Inflation and The Housing Market: Problems and Potential Solutions

Inflation and the
Housing Market:
Problems and
Potential Solutions
Donald Lessard and Franco Modigliani*

I. INTRODUCTION
Recent years have witnessed increasing rates of inflation accompanied
by high and volatile interest rates. Although these factors have affected
the entire economy, their most drastic effect has been on housing as
shown by wide swings in construction activity and in the turnover of the
existing stock of housing as well as by a growing feeling that adequate
housing is out of the reach of an increasingly large number of households.
The conclusions of the M.I.T. study are that (1) these effects can be large-
ly traced to shortcomings of the standard mortgage and the institutional
arrangements that surround it which, in an inflationary environment, have
had a serious destabilizing impact on both the demand for and supply of
housing and that (2) this instrument, in many ways obsolete, should be
supplemented by alternative mortgage designs.
Given the persuasive case that the standard mortgage instrument is a
major culprit, the study examines a variety of possible modifications of
the traditional mortgage in order to assess the effectiveness of alternative
designs in reducing or eliminating the demand and supply effects resulting
from inflation and its variability. The alternatives examined include de-
signs which have been advocated within the United States or actually im-
plemented either here or abroad as well as a set of novel designs aimed di-
rectly at the two major types of inflationary effects.
Outline of the Study

The study was broken down into five subtasks, each of which appears
as a paper in this volume.
The paper by Cohn and Fischer provides a detailed description and
microeconomic analysis of the major alternative mortgage instruments
from the perspectives of both borrowers and lenders. The types of mort*
gages considered include variable interest rate mortgages, which resolve
*Donald Lessard is an Assistant Professor of Management and Franco Modigliani is

an Institute Professor of Economics and Finance, both at the Massachusetts Institute of
Technology.
13
14 NEW MORTGAGE DESIGNS
the lender mismatch problem but in their standard form do not eliminate
the inflation-related distortions in the time pattern of payments; gradu-
ated-payment mortgages with a fixed graduation geared to expected in-
flation, which do adjust the time stream of nominal payments for antici-
pated inflation but have no flexibility to cope with subsequent changes,
and do nothing to resolve the supply problem; price-level-adjusted mort-
gages, where the outstanding principal is adjusted in line with changes in
the price level, which address both problems; and a class of novel mort-
gage designs which are roughly as effective as the price-level-adjusted
mortgage but which are deemed to be more easily implementable within
the current institutional setting.
Kearl, Rosen, and Swan review existing empirical evidence regarding
the potential impact on the demand for housing of changes in the mort-
gage instrument. They find general confirmation that basic elements of the
mortgage do matter, though are led to conclude they find that existing
studies are inadequate to provide quantitative information about the likely
impact of the various proposed changes.
Experience in six countries with alternative types of mortgages is re-
viewed by Lessard with the collaboration of Anderson, Cohen,
Cukierman, and Kouri. The countries studied include the United King-
dom and Canada, which employ variable-rate mortgages with level money
payments; Brazil and Israel, which have adopted price-level-adjusted
mortgages; Sweden, which has combined variable rates with a time stream
of patterns tailored to remove inflationary distortions; and Finland, which
has a hybrid scheme lying somewhere between that of price-level index-
ation and variable nominal interest rates.
Jaffee and Kearl present an examination, through simulation, of the
macroeconomic impacts of various mortgage innovations. Relying on the
MPS econometric model they estimate how construction activity, the
profitability of thrift institutions and other related variables would have
fared over the last ten years if the traditional mortgage had been replaced
by a number of alternative mortgage designs.
Various tax, legal and regulatory barriers to innovation in the mort-
gage instrument are examined in the final paper by Holland.
The remainder of this paper provides an overview and synthesis of the
results of the five studies.
II. THE CAUSES OF RECENT INSTABILITY IN THE

HOUSING SECTOR
The recurrent crises which have plagued the housing industry in the
last decade can be largely traced to the interaction of a rising and variable
rate of inflation with two major institutional features which have char-
acterized the financing of housing in the United States in the postwar pe-
riod. These are (1) almost exclusive reliance on the traditional fully amor-
tized, level-payment mortgage as the vehicle for financing the acquisition
PROBLEMS AND SOLUTIONS LESSARD-MODIGLIANI 15
of single-family houses; and (2) overwhelming dependence for mortgage

funds on thrift institutions which secure the bulk of their funds through
relatively short-term deposits. This framework could and did work reason-
ably well in the period of relative price stability that prevailed until 1965,
but has been a source of serious problems in the environment of rising
and variable rates of inflation which have prevailed in the last decade
through their effect on both the demand by potential buyers and the
supply of mortgage funds.1
A. The Effect of Inflation on Demand: Shortcomings of the Traditional
Mortgages
Our conclusions that inflation has an unfavorable effect on the

demand for houses financed by mortgages and that fluctuations in the
rate of inflation tend to lead to corresponding fluctuations in construction
activity rests on the following considerations which are spelled out in the
rest of this section.
1. Inflation and the anticipation of its continuation tends to raise in-
terest rates, including mortgage rates, by an "inflation premium"
needed to compensate the lender for the anticipated erosion in the
purchasing power of his claim. The rise in interest in turn raises
the annual payment needed to acquire a house of given value.
2. This higher interest rate and resulting annual payment do not per
se change the real cost of carrying a house in that they are offset
by the gain to the debtor resulting from the gradual decline in the
purchasing power of his debt and of his annual payment.
3. Nonetheless the rise in interest rates resulting from inflation has
an important effect on the time profile of the stream of annual
payments, expressed in terms of constant purchasing power.
Whereas in a world of constant prices these payments are con-
stant over the life of the mortgage, the inflation-induced increase
in interest rates results in an increase in the level of real payments
in the early years of the contract with a commensurate reduction
in the later years.
4. In a world in which the household’s ability to meet the annual
payment is constrained by its current income (there being no sig-
nificant opportunities for second mortgages and the like) the in-
crease in the annual payment in the early years of the contract is
bound to have an unfavorable effect on the demand for housing
by forcing many households to postpone or forego home-
ownership or scale down their demand.
These propositions are illustrated by Table 1 and Figures 1 and 2.
Column 2 of the table shows the effect of alternative rates of inflation on
the annual payment for a $20,000 30-year mortgage. Assuming a 3 percent
~Most studies of inflation and housing have focused on the supply effects. Only Poole
[1972] and Tucker [1975] have addressed the demand effects in any detail.
Table 1
EFFECT OF INFLATION ON
THE STREAM OF PAYMENTS
FOR DIFFERENT RATES
OF INFLATION
$20,000 -- 30-Year Mortgage
$10,000 Initial Annual Income Increasing at Inflation Rate
(1) (2) (3) (4)
Annual Real Payment/
Year Payment Payment Income (%)
Case A: 0% Inflation 3% Interest Rate
1 1,020.39 1,020.39 10.00
5 1,020.39 1,020.39 9.24
10 1,020.39 1,020.39 8.37
15 1,020.39 1,020.39 7.58
20 1,020.39 1,020.39 6.87
25 1,020.39 1,020.39 6.22
30 1,020.39 1,020.39 5.63
Case B: 2% Inflation 5% Interest Rate
1 1,301.02 1,275.52 12.50
5 1,301.02 1,178.38 10.67
10 1,301.02 1,067.30 8.79
15 1,301.02 966.68 7.22
20 1,301.02 875.56 5.94
25 1,301.02 793.02 4.88
30 1,301.02 718.26 4.01
Case C: 4% Inflation 7%Interest Rate
1 1,611.73 1,549.74 15.21
5 1,611.73 1,324.72 12.04
10 1,611.73 1,088.83 8.99
15 1,611.73 894.94 6.73
20 1,611.73 735.57 5.03
25 1,611.73 604.59 3.76
30 1,611.73 496.93 2.81
Case D: 8% Inflation 11% Interest Rate
1 2,300.49 2,130.01 20.91
5 2,300.49 1,565.68 14.28
10 2,300.49 1,065.59 8.87
15 2,300.49 725.21 5.50
20 2,300.49 493.57 3.42
25 2,300.49 335.91 2.12
30 2,300.49 228.62 1.32
Assumes 2% real growth in income
16
Figure 1
REAL VALUE OF MONTHLY PAYMENTS
150
No Infl ation
100
4% Inflation
8% Inflation
5 lO 15 20 f25 30
Years Elapsed
Source: Donald Tucker, "The Variable-Rate Graduated-Payment Mortgage"

