The full text on this page is automatically extracted from the file linked above and may contain errors and inconsistencies.
Improving housing finance in an inflationary environment: alternative residential mortgage instruments CONTENTS R ecen t increases in the leve l and volatility o f m arket interest rates have made fixed-rate, fixed-paym ent m o rt gage contracts less d esirable than they used to be. A lternative m ortgage instrum ents have b een d esig n ed to m eet the cu rren t needs o f residential m ortgage lend ers and b o rro w ers. This article d escrib es the alternative m ort gage plans and analyzes their advan tages and disadvantages. E C O N O M IC PERSPECTIVES July/August 1981, Volum e V , Issue 4 Economic Perspectives is published bim onthly by the Research D epartm ent of the Federal Reserve Bank of C h icag o . The p ub licatio n is produced under the d irectio n of H arvey R osen blu m , Vice P resident, and is edited by Larry R. M ote, Assistant V ice President, w ith the assistance of Sandra C o w e n (e d ito ria l), Roger Thryseliu s (artw o rk and graph ics), and Nancy Ahlstrom (typesetting). The view s expressed in Economic Perspectives are the au th ors’ and do not necessarily reflect the view s of the m anagem ent of the Federal Reserve Bank of Chicago or the Federal R eserve System. Single-copy sub scrip tion s of Economic Perspectives are available free of charg e. Please send requests fo r single- and m ultip le-co p y sub scriptions, back issues, and address changes to Public Info rm atio n C e n te r, Federal R eserve Bank of C h icago , P .O . Box 834, C h icag o , Illinois 60690, or telep h o ne (312) 322-5112. A rticles may be rep rinted provided source is credited and Public Inform ation C e n ter is provided w ith a copy of the published m aterial. C o ntrolled circu latio n postage paid at C h icago , Illino is. Improving housing finance in an inflationary environment: alternative residential mortgage instruments G eo rg e G. Kaufman and Eleanor Erdevig A lternative mortgage instrum ents are m ort gage plans designed to accom m odate better than traditional mortgages the current needs of residential mortgage borrow ers, mortgage lenders, or both. The long-term fixed-rate, fixed-paym ent mortgage becam e the preval ent type in the U nited States in the 1930s and served both borrow ers and lenders w ell as long as p rice and interest rate movem ents w ere relatively sm all. But recent increases in the level and vo latility of m arket interest rates have made this mortgage contract less desir able to lender and borrow er alike. A hom e is the largest single purchase that most A m ericans m ake and the largest single item in their w ealth portfolio. Because of the m agnitude of the expense— generally some tw o to t h r e e times th e p u rc h a s e r ’s a n n u a l incom e— a large portion of the purchase price of most homes is borrow ed. The ow ner's dow np aym ent, w hich represents the initial e q u ity, is generally only 10 to 30 percent of the purchase price. As a result, the cost, term s, and availability of mortgage credit are an im portant part of the hom e purchase decision. In recent years the cost of obtaining mortgage cred it has increased sharply. For exam p le, in 1971 the average interest rate on a 30-year single-fam ily fixed -coup on, fixedpaym ent mortgage with a loan-to-value ratio of 90 p ercent was about 7Va percent. In 1975 the rate on this mortgage had clim bed to 9 p ercen t, and in 1980 it reached 121/2 percent. O n a 90 p ercent fixed -rate, fixed-paym ent mortgage on a m edian price hom e of $24,800 in 1971, this represented m onthly payments of $160. The paym ents on a new mortgage on a hom e of the same price would have in creased to $182 in 1975 and $238 in 1980. But Federal Reserve Bank of Chicago the median price of existing homes also increased, from $24,800 in 1971 to $62,200 in 1980. Thus, a rate of 12V2 percent on a 90 percent fixed-rate, fixed-paym ent mortgage of $55,980 translates into m onthly payments of $597. M onthly payments have increased con siderably faster than household incom e. In 1971 the annual sum of m onthly payments was equal to 19 percent of the median family incom e of $10,285. In 1980 this percentage had almost doubled to 34 percent of an esti mated m edian fam ily incom e of $21,652.1 Some households found this percentage too high to pay and, as a result, decided not to purchase a house. Because most hom ebuyers are already housed and w ill sell their homes when buying ano th er, the higher mortgage rates reduce the turno ver of existing homes and thus labor m obility, but have relatively little effect on the overall stock of housing. The dem and for new ly constructed housing is m ore severely im pacted, but it accounts for only a small proportion of total housing. In the 1970s the average 1.7 m illion new units co n stru cte d a n n u a lly acco u n ted for only about 2.2 percent of the total housing stock of some 79 m illion units. The heaviest burden of the increase in mortgage payments falls on first-tim e home b uyers, w ho did not share in the rapid appre ciation in hom e prices and the perm issible tax-free transfer of gains from one home to another, and, in particular, on younger house holds, whose cu rren t incom es tend to be below the average for all households. In addi1Assumes an increase of 10 percent in 1980 over the reported median family income of $19,684 in 1979. Of course, the sizes and amenities of homes purchased in different years may be different. 3 tion, for these as w ell as other households, the burden of the mortgage payments may be expected to be greater at the beginning of the loan than later w hen the expected rate of inflation incorporated in the mortgage rate and in the m onthly paym ent m aterializes and increases incom es. This paym ent pattern, w hich is generally referred to as the “ tilt,” will be analyzed m ore closely in later sections of the article. Because the tilt is com m on to many mortgage plans, a large num ber of bor rowers w ould benefit from mortgage plans that w o u ld re d u c e th e in itia l m ortgage burden. At the same tim e traditional mortgage lenders, such as savings and loan associations, com m ercial banks, and mutual savings banks, have becom e increasingly reluctant to make long-term, fixed-rate mortgage loans. Because of the operating characteristics of these insti tutions, many of the fixed-rate loans they have made in recent years have proven unpro fitable. O n the w hole these institutions raise funds through short-term deposits. Through tim e, a series of successive short-term depos its, each at a fixed interest rate, is equivalent to a single long-term deposit at a variable interest rate. In order to charge an interest rate sufficient to cover the cost of the funds, other operating costs, and a com petitive profit when they relend the funds as a long term loan with the coupon or contract rate fixed for the m aturity of the loan, the institu tions must predict their future cost of funds over the life of the loan. If they predict their future deposit costs in co rrectly, the loans may be unp ro fitab le. Borrow ing (or lending) on a fixed-interest rate basis for one maturity w hile lending (or borrow ing) on a variablerate basis for the same m aturity is term ed interest rate interm ediation. Interest rate intermediation Econom ic theory suggests a relationship between short- and long-term interest rates. In eq u ilib riu m and assuming certainty of forecasts investors w ould be ind ifferent be tween buying two fixed-coupon rate secur 4 ities differing only in term to m aturity if the two securities w ere expected to yield the same interest return over the investors' ex pected holding period. If one fixed-rate se curity had a m aturity less than that holding period, the proceeds w ould have to be rein vested in one or m ore successive securities. It follow s that the investor must predict the rates on these successive securities in order to m ake a fully inform ed investm ent d ecision. If one alternative prom ised a higher return than the other, the funds w ould be invested in the one prom ising the higher return. By bidding up the price of the securities, this w ould serve to low er the expected interest rates on this alternative until the two alternatives promised the same average expected return. In m athem atical terms the yield on the long-term , fixed-coupon security is equal to an average of the current yield on the short term security and the yields predicted on suc cessive short-term securities over the life of the longer-term security. For a depository institution this relationship im plies that the rate it charges on a long-term , fixed-coupon mortgage loan should be equal to the average of the rate it cu rren tly pays on short-term deposits and the rates it expects to pay on these deposits in the fu tu re plus operating costs and allowances for both the risk of default on the mortgage and a com petitive profit. This relationship is illustrated in figure 1. Interest rates are plotted on the vertical axis and tim e on the horizontal axis.2 Assume that in the current p eriod, N, the institution pays com petitive interest rate A for one-period deposits and that this rate includes an allow ance for operating costs and a com petitive profit. The institution wants to make a fixedrate loan for q periods to period M . W hat rate should it charge? The institution first needs to predict its one-period deposit rate in each period from N + 1 to M . If it expects the costs 2Because the interest on most securities is com pound interest, it is necessary to plot the logarithm of (1 + i) on the vertical axis. But this may be interpreted as simply the interest rate. For the sake of simplicity, the diagram also assumes a nonamortized loan. Econom ic Perspectives Figure 1. Determination of interest rate on fixed-rate mortgages depends on projection of future deposit rates interest rate* •Technically, interest rate on the vertical axis is measured in the form of ln(1+i). of deposits to rise steadily to C along the straight line A C , the average deposit cost, B, is the rate it should charge on a long-term loan in order to m ake a com petitive profit. The expected total cost of the deposits is the area N A C M . The expected revenue on the loan is the area in the rectangle N BDM . The two areas are eq ual, so that total ex pected