp. 225-49. Denison, Edward F. “ A Note on Private Saving,” The Review of Economics and Statistics (August 1958), pp. 261-67. Dolan, Edwin G. Basic Economics, 3rd ed. (CBS College Publish ing, The Dryden Press, 1983). Evans, Paul. "Do Large Deficits Produce High Interest Rates?” American Economic Review (March 1985), pp. 68-87. Feldstein, Martin. “ Government Deficits and Aggregate Demand,” Journal of Monetary Economics (January 1982), pp. 1-20. Feldstein, Martin and Otto Eckstein. “The Fundamental Determi nants of the Interest Rate," Review of Economics and Statistics (November 1970), pp. 363-75. 16 McConnell, Campbell R. Economics, 9th ed. (McGraw Hill, 1984). Office of Management and Budget. Government 1986 (GPO, 1985). Budget of the United States Plosser, Charles I. “ Government Financing Decisions and Asset Returns, " Journal of Monetary Economics (May 1982), pp. 325-52. Samuelson, Paul E., and William D. Nordhaus. (McGraw Hill, 1985). Economics, 12th ed. Sargent, Thomas J. “The Fundamental Determinants of the Interest Rate: A Comment,” Review of Economics and Statistics (August 1973), pp. 391-93. Tanner, J. Ernest. “An Empirical Investigation of Tax Discounting,” Journal of Money, Credit and Banking (May 1979), pp. 214-18. Tatom, John A. “A Perspective on the Federal Deficit Problem,” this Review (June/July 1984), pp. 5-17. U.S. Department of the Treasury. The Effects of Deficits on Prices of Financial Assets: Theory and Evidence (GPO, March 1984). The Status of Farm Lenders: An Assessment of Eighth District and National Trends Michael T. Belongia and Kenneth C. Carraro c L-7IN C E 1982, sharp declines in real farm incom e a n d the asset values supporting over $200 billion in farm debt have created substantial increases in farm loan defaults a n d farm bank failures. Large an d in creasing loan losses have generated a great deal of concern that the rapidly deteriorating quality o f farm debt m ay have severe consequences for the long-term structure of American agriculture and adverse short term effects o n the aggregate econom y as well.' This article reviews a variety of performance in dica tors for the three m ajor lenders to the farm sector and assesses both the tim ing an d breadth of portfolio deterioration. The lending institutions exam ined are agricultural banks, the Farm Credit System IFCS) and the Farmers Home A dm inistration IFmHA). The perMichael T. Belongia is a senior economist and Kenneth C. Carraro is an economist at the Federal Resen/e Bank of St. Louis. James C. Poletti provided research assistance. The following individuals and their insti tutions are gratefully acknowledged for their assistance in providing data used in this article: Kenneth Obrecht of the Farm Credit Banks of St. Louis, Lyle Stucki of the Farm Credit Banks of Louisville, David Meads of the FmHA, and Arnold Miller of the Farm Credit Admin istration. ’ Estimates of farm loan defaults under different scenarios are pro vided by Bullock (1985). Ranges commonly cited for losses on all farm loans are $25-50 billion over the next four years. Losses within the Farm Credit System alone have been estimated to be $350-400 million in 1985 with additional losses in 1986. Also, an estimated 12 percent of the Farm Credit System’s loans are not adequately secured by property and assets and could produce "significant” future losses; see “ Farm Agency Estimates” (1985). Schink and Urbanchuck (1985) describe the channels through which farm loan defaults could affect interest rates, GNP and employment and estimate these effects for different magnitudes of loan losses. An alternative assessment of how farm loan defaults might affect the aggregate economy is provided by Belongia and Gilbert (1985). formance of these lenders in the Eighth Federal Re serve District is com pared w ith their performance in the rest of the U nited States. FA RM L E N D E R S : A B R IE F P R O F IL E An agricultural bank is defined as a commercial bank w ith a ratio of farm loans to total loans that is above the average farm loan ratio at all banks. At the end of 1984, the average farm loan ratio was approxi mately 17 percent. Currently, there are 589 banks in the “official” boundaries of the Eighth Federal Reserve District an d 1,383 agricultural banks in the region defined more broadly that have a higher farm loan ratio and meet the current definition of an agricultural bank. The broader definition of the Eighth District is used to make com parisons w ith Farm Credit Bank districts.