View original document

The full text on this page is automatically extracted from the file linked above and may contain errors and inconsistencies.



March 1, 1991

Consumer Sentiment
and the Economic Downturn
From August through October 1990, the University of Michigan's survey of consumer sentiment
recorded the biggest decline in any three-month
period in its 44-year history; and it has stayed
low since then. A similar index constructed by
the Conference Board began to decline in April
and has dropped even more. Analysts have used
these indexes to forecast future consumer expenditures, to forecast the future of the economy as
a whole, and to understand the motives and attitudes underlying consumer behavior. Indeed, the
large drop in these indexes accurately forecast
the current consumer-led downturn.
This Weekly Letter focuses on the usefulness
of the sentiment indexes for forecasting consumer
expenditures. It discusses the concepts that underlie the sentiment indexes and attempts to
reconcile them with standard economic theory.
A simple test of the indexes' ability to contribute
to forecasts of consumer expenditures is then
The psychological approach
Indexes of consumer sentiment are grounded
in the discipline of psychological economics.
Standard economic theory assumes that households will react the same way to economic
stimuli at different points in time; psychological
economics attempts to establish relationships
between specific conditions and specific forms
of behavior at specific times. For example, sometimes inflation may cause households to speed
up their spending, while at others inflation may
depress spending.

Psychological economics may be especially
useful in explaining and predicting expenditures
on consumer durables as well as other discretionary purchases, such as travel, recreation,
and entertainment, which are neither habitual
nor generally made on the spur of the moment.
It suggests that consumer purchases of important
durable items depend not only on consumers'
ability to buy-represented in standard economic theory by income, assets, and availability

of credit-but also on a willingness to buyrepresented by favorable expectations and
The index of consumer sentiment uses sample
surveys to measure consumer willingness to buy.
Respondents answer questions about their individual financial situat'ions and their views of general business conditions. George Katona and his
associates at the University of Michigan found
that while no single survey question could provide a reliable indicator of subsequent purchases,
a set of questions taken together could do so.
Economists have tried to explain consumer sentiment by relating it to such things as oil prices,
inflation, and unemployment. However, these
relationships have not proved to be very stable.
The economic approach
The life-cycle hypothesis is a widely accepted
economic theory used to predict consumption.
According to this theory, households consciously
attempt to achieve a preferred distribution of
consumption over their lifetimes, subject to the
amount of resources they expect to have. These
resources consist of total expected future labor
incomes plus current net worth. This theory
subsumes a concept like "consumer willingness
to buy" in the notion of "tastes" which are
assumed to be relatively constant.

Expectations about future labor incomes are
usually assumed to be formed "adaptively" on
the basis of recent experience. In other words,
expectations of income are measured by an
average of current and past incomes. As far as
net worth is concerned, although stock market
wealth can be measured at its current market
value, it is more difficult to measure the current
value of other forms of wealth, such as owneroccupied housing.
The life-cycle model makes an important
distinction between consumption and consumption expenditures. Total consumption refers
to the current flow of satisfaction from consumer

goods, including the existing stock of durables.
This flow of satisfaction can be quantified as the
rental value of the current stock of consumer
durables plus current household expenditures
on nondurables and services. In contrast, total
consumption expenditures are simply the actual
purchases of consumer durables plus nondurables and services during a specified period
of time.
In this theory, the stock of durables that
consumers demand depends upon total current
consumption and the relative price of durables,
which in turn depends on interest rates. When
interest rates are high, the relative price of durables increases because of either higher borrowing
costs or the larger amount of income that is forgone by purchasing durables. Actual purchases
of consumer durables then result as households
gradually adjust the actual stock of consumer
durables to the stock demanded.

A reconciliation
One possible avenue for reconciling the lifecycle theory of consumption and psychological
economics is that a household's perception of its
total resources is partly psychological. Katona
and his associates found that the households
were not able to predict their incomes in the
near future very well. Also, general attitudes
about their financial well-being predicted their
near-term expenditures on durables better than
their expectations of future incomes did. Consumer attitudes thus appear to be of some
If consumer sentiment matters, then it ought to
help to explain a portion of consumer spending
at any point in time that is not accounted for by
the life-cycle model. More specifically, it ought
to help to explain the life-cycle model's prediction errors.
The accompanying chart plots the two indexes
of consumer sentiment in the upper panel, and
the prediction errors (as a percent of trend) of
the life-cycle model for expenditures on durables, expenditures on all consumption goods,