revenues are equal to expected total deposit costs. This equality may also be seen by exam ining the d ifferences between the interest rate earned on the loan and that expected to be paid on the deposits. In period N the profit is B - A. This profit, w hich is in addition to the com petitive profit in cluded in the deposit cost, is reduced steadily as deposit rates are expected to rise until the two rates are equal at E at tim e P. Th ereafter, the deposit rate is expected to rise above the loan rate up to C - D at period M . Although the institution w ill exp erien ce losses from time P through tim e M , it breaks even over the en tire period N - M as the loss triangle C D E is exactly equal to the earlier gain tria n gle ABE. These losses must be charged against the previous extra profits, w hich would be classified m ore accurately as reserves against future expected losses than as profits. Federal Reserve Bank o f Chicago Note that if deposit rates had been e x pected to d e clin e , the interest rate on the long-term , fixed-coupon loan would have been low er than the initial deposit rate. The institution w ould have experienced losses at the beginning of the p erio d , but would have expected to recoup them later as deposit rates d eclin ed below the loan rate. During the loss period the institution is said to be exp erien cing a liq u id ity p ro b le m as its cash inflows are insufficient to satisfy its cash out flow s. But this is expected to be only a tem porary problem and may be accom m odated by using the reserves accum ulated in past periods of greater than com petitive account ing profits. Now assume that future deposit rates w ere predicted in co rrectly. After a loan was m ade, the cost of deposits actually increased along line A G rather than A C , so that the cost of deposits was underestim ated. The loan rate charged, B, is now insufficient to cover the actual cost of the deposits. The gain triangle ABH is sm aller than the loss triangle DH G. This is a lasting loss and, if not offset quickly by unexpected gains experienced in periods in w hich the increase in deposit rates was overestim ated and the loan rate charged was higher than necessary, w ill result in a solvency p ro b le m . In retrospect, the institution should have charged a loan rate of J. T h u s, the im p o rta n ce fo r lon g-term , fixed -rate len d in g of co rre ctly predicting future deposit rates is clear. In making fixedrate loans, the lender assumes all the risk of unfavorable interest rate changes over the life of the loan. It is effectively selling interest rate in su ra n ce to th e b o rro w e r. Likeany insur ance co m pany, it may be expected to charge a prem ium for this insu ran ce, the size of w hich is dependent on the estimated degree of risk in cu rred . This prem ium is sim ply included in the interest rate charged the borrow er. M arket interest rates also incorporate an inflation prem ium to com pensate lenders for the expected loss in the value of their p rin ci pal due to inflation over the period that the credit is outstanding. In recent years higher 5 and more volatile rates of inflation have caused m arket interest rates to be higher than most market participants— including most depository institutions— expected. As a re sult, many long-term loans made at fixed interest rates expected to be profitable at the time of origination have turned out to be unprofitable. That is, in retrospect the insur ance prem ium charged was insufficient to com pensate the len d er for the loss incurred . In addition, the increased volatility of both inflation and interest rates has im paired the co nfid ence of depository institutions in their ability to predict future deposit rates and therefo re to estim ate accurately the insu r ance prem ium to be charged. Not only have depository institutions been faced with higher and m ore volatile market interest rates, but regulatory ceilings have lim ited the rates that they are permitted to pay on deposits. W hen m arket interest rates rise above the m axim um interest rates permitted on deposits, savers w ithdraw their deposits to purchase unregulated financial instrum ents in the m arkets. To enable com m ercial banks and thrift institutions to co m pete more effectively for funds in the open market and to provide a m ore even flow of funds for home mortgage lending, the regu latory agencies authorized m oney market certificates (M M Cs) in 1978 and small savers certificates (SSCs) in 1979, the rates on w hich are tied to market rates on Treasury securities. These moves w ere carried to their logical conclusion by the D epository Institutions Deregulation and M onetary Control Act of 1980 (D ID M C A ), w hich provides for phas ing out interest rate ceilin g s on all tim e and savings deposits at depository institutions by M arch 31, 1986.3 C o n seq u en tly, the aver age cost of funds of depository institutions will increasingly reflect changes in market rates of interest and therefore be even more unpredictable in the future. Because of the increased volatility in 3See “The Depository Institutions Deregulation and Monetary Control Act of 1980,” Econom ic Perspectives, Federal Reserve Bank of Chicago (September/October 1980). 6 deposit rates, many institutions have become increasingly reluctant to engage in interest rate interm ediation and to make long-term , fixed -rate loans. These institutions w ould benefit from long-term loans whose interest rates could vary more closely with their cost of funds. U nexpected increases in the cost of funds w ould be passed through to the b or rower and leave the institution u naffected.4 Alternative mortgage instrum ents are de signed to accom m odate the new needs of both mortgage borrow ers and lenders. Be cause the problem s of borrow ers and lenders d iffer, the mortgage plan best suited to one may not be the one best suited to the other. Risk averse mortgage lenders would benefit most from mortgages whose rates fluctuate closely with their cost of deposits, w hile younger mortgage borrow ers stand to gain from mortgages whose initial m onthly pay ments are low relative to their incom es. U n fo rtu n ately, these two objectives are not always m utually consistent. M ortgage plans with increased rate volatility may not reduce early m onthly payments and w ill transfer interest risk from the lending institution to the b o rro w er, w ho is often less w illing and able to assume it. M ortgage plans with low er initial m onthly payments need not increase rate flexib ility and could result in monthly payments to the depository institution smaller than necessary to cover the m onthly interest cost of the loan. Any shortfall would be added to the p rincipal of the loan in the form of negative am ortization and, by increasing the unpaid balance of the loan, w ould increase the risk of default. The increase in default risk may offset part or all of the benefit to the lender of the d eclin e in interest rate risk from rate flexib ility. The fo llow ing sections des cribe a num ber of “ alternative” mortgage plans, either already in use or proposed, and their advantages and disadvantages to lend ers and borrow ers. “•Regulations adopted by the FHLBB, effective July 10,1981, giving S&Ls broader latitude to engage in inter est rate futures transactions, are also expected to reduce their net interest rate exposure. See Am erican Banker, “ Broad Powers Given S&Ls to Use Futures,’’ Vol. 146, No. 130, Monday, July 6, 1981, page 1. Econom ic Perspectives Adjustable-rate mortgages Adjustable-rate mortgages (A R M s), as their name im plies, are mortgage plans whose contract or coupon interest rates change p erio d ically after origination according to some agreed upon conditions. Pure A R M s, that is, A R M s w ithout any restrictions on rate changes, w ould elim in ate th e lending institu tio n ’s interest rate exposure com pletely if the loan rate changed im m ediately every tim e the deposit rate changed and by exactly the same am o u n t.5 The change in the cost of funds effectively w ould be passed through to the mortgage b orro w er and not im pact the institution at all. All the risk of unfavorable interest rate changes that was borne by the lender under fixed-rate mortgages (FRM s) w o uld now be borne by the b orro w er. In term s of figure 1, the mortgage loan rate on an A R M w ould be superim posed on the d ep osit rate lin e. The institution would be freed of any need to forecast interest rates in order to price its mortgages co rrectly.6 But the A R M is not costless to either le n d e r or b o rro w e r. A lth o u g h m ortgage lending institutions incurred risk in making long-term , fixed-rate mortgages, they charged a prem ium for this service w hich until recent years was su fficien tly high to make this activ ity profitable. If the institutions elim inate this risk, they w ill also elim inate a line of business that could again be potentially profitable. Financial institutions, w hich are in the busi ness of dealing with interest rate changes every day and are by nature fam iliar with finance and eco n o m ics, may be expected to be better situated than individual households to m ake m eaningful interest rate predictions and assume interest rate risk. Although house hold mortgage b orrow ers would benefit from alternatively, the institution could eliminate inter est rate risk by altering its deposit structure to match the characteristics of its fixed- and/or variable-rate mortgages. 