- Figure 1 represents these alternative desig nations of the District’s borders. Nationally, 4,970 banks, or 35 percent of all com m er cial banks are defined to be agricultural banks. Collec tively, they h old $30 billion, or 60 percent, of the total farm debt held bv com m ercial banks. Of the farm debt held by agricultural banks, 83 percent ($24.8 billion) is O fficially, the Eighth Federal Reserve District includes Arkansas, northern Mississippi, southern Illinois and Indiana, western Ken tucky and Tennessee and eastern Missouri. Farm Credit districts, however, cover entire states. The Fourth and Sixth Farm Credit districts, headquartered in Louisville and St. Louis, do not include Mississippi but cover all of the remaining states and Ohio. To provide a representative, but unofficial Eighth District, we define it as the states covered by the Fourth and Sixth Farm Credit districts: Arkansas, Illinois, Indiana, Kentucky, Missouri, Ohio and Tennes see. Comparisons with the nation refer to data for all states other than those included in this definition of the Eighth District. 17 FEDERAL RESERVE BANK OF ST. LOUIS FICBs function as intermediaries that package these loanable funds for, as of October 1985, 318 Production Credit Associations (PCAs), w h o in turn, make loans directly to farmers for annual operating expenses. The FLBs make loans to farmers for the purchase of farm land through a network of 390 Federal Land Bank Associations that function as loan originating offices. Banks for Cooperatives make loans to farmer-owned cooperatives, such as supply stores. As of December 31, 1984, the Farm Credit System, exclusive of the Banks for Cooperatives, held $67.9 billion, or 32 per cent, of total farm debt. O f this total, FLBs held $49.1 billion an d PCAs held $17.9 billion. FICBs held the rem aining $0.9 billion in the form of loans to other financial institutions. F ig u re 1 KY MS □ □ OCTOBER 1985 O ffic ia l Eighth D is tric t The a r e a used to re p re s e n t the Eighth The Farmers Home A dm inistration (FmHA) is the so-called "lender of last resort” to farmers. It extends credit to farmers through direct loans, guarantees of farm loans m ade an d seiviced by com m ercial banks, and various emergency loan programs. FmHA, for the most part, lends to farmers w h e n they have trouble servicing debt acquired from other lenders or if credit is not available at "reasonable” interest rates from their current lenders. As of 1984, FmHA held $25.7 billion, or 12 percent, of total farm debt. D istrict fo r this a rtic le in c lu d es th e 4th a n d 6th Farm C r e d it D istricts o f Louisvil an d St. Louis an d is b o u n d e d by the h e a v y b la c k lin e . non-real estate debt, or operating debt, associated w ith the variable costs of farm production. Because farm lending by com m ercial banks is prim arily for short-term operating debt, their chief source of co m petition is the Production Credit Associations (PCAs) of the Farm Credit System. The cooperative Farm Credit System (FCS) is a sys tem of federally chartered, but privately owned, banks and associations, w h ich are organized as coopera tives. These banks are supervised a n d exam ined by the Farm Credit Adm inistration, an independent agency of the United States government, and are m andated by their charter to make loans only for purposes directly related to agriculture. The FCS consists of 12 districts and 37 banks: 12 Federal Land Banks (FLBs), 12 Federal Intermediate Credit Banks (FICBs) a n d 13 Banks for Cooperatives (BCs). The FCS obtains loanable funds bv the sale of securi ties through the system’s W all Street fund in g arm, the Federal Farm Credit Banks F unding Corporation. The Digitized for 18FRASER T R E N D S IN T H E A LLO C A T IO N O F FA RM D E B T A convenient place to begin a review of farm debt holdings an d problem s is an analysis of trends in loans outstanding at the various lenders. Table 1 presents the market shares of non-real estate agricul tural debt held by the m ajor lenders since 1970 for both the Eighth District, broadly defined, a n d for the rem ainder of the U.S. Non-real estate debt represents financing for an n ual operating expenses such as feed, fertilizer an d seed, as well as for the purchase of farm m achinery an d livestock. The categoiy of "All Others" includes such