Real Estate Review, Spring 1975.
]?
interest rate in the absence of inflation, the annual payment is $1,020. As

the inflation rate rises to 2, 4, and 8 percent, raising the mortgage rate by
corresponding amounts, the annual payment is seen to increase by 30, 60,
and 130 percent respectively.
The reason for the higher annual payment is that the payments are
spread over a long period of time and, in the presence of steady inflation,
these payments are made in dollars which are worth less and less in terms
of purchasing power. This proposition is illustrated in column (3) of the
Table, which expresses the annual payment in dollars of "constant pur-
chasing power." This column is obtained by dividing the figures of col-
umn (2) by the price level relative to that prevailing in the year the con~
tract was initiated, which is implied by the assumed rate of inflation for
each of the years indicated in column (1).
In Case A, where no inflation is assumed, the figures of column (3)
are of course identical to those of column (2) with stable prices, a stan-
dard mortgage calls for a stream of payments which is constant both in
current dollars and in terms of purchasing power.
In Case B, with a 2 percent rate of inflation, the payments of column
(3) decline at a rate of 2 percent per annum; thus while they start higher
than in Case A, they end appreciably lower, with the terminal rate of pay-
ment only about half as high as the initial rate. This effect of inflation in
"tilting" the real stream of repayments becomes more and more pro-
nounced as we move to 4 and 8 percent rates of inflation in Cases C and
D. In this last case, the payments start twice as high, but end up one-fifth
as large.
This tilting effect of a rising rate of inflation on the stream of annual
payments expressed in constant purchasing power is brought out vividly
in Figure 1 which shows a graph of the real payment required in each
year of the contract (the information reported in column (3) is only for se-
lected years), for zero inflation, 4 percent inflation, and 8 percent
inflation.
The Level of Inflation and the "Real" Cost of Housing. While the
payment streams corresponding to different rates of inflation differ radi-
cally in shape, they do have one feature in common: the present value of
each of the payment streams measured in dollars of constant purchasing
power is the same -- $20,000 -- when discounted at the rate of 3 percent
which we have assumed represents the interest rate which would prevail in
a world of no inflation (and hence is appropriate to dollars of constant
purchasing power).
It is precisely in this sense that the higher rate of interest and the
higher initial level of money payments resulting from inflation merely
compensate the lender but do not per se increase the real overall cost of
acquiring a house. The same conclusion may be arrived at in a different
way. The cost of owning and using a house for a determined period con-
sists of the outlays to acquire the house less the value of the house when
sold. As long as the value of the house maintains a reasonably close cor-
respondence to the general price level (or better yet exceeds it), the in-
flation premium paid to finance the house will be recaptured through an
Figure 2
REAL VALUE OF OWNER’S EQUITY AND UNPAID BALANCE
Owner’s Equity
Unpaid Balance
825,000
20,000
15,000
10,000
5,000
8% Inflation
5 10 15 20 25 30
Years Elapsed
Source: Donald Tueker, "The Variable~Rate (~raduated-Payment Mortgage"

Real Estate Review, Spring 1975.
]9
eventual capital gain. In fact, taking into account the assymetric tax treat-
ment of interest charges (fully deductible) and capital gains on a primary
residence (totally exempt if reinvested in another residence and taxed at
the capital gain rate otherwise), inflation should actually lower the real
cost of home ownership.
The Level of Inflation and the Ability of Households to Purchase
Housing. Even though inflation does not increase the sum of discounted
payments, it will have an effect on the value of housing which a house-
hold is able to acquire, for this depends not only on the sum of payments
but also on their time profile. The typical household must meet payments
from current income and lenders generally limit the size of a mortgage in
order to maintain a desired payment-to-income ratio in the early years of
the contract. Thus, the amount of housing which a household can acquire
will be limited by its current income and the fraction thereof it can devote
to housing.
It can be seen from Table 1 that a household with an annual income
of say $10,000 and a mortgage of $20,000 could, in the absence of in-
flation, service the debt throughout the life of the mortgage with 10 per-
cent of its income; with a 2 percent inflation, the initial payment would
require over 13 percent of his income; with 4 percent inflation nearly 16
percent; and the figure would rise to nearly 23 percent if inflation reached
8 percent. Furthermore, as is apparent from Table 1 and Figure 1, with in-
flation the traditional mortgage will require higher real payments through
most of the first half of the contract than would be required in a world of
no inflation for which the mortgage instrument was designed. Looked at
from a different angle, the traditional mortgage requires of the borrower
quite different time shapes of repayments of his "real" debt, depending on
the rate of inflation. This point is illustrated in Figure 2, which compares
the behavior over time of the unpaid balance, measured in terms of pur-
chasing power, for alternative rates of inflation. As one would expect
from Figure 1, a higher inflation results in a more rapid decline in the
outstanding debt. Correspondingly, the owner’s equity also builds up
more rapidly, if the value of the house remains constant in real terms.
Our conclusions about the unfavorable effects of the tilting induced
by inflation are reinforced by the consideration that a major group of
potential home buyers are young households who can look forward to an
increase in income even in the absence of inflation, both because of the
general effect of productivity growth, which tends to raise all incomes,
and because typically, even in the absence of productivity growth, income
tends to rise with age, at least for a while. For such households, the op-
timal time profile of real payments might be one rising over time; in-
flation instead will tilt the ratio of mortgage payments to income even fur-
ther than indicated by Figure 1.2
2For this reason, young families with expectations of rapid increases in income might
prefer a mortgage with rising real payments while other families, facing retirement and a
drop in income, might prefer the opposite.
A faster repayment schedule and the resulting higher ratio of payment

to income in the early years of the contract need not of course be a prob-
lem for those households who had intended to save at a rate sufficiently
high to satisfy the schedule; but would be a problem for other households
and their number would grow rapidly with rising inflation and the
sulting speedup of repayments.
Even for these households, the problem could be handled in a world
of perfect markets, no money illusion, and infinite ingenuity in devising fi-
nancial instruments suited to changing circumstances. In this ideal world,
the borrowers would be able to raise otherwise the funds needed for the
high early payments, for example, through second mortgages or unsecured
personal loans. But, obviously, our world does not meet these ideal speci-
fications. Indeed, there is little evidence of any significant tendency on the
part of lenders to make full use even of the flexibility in the existing mort-
gage contract to counteract the higher initial payments resulting from in-
flation by lengthening maturity or by raising the loan-to-value ratio. In
any event, these devices would not go very far in counteracting the effect
of inflation on the early payments.3
On the basis of this analysis, we conclude that, under traditional
mortgage financing, inflation is likely to affect adversely the demand for
houses by inducing potential buyers -- especially first owners -- to scale
down their demand in terms of quantity and/or quality or to forego
quisition, at least until they have accumulated enough assets for a larger
downpayment. It also follows that marked fluctuations in the actual and
anticipated rate of inflation such as have occurred in the last decade, tend
to change the demand for housing and thus contribute to the observed
swings in residential construction activity. The Kearl, Rosen, and Swan
paper endeavored to find empirical evidence on the quantitative mag-
nitude of this impact, through a search of the empirical literature of fac-
tors controlling the demand for housing. Unfortunately, the existing con-
tributions did not provide this evidence as none of the authors of previous
empirical studies has explicitly treated these effects.
Uncertainty About the Level of Inflation and the Demand for Hous-
ing. In addition to the effects just discussed, which depend on the level of
inflation, the demand for housing may also be affected by uncertainty
about the future of inflation. Consider, for instance, the case illustrated in
Case D of Table 1, when the rate of inflation anticipated over the 30
years of the contract is 8 percent, and on this basis the mortgage rate is
set at 11 percent. If the actual path of inflation turned out to be appre-
ciably different from 8 percent, the path of actual payments expressed in
terms of purchasing power would also be different from that of column
(3). In particular, if the deviations were prevailing in one direction, the
3Lengthening maturity from prevailing practices could accommodate, at best, only ~

very moderate rate of inflation and changing the loan-to-value ratio increases the lender’
risk in the early years of the contract.
present value of the stream of real payments could deviate significantly