6The appropriate mortgage rate would be the rate on deposits with a maturity equal to the period between permitted changes in interest rates on the mortgage, plus a premium for default risk and other costs associated with originating and servicing mortgage loans. Federal Reserve Bank o f Chicago A R M s if interest rates w ere to fall, many tend to be risk averse, putting greater weight on interest rate increases than decreases of the same m agnitude. They are generally w illing to pay some prem ium to insure themselves against the possibility of paying unexpectedly higher rates during the life of the mortgage. This is so even though the average borrower's incom e rises w ith the interest rate as might be expected if market rates incorporate a pre mium for expected inflation. To achieve a com prom ise between these two positions, the cum ulative change in in terest rates on A R M s can be limited to a predeterm ined band, e.g ., 2Vi or 5 percent age points or 30 or 50 percent above and below the initial contract rate. M oreover, individual rate changes also can be restricted to some m axim um am ount— e.g ., Vi or 1 per centage point—and lim ited as to frequency— e .g ., once every six m onths, every year, or every five years. Thus, the interest rate risk is shared by the b orro w er and the lender. The cost of changes in market interest rates within the overall band is borne by the b orro w er; outside the band by the lender. Because the width of the band affects the p re m iu m the borrow er pays for “ interest rate insurance’’— the low er the p rem iu m , the w ider the band— the band acts like a deductability clause in an accident or fire insurance policy. A R M S do not mitigate a particular form of the “ tilt” problem that arises when market interest rates increase in response to upward revisions of inflationary expectations. These rate increases are translated im m ediately into increases in m onthly paym ents, w hile the b o rro w er’s incom e increases only slowly with the realized rate of inflation. Thus, the bur den of the mortgage increases im m ediately. Through tim e, as the b o rro w e r’s incom e rises and m onthly payments remain unchanged, the burden w ill declin e. N evertheless, the tilt of the burden to the early months of the mortgage may price some potential home buyers out of the market. A variety of A R M plans have been used or proposed. In January 1979 the Federal Hom e Loan Bank Board (FH LBB) permitted 7 federally chartered savings and loan associa tions to make variable-rate mortgages (VRM s) on w hich the interest rate could change after origination with the cost of funds to all fed e r ally insured savings and loan associations, but by no m ore than Vi percentage point once a year, and by no more than 2Vi percentage points over the life of the m ortgage.7 M onthly payments would change accordingly. In 1980 the FH LBB perm itted a variation on this them e term ed the renegotiable-rate mortgage (R R M ) in w h ich the interest rate could again change by V2 percentage point annually but the m axim um change over the life of the loan increased to 5 percentage points. M o reo ver, the changes in the rate could be translated into changed m onthly payments only once every three to five years and w ere tied to changes in the national aver age rate on mortgages as m easured by the FHLBB contract mortgage rate index for co n ventional mortgages for the purchase of exist ing hom es, rather than to the cost of funds to the lenders. Because the new mortgage rate was based on loans by all lenders, not just savings and loan associations, and because it is highly publicized, the causes of rate changes on loans could be better understood by bor rowers. H ow ever, the inability of the new mortgage rate to change in perfect synchron ization with the cost of funds means that the lender is not fully protected from interest rate risk. In ad d ition, as the rem aining life of the mortgage declines through tim e, the appli cable rate on the mortgage may be expected to differ m ore and m ore from the prevailing rate on new fixed-rate m ortgages, w hich c u r rently dom inate the new mortgage index. O n M arch 23, 1981, the C o m p tro ller of the C u rre n cy (C O C ) authorized national banks to offer, w ithin specified guidelines, adjustable-rate mortgage loans for the pur chase of one- to four-fam ily ow ner-o ccup ied homes. U nder the regulation the interest rate on a mortgage loan may change in accor dance with any one of three specified refer7Federally chartered associations in California were granted this authorization in 1978. 8 Federal regulation ot mortgage lending Some have questioned w h e th er the C o m p tro lle r of the C u rre n c y has the authority to prom ulgate adjustable-rate mortgage regula tions for national banks and to preem pt state laws that restrict such mortgages. Tw o recent federal court decisions support Federal Hom e Loan Bank Board preem ptive regulations estab lishing uniform standards for real estate lending by fe d e ra lly c h a rte re d saving s and loan associations. In C o n fe r e n c e o f Federal Savings and Loan Associations v. Stein, 604 F.2d 1256 (9th C ir. 1979), a ffirm ed 445 U.S. 921 (1980), the state of C alifo rn ia had req uired federal savings and loan associations to abide by the provisions of the state's an ti-red linin g act. The court co n cluded that w here federal regulation such as the regulatory control of the FH LBB over fed eral S&Ls is so pervasive as to leave no room for state regulatory co n tro l, im plicit preem ption can be fo und. In G len dale Federal Savings and Loan A sso ciation v. Fox, 459 F. Supp. 903 (C .D . C al. 1978), appeal p e n d in g , the co urt reasoned that the Congress had given the FHLBB com plete author ity to charter and regulate federal S&Ls so as to proh ibitstatesfro m reg ulatingthem . (A num ber of other circu it courts have reached sim ilar co nclusio ns.) T h e re fo re , w ith respect to m ort gage loans extended by federal S&Ls, federal law can preem pt state regulation of the validity and e xercisab ility of “ due-o n-sale” clauses re qu iring com plete repaym ent of the mortgage w hen the hom e is sold. Federal law governing m ortgage lending by national banks is contained in the general and sp ecific rule-m aking authority granted by Title 12 of the United States C ode on banks and banking, 12 U .S .C . §1, et seq. (especially §371 (g), w hich states that loans are subject to conditions and lim itations prescribed by the C o m p tro ller of the C u rre n c y by rule or reg u latio n ), the National Bank A ct, other federal banking laws, and the Housing and C o m m u n ity D evelopm ent Act of 1974 (w hich liberalized the powers of national banks to m ake real estate loans). Taken to gether, these provisions suggest that it was the intent of the Congress that the C o m p tro ller of the C u rre n cy regulate real estate lending by n a tio n a l b a n k s , th u s p r e e m p tin g sta te regulations. Econom ic Perspectives ence rates by a m axim um of 1 percentage point every six m onths with no limit on the cum ulative change over the life of the m ort gage. At the option of the bank, monthly paym ents may be m aintained at a fixed dollar am ount for a specified tim e up to five years, regardless of changes in the interest rate. If the bank chooses this o p tio n , increases or decreases in the interest rate w ill result in changes in the proportions of the m onthly paym ent credited to interest and repaym ent of prin cipal. If the required interest payment is greater than the am ount of the fixed m onthly paym ent, the differen ce may be added to the outstanding principal in the form of negative am ortization within a speci fied lim it. Upon expiration of a period of fixed m onthly paym ents, the m onthly pay ment is adjusted up or dow n for the next period to provide full am ortization of the outstanding p rincipal balance w ithin the re maining tim e to m aturity of the loan. The FHLBB adopted regulations A pril 30, 1981, that granted federally chartered thrift institutions broader flexib ility in thedesign of adjustable-rate mortgages. Interest rate ad justm ents on the loans may be tied to any referen ce rate provided that it can be verified easily by the b o rro w er and it is not controlled by the len d er. The interest rate may be adjusted through changes in the m onthly paym ent and /or in the loan term , subject only to the conditions that the loan term from the date of closing be lim ited to 40 years and that the paym ent am ount be adjusted at least every five years to a level sufficient at the existing interest rate to am ortize the loan fully over its rem aining life. (The restrictions im posed by the two agencies are sum m arized on page 22.) In allow ing such a broad range of options, the FH LBB and the C O C are clearly assuming the developm ent of a large variety of individual mortgage plans with varying degrees of risk sharing. Because many bor row ers may not wish to assume the risk of a fully adjustable mortgage rate and many m ort gage lenders may not wish to spin off their interest rate risk insurance business alto gether, this forecast is likely to be correct. Federal Reserve Bank o f Chicago Developm ent of mortgage plans The new regulations of both the C O C and the FHLBB perm it lenders to develop mortgage plans designed to meet both their own needs and those of borrow ers. The var iety of mortgage plans that can be designed and offered under the regulations of both the C O C and the FH LBB is virtually infinite. In p ractice, h o w ever, variations in the design are likely to depend prim arily on differences in a relatively few key characteristics. These in clu d e : (1) the choice of a reference rate; (2) the frequency and amount of adjustment of the interest rate; (3) the frequency and am ount of change in the m onthly paym ent; (4) the lim itations on additions to the out standing principal b alance, i.e ., negative am ortization; and (5) perm itted extensions of the m aturity of the original loan. Using a co m puter, it is possible to sim u late the behavior of several critical variables such as the frequency and amount of interest rate adjustm ents of different mortgage plans. Such sim ulations are useful in evaluating the impact of each mortgage plan on both the borrow er and the lender. For exam ple, the sim ulations allow the borrow er to compare the b e h av io r o ver tim e of the exp ected m onthly paym ent and the outstanding p rin cipal balance of each plan. In addition, they enable the lending institution to project its cash flow and interest incom e over the life of the loan. Sim ulations can be run using either data from some past p erio d — in w hich case they show how particular types of mortgages would have perform ed if they had been made at the beginning of that period— or hypothetical data believed useful in illum inating the likely future p erform ance of mortgages with d iffer ing characteristics. The relative desirability of the various types of mortgages depends heav ily on the actual behavior of inflation and m arket interest rates in the fu tu re, and that may not resem ble either the past or any other assumed behavior. N evertheless, given one's expectations about the fu tu re, simulations provide a useful way of exam ining