lenders as private individuals, dealei'S an d merchants. The trends in the District an d the United States are roughly parallel a n d indicate that both commercial banks an d the FCS gained their highest market shares in the mid-1970s a n d u n til recently have been steadily losing market share to the FmHA. By the end of 1984, commercial banks at both the District a n d national levels reversed the 10-vear dow ntrend, show ing sig nificant market share gains over 1983. Table 2 presents the market shares held by lendei-s for farm real estate loans. The lender categoiy of " In surers” has been added to reflect the significant pres- OCTOBER 1985 FEDERAL RESERVE BANK OF ST. LOUIS Table 1 Farm Non-Real-Estate Debt Outstanding at Major Lenders (percent of market held by each lender) Banks PCAs FmHA U.S. District U.S. 43.1% 19.1% 22.5% 3.5% 45.7 22.2 25.6 3.5 46.2 43.4 22.3 26.2 1973 48.6 46.2 21.9 1974 51.7 46.8 1975 50.0 1976 A ll Others U.S. District 2.6% 34.0% 31.8% 2.6 28.3 26.1 3.0 2.2 28.5 28.2 28.0 2.8 2.1 26.8 23.8 23.0 28.5 2.7 2.0 22.5 22.6 46.4 25.6 30.8 2.9 2.4 21.5 20.3 49.1 44.0 25.1 31.6 4.3 4.0 21.5 20.4 1977 48.7 44.1 24.3 31.5 4.0 3.3 23.0 21.1 1978 43.8 41.3 21.6 29.6 5.7 3.6 28.9 25.5 1979 40.6 41.3 20.3 29.6 9.2 4.6 29.8 24.5 1980 38.5 39.1 21.7 29.8 12.0 7.5 27.8 23.6 U.S. District 1970 43.4% 1971 46.1 1972 District 1981 36.6 36.0 22.5 28.3 14.4 10.3 26.5 25.4 1982 34.4 33.7 21.5 28.1 15.5 13.3 28.6 24.9 1983 33.9 33.7 19.5 20.2 13.9 13.5 32.8 32.6 1984 37.5 39.7 19.1 18.4 14.0 15.0 29.3 27.0 NOTE: Due to rounding, percentages may not add to 100. SOURCE: U.S. Department of Agriculture. Table 2 Farm Real Estate Debt Outstanding at Major Lenders (percent of market held by each lender) Banks FLBs U.S. District U.S. District FmHA U.S. District Insurers U.S. All Others District U.S. District 1970 10.3% 18.9% 23.1% 21.9% 7.9% 7.4% 19.7% 19.4% 39.0% 32.4% 1971 10.5 19.5 23.9 22.1 8.1 7.7 18.6 18.1 38.8 32.7 1972 11.1 20.6 25.0 22.5 8.2 7.8 17.5 16.6 38.2 32.5 1973 11.5 21.6 26.5 23.3 8.2 7.5 16.3 15.1 37.5 32.5 1974 11.5 22.4 28.4 24.6 7.8 7.1 15.5 13.7 36.9 32.3 1975 11.0 22.0 30.7 27.6 7.3 6.8 14.7 12.0 36.3 31.6 1976 10.4 21.1 32.7 30.1 6.9 6.4 14.3 11.0 35.7 31.4 30.6 1977 9.9 20.9 34.0 31.6 6.8 6.1 14.1 10.8 35.2 1978 9.8 20.8 34.2 32.3 6.5 5.6 14.6 11.5 34.8 29.7 1979 9.6 19.9 34.8 33.3 6.0 5.1 15.2 12.8 34.3 28.9 1980 8.2 16.7 34.9 33.9 8.5 7.8 14.6 13.0 33.8 28.6 1981 7.5 14.9 37.9 36.6 8.2 7.6 13.7 12.9 32.6 28.0 1982 6.4 13.3 41.7 39.8 8.4 8.0 12.5 12.0 31.0 26.9 1983 6.3 12.6 43.7 41.0 8.3 8.4 11.8 11.5 30.0 26.5 1984 7.0 13.3 43.7 40.2 8.2 8.9 11.5 11.0 29.6 26.6 NOTE: Due to rounding, percentages may not add to 100. SOURCE: U.S. Department of Agriculture. 19 FEDERAL RESERVE BANK OF ST. LOUIS ence of a n um b er of insurance com panies in farm land lending. The “All Others” category for real estate le n d ing m ainlv represents debt held bv individuals. The trends are again consistent across the District and the United States. A lthough com m ercial banks in the Dis trict h old a larger share of the farm real estate debt than banks in the rem ainder of the U.S., banks in both areas have seen steady declines in market share after initial gains in the early 1970s. Insurance com panies and the category of “All Others” also have exhibited secular declines in market share in both areas. The share losses of these three lender groups have accrued almost entirely to the Federal Land Banks of the FCS. At both the District a n d U.S. levels, the FCS has nearly doubled its market share w ith steady growth over the period since 1970. The FmHA share, however, has rem ained largely unch ange d over the same period, although m in o r gains are evident in the District. In summary, the market share data indicate that, for farm operating debt, the FmHA has posted sharp gains since the mid-1970s at the expense of commercial banks an d the PCAs. The farm real estate market, however, has been dom inated by the sharp gains m ade by the Federal Land Banks relative to the share losses of most other m ajor farm lenders. M E A SU R E S O F P O R T F O L IO QUALITY The m ajor causes of the recent farm debt defaults are erroneous forecasts — both by farmers and their creditors — of continue d high an d accelerating in flation a n d increased real returns to assets em ployed in agriculture:1So long as land prices continued to rise w ith inflation, the collateral