from the intended $20,000. Thus, in the presence of significant uncertainty
about the future rate of inflation, the mortgage instrument, as a fixed
long-term contract, becomes a risky one for the borrower as well as the
lender. If inflation turns out to be higher than expected, the borrower
reaps a windfall gain (and the lender suffers a windfall loss); and if lower,
the opposite occurs.
In our recent history, inflation typically has turned out to be higher
than expected and, in addition, interest rates have frequently been kept
artificially low by government policy, all of which has worked out to the
advantage of the borrower. Thus there has been a tendency to assume
that inflation is detrimental to the lender, but is good for the borrower
and has a favorable impact on housing demand. Actually, once inflation
has developed for a while, and interest rates are left free to incorporate
expectations of hefty rates of inflation, anyone borrowing on a long-term
basis to invest in a house bears a substantial risk of inflation turning out
lower than anticipated.
This risk is mitigated to some extent by the prevailing early re-
payment provisions on mortgages, mandated by law in many states. Often
borrowers are allowed to repay ahead of schedule with minimal penalties.
This is viewed as a social necessity to allow people to buy and sell houses
freely; but it also results in a "one-way option’’ in which the borrower can
always get out of the original contract if interest rates fall, thereby reduc-
ing his risk of a lower than expected rate of inflation -- but the lender
cannot get out if they rise. Of course, a rational financial intermediary
that recognizes this assymetry should exact a premium for this option
during periods of high and uncertain inflation and interest rates with the
result that borrowers would have to pay for the reduction of risk inherent
in the prepayment clause in the form of an even higher interest rate.
One might conclude that insofar as households are prevailingly averse
to risk, prepayment options are correctly priced, and interest rates freely
reflect expected inflation. A high and uncertain rate of inflation could
tend to reduce the demand for housing through its effect on the expected
cost and risk to the borrower. It must be acknowledged however that,
since these circumstances also increase the risk of investment in long-term
fixed-rate financial assets, they may encourage wealth holders to invest in
physical assets such as houses, especially since much evidence suggests
that equities are not a particularly good hedge against inflation. The em-
pirical relevance of this phenomenon is supported by the experience of
countries with high rates of inflation.
These considerations make it hard to reach firm conclusions about
the overall impact of uncertainty about the future of inflation on the de-
mand for houses, especially since this depends in part on the nature of fi-
nancial instruments available to investors. One conclusion that seems war-
ranted, however, is that, if alternative instruments could be devised to
finance housing which reduce the price-level risk inherent in the standard
mortgage, this would also have some favorable impact at least on the de-
mand for owner-occupied housing. However, the shortcomings of the tra-
ditional mortgage arising from the uncertainty of inflation are likely to be
of secondary importance compared with those arising from the tilting of
the stream of payments discussed earlier.4
Inadequacies of Current Remedies. Several countries which at one
time or another experience double-digit inflation have come to realize that
at these high rates the traditional mortgage instrument requires such an
exorbitant initial rate of repayment of principal that it becomes practically
useless as a financing device. They have accordingly been led to try out
basic reforms in this instrument involving some form of "price-level ad-
justment" along lines detailed in the reviews of Finland, Israel and Brazil
and discussed further in IV.D below. Many other countries, including the
United States and the United Kingdom have tried to relieve the problem
by holding down interest rates through ceilings or by providing interest
rate or housing subsidies. Only a few countries, notably Sweden, have
tried to combine subsidies with financial innovation and government
guarantees.
Typically, the approaches implemented or proposed in the United
States have aimed at making mortgages available to qualified borrowers
at below equilibrium interest rates. It should be apparent from our anal-
ysis that such schemes constitute an inefficient approach: they would be
unnecessary if the right cure were provided.
If our analysis is correct, the problem does not arise from the fact
that, with a higher inflation, the borrowers can no longer afford to pay
the interest rate on the principal and amortize the debt at a reasonable
and prudent rate. Indeed, we have shown that higher interest rates arising
from inflation do not change the overall real cost of the house; hence, in-
flation per se should not be a ground for subsidies, especially to potential
home purchasers who on the average do not come from the poorest clas-
ses of society. The problem arises instead from the fact that, with high in-
flation, use of the standard mortgage requires borrowers to repay the debt
at an unreasonably fast pace.
The true solution to this demand effect must therefore lie in devising
instruments such that the path of repayment of the loan (measured in
terms of purchasing power) will be independent of the rate of inflation --
say the same as it would be under a traditional mortgage in the absence
of inflation thus eliminating the tilt effect of the standard mortgage.
B. Effect of Inflation on Housing Through the Supply of Mortgage
Funds
Supply effects can be dealt with briefly as they are already fairly gen-
erally understood and agreed upon. They arise not from the rate of in-
flation as such, but rather from its variations, and from its interaction
4To the extent that the risks transferred to the lender via the "one-way" prepayment op-
tions result in higher mortgage interest rates, they will exacerbate the "tilt" effects.
with interest rate ceilings. Both are intimately related to the rather unique
and not altogether satisfactory structure through which the bulk of funds
to finance mortgages have been raised in the United States in recent
decades.
As is well known, by far the largest share of private mortgage funds,
especially those financing owner-occupied housing, has come from the
thrift institutions -- savings and loan associations and mutual savings
banks -- and to some extent from commercial banks and life insurance
companies.5 These institutions in turn have obtained the funds almost en-
tirely from deposits. Through much of the postwar periods these deposits
were almost entirely short term and highly liquid -- indeed, l%r all prac-
tical purposes they could be and were regarded as demand liabilities. This
practice was on the whole looked upon with favor, as one of the basic
functions of these institutions was viewed as that of providing the public
with a highly liquid investment. Only recently has this type of liability
been supplemented to a growing extent by deposits with longer maturities.
Consequences of Maturity Mismatching. As a result of these prac-
tices, thrift institutions acquired an extremely unbalanced or mismatched
financial structure, consisting of very long-term assets and very short-term
liabilities. This unbalanced portfolio did not reflect a conscious endeavor
to speculate on the term structure, which would have involved shifting as-
set and liability maturities at various points in time. This becomes clear
when one recognizes that mortgages are not very attractive instruments
for speculating on the term structure since in many states the borrower
can easily avail himself of the option to repay at no significant penalty in
the event that interest rates fall. Rather, at least in the case of S&Ls,
mortgages were one of the few investments that regulation allowed them
to make. That thrift institutions were thus forced into an unbalanced as-
set-liability structure must be regarded as unfortunate since it would hard-
ly seem socially desirable for these institutions to incur the risks of failure
associated with extensive term-structure intermediation.
This portfolio imbalance did not create any difficulty during the peri-
od of relative price stability which lasted until the mid-60s as interest rates
changed slowly, the term structure was prevailingly a rising one, and in
addition, deposit rates were not under serious competitive pressure thanks
to the low ceilings imposed on commercial bank time deposits. Ac-
cordingly, the thrift institutions and the S&Ls in particular were able to
attract a large flow of funds and provide ample financing for residential
mortgages. They were in fact so successful that home mortgages became
less attractive for other intermediaries, such as life insurance companies,
causing the market to rely on thrift institutions to a growing extent. Thus,
the thrift institutions’ share of all privately held home mortgages increased
from the early 50s to the early 70s from roughly one- to two-thirds; and
5Government funds, in particular purchases of mortgages by the FNMA, have played

an increasingly important role in mortgage financing in recent years.
because of this growth, their share of the annual flows was even more im-
pressive, frequently reaching 80 percent and over.
But the weaknesses inherent in such a structure become apparent in
the era of rising and variable inflation that began in the mid-60s. Rising
interest rates during periods of monetary stringency made it difficult to at~
tract depositors at rates of the earlier period. And the problem became
more acute at each successive monetary crunch -- 1966, 1969-70, 1974 --
when short-term rates rose even more than long-term ones. Supervisory
authorities became concerned that if institutions competed to retain de-
posits, they would have had to offer rates which would have resulted in
severe losses and ultimate collapse -- especially in view of the reduced
market value of their portfolios which were very illiquid anyway. To pre-
vent this outcome the regulatory authorities imposed ceilings on all de-
pository intermediaries.
Since the level of ceilings was constrained by what the thrift in-
stitutions could afford to pay, it was frequently well below short-term
market rates. Because no other assets of similar characteristics yielded
more, the thrift institutions were spared a mass withdrawal. Nonetheless,
their liabilities lost attractiveness for savers and their net inflows slowed
down dramatically and even became negative for brief periods (the so-
called "disintermediation"). Furthermore, this unfavorable response of de-
positors tended to become more pronounced at successive "crunches" as
they became sensitive to rate differentials and as financial innovations
provided them with better alternatives, such as the shoi:t’-term money-
market funds. These periods of famine were typically followed by periods
of heavy inflows as each crunch was followed by a period of very low
short-term rates as monetary policy eased off. The wide swings in deposit
inflows resulted in similar swings in the supply of mortgage funds which
played a major role in the wide fluctuations in construction activity and
housing markets.
Solutions to the Supply Problem. The lessons to be learned from this
experience are fairly obvious and broadly agreed upon: if there is a sub-
stantial risk of inflation, the institutions financing housing must not be al-
lowed to continue the present practice of lending through traditional
mortgages -- a very long-term instrument -- while relying on very short-
term liabilities as a source of funds.
Hence, if the thrift institutions are to continue to provide the public
with a highly liquid, conventional, deposit-type asset and to use the bulk
of the funds so obtained to finance housing, they must have a financing
instrument which will allow them to earn a return commensurate with
changing short-term market rates. If they continue to invest part of their
portfolio in instruments of long maturity with fixed interest rates, they
should hedge them by liabilities of commensurate maturity, as well as
matching characteristics in terms of prepayment options and the like. If
instruments of an entirely new type were made available to them (such as
the price-level-adjusted mortgages discussed below), they should again fi-
nance investments in this asset with liabilities of similar characteristics. It
should be added that the basic principle that prudent financial structure
requires matching the characteristics of assets and liabilities has long been
a tenet of financial theory and practice and is recognized by the in-
stitutions which finance housing in other countries. Thus (1) where con-
ventional mortgages are used, they are typically financed by mortgage
bonds, (e.g., Sweden, as well as many other countries); (2) where rnort-
gages are financed by short-term deposits, their interest rate is subject to
change (e.g., United Kingdom); (3) where the mortgage is financed by li-
abilities of intermediate term, the balance still due at the end of that term
is refinanced at the then prevailing rate (e.g., Canada).
In the next section, drawing on the more detailed and rigorous anal-
ysis of the Cohn-Fischer paper, we review a number of alternative mort-
gage designs, assessing how well each design could fit into the portfolio of
thrift institutions in terms of matching requirements, how well it would
suit borrowers’ interests, and how effective it would be in eliminating or
reducing the demand effect of inflation-induced changes in interest rates,
and hence in reducing instability in construction activity resulting from
such changes.
III. ALTERNATIVE MORTGAGE DESIGNS AND THEIR
EFFECTIVENESS IN ELIMINATING DEMAND AND
SUPPLY EFFECTS OF INFLATION
A. The Basic Elements of the Mortgage Contract
A mortgage is simply a loan contract which specifies a rule for (1) de-
termining the interest rate applying in any year to the debit balance then
outstanding, called hereafter the debiting rate, and (2) calculating the
periodic payments through which the debtor is to pay the interest and
amortize the principal over the life of the contract. The traditional mort-
gage can thus be viewed as a special case of a much broader class, and a
large number of alternative designs can be constructed by varying the var-
ious parameters characterizing the instrument. In the course of the Cohn-
Fischer study, as well as in the Jaffee-Kearl simulations, many designs
have been given at least passing consideration. In what follows, we con-
centrate on a. few of these, chosen on the basis of two criteria: (1) the ex-
tent to which they have already received attention and are being applied
here or abroad, or are at least being actively promoted, and (2) the extent
to which they appear to provide a viable solution to the problems dis-
cussed in Section II.
B. The Variable-Rate Mortgage
The alternative to the traditional mortgage that has received by far