the impli- 9 cations of those expected conditions for the perform ance of different mortgages. Choice of reference rate Both the C O C and the FHLBB require that adjustm ents in the mortgage interest rate be tied to a sp ecific referen ce rate. National banks are authorized to use the six-month Treasury bill auction rate, the three-year co n stant m aturity Treasury note rate, and the FHLBB national average contract mortgage rate on conventional mortgages for the p ur chase of existing hom es. A federal S&L may use any agreed-upon published rate, but not its own mortgage rate or its own cost of funds. Sim ulations of mortgage plans with the interest rate tied to each of the three re ference rates authorized by the C O C for the period from January 1970 through Decem ber Table 1 Sim ulations of A R M plans with alternative reference rates 1970-1980 ($20,000 m ortgage)1 FHLBB contract rate*2 Six-month period Monthly payment Ending balance Six-month T-bill3 Monthly payment Three-year T-note3 Ending balance Monthly payment S&L cost of funds3 Ending balance Monthly payment Ending balance 19,920 19,839 19,739 19,605 19,476 19,336 19,201 19,078 18,961 18,847 18,737 18,604 18,480 18,335 18,171 17,999 17,835 17,684 17,543 17,406 17,276 17,160 148.57 152.76 154.43 154.70 154.57 154.84 155.24 155.78 159.23 162.94 166.64 167.01 167.43 167.55 168.19 168.07 169.18 169.92 172.97 178.31 184.12 197.03 19,920 19,842 19,762 19,679 19,593 19,503 19,409 19,312 19,215 19,119 19,022 18,922 18,817 18,707 18,592 18,472 18,348 18,218 18,087 17,956 17,827 17,708 (dollars) 1970.1 .2 1971.1 .2 1972.1 .2 1973.1 .2 1974.1 .2 1975.1 .2 1976.1 .2 1977.1 .2 1978.1 .2 1979.1 .2 1980.1 .2 148.57* 149.68 145.97 140.06 139.40 138.36 140.42 142.60 152.54 157.09 163.61 158.05 160.34 157.58 157.46 157.34 158.78 164.29 172.35 179.47 192.96 198.47 Average 157.97 19,920 19,838 19,749 19,647 19,541 19,430 19,317 19,203 19,000 18,997 18,899 18,789 18,676 18,555 18,428 18,295 18,159 18,025 17,897 17,772 17,658 17,544 148.59 147.07 132.75 109.56 117.80 107.64 118.93 135.98 156.75 155.43 156.39 130.45 137.12 125.81 122.87 120.89 132.77 141.21 157.87 170.45 184.86 196.45 141.26 19,920 19,835 19,727 19,572 19,430 19,261 19,113 18,993 18,898 18,797 18,692 18,541 18,397 18,224 18,040 17,844 17,671 17,508 17,369 17,243 17,129 17,023 148.57 150.03 138.19 120.16 125.27 120.83 126.20 135.47 142.12 146.77 152.71 141.68 149.48 140.95 133.32 132.87 139.35 149.41 159.03 166.19 175.13 189.93 144.71 164.98 'Assumes 30-year $20,000 adjustable-rate mortgage at 8.13 percent closed in January 1970 and tied to indicated reference rate. Mortgage interest rate and monthly payment are adjusted every six months without restriction on either. -Mortgage loan interest rate for each six-month period is the actual FHLBB mortgage loan contract rate on loans closed in January and July. 'Mortgage loan interest rate for each six-month period after the first is the average rate during the prior six months plus the spread between 8.13 percent and the average rate during the last six months of 1969. ■'Monthly payment for comparable fixed-rate mortgage; ending balance in 1980.2 would be $17,228. 70 Econom ic Perspectives 1980 are shown in table 1. For the sake of sim p licity, all of the sim ulations in the table assumed that a 30-year mortgage loan for $20,000 was closed in January 1970 at 8.13 p er cen t, the actual FH LBB national average co n tract rate on fixed-rate mortgage loans for the purchase of existing homes at that tim e. The mortgage loan interest rate was adjusted every six m onths and the m onthly payment changed accordingly w ithout restriction. The am ount of the adjustm ent in the interest rate was equal to the change in the reference rate.8 The sim ulations indicate that an adjust able-rate mortgage closed in January 1970 and tied to the six-m onth Treasury bill rate would have required the lowest average m onthly paym ent during the 11-year period and w ould have resulted in the lowest ending principal balance of the three adjustable mortgage plans. The average m onthly pay ments and the ending principal balances for the mortgage plans with interest rates tied to the six-m onth Treasury bill rate and to the three-year Treasury note were both lower than those for the fixed-rate mortgage. The results of a sim ulation of a mortgage plan with the interest rate tied to the FHLBB mortgage contract rate w ere very sim ilar to those of a mortgage plan tied to the S&L’s average cost of funds. S&Ls' cost of funds has increased steadily since 1970, and mortgage interest rates, although declining over short periods of tim e, have also been in a generally upward trend. Both of these plans w ould have had paym ents higher than those on a fixed-rate mortgage. These results are specific to the period used, h o w ever. Interest rates w ere at a rela tively high level in 1970 and the subsequent d eclin e in rates w ould have required a co r responding reduction of the interest rate on 8The monthly rates were averaged over the sixmonth periods in the above simulations except when the reference rate was the FHLBB mortgage contract rate. The use of averaging generally reduces the amount of fluctuation in the reference rate and may also prevent tying a mortgage loan interest rate to a reference rate, particularly a sh'ort-term interest rate, that is temporarily distorted. Federal Reserve Bank o f Chicago an adjustable-rate mortgage. The w ider flu c tuations of the six-m onth Treasury bill rate require greater changes in the mortgage loan interest rate. Sim ulations of mortgage plans beginning w hen interest rates w ere relatively low and for shorter periods of time w ill show different results. In addition, all mortgage plans w ould not be expected to have the same initial interest rate. Adjustment of the interest rate Although the above sim ulations adjusted the mortgage loan interest rate every six months by the full am ount of the change in the referen ce rate, the interest rate may be adjusted more or less frequently and the am ount of the adjustm ent may be limited both at each adjustm ent period and over the term of the loan. M ore frequent adjustment translates changes in market interest rates into changes in mortgage loan rates more quickly. Restrictions on the amount of the adjustm ent w ill reduce the correlation of mortgage loan rates with market interest rates. The regulations of the C O C limit the adjustm ent of the mortgage interest rate in either d irectio n to 1 percentage point every six months with no limit on the cum ulative change over the life of the mortgage. In the above sim ulations this restriction would have had little effect on the use of the FHLBB co n tract mortgage rate and the three-year Trea sury note rate. It w o uld , how ever, have re duced the am ount of perm itted change if the six-m onth Treasury bill had been used as the referen ce rate. A com parison of mortgage plans with the six-m onth Treasury bill rate as the referen ce rate with and w ithout the lim itation to a 1 percentage point change every six months is shown in table 2. During the period from 1970 to 1980, the limitation on the adjustm ent to the interest rate in this instance resulted in a slightly lower average m onthly payment and ending balance out standing because the higher market interest rates w ere not fully translated into higher mortgage loan interest rates. The FHLBB regulations impose no spe- 11 Table 2 Simulation of ARM plan with C O C limitation on adjustment of interest rate1 1970-1980 No limitation Six-month period Monthly payment C O C limitation Ending balance Monthly payment Ending balance (d o lla r s ) 1970.1 .2 1971.1 .2 1972.1 .2 1973.1 .2 1974.1 .2 1975.1 .2 1976.1 .2 1977.1 .2 1978.1 .2 1979.1 .2 1980.1 .2 148.57 147.07 132.75 109.56 117.80 107.64 118.93 135.98 156.75 155.43 156.39 130.45 137.12 125.81 122.87 120.89 132.77 141.21 157.87 170.45 184.86 196.45 Average 141.26 19,920 19,835 19,727 19,572 19,430 19,261 19,113 18,993 18,898 18,797 18,692 18,541 18,397 18,224 18,040 17,844 17,671 17,508 17,369 17,243 17,129 17,023 148.57 147.07 133.61 120.81 117.95 107.77 119.07 131.59 144.06 155.42 156.38 143.90 137.28 125.96 123.02 121.04 131.78 141.36 153.96 165.35 176.91 188.60 19,920 19,835 19,729 19,596 19,454 19,285 19,137 19,009 18,896 18,795 18,691 18,563 18,418 18,246 18,062 17,866 17,690 17,527 17,381 17,247 17,123 17,008 140.52 'Assumes 30-year $20,000 adjustable-rate mortgage at 8.13 per cent closed in january 1970 and tied to six-month Treasury bill rate as referencerate. Mortgage interest rate and monthly payment are adjusted every six months. cific lim itations on the freq uency or am ount of the adjustm ent of the mortgage loan in te r est rate. The frequency and am ount of change w ill thus depend on the reference rate used and how often it is available and on any lim ita tions and term s that are incorporated in the mortgage contract by the lending institution. Changes in the monthly payment Changes in the m onthly payment may be made w h en ever the interest rate is changed or at some other agreed upon tim e. In the latter instance the m onthly paym ent may be held constant for a given period during w hich the interest rate is perm itted to fluctuate in accordance with the specified reference rate and any sp ecific lim its in the contract. If the required m onthly interest payment exceeds 12 the total m onthly paym ent, the m onthly pay ment must be increased to cover the interest, unless negative am ortization is perm itted. W hen the m onthly paym ent is changed, it is generally increased or decreased, as neces sary, to am ortize the loan fully over the re maining term of the mortgage. The earlier FHLBB regulations auth o riz ing VRM s and RRM s required a co rresp o nd ing change in the m onthly payments w h en ever the interest rate on the mortgage was increased or decreased. The dollar amounts of the changes in the m onthly paym ent were limited by the restrictions on the periodic and overall increases in the mortgage loan inter est rate. H ow ever, no such interest rate lim ita tions are included in the FHLBB regulations issued in A pril 1981, and the C O C regulations lim it the change in the mortgage interest rate to 1 percentage point every six m onths. Thus, unless restrictions are imposed on changes in m onthly paym ents, interest rate changes will be translated im m ediately into changes in monthly payments. Restrictions