base against w hich farmers could borrow increased, an d the value of dollars used to repay the debt decreased. In c o n ju n c tion w ith tax advantages for land ow nership an d the availability of subsidized credit for land purchases, it m ade sense to buy farm land at prices of $3,000-S4,000 per acre — i f the purchaser believed the land could be resold at a higher price. Similarly, u n d e r the expecta tion of w orld food shortages an d increases in real com m odity prices, the price of land in agricultural production w o uld be expected to rise.1 Under these conditions, both farmers an d their lenders agreed that extending more credit on a rising n om inal asset base was a p rudent business decision. Unfortunately for both parties, however, their forecasts of inflation and com m odity prices were seriously in error. 3See Belongia (1985). 4See, for example, Will There Be Enough Food? (1981). Digitized for 20FRASER OCTOBER 1985 This description of events suggests that institutions w ho increased their lending to agriculture sharply between 1974-81 — w hen inflation, foreign dem and for U.S. farm products an d real com m odity prices were increasing or were expected to increase sharply — should be experiencing the greatest deterioration in portfolio quality."' O n the basis of this criterion, the Farm Credit Sys tem and FmHA should be experiencing relatively more trouble w ith portfolio performance th an other farm lenders. To assess this thesis, we n o w turn to a dis cussion of measures of loan quality an d portfolio performance. Portfolio Quality at FLBs and PCAs A com m on measure of loan quality is the percent age of loans on w hich payments are delinquent. This percentage tends to be a leading indicator of ultim ate loan losses because borrowers w ho eventually default on debt fii'st experience problem s w ith m aking their scheduled pavments. If efforts to reschedule the loan and to service only its interest obligation fail, the delinquent loan becomes, after some lag, a loan loss. The data required for this analysis are difficult to obtain an d are not entirely com parable across differ ent lenders an d even across different geographical areas for the same lender group. The shaded insert discusses the data used in this article a n d some cave ats that should be exercised w h e n m aking com pari sons or draw ing inferences from these series. Chart 1 plots loan delinquency rates for FLBs in the U.S. an d the Eighth District, broadly defined; chart 2 plots the loan loss series for FLBs. In each case, these series are defined to be the dollar value of loans on w hich pavm ents are delinquent or the dollar value of loan losses as a percent of total loans outstanding. The FLB series indicate that these institutions have experienced sim ilar patterns an d rates of loan d e lin quencies an d losses both in the District an d in the rem ainder of the U.S. Loan losses at District FLBs, 5Market share data as a proxy for loan quality should be applied with some caution. Moreover, it should be noted that this measure is better suited to long-term land mortgages than to short-term operat ing loans. The reasoning is as follows: If market share for mortgage lending declined in the 1970s, subsequent portfolio quality might be improved because fewer new loans (that turned out to be poor loans) were extended and the old loans carried forward were of higher quality if, for no other reason, because a larger share of the principal had been repaid. For annual operating loans, however, market share in a given year may be unrelated to loan quality. In fact, the lower interest rates offered by PCAs in much of the 1970s may have attracted the more creditworthy farmers. OCTOBER 1985 FEDERAL RESERVE BANK OF ST. LOUIS The Data Agricultural Banks All FDIC-insured com m ercial banks are re quired to file a quarterly Report of Incom e and Condition. These reports are com m only referred to as "call reports” a n d are roughly equivalent to a bank’s balance sheet an d incom e statement. Two loan item categories identify the volum e of agricultural production loans and farm real es tate loans outstanding on the reporting date. Banks have been required to report loan delin quency data only since 1982. Bank loans are con sidered past due w hen interest or principal pay ments are more than 30 days delinquent. These data are generally verified in the course of norm al inspections