^CHff^N
Digitized by the Internet Archive
in
2011 with funding from
Boston Library Consortium Member Libraries
http://www.archive.org/details/stockownershippaOOpote
DEWEY
HB31
.M415
|
working paper
department
of economics
STOCK OWNERSHIP PATTERNS, STOCK MARKET
FLUCTUATIONS, AND CONSUMPTION
James M. Poterba
Andrew A. Samwick
96-2
Oct. 1995
massachusetts
institute of
technology
50 memorial drive
Cambridge, mass. 02139
J
STOCK OWNERSHIP PATTERNS, STOCK MARKET
FLUCTUATIONS, AND CONSUMPTION
James M. Poterba
Andrew A. Samwick
96-2
Oct. 1995
mas:
m
13 1996
STOCK OWNERSHIP PATTERNS, STOCK MARKET
FLUCTUATIONS, AND CONSUMPTION
James M. Poterba
MIT and NBER
Samwick
Dartmouth College and NBER
Andrew
A.
October 1995
ABSTRACT
The market value of corporate stock in the United States increased by nearly one
between December 1994 and July 1995. This paper explores the distribution
of the stock ownership, and hence the gains from the stock price rise, and what the rise in
stock prices implies for consumer spending. It begins by noting the substantial change in the
trillion
dollars
pattern of stock ownership during the postwar period. Individual investors,
who directly
held
the early 1950s, have gradually replaced their direct stock holdings
with indirect holdings through mutual funds, pension funds, and other financial intermediaries.
It then documents the substantial predictive power of stock price fluctuations for future
consumption growth, and considers two potential explanations for this relationship. The first,
or "leading indicator," view, holds that the stock market responds to news that suggests
most corporate stock
in
consumption will rise in the future. This does not suggest any causality between stock price
changes and subsequent consumption movements. An alternative and not necessarily
exclusive view, the "wealth effect," holds that higher stock prices raise consumption by
raising household net worth, and thereby expanding consumption opportunity sets. We test
for the importance of the wealth effect by studying the effect of stock price changes on the
share of consumption devoted to luxury items, and we test for effects of changing stock price
ownership patterns on the link between stock price fluctuations and consumption growth.
We find virtually no evidence to support important wealth effects associated with stock price
changes. We also explore whether the source of stock price fluctuations, in particular
fluctuations that are related to changes in dividends or earnings rather than to changes in
discount rates, affects the predicted change in consumption that follows a stock price change.
We are grateful to Rochelle Antoniewicz, William Brainard, John Campbell, Joel Dickson, Jerry
Hausman, Greg Mankiw, Tom Morley, George
Perry, Chris Probyn, James Shapiro, Martha
Andrei Shleifer, David Shulman, Chris Sims, David Wilcox, and
participants in the Brookings Panel meeting and at a seminar at the Federal Reserve Board, for
helpful discussions, to Alex David for data on share repurchases, and especially to Arthur
Kennickell for assistance with the 1992 Survey of Consumer Finances. This research was
supported by the National Science Foundation.
Starr-McLuer, Robert
Shiller,
The
market
bull
has raised the
of the last year
United States by nearly a
many
While
trillion dollars.
corporate stock
total value of
in
the
analysts have tried to explain or interpret
the stock market's recent movements, there has been less attention to the link between rising
stock prices and real economic activity.
What
distributed across households?
investors?
How
do
rising
in
share prices
stock prices affect consumer spending?
depends on labor income and
is
are the gains from an increase
fraction of these gains accrues to a small set of wealthy
The standard textbook treatment
wealth
How
of aggregate
financial wealth.
consumption holds that consumption
1
The marginal propensity to consume out
typically taken to be approximately .04 per year.
In this
of
framework, the wealth
effect of a stock market rally should have an important stimulative effect on consumption.
Although
this
stock prices
consumption
view neglects some
may
rise
directly,
potentially important factors, notably the possibility that
as a result of a decline
many economic
in real
discount rates, which might also affect
forecasters embrace the textbook view. For example,
a recent Wall Street research report explains that "as long as asset prices are rising, the risk
of a significant drop in
consumer spending
is
small,"
and notes that the
rising
number
of
stockholders has "made real economic activity more tied to the performance of financial
assets than ever before." 2
This paper describes the changing pattern of stock ownership during the last three
decades, investigates whether changing ownership patterns have
in
fact altered the links
between stock values and consumption, and explores the wealth effect
fluctuations.
At the outset,
following a rise
prices
may rise
in
in
it
is
important to recognize that an increase
share prices could be attributable to either of
anticipation of strong
economic
activity, including
'See for example Dornbusch and Fischer (1994).
2
Shulman{1995).
two
in
of stock price
consumer spending
factors.
First,
stock
consumer spending. The
2
role of
share prices as a leading indicator
is
well
documented.
market values are not a source of subsequent changes
that subsequent changes are expected.
stock prices and consumption
values cause changes
in
is
A second, and
in
case changes
In this
not necessarily exclusive,
link
between
in
share
consumption by expanding households' resource constraints. Over
must be
operative;
we
consider whether
important effect on consumption at horizons of one to four quavers.
between the leading
prices and consumption,
stock
consumption, but merely an indicator
the "wealth effect," the possibility that changes
long horizons, such a wealth effect
distinguish
in
indicator
It
it
also has an
is
difficult to
and causative view of the relationship between share
because doing so requires identifying autonomous movements
in
share prices that are not attributable to changing expectations of future dividends or discount
rates.
This paper presents
and consumption.
substantial
It
increases
new evidence on the
summarizes the changes
in
association
in
between share
is
stock market values, and presents several tests directed at
correct, then the pattern of
If
the "leading indicator"
consumption changes following stock
should be independent of the distribution of stockownership, and there
differential
movements
consumption that have typically followed
disentangling the "leading indicator" and "wealth effect" views.
view
price
is
price fluctuations
no reason to expect
consumption responses by households that do and do not own corporate stock.
This paper presents empirical tests of both of these propositions.
We
find
little
support for
an important wealth effect of share prices on consumption. The strong positive correlation
between consumption growth and lagged stock market returns therefore appears due
to the "leading indicator" feature of stock price
We
begin by placing the
primarily
movements.
1995 stock market increase
in
context.
We
report on the
evolution of price-to-dividend and price-earnings ratios, Tobin's q, and the ratio of stock
1
3
market value to
q,
GDP
during the postwar period.
suggest that the stock market of
1
995
is
at a
Some
of these
measures, notably Tobin's
postwar valuation high. Others, notably the
price-earnings ratio, suggest a less extreme situation.
We
next investigate the fraction of stock market capital gains that accrue directly to
individual investors, in contrast to gains that accrue indirectly
such as defined benefit pension plans or
life
through financial intermediaries
A
insurance companies.
range of recent
"behavioral" models of consumption suggest that the marginal propensity to spend out of
depends not only on
different types of assets
the
way they
are held.
of capital gains
3
their risk
and return characteristics, but also on
Households may exhibit lower marginal propensities to spend out
on assets held
in
retirement plans than on assets held directly.
Popular discussions sometimes note that the fraction of corporate stock
households has declined during the postwar period and
fact, the principal
postwar trend has been from direct
is
We
Funds data that show households owning less than
fifty
combine
plans,
least
individual
962 Survey
s,
May
See the accounts
1
We
defined contribution pension
find that individuals
of stock ownership,
of Financial Characteristics of Households,
Thaler (1994) provides a
4
indirect
percent of outstanding shares.
and the
Consumer Finances. While share ownership patterns changed
3
In
have direct control over at
thirds of outstanding corporate stock.
To describe the changing incidence
1
ownership to
4
re-analyze the widely-cited Flow of
ownership of equities through mutual func
and other financial intermediaries, and
two
presently below fifty percent.
individual stock
financial intermediaries.
ownership through various
owned by
992
for
in
the
examples.
summary
we examine
1
983 and
relatively
1
little
data from the
992 Surveys
between
1
of
962
of this literature.
New York Times
.
18 July 1995 p.D21,
or Wall Street Journal
1
4
and 1983, there has been a substantial increase
the last decade.
Recent shareownership growth
in
is
the prevalence of share ownership during
the result of rising rates of indirect share
ownership. The fraction of stock held by the largest stockholders, those
of
the top one half
one percent of the distribution of equity investors, has also declined during
To develop evidence
affect consumption,
of
in
that can distinguish
between the two views
of
this period.
how
share prices
we explore the correlation between stock returns and the subsequent mix
consumer spending between luxury items and
all
other goods.
An
operative wealth effect
should imply that positive stock returns increase the share of consumption accounted for by
luxury goods.
We
consider aggregate data on several categories of consumption that are
purchased disproportionately by high-income households, including purchases of "upper
luxury" vehicles, and find
little
evidence that luxury spending rises
in
the aftermath of rising
stock prices.
We
next consider whether changing patterns of stock ownership affect the linkages
between consumption and stock market
fluctuations.
The "leading
indicator"
view suggests
that ownership patterns should not affect this relationship, while the "wealth effect" at least
admits this
in
possibility.
We explore
the effect of changes
in
stock prices, as well as changes
the dividend-price and earnings-price ratio, on various measures of consumption. 5
We
recognize that stock prices and consumption are jointly determined, and simply try to describe
the typical pattern of economic activity following substantial stock price movements.
results suggest that
changes
consumption spending.
A
in
5
Fama (1981),
equations
stock prices have significant predictive power for future
permanent
magnitude as the price increase
Our
in
the
17%
stock price
first six
months
of
rise,
one of roughly the same
1995, forecasts an increase
in
Fischer and Merton (1984), and Barro (1989) also estimate reduced form
the predictive power of stock price movements for various
measuring
macroeconomic aggregates.
5
consumption by about 1.1%
in
mid- 1996 relative to what
would otherwise have been. The
it
stock market increase forecasts particularly large increases
in
consumer spending on new
autos and other durables.
We
find
little
evidence to suggest that the
shift
from direct to indirect ownership of
corporate stock has altered the link between stock price fluctuations and consumption
spending, and
little
evidence more generally of an important wealth effect on consumption.
Unfortunately, the time-series variation
the pattern of corporate stock ownership yields
in
To address
tests with low statistical power.
this limitation,
we
also use household survey
data from the Panel Survey of Income Dynamics to compare the correlation between
consumption growth and stock market fluctuations
We
holdings.
through
find
thrift plans,
some evidence
such as 401
with direct and indirect
that the consumption of individuals
(k)s,
movements than the consumption
for stockholders
403(b)s, and ESOPs,
of those
who do
is
more
who
hold stocks
sensitive to stock price
not hold any stock, but once again, the
avaiable tests have low power.
We
do not
find
any evidence that the effect
on the source of stock price movements.
substantial
body of research
in financial
change the value of the dividend-price
several years.
for further
1
.
We conclude
of share prices
This finding
somewhat
economics that suggests
surprising, given the
price fluctuations that
or earnings-price ratio are reversed over a period of
with a brief
summary
of our findings
and a discussion of topics
work.
Recent Stock Market Fluctuations
in
Perspective
The stock market has climbed to record heights
after the
is
on consumption depends
Dow
Jones
Industrial
Average
first
in
the last year.
reached the historic
4000
In
the six months
level,
it
climbed
6
another
700
rose nearly
points.