on changes in the m onthly paym ent—so-called payment caps— are an alternative to or may be used in conjunction with lim itations on changes in the interest rate. For exam ple, the m onthly payment may be held constant for a given period, say three years, but the interest rate on the loan may change every six m onths. O r the monthly payment may change sim ultaneously with changes in the interest rate, but any increase may be lim ited to some dollar or percentage increase over the payment in the prior period. Specific restrictions on payment caps w ere not included in the recent regulations of either the C O C or the FH LBB. H ow ever, several com m ercial banks and thrift institu tions have been offering 30-year A R M s since the fall of 1980 with lim itations on payment changes. Negative amortization Any upward adjustm ent in the mortgage interest rate that is not accom panied by a change in the monthly payment sufficient to Econom ic Perspectives am ortize the loan over the rem aining life of the loan may req u ire negative am ortization, i.e ., an addition to the outstanding loan bal ance. As noted above, this may be necessary w hen the m onthly payment is held constant fo ra given period of tim e but the interest rate increases or w hen the increase in the monthly paym ent is lim ited to a specified percentage or am ount. Although the regulations of both the C O C and the FH LBB perm it negative am orti zation in A R M s, the two differ som ewhat. The FHLBB regulation requires that the m onthly payments be adjusted at least once every five years to am ortize fu lly, w ithin 40 years from the date of closing, the outstanding principal at the interest rate indicated by the reference rate. The C O C regulation limits negative am o rtizatio n , for periods during w hich the m onthly paym ents are fixed , to no more than 1 percent of the principal outstanding at the beginning of the fixed-paym ent period times the num ber of six-m onth intervals w ithin the fixed-paym ent period. M onthly payments must be adjusted at least every five years to an am ount su fficien t to am ortize the outstand ing p rincipal over the rem aining term . Thus, the m axim um negative am ortization perm it ted over a five-year period would be 10 per cent of the p rincipal outstanding at the be ginning of the period. D uring the period of negative am ortiza tion the ratio of the loan to the value of the house increases and may becom e greater than 1.00. W hen this o ccu rs, the lending insti tution incurs significant risk of default be cause the b orro w er may “ walk away” from the loan. Because housing prices tended to rise sharply in recent years, greater than 100 percent loan-to-current-value ratios would have been u n like ly. N evertheless, all housing prices did not rise and even the average house price may not rise as rapidly in future years. As a result, the initial loan-to-value ratio takes on additional significance. Sim ulations of mortgage plans that incorp oratethe lim itationson interest ratechanges and negative am ortization prescribed by the C O C regulations indicate that, with an inter Federal Reserve Bank o f Chicago est rate of 14 p ercen t, the lim it on negative am ortization is not breached, even when the interest rate changes by the maximum allow able 1 percentage point every six months, unless the m onthly payment remains con stant for at least two years. In table 3 the interest rate on a $1,000 mortgage loan with an original rate of 14 percent was increased 1 percentage point every six m onths, the max imum perm itted. W hen the monthly pay ment was held constant for at least one year, negative am ortization always occurred in the second and subsequent six-m onth periods until the m onthly payment was adjusted. H o w ever, the perm itted lim it on negative am ortization was not reached unless the m onthly paym ent was unchanged for at least two years and then not until the second half of the sixth year of the mortgage term. In all sim ulations, when the m onthly payment was held constant for a year or m ore, the out standing balance of the loan was above $1,000, the original am ount of the loan, at the end of 12 years. Extension of loan maturity In order to lessen the impact of an increase in the interest rate on the required m onthly paym ent, the FHLBB regulations perm it extension of the m aturity of the loan up to a m axim um of 40 years from the date of closing. H ow ever, when interest rates are high, an increase in the mortgage interest rate, by even a small am ount, quickly results in the extension of the loan term to the full 40 years, as shown in figure 2. This is especially true in the early years when interest rep re sents substantially all of the monthly pay ment. The im m ediate benefit to the borrower is, th erefo re, lim ited , and the monthly pay ments continue for 40 years instead of the original 30. Graduated-payment mortgages Graduated-paym ent mortgages (GPMs) are adjustable-paym ent mortgages on which the m onthly mortgage payments increase 13 Simulations of ARM plans with C O C negative amortization limit and selected constant payment periods b. end-of-period outstanding balance ($1,000 mortgage) a. average monthly payment ($1,000 mortgage) Six-month period 1.1 .2 2.1 .2 3.1 .2 4.1 .2 5.1 .2 6.1 .2 7.1 .2 8.1 .2 9.1 .2 10.1 .2 11.1 .2 12.1 .2 Interest rate Constant payment period 1 year 2 years 3 years 6 mos. (percent) (dollars) 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 11.85 12.64 13.44 14.24 15.05 15.85 16.67 17.48 18.29 19.11 19.92 20.74 21.56 22.37 23.19 24.01 24.83 25.65 26.47 27.29 28.11 28.93 29.75 30.57 11.85 11.85 13.51 13.51 15.19 15.19 16.92 16.92 18.66 18.66 20.43 20.43 22.22 22.22 24.03 24.03 25.86 25.86 27.72 27.72 29.59 29.59 31.48 31.48 11.85 11.85 11.85 11.85 15.52 15.52 15.52 15.52 19.49 19.49 19.49 19.73* 23.71 23.71 23.71 24.21 ‘ 28.19 28.19 28.19 28.89* 32.93 32.93 32.93 33.82* 11.85 11.85 11.85 11.85 13.32* 14.93* 17.60 17.60 17.60 17.71* 20.06* 21.20* 23.98 23.98 23.98 24.51* 26.82* 28.08* 31.00 31.00 31.00 31.81* 34.22* 35.60* Six-month period 1.1 .2 2.1 .2 3.1 .2 4.1 .2 5.1 .2 6.1 .2 7.1 .2 8.1 .2 9.1 .2 10.1 .2 11.1 .2 12.1 .2 Interest rate Constant payment period 1 year 2 years 3 years 6 mos. (percent) (dollars) 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 999 998 997 996 996 995 995 994 994 993 993 993 993 992 992 992 992 991 991 991 991 990 990 990 999 1003 1002 1006 1006 1010 1010 1014 1014 1019 1018 1023 1023 1028 1028 1033 1033 1038 1038 1043 1043 1048 1048 1053 999 1003 1012 1028 1027 1032 1042 1059 1059 1064 1075 1092 1092 1097 1109 1126 1126 1131 1144 1161 1161 1166 1180 1197 999 1003 1012 1028 1041 1051 1050 1055 1066 1084 1094 1105 1104 1110 1122 1139 1150 1161 1161 1167 1180 1197 1209 1220 ‘ Limit on negative amortization requires increase in monthly payment. according to a p red eterm ined sched ule. These plans are designed prim arily for young er fam ilies whose incom es are cu rren tly low but may be expected to rise faster than aver age as they enter the m ore productive years of their lives. G raduated-paym ent mortgages were originally authorized by the Housing and C o m m unity D evelopm ent Act of 1974, and five plans w ere subsequently offered by H U D -FH A . The plan in w hich monthly paym ents could increase by IV.2 percent annually for five years proved to be the most popular. Since th en , a num ber of private lending institutions have offered co nventio n al graduated-paym ent mortgage plans sim ilar to this plan. U nder many of these plans, the low early m onthly paym ents may be sm aller than the monthly interest on the outstanding balance. 14 If so, the d ifferen ce is added to the principal of the mortgage loan. As the monthly pay ment increases, it eventually rises above the am ount necessary to pay the interest on the outstanding principal and the principal be gins to am ortize. The later payments are suf ficien tly higher than the required interest in o rd e r to o ffse t th e e a rlie r n e g a tive am ortization. Default risk is potentially a serious prob lem with graduated-paym ent mortgages. The higher the initial ratio of loan to house value, the low er must be the graduation in order to avoid having the loan value rise above the house value. The risk of default may also be greater on GPM s than on other mortgages because the borrow er's incom e may fail to increase as rapidly as the m onthly payments. A lth o u g h the d o lla r am o u n t of the Econom ic Perspectives Figure 2. Increase in mortgage interest rate necessary to extend loan term from 30 to 40 years percent m onthly paym ent varies on all G PM s, the interest rate may or may not. Alm ost all GPM s that have been offered are fixed-rate m ort gages. Th u s, these plans do not assist lending institutions in reducing their interest rate risk. G rad u ated -p aym en t adjustable-rate m ort gages (G PAM s) have recently been proposed by the FH LBB. Although G PAM s would offer institutions better protection from interest rate risk, rapid increases in interest rates in the early years w ould raise sharply the e ffe c tive rate of graduation, thereby increasing default risk. In addition, the variability in m onthly paym ents w ould be increased to a greater extent than with regular ARM s under like circum stances. N evertheless, GPAM s may be a w o rkab le com prom ise instrum ent that reduces both the interest rate risk exposure of thrift institutions and the early monthly payments burden of econom ically promising younger hom e buyers. Price level-adjusted mortgages In periods of rapid inflation, the m ort gage plans discussed so far (except GPM s) req u ire, at the tim e of their origination, large m onthly paym ents relative to the mortgage borrow er's incom e. As the borro w er’s incom e increases through tim e in line with the rate of in flatio n , the burden of the m onthly pay Federal Reserve Bank of Chicago ments declines and offsets the previous in crease so that, for the en tire life of the m ort gage, the burden to the borrow er is no higher. As discussed e a rlie r, the mortgage paym ent-to-incom e ratio is said to be tilted dow nw ard. This troublesom e “ tilt” problem can be reduced or elim inated by a price leveladjusted mortgage (PLAM ) w hich ties the m onthly paym ents on the mortgage to the price level rather than to the interest rate. Thus, accelerations in the rate of inflation are