bv bank regulatory agencies such as the FDIC, Federal Reserve or Com ptroller of the Currency. Farm Credit Svstem %/ Most data used in this article are derived from statistics published in the Farm Credit A dm inis tration’s an n ual reports. In m any cases, the Farm Credit Banks of St. Louis an d Louisville cooper ated to provide data not available elsewhere. Un til 1984, no guidelines were available to assure that reporting standards for loan delinquency inform ation were consistently applied across the 12 Farm Credit Districts. This fact introduces an inconsistency in the data because 1984 data were collected using definitions that were possibly different from those used in earlier periods. Farmers H om e Administration Data on the FmHA were derived from the FmHA report 616 an d include only loans m ade un d e r the following farm programs: Farm Ow ner ship, Farm Operating, Econom ic Emergency and Disaster Emergency. FmHA judges a loan as de linquent w hen principal or interest payments are 15 days or more past due. Percent of Total Loans D elinquent Loan Losses as a Percent o f Total Loans federal Land B a nks Federal Land Banks P e rc e it .4 0 P e r c e il .4 0 r # i # .35 « .35 i i i i i .3 0 .3 0 i i i i i i .25 .25 i i i .20 .20 1 1 1 1 1 I .15 1 D i s t r ic t / « i .10 .10 i i .0 5 # i i # i .0 5 / j ^ / / , / / U.S. .00 .00 1981 1 98 2 1983 1984 21 FEDERAL RESERVE BANK OF ST. LOUIS OCTOBER 1985 C h a rt 3 Percent of Total Loans D elinquent Loan Losses as a Percent of Total Loans Eighth District Production C redit A sso cia tio n s Production Credit A sso cia tio n s Percent 1970 72 74 76 78 tO »2 1984 however, have risen slightly more sharply than at FLBs in the rem ainder of the U.S.'1Moreover, there appears to be a lag of about two years in each case between the time delinquencies rise sharply (1982) a n d later reveal themselves in h igherloan losses (1984). Note, however, that while the patterns of delinquency an d loan loss rates have been similar, loan losses have been about one-tenth of prior years’ delinquencies. Chart 1 also shows that delinquency rates were nearly constant between 1970-81. Prior to 1981, losses at FLBs were less than two-tenths of one percent of all loans o u t standing. PCA loan delinquency data are not available on a consistent basis for both the District and the rem ain der of the U.S. For this reason, only the District delin quency data are show n in chart 3. They also reveal a dramatic increase in delinquency rates beginning in 6U.S. and District FLB delinquency rates are derived from Farm Credit Administration annual reports and include the items of “ non accrual loans" and "delinquent principal and advances.” Digitized for 22FRASER 1982. It m ust be pointed out that the absolute levels of delinquency rates for PCAs are not com parable w ith the FLB rates portrayed in chart 1. Loan loss data for PCAs, however, are available for both the District an d the U.S. a n d are presented in chart 4. Write-offs in 1984 as a percent of total loans were eight times higher than the percentage in 1981. In contrast to FLB loans, however, there appears to be almost no lag between the time these delinquencies are reported and the tim e they result in loan losses. The likely reason for this difference is that PCAs make 'District PCA delinquency rates were obtained directly from the Farm Credit Banks of Louisville and St. Louis on the following basis: for 1970 and 1971 the information includes loans 30 days past due; from 1972 to 1983 the information is for loans 60 days past due; the 1984 data are calculated according to the new FCA standards for non performing loans. The national data available for the 1970 to 1983 period include loans which are termed “ loans in process of liquida tion” and which are not comparable to the District data. In spite of the different nature of the two series, the data exhibit very similar behavior when plotted against each other on the basis of annual percentage changes rather than as absolute levels of delinquency. FEDERAL RESERVE BANK OF ST. LOUIS OCTOBER 1985 C h a rt 5 C h a rt 6 Return on Equity Return on Equity Federal Land Banks P rod uction Credit A sso c ia tio n s Percent P trciat loans for an individual year's operating expenses and usually schedule repayment shortly after the year’s harvest. For this reason, unlike the FLBs' multi-vear loans for lan d purchases, PCAs tend to exhibit a