Has the Deregulation ofDeposit Interest RatesRaised Mortgage Rates?R. Alton Gilbert and A. Steven Holland

EGISLATION enacted in 1980 calls for the gradualphase-out of interest rate ceilings on deposits at banksand thrift institutions by 1986.1 This legislation wasintended to increase the efficiency offinancial markets,which a deregulated financial environment provides,and permit small savers to earn more competitive rateson their savings. Many of these interest rate ceilingsalready have been removed.

Some economists have suggested that the paymentof higher interest rates to depositors has contributed tothe high rates of interest in this country over the last

fewyears. According to Arenson (1983) in the New Yorklimes, “Economists estimate that the higher cost ofbank funds probably has raised the genet-al level ofinterest rates by about 1½percentage points.” Bacon

(1983), in the Wall Street Journal, quotes Lat~’enceChimerine of Chase Econometrics as estimating thesame effect on long-term real rates of interest. Thebasic argument is that the phase-out of Regulation Qhas raised the interest expense of depository institu-tions; in response, these institutions have raised theinterest rates they chat-ge borrowers.

This article assesses the effects of the removal ofdeposit rate (Regulation Q) ceilings on the interestrates chaiged on mortgage loans. While the analysisdeveloped here applies to all interest rates, we empha-size mortgage interest rates because large proportions

R. Alton Gilbert is a Research Officerand A. Steven Holland is aneconomist at the Federal Reserve Bank of St. Louis. Jude L. Naes,Jr., provided research assistance.‘Depository Institutions Deregulation (1980).

of the deposit liabilities of mnajor mortgage lenders,such as savings and loan associations (S&Ls) andmutual savings banks, have been subject to RegulationQ ceiling rates; indeed, one reason for the removal ofthese ceilings was to increase the ability of these thriftinstitutions to attract deposits to use for mortgagelending.z Furthermore, some analysts have suggestedthat such deregulation has caused mortgage rates toincrease more than other long-term interest rates.3

STEPS IN PHASING OUTDEPOSIT RATE CEILINGS

Table 1 describes the steps that already have beentaken in eliminating deposit interest rate ceilings.Many of these steps created new types of accounts,with ceiling rates higher than those on passbook sav-

ings accounts or with no ceilings at all. The first signifi-cant steps in the i-elaxation of Regulation Q occurredeven before the passage of the Depository Institutions

2Thrifts currently hold around 40 percent ot the one- to four-familyresidential mortgage debt in the United States. They originate amuch greater percentage, however, selling a large proportion of theirmortgages to investors in the form ot mortgage passthrough certifi-cates. See McNulty (1983) for a discussion ot mortgage originationand investments of thrift institutions.

3For instance, Edward Friedman (1983), pp. A.40—A.41, of ChaseEconometrics maintains that:

The other ma)or effect of the new deposit structure atthrifts and banks isthe permanent rise in borrowing costs for deposit institution borrowersrelative to open-market rates . . . . The implication is that it, for example,bond rates were to tall to much lower levels, home mortgage rates wouldnot necessarily follow point for point.

5

FEDERAL RESERVE BANK OF St LOUIS ~ 1984

/ ‘ 5,’,’,’ ‘‘64’ A’ ‘/‘/< “,6’/4’/’’ ~/,45y “A‘‘~‘‘A ‘ s “ ‘‘‘A,, ~/‘/~~‘P ,/“ ““A‘ ~ A , ~6’ ~

~fitLt~ A ~ 4 N44’,’ ~ /4 fAr”’ ~~i’ ~ /L~:~

~ ~A’A ~ ~ ‘~‘ 1:’ /~ /7A6// 6’,, ~ // ~< ~N/ ~

~ ~i ‘A/’:’II, /44 <‘A NA

~ ~“AIA ~ <~: :cIA16”~1~?]~~i 1~ ½&//’4,*at~/ 4 //~ / ~

/‘A6:~2~A’/ ‘A446~*~~é// 6< ‘A~!t~7~tA: ~ 1AA’~~,AA’1~/< ~IA’~~:’<41~‘1 “ P::’, ‘1 ,, ~ ‘‘A~,:nts :~:: I.