the greatest attention and already has been adopted in some parts of the
country, is the variable-rate mortgage (VRM). It is being promoted pri-
marily by lending intermediary interests as a solution to their problem
and thus also to the supply component of the housing problem
The essential characteristic of the VRM is that the debiting rate
charged on the borrower’s outstanding balance is not fixed at the outset
but is allowed to float up or down, being tied to some agreed "reference
rate." This specification is consistent with a variety of designs in terms of
(1) choice of specific reference rates such as a short, intermediate or long-
term market rate, or the deposit rate of the intermediary originating the
loan, (2) frequency with which the debiting rate is changed, (3) lim-
itations, if any, on the maximum permissible change at revision points or
over the life of the contract, and (4) methods for computing the periodic
payments.
Within this class, two major alternative designs have received con-
sideration. In one design, which has been adopted in the United Kingdom
and elsewhere, the periodic payment is fixed at the beginning of the con-
tract as in the traditional mortgage. Because a discrepancy between the
debiting rate and the rate used to compute the payment at the outset leads
to a corresponding discrepancy between the amount available for the
amortization of principal and the amount scheduled for that purpose, the
payments do not necessarily terminate at the original scheduled maturity,
but only when the principal has been fully amortized. Thus, the in-
strument is of variable maturity. In the alternative design, the maturity is
fixed and the periodic payments change with the debiting rate, that rate
being used to recompute a new level payment over the remaining life of
the contract which applies until a new change occurs in the debiting rate.
The adoption of the VRM could be expected to alleviate, if not solve,
the intermediaries’ mismatching problem and, hopefully, the supply aspect
of swings in housing markets, especially if the reference rate were of the
same maturity as the funds used to finance the mortgage. For inter-
mediaries financed by short-term deposit liabilities, whose market value is
always par, the appropriate rate would be a short-term rate or the deposit
rate itself provided it was not distorted by ceilings. This would keep the
value of the mortgages close to par. In terms of its effect on the borrower,
however, the VRM appears to offer little relief to the housing problems
and in fact is likely to make matters worse. This is because the rate used
to compute payments with a VRM is a nominal rate which responds to
the rate of inflation and hence does not eliminate the tilt effect. Actually
the relatively wider variations in short-term interest rates are likely to ex-
acerbate swings in demand due to changes in initial periodic payments, al-
though the generally lower level of these rates, relative to long-term rates,
may stimulate demand over the long term (Cf. the simulation results of
Jaffee and Kearl).
A.more common criticism of the VRM advanced by consumer advo-
cates has been that making the interest rate variable increases the bor-
rower’s risk. This conclusion is open to question. It is true that if the ref-
erence rate should turn out to rise above the initial mortgage rate, the
debtor would end up paying more, but presumably this would tend to
happen if inflation were also higher than the expectation built into the
long-term rate, in which case the debtor’s money income would also tend
to be higher in the long run. On the other hand, the reference rate could
also decline, reducing payments, and this would tend to happen in the
event that the rate of inflation turns out lower than anticipated and hence
less growth in money income is realized than was expected. In other
words, some of the risk of the VRM is offset by the long-term positive
association between the borrower’s money liability and his money income.
However, with the fixed maturity version some of this risk remains.
Although variations in the periodic payment are broadly associated with
those in the rate of inflation and money income, in the short run the asso-
ciation is not close, in part because of the jerky nature of payment
changes, and as a result the ratio of payment to income could be subject
to substantial variability. This can be seen by inspecting columns (1) and
(2) in the last row of the VRM block. If the rate of inflation rises from 3
to 5 percent, the scheduled payment under VRM rises from $1,453 to
$1,798 or by 24 percent, whereas the effect on the average homeowner’s
nominal income would be more like 2 percent. The reason for this much
higher percentage change is that the higher inflation, by raising the nomi-
nal rate used in computing the constant payment for the rest of the con-
tract, implies a further tilting of the real repayment schedule. For similar
reasons, an absolute decline in inflation produces a much larger per-
centage decline in the scheduled payment. (Cf. col. (3) and (4)).
The potentially large fluctuations in payments over time with VRMs
could be relieved by a variety of modifications. One modification is the
fixed-payment variable-maturity version of VRM. But this version can af-
ford only limited relief when the maturity is long, as is the case in the ear-
ly years of the contract, and most of the periodic payment consists of in-
terest. Even small upward revisions in the debiting rate produce large
changes in the scheduled maturity, and the point is soon reached where a
fixed payment proves insufficient even to amortize the debt.6 Thus the
variable-maturity VRM is capable of "smoothing" minor fluctuations in
the interest rate, but not major shifts such as those observed in recent
years.
Various other modifications have been proposed for the variable pay-
ment VRM such as using as reference a longer-term interest rate which
presumably is less volatile than a short-term rate; limiting the frequency
with which the debiting rate can be changed; allowing the maturity to
vary and limiting its maximum permissible change at revision dates or
over the entire life of the contract. But while such modifications would
certainly improve the borrower’s lot, they might by the same token reduce
the benefits of VRM to the lender, and hence also its effectiveness in
solving the supply problem. Indeed, any of these proposals increase the
probability that the market value of the mortgages will vary relative to
their par value and thus deviate from the value of intermediaries’
liabilities.
6For example, an increase in the debiting rate to 8.8 percent from an initial level of 8
percent would result in the entire payment going toward interest.
All of these proposals relate directly or indirectly to a basic dilemma

in VRM design. From the perspective of the lender who obtains a signifi-
cant proportion of funds with short-term liabilities, a short-term debiting
rate is desirable while from the borrower’s perspective a longer-term rate
is desirable because of its lower volatility. This dilemma, and the extent to
which the various proposed modifications of the basic VRM instrument
resolve it, can be best understood by considering a novel variant of the
VRM which emerged during the course of our study.
C. The Dual-Rate VRM and Other Approaches to the VRM Dilemma
The dual-rate VRM endeavors to resolve the above dilemma by using