Between January
17%. 6 Although
the
news media has
recent returns are not extraordinary.
1
993, the
real return
In
To provide background
1
twenty
presents several
for analyzing the
summary
annual percentage change
on the market because
real
it
in
1
historic,
926 and
In five
years,
The
statistics.
in
aggregate effects of stock price movements.
1
994
first
column shows the
dollars.
real
value of the
The second column presents the
the index. This column does not correspond to the return
excludes income from dividends. The entries
($1994) value of corporate shares
at the
in
in
column two confirm
1995.
The
third
column
end of each year, as reported
Flow of Funds accounts. Real equity values increased by more than
six
market as
between
of the sixty-eight years
the presence of other years with returns comparable to those
shows the
bull
7
Standard and Poor's 500 Index, measured
real
depicted the recent
on stocks of large corporations exceeded twenty percent.
the real return exceeded forty percent.
Table
and June 30, 1995, the Standard and Poor's 500
1
1 .1
trillion
in
the
dollars in the
months ended June 30, 1995. 8
While the increase
in
share prices during the last year
is
not unprecedented,
measures of stock market valuation do suggest that the stock market
war
high. Table 2 reports four different valuation measures.
of the
The
first
is
some
currently at a post-
column shows the
market value of corporate stock to gross domestic product. On June 30, 1995,
"We
calculate this as ln(539.4/455.2)
7
These
8
The change
statistics are
ratio
this
= .1697, which we approximate as 17%.
based on Ibbotson Associates (1994).
in the market value of equity during any period reflects the change in the
value of the shares that were outstanding at the beginning of the period, plus the value of any
new shares issued during the period. If firms are issuing substantial amounts of new equity,
then changes in the market value of stock can overstate the rate of share price appreciation.
In each quarter of 1994 and the first two quarters of 1995, however, nonfinancial
corporations were net repurchasers of shares, so this concern does not apply.
7
was
ratio
1
.039, a level that has been exceeded only once
This ratio has more than doubled
of similar
in just
magnitude, between the early
950s.
over ten years. The table records one previous
move
1
950s and
the early
1
it
also plotted
is
in
Figure
The recent stock
historic highs.
1
.
late
it
is
ratio for the
price rise
mid-1995 (16.3)
in
substantially lower than at the end of
1
991 (26.2)
Because earnings fluctuate substantially from year to year,
or
992
1
15.7
P/E„
in
although
was 22.3
1990.
it
is
it
at the
The June 1995 value
is
1
in
in
1980,
the postwar period.
965. Nevertheless,
postwar
(22.8). Since
80%.
of real earnings.
S&P 500. The
1.9
1
higher than any year-end value
not the highest value recorded
reached 23.3 at the end of
Stock
can be helpful to construct
of real earnings for the
end of June 1995, up from 9.8
its
risen almost
measures that divide share prices by a moving average
We did this with a ten-year arithmetic average
ratio of
S&P 500
above
is
,
alternative valuation
doubled. 9
has coincided with rapidly increasing corporate
991 share prices have increased by nearly 30%, while earnings have
1
ratio
P/E ratios do not suggest that stocks are currently at
earnings, so that while the price-earnings ratio
average value,
1
960s, when the
The second column shows the year-end price-earnings
Index;
968) since the early
1
(in
in
in
resulting
1985, and
the
1990s 10
,
This measure of P/E avfl
this valuation indicator
suggests that stock
prices are high relative to historical patterns.
The
rise in
corporate earnings
is
evident
in
the national income accounts, where
corporate profits adjusted for capital consumption and inventory valuation have increased as
a share of
GNP
from 5.6%
in
1991 to 7.2%
in
1994. The rate of return on tangible assets
has also increased. Table 3 presents recent estimates of a standard measure of the pretax
9
A similar pattern emerges if we consider the market value of corporate equity plus an
estimate of the market value of corporate debt, computed by capitalizing corporate interest
payments by the BAA bond rate, relative to GDP.
10
The values
of
P/E„B for
1
990-1 994 are
1
5.7,
1
8.4, 20.3, 21 .2, and
1
9.9, respectively.
Figure
o
Price-Earnings
&
1
Pnce-Qmaenc
40
35
30
25
20
15
10
5
H
1950
1955
1960
19B5
1970 1975
Year
19B0
p/E and P/D Ratios, 1947-1995
1985
1990
1995
8
on the tangible assets
rate of return
shown
of return
in
the
first
of nonfinancial corporations.
11
In
addition to the rate
column, the table also reports a cycle-adjusted rate of return. To
construct the adjusted return,
we
first
regress the rate of return on the civilian
rate, using a first-order serial correlation correction
unemployment
with an autocorrelation coefficient p. This
yields
= 0.124
Return
0.0052*RU
-
p = .845
t
(0.0099) (0.0013)
We
then compute
fitted
12
The
values at the sample average unemployment rate of 5.61 %.
adjusted and unadjusted series are plotted
The unadjusted
1
it
is
1
989 and
1
in
1
995
is
not as high as
it
between 1991 and
was throughout
the last twenty-five years. The change
994
is
the
the cycle-
in
less dramatic than that for the unadjusted series,
suggests an increase of 1.7 percentage points. 13
still
In spite of this
last
Figure 2.
994 and
in 1
higher than at any point
adjusted return between
but
in
rate of return rose nearly three percentage points
994, and while the rate of return
1960s,
(.085)
few
years.
The
sharp increase
third
column
in
earnings, corporate dividends have not risen during the
of Table 2,
"This measure of the rate of return
and Figure
1
,
shows the
price-dividend ratio on
was
analyzed by Nordhaus (1974) and Feldstein and
Summers (1977). For a discussion of alternative measures of the rate of return, and of the
effective tax rate on corporate earnings, see Feldstein, Poterba, and Dicks-Mireaux (1983).
"Adding a time trend to this equation changes the coefficient on RU to -.00521
coefficient,
-.00051 (.00034), provides weak support for a secular decline
in
The time
the corporate
.
profit rate.
13
We
have also estimated the cycle-adjustment equation allowing for a
1994 and 1995. The results are:
Return = .123 - .005*RUC + .007 # DUM94&95
p = .842.
shift in the level
of profits in
(.010) (.001)
(.009)
also include a time trend, the coefficient on the trend
coefficient on DUM94&95 rises to .0136 (.0093).
If
we
(.086)
is
-.00067 (.00031), and the
Figure 2
o
NFC Return
:ycle-A:
13
11
"
c
OJ
u
c
01
a.
7
-
1950
1955
1960
1965
1970
1975
Year
1980
1985
1990
1995
Return to Capital: Nonfinancial Corporations, 1948-1995
9
the
S&P 500, which
reached 39.2 at the end of June,
1
995. This value
is
end of any year during the post-war period. Values of the price-dividend
thirty
have been recorded only a few times during the
the late 1960s, and then
One
higher than at the
excess
ratio in
last forty years: in the early
1
is
a growing reliance on non-
dividend alternatives, such as share repurchases and cash purchases of stock
for returning
cash to shareholders.
became an important source
limited,
declined
the late
in
mid-1980s,
in
1980s and
the last
two
in
the recent stock price increase.
in
potential explanation of rising price-dividend ratios
companies,
960s,
of
early
years.
of
in
other
Repurchases, which were historically very
cash payout during the mid-1980s. Their importance
1990s, and then increased, but not to the
level of the
14
Table 4 tracks the role of non-dividend cash payouts over the last fifteen years, and
suggests that these payouts cannot explain the recent
column
in
Table 4 shows the ratio of
cash payouts to cash dividends
all
1980s very close
corporations. This ratio, which began the
in
the mid-1 980s.
years.
declined
The second and
cash payout
more
It
ratio, for
third
the
in
the early
rise in price-dividend ratios.
1
columns show the price-cash dividend
S&P Composite
Index.
The
rapid increase in share prices relative to total
ratio,
ratios in the third
1
first
for nonfinancial
to unity, rose to
990s, and has ranged between
The
more than two
and
1
.5 in recent
and the
price-total
column show an even
cash payouts than
in prices relative
to
dividends, because non-dividend cash payouts have declined relative to cash dividends during
the mid
1
990s.
The fourth column
of Table 2, as well as Figure
3,
shows Tobin's q
ratio for
nonfinancial corporations, the ratio of the market value of their equity to the replacement cost
14
Bagwell and Shoven (1 989) describe the growth of share repurchases
the tax incentives for repurchases rather than cash dividends.
in
the
1
980s, and
Figure 3
o
Stock Market/CD*
a
ionin'5
i
i
a
o
y
c
ao
U
o
en
-0.25
1950 1955 1960
1965 1970 1975 1980 1985 1990 1995
Year
Stock Market/GDP and Tobin's
Q,
1952-1995
10
of their tangible assets net of outstanding debt.
several years
1995. 16
in
the late
1
970s, was
The recent change
in
between
q
is
which
ratio,
fell
to
below .40
for
.47 at the end of June
unusual by historical standards.
The only comparable
.27 at the end of
1972 and
is
when
more than .50 during
q declined by
This
1
1
percentage change
late
15
a
late
1
994, and
1974, the period of the
two year
first oil
embargo,
The recent values are the highest
period.
recorded values of q during the postwar period.
There are good reasons
"average q" measures such as those
for suspecting that
Table 2 are poor measures of relative value for
some types
the market value and replacement cost of corporate assets
illustrate this.
17
At the end
of
of firms.
in
An
the 1980s
in
extensive study of
makes
it
possible to
1987, when the aggregate value of q was .574, many high
technology and high-growth stocks exhibited q's
many times
greater than the aggregate.
Several illustrations are provided below:
Company
Coca-Cola
Compaq
IBM
Average o Value. December 1987
2.89
2.65
1.53
15
This ratio is computed by the Federal Reserve Board and published in the Balance Sheets
of the U.S. Economy
Tangible assets include plant, equipment, and residential structures,
the replacement values of which are estimated by the Commerce Department using a
perpetual inventory method with adjustment for changing investment good prices, inventories,
and land. The market value of land is estimated by the Federal Reserve Board. The book
value of debt is subtracted from this estimate of asset replacement cost. This measure of q
suffers from several limitations, notably the failure to compute the market value of debt and
the absence of any correction for the presented discounted value of future tax shields as in
.
Summers (1981).
movements in q.
These factors are
unlikely to result in large
changes
in
the short-run
16
Revisions to the Flow of Funds that are expected in December 1 995 will reduce the
estimated market value of equity in nonfinancial corporations, thereby reducing the estimate
of q for recent years. The rise in the value of q in 1 994-95, relative to other recent years, will
not be affected by the revision.
17
The q values shown below are drawn from the
in Hall
(1993).
R&D
Master
File
described and analyzed
11
2.79
1.53
1.59
Intel
Kodak
Motoro'a
If
a rising fraction of firms' high-value assets are intangibles,
such as patents, specialized
workforces with particular human capital attributes, or brand loyalty, rather than the property,
plant,
and equipment, q
may become
increasingly difficult to interpret.
