FRBSF Economic Letter
2002-13; May 3, 2002
House Price Dynamics and the Business Cycle
 |
It is somewhat surprising that house prices in most parts of the nation
have stayed high despite the downturn in the economy. As Figure 1 shows,
real (inflation-adjusted) house price changes became negative with GDP
growth during the last recession in 1991; but this time, they have remained
positive and appear to be firm. The strength in the housing market also
contrasts vividly with the declines in the stock market over the past
couple of years. Indeed, many of the regions where house price appreciation
has been strongest over the past two years have large concentrations of
high-tech firms; tumbling share prices and job losses in the high-tech
sector would seem to represent a large shock to housing demand, yet year-over-year
price changes in places like San Francisco remain stubbornly positive.
Declining mortgage interest rates may have helped to support prices. But
interest rates also fell around the time of the last recession, and that
did not prevent real house prices from declining.
In this Economic Letter, I outline some of the empirical facts
about house prices that have been documented in the real estate literature
and explore some of the determinants of supply and demand that lead to
house price dynamics. After focusing particularly on demand, I attempt
to reconcile a current estimate of housing demand with our position in
the business cycle.
House price dynamics
Figure 2 helps illustrate two well-known facts about house prices. The
figure plots real and nominal quarterly house price changes for the U.S.,
based on data from the Office of Federal Housing Enterprise Oversight's
house price index. First, although it is relatively unusual to observe
declines in nominal house prices—at least at the national level—declines
in real prices are relatively more common. One important point to keep
in mind, however, is that while declines in nominal house prices are relatively
rare, the volume of housing market transactions tends to be more responsive
to a slowing economy. A flattening out of an observed price series may
in fact mask a buildup of inventory of unsold houses.
 |
Second, real house price changes appear to be persistent; that is, positive
price changes tend to be followed by more positive price changes, and
vice versa for negative price changes. This observation has stoked considerable
interest amongst researchers because, normally, we expect asset prices
to adjust immediately to reflect new information about fundamental value,
not gradually over time (see Meese and Wallace 1994 for some important
empirical work on this issue). Persistence in house prices could indicate
that housing markets are inefficient, either in the sense that the market
takes time to clear, or that prices and expectations of future price changes
are set in a backward-looking manner. An alternative explanation for the
persistence in house prices is that prices depend directly on economic
variables, such as job growth and changes in personal income, that are
themselves persistent.
Supply and demand in housing markets
Simple economic models demonstrate how changes in house prices come about
from changes in supply and demand. Disentangling supply and demand shocks
and assessing their impact on house prices is, of course, a challenging
empirical problem in general. For assessing the current situation, however,
this "identification" problem may be mitigated by the fact that
there appears to be little evidence of sudden supply shocks over the past
couple years. Figure 3 shows that the supply of newly completed single
family homes rose steadily through the 1990s, peaking at the end of the
decade, and has remained close to its long-run average ever since.
 |
One of the most interesting features of Figure 3 is the apparent change
in the behavior of new construction beginning in the early 1990s. Before
then, the series was volatile, and thereafter it has been relatively smooth.
This 1990 breakpoint may not be coincidental. Following the banking crisis
in the late 1980s and early 1990s, capital market discipline and scrutiny
from banking regulators may have acted as a check to the boom and bust
dynamics that seem to characterize development in prior periods. In that
case, it is possible that developers have been constrained in their ability
to bring new housing stock to market and may not have overbuilt during
the economic expansion. If housing supply has leveled off while housing
demand has remained strong, then that interplay helps explain why house
prices may have remained firm at this point of the business cycle.
What about demand? The demand for housing likely depends on variables
such as job growth, growth in personal income, changes in demographics,
and the cost of buying housing. This cost includes the actual price and
the financing cost. Most studies of housing demand recognize the user
cost of housing capital (for example, Poterba 1984) as an important determinant
of financing costs. Decreases in the user cost are thought to accompany
increases in the demand for housing and vice versa. Technically, the user
cost measures the willingness of a household to trade off housing consumption
for nonhousing consumption over the course of a single period (i.e., the
user cost is a marginal rate of substitution). In competitive settings,
the user cost is simply the per-period rental cost of the housing asset.