translated into higher mortgage payments only as they occu r rather than all at once when they are first anticipated and become em bodied in interest rates as an inflation prem ium . In addition, m onthly payments under a PLAM reflect actual rates of inflation, rather than expected rates of inflation. As a result, neither mortgage lenders nor mortgage bor rowers are injured (or rew arded) financially if the expected rate of inflation im pounded in the interest rate is not realized. Like almost everyone else, participants in mortgage mar kets badly underestim ated the future rate of inflation throughout most of the 1960s and 1970s. In retrospect, the rates on fixed-rate mortgages in these years were lower than necessary to m aintain the purchasing power of the p rincipal. The resulting loss to the lend ing institutions is the major cause of the financial difficulties these institutions are now exp erien cing . In contrast, mortgage borrow ers enjoyed w indfall gains. Because PLAM s both protect lenders from unexpected increases in the rate of inflation and reduce the initial burden to mortgage b orrow ers, it might appear that they are sup erio r to A R M s, w hich do only the first. But this is not necessarily so. PLAM s pro tect lenders from unexpected interest rate increases only to the extent that those in creases are attributable to unexpected in creases in the rate of inflation. ARM s protect lenders from u n expected interest rate in creases regardless of the cause of those in creases. Thus, A R M s provide greater interest rate risk protection to lenders than PLAMs but less protection to borrow ers. 15 T h e o p e ra tio n of th e Table 4 price level-adjusted mortgage Simulations of price-level adjusted mortgages ($60,000 mortgage) plan is shown in table 4. As sume that in the absence of inflation the rate on a 30-year V* expected inflation3 V i expected inflation2 No tilt 1 fix e d - ra t e , fix e d - p a y m e n t Ending M onthly M onthly Ending M onthly Ending mortgage would be 5 percent. balance payment balance Year payment balance payment The m onthly payments on a (d o lla r s ) loan of $60,000 on an $80,000 59,730 571.39 59,391 399.18 1 59,115 322.09 hom e w ould be $322. If the 60,617 347.86 582.82 62,261 62,839 423.13 2 61,480 594.48 borrow er's incom e rem ained 65,209 448.52 375.69 66,725 3 606.37 62,311 68,227 4 405.74 475.43 70,768 at, say, $30,000 per year, the 63,104 618.49 503.96 71,303 438.20 74,959 5 m onthly mortgage payments 63,851 630.86 74,425 79,287 534.19 6 473.26 64,541 643.48 77,576 511.12 566.25 7 83,735 would be equal to 12.9 p er 65,164 80,737 656.35 88,282 600.22 8 552.01 cent of pretax incom e. Ignor 65,708 669.48 83,883 636.23 596.17 92,903 9 ing taxes, if the rate of infla 66,157 682.87 86,987 674.41 643.87 97,562 10 66,496 696.52 714.87 90,012 695.38 11 102,219 tion w ere suddenly expected 66,706 710.45 106,824 757.76 92,920 751.01 12 to accelerate to 8 percent 66,766 724.66 95,661 811.09 111,313 803.23 13 annually, the mortgage rate 66,651 739.16 98,179 875.97 115,612 851.42 14 66.333 753.94 100,409 902.50 946.05 119,633 15 on new loans might jum p to 769.02 65,781 102,272 123,268 956.65 16 1021.73 13 percent in order to m ain 64,959 784.40 103,679 1103.47 126,392 1014.05 17 63,825 800.08 104,523 1074.90 1191.75 128,853 tain the purchasing pow er of 18 62.333 104,684 816.09 130,477 1287.09 1139.39 19 the loan and the m onthly pay 832.41 60,429 104,019 131,057 1207.75 1390.06 20 ment on new loans would 58,052 849.05 102,366 130,349 1280.22 21 1501.26 55,132 866.03 99,535 1621.37 128,071 1357.03 22 rise to $664. Assum ing that 51,590 883.36 95,308 1751.07 123,892 1438.45 23 the new borrow er's incom e 47,337 901.02 89,434 117,427 1524.75 24 1891.16 does not increase im m ediate 919.04 42.270 1616.24 81,623 2042.45 108,231 25 36.270 937.42 71,544 1713.21 2205.85 95,785 26 ly, the payments would rep re 29,206 956.17 58,814 1816.00 2382.32 79,488 27 sent 26.5 percent of incom e. 20,926 975.29 42,994 1924.96 2572.90 58,646 28 994.80 11,256 A sim ilar pattern exists for 23,582 2788.73 2040.46 32,459 29 0 1014.70 3001.04 0 2162.90 0 30 ARMS. The P L A M , h o w e v e r, 'Assum es 5 percent real interest rate and 8 percent inflation. would increase the m onthly ^Assumes 5 percent real interest rate and 8 percent inflation but incorporates 2 percent expected inflation in mortgage rate. paym ent by only the actual 'Assum es 5 percent real interest rate and 8 percent expected inflation but in co rp o inflation realized in the p eri rates 6 percent inflation in mortgage rate. od. If this w ere also 8 percent, the payment would increase by 8 percent from the contract am ount of pal rises sharply through the early and m iddle $322to $348 after 12 m onths. At the same time years of the mortgage and declines sharply in the principal on the loan also would be the later years. This occurs because a large adjusted by the 8 percent rate of inflation. If part of the early payments represents interest. the borrow er's incom e increased at the aver Thus, little of the principal is paid back and age rate of inflatio n , the paym ent would the principal increases by almost the full rem ain at 12.9 percent of incom e. In the next am ount of the rate of inflation. Near the end, period both the mortgage paym ent and in almost all of the m onthly paym ent represents com e would again increase by equal p ercent principal reduction and the outstanding p rin ages so that the mortgage burden remains cipal declines sharply. constant. PLAM s have three potential problem s. As indicated by figure 3, the loan princiFirst, the ratio of the m onthly payment to the 76 Econom ic Perspectives Figure 3. Comparisons of monthly payments and outstanding balances of FRM* and PLAM** monthly payments (dollars) ending balance (dollars) years •Assumes $60,000 mortgage loan with loan term of 30 years and 13 percent mortgage interest rate. ••Assumes $60,000 original mortgage loan with 5 percent real interest rate and 8 percent rate of inflation. incom e of a given borro w er w ill rem ain co n stant only if that b o rro w er’s incom e changes m ore or less in line w ith the particular price index used. If increases in the b o rro w er’s incom e lag substantially behind the index, the mortgage burden w ill increase through tim e and might increase the risk that the bor row er w ill default. The possibility of default might be reduced by choosing an appro priate price index and, possibly, by modifying the paym ents sched ule to reintrod uce some but not all of the tilt. The p rice index to w hich the mortgage paym ents are tied should reflect the ability of households to service their mortgage debt. Th u s, it w o uld be appropriate to use a wage index. But, as noted, everyo ne's incom e does not change by the average am ount. Further m ore, under most circum stances, some in creases in labor productivity might be e x pected through tim e. If so, prices will increase m ore slow ly than wages. Because incom e is the desirable characteristic to track, it would be m ore desirable to use a price index asso ciated with inco m e, such as the GN P deflator, Federal Reserve Bank o f Chicago than a price index associated with a fixed market basket of goods, such as the consum er price index. The m aintenance of some tilt in the mortgage paym ent-to-incom e ratio may be desirable as a means of reducing the risk of default. This can be achieved by indexing only part of the d ifferen ce between the nom inal market and real mortgage interest rates rather than all of the d ifferen ce. The real mortgage rate may be defined as the constant purchasing pow er mortgage rate or the rate that w ould exist if prices at the end of the loan period w ere expected to be the same as at the beginning. Assum e, as before, that the annual rate of inflation over the next 30 years is expected to be 8 p ercent, so that in the absence of incom e taxes the market rate might be 13 p ercent if the real rate is 5 per cent. The mortgage plan might add to the real rate one-quarter of the interest rate com po nent attributable to the expected rate of infla tion. This w ould increase the initial m onthly payment and change the subsequent monthly paym ent and p rincipal amounts by only three-quarters of the actual rate of inflation. For the exam ple above, the initial interest rate w ould be 5 plus 2, or 7 percent. The monthly payments on this “ m o d ified ” PLAM would be higher initially than on the “ p u re” PLAM , but would increase m ore slow ly through time and eventually becom e sm aller. They would also increase m ore slow ly than average in com e, so that for most borrow ers the burden of the mortgage payments would decline. Because the initial m onthly payments are higher, the principal outstanding on the loan w ould not increase as rapidly. The amounts for the $60,000 mortgage loan discussed ear lier aresh o w n in fig u re 4 . Note that, unlike for the pure PLA M , the b o rro w er’s payment-toincom e ratio d eclin es, but by less than for the fixed- or adjustable-rate mortgage plans. The second difficulty concerns the cash flow to mortgage lenders. If the len d er’s prim ary source of funds is regular interestbearing deposits, then a liquidity problem may arise. In the absence of deposit rate ceil ings, an acceleration in the expected rate of 17 Figure 4. Comparisons of monthly payments and outstanding balances of PLAM* and modified PLAMs** 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 years ♦Assumes $60,000 original mortgage loan with loan term of 30 years and 5 percent real interest rate and 8 percent rate of inflation. ♦♦Assumes $60,000 original mortgage loan with loan term of 30 years and 5 percent real interest rate and 8 percent rate of inflation, but PLAM i incorporates 2 percent expected rate of inflation and PLAM2 incorporates 6 percent expected rate of inflation in mortgage interest rate. inflation w ill im m ediately increase the cost of deposits by the full amount of any accom pany ing increase in interest rates. But revenues from PLAM s w ill increase m ore slow ly in line with the actual rate of inflation. O ver the life of the m ortgage, of course, the increases in incom e inflow and exp en d itu re outflow will be equal. Thus, the “ tilt” risk has been shifted from the b orrow er to the lending institution. The institution may protect itself by offering similarly indexed deposits— price level-adjust