In addition,
an
increasing fraction of corporate earnings are also generated overseas, and the replacement
cost of foreign assets
domestic assets.
is
probably measured with more error than the replacement cost of
18
The market value
of equity relative to
GDP
has increased more slowly than Tobin's q
during the recent stock market rise. This implies a decline
assets to GDP.
value
in
ratio five
the early
fluctuated
the ratio of tangible corporate
At the end of 1994, the replacement cost of tangible assets
nonfinancial corporate sector
below the
in
was 0.77 times GDP.
This value
was more than twenty
for the
percent
years earlier (1.016 at the end of 1989), and forty percent below the
1980s (1.23
between .86 and
1982). During the postwar period prior to 1973, this ratio
in
.96.
rose between
It
1
974 and 1 982, and has been
declining for
the period since then.
It
is difficult
to
distill
a simple conclusion from Table 2.
While P/E ratios are not
unusually high at present, other measures of stock price valuation are at or near historical
highs.
Conclusions about whether current stock prices can be justified by fundamentals are
beyond the current
18
project.
19
Table 2 does suggest, however, that
Hines (1991) documents the
rising
it
may
be important to
share of international earnings for U.S. nonfinancial
corporations.
19
One
concluded in French and Poterba (1991) that fundamental
Nikkei stock index was approximately 39,000 in 1 989
(high real estate values for corporate land made Tobin's q for Japanese shares was nearly
one), is hesitant to venture again into analyzing stock market fundamentals!
of the authors, having
factors could explain
why Tokyo's
12
distinguish
or
P/D
between stock
ratio,
price fluctuations that are associated with
and fluctuations that are not,
in
movements
in
the P/E
assessing the macroeconomic consequences of
stock price movements.
A number
of recent studies
suggest that variations
in
the earnings-price ratio are
correlated with prospective stock market returns, and one concludes that "shocks to [stock]
prices holding dividends constant are almost entirely transitory (Cochrane,
Sharp increases
in
the P/E or P/D ratio, other things equal, are associated with lower
prospective returns.
increases
in
If
households view increases
Aggregate Trends
All
in
share prices that are not supported by
dividends or earnings differently from those that are, this could result
effects on consumption from these
2.
in
two types
is
ultimately
owned by
individuals.
With the exception
of shares
all
U.S. equities represent net
Yet the form of individual equity ownership
share price fluctuations on household behavior.
in
divergent
the Ownership of Corporate Stock
corporate stock
of U.S. citizens.
in
movements.
of share price
held by foreigners, currently about five percent of the total,
worth
1994, p.241)." 20
response to fluctuations
in
If
may
affect the impact of
individuals adjust their
the price of shares that they
shares that they hold through financial intermediaries or
own
in
consumption more
directly than in
response to
accounts that are dedicated to
retirement saving, then the causal channel by which stock prices affect real economic activity
may depend on
stock ownership patterns. 21
Differential transactions costs associated with
20
Campbell and Shiller (1988, 1989) present closely related evidence on E/P, D/P, and
stock returns.
21
Behavioral models, developed for example in Thaler (1994), suggest that the form in
which shares are held, and even the record-keeping convention applied to them, may affect
the magnitude of the wealth effect on consumer spending. Because mutual funds and other
financial intermediaries mail their investors quarterly statements, individuals may be more
13
different types of equity accounts,
401
(k)s,
such as penalty taxes
can also induce divergences
in
for early
withdrawals from IRAs
or
the consumption response to capital gains on stock
held different ways.
Accumulation
in
accounts that are
tax-deferred saving plans,
may
"off-limits,"
not lead to the
as
many
people view their IRAs or other
same spending response as
increases
in
the
value of directly held assets that can be tapped for current consumption. Investment through
these accounts has
of
become
particularly important in recent years.
the
first
eight
months
1995, more than two thirds of the cash inflows to leading mutual fund managers were
directed to mutual funds held
nearly one third of
For
some
all
in
retirement plans.
and
local
complicated. Consider the example of equity
government retirement
plan, a defined-benefit plan for the retirement
employees. Individuals as taxpayers are the ultimate beneficiaries of gains
not be aware of the increase
in
the value of their locality's pension portfolio, and they
may
funding future pension
The perceived change
aware
owned
of their gains
imply that future
may
still
not be confident enough that their future taxes
the value of stock
will
Yet individuals
tax burdens can be lowered while
23
represently
is
the value of this pension fund's holdings, since higher asset values
to such gains.
now
categories of indirect stock ownership, the link between a current capital gain
held by a state and local
of state
Retirement plan assets
mutual fund assets. 22
and benefits to the indirect individual holders
in
In
directly, or
in
will
decline to raise consumption
net worth
stock
liabilities.
may
in
response
be quite different for changes
owned through mutual funds which
in
continually
on these investments than on direct stock investments that require action
to evaluate.
"Schultz (1995).
"There is an inconclusive literature on the extent to which unfunded state and local
government pension liabilities are capitalized into house values; see Epple and Shipper (1 981 ).
14
provide information on net asset value.
Before considering whether changes
in
stock ownership patterns have affected the
between share prices and consumer spending, we summarize the postwar history
stock ownership. The standard claim that individual investors
of equity
ownership
in
the United States
Flow of Funds Accounts.
declining from nearly
90%
24
is
now account
the
1950s
of individual
for less than half
based on data from the Federal Reserve Board's
These data show "household" ownership
in
link
to less than
50%
of corporate stock
today.
Although widely used, the Flow of Funds data do not measure what many analysts
think they do.
They do not apply
to listed equity on stock exchanges, but rather to a broader
concept of corporate equity, including stock
economic actors
institutions.
a group of
25
in
closely-held companies, by a group of
(the "household sector") that includes individuals as well as nonprofit
Moreover, they do not describe holdings of individual investors, but rather of
economic actors, the "household sector," which includes nonprofit
The entry
for
household sector holdings
in
institutions.
the Flow of Funds table for corporate equity
balances also excludes equity held through mutual funds, through defined contribution pension
plan accounts, and through other financial products such as variable annuities.
of individual equity
ownership are allocated to other sectors
have become more important
in
in
These forms
the Flow of Funds, and as they
the last decade, the potential for misinterpretation of the
"household sector" data has grown. The growth of institutional holdings does not necessarily
imply that shocks to stock market values
than
in
now have
smaller effects on individual net worth
previous periods.
"See
for
example Blume and Zeldes (1994) and Friedman (1995).
25
The total market value of corporate stock in the Flow of Funds exceeds that on the
NYSE, AMEX, and NASDAQ. The value of closely-held shares at the end of 1994 was
approximately $1.2
trillion.
.
15
Table 5 summarizes the Flow of Funds data on the share of outstanding equity held
by various classes of investors.
column
is
shown
in
the
first
column. 26 The
mutual funds combines ownership by open-end and closed-end investment
for
companies, and that
local
The household sector
pension funds includes private pension funds as well as state and
for
government retirement systems.
contribution (DC) plans,
in
The
private plans
which the plan participants have
column includes both defined
distinct
accounts that change
in
value along with the price of underlying assets, as well as defined benefit (DB) plans, which
promise particular benefit streams to retirees as a function of their age, years of service, and
wage
history at retirement.
To estimate the share
indirectly,
(i)
we make
We
five
of corporate stock that individuals hold either directlv or
adjustments to the Flow of Funds household sector data:
subtract the equity holdings of nonprofit institutions from the Flow of Funds
household sector.
Experimental data presented
in
equity holdings of nonprofit institutions for the period
1
5.7%
of
of the
we
in
the
987-1 992. Nonprofit holdings average
this period,
we
so
each year between
1
multiply the Flow
952 and
1
994 when
need to remove these holdings.
(ii)
of
1
household sector's equity holdings during
Funds household sector equity value by .843
shows
the Flow of Funds accounts
all
add stock held by bank personal
trusts, since individuals are the beneficiaries
of these accounts.
(iii)
holdings.
stock,
26
We
We add
equity held
in
defined contribution pension plans to the household sector
At the end of 1993, when private pension plans held $1075
$481
billion
was
held
in
these plans.
billion in
corporate
Since individuals are the owners of these
The sharp decline in the share of equity held by households between 1 968 (81 .9%) and
969 (69.1 %) is due to the creation of a separate Flow of Funds category for bank personal
trusts, which accounted for 10.5% of equity holdings in 1969.
1
16
accounts,
held
the
in
we
attribute this equity to them.
The share
defined contribution plans has increased from just over a quarter at the beginning of
1980s
to nearly half in the mid-1990s.
We
(iv)
add equity held
in
variable annuity reserves at
Variable annuities, which have been one of the
the last decade, provide a
Total assets held
991 to $1 76.4
invested
in
(v)
in
rapidly
for individuals to defer
variable annuity accounts
the end of
billion at
stock.
means
most
life
insurance companies.
growing insurance products
taxes on capital income, at the price
1
have grown from $47.7
billion at
the end of
994, and nearly three quarters of variable annuities are
27
Finally,
we
add household ownership of open- and closed-end mutual funds that
invest in corporate stock.
Individuals are the ultimate
owners
of
most mutual fund shares.
At the end of 1994, 66.2% of mutual fund shares were allocated to households
of Funds, with another
households.
of
and some limitations on investment options and withdrawal provisions.
of insurance loads
1
of private pension fund equity assets
1
3.2%
allocated to bank trust departments,
Given the growth
in
in
the Flow
which we aggregate with
mutual fund holdings of corporate equity over time,
it
is
increasingly important to recognize this channel for individual equity ownership. 28
The magnitude
of
each of these corrections to household equity ownership, at the end
"Gentry (1994) and Poterba (1995a) discuss the growth
of variable annuities.
"Retirement accounts comprise roughly two thirds of household mutual fund holdings.
At the end of 1994, household ownership of mutual funds totalled $1066 billion. Data from
the Investment Company Institute show that $361 billion of this total was held in Individual
Retirement Accounts, $161 billion was in 401 (k) plans, $76 billion was in non-401 (k) defined
contribution pension plans, and $98 billion was in 403(b) plans.
17
29
1994, are shown below.
of
entries
All
are
percentages of
corporate stock
total
outstanding:
Flow
of
47.7%
Funds Household Sector
Nonprofit Holdings
-
(7.5)
+ Bank Personal Trusts
+ Pension Plan Assets
+ Variable Annuity Accounts
+ Mutual Fund Holdings
2.7
7.7
2.0
1 1
The net
that
all
is
effect of these adjustments
attributed to individuals.
.0
63.7%
Adjusted Individual Holdings
is
to substantially raise the fraction of corporate equity
Rather than suggesting that individuals hold less than half of
corporate stock, the modified calculations suggest that individual investors hold
two two
thirds of outstanding stock, either directly or through a fiduciary.
Table 6 shows these adjustment terms for the period 1952-1994; Figure 4 plots the
adjusted and unadjusted time series for individual stock ownership.
the trend
in
stock ownership patterns. The secular decline
individuals that
emerges
individual direct
and
declines from
76%
but then rebounds.
years. This
is
in
the
first
column
indirect ownership.
of the
It
market
in
1970
of Table 6
Individual
is
in
The adjustments change
the share of equity
owned by
not supported by the data on
ownership on our expanded definition
to just over
60%
of the
market
in
the late 1980s,
has grown by more than three percentage points during the
last four
largely the result of the diffusion of tax-deferred saving plans, particularly tax-
deferred saving programs such as 401 (k)s and 403(b)s, through the employed population.