A’ “’ ‘~‘ ‘sA ‘ ‘,,SS’ A, 4 , , ~ 4,‘6’, ‘A’, ‘5 ,‘ ‘,6,5,~ /A’4 ~‘ ‘ ‘A,, A A’,, , ~ 4N/,‘A ~ ~ ~ “’‘A’~ ,~A4’,<’ ~ ‘

4,’NA64’A~ ~ ‘A’ A~~A’,,’/’A’4 ~~4~A/ >~‘“4’’A A ‘

s;~ ;~,‘A”N~ 4,45~ ‘ ‘54,45 ‘ ‘~‘ A/15A< ‘‘ “‘4’ ~‘A ~ ~,“4 ., ‘ ‘A~~AA/“ ~ ‘4 ‘ ‘,

A, A A’A~ ,N ,‘,‘4” “,“ , / “A ‘1’ ‘<4” /

A/”5’ ‘I4S/”A”’<’ t’A<<’~<” ‘,A/’ A~’’ ‘<‘, ‘/t’,~’”I, <<6’’ c //4

NA / //SN,/ ‘1’>‘A

1 ~ / / ‘~, / As~’<’~ 4S4 / 5 ‘ 44/~ /<ACA,S/45

/

½ / ‘A’” “~

$A1A/4~’A ,<~esç,~ ‘ 1/’A’A ‘A”<, ~ ,,~>A’ N”’/A/ ,,, 1.4,, “/,</Ak,,A’’~ ‘~~:~

- ,~ A A’ A~

/ A / A,4~4/ / , — / / / ,/‘ ,1446’ N A /7S½”’~

,/4 ‘ / /15 / / ‘ ‘‘/4 4* **A~ A’,A / - /*CA” /

5v’A½~A’ ‘444

‘A A/4 A s, / ~ /AA/ A’ ~ :~,:“<~*.,s~,e:4’/~’~’AZE:

/ / 4* / ‘ A’ A’ “/, 4 ‘ ‘~

,E~c”’t AN ~ ~ ~~‘<&‘ ~‘6”’A’’A ‘N ‘A IAfl4,

5” AA’,’<’5’44, A / / / / “

‘A ,‘‘<,½”’ / “ A’A”5A’< A’ ‘/N’A NA’<’A “’$‘(/4*AA, , <“NA ‘~“/ ,

‘AN,’ ‘<A%<’ ‘ ,.‘ ‘‘N ‘‘A ‘/“A’A’ /‘‘~“ N’’, “A“’ 6”~SN: a~~sa1e~wafrm~sAtS.* ~ ‘A

‘‘~‘< ,“‘A N;,1, / ,,, ~ ,,‘~ 55 / 4 ‘5’ ‘ A /

‘<~‘ IN,,,,, “~ ‘: ‘A’ A ~ ,!!~s :1’,’’s/S \N,/’ 44,5 A’ ‘A”A’; ‘A’ “Z ~“’

‘A A’ ‘~ “~ IA~$fI*~ss*~*~‘gf~t”‘A /

‘<‘A’A~””’4”’6” 1, ‘ 4”’’/’,1”,//A’ .4*’ ,,/sc//4T,’~’/’/ ,,A N~’~’, “,,‘ S , / / / 4’ /

/ / SS.mnSMw4W~ IA ~ ‘~ $~S~’ Ü~MOS~S~

‘A 4’ “ 4’/~ A .p*~* / ~,,, / A

/4 ajgiiias tast “1 ‘ 4” /‘

/ /4/ ‘A / I~~”’~N,“ ~ A’ A/6’~47 ~ A”// / / A’ / f” <—,I ,‘

/ ‘ IA A 4/ ~ ‘ ‘4 1 N /

““’““,,‘//,,ft’/’5 ~/4 / ~i” ,‘A ~‘5 ‘ / / .4 44 6/ ,‘ ““‘~ 5/ /4

/4 / “‘ / /44

N4 / ‘A “A 4”/’ ~‘/ 6’,’, “4 /“//‘/ /<0 , / “ // ‘~~/ /4 “.4

6

FEDERAL RESERVE BANK OF ST. LOUIS MAY 1984

Deregulation and Monetary Control Act of 1980 (MCA),with the establishment of money market certificates(June 1978), automatic transfer service accounts(November 1978) and small saver certificates (June1979). The introduction of NOW accounts nationwide(January 1981) was the first major change in depositinterest rate ceilings put into effect under provisions of

the MCA.4

The Depositosy Institutions Deregulation Commit-tee has the responsibility for complete removal of de-posit rate ceilings by 1986. The committee meetsperiodically during the transition period, and most ofthe changes described in table I represent the out-comes ofthese meetings. Currently, the only ceilings ineffect apply to passbook savings deposits and NOWaccounts.5

THE DETERMINATION OFMORTGAGE INTEREST RATES

In analyzing how mortgage rates are determinedand how they might be affected by the deregulation ofdeposit interest rates, we assume that lenders, deposi-tors and borrowers are all wealth-maximizers. In par-ticular-, we assume that lenders attempt to maximizetheir profits, depositors attempt to get the highest in-terest return they can for a given degree of nsk, andborrowers search fot- the lowest interest rates, givenother contractual characteristics of the loan.