closer short-run relationship between delinquencies and losses. As in the case of FLBs, it is necessaiy to note that, w hile the patterns of delinquency an d loan loss rates are similar, losses have been one-fifth of delinquencies. W ith rising rates of d e lin q u e n c ie s a n d loan losses one also w o uld expect the returns to equity and assets held by these lenders to decline. Charts 5 and 6 plot the returns to equity for FLBs an d PCAs, respectively, in the U.S. a n d the District. In chart 5, similar patterns for returns in the U.S. and the Dis trict are revealed w ith the District show ing lower average returns since 1979 a n d a sharper decline since the 1982 peak of 10.4 percent. Returns to equity for PCAs (chart 6) peaked in 1980 for the U.S. and 1981 for the District a n d have fallen sharply in just two years. Returns to assets have followed sim ilar patterns for these lenders at both District and national levels." Agricultural Banks The conjecture was that agricultural banks, w hich d id not increase market share or dollar volum e of farm loans as aggressively in the 1970s, w o uld show som e what lower measures of loan delinquencies an d losses an d better returns to assets an d equity than members of the Farm Credit System. Another im portant factor supporting this expectation is the fact that the loan portfolios of agricultural banks are diversified outside of agricultural lending. This loan diversity could help protect bank earnings from the wide swings of returns on equity experienced bv the FCS lenders w ho extend credit only for purposes directly related to agriculture. 8Returns to assets for FLBs peaked in 1982 at 1.12 percent nationally and .99 percent in the District. By 1984, these values had declined to .42 and -.0 3 percent, respectively. Between 1981 and 1984, returns to assets at District PCAs fell from 1.07 to 0.05 percent. 23 FEDERAL RESERVE BANK OF ST. LOUIS OCTOBER 1985 C h a rt 8 C h a rt 7 A gricultural Loans Past Due as a Percent of Production Loans Total L o a n Losses a s a Percent o f Total L o a n s Agricultural and Honagricultural Banks P e rc e n t A gric u ltu ra l B a nks Percent 4 .2 5 P e rc e n t Pe rce nt 4 .0 0 3 .7 5 3 .5 0 3 .2 5 3 .0 0 2 .7 5 2 .5 0 2 .2 5 1982 1 98 3 1984 Agricultural loan delinquency data have been col lected from com m ercial banks since 1982. O nly banks w ith total assets greater than $100 m illion, however, are required to report the volum es of agricultural loans considered in nonaccrual or renegotiated sta tus. These two categories include loans on w hich interest payments arc not being paid or are being paid more slowly than originally established. Since a large majority of agricultural banks are smaller than $100 m illion an d therefore do not report nonaccrual and renegotiated agricultural loans, chart 7 plots only the percentage of agricultural production loans that are considered past due by 30 days or more at agricultural banks.3These data, therefore, are not directly c o m p a rable w ith the PCA data sum m arized earlier in chart 3. Nonetheless, these lim ited data suggest that agricul tural banks also have experienced rapid increases in 9The set of banks defined as agricultural banks will change over time as the shares of agricultural loans in some banks’ portfolios become less than or greater than the cutoff point. Over time, however, the number of agricultural banks has remained fairly constant, ranging between 5,668 in 1974 and 4,970 in 1984. Digitized for 24FRASER delinquent farm debt both in the District an d in the rem ainder of the U.S.1" Loan loss data for agricultural banks provide a d d i tional inform ation to supplem ent that provided bv past due rates. Because loan loss data are not available specifically for agricultural loans, chart 8 is a plot of all loan losses at agricultural banks an d at small nonagricultural banks in the nation (less than $100 m illion in total assets) expressed as a percentage of all loans at these banks since 1976. It indicates that loan losses at agricultural banks have been increasing steadily since 1979. The rate of increase has been such that the percent of loan losses has risen by a factor of nearly seven since 1979. Losses at com parably sized nonagricultural banks were larger than losses at agricultural banks until 1981. Now the rate of losses at nonagricul- ,0For the small number of agricultural banks reporting all agricultural