29
0ne
of the revisions that is due to be incorporated in the December 15, 1995 revision
Flow of Funds accounts will involve some reallocation of variable annuity assets
between the mutual fund and insurance company sectors. The current Flow of Funds
procedure includes variable annuity equity assets in both the mutual fund and insurance
sectors, and consequently subtracts these assets twice from total equity outstanding in
computing household equity holdings. The correction presented in the text, adding back
variable annuity assets, corrects for this. The December 1 995 Flow of Funds revision is also
expected to decrease the amount of stock held by corporate pension plans.
of the
Figure 4
o
FOF House r olc:s
.rc;v:cj3.3
90 -
BO -
01
en
to
—
c
70
QJ
(J
C
(U
a
60 -
50
1950
1955
1960
1965
1970
1975
Yean
1980
1985
1990
Share of Stock OwneO by Households, 1952-1994
1995
18
The adjustments
potential
to the Flow of
consumption effects
framework.
A key concern
an increase
of
in this
Funds data are necessary
context
is
in
share values
ownership
traders'* in security
may
understanding the
the "mental accounts"
in
the degree to which individuals recognize capital
gains on equities as a potential basis for higher consumption.
individual stock
for
The adjusted measures
also be important for gauging the significance of "noise
markets. 30
For other issues concerning stock ownership, however, these corrections
relevant,
and the standard view that individuals "own" less than
is
may
not be
half of corporate stock
may
Since corporate stock held through mutual funds or defined contribution
be appropriate.
pension plans
of
who
voted by the fiduciaries, not by the individuals
of these shares, the rise of indirect
ownership
may have
are the beneficial holders
altered the balance of
power within
corporations.
The factors that explain the evident trend away from
direct
ownership of corporate
stock and toward ownership through financial intermediaries are not well understood.
Tax
considerations actually incline individuals toward direct ownership of shares. Individuals can
more
gains
efficiently invoke tax planning strategies that realize capital losses,
if
they
own
shares directly rather than through a mutual fund.
explain part of the growth of corporate pensions.
and defer
capital
Tax incentives may
By investing through 401
(k)
plans and
defined contribution pension plans, individuals can defer taxes on both capital gains and
dividend income.
Many households
only hold equity
in
these tax-deferred forms.
vailing incentive is the opportunity for greater diversification that is afforded
relative to the
30
Shleifer
traders,
purchase of securities
and Summers
and explore
(1
in individual
companies.
990) describe models of
their allocative effects.
financial
Further
A
counter-
by mutual funds
work
is
needed to
market equilibrium with noise
19
understand the other factors
in
the financial services marketplace that have led investors
away
from direct stock holding.
3.
Evolvino Patterns of Individual Stock Ownership
One
of the salient features of stock
the population.
In light of
ownership
is its
in
31
the analysis of portfolio behavior.
who
understanding the consumption effects of
changing
"who
share prices, and
role of direct vs. indirect
we
stock ownership
may
it
is
be important for
also a key input to
may
Because the
be associated with shifts
in
the
interrupt our analysis of aggregate trends to present
information on the changing cross-sectional pattern of equity ownership.
We use data from the 962 Survey of Financial
1
1983 and 1992 Surveys
of
Consumer Finances,
during the last three decades.
random sample
of U.S.
to gather detailed
Beginning
This
gains or loses" from share price fluctuations.
distribution of stock holdings,
summary
rising
hold no equity represents a
Holdings of corporate stock are more
concentrated than most other components of net worth.
standard analyses of
a sub-set of
the higher historical average return on stocks than on other
investment assets, the substantial number of households
puzzle
among
concentration
in
32
The Survey
of
to
Characteristics of Households, and the
summarize
individual equity
Consumer Finances (SCF)
households administered by the Federal Reserve Board.
information on assets,
liabilites,
is
It
ownership
a stratified
is
designed
and demographic characteristics.
1983, the SCF has been conducted every three years.
In
recognition of the
31
Haliassos and Bertaut (1995) provide a recent survey of the related literature. They,
along with King and Leape (1984), present careful econometric treatments of the incidence
of stock ownership.
32
A systematic survey of trends in share ownership in the early postwar period may be
found in Blume, Crockett, and Friend (1974). Projector and Wise (1966) describe the 1962
Survey of Financial Characteristics of Households.
20
highly
skewed
distributions
of
many types
of
and
financial
assets, each
real
survey
oversamples high income households. Each SCF contains an area-probability sample, which
is
a stratified
stratified
random sample
random sample
of
households chosen from the population at large, and a
of
households drawn from a set of high-income tax returns. Both
samples are surveyed using the same questionnaire, but missing value imputations
public release versions are typically
households
in
the
SCF were from
done separately.
In
1983, 438
of the
the high-income sample, compared with
1450
in
the
4103 SCF
of
3906
in
1992. 33
Our analysis focuses primarily on the
because
in
1
983 sample and was
previous survey.
The design
983 and
1
992 Surveys
these two years the specific goal of the
cross-sectional sample of wealth holdings.
the
1
35
One
of the
not as complete
of the
1
989
34
The 1986 SCF was
in its
priorities
SCF was
was
of
Consumer Finances,
to provide a detailed
entirely a re-interview of
gathering of stock and pension data as the
establishing a panel with the
1
983
survey.
1983 and 1992 surveys was not encumbered by considerations
of
preserving a panel data set. 36
Table 7 reports the number and percent of households owning stock
and 1992.
33
The successive rows
of the table contain progressively
in
1962, 1983,
more comprehensive
Kennickell and Starr-McCluer (1994).
"The 1 983 survey instrument and sample are described
1992 survey is described in Kennickell (1995).
in
a very and
Elliehausen
(1
988),
while the
35
36
Heeringa, Connor, and
Woodburn (1994).
and Morgan (1 989) compare different wealth surveys, and conclude that
the only one with enough high-income households to permit tabulations of detailed
asset categories.
Avery, Elliehausen, Canner, and Gustafson (1984) and Avery and
Elliehausen (1 986) tabulate basic results from the 1 983 SCF. The 1 992 data are summarized
in Kennickell and Starr-McCluer (1 994). Antoniewicz (1 994) describes the link between SCF
data and aggregate data from the Flow of Funds accounts.
the
Curtin, Juster,
SCF
is
21
measures
The
of stock ownership.
The next row
traded companies.
The next row adds shares held
row focuses on
first
also includes stock held indirectly through mutual funds.
IRAs or Keogh accounts, then shares held through tax-
in
deferred saving plans such as 401 (k)s, and
plans.
finally,
equity held
in
defined contribution pension
37
The upper panel
increased between
to calculate
directly or
196^
direct holding of shares, in publicly
two
1
of Table 7
962 and
1
definitions of
983.
shows
Institutional
ownership
through mutual funds.
that the
in
number
of
households owning stock
changes and data
1962:
limitations only permit us
directly held stock,
Neither IRAs nor 401
(k)
and stock held either
plans existed
in
1962, and the
£>CF did not collect information on detailed pension plan attributes, so there
is
no
information on equity held through defined contribution pension plans.
The percentage
of
households with direct stock ownership declined between
1992, while the percentage with
indirect holdings increased.
corporate stock accounted for less than half of
all
By 1992,
equity holders.
1
7.8%
stock
of
households reported direct stock holdings. For
ownership,
direct
holdings,
the
data
all
show an
but th
most
increase
in
mutual funds,
for
37.4%
of
we consider,
limited
the
stockownership between 1983 and 1992. The percentage of households
either directly or through
983 and
direct holders of
While
households owned stock either directly or through one of the intermediaries
1
only
measure
incidence
who own
all
of
of
stock
example, rises from 20.1 to 22.0 percent. The
share of households holding equity under our most expansive definition increased by 4.2
37
Some equity holdings may still be omitted in our analysis, because the SCF does not
provide detailed information on equity held in trusts for which the respondent is a beneficiary,
or in variable annuity accounts. We have imputed the share of mutual fund holdings in 1 983
that are accounted for by equity mutual funds. The 1983 SCF questionnaire did not collect
We
information on the type of mutual fund investors held.
therefore divided reported 1983
mutual fund assets between stock and bond mutual funds on the basis of the mix of these
funds in the 1989 SCF.
22
38
percentage points from 33.2 to 37.4 percent.
12.7%
of
Between 1983 and 1992, an
additional
households became stock owners.
The information
in
the
first
part of Table 7 indicates the total
number
of
households
with any exposure to stock price fluctuations. For analyzing consumption decisions, however,
it
may
be more appropriate to focus only on those households with substantial exposure,
defined by an absolute level of equity holding.
$2000
to only those individuals with at least
own some
falls
invested
limit
in
our definition of stockholders
common
stocks
stockowners declines to 12.9%
fraction of households categorized as
extended measure of ownership
When we
to
corporate stock, but less than
29.3%. Thus roughly
$2000 worth. Stock
in
in
1992 39 the
,
1992, and the
eight percent of households
price fluctuations are not likely
to have large absolute effects on the net worth of these households, although given the low
levels of asset holdings for
movements may be
One
provide
holders.
many
substantial.
households, the proportionate effects of stock price
40
of the reasons for exploring the disaggregate pattern of stock
some
information on the marginal propensities to
One important dimension
for
such analysis
is
consume out
stor '(holder nge.
ownership
is
to
of wealth for stock
41
Table 8 presents
38
SCF staff report some concern that some closely held stock was classified as publicly
traded stock in the 1 983 SCF, thereby overstating the number of households owning traded
stock and understating the growth of shareholdings between 1 983 and 1 992.
39
All
monetary amounts
in
our analysis of the Survey of
Consumer Finances
refer to
1
992
dollars.
40
Poterba, Venti, and Wise (1994) present data on the distribution of financial asset
holdings for households approaching retirement age.
41
Bosworth, Burtless, and Sabelhaus (1991) and Attanasio (1994) present information on
age-specific saving rates.
23
information on stockownership by the age of household head.
of
65 hold more than
The highest
40% (50%
probability of
in
1983)
of the
owning stock occurs
in
42
Households over the age
household sector's publicly traded stock.
the years preceding retirement.
Table 8 illustrates the dramatic growth of indirect stock holdings.
17%
example,
of
households
in
the 45-54 group held stock only indirectly.
analogous percentage had grown to 28%. The comparable
20%
from
to
24%. Comparing
that the critical
the entries
growth has taken place
in
statistic for
1983,
In
for
By 1992, the
those aged 35-44 rose
the middle and right panels of Table 8 suggests
mutual fund holdings and IRA/Keogh holdings.
in
Including indirect holdings has a larger effect on the estimated rate of stock ownership for
younger households than
that
own
shares directly
indirectly.
The increase
in
for older ones.
is
only
1992, the share
In
of
35-44 year
old
households
20.2%, compared with 44.1% owning stock
ownership probability
is
smaller
(1
8.4%
to
26.5%)
for
directly or
households
over the age of 65.
To
link
the disaggregate information on stock ownership to our discussion of
consumption and stock
price fluctuations,
we
can use the information
in
Table 8 to compute
the age distribution of capital gains on corporate stock. The market value of corporate stock
was $6048.8
billion at
the end of
1
994, and
1994 and June 1995. Since 63.7%
it
increased by $1
of outstanding equity
1 1
was
9
billion
held
in
between December
forms that
we
define
as providing individual control over these assets, individuals therefore received a capital gain
of
$713
billion.