We also make two alternative assumptions aboutcompetitive forces in the market for residential mort-gages. Under the first assumption, interest rates onresidential mortgages are determined in a competitivenational mar-ket by the interaction of the total demandfor- and supply of long-term credit. Under the secondassumption, each depository institution has some

market power- that permits it to choose the interest i-ateat which it lends.

In the first case, the phasing out of Regulation Qwould increase the supply of long-term credit, due toan increase in savings by those whose returns from

saving previously were limited by Regulation Q ceilingrates. The increase in the supply of credit would causelong-term interest rates to fall. This is illustrated infigure 1 as a rightward movement in the supply curve

fiom S~to ~2 and a reduction in the rate of interest from

4NOW accounts were available for many years in New England beforetheir introduction nationwide.

5The prohibition of interest payments on demand deposits is notaffected by the MCA.

Fig,,,e I

Effect of Eliminating the Regulation 0 teiliag Rate ona Competitive Market br tong-Term Credit

Interestrates

~I °20 ‘ Demand to, long’tnrm credit

/ Supply of long-teem c, edit betore elimination, of Regulation 0

Supply of long-te,m c,ed’,t otter eI’.minotioe of tegulo,ion c

i1 to L. Of course, the decline in rates may be small; itdepends on the extent to which deposit rate ceilingslimited the incentives for- saving. There would be nochange in the relationship between mortgage andother long-term interest rates, since differences in risk

and liquidity that determine the spreads in interestiates between various types of long-term secur tieswould not be affected by the phase-out of Regulation Q.

The conclusion is not dramatically different if resi-dential mortgages are made by specialized lendinginstitutions that have some market power. Ifa firm with

mat-ket power- raises its mor-tgage i-ate, it will makefewer loans than if it offered mortgage credit at lowerinterest ratesY This is illustrated by the downward-

sloping demand curve (L)~~)in figure 2. We also assumethat the firm must raise the interest i-ate it pays onsmall-denomination deposits if it wishes to attractmore of these deposits~.~~hisis illustrated by the up-wai-d-sloping supply curve ~ In conti-ast, the firmcan attract all the large-denomination deposits itwants by selling certificates of deposit at the rate of

eLenders might have such market power if most borrowers werelimited to borrowing from institutions with offices in their local areaand if the government restricted the number of institutions that mayestablish offices in each area.

-SI

Ii

0

Quantity ($)

7

FEDERAL RESERVE BANK OF ST. LOUIS MAY 1984

o Ceiling Rate onPale Set by a Lender with Market Power

03 03 04

/ Demand to, mortgagesMR / Moeginot revenue

Ss~/ Supply of smolf,denominotion depositsMC

50’ Marginal cost of smoll,ctenomin otioe deposits

interest determined in a competitive national market.With no Regulation Qceilings in effect, we assume thisinterest r-ate is ~

‘the lender maximizes profits by lending the amountof mortgages at which the marginal cost (the increasein total cost due to the last dollar increase in mortgage

lending) equals the marginal revenue (the increase intotal revenue from the last dollar increase in mortgagelending). Marginal cost and marginal revenue are illus-trated by MC (the hea~yblack line) and MB, respective-ly, in figure 2.

The relevant matginal cost curve has two portions:(1) For deposit levels below Q, it is the marginal cost ofattracting small-denomination deposits )MC5~3),sinceMC50 is less than the interest rate on large-denomina-tion deposits, ir~,i. (2) For deposit levels above Q~,it isequal to i’<01. If the lender- wants to attract more de-posits than ft for mortgage lending, it will attract 0,2 assmall-denomination deposits and any additionalfunds as large-denomination deposits. In figure 2, ifthei-e aie no ceilings on deposit rates, the profit-maximizing quantity of mortgage loans is ft with a

mortgage rate of ~M1 and a rate on small-denominationdeposits of i~131.

Suppose regulators impose amaximum interest ratethat may be paid on small-denomination deposits of

i~0~.7The lender will be able to attract only ft of small-denomination deposits and will have to atttact any

additional hinds in the market for large-denominationdeposits. Each lender increases its demand for large-denomination deposits, causing the interest i-ate onthese deposits to rse (to lLuz, for instance). By con-

structing a new marginal cost curve in the same man-ner as before (not shown), we find that the new equilib-rium mortgage rate rises to ~M2, and the amount ofmortgage lending falls to ft. Thus, the theory indicatesthat a binding ceiling on the interest rates paid onsmall-denomination deposits results in a higher in-terest rate on mortgage loans, less mortgage lending,and a higher interest rate on large-denominationdeposits.5 Therefore, the elimination of Regulation Q

MC ceilings should result in lower mortgage interest rates.