two distinct interest rates; one, which we call hereafter the debiting rate, is
used to compute the interest on the outstanding balance; the other, which
we term the payment factor, is used to compute the periodic payment.
For the debiting rate, one would use as reference a short-term rate or the
deposit rate; the latter would seem preferable because it is directly related
to the cost of funds to the intermediary, and because this mechanism is
likely to be more readily understood, verified, and accepted by borrowers.
The periodic payment, on the other hand, is recomputed at fixed intervals
by applying to the principal still outstanding with the standard annuity
formula using some longer-term rate, say an intermediate rate or the rate
on the longest-term deposit offered by the intermediary.7 Using longer
rates for computing the periodic payment would have the effect of reduc-
ing the magnitude and, possibly, the frequency of changes in the payment.
If the debiting rate differs from the payment factor, the actual amor-
tization of the debt may differ from that implied by the payment factor.
Thus when a new periodic payment is computed, it could differ from the
previous payment because of the aforementioned discrepancy in principal
and because of a change in the reference rate for the payment factor.
Nonetheless the variations could be expected to be appreciably smaller
than for a standard VRM which used the same debiting rate for three rea-
sons. First, the discrepancy in principal should not be large since the aver-
age debiting rates -- short-term rates -- should not differ markedly from
the longer-term rate which is, after all, a forecast of the average short-
term rates. Second, the discrepancy, if any, is spread over the remaining
life of the contract and thus will not have a major impact on the payment.
Finally, the payment rate, a longer-term rate, should be smoother than
the debiting rate.
Thus a dual-rate VRM, with appropriately chosen reference rates and
frequency of adjustment, can both enable the lending intermediary to earn
a rate adequate to keep its deposit rate competitive with other short-term
market instruments and still result in a smooth path of periodic payments
in money terms. Its primary drawback, however, is its complexity.
7Section III.F and the Cohn-Fischer paper illustrate mechanics of this design.
3O NEW MORTGAGE DESIGNS
Another approach to the dilemma is simply to use a longer-term rate
for debiting as well as computing the payment. Insofar as its liabilities are
of shorter term, this approach, as noted earlier, again exposes the inter-
mediary to the danger of its revenue not keeping up with the rate it must
pay on its liabilities or equivalently to the risk that the market value of its
assets will fall short of that of its liabilities. Ideally, this risk would be
avoided if the liabilities were themselves term deposits with maturities
matching that of the debiting rate. This approach is actually used in Can-
ada, where mortgage rates are adjusted at five-year intervals and funding
is obtained through five-year term certificates. As a result, Canadian in-
stitutions are perfectly hedged, that is, changes in the market value of
sets are perfeclty matched by changes in the value of liabilities. Because of
this, they have been able to avoid most of the supply (but not the de-
mand) problems which have plagued U.S. housing markets.
If the debiting rate were a three-year rate fixed for three years, the
risk to an intermediary financed by short-term liabilities might not be ap-
preciably larger than if the debiting rate were a short-term one (a three-
year instrument is unlikely to fall significantly below par), while the
smoothing from the point of view of the borrower would be appreciable.
It may be argued that bearing this limited risk is an appropriate function
of the intermediary in order to reduce the borrower’s risk.
The Federal Home Loan Bank Board has recently proposed a mod-
ification of this approach in which the debiting rate would be a three- to
five-year rate, but instead of being fixed for this term, it would be adjust-
ed every six months in accordance with movements in this same rate.
There would also be a limitation to the maximum change in the debiting
rate to one-half,of 1 percent every six months and 2.5 percent over the life
of the contract. This instrument is a hybrid that is neither short nor inter-
mediate term. By adjusting the rate at more frequent intervals than the
term of the rate, it would appear to create situations where market values
would fluctuate around par and might provide borrowers with arbitrage
opportunities. However, the more frequent adjustments would insure that
mortgage yields would be sensitive to general shifts in the level of interest
rates, thus reducing the chance of the mortgage portfolio going to a sig-
nificant discount.
To summarize, the VRM would be helpful to lenders and with in-
genuity might not impose too great a burden on borrowers as compared
with the standard mortgage. The dual-rate VRM appears to go furthest in
mitigating the disadvantages to the borrower for a given gain to the lend-
er by using a short-term debiting rate such as the deposit rate, wtiile elim-
inating much of the inconvenience and risk placed on the borrower
through large, sudden changes in the periodic payment.
However, the VRM in any form still fails to resolve and at least to
some extent would worsen what we have called the demand effects of in-
flation, namely the capricious changes in initial level of payments due to
inflation-swollen interest rates.
A quite different foreseeable shortcoming that might result from wide-
spread adoption of the fixed maturity VRM is of a macroeconomic char-
acter. A change in the debiting rate would result in an increase of the
periodic payments for millions of homeowners. If the reference rate is a
market rate, a great deal of public pressure might be brought to bear for
the central bank to hold down that rate when stabilization considerations
would, on the contrary, call for higher rates (reflecting, e.g., inflationary
expectations). This sort of pressure, which even now interferes with ap-
propriate policy, would certainly be greatly magnified under the VRM.
And if the VRM were the deposit rate, the same pressures would be direc-
ted toward holding that rate down in the face of rising market rates. This
pressure, if successful, would, much like the imposition of ceilings, cause
the intermediaries’ deposits to lose attractiveness, and thus recreate the
very supply effect that VRM was designed to solve. The recent experience
of the United Kingdom provides an enlightening illustration of this
scenario.
D. The Graduated-Payment Mortgage (GP)
Since a major impact of inflation on the homebuyer is the tilting of
the time-stream of payments -- one obvious solution to this problem is a
mortgage which involves relatively lower money payments in early years.
Clearly, unless such a mortgage is subsidized or of longer maturity, it
must involve relatively higher money payments in later years in order to
fully amortize the loan and provide the required return to the lender.
Graduated-payment mortgages, with contractually rising payment streams,
have been advocated in the United States and have been implemented in
some other countries including the United Kingdom, where they are
known as "low-start" mortgages, and Germany. The Federal Home Loan
Bank Board moved part way in this direction when it authorized S&Ls to
write mortgages with payments covering only interest for the first five
years and amortizing the principal over the remaining term of the
mortgages.
In a world with a steady rate of inflation, a graduated-payment mort-
gage with payments which increase over time at a rate equal to the rate of
inflation would eliminate the tilt effect in terms of constant purchasing
power dollars and restore the basic feature of the traditional mortgage in
a noninflationary environment -- level payments over the life of the mort-
gages. By and large, this would imply the same ratio of mortgage pay-
ments to household incomes and the same equity buildup (measured in
real terms or simply as a ratio of the value of the property to the loan
outstanding) as the traditional mortgage instrument, since wages and
house values should, on average, also increase at the rate of inflation rela-
tive to their levels in the noninflationary environment.
One feature of the graduated-payment mortgage which might generate
resistance on the part of both borrowers and lenders is that the out-
standing principal in the early years of the contract would actually in-
crease. For example, if the rate of inflation was 6 percent and the current
nominal interest rate 9 percent, reflecting an interest rate of 3 percent in
dollars of constant purchasing power, a $20,000, 30-year graduated-pay-
ment mortgage with payments geared to rise at the rate of irfflation would
call for a payment of $1,020 in the first year.8 The interest charge on the
other hand would be $1,800. The "shortfall" of $780 would be added to
the loan balance. The principal would continue to increase for several
years, although the rising payments would eventually exceed interest
charges and would fully amortize the principal by the end of the contract
period.
While this situation raises some interesting tax questions, which are
discussed by Holland, it should not be a cause for alarm on the part of
either borrower or lender. The value of the house, and hence of the bor-
rower’s equity and the lender’s collateral, can be expected to rise along
with the loan buildup. In fact, if the rate of increase in the property value
was exactly 6 percent greater than under noninflationary conditions, the
borrower’s equity position every time, measured by the ratio of out-
standing debt to the value of property, would be identical to that in the
zero inflation environment.
Any resistance, then, would be the result of a failure to take into ac-
count the changing value of the dollar due to inflation. This is not to say
that this "money illusion" will not be present or hard to overcome; hope-
fully it should be possible to overcome it through information and
education.
Unfortunately, the GP mortgage suffers from several serious short-
comings. First, with uncertainty about future rates of inflation, a contract
calling for payments rising at the expected rate of inflation would be risky
for both the borrower and lender. If inflation turned out to be less than
anticipated, the borrower would face payments rising relative to income
and a slimmer equity position. This, of course, would also increase lender
risk. For this reason, the graduated payment mortgage with a rising
schedule of payments set forth at the outset is generally viewed as appro-
priate only for young families with expectations of wage growth sub-
stantially in excess of the rate of inflation. While it is true that the risk is
less for such families, this view confuses two issues -- the need for a non-
level payment in money terms simply to remove the distortions in the pay-
ment pattern of the traditional mortgage resulting from inflation and the
need for a nonlevel payment in real terms, either rising or falling, to
match a household’s position in the life cycle.
Finally, a graduated-payment mortgage with a fixed interest rate over
its entire life, being a long-term instrument, would do nothing to solve the
supply problem stemming from thrift instituions’ reliance on short-term
deposits as a source of funds. In fact, it would exacerbate the problem
since it would lengthen the duration of the mortgage, i.e., a larger balance
8This payment is equivalent to the payment required to amortize the loan with level
payments at 3 percent, the difference between the debiting rate and the rate of graduation.
would be outstanding at each payment date than would be the case with a
standard mortgage.
We must conclude that neither the VRM nor the GP is an attractive
solution to the distortions in mortgage financing brought about by in-
flation and the accompanying high and uncertain interest rates. Each is a
partial solution that benefits either the lender or the borrower, but at the
expense of the other party. One mortgage design which, in the abstract at
least, has the potential of satisfying these requirements is the price-level--