Indexes of Consumer Sentiment and Forecasting Errors *


Conference Board

40 -i-~;--~I"""""'"'f~i.........r-~-r--r--r
Forecast Errors
as Percent of Trend


Spending on
Consumer Durables



Total Consumption




Total Consumption


* Shaded areas indicate economic recessions.




and total consumption in the lower panels. The
prediction errors are measured as actual values
less the values predicted by the life-cycle model.
The chart shows, and statistical regression analysis confirms, that the errors in predicting spending on consumer durables and total consumption
expenditures are significantly related to the
sentiment indexes, both contemporaneously
and up to three quarters later. But the error
in predicting total consumption is not.
In the life-cycle model, total consumption
depends upon perceived income and wealth.
Since the sentiment indexes do not help to explain the life-cycle model's errors in predicting
the level of total consumption, they appear to be
something other than just a better reflection of
perceived income and wealth. The important
thing that they seem to measure is household
perceptions of uncertainty, or confidence, and
the corresponding probability of financial distress. If the probability of financial distress is
high, households would prefer holding liquid
financial assets and spending the income from
them on nondurables and services rather than
holding illiquid consumer durables. They would
therefore allocate their consumption away from
the satisfaction provided by illiquid consumer
durables and towards nondurables and services.
The life-cycle model does not capture this aspect
of household behavior because it abstracts from
uncertainty and risk.
Even though consumer sentiment does not
affect the level of total consumption, the fact
that it leads spending on consumer durables is
useful not only for forecasting durables but also
for forecasting GNP. The reason is that aggregate
spending in the economy, and hence income
and employment, depend upon total consumption expenditures rather than total consumption.
As the chart shows, and statistical regression
analysis confirms, the indexes of consumer
sentiment also lead the life-cycle model's errors
in predicting total consumption expenditures. As
a result, an index of consumer sentiment can be
used to improve the forecast of total consumption
spending; and since consumption expenditures
are about two-thirds of GNP, forecasts of GNP
would be improved also.
The finding that consumer sentiment leads total
consumption expenditures is consistent with our
other findings. Suppose the average consumer is
afflicted by a decline in confidence. As a result,

he spends less on consumer durables, acquiring
a smaller stock than he would have otherwise,
and also getting proportionately less satisfaction
from his stock of durables. Since the decline in
confidence does not affect total consumption,
then our consumer must increase his spending
on nondurables and services. But the increased
spending on nondurables and services isn't
matched dollar for dollar with the lower spending on durables. Instead, the consumer increases
his spending on nondurables and services by the
amount of satisfaction lost due to his lower stock
of durables. The value of the lost satisfaction from
durables is only a fraction of the decline in expenditures on durables. (This fraction equals the
gross rate of return on the stock of durables, or
the depreciation rate plus a real rate of interest.)
So the increase in spending on nondurables and
services tends to be less than the decrease in expenditures on durables. As a result, the decline
in consumer sentiment tends to lead to a decline
in total consumption expenditures.

Swings in consumer sentiment produce lagged
changes in consumer spending on durables that
are not fully offset by opposite changes in spending on nondurables and services. As a result,
consumer sentiment indexes are useful for forecasting total consumption expenditures and GNP.
Although the Conference Board's index shows
wider swings than the Michigan index, the estimated sensitivity of consumer spending to it is
smaller, so that the predicted swings in consumer
spending are similar for the two indexes. Sharp
drops in the two indexes in the third quarter
accurately forecast large declines in expenditures
on consumer durables, and hence reductions in
total consumption expenditures and GNP in the
fourth quarter. Also, because of the lags involved,
the indexes forecast further declines in consumer
spending in the current quarter.
The evidence examined here suggests that
consumer sentiment is more than just a better
reflection of perceived income and wealth. Instead, it helps to measure household perceptions
of uncertainty and the corresponding likelihood
of financial distress, which affect the time patterns of consumer purchases of durables. Thus,
the most important attitude being measured by
the senti ment indexes appears to be confidence.

Adrian Throop
Research Officer

Opinions expressed in this newsletter do not necessarily reflect the views of the management of the Federal Reserve Bank of
San Francisco, or of the Board of Governors of the Federal Reserve System.
Editorial comments may be addressed to the editor (Judith Goff) or to the author.... Free copies of Federal Reserve
publications can be obtained from the Public Information Department, Federal Reserve Bank of San Francisco, P.O. Box 7702,
San Francisco 94120. Phone (415) 974-2246.

Research Deportment

Federal Reserve
Bank of
San Francisco
P.O. Box 7702
San Francisco, CA 94120