Given this conclusion, what are we to make of the

argument that the phase-out of Regulation Q ceilingrates has caused mortgage interest rates to rise? It is anassertion that is inconsistent with standard economic

analysis, which is based on the wealth-maximizingbehavior of business firms and individuals.

WHAT HAS HAPPENED TOMORTGAGE HATES?

We now compare the recent behavior of mortgageinterest rates with movements in other- maiket rates

and the average cost of hinds for S&Ls. The objective isto determine whether- the evidence supports the argu-ment that deregulation of deposit interest rate ceilingshas caused mortgage interest rates to rise relative to

other market interest rates of comparable duration.erhe mortgage interest rate series used is published bythe Department of Housing and Urban Development:

the aver-age interest rate at which residential mortgagelenders make commitments to lend for long-term,fixed-rate conventional loans. The insert on pages 10

and 11 describes several series on residential mortgageinterest rates and discusses the basis for choosing thismeasure.

in the theoretical analysis illustrated in figure 2, Regulation Q ceilingrates are assumed to apply only to small-denomination deposits.This assumption corresponds to the actual structure of ceiling in-terest rates under Regulation 0, which have exempted deposits indenominations of $100,000 or more for many years.

tmThe general conclusions would be the same if all deposits weresubiect to a Regula:ion 0 ceiling rate. Imposing a ceiling interest rateon all deposits that is below the unregulated market interest ratewould reduce the amount of deposits the lender could attract. Theprofit-maximizing lender with market power would raise its mortgageinterest rate to ration the reduced supply of mortgage credit amongits customers.

Fignrn 2

Effect of a Regulationthe Mortgage Interest

ratesInterest

52

Mt

MC-

5Lol

iso’

~502

8

FEDERAL RESERVE BANK OF ST. LOUIS MAY 1984

Chart 1

Semiannual Comparison of Mortgage Interest Rate withCost of Funds to S&Ls and 10-Year Treasury Bond YieldPercent18

1966 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 1983

The yield on 10-year U.S. ‘I g-easury bonds is used as ameasure of the inteiest rate on long-term debt obliga-tions other than residential mortgages.” The 10-year

maturity approximates the average length of tinie thatresidential mortgages are outstanding. This is muchshos-ter than the stated matur-ities of conventionalloans because of the prepayment of a substantial num-

ber of mortgage loans before their matut-it’v.

Chart 1 indicates that semiannual averages of thecost of funds to S&Ls, the mortgage interest rate, andLhe yield on 10-year U.S. Treasury bonds tend to movetogether over time.’° The relationship between

changes in the mot-tgage and bond rates is somewhatcloser- (correlation coefficient of 0.897) than betweenchanges in the mortgage rate and the average cost offunds (correlation coefficient of 0.816).

All three series were substantially higher in the late1970s and 1980s than they had been earlier. ‘thus, thephase-out of Regulation Q ceilings allowed S&Ls to bidfor funds by offering rates that kept pace with rises inmat-ket interest rates. One indicator- of how rising mar-ket interest rates and the phase-out of Regulation Qaffected the average cost of funds for thrift institutionsis the decline in the share of their deposit liabilitiesheld in the fos-mof passbook savings deposits. Between

TMMayer and Nathan (1983) use the 10-year Treasury bond rate for thesame purpose.

‘°Theaverage cost of funds for S&Ls, obtained from the FederalHome Loan Bank Board, incorporates not onlythe interest S&Ls pay

on deposits, but also the interest they pay on advances from theirFederal Home Loan Banks and other borrowed funds. The averagecost of funds is somewhat higher than the average interest rate paidon deposits.