adjusted mortgage (often referred to as a price-level-indexed or index-link-
ed mortgage).
E. The Price Level-Adjusted Mortgage (PLAM)
The basic mechanics of the PLAM involve a contractual interest rate
which abstracts from inflationary anticipations, and a periodic revaluation
of the outstanding principal in accordance with the change in the price-
level index to which it is tied. In effect, the debiting rate on the PLAM is
a real rate of interest, differing from the current money rate by the ex-
clusion of the inflation premium, which reflects the anticipated change in
the price level over the period of the contract. Payments are recomputed
whenever the principal is revised, using the contract rate as the payment
factor. As a result, the PLAM payment stream changes exactly in line
with the reference price level.
This is illustrated in Case C of Table 2, which also shows the me-
chanics of the calculations. The contract rate is taken as 3 percent, the
rate assumed to hold in the absence of inflation premia, and this results in
an initial payment of $1,020, as compared with $1,453 at the 6 percent
rate for the standard mortgage at the market rate shown in Case A of the
exhibit. This payment is subtracted from the sum of the beginning prin-
cipal plus interest plus the revaluation of principal (the rate of inflation
times the beginning principal). Thus, at the end of the period the bor-
rower owed the amount shown in row 5, an amount greater than the be-
ginning principal much as with a GPM.
When account is taken both of the 3 percent interest charged on the
oustanding principal and of the 3 percent writeup of the debt to reflect in-
flation, the total return to the lender and cost to the borrower is 6 per-
cent, the same as the nominal rate.9 The low contract rate, however,
makes it possible to hold the initial payment down. Moving to the second
year, the revalued principal is used to compute the next year’s payment at
the 3 percent rate. Because the principal has been increased precisely by
the rate of inflation the new payment based on it also increased at that
9More precisely the return is (1 + payment rate) x (1 + rate of change in reference price
index) -1.
Table 2
EXAMPLES OF COMPUTATION OF ANNUAL
MORTGAGE PAYMENTS UNDER STANDARD MORTGAGE
AND THREE ALTERNATIVE TYPES
Year l 2 3 4
Real Interest Rate 3% 3% 3% 3%
Rate of Inflation 3% 5% 5% 4%
Nominal interest rate~ 6% 8% 8% 7%
Years to Maturity 30 29 28 27
A -- STANDARD MORTGAGE
1. Beginning Principal 20,000.00 19,747.00 19,478.82 19,194.55
2. Plus Interest (6%) .2 1,200.00 1,184.82 1,168.73 1,151.67
3. Less Annual Payment 1,453.00 1,453.00 1,453.00 1,453.00
4. Ending Principal 19,747.00 19,478.82 19,194.55 18,893.22
5. Memo: Annual Payment
in Constant
Purchasing Power
as of Beginning
of Each Year3 1,453.00 1,410.68 1,342.89 1,279.53
B -- VARIABLE-RATE MORTGAGE (VRM)
2. Plus Interest (nominal2rate) 1,200.00 1,579.76 1,564.57 1,354.64
in Constant
Purchasing Power
as of Beginning
of Year 1,453.00 1,718.16 1,636.34 1,421.70
C -- PRICE-LEVEL-ADJUSTED MORTGAGE (PLAM)
2. Plus Interest (3%) 600.00 605.39 622.27 638.89
3. Plus Revaluation of
Principal for Inflation 600.00 1,008.98 1,037.12 851.85
4. Less Payment2 1,020.39 1,051.65 1,105.43 1,162.02
in Constant
Purchasing Power
as of Beginning
of Each Year~ 1,020.39 1,021.02 1,021.65 1,023.29
D -- CONSTANT-PAYM.ENT-FACTOR
VARIABLE-RATE MORTGAGE
2. Plus Interest (nominal2rate) 1,200.00 1,614.38 1,659.39 1,490.74
in Constant
Purchasing Power
as of Beginning
of Each Year~ 1,020.39 1,021.02 1,021.65 1,023.29
~For simplicity, we simply add the rate of inflation, q and the real rate of interest, r, to
obtain the nominal rate of interest, i. Tbe precise rate is tile product of the two --
2The payment due at the end of each year is calculated at the beginning of the year by
applying the appropriate payment factor, either a constant or the nominal rate of interest, to
the principal outstanding at the beginning of the year. This is done in order to provide the
borrower with adequate notice of a change in payments. In practice, such a "notification"
lag would more likely be on the order of three months.
rate. This of course means that the payment expressed in constant pur-
chasing power, shown in row 6, remains at the initial level.1° This result
holds for all remaining years of the contract.
Advantages of PLAMs for Borrowers. PLAM has a number of ad-
vantages for borrowers. First and foremost, it completely eliminates the
tilting effect of inflation on the stream of payments in purchasing power
terms which results from the traditional mortgage (or the YRM); under
PLAM the stream of payments is constant over the life of the contract
and is, in fact, equal to the payment required by a traditional mortgage in
the absence of inflation. In terms of the example of Table I, the initial
payment would be $1,020, whether the expected inflation be 0, 2, 4 or 8
percent. Second, a constant stream of payments in real terms, in contrast
to one decreasing at a rate capriciously determined by the happen-chance
of the rate of inflation, could be expected to suit the bulk of potential ho-
meowners, particularly the younger households, whose real income is
largely independent of the rate of inflation.11
A third important property of PLAM is that, by contractually estab-
lishing the total payment in terms of purchasing power, it eliminates the
risk to borrowers associated with unanticipated variations in the price
level. As pointed out earlier, though in the past these unanticipated vari-
ations have tended to benefit borrowers, this need not be the case in the
future as interest rates have adjusted to reflect expectations more ade-
quately. As shown by Cohn and Fischer, this property is again especially
important in reducing uncertainty for those households who can expect
their real income to be largely independent of the rate of inflation.
To summarize then, PLAM (in contrast to VRM or GP) does appear
to offer a more complete solution to the range of problems which we have
labeled the demand effects of inflation. It does so through a contract
which, in effect, produces the same real consequences for the borrower
(and the lender) as would the traditional mortgage in the absence of in-
flation -- and does so no matter what the rate of inflation either antici-
pated or realized.
Feasibility of PLAMs. Some form of PLAM has actually been adop-
ted in several countries, most notably Brazil, but also Israel, Finland,
Colombia, and Chile. Experience with PLAMs appears to have been ex-
tremely successful in a few cases, though they have been abandoned in
~°The payments in the example are constant in purchasing power as of the time when
they are scheduled, i.e., at the beginning of the year. As noted in footnote 1 of Table 2, pay-
ment could be scheduled so as to be constant in terms of purchasing power at the time of
payment. However, this would not provide the borrower with any prior notice regarding the
exact money amount of his payment. The problem is exaggerated in our example since there
is a one-year interval between the scheduling date and the date when the payment is due.
l~Actually, as noted earlier, for many households~ real income may be expected to have
a rising trend over time, and to this extent, even under PLAM the ratio of periodic pay-
ments to income would tend to decline over time. In principle, this variation too could be re-
duced by combining the PLAM with the GP mechanism.
others (but the reviews of country experience included in this volume sug-
gest that this occurred for reasons largely unrelated to the basic mortgage
instrument itself.)
Unfortunately, as a practical short-run solution to the U.S. problems,
the novelty of the PLAM is a drawback. Borrowers and lenders are used
to contracting in money terms with nominal rather than "real" rates and
to payments fixed in nominal terms. Rates of inflation have not been so
high and persistent in the United States as to make people fully aware of
the pitfalls of money illusion. Thus fixing the payments in real terms with
the actual payment depending on inflation may be regarded by many as
increasing rather than decreasing risk. This hurdle could be surmounted
as it has been in other countries, but it might require an education effort.
To the extent that consumers are acquainted with wage escalators and
other such price-level-indexed contracts, this task will be made somewhat
easier.
There is, however, one further, and in the short run, more serious dif-
ficulty. Reaping the full benefits from PLAM would require substantial
changes in the type of liabilities issued by financial intermediaries -- as
well as possibly some changes in existing laws. Specifically, if thrift in-
stitutions are to be encouraged to offer PLAMs, they should be enabled
to hedge this asset by a price-level-adjusted deposit -- or PLAD -- that
is, a deposit whose principal would be reva!ued periodically on the basis
of the reference price index, and which accordingly would pay a real rate.
In our view, the addition of PLADs to the menu of presently existing
assets would be highly desirable in the presence of substantial and un-
certain inflation, as it would make it possible for savers to hedge against
the risk of price level changes. Such an opportunity is not presently avail-
able, especially where small savers are concerned.
One further advantage of empowering thrift institutions to offer
PLADs is that it would go a long way toward also solving the supply
problem -- assuming of course that supervisory authority would refrain
from placing ceilings on PLAD rates. Indeed, there are sound reasons for
supposing that PLADs could effectively compete with other instruments
even in periods of high interest rates. The U.S. experience suggests that
much of the variation in interest rates, especially longer-term ones, can be
traced to variations in actual and anticipated inflation. Thus keeping
PLADs competitive with other assets would not require appreciable
changes in the rate offered depositors, even in the face of large changes in
market rates. ~2
Unfortunately, the straightforward solution involving PLAMs hedged
by PLADs, despite its great attractiveness in principle, is likely to face se-
rious obstacles and resistance, at least in the near future. First, as already
~2Cohn and Fischer point out that thrift institutions could even finance PLAMs with
short-term PLADS or with ordinary deposits with a risk substantially smaller than they
presently incur in financing the traditional mortgage with short-term deposits.
indicated, this solution would require substantial changes in the thinking
of both borrowers and lenders, as well as substantial changes in reg-
ulations affecting thrift institutions. Second, authoritative financial circles
have frequently expressed strong opposition to the introduction of price-
level-adjusted deposits for fear that this would disrupt the market for
other instruments and/or force widespread adoption of price-level-adjust-
ed securities. They further argue -- though wrongly in our view -- that
any reform that would reduce the pains of inflation should be opposed, as
it would sap the will to fight inflation. Finally, the adoption of PLAMs
and PLADs might well require some changes or reinterpretation of the
tax laws. Thus, for a PLAM borrower, the revaluation of principal would
have to be treated, for income tax purposes, as a deductible expense on a
par with interest if he is not to be at a disadvantage vis-h-vis a borrower
relying on the standard mortgage; and if this treatment were accorded to
him, then the revaluation of principal of a PLAD would have to be trea-
ted as ordinary income to avoid a special advantage to this asset and min-
imize disruption of capital markets, as well as avoid a net loss of revenue
to the Treasury. These issues and related ones are reviewed in the Holland
paper.
For all of these reasons, we believe that a more promising solution to
the problem may be found in the adoption of a somewhat different in-
strument which we label the "constant-payment-factor VRM." This in-
strument, described in the next section, combines most of the advantages
of the PLAM-PLAD approach, while requiring a minimum of in-
stitutional changes.
F. The Constant-Payment-Factor Variable-Rate Mortgage
This instrument may be thought of as a variant of the dual-rate VRM