ed deposits or PLADs— or by tem porarily dipping into reserves to finance the d iffe r ence. Because the institutions are larger, bet ter positioned in the financial markets to b or row funds, and m ore know ledgeable in the area of fin an ce , they may be expected to be better equipped to manage the tilt problem than most mortgage borrow ers. The third d ifficulty arises because most mortgage lenders and borrow ers are subject to incom e taxes. Lenders must share part of their interest incom e with the governm ent, 18 w h ile borrow ers may deduct their interest costs from their taxable incom e. Thus, as mortgage paym ents increase w ith the rate of inflation , the aftertax incom e of taxable len d ers from the mortgage loan w ill increase m ore slow ly and be less than is required to maintain the purchasing pow er value of the payments. At the same time the aftertax cost of the mortgage paym ents to taxable b orro w ers w ill increase by less than is required to maintain the purchasing pow er value un changed. That is, on an aftertax basis, the lenders are not fu lly protected, w h ile the bor rowers benefit. This asymmetry may be co r rected by indexing the mortgage payments and principal payments to the price index by a factor to correct for the tax effects.9 Although the incom e tax brackets of the individual participants in the mortgage m ar ket may be expected to vary, a factor based on the estimated average marginal tax rate for all participants may be a w o rkab le solution in practice. Sim ulations of m onthly payments and principal amounts for a PLAM plan assuming an average marginal tax rate of 25 percent are shown in table 5. In these sim ula tions the aftertax ratio of mortgage payments to incom e rem ains constant w h ile the pretax ratio increases. Shared-appreciation mortgages The recent period of high rates of in fla tion and interest has also been characterized *1 9The appropriate factor for bonds is 1/1-t where t is the marginal tax rate; the factor for amortized mortgages is more complex because the monthly payments are divided between interest and principal. For the ratio of aftertax mortgage payments to income to remain con stant for a PLAM plan, the factor, g', for increasing the outstanding mortgage balance must be less than the inflation factor and is: , 1 - t + t (1+i)-n a — ---------------- -------1---------- where x R 1 - t + t (1+i)-n+1 t= marginal tax rate i = interest rate n = number of payments g = 1 + inflation rate. The derivation of this factor has been provided by Henry J. Cassidy, Office of Policy and Economic Research, FHLBB. Econom ic Perspectives interest must be specified as a percentage, say, 30 or 50 per cen t, of the amount of appre ciation accrued, either at the Ratio of tim e the hom e is sold or at aftertax payments some designated earlier date. to income This interest is received only at that tim e. Thus, shared ap 10.5 preciation represents contin 10.5 10.5 gent deferred interest. Until 10.5 th e s e t t le m e n t d a te th e 10.5 10.5 am ount of this interest is un 10.5 10.5 certain. The other component 10.5 of the interest is determ ined 10.5 10.5 at the origination of the loan 10.5 and is paid regularly through 10.5 10.5 e ith e r fix e d or graduated 10.5 10.5 m onthly payments. 10.5 The two com ponents of 10.5 10.5 the interest payments are re 10.5 lated. The higher the amount 10.5 10.5 of contingent interest expect 10.5 10.5 ed at origination, w hich de 10.42 pends both on the expected 10.4 10.4 am ount of appreciation and 10.4 10.4 on the appreciation sharing 10.4 ratio, the lower the im m e diate interest rate and the corresponding m onthly pay ments. Co nversely, the lower the expected contingent interest, the higher the im m ediate interest rate and monthly paym ents. As eith er the am ount of apprecia tion expected or the sharing ratio declines, the SAM approaches a fixed-rate, fixed- or graduated-paym ent mortgage. Potential hom e buyers may find a SAM d esira b le if th ey can no t affo rd the high m o n th ly paym ents associated w ith other mortgage plans. They w ould be able to defer a part of their interest payments until the necessary liquidity could be obtained through the sale of the house or through an im prove ment in their cash flow situation. Although appealing, the SAM presents a n u m b er of sig n ifican t d raw b acks for the lender, particularly if it is a thrift institution, for the b o rro w er, and for the relevant regula tory agency. W ith respect to the lender, SAMs Table 5 Simulation of PLAM with constant ratio of aftertax mortgage payments to income1 ($60,000 mortgage) Annual income Annual pretax payments 30,000 32,400 34,992 37,791 40,815 44,080 47,606 51,415 55,528 59,970 64,768 69,949 75,545 81,589 88,116 95,165 102,778 111,001 119,881 129,471 139,829 151,015 163,0% 176,144 190,235 205,454 221,891 239,642 258,813 279,518 3,903 4,200 4,518 4,859 5,226 5,619 6,040 6,491 6,975 7,493 8,047 8,641 9,276 9,954 10,680 11,455 12,283 13,167 14,110 15,115 16,187 17,329 18,545 19,838 21,214 22,676 24,230 25,879 27,628 29,483 Ending balance Annual aftertax payments Ratio of pretax payments to income (dollars) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 59,097 62,566 66,155 69,855 73,651 77,529 81,469 85,444 89,426 93,377 97,255 101,007 104,574 107,884 110,855 113,392 115,384 116,703 117,203 116,718 115,056 112,001 107,306 100,692 91,845 80,409 65,983 48,119 26,313 0 3,153 3,405 3,677 3,970 4,287 4,629 4,998 5,397 5,827 6,292 6,794 7,335 7,920 8,552 9,233 9,%9 10,763 11,621 12,547 13,546 14,625 15,789 17,046 18,403 19,868 21,449 23,156 24,998 26,986 29,132 13.0 13.0 12.9 12.9 12.8 12.7 12.7 12.6 12.6 12.5 12.4 12.4 12.3 12.2 12.1 12.0 12.0 11.9 11.8 11.7 11.6 11.5 11.4 11.3 11.2 11.0 10.9 10.8 10.7 10.5 ’Assumes 5 percent real interest rate and 8 percent inflation. 2Change due to internal computer rounding. by very high rates of appreciation in residen tial property values. To the extent that m ort gage lending institutions underestim ated the rate of inflation and, in retrospect, charged a low er fixed mortgage rate than w arranted, a greater share of the appreciation in value accrued to the hom e ow ner. In light of this exp e rie n ce , some mortgage lenders have begun to view appreciation in hom e values as protection against the risk of unexpected inflation and w ould prefer to index their mortgage interest rates to this appreciation. This may be achieved in a shared-appreciation mortgage (SAM ). The SAM is m ore com plex than the other types of alternative mortgages. The interest rate consists of two com ponents. Because the am ount of the app reciation , if any, is un cer tain at the tim e a loan is extended, part of the Federal Reserve Bank o f Chicago 19 do not alleviate the liq uid ity problem arising from interest rate interm ed iatio n. Ind eed , to the extent that the regular m onthly mortgage payments are low er than on other mortgage plans, higher deposit costs from inflation are more d ifficu lt to pay out of cu rren t revenues, even in com parison with fixed-rate m ort gages. The liquidity problem s of the institu tions are exacerbated. SAM s are also likely to increase substan tially the degree of credit risk assumed. The future value of residential real estate is not known and the degree of co nfid ence with w hich values can be predicted d eclines as the time to m aturity of the mortgage increases. Thus, at the tim e of o rig in atio n , the actual return is u ncertain. M o reo ve r, the price of real estate has not always increased rapidly and, over some periods, it has not increased at all. Between 1947 and 1978, the average return realized on residential real estate was about 7 percent per year. This was tw ice as great as the average annual increase in co n sum er p rices, but less than half the average return on the stock m arket. By decades, the return on residential real estate was 5 percent in the 1950s, 6 percent in the 1960s, and 10 percent in the 1970s through 1978. Thus, the double-digit rates of hom e appreciation in recent years w ere not typical of past exp e rience. Most of us know of individual homes or individual neighborhoods that have d e clined in value. If the lend er's expected appreciation is not realized , the im m ediate interest rate charged on the loan w ill prove in retrospect to have been too low and the lender's real ized return may well be lower than on other mortgage plans. D eterm ination of future housing values by geographical and neigh borhood area is also likely to require person nel with d ifferen t skills than those the institu tions currently em ploy. Because of neighborhood effects, two otherw ise identical homes owned by eco nom ically com parable households may have different expected rates of appreciation. Thus, at origination, the same lending institution would charge d ifferent im m ediate interest 20 rates on the two properties. This is likely to lead to resentm ent on the part of the bor row er being charged the higher rate and to charges of "re d lin in g ” if entire n eighbor hood areas are charged higher initial m ort gage rates. Lenders can generally reduce the degree of default risk they assume by diversifying among a num ber of different loans whose default probabilities may not be expected to be closely related. But many thrift institutions make loans in a lim ited geographical area. All the houses in the area may reasonably be expected to be subject to sim ilar underlying changes in value. Thus, any one lender may not be able to diversify this risk away. Real estate is also relatively illiquid so that the estimated market price may not be realized q uickly. In terms of a hedge against u n ex pected inflation, SAM s may not be much less risky than equities, w hich have generally b e e n c o n s id e re d to o ris k y fo r th rift institutions. Because the lender's return is dependent on the am ount of appreciation in the home's value after the loan is originated, anything that affects the potential for appreciation is of concern to the institution. Thus, the lender is likely to require m inim um m aintenance stan dards and to engage in p eriodic surveillance to see that they are met. If the standards are not m aintained, then costly en fo rcem ent and legal actions may be necessary. Capital im provem ents made by hom eow ners generally add value to the house that is greater or sm aller than the cost of the im provem ents