1

1

Percent18

0TTTTTTTTTTTTTTTTJ~ThX 0

9

/C CC I’ ,11’C C-/, ~ 1’~’ I ‘‘ ‘ C ,C, C, ~iC C CICC~1C’C-z2G ’/1CCC~CC:)

1I

CC-/CC C C ~‘ ‘ C\CCC/C CCCX CC/CC CC, / Ik~C ICCC CCvKCC C-~//

CCC ~ CCC CCC/// ~CCC C~C~CIZ K~//~6C CCC

//CCCC/ I CCC CCC C, CC C ‘CC C , /C

C Clv ~1~\ C CC CC//C/C/CC-C CC /C CC/CC/1 CC ~‘ C CC C CC u, CC

1C C~ CCC CC/C ~ ~

CCI C ~ ~ ~ C C C C /CI/ C-C ~ C- C CCC’ CC/ ~ C/~ C AC CC

C CO C C\ ‘C~ ~ C C ~ C C C /C CC C / /

C ~CCC ~CC C ~ ~/ I CCC ‘ CCC K C CC C C

1C/ CC C ~: ICC /CCC / CC C C ICC CCCC C/

~C C~C ~I C

~ 11 ~ C C 111/~CC C ~ /~CC~~C CC- ~ C CI

~ C I C C-/ C: C C CCvC/~// C ~ CCICC C

CCII CC C C~C C /1CC/CC ~ I CC CC cC/-C C

C CC CC Ct/C ~<~C C ~ CCC ‘CC C C/C:C~IC~I C~ ~ ~, CCC

CCC ~ C ~ C~ CCC C SC-~v-~CC CC /CCCCCC C CICCC SIC/C C ICCCC CC 1//C C C I C I C , C C~C- CC //

C C C CCC

CC CI CC C/C CC>:

C ~ 11 /1CC/C\ C CC. . CC ~C CCCC v ~IC ~~///C C vC~ I’/C

/C/C/

C C C ~ CCC / C////C CCC

Cv’ C C/C /CC/CC

C~ CCCCCC <IC

CCC

//C- ~I C

/ CC—C ‘~C1C ~

CCI

:6 ~// CC C C

C CC~~/ C CC ~ CC C/ v~CC~ CCC

C / v111

’~’ICC~~_a~ C CC CC CC C1~/ CCCCC11~CCCI ICCCICCC%IC C C C

C ,CC IC C CCC’

~CCCC&~C ~ C C<CCC C~ CC~I: C C CI CC/6C C /11< ~C ICC ~ CCC: C ICC

III%~11CC C C C C ::CCCCCCICCCCCC C~ C \CCCICC11ICCCCCCK ~ ~C K/C

CC~ CI ‘ CC CCC CC C C C C ~ CC CIIC CCIC/CC//C C/CCCC CC

~C C CCC ~‘CC/ CCC CCCCCC CCC C/~ ~ ~ CC//CC, C C C CCCCCC6CCC/CC / C

ICC11C/CCC/ CCC C C C ~ C’ 1/ CC CC CCC C C // CCC CCCC CCC CC CC,

: . I CC : CCC’ ~ CC’ C ~‘~ C CCC CC ~ C ~ CCC C CCC ,~

KCC C C

CC CC C~ ~ CC C CC C C C ~C-* ~ CC CCCI CC C >(~~ C’ C C CKC C / C~ C~C C ‘ C-C-~ ~ ‘I ‘I C I CCC4C C C 4//ICC” C>I~CC CC C~’/ C- /,CC C \/C C C C ~ ~ ~,

~CCC/IC C/ N~CC CC ~ C CCC/ CCC C’ CC~/C C/C~ I CC C /CCC C CCC C ~ C/CC C CC CCCCCCCCCC C CC/C CCC ~ C~ /C C ~IC /C C / CC

C C C/C C CC CCCCCC CCCC/CCIC,CC C CC CC> CC11CCC ~ CC

/CCCCCCC C C C CC C C ‘I C C IC

CCC~1I/C C C-11ICCCCCC C CCC CCCCC/// /CCC : :~1 I “ ~ C/ C IC CC

CC CC C CCC C arae” ~tt~ CC CC

/ CCC CC> CC C~O/S~SCCCC C/C CC C

C~C ~ CC C CCC/ CC/ ~ C CC CCCC ~ C/CC CC C / / C C ~gC C ICC CC CC ~ CC- ~v CC ~,C ~/C ~~~CCCCC

~CCC’ CCCCCCCCC CC CC/CC CC~C/y ~_/C C/I/CC CCCCC~C C/Cl/C/CC : CC / CC C C/ CC ~ CC4t*t~C CC 4ø~~~$11:

1/CC CC

CCK ,~ ~K CC C ~ ~ /‘C C/CC CC tot CC

:: ~ C-CC CC/C/C C:C~CCCC C/C//CC C/CC. IC>/C C C C sass.ICC/CCC/C

;C ~ ~ CCCCCCCCCICCCCC ~~/C’ CC/CC/CC CCC C C CC CC/CC C/ CC CC,CCC~, C/Cl/ICC/C CC/CC I CC/ C/CC CC~/~ ~ ‘C ~ ~/C CCC C C ~ ~C

~i:~:11: 2:1CC/CC CCVCCCC ~ I CC ~ IC

~ CC 1111 ~

FEDERAL RESERVE BANK OF ST. LOUIS MAY 1984

Chad 2

Selected Mortgage Rates

0

1978 and 1983, savings deposits (subject to fixed in-terest rate ceilings) fell from over 35 percent of totaldeposits to less than 15 percent. Meanwhile, the newmoney market certificates and money masket deposit

accounts each grew to represent about 17 percent oftotal deposits.