The
distribution of this gain
Age < 35
Age 35-45
Age 45-54
42
$
by age
37
is
shown below:
billion
$ 121
billion
175
billion
$
head of household is that when the survey respondent is
couple, the head is the spouse with higher wage, salary, and
self-employment income. If neither spouse reports labor income, the head is the older spouse.
part
Our convention
of
a
married
for selecting a
24
Age 54-64
Age 65+
The key conclusion from
this calculation
$
161
billion
$
219
billion.
is
that the majority of wealth changes from stock
43
market fluctuations accrue to older households.
Some have
because they
will
argued that younger households benefit indirectly
receive
in
among households
that do not hold stock.
Ultimately, the
bequests any assets that are not consumed by older households.
develop this argument
more
in
actual mortality tables to
detail,
we
consider the timing of expected bequests.
compute the expected percentage
bequests are not a
critical factor in
The
Over a
Yet to
We
we
estimate that
fifteen year horizon the share is
receive a large fraction of outstanding equity through the bequest channel, but
that the expected bequest of equity during the next fifteen years
of equity currently held
24.4%,
can translate
is
it
is
will
notable
comparable to the amount
by households under the age of 45.
The concentration
We
use
results suggest that
the near term. Over the next five years,
of corporate stock will be bequeathed.
young
and over 25 years, 45.3%. These calculations do not suggest that younger generations
43
rise
of corporate stock that will be
bequeathed to younger generations over various horizons. 44
5.7%
share prices
receive substantial bequests from the current elderly. This could stimulate
higher consumption even
will
when
of stock ownership, as well as its
age
distribution,
can affect the
consumption metric with the crude assumption that
households exhibit marginal propensities to consume out of wealth equal to 1 /(T-age), where
T is expected end of life. We set T = 80, assume that all over-65 households are age 72 and
that all under 35 are 30, and that the households in the other age brackets are all at the
bracket midpoint (i.e., 50 for the 45-54 bracket). This implies a "predicted" consumption
response of $45 billion (6.3%) to the $713 billion share price increase.
this into a
"For married couple households that own stock, we define a bequest as occurring when
members of the couple have died. The average mortality table may understate life
expectancy for stockholders, since age-specific mortality rates are negatively correlated with
both
wealth holdings.
25
between stock
linkage
price fluctuations
and consumption spending. The skewed nature
ownership underpins the view that consumption adjustments by the small set
of substantial
stock owners cannot have detectable effects on aggregate consumer spending.
the concentration of share ownership
1
983, and
new
1
992 SCFs. The
In
least
in
of
all
of concentration
is
owned 55.1%
To be included
equities in
10.3%
stockholders, which held
962,
well-known cross-sectional patterns, but suggest
both direct and indirect holdings,
$800,000
1
1983, the 0.5% of stockowners with the largest equity
the analogous group held only 36.8%.
needed at
We describe
Table 9, which presents information from the
results confirm
conclusions about trends.
portfolios, including
in
of
in
1%,
top 1/2 of
The next one
1992.
equity
in this
of the stock.
half of
In
1992,
a household
one percent of
1992, had equity portfolios worth between
$500,000 and $800,000.
The degree
publicly traded stock.
with the largest holdings
between the
66%
Over
"All Equity"
in
1
even greater
we
if
was
of directly held stock
as IRAs
in
in
and "Non Pension Equity" entries
0.5%
in
1
of stockholders
992. Comparison
Table 9 suggests that growing
been less important than growing
reducing the concentration of equity ownership. 45
holding, and net worth, in
two
variants of
in
equity holdings, non-equity financial asset
1962, 1983, and 1992. Because some of the variables that
use to construct net worth
45
held by the
mutual funds and the expansion of tax-deferred retirement saving vehicles such
Table 9 permits comparisons of inequality
report
our attention to directly-held
983. This percentage declined to 58.6% by
participation in defined contribution pension plans has
investment
limit
in
1992
are not available
1983 wealth
inequality
in
for
the
1962
data, and vice versa,
we
we
comparison with 1992 and 1962
Some assets that are accumulated in defined contribution pension plans may show up
as assets in IRAs if these pension assets have been "rolled over" in a lump sum distribution.
For information on the importance of such rollovers, see Poterba, Venti, and Wise (1995).
26
respectively.
1983
We then focus
on pairwise comparisons across years. The calculations
1962
data, on the basis of
definitions, are
shown
in
the last
The comparison between 1962 and 1983 suggests
two columns
relatively
little
of Table 9.
change
1
983. Total non-pension equity also became
The substantial decline
(62.3%
to
43.2%
in
1962 was 63.3%, compared with 66.2%
in
0.5%
with stock portfolios
in
directly held stock
more unequal between
slightly
0.5%) thus represents
The data show
trend of the previous period.
that
1
962 and
by households
in
983.
a significant departure from the
between 1983 and 1992, the share
the 90th to 99th percentiles increased substantially.
for
1
1983 and 1992
of the stockholding population declined, while that of
accounted
the
of
the inequality of non-pension equity between
held by the top
equity held by the top
in
in
0.5%
concentration of equity ownership. The share of publicly traded stock held by the
households with the largest stock portfolios
for the
of
households
The share
of
the 80th through 90th percentiles declined
slightly.
The
central
message
of Table 9
is
that
more than one
third of the gains or losses
on
corporate stock accrue to the roughly half a million households with the largest equity
holdings, with another forty percent of the gains accruing to the 4.5 million households with
the next largest equity stakes.
If
the linkage between stock returns and consumption turns
on directly held equity, then the concentration of holdings
is
even more dramatic: nearly sixty
percent of the capital gains on directly-held corporate stock accrue to the half a million
households with the largest portfolios of corporate stock.
Table 9 also presents distributions of non-equity financial assets, real assets such as
owner-occupied
48
real estate,
The tabulations
because the
and
total net
worth. 46
These tabulations use the data and
1992 should be viewed as preliminary and subject to revision
release of the Survey of Consumer Finances has not been made. The
for
final public
data underlying these tabulations contain no missing values but use a preliminary sample
27
sample weights that underlie the
first
sample weights are subject to revision
published tabulations from the
the
in
final
1992 data
between 1983 and 1992.* 8 The seventh column
distribution in
1
983 and
1
992 as computed from the Survey
a rising share (9.7% versus
1
%
of
income
of the
7.6%)
distribution,
concentration of housing equity,
is
in
of
became more
is
80%
Consumer Finances.
income accruing to households
of the distribution.*
important
not simply an artifact of the Survey of
computational algorithms. Net worth
financial assets are distributed
is
9
in
It
shows
the top 1/2 of
in
the share
falling inequality of equity
Consumer Finances data
more unequally
equally
This finding, and the rising
showing that the
in
The
Table 9 shows the income
and a substantial decline, from 49.6% to 43.4%,
income reported by the bottom
holding
of family
47
public-use version of the data set. With the
exception of owner-occupied real estate, each of these asset categories
distributed
set.
set, nor of our
distributed than income, non-equity
more unequally than net worth, and equity holdings are
distributed less equally than non-equity financial assets.
The
and top
of
1
finding that the fraction of corporate stock
% of the distribution declined between
1
and of net worth held by the top 0.5%
983 and
1
992
contrasts with recent studies
wealth inequality suggesting growing inequality of financial asset holdings increased during
weight to construct the asset distributions.
47
Kennickell and Starr-McLuer (1994).
48
Our net worth calculations, and most others directed at measuring the inequality of
wealth, exclude the actuarial present discounted value of defined benefit pension benefits.
Including these benefits would probably reduce the share of net worth held by the highestwealth households.
49
Levy and Murnane (1992) describe and discuss recent changes
distribution.
in
the U.S. income
28
the
1983-1989
period.
50
The
result
is
striking given the rise in share prices during this
and the unequal distribution of share ownership. An increase
period,
distributed less equally than net
an asset that
is
worth;
principle
it
is in
more equal, while the
changes.
even possible
the relative value of
worth should increase the inequality
for the distribution of
distribution of net
in
of net
each component asset to become
worth becomes less equal, with such asset price
51
To assess the importance
estimated the change
resulted only from
in
the inequality of net worth between
changes
in
52
asset prices.
We
1
983 and 1 992
adjusted the reported
that would have
1983 values
corporate stock holdings and closely-held businesses by the real appreciation of the
Composite share
the
price index,
Commerce Department's
we
of rising share prices in contributing to inequality,
and the value of owner-occupied housing by the
homes. The
price index for constant-quality
real
of
S&P
change
in
resulting shares of
50
Wolff (1994, 1995) reports the changes in wealth inequality between 1983 and 1989.
These studies adjust the SCF data to align the total reported assets with aggregate totals in
the Flow of Funds accounts. Because the SCF totals are typically below the Flow of Funds
values, these corrections inflate the amount of each asset held by each household that reports
the asset. These corrections do not change the set of households who have a given asset,
or the inequality within asset categories, although they can affect the measured inequality of
broader composite measures of financial assets or net worth.
Consider an economy with two assets, A and B. Households in the top 1 % of the
wealth distribution own sA and s B percent of these assets respectively, the market value of
asset A is VA and that of B is VB Let w A = VA /(VA + VB ) and w B = VB /(VA + VB ). The
percent of net worth held by the richest 1 % of households is sA *w A + s B *w B Assume that
A is distributed less equally than B (sA > sB ). On a different date, the top 1 of households
hold sA and sB percent of A and B, respectively. Assume sA < s A and s B < Sj,. It does not
follow that net worth is more equally distributed. If w A > w A wealth could still be more
equal at the second date then the first. If the simple case of s B = sB net worth inequality
rises if s A 7s A > SB/ s A + n-s B /s A ) # (w A /w A '). If s B /s A = .25, and asset A appreciates 50%, so
w A /w A = .67, then net worth inequality will rise for any sA 7sA above .75.
51
.
.
%
'
'
'
'
'
,
'
,
'
"Weicher (1995) explores the effect of rising share prices on net worth inequality during
the 1983-1989 period, and concludes that favorable stock returns did not substantially
exacerbate inequality then, because they were parallelled by rising real estate values.
29
net worth are
The
in
shown below:
1992
Actual 1992
Wealth Percentile
Actual 1983
Predicted
Top 0.5%
23.74%
25.63%
21.68%
Next 0.5%
Next 4.0%
Next 5.0%
Next 10.0%
Remaining 80.0%
7.12
22.86
12.12
13.08
21.08
7.70
23.11
11.71
12.10
19.75
6.99
24.37
13.77
14.10
19.20
results
show
1992 was
that the actual share of wealth holdings by the highest-wealth households
substantially less than the extrapolation of the
have predicted.