outlined in Section IV.C, or also as a hybrid of the variable-rate and the
graduated-payment mortgage.13 Just like the dual-rate mortgage, the con-
stant-payment-factor VRM makes use of two separate rates: a debiting
factor which is charged on the outstanding balance and a payment factor
which is used to recompute the periodic payment at regular intervals by
applying to the balance still outstanding the standard annuity formula. As
in all VRMs, the debiting rate varies in accordance with an appropriate
reference rate reflecting market conditions. There is some room about the
choice of this reference rate, but ideally it should be chosen with reference
to the frequency with which the rate is adjusted and to the term of the in-
strument with which the mortgage is financed. The basic difference with
~Donald Tucker [1975] advocates a graduated-payment VRM with either a fixed

schedule of graduation or a schedule which varies with changes in the reference interest rate.
Cohn and Fischer show that the latter version, which calls for recomputing the entire stream
of payments whenever the reference rate changes, is identical to the constant-payment-factor
VRM.
respect to the dual-rate VRM, and also the essential ingredient of the in-
strument, is in the choice of the payment factors. This factor would be
chosen to approximate the "real" rate and would be kept fixed for the du-
ration of the contract. To the extent that the real rate is reasonably stable,
or equivalently that the debiting rate less the rate of inflation does not de-
viate widely or systematically from the payment factor, as past evidence
suggests, the payments on a constant-payment-factor VRM will approxi-
mate those of a PLAM. The initial payment, being based on the real rate,
will be the same as under PLAM -- a "low-start" payment which regard-
less of the actual rate of inflation is equal to what it would be in the ab-
sence of inflation. The behavior of the periodic payment in subsequent
years will depend upon the difference between the payment factor and the
debiting rate. If there is inflation, the debiting rate will exceed the pay-
ment factor. Therefore, principal will not be amortized at the rate implied
by the annuity formula and under highly inflationary conditions it would
actually rise. Thus, when the payment is recomputed, it will rise, even
though the payment factor remains unchanged. In fact, if the debiting rate
exceeds the real rate by a differential roughly equal to the rate of in-
flation, then the periodic payment will also rise roughly at the rate of
inflation.
This conclusion is illustrated by a specific example in Part D of Table
2. In the first year the payment is $1,020.39, the annual amount required
to amortize the $20,000 balance over 30 years with a 3 percent interest
rate. (Note that it is the same as the initial PLAM payment, also com-
puted at the 3 percent rate.) This $1,020.39, however, falls short of the ac-
tual interest charge of $1,200 at the 6 percent debiting rate. Hence, the
"shortfall" or $179.61 is added to the principal (the amortization is nega-
tive). In the second year, a new payment of $1,051.65 is computed by ap-
plying the 3 percent annuity factor for 29 years to the $20,179.61. Note
that the payment is 3.06 percent greater than the payment in year 1, ap-
proximately the rate of inflation.~4 15
DiJ]erences from the PLAM. The debiting rate could not be expected
to equal the rate used as payment factor plus the rate of inflation in every
~4It would have been exactly the same if we had used the precise nominal rate rather
than the simple sum of the real rate and the rate of inflation.
~SIn order to be consistent with the PLAM illustration, the payments are computed so
as to be constant in terms of purchasing power at the time they are scheduled (the beginning
of the year in our example). In the case of the PLAM this was necessary if the borrower re-
quired one-period advance notice of the exact money payment. With the constant-payment-
factor VRM, however, payments could be set to be roughly constant in purchasing power
terms at the time of payment since the interest rates, which implicitly forecast inflation, are
known at the start of the period. Payments would vary only in that debiting rates did not
properly forecast inflation. The minor increases over time in the payment stated in terms of
purchasing power result from the fact that the compounding of interest and inflation is ig-
nored in the example and thus payments rise to adjust for the minor discrepancy.
year, though one would expect this relation to hold approximately, and
on the average, as long as the payment rate had been chosen judiciously.
If, in fact, the debiting rate for a given year differs from the constant pay-
ment factor plus the inflation that actually materialized in that year, the
annual payment in successive years will differ from that under PLAM --
and hence will not quite be constant in terms of purchasing power -- but
the differences will tend to be small.~6 Furthermore, such differences
would not produce serious consequences since the effective cost to the
borrower would be unaffected. The interest rate paid on the balance is the
same, namely the debiting rate. The choice of the payment rate affects
only the path of periodic payment and hence the path of repayment of
principal. If the rate chosen is too low, the repayments are more gradual
than expected, being initially lower and eventually high -- and conversely,
if too high. The evidence for the United States cited earlier, indicating a
reasonably stable real interest rate over the postwar period, suggests that
it should not be difficult to select a payment factor such that the resulting
stream of payments will be approximately level in real terms.
The conclusion that moderate "errors’~ in the choice of the payment
factor would not produce serious consequences for either the lender or the
borrower, when combined with the evidence that the real rate is quite sta-
ble, has one implication of considerable practical importance: an in-
stitution that chose to offer a constant-payment-factor VRM could afford
to post a payment rate that changed at very infrequent intervals if at all.
The convenience of such an arrangement should be obvious.
If the lender were anxious to avoid the risk of too slow a rate of re-
payment and/or the borrower were anxious to avoid the risk of his pay-
ment stream rising in time, one could readily reduce the risk to any de-
sired extent by choosing for the payment rate an upward-biased estimate
of the real rate. This would of course imply a higher initial payment, and,
on the average a correspondingly declining real payment stream.17 Fur-
ther, this option would be greatly preferable to the traditional mortgage in
which both the initial payment and the anticipated rate of decline are de-
termined by the happen-chance of inflationary expectations.
Flexibility of the Constant-Payment-Factor VRM. One further fea-
ture of the constant-payment-factor VRM should be noted. By in-
tentionally setting the payment rate different from the estimated real rate,
~6To a very good approximation, a 1 percent deviation of the debiting rate from the
sum of the payment rate plus the rate of inflation will result in the annual payment rising by
1 percent relative to the PLAM payment.
~TTucker [1975] advocates this approach as a means to gain acceptance of this type of
mortgage,
4O NEW MORTGAGE DESIGNS
one can approximate any desired rate of graduation in real terms. Setting
the
a rea~l
payment
8 payment
rate atstream
x percent
withabove
a declining
or below
or the
rising
realtrend
rate would
of x percent
result in
per
year.
Further flexibility is to be obtained through appropriate choice of the
debiting rate. If intermediaries issued term deposits of substantial length,
say three to five years, then they could afford to offer a borrower anxious
to minimize changes in the debiting rate, a contract in which the debiting
rate would itself be fixed for that length of time. If that length were say
five years, then over that period the contract would behave precisely like a
GP mortgage in nominal terms, with the annual payment rising over the
term of the debiting rate at a predetermined rate equal to the difference
between the fixed payment rate and the fixed debiting rate. Of course
while this arrangement would eliminate uncertainty about money pay-
ment, it would correspondingly enhance uncertainty about real payments;
yet for reasonably short periods of time, the uncertainty of inflation may
be fairly limited and households may be more able to estimate their
money income over such a span. In such circumstances the use of a medi-
um-term fixed debiting rate may serve to reduce risk.
It is apparent that with such arrangements, thrift institutions could
themselves offer an array of short-term and longer-term deposits, match-
ing their asset maturity structure, and could always afford to pay rates
competitive with the market, as these would be the rates they would in
turn earn on their assets. The scheme is thus fully consistent with the in-
termediaries performing the function for which they were designed, while
eliminating the supply effects of inflationJ9
To summarize, the constant-payment-factor VRM relies on two basic
ingredients: a payment factor related to the "real" rate and hence inde-
pendent of the rate of inflation, and a variable-debiting rate tied to an ap-
propriate market rate, with maturity related to the frequency of rate re-
visions. By combining these ingredients in different ways one can readily
put together a wide variety of specific contracts capable of suiting the
needs and preferences of both borrowers and lenders, providing thereby a
solution to many of the present problems of housing and of the thrift in-
stitutions. The instrument achieves this result because it combines the de-
sirable features of a VRM from the point of view of the lending
~SAs noted earlier the same result could be achieved with a PLAM.
~gA variant of this instrument considered in the Cohn-Fischer paper involves a periodic
payment which is fixed at the outset in terms of purchasing power and thus a variable matu-
rity. Because large sustained discrepancies between the rate of inflation and the inflation pre-
mium reflected in the debiting rate are unlikely, this variable-maturity instrument does not
suffer from the difficulty which was outlined in connection with variable-maturity VRM in
Section III.B.
intermediaries with.the main positive aspects of the PLAM from the point
of view of the borrowers.2°
These considerations lead us to conclude that while the PLAM is in
some ways the most straightforward, rational solution of the problem in
an abstract sense, the constant-payment-factor VRM provides an alter-
native which is not significantly inferior in any sense and is superior in
many respects, particularly in terms of its ease of implementation in the
light of existing institutions and attitudes.
IV. TRANSITION PROBLEMS
The adoption of either the PLAM or the constant-payment-factor
VRM (or any other VRM for that matter) would allow lenders to better
match asset and liability maturities, thus reducing the periodic profit
squeezes and related problems that have contributed to interruptions in
mortgage supply.~ However, supply difficulties will be resolved fully only if
deposit rates paid by institutions are competitive, i.e., sufficiently high to
attract the deposits needed to satisfy mortgage demand at the deposit rate
plus an equilibrium spread without resorting to outright rationing or to
indirect rationing .devices such as very high downpayments and excessively
strict lending standards. If rate ceilings continue, or if rates are repressed
in any other fashion, fluctuations in supply will continue although thrift
institutions might no longer bear much of the brunt.
A major obstacle to competitive deposit rates is that most thrift in-
stitutions still have large proportions of their assets tied up in low-yielding
fixed-interest rate mortgages. Therefore, an immediate shift to fully com-
petitive -- and presumably on the average higher -- deposit rates, would
worsen their profit position and would threaten the solvency of many of
them. It is for this reason that we have seen a number of proposals, such
as tax exemptions for interest paid on thrift institution deposits, which
would increase their ability to compete for funds without threatening their
profitability or their solvency.
2°There are some differences between the constant-payment-factor VRM and the
PLAM which should be recognized, and which depend in part on the specific form of the
constant-payment contract. If the borrower opts for a short-term debiting rate, he ends up
by paying over the life of his contract a real rate equal to the average rate which actually
materializes over that life. That rate is of course uncertain and need not coincide with the
payment factor. By contrast, under a PLAM the real rate ~’s the payment factor and is thus
fixed and known in advance. Furthermore, under PLAM the periodic payments, are by con-
struction, constant in terms of purchasing power (as measured by the reference price index)
whereas with the alternative instrument they would exhibit at least some fluctuations because
of fluctuations in the realized real rate. Accordingly, the PLAM might be a somewhat pref-
erable instrument for the majority of borrowers in that it would enable them to hedge
against future movements of the real rate. The alternative contract would be superior only
for those who had reason to expect a positive association between their real income and the
real rate. While this disadvantage relative to PLAM should be acknowledged, we do not be-
lieve that it is a major one.
Such proposals -- as well as attempts to protect thrifts by main-
taining deposit rate ceilings even if alternative mortgages are adopted --
create distortions in current financial transactions in order to avoid the
consequences of past errors. Further, they would very likely be planting
the seeds for future supply crises if conditions changed. A superior ap-
proach would be to deal directly and separately with the problems arising
from past practices and allow current transactions to take place on a
sound basis. It seems clear to us, at least, that the entire burden of this
adjustment should not be imposed on the thrift institutions. While part of
the current problem no doubt can be blamed on their shortsightedness, it
is quite clear that it resulted primarily from behavior patterns forced on
them by government regulation as well as major changes in the economic
environment over which they had no control.
It would seem that to achieve a rapid phasing-out of rate ceilings
would require not only the adoption of new types of mortgages along the
lines presented in section IV but also some form of one-time government
transfers to compensate institutions for the losses they would incur in the
short run and thus maintain their solvency. While such a subsidy program
might appear to be expensive, its cost would probably be modest when
measured against that of wild gyrations in construction and the fact that
an increasing proportion of Americans cannot acquire adequate housing.2~
Clearly, there are many issues which will have to be dealt with in the
transition to new mortgage lending patterns. The new instruments would
have to be described in tetans intelligible to consumers so that they can
make appropriate choices. In particular, since they would presumably face
a variety of choices, they would have to give careful attention to the bene-
fits and costs of alternative features including prepayment provisions, the
level of initial payments and the potential variability of payments. In a
similar vein, lenders would have to rethink credit standards, down-
payments, and desirable real payment patterns for different types of
households. Since thrift institutions would face a situation in which cash
inflows might fall short of accounting income, especially during early
years of the transition, changes would be required in liquidity planning
and might require further recourse to advances or secondary market oper-
ations. Along similar lines, regulatory authorities would undoubtedly have
to rethink reserve and liquidity requirements for thrift institutions in re-
sponse to different asset characteristics. Details of mortgage design, in-
cluding the choice of appropriate reference rates for VRMs or price indi-
ces for PLAMs, adjustment intervals, and so on would have to be worked
out.
Z~Such a subsidy would have a more favorable impact on the distribution of income
than tax exemptions on thrift deposits. It would benefit all depositors proportionately rather
than providing the greatest benefits to those in the highest income brackets.
Figure :3
CLASSiFiCATiON OF MORTGAGE DESIGNS BY EXTENT TO WHICH THEY ELIMINATE
Alleviate ’
PLAM~
Demand Effects
Constant-
Payment-Factor
VRM
(~raduated-Payment
t
Mortgage
Worsen Supply Standard Mortgage Alleviate