loan delinquency items (29 banks in the District for 1984 and 75 in the remainder of the U.S.), the delinquency rate rose from 4.2 percent in 1982 to 9.0 percent in 1984 for the District. In the remainder of the U.S., however, the overall delinquency rate rose only from 6.1 percent in 1982 to 6.3 percent in 1984. Given the small number of agricultural banks reporting these data, caution in their interpretation must be used. FEDERAL RESERVE BANK OF ST. LOUIS OCTOBER 1985 C h o rt ♦ Return on Equity Agricultural and N onagricultural Banks Percent P e rc e it loss reserves." The latter two items represent the re sources of a bank to absorb loan losses. At the national level, the num b e r of agricultural banks that fall into the threatened bank categoiy has nearly tripled from 1982 to 1984, going from 76 to 202. In the District, however, the n um b e r of such threatened institutions has doubled from 23 to 46 over the same period. Thus, concern about rapid increases in farm bank failures, although certainly im portant in the Eighth District, appears to be even more im portant for institutions beyond the borders of the Eighth District. In sum m aiy, agricultural banks have suffered rising rates of delinquency and loan losses an d declining profitability. W hen com pared w ith PCAs, w hich repre sent their most significant com petitors in the agricul tural lending arena, however, banks appear to have suivived recent dow nturns in the agricultural econ om y in m uch better fashion. Although loan loss rates for PCAs and agricultural banks are comparable, d e lin quency rates have increased, an d profitability de creased more quickly at PCAs than at agricultural banks. The Farmers Hom e Administration tural banks is only half that experienced by agricul tural banks. Based on returns to equity data at agricultural banks an d at small nonagricultural banks of the District (chart 91, profitability in the District peaked in 1974 an d has been falling steadily since then, with the exceptions of 1979 an d 1980. At the national level, agricultural bank profitability peaked in 1980 before starting a sharp decline. As in the case of loan losses, agricultural banks’ declining profitability was greater than that experienced bv nonagricultural banks. Re turns to equity at District agricultural banks have fallen by nearly 34 percent since their recent peak in 1979, while the returns to equity ratio for nonagricul tural banks in the District has fallen by only 16 percent since 1979. One further means of assessing the viability of agri cultural banks is by com paring the volum e of “risky" loans for w hich repayments are uncertain w ith a bank's ability to absorb the potential loss. W ith this in m ind, banks whose viability may be threatened can be defined as those for w h ich the volum e of delinquent loans exceeds the sum of total bank capital an d loan The FmHA’s role as “lender of last resort" dictates that its borrowers are from a high-risk category. Loan delinquency data for both real estate a n d non-realestate farm loans bear this out. Charts 10 an d 11 docum ent the steady rise in delinquency rates for both loan categories. Com parisons of delinquency rates at FCS lenders (charts 1 an d 3) w ith those of the FmHA are instructive. The FmHA appears to have experienced rising delinquency rates earlier than 1981 w hen the FCS lenders began to show marked in creases in delinquencies. This finding is to be ex pected given the character of the FmHA’s borrower clientele. FmHA borrowers w o u ld be more likely to exhibit repayment problems w hen a dow nturn in the agricultural econom y occurs than w o uld the more creditworthy borrowers of the FCS or of agricultural banks. This also highlights an im portant aspect of the "The FDIC compiles the official list of “problem banks” by rating all insured banks on the basis of five categories: capital, assets, management, earnings and liquidity. Banks receiving a rating of four or five on a scale from one to five are placed on the problem bank list. Our definition, which focuses on capital and asset quality, is likely to provide a parallel indicator of banks threatened by bank ruptcy should a large share of delinquent loans become loan losses. ,2Melichar and Inwin have reported that one-half of potentially vulnera ble agricultural banks are located in five states: Iowa, Nebraska, Kansas, Minnesota and Missouri; only Missouri is in the Eighth District. 