Chart 3 plots the same three interest rate series on amonthly basis since May 1979.” The relationshipsamong the three series enable us to see that changes inthe cost of funds to S&Ls cleasly lag changes in themortgage t-ate and the Treasury bond rate, usually byabout two months. A simple statistical analysisconfirms the visual pattern in chart 3. The contempo-raneous correlation between changes in the cost of

funds and the other two series is actually negative,though not statistically significant. However, the cor-

‘1See Chamberlain, Olin and Mckenzie (1983) for a discussion of themonthly cost of funds data. This series is actually the median cost offunds rather than the average.

relation between the current change in the mortgagesate and the change in the cost of hinds two monthslater is 0.61212

The Rise of Mortgage Rates Relative to

Other Long-Term Interest Rates

The behavior of mortgage rates since 1980 appears tolend empirical support to the hypothesis that dereg-ulation has resulted in higher mortgage rates relative toother long-term rates. The average spread between themortgage rate and the 10-year Treasury bond rate from1966 to 1979 ranged generally from 1 to 1.75 percentagepoints; in the 1980s, it has ranged from 2 to 3 percent-

age points.

‘2The contemporaneous correlation between changes in the mort-gage rate and changes in the yield on 10-year Treasury bonds is0.794, indicating that interest rates on both kinds of long-term debtinstruments are affected simultaneously by the same credit-marketinfluences.

CCLT ~ ~

1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983NOTE, FHLMC c omtnitmmtt eats begint in .&tgntst 978. Nt.88 commifmentroft seas clnatged alter Augunt 983.

11

FEDERAL RESERVE BANK OF ST. LOUIS MAY 1984

Chart 3

Monthly Comparison of Mortgage Interest Rate withCost of Funds to S&Ls and 10-Year Treasury Bond Yield

Since 1980, however, the average spread between themortgage rate and the avet-age cost of funds tbr S&Ls

also has increased, by roughly the same order of mag-nitude as the increase in the aver-age spread betweenmortgage sates and the i-ate on lo-ear-Treasus bonds.

The gap between mortgage interest r-ates and the aver-age cost of funds stayed mostly between 2 and 3.5percentage points before 1980; since then, it has variedbetween 3.25 and almost 6 percentage points. There-fore, the widening in the spi-ead between mortgager-ates and the Treasury bond i-ate does not appear to bethe result of a higher average cost of funds to S&Ls.

Why, then, did mortgage rates rise relative to rates

on Treasury bonds of compasable term to maturityafter 1980? The answer appears to involve differencesbetween conventional residential mortgages andTreasury bonds as debt instruments. ‘the two majordiffes-ences are: II) Most mortgages allow the borrower

to pay off his debt befos-e maturity without penalty; and(ZiTher-c is risk of default on mortgage loans. Ti-easurybond holders face neither prepayment risk nor default

risk.

Mortgage Rates and thePrepayment Option

Investors must be compensated with higher interestrates on residential mortgages than on Treasury bondsto compensate for- the r-isk of prepayment by debtors.’1

Mortgage borrowers must pay a higher interest rate forsuch a call option.” The value of this option need not

remain constant over time. In particular, its value willbe higher during periods of more volatile long-term

‘3For a more thorough analysis of the role of the prepayment option indetermining the spread between mortgage interest rates and Trea-sury bond rates, see Hendershott, Shilling and Villani (1982).

1979 1980 1981 1982 1983

12

FEDERAL RESERVE BANK OF ST. LOUIS MAV 1984

interest rates than during periods of stable rates, be-cause of the incseased likelihood that the ps-epayment

option will be exercised. Long-term interest rates wereextremely variable by historical standards from 1980 to1982. Thus, we would expect mortgage rates to riserelative to Treasury bond rates during this period.