Ceteris paribus
the relative rise
,
in
1983 wealth
distribution
would
corporate stock prices would have
Other changes, however, principally the diffusion of
contributed to increasing inequality.
substantial stock holdings to households near the top, but not at the top, of the wealth
distribution,
were important than asset
of equity holdings
The source
changes
price
in
generating changes
of the difference
1
between our
%
findings,
of asset holders,
which show a decline
and other findings of
appears to be due to our reliance on 1992 rather than 1989 data.
methods to the 1989 SCF data, we
find rising
Our estimates suggest that the share of
the highest net worth increased from
is
the distribution
and net worth.
holdings and net worth of the top
the decline from
in
26.1%
whether the change
Consumer Finances
is
21.7%
the equity
rising inequality
When we
apply our
wealth inequality between 1983 and 1989.
total net
23.7%
in
worth held by the 0.5% of households with
to 26.1
1992
%
between
1
983 and
1
989. This makes
A
in
1989
in
wealth inequality recorded by the 1989 and 1992 Surveys of
to
in
a reliable indicator of actual
all
the more striking.
changes
in
critical
question
the U.S. wealth distribution.
53
"Even though the SCF is the best available data source on the distribution of wealth, the
number of high-wealth households on whom the SCF results are based still makes these
tabulations potentially sensitive to outliers. Analysis of the 1983-1989 SCF panel might
provide further information on the changing patterns of asset holdings between these years.
small
30
The asset
1
992
for
raises
among
distribution
some questions about
the highest net worth households
the
1
989 SCF
example, the share of net worth held
1989, and 14.4%
in
in
Among
data.
corporate stocks
1992. The share of net worth
in
the top
in
1983, 1989, and
0.5%
was 21.2%
in
of
households,
1983, 8.0%
closely held businesses
was 34.6%,
42.1 %, and 38.4%, respectively. At a minimum, the negatively-correlated fluctuations
two assets suggests
share of these
equity
in
some
years.
It
is
that there
may
be
some
in
in
the
misclassification of closely held
not clear whether this could have any effect on the measured
inequality of net worth, but
could affect the inequality of
it
component assets such as
corporate stock.
The 1989 and 1992 Surveys
of net
worth held by the
of net
worth
from 21
fell
80%
.
1
%
of
in
1
of
Consumer Finance both show
a decline in the share
households with the lowest net worth. This group's share
983
to
1
9.2%
in 1
992. Roughly half of the decline
group accrued to households between the 80th and 90th percentiles
distribution,
and the remainder accrued to those
from Table 9
is
among those
high
in
changes
analysis of stockownership,
who own
stock.
54
information on these dimensions of stock ownership,
992 SCFs. The
the
wealth
A key conclusion to emerge
in
the distribution of wealth
the net worth distribution.
non-equity wealth holdings of households
1
the top decile.
that there have been non-trivial recent
To complete the disaggregate
summary
in
in
for this
tables
show
a strong positive relationship
we consider the income and
Tables 10 and 11 provide
drawn from the 1983 and
between income,
financial assets
other than equity, and the probability of stock ownership.
Table 10
^Poterba
stockholders.
shows
(1995b)
that
60.5%
presents
of
more
households with a family income of more than
detailed
information
on the
characteristics
of
31
$250,000
in
was 79.3%
when
1
992 owned stock
for this group.
directly.
The
ownership
probability of direct or indirect stock
much
Both of these probabilities were
higher a decade earlier.
the probability of direct or indirect equity ownership for a household with an income of
$250,000 (1992
dollars)
was 92%. Approximately 23%
of corporate stock
was owned by
23% was
households with 1992 family incomes of more than $250,000; another
held by
households with incomes between $100,000 and $250,000. Thus, consumption decisions
by these households play an important part
in linking
stock price fluctuations to overall
consumer spending. Table 10 also shows, however, the
In
rising equality of
1983, those with family incomes of $100,000 (1992
directly-held equity; that share
Table
households
1 1
had declined to
50%
dollars)
share ownership.
and above held
75%
by 1992.
who own
corporate stock.
we
Publicly traded stock
consider, with
44%
is
the most unequal of the
of directly-held stock
owned by households
with non-equity financial asset holdings above $250,000. The comparison between
reveals a substantial increase
ownership
in
of stock at
expansive definition of equity holdings, the one that includes equity
plans as well as shares held through financial intermediaries,
held by households with less than
$50,000
in
in
1983 and
lower wealth levels,
consistent with the previous data on the family incomes of stockholders.
is
all
presents similar information on the non-equity financial assets of the
various equity measures
1992 again
of
On
the most
defined contribution
27 percent
of corporate stock
other financial assets, and
34 percent by
those with non-equity financial assets between $50,000 and $250,000.
The conclusion that emerges from
ownership
is
concentrated.
households
is
that
this analysis of cross-sectional data
on stock
ownership has become more equal over time, but remains highly
The proposition that equity
capital gains
not supported by the data, since
37.4%
of
accrue to only a small set of
households owned some corporate
32
stock
in
1
992. The concentration of stock holdings nevertheless implies that a small subset
of the population,
on the order of
five
percent of
households, receives roughly three
all
quarters of the capital gains and losses associated with stock price
4.
movements.
Stock Market Fluctuations and Consumption
We now consider the relationship between stock market returns and consumption. We
organize our analysis as a test of whether the stock market has a causal wealth effect on
consumption, or
spending. 55
steps.
We
is
simply a leading indicator that forecasts future changes
in
consider the effect of rising stock prices on consumption outlays
We
illustrate
several
the difficulty of interpreting these time
series relationships by reference to discussions of the
market crashes of 1987 and 1929.
We
consumption effects of the stock
then study four issues that are motivated by our
previous discussion of share ownership patterns.
First,
fluctuations affect the mix of consumption spending
disproportionately by high-income households, and
all
we examine whether
closely correlated with
households.
Third,
we
changes
in
share prices than
is
stock price
between "luxury goods" consumed
other goods. Next,
survey data to investigate whether the consumption of households that
we
own
use household
stock
is
more
the consumption of non-stockholding
investigate whether the changing pattern of direct versus indirect
stock ownership affects the relationship between stock market fluctuations and
consumption spending. This amounts to testing whether stock price changes
55
in
We begin by summarizing the time series relationship between stock price changes and
subsequent consumption fluctuations.
in
consumer
movements
in
the early
This leading indicator view closely resembles the "passive informant" hypothesis that
Morck, Shleifer, and Vishny (1990) develop with respect to stock price movements and
investment spending. Detailed evidence on the predictive power of stock returns as leading
indicators may be found in Stock and Watson (1990).
33
post-war years had greater predictive power for consumption growth than analogous
fluctuations in r.iore recent years,
when
individual direct stock
we
share of total market capitalization. Finally,
changes
are associated with
ratio,
in
dividends,
i.e.
ownership represents a smaller
explore whether changes
in
share prices that
price fluctuations with a constant dividend price
have different effects on consumption spending than fluctuations that are not supported
by dividend movements.
4.1
Aggregate Statistics
We
c
t
)
begin by estimating regressing the growth rate
on lagged changes
Aln
(1)
c,
in real
share prices, Aim
= a + a(L)*Aln P
t .,
+
et
in real
per capita consumption (Aln
P,.,:
.
We consider equations with only the most recent lagged stock price change
side, as well as equations with a fourth-order lag polynomial o(L).
using seasonally adjusted quarterly data for the period 1947:2 to
Table
aggregates.
12 present the
The
results
of
estimating
results r the first row, for total
i
(1)
1
We
right
estimate equations
broad consumption
consumption and including only a single
lagged stock return, suggest that stock market fluctuations forecast increases
consumption outlays.
A
ten percent rise
in real
hand
995:2. 56
several
for
on the
stock prices predicts an increase
capita consumption of approximately 0.3 percentage points.
57
in
in real
real
per
Further lagged changes
in
56
Fischer and Merton (1 984) report some results for consumption growth as a function of
lagged stock returns. Hall (1978) found that lagged stock market returns were the only
variable known at the beginning of each quarter with predictive value for future consumption
changes.
"We tested for the possibility that stock price increases are associated with proportionally
different
changes
in
subsequent consumption than stock price decreases, but found no
evidence of such an effect.
34
share prices also have predictive power for consumption growth. The entry
of Table
2, also for total
1
consumption,
is
the
sum
values of stock price changes. The cumulative change
after a stock price increase
We
is
values of Aln
durables.
58
.064, more than twice the first-quarter effect.
We
P,.,
The
for the three
results in the lower
in
rows
of Table
12 show that stock price changes predict
spending for consumer durables.
in
durable outlays of
1.4%
the
ten percent increase
first
quarter, and
in
2.9%
six
times larger than that on nondurables, which
in
turn
is
several times
than the effect on consumption of services.
The
results in Table
change
in
1
2 focus on the change
in
stock prices, to avoid the simultaneity
consumption beginning
in
in
the quarter
contemporaneous stock returns and
consumption growth. This may lead us to underestimate the
is
in
A
Rising stock prices predict a proportionate increase in durable outlays
after four quarters.
between four and
for various
present results with both one lagged value and four lagged
share prices predicts an increase
after a
make
major subcategories of consumption: nondurables, services, and
the largest percentage change
larger
aggregate consumption four quarters
next consider the predictions that stock price fluctuations
categories of consumption.
the second row
on the four lagged
of the coefficients
in
in
total
change
in
consumption that
predicted by a stock price change, although our use of quarterly data should mitigate this
problem.
We have
explored the sensitivity of our findings to including the contemporaneous
stock market return.
of the coefficients
In
an equation
like
ln
in
the second row of Table
on four lagged stock market returns
market return has a coefficient of .01
58
that
1
equations not reported here,
(.007).
we
The
is
1
2,
where the sum
.064 (.014), the current stock
coefficients on the lacked stock market
also included four lagged values of the real
consumption growth rate in the specification. Only one of the four lagged values of
consumption enters with a statistically significant coefficient, and the long-run predicted
effect of stock price growth on consumption is very similar to that from the equation in the
second column.
35
returns are virtually
largest
unchanged when we add the current return
contemporaneous
correlation
is
point estimates
in
Table
1
is
of
durable spending
1
995
1.09%, and an increase
in
1994 was $591.5
outlays, and to a
$50.4
billion
prices translates into just over one
consumption spending
is
negative.
2 suggest that the increase
December 1994 and June 1995 (17.0%)
consumption
share prices between
predicts, after four quarters,
durable consumption of
in
billion, this
increase
trillion
in
4.92%.
corresponds to a $29
in total
dollars of
an increase
billion
consumption. Since a
1
in total
Since total
increment to
7%
rise in
potential
"control"
approximately .05 times the change
variables,
is
share
wealth creation, the predicted change
in
59
net worth.
in
Thus, the
conclusion that emerges from these consumption growth equations, which exclude
other
The
between stock returns and nondurable consumption;
the correlation with outlays on durables
The
to the specification.
many
very similar to that from traditional aggregate
consumption function analysis. The open question
is
whether these results
reflect the stock
market's role as a leading indicator, or whether they are partly due to a wealth effect
associated with stock price fluctuations. 60
4.2 Consumption and the Stock Market
In spite of
in
1987 and 1929
the long tradition of modelling aggregate consumption as a function of labor
59
One component of the link between stock price fluctuations and consumption involves
the "target saving" of defined benefit pension plans. When share prices rise, corporations do
not need to contribute as much to their pension plans to cover prospective pension liabilities.