25 FEDERAL RESERVE BANK OF ST. LOUIS OCTOBER 1985 C h a rt 10 C h a rt 11 Percent of Total Farm O p eratin g Loans D elinquent Percent of Total Farm O w n e rs h ip Loans D elinquent Farm ers Hom e A d m in istratio n Pe rc e n t 6 0 , -------- Farm ers Hom e A d m inistratio n Pe rc e n t ,6 0 FmHA — as a lender of last resort, the FmHA provides an informal subsidy to the FCS an d agricultural banks through its direct lending programs. Bv lending in some cases to farmers w ho h ad received FCS or bank financing but w ho are no longer considered credit worthy, the FmHA allows these lenders to delay fore closure and to continue to receive loan payments from such borrowers. Moreover, u n d e r the Econom ic Emergency Credit Act, FmHA refinanced loans origi nally m ade by the FCS and com m ercial banks and repaid the original lenders from proceeds of the FmHA loan. Given the extremely high an d rapidly growing delin quency rates on FmHA loans, one w o uld expect, other things equal, to find com m ensuratelv higher levels of loan losses. Loan loss data for FmHA farm loan pro grams are available only on a consolidated basis (i.e., farm ow nership and operating loan losses are not segregated). Chart 12, however, shows low, although rising, rates of loan write-offs. For example, the 1984 delinquency rate on FmHA farm ow nership loans was near 25 percent, but only 0.22 percent of all FmHA Digitized for26 FRASER P e rc e n t 3 5 . -------- P e rc e n t 35 loans were charged off. This contrasts w ith the FLB s 1984 delinquency rate of 3 percent an d loan chargeoffs of 1.5 percent. This discrepancy between institu tions can be explained by the greater degree of for bearance that the FmHA has exhibited w ith respect to its delinquent borrowers. As evidence of this forbear ance, data on the length of tim e that loans are carried in the d elinquent status can be examined. W hile not available on a District scale, national FmHA data in d i cate that, as of Ju n e 30, 1985, more th an 45 percent of the volum e of delinquent FmHA farm loans has been in that status for more than four years. Only 9 percent of the delinquencies nationw ide were less than one year past due. C O N C LU SIO N S The Farm Credit System an d agricultural banks did not show substantial deterioration of loan portfolio quality un til 1982. The FmHA, however, began to ex hibit rising delinquency rates in 1979. The deteriora tion has been more p ro n o un ce d am ong those lenders FEDERAL RESERVE BANK OF ST. LOUIS OCTOBER 1985 Bullock, J. Bruce. “ Farm Credit Situation: Implications for Agricul tural Policy,” FAPRI #4-85, Food and Agricultural Policy Research Institute (February 1985). C h a r i 12 Farm Loan W rite-offs as a Percent of Total Farm Loans “ Farm Agency Estimates Loss.” 1985. Farm ers Hom e A d m inistratio n New York Times, September 17, Melichar, Emanuel, and George D. Irwin. “ Condition of Rural Finan cial Intermediaries,” paper presented at the meetings of the Ameri can Agricultural Economics Association, Ames, Iowa, August 6, 1985. Schink, George R., and John M. Urbanchuck. “ Economy-Wide Impacts of Agricultural Sector Loan Losses," Wharton Economet ric Forecasting Associates: Philadelphia, Pennsylvania (July 1985). U.S. Department of Agriculture. " Will There Be Enough Food?" The 1981 Yearbook of Agriculture (U.S. Government Printing Office, 1981). 1980 81 82 83 84 1 985 w ho aggressively expanded their lending to agricul ture in the 1970s. In particular, the Federal Land Banks of the Farm Credit System expanded their market share during the 1970s an d experienced some of the sharpest declines in portfolio quality in recent years. The m andate of the FCS to lend m ainly to agricultural interests inhibits its ability to diversify its portfolio and raises the risks associated w ith concentrated lending to one sector. Hence, sharper declines in overall port folio quality for branches of the FCS, relative to agricul tural banks, took place. Finally, the data revealed little difference in the measures of portfolio quality or insti tutional earnings at the Eighth District and national levels. REFERENCES Belongia, Michael T. “ Factors Behind the Rise and Fall of Farmland Prices: A Preliminary Assessment," this Review (August/Septem ber 1985), pp. 18-24. ________ and R. Alton Gilbert. “ Farm-Bank Fallout? Recall the 20s," The Wall Street Journal, September 25, 1985. 27