The size of the interest rate premium necessary tocompensate investors for the prepayment option onresidential mortgages can be gauged by examining thespr-ead between the yield on passthrough securitiesissued by the Government National Mortgage Associa-tion IGNMA} and the yield on 10-year Treasury secu-rities. The risk of prepayment is the major- differencebetween investing in GNMA passthr-oughs and Trea-

sury bonds. Investors who purchase these pass-through securities receive the interest and principalpayments from a pool of FHA-VA government-guaranteed residential mortgages. Thus, there is nomore risk of default on the interest and principal pay-

ments on GNMA passthroughs than there is for bondsissued by the U.S. Treasury. Any prepayment of themortgages, however, is ‘passed through” to theholders of the passthrough securities.’4

This feature reduces the probability of a capital gain

on GNMA passthrough securities compared with aninvestment in Treasury bonds. If long-term interestrates decline after an investor buys Treasury bonds,their market value rises, and the investor receives a

capital gain if he sells them. In contrast, if long-terminterest sates decline after an investor buys GNMApassthrough securities, the mortgages in the invest-ment pool are more likely to be prepaid. Because suchprepayments i-educe the size of the potential capitalgain, a premium in the form of a higher yield on mort-gage passthr-oughs is required to make investors indif-ferent between them and Ts-easusy bonds.

Chart 4 indicates that the spread between yields onGNMA passthrough securities and 10-year- Treasurybonds rose during 1980 through early 1983. Thus, onereason for the relative mci-ease in mnos-tgage interestrates during those years was a rise in the rate premiumnecessary to compensate investors for the prepaymentoption on moi-tgages.

‘4Another factor that accounts for a small portion of the spread be-tween rates on GNMA passthrough securities and Treasury bondsis the effect of stale and local taxes. Interest earned on Treasurybonds is exempt from state and local taxes, but earnings on mort-gage passthroughs are not. There is no reason to suspect that thisfactor has increased in importance during recent years. There alsocould be differences in yields on these two assets if investors do notview them as being of roughly equal term to maturity, as we areassuming.

Mortgage Rates and Default Risk

Anothet- reason for the rise in intes’est rates on con-ventional mortgages since 1980 appears to be a generalrise in interest rates on privately issued debt securitiesrelative to yields on securities issued or guaranteed bythe federal government. Table 2 shows that the aver-agespread between interest rates on privately issued debtinstruments and “t’reasury securities is higher in thegenerally recessionary period, February 1980 toNovember 1982, than in the expansionary period, April1975 to January 1980.15 This is a reflection ofthe greaterdefault risk associated with privately issued securitiesduring recessionary periods. In each case, the differ-ences in mean spreads between the time periods arestatistically significant at the 1 percent level.’” Thepattern of spreads between mortgage and Treasurybond rates is very similar to the pattern of spreadsbetween yields on other privately issued securities andTreasury securities of comparable duration.

‘rable 2 also indicates that the spreads betweenyields on privately issued and U.S. Treasury securitiesdeclined to near their- ps-c-I 980 levels a few monthsafter the economic s’ecoveiy began in December 1982.The decline in the spread between the mortgage com-mitment rate and the et~reasurybond rate occurreddespite the authorization of money market depositaccounts — a majos- relaxation of Regi,rlation Q ceilingrates that occuri-ed in the first month of the currentrecovery.

These observations are supported by the behavior of

delinquency mates for mortgages. The percentage ofconventional mortgages with payments delinquent for60 days or more rose steadily from 0.61 per-cent in thesecond quarter of 1979 to 1.37 percer~tin the first quar-

ter of 1983, then began to decline. Delinquency rates inthe 1980s have been substantially higher than in theperiod 1964—79, which undoubtedly accounts for a

substantial pos-tion of the higher- n~ostgagerates rela-tive to Treasury bond rates observed since 1980.”

“The period from July 1980 to July 1981 is officially classified as aneconomic recovery. The financial markets, however, did not re-spond during that period as they typically do during expansionaryperiods. Stock price indexes were little affected, and the spreadbetween corporate Baa and Aaa bond rates (known to be influencedby cyclical factors) changed little. The lack of response is un-doubledly due to the weakness and short duration of the recovery.

leSome corporate Baa bonds grant a call option to the issuer. Part ofthe increase in Ihe spread between the Baa bond rate and long-termTreasury securities, therefore, is accounted for by an increase in thevalue of this prepayment opfion.

‘7The average quarterly delinquency rate (60 days or more) for con-ventional mortgage loans between /1964 and V/I 979 was 0.58percent; between 1/1980 and IV/1983, it was 1.01 percent. Thisdifference is statistically significant at the 1 percent level.