This leads to a reduction in the flow of contributions to these plans; tnese contributions are
income accounts. Bernheim and Shoven (1988)
discuss this linkage between asset prices and saving in more detail.
classified as personal saving in the national
S0
The ideal test for distinguishing these views would study the reaction of consumption
autonomous changes in stock prices, changes that were not explained by revisions to
expectations about future cash flows or discount rates. Morck, Shleifer, and Vishny (1 990)
to
attempt a related test
in their
analysis of
how
the stock market affects investment.
36
61
there appears to be some reluctance to apply
income and household net worth
,
to analyzing the
evident
in
consumption effects
of large stock
market movements. This
discussions about the economic effects of the 1987 and
is
this
model
particularly
1929 stock market
crashes.
The stock market crash
of
October
effect of stock price fluctuations on
analysis of their effects.
August
1
987
1
987 provides
consumption spending, and to review the economic
Real share prices declined nearly thirty percent from their peak
Mechanisms concluded
that
"it is
of the
Presidential
Task Force on Market
unlikely that a direct wealth effect along the straightforward
described stands behind
movements and
fact that
in
to their value after the crash. In evaluating the potential effect of such a price
change on consumer spending, the report
lines usually
a valuable opportunity to study the
...
the observed relationship between stock price
aggregate-level consumer spending."
62
This conclusion
was based on
most households do not own stock, the highly-concentrated
ownership among those
who do own shares, and the view
that those
distribution
the
of
who do own substantial
stock have enough wealth to insulate their consumer spending from short-run shocks. 63
Popular accounts noted that a feared consumption collapse failed to materialize
after the crash,
61
in
the months
and that surveys showed that most consumers reported that they had not
and Deaton (1985) provide a recent discussion of aggregate consumption
Modigliani (1963) is the seminal paper on the empirical modelling of
aggregate consumption as a function of labor income and net worth.
Blinder
functions.
62
Ando and
U.S. Presidential Task Force on Market
Mechanisms (1988,
p. VII-2).
"Mankiw and Zeldes (1991) note that households that report owning stock in the Panel
Survey of Income Dynamics account for 32% of total food consumption in this data base.
Because the budget share of food is smaller for high-income, high-wealth households than for
lower income households, the fraction of total consumption accounted for by stockholders is
presumably greater than this. The fraction of consumption done by households with
substantial equity holdings is likely to be substantially smaller.
37
adjusted their spending patterns
Two important features
of the
stock price increase that preceded
crash.
The growth path
response to the crash.
in
1
987 stock market crash were the
and the rebound
it,
of both total
64
in
share prices
in
short duration of the
the quarter after the
and durable consumption, as well as the pattern
of
stock market returns, for the seven quarters centered on the October 1987 stock marekt
crash are
shown below:
.148
.019
.046
-.290
-.002
-.067
.010
.007
1988:1
.091
1988:2
1988:3
.006
-.023
.015
.004
.005
.032
.033
-.034
.047
.002
1987:1
1987:2
1987:3
1987:4 (Crash)
The stock
-.003
price increases in the three pre-crash quarters
prices finished
first
Growth Rate of Per Capita:
Consumption Durables Expenditure
Aln (Stock Price)
Quarter
quarter of
987
only
1988
left
1
The columns
slightly negative,
for
7.7% below
the market
were reversed by the crash, but stock
their value a year earlier.
in
March 1988 above
it's
The stock market
value
in
rally in
the
January 1987.
consumption growth show that per capita consumption growth was
and the growth
quarter of the crash. 65
-.010
If
we
in
spending on durables
was
include indicator variables for
regression equations for total consumption outlays reported
substantially negative, in the
1987:4 and 1988:1
in
in
the
Table 12, the resulting
"Pennar (1988) reports evidence of a Business Week /Harris Poll in which 85% of the
respondents indicated that the crash did not affect their finances, and noting that some
macroeconomic forecasters had revised downward their view of the "wealth effect" of stock
prices on consumption.
65
Dornbusch and Fischer (1994) note that consumption grew slowly
in support of a wealth effect on consumption.
use this as evidence
after the crash,
and
38
Thus the
coefficients are -.0096 (.0073) and .018 (.008), respectively.
first
quarter of
1
988
experienced more rapid consumption growth than would have been predicted by simple
models with four lagged quarterly values of stock returns.
hypothesis of no unusual effect on total consumption
patterns are slightly different.
is
The
effect of
1987:4
.071 (.042). Expenditure on durables did decline
but
it
was
unusually strong, given the decline
in
is
in
in
1
We
cannot reject the
987:4. 66
null
For durables, the
negative, but the 1988:1 coefficient
the quarter of the stock market crash,
share prices, during the
first
quarter of
1
988.
The data thus suggest that the 1987 stock market crash had a smaller negative effect on
consumption growth than the regression equations
predicted.
in
Table 12 would otherwise have
67
The events
of
1929 and the
early
1930s provide another opportunity to study the
effect of stock price fluctuations on consumption.
The data
for this period are less detailed
than for 1987, and the strength of the conclusions that can be drawn
lower.
Calculations that
assume
a stable marginal propensity to
is
correspondingly
consume out
of
wealth
suggest that the wealth effect of the 1929 crash on consumer spending should have been
small, both
because the stock market accounted
for a relatively small share of
household net
worth, and because the marginal propensity to consume out of wealth appears to be small
e6
The models in Table 1 2 related consumption growth to lagged returns. Since the stock
market crash occurred only three weeks into the fourth quarter of 1987, it is therefore
plausible to expect unusually low consumption growth in this quarter. While the coefficient
estimate on the dummy variable for 1987:Q4 confirms this, we are unable to reject the
hypothesis that this quarter's consumption growth is explained by the model that excludes
current returns.
and Miller (1987) conclude that the evidence suggesting that stock market
fluctuations cause consumption changes is weak at best. They present evidence of a very
weak association between the prices of New York City condominiums and changes in stock
market values, despite the fact that this is a luxury consumption item that might be demanded
by stock owners. This evidence is similar in spirit to our tests below for whether stock market
fluctuations affect the share of luxury consumption.
"Birinyi
39
during the inter-war period.
68
Household spending on durables declined more than that on
non-durables, which remained robust
of a stock price decline
until
1932. 89 Thus
it
appears that the negative effect
on consumers, through the wealth effect, was muted.
4.3 Stock Returns and Spending on "Luxury Goods"
We
investigate whether stock price fluctuations affect
consumer spending through
wealth effect by examining whether stock returns forecast changes
consumer spending.
We
in
a
the composition of
70
(CEX) to identify several
use the Consumer Expenditure Survey
groups of goods that are disproportionately consumed by high-income households that are
likely to
own
stock, and test whether the share of these goods
after stock prices increase.
luxury products
The
in
in
aggregate consumption rises
Anecdotal evidence suggesting a very strong market for some
1995, possibly related to the
results of our analysis of
rise in
share prices.
consumption patterns are shown
both the share of spending on particular items that
is
accounted
before-tax annual incomes of $70,000, the value at which income
and the
68
ratio of
71
in
for
in
Table
1
3.
We report
by households with
the
CEX
is
top-coded,
spending by this group to spending by households with before tax income of
A
constant marginal propensity to consume out of wealth is a specialized result that
when a consumer maximizes a time-separable utility function with perperiod utility given by log(C). More generally, the marginal propensity to consume out of
wealth depends upon the available rate of return.
obtains for example
69
Temin (1976) discusses the effect
Romer (1990) draws particular attention
1929 stock market crash on consumption.
consumer uncertainty engendered by
the crash in depressing household durables purchases. Wigmore (1985) examines the
behavior of the earnings and share prices of companies in various sectors of the economy, and
of the
to the role of
notes the relatively stable earnings of retailers until 1 932. Durables producers, notably auto
companies, experienced sharp downturns in profits and share prices much earlier.
"U.S. Bureau of Labor
71
Bird (1995).
Statistics (1994).
40
$20-30,000.
31
.4%
spending on
of
$70,000 and above account
For example, households with incomes of
new
cars, while they
account
23.6%
for
of
spending on
all
for
goods.
Table 14 reports the results of estimating consumption share equations that are
designed to evaluate whether increases
share prices
in
tilt
the mix of consumption toward
goods that are consumed by higher-income, stock-holding households.
new
considers the case spending on
predicted to raise spending on
is
next quarter, and by
6.3%
share prices, similar to that
new
A 10%
cars.
new
the
share prices
in
The
four-quarter effect of a
1995, would be a
of
first half
72
rise
row
first
the current quarter
cars as a share of total consumption by
after four quarters.
in
rise in
The
2.5%
20%
in
the
rise in
from 2.3% to 2.7%
in
car spending as a percentage of total spending.
Whether these
findings for
new
car sales reflect the operation of a wealth effect, or
simply the stock market's forecast of strong consumer demand, can be evaluated by
demand
considering the relative
for different
types of automobiles.
Ward's Automotive
Yearbook allocates new cars to several different categories, one of which
includes
1
most cars with
prices
above $25,000
.22 million were classified as luxury cars.
accounted
5-, 6-,
for .1
7 million cars
in
1
1994,
additional "upper luxury" category,
which
and 7-series, three Jaguars, and Mercedes
prevalent, and
many
in
0ne
difficulty
in
new
and S-class cars. The households
are likely to have substantial equity portfolios.
new
with interpreting these results
is
who
may
rise
we
is
shows both
1980-1995
period.
estimated a regression
that automobile manufacturers and their
market value. If investors foresee
a nontrivial fraction of total stock
car sales, stock prices
Figure 5
car sales for the
effect of share prices on luxury car purchases,
components suppliers are
an increase
E-
BMWs in the
the wealth category where stock ownership
luxury and upper-luxury cars as a percentage of
To study the
This
million cars sold in
in
An
"luxury."
994, consist of only fifteen models, including
purchase these vehicles are almost surely
72
1994. Of 8.99
is
as a
result.
Figure 5
o
Luxury Cars
0.16 -
0.14 -
-0.03
en
L.
in
0.12
ID
-
C-J
c_
>.
ID
C
3
C
3
0.10 -
-0.02
C
CD
Q.
Q.
3
-0.01
1980
1983
1989
1985
1992
Year
Luxury Can Sales, 1980-1995
1995
41
equation for the luxury fraction of
consumer spending through
new
fraction of
for the period
a
new
car sales
in
each quarter.
If
share price changes affect
wealth effect, then stock returns should positively affect the
car sales that are classified as luxury vehicles.
We
estimate this relationship
1980:1-1995:2, the sample over which the Ward's data were
regression equation, which includes unreported seasonal indicator variables,
= -0.283
log(luxury t /allcar,)
(0.130)
These
results
-
0.084 *Aln P
t .,
+ 0.821
available.
is
The
shown below:
R 2 = .763
•log(luxuryt.,/allcar,.,)
(0.064)
(0.098)
do not support the operation
of
an important wealth effect. The unreported
seasonal coefficients suggest that purchases of luxury cars reach their highest share of
vehicles
the fourth quarter of each year, and on average account for a
in
of total vehicle sales in
Q4
We also estimated
The
autos sold.
than
in
Q1
,
and 5.4% more
in
Q4
than
in
14.5%
greater share
Q3.
a similar equation for the share of "upper luxury" cars in the
results are
shown below, and
provide very
little
all
mix
of
support for an operative
wealth effect:
log(upp luxt /allcar
t
)
= -0.223 + 0.017'Aln
(0.244)
P,.,
(0.046)
+ 0.900*log(upp
lux^/allcar,.,)
is
= .824
(0.056)
While stock prices are positively correlated with the upper luxury share of
the effect
R2
new
vehicle sales,
not statistically significant. These results must also be viewed as providing
littel
support for the wealth effect, as opposed to the alternative "leading indicator" explanation of
the correlation between contemporaneous stock returns and future consumption growth.