them selves. Separation of the appreciation in the value of a hom e between the unim proved hom e and the capital im provem ents is d iffi cult and may also result in controversy and litigation. Lastly, to reduce the liquidity strains, SAM s generally provide for the lender to receive instalm ents on the deferred co ntin gent incom e from appreciation before the final m aturity, say, every five or ten years. (Regulations recently proposed by the Fed eral Hom e Loan Bank Board effectively re quire that the m axim um period for adjust- Econom ic Perspectives merit be no longer than ten years.) If the hom e is not sold before this period ends, interim settlem ent is based on the appraised value at that tim e. The accrued interest may be paid in cash as a lum p sum or be refi nanced by a mortgage not providing for shared appreciation. These features present two potential problem s for the len d er. First, the am ount of the appraised value may not be accepted by the b o rro w er, causing ill feelings, loss of good w ill, and possibly litigation. (O f course, the appraised value at times may be less than the institution w ould like.) Even sale prices may be co ntro versial. A seller may sell the house at an artificially low price to a friend for other com pensation to avoid paying the lender its share. The lender can protect itself against this possibility by reserving a right of first refusal. Second , the refinancing of the accrued interest delays further the cash inflows to the lending institution, intensifying any liquidity pressures. M o re o ve r, if the appreciation is large so that the accrued interest is large, the increase in m onthly payments may cause finan cial strains on the b orrow er at the tim e the mortgage is refinanced . This increases the risk of default on top of any increase in liquid ity pressures. The Federal Hom e Loan Bank Board proposes that the refinancing be guar anteed by the lending institution “ without regard to the . . . b o rro w er’s incom e . . . for a term of not less than thirty years.” The disadvantages of SAM s to the lend ing institutions are not necessarily advantages to the b o rro w er. Because of the lender's interest in the appreciation of the hom e, the b orrow er loses partial control over the man agem ent of the hom e. Changes in the house, capital im pro vem ents, and even color of paint may need to be approved by the lender in advance. Freedom to make decisions of this nature ind epend ently is often view ed as one of the m ajor advantages of hom e o w n e r ship. Disagreem ents with respect to m ainte nance, co ntribu tion of capital im provem ents, and assessed value at interim adjustm ent dates are lik ely to be, at m inim um , nuisances Federal Reserve Bank o f Chicago and, at w o rst, may involve the borrow er in exp en sive, tim e consum ing, and unpleasant litigation. Lastly, SAM b orrow ers e xp e ri encing a large appreciation in the value of their homes and wishing to sell and move may not be able to purchase another home of equal value w ithout financial strain because part of any appreciation must be paid to the lender on sale. Th e reg u la to ry agencies sup ervising banks and thrift institutions are charged with the responsibility for protecting the financial solvency of the institutions. To do so, they evaluate the quality of the institutions’ loans. It is d ifficu lt, if not im possible, to judge accu rately the correctness of the institution’s pre dicted rate of appreciation for a large number of w id ely d ifferen t hom es and, th erefo re, the quality of the loans. M o reo ve r, even to try to do so w ould requ ire a staff with both skills and training different from those that exa m iners cu rren tly possess, as well as substantial fam iliarity with the characteristics of the geo graphical areas served by each lending insti tution. This w ould increase significantly both the cost and d ifficulty of the agency’s task. In light of these problem s, the SAM is probably not a viable alternative mortgage instrum ent at this tim e for depository institu tions. Further w o rk is required to correct some of these lim itations before it can make a major positive co ntribu tion to the mortgage market. Conclusions It is evident that the traditional fixedcoupon rate, fixed-paym ent mortgage is no longer king of the hill. Prim arily because of the dual im pact of rapid inflation and high and volatile interest rates, it no longer serves the needs of all b orrow ers and lenders. Some borrow ers find the high initial m onthly pay ments required by this mortgage an undue burden, w h ile many lenders view making long-term loans at a fixed rate financed by deposits at an adjustable rate as too great a risk. Thus, fixed -rate, fixed-paym ent m ort gages are being supplem ented by a large 21 num ber of alternative mortgage plans tai lored to the sp ecific needs of individual mortgage borrowers and lenders. These m ort gage plans include adjustable-rate mortgages, graduated-paym ent m ortgages, price level- adjusted m o rtgages, sh a re d -a p p re c ia tio n mortgages, and a variety of com binations thereof. All of these mortgage plans differ in at least one com m on way from the fixed-rate, Major characteristics of recent federal regulations governing adjustable-rate home mortgage lending Federal savings and loans and mutual savings banks National banks Requirem ent to offer fixed-rate mortgage instrument to borrower None None Limit to amount of A R M s that may be held None None Any interest rate index that is readily verifiable by the borrower and not under the control of the lender, in cluding national or regional cost-offunds indexes for S&Ls. O n e of three national rate indexes— a long-term mortgage rate, a Trea sury bill rate, or a three-year Trea sury bond rate. M ajor characteristics Indexes governing mortgage rate adjustments Limit on frequency of rate adjustments None Limit on size of periodic rate None Not m ore often than every six months. 1 percentage point for each sixmonth period between rate adjust adjustments ments, and no single rate adjustment may exceed 5 percentage points. Limit on size of total rate adjustment None None over life of mortgage Allow able m ethods of adjustment to rate changes Limit on amount of negative am orti zation Any com bination of changes in monthly payment, loan term , or principal balance. No limit, but monthly payments must be adjusted p e rio d ically to amortize fully the loan over the remaining term. Advance notice of rate adjustments Prepayment restrictions or charges Disclosure requirem ents 22 Limits are set, and m onthly payments must be adjusted periodically to amortize fully the loan over the rem aining term. 30 to 45 days prior to scheduled 30 to 45 days prior to scheduled adjustments. adjustments. Prepayment without penalty per mitted after notification of first scheduled rate adjustment. None Full disclosure of ARM character istics no later than time of loan application. S O U R C F : David F. Seiders, “ Changing Patterns of Housing Finance,” the Federal Reserve System (June 1981), p. 468. Changes in m onthly payment or rate of amortization. Full disclosure of ARM character istics no later than time of loan application. F e d e r a l R e s e r v e B u lle t in , Board of Governors of Econom ic Perspectives fixed-paym ent mortgage— they are consid erably m ore co m plex. This hampers both the design and operation of efficient plans and their acceptance by household borrow ers. Partially in response to this problem , the agencies w ith prim ary resp o n sib ilities for regulating the m ajor residential mortgage lending institutions— the Federal Home Loan Bank Board for savings and loan associations and the C o m p tro ller of the C u rre n cy for national banks— have issued regulations pre scribing the m ajor features and restrictions of these mortgage plans. In itially, the regula tions specified very carefu lly the perm issible characteristics of each mortgage plan, in par ticu la r, the degree of flexib ility in changing interest rates and/or monthly payments. M ore recen tly, the agencies have issued more gen eral regulations with greatly reduced restric tions, in p articular, on rate and payment flex ibility for adjustable rate mortgages. These changes reflect, in part, the current trend tow ard deregulation of firm s and markets and, in part, the b elief that com petition in the m arket w ill foster the developm ent of m ort gage plans satisfactory to both borrow ers and lenders. The more liberal regulations place the burden of developing m arketable mortgage products on the lenders. For exam ple, to the extent that b orrow ers are risk averse and do not wish to assume the total interest rate risk, Federal Reserve Bank o f Chicago lenders w ill need to design plans that share that risk. To the extent that borrowers' in com es do not change as rapidly as interest rates, lenders w ill need to design plans that m aintain m onthly payments unchanged for some time when interest rates change w ith out incurrin g undue default risk from in creases in the value of the loan. It is likely that individual lending institutions w ill design a num ber of differentiated mortgage plans, as is the case with other products sold by a large num ber of producers in a com petitive market. The design of the mortgage plans will also be greatly influenced by the policies established by the Federal Hom e Loan Mortgage C o rp o ration and the Federal National Mortgage Association for purchases in the secondary m arket.10 The sim ulations presented in this article illustrate the major characteristics of the mortgage plansthat have been permitted or proposed and may be useful to institutions in designing their plans. 10The Federal National Mortgage Association an nounced on June 25, 1981, that it would make commit ments to purchase eight types of adjustable rate mort gages based on five different indices beginning in late July. The program announced by the Federal Home Loan Mortgage Corporation in late May was more limited. See Am erican Banker, “ FNMA Unveils Adjustable Rate Mortgage Plan," Vol. 146, No. 125, Friday, June 26,1981, page 3, and “ FNMA’s Adjustable-Rate Mortgage Pro posal,” Vol. 146, No. 132, July 8, 1981, page 2. 23 Public Information Center Federal Reserve Bank of Chicago P.O. Box 834 Chicago, Illinois 60690 Please attach address label to correspondence regarding your subscription.