13

FEDERAL RESERVE BANK OF ST. LOUIS MAY 1984

Chart 4

Selected Interest Rate Spreads

/ \\\,. //~~ 5. /555 \~, 4 \,,\ 5 544

/5 5”.r’, 4// \, ///‘\/<~/~ / /5 / ~ (,~51 , 5555

‘~r~bIa~5/ 5/ 5 “.5 5 5 5

$preadsS~\n*est$Itetrwate1y’.\ 00* IThStTUm*M* rnnd’.~.’.treasury$i~UriUes4mânthI~aveis spteads 1~percentage POints)intevIS$13a1e’$preaø ‘.‘., ap~t9tk3an980 Fe~1$S44ei~$&t ta 982h1u1%e IS Mge I$34~Y~’.I984

M.ont~a*ecates~f~ttsa s s s*nrTasuryM~M~ ‘.I4~ 7’.

tømpq~afröaaS S

67 90 2 ¶38

4smnommerSpapør ‘.5 5’.’. ‘.55’. 55

/ ttsplltitis Sonflflreastfl ‘.5 5’. 55 5’. 555 5

$3 . ‘.‘.* 026.

5555 5 55 5 5 5’. 5 5 55

1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983

14

FEDERAL RESERVE BANK OF ST. LOUIS MAY 1984

The effects of the major factors that appear toaccount for the s’ise in mortgage rates relative to Trea-sury bond rates can be seen in table 3 (and also in chart41. For- the period 1980—82, the premium to compensate

for the risk of prepayment (approximated by thespread between the yield on GNMA passthi-ough secu-rities and 10-year ‘treasury bondsl was about 73 basispoints higher on average than in 1975—79. The defaultrisk premium on privately issued securities notguaranteed by the government (approximated by thespread between interest rates on new conventionalresidential mortgages and the yield on GNMA pass-through securitiesl was approximately 50 basis pointshigher on average dusing this period. Therefose, botheffects appear to share in the responsibility for highesmortgage interest rates s-dative to Treasury secui-itiesin the early 1980s. Both have declined during the cur-rent economic expansion.

CONCLUSION

Economic theory suggests that the der-egulation of

deposit intes-est rates does not cause mortgage rates torise and may, in fact, result in lower mortgage interest

rates than would otherwise be observed. Nonetheless,many believe that the higher aver-age cost of obtainingloanable funds that results from deregulated deposit

r-ates have led to higher- mortgage rates.

Since the intsoduction of new types of deposits with

flexible inter-est ceilings (or no ceilings at alli, the aver-age interest r-ate on mortgage loans, the average cost offunds for- savings and loan associations, and market

interest rates in genes-al have risen substantially. Thenotion that higher mortgage s-ates are clue to the re-moval of deposit interest rate ceilings, however, is not

supposted.

Although mortgage rates have moved higher relativeto government bond rates of similar duration following

the beginning of deregulation, that pattern appears tobe unrelated to the deregulation of deposit rates. In-

stead, it was the result of more variable interest rates,which caused a higher premium for the option of pre-

paying a mortgage loan, and the economic downturnin the early 1980s, which raised the premium for therisk of default on mortgages. Since interest rates havebecome less variable and an economic expansion hasbegun, the spreads between mortgage rates and gov-ernment bond rates have fallen over the last year toclose to their pre-1980 level.

REFERENCES

Arenson, Karen W. “Why High Interest Rates Are Here to Stay,”New York Times, October 9, 1983.

Bacon, Kenneth H. “Flexible Interest Rates May Add to Stability,”Wail Street Journal, November 7. 1983.

Chamberlain, Charlotte A., Virginia K. Olin and Joseph A. Mcken-zie. “Monthly Cost of Funds,” Federal Home Loan Bank BoardJournal (January 1983), pp. 34—37.

Depository Institurions Deregulation and Monetary Control Act at1980,5. Rept. No.96-640,96 Cong. 2 Sess. (GPO, 1980), title II,sec. 202(a).

Friedman, Edward A. “Key Consumer Markets of the 1980s: Autos,Energy, Finance and Health Care,” in Long-Term Special Report:Consumer Markets in the 19805 (Chase Econometrics, July1983).pp. A.29—A.46.

Hendershott, Patric H., James D. Shilling and Kevin E. Villani, ‘TheDetermination of Home Mortgage Rates: Empirical Results for the1975—81 Period” (paper presented at a joint session of the meet-ings of the American Finance Association and the American RealEstate and Urban Economics Association, New York, December29, 1982).

Mayer, Thomas, and Harold Nathan. “Mortgage Rates and Regula-tion 0,” Journal of Money, Credit and Banking (February 1983),pp. 107—15.

McNulty, James E. “Secondary Mortgage Markets: Recent Trendsand Research Results,” Federal Home Loan Bank of Atlanta Re-view (December 1983), pp. 1—5.

15