The principal advantage of looking at the user cost, as opposed to looking
only at changes in rents, is that the user cost offers more insight into
which variables are driving changes in costs and, hence, driving changes
in demand.
A simple model of the user cost
A typical specification of the user cost includes a mortgage interest
rate, a property tax rate, a maintenance cost, a deduction for depreciation,
and a term reflecting expectations about future price appreciation. The
interest payment and property taxes terms are calculated on an after-tax
basis. Expected appreciation serves to lower the user cost, because, if
that appreciation eventually comes to pass, it will benefit the owner
when she sells the house and can conceivably allow her to alter her current
decisions about saving and consumption.
The user cost would be extremely easy to measure were it not for the
inherently unobservable expectations term. In this model, I define expectations
as the four-quarter moving average of past price changes. Clearly, these
are naïve expectations. However, I have found that the results based
on more sophisticated measures are not markedly different from those based
on the crude measure used here.
The analysis below is based on a "stripped-down" model of the
user cost that emphasizes the three key variables that contribute to its
dynamics: taxes, mortgage interest rates, and expectations. For this exercise,
I focus on quarterly changes. Thus, the interest rate series is converted
to a quarterly rate. The marginal tax rate used is the top marginal rate
prevailing at the time of the observation.
 |
Figure 4 presents estimates of this stripped-down version of the user
cost. Much of the short-run variation in the user cost comes from changes
in mortgage interest rates and expected appreciation rates. High mortgage
rates in the early 1980s put significant upward pressure on the user cost.
But note that the higher mortgage rates in the late 1980s can be partially
offset by expected appreciation rates. For the current situation, it is
interesting to note that this particular specification of the user cost
indicates much higher demand at present than during the last recession
in 1991. This is not entirely due to the decline in mortgage interest
rates. Indeed, after-tax mortgage interest rates fell faster during the
last recession than during the current one. In this model, the strong
demand at the current time is due primarily to strong expectations for
future price appreciation.
This result raises the question of whether house prices are, in fact,
backward-looking or forward-looking. In the model, they were assumed to
be backward-looking; specifically, expectations for appreciation in the
next quarter were defined to be equal to the average appreciation over
the past four quarters (as argued above, more sophisticated techniques
that use different moving average lengths or incorporate additional economic
data yield the same basic results). Short of conducting a survey of housing
market participants, economists are forced to estimate expectations using
past data, and this reliance forces us to contend with the strong persistence
in the data as seen in Figure 2.
The question of whether true expectations are forward-looking or backward-looking
is extremely important, however. If expectations are forward-looking,
then today's prices are likely to be justified and will continue to hold
up until news arrives that makes people change their expectations. If
not, then the market will eventually "learn" that fundamentals
are weak and prices will adjust.
Conclusion
House prices in the U.S. appear to be more firm than they were at this
phase of the last business cycle. According to one measure of housing
demand, the reason house prices are firm is not so much because of the
drop in mortgage interest rates as because of expectations. On its own,
this conclusion—prices are high today because they are expected to be
high in the future—would be unsettling. But the strong expectations in
the housing market are mirrored in other parts of the economy. These expectations
could, of course, be proven wrong. However, indicators ranging from interest
rate differentials to trends in consumer savings (see Marquis 2002), provide
a substantial amount of economic data suggesting that investors and consumers
foresee a relatively mild recession and a return to solid economic growth.
John Krainer
Economist
References
Marquis, M. 2002. "What's
Behind the Low U.S. Personal Saving Rate?" FRBSF Economic
Letter 2002-09 (March 29).
Meese, R., and N. Wallace. 1994. "Testing the
Present Value Relation for Housing Prices: Should I Leave My House in
San Francisco?" Journal of Urban Economics 35, pp. 245-266.
Poterba, J. 1984. "Tax Subsidies to Owner-Occupied
Housing: An Asset Market Approach." Quarterly Journal of Economics
(November) pp. 729-752.
|