Supply Effects
Effects
Dual-rate VRM-~e
FHLBB VRM~
E;tand erd VRM~e
Worsen Demand
Effects
price-level adjusted deposits are issued.

Inasmuch as these transition issues were not part of the study, we do
not pretend to present a concrete set of recommendations. However, it is
Clear that they must be dealt with in relation to any potential changes in
patterns of mortgage lending.
V. CONCLUSIONS AND RECOMMENDATIONS

The analyses summarized in this introduction and detailed in the fol-
lowing five papers support the conclusion that the standard mortgage has
been a major contributor to the problems which have plagued housing
during the recent inflationary period. Further, they provide the basis for
the hopeful conclusion that innovations in mortgage financing could sub-
stantially alleviate these problems, eliminating the need for further resort
to housing subsidies or to greater direct government intervention,z2
Alternative mortgage designs were analyzed along two dimensions: 1)
the extent to which they resolve the demand problem by eliminating in-
flation-related distortions in the time pattern of real payments and 2) the
extent to which they resolve the supply problem by allowing closer asset-
liability matching. The position of the various instruments along these di-
mensions is shown in Figure 3. Of all the mortgage innnovations studied,
only the price-level-adjusted mortgages and the class of variable-rate
mortgages with smoothed real payment streams (of which the constant-
payment-factor VRM appears to be best) rate well on both dimensions.
Based on these analyses, we offer four recommendations which should
be considered as a package. These are:
1. Price-level-adjusted mortgages and/or variable-rate mortgages
with constant-payment factors should be offered to the public.
Federal and state regulations, as well as institutional practices,
should be changed where necessary to allow for these instruments.
2. Thrift institutions should maintain a much closer balance between
asset and liability maturities by both shortening effective asset
maturities through PLAMs or VRMs (hopefully with constant--
payment factors), and lengthening liability maturities through
more extensive use of term deposits and mortgage bonds.
3. Regulation Q ceilings should be abandoned as quickly as possible
in order to restore the allocative mechanism of financial markets
and reduce fluctuations in the supply of funds through traditional
mortgage lenders.
2~We refer here to general mortgage subsidies which are likely to benefit largely those
groups that are able to borrow most as opposed to subsidies or other mechanisms targeted
at income groups which could not afford adequate housing even with appropriate in-
novations in mortgage financing. We wish to stress that subsidies should not be wasted in
correcting problems which can be dealt with more efficiently and at lower cost through fi-
nancial innovation.
Some form of once and for all subsidy (or other form of public
intervention) should be granted to thrift institutions which will
erase past mistakes and will not penalize housing and depositors
of these institutions for past errors of financial policy.
REFERENCES
Modigliani, Franco. "Some Economic Implications of the Indexing of Fi-
nancial Assets with Special Reference to Mortgages." Forthcoming
in Proceedings of the Conference on the New Inflation (Milano,
Italy, June 1974).
Poole, William. "Housing Finance Under Inflationary Conditions." In
Ways to Moderate Fluctuations in Housing Construction, Board of
Governors of the Federal Reserve System (1972).
Tucker, Donald. "The Variable-Rate Graduated-Payment Mortgage."
Real Estate Review (Spring 1975).

Inflation and The Housing Market: Problems and Potential Solutions

Uploaded by

Copyright:

Available Formats

Inflation and The Housing Market: Problems and Potential Solutions

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Inflation and The Housing Market: Problems and Potential Solutions

Uploaded by

Copyright:

Available Formats

Inflation and the

Donald Lessard and Franco Modigliani*

Outline of the Study

*Donald Lessard is an Assistant Professor of Management and Franco Modigliani is

II. THE CAUSES OF RECENT INSTABILITY IN THE

of single-family houses; and (2) overwhelming dependence for mortgage

Our conclusions that inflation has an unfavorable effect on the

REAL VALUE OF MONTHLY PAYMENTS

Source: Donald Tucker, "The Variable-Rate Graduated-Payment Mortgage"

interest rate in the absence of inflation, the annual payment is $1,020. As

REAL VALUE OF OWNER’S EQUITY AND UNPAID BALANCE

Source: Donald Tueker, "The Variable~Rate (~raduated-Payment Mortgage"

A faster repayment schedule and the resulting higher ratio of payment

3Lengthening maturity from prevailing practices could accommodate, at best, only ~

present value of the stream of real payments could deviate significantly

5Government funds, in particular purchases of mortgages by the FNMA, have played

The alternative to the traditional mortgage that has received by far

All of these proposals relate directly or indirectly to a basic dilemma

The dual-rate VRM endeavors to resolve the above dilemma by using

This instrument may be thought of as a variant of the dual-rate VRM

~Donald Tucker [1975] advocates a graduated-payment VRM with either a fixed

CLASSiFiCATiON OF MORTGAGE DESIGNS BY EXTENT TO WHICH THEY ELIMINATE

Worsen Supply Standard Mortgage Alleviate

price-level adjusted deposits are issued.

V. CONCLUSIONS AND RECOMMENDATIONS

You might also like