The
results for the other luxury items
luxury cars.
1
4.
in
The regression
we
consider are consistent with the findings for
coefficients on each consumption category are
shown
in
Table
We cannot reject the null hypothesis that stock price changes do not predict any change
the share of aggregate consumption accounted for by education, domestic services, hotel
and motel spending, and entertainment spending.
In
most cases the standard
errors on the
42
estimates are large, admitting large positive or negative effects, but the point estimates are
not supportive of wealth effects.
The two
"anti-luxuries" considered at the
bottom of Table 14 provide more support
the proposition that stock returns affect the consumption mix.
associated with a decline
in
the share of total consumption that
is
for
Rising share prices are
devoted to rental housing,
and with a weak negative effect on the share of tobacco spending. These results provide
more support
for operative
wealth effects associated with stock price fluctuations, and leave
us with mixed results from our analysis of luxury goods.
4.4 Stock Prices and Consumption: Household-Level Evidence
Aggregate data appear to have limited power to resolve whether stock returns exert
a significant wealth effect on consumption.
is
An
alternative source of information
household survey data. One of the standard problems with such analysis
of data sets that collect information
on both household
portfolio holdings
is
on
this issue
the absence
and consumption.
The Panel Survey
of
Income Dynamics, an annual survey that collects information during the
few months
of
each year,
first
information on consumption
help resolve the "equity
is
is
one survey that does contain such data, although direct
limited to outlays
premium puzzle"
stratified
on food.
Ar,
important study designed to
PSID households
into those that
own stock
and those that do not, and then computed the correlation between per capita food
consumption growth and excess returns on the stock market
households. 73 The results suggest that this correlation
own
is
for
these two groups of
substantially higher for those
who
stock than for others. These results could be due to differences between stockholders
and non-stockholders that are not related to
73
Mankiw and Zeldes (1991).
their share
ownership per se but they at least
.
43
consumption through wealth effects.
raise the possibility that stock returns affect
We
revisit
the Mankiw-Zeldes analysis, but introduce three modifications.
addition to considering food consumption growth,
consumption from the PSID.
It
as well as spending for food at
is
of the variation in total
also analyze a broader
home and food away from home
74
we
Second,
upon food consumption as an
exploit information that has
participation in pension plans
have
72%. Thus
it
seems
possible to
indicator of household consumption.
been released since the Mankiw and Zeldes study,
and employer
thrift
indirect holdings of corporate stock, but
of traded equity
to obtain a proxy for total
consumption the Consumer Expenditure Survey, a consumption
notably the additional years of consumption data.
to
of total
While food consumption alone explains only
proxy including house value and rental payments explains
substantially improve
measure
possible to use information on rent payments, house value,
household consumption for PSID households.
26%
we
First, in
Third,
we
use PSID information on
plans to identify households
who
are likely
no direct holdings. While direct ownership
and mutual funds cannot be distinguished
between those with such equity holdings and those with equity
in
in
the PSID, the distinction
401
(k)
plans, 403(b) plans,
or defined contribution pension plans is relevant for our analysis of direct vs. indirect stock
holdings.
75
Our analysis follows previous work
74
in
defining the sample for analysis.
76
Skinner (1987) develops this broader consumption measure, and explores
In particular.
its statistical
properties.
75
The PSID survey asks households whether they own "any shares of stock in publicly held
corporations, mutual funds, or investment trusts, including stocks in IRAs."
76
We
follow the sample selection and data definition rules described in the appendix to
(1 991 and Zeldes (1 989). When we test for differences in the correlation
between consumption growth and stock returns between households that hold stock, and
households that do not, for the sample period considered by Mankiw and Zeldes (1991), we
Mankiw and Zeldes
)
obtain results that are broadly similar to theirs.
44
we
exclude from the PSID families
who were
that are living with other families (which
family-years
in
which the
family-years
in
which any component
makes
identity of the
of
it
1968 poverty subsample,
difficult to distinguish
consumption
families
outlays),
household head or the head's spouse changed, and
consumption was topcoded. Because the PSID
not collect information on food consumption
period
part of the
in
did
1973, 1988, or 1989, our sample spans the
1970-1992 but excludes 1973, 1974, 1988, 1989, and 1990.
We
PSID households along three dimensions. The
stratify
We
report ownership of corporate stock or mutual funds.
first is
distinguish
whether
or not they
between households
with different levels of equity ownership, considering "stockholders" to be those with any,
more than $1000,
consider
and the
is
or
more than $10000
whether the head
in
corporate stock. 77
of household or the head's
third is participation in a
78
pension plan.
We
The second dimension we
spouse participates
in
a thrift plan,
then consider the correlation between
the year to year growth of the Skinner consumption aggregate, and stock market returns, for
groups of households with different combinations of direct equity holdings,
pension coverage.
related
we
changes
in
thrift plans,
and
While our previous analysis focused on quarterly time series data, and
consumption to lagged stock market returns, the PSin data are annual and
therefore focus on contemporaneous changes
"These thresholds
for the definition of equity
in
share prices and consumption. 79
ownership are measured
in
1984
dollars.
78
The PSID only collects detailed information on pension plan characteristics for workers
over the age of 45. For younger workers, we cannot distinguish between defined benefit and
defined contribution plans. We therefore use only an indicator variable for "pension plan
participant" in our analysis, recognizing that this combines some DB plan participants with no
equity stake with equity holders through defined contribution plans.
We
79
have estimated the correlation between current consumption growth and lagged
stock returns for the PSID households, and the resulting correlations are indistinguishable from
zero. This is due to our choice of sample period; a similar result emerges from annual national
accounts data for the PSID sample years, even though our findings in Table 1 2 show that over
a longer period, quarterly national accounts data suggest a positive correlation between stock
45
Table
1
5 presents the results of our analysis of these consumption correlations.
upper panel reports results for food consumption growth, while the lower panel
with growth
in
the Skinner consumption index.
small differences
80
who
stock holders than for those
is
greater for those
"stockholders" defined as having
between the growth
who
more than $10000
statistical tests.
rate of the Skinner
It
is
disturbing that
non-stockholders, those
who do
is
positive.
ownership are suggestive, but not conclusive.
not have
which are defined contribution
correlation than those without
such plans. 81
that even indirect stock holding matters
price
of the
when we
consider
who have thrift plans exhibit greater correlation between consumption
and stock returns than those
of
equity,
in
many
between food consumption growth and excess returns
results with respect to indirect stock
some
of these
consumption index and excess stock
For both food consumption and the Skinner index of consumption growth,
plans,
are classified as
This appears to be due to the non-food components of the Skinner
index, since the correlation
The
find
correlation for stockholders
With the Skinner index, only one
are not.
approaches significance using conventional
returns are negative.
we
For each definition of stock ownership, however, the correlation
between consumption growth and stock returns
differences, that for
concerned
use the Skinner index,
between the consumption growth-stock return
and non-stockholders.
correlations
When we
is
The
in
thrift plans.
Similarly, those with pension
plans, also exhibit a higher
These
consumption
results are consistent with the notion
household consumption planning, but they do not
changes and consumption growth.
80
The differences in the stock returns-food consumption growth correlations between
between stockholders and non-stockholders parallel those in Mankiw and Zeldes (1 991 ). Our
results from the longer PSID sample period yield less statistically significant differences in
these correlations than in the Mankiw and Zeldes (1991) study.
81
For consumption growth, these findings emerge for two of the three definitions of
"stockholding," any equity held and stock worth more than $10,000.
46
reject the null hypothesis of
no effect. For those households
Those
of the patterns are difficult to explain.
correlations than those
who
are not
in
in thrift
such plans
who own
plans exhibit
for
two
stock directly,
some
weaker consumption
of the three definitions of
stockownership.
The
they are
results of this analysis are stronger than the findings using aggregate data, but
not conclusive. Stock returns are more closely correlated with the consumption
still
growth of households
who own
who do
stock than of those
The PSID data do not have
not.
enough power, however, to distinguish between the view that only
direct stock ownership
matters, and the possibility that either direct or indirect stock ownership has a similar effect
on consumption growth. The results are consistent with the presence of a wealth effect, but
they are not definitive support for
those with larger stock portfolios
consumption
it.
is
There
is
only
weak evidence
that the consumption of
more closely correlated with the stock market than
is
the
of those with small portfolios.
4.5 The Ownership of Shares and the Consumption-Stock Price Nexus
The
third
broad issue that
ownership patterns on the
link
we
consider concerns the effect of changing stock
between stock returns and consumption.
question by interacting the lagged stock return variable
in
equation
the fraction of corporate stock held by households. 82
corresponds to direct individual ownership,
stock ownership (excluding
82
nonprofit
is
The
(1
first
)
with
We
study this
two measures
such measure, which
the ratio of the Flow of Funds household sector
holdings)
to
the
total
market value of shares
We also tested the sub-sample coefficient stability of the equations reported in Table
sample before and after 1970:1, and before and after 1985:1. We could
dividing the
any of the consumption categories;
the trending shareownership variables.
reject the null hypothesis of constant coefficients for
foreshadows our weak results
for
of
1
2,
not
this
47
outstanding. This time series declines substantially during the postwar period, from
1960, to 51.3%
in
in
1980, to 40.2%
in
1994.
We amend
equation
(1)
72.3%
to include lagged
stock price changes as well as the lagged price change interacted with the direct share
ownership measure (SHARE1,.,):
Aln c t
(2)
Because Aln
P,.,
= o + a/AIn P
and SHARE1,., *Aln P ., are
t
that only include the interaction term.
addition to our
In
equation
in
(2)
measure
directly or through
it
.,'Aln P
+
t .,
we
collinear variables,
ev
also estimate equations
83
of direct shareownership,
series for
SHARE2,
bank personal
SHARE2,
is
is
SHARE 1,, we
the percentage of corporate stock
trusts,
also estimate
shown
owned by households
mutual funds, and defined contribution pension plans.
in
the last column of Table 6. This time series declines
more gradually over the postwar period than SHARE1
1994,
t
with a variable that captures direct as well as indirect share ownership, SHARE2,,
the interaction term.
The time
+ o2 # SHARE1
t .,
.
In
1970,
in
Table
SHARE2 was 75%, and
by
had declined to 63.7%.
The
results of estimating equation (2) are
shown
1
6.
The
first
row shows an
equation relating consumption growth to lagged stock returns without any allowance for share
ownership effects. The second row shows the effect of including the interactive term with
SHARE 1,. The
coefficient on the lagged stock price
interaction term
change
in
the
is
first
positive
row.
change
is
negative, and that on the
and roughly twice as large as the coefficient on the stock price
Unfortunately, the collinearity of these
impossible to reject the hypothesis that either
o, or
a2
in (2) is
two
variables
zero, although
we
makes
it
reject the
hypothesis that these coefficients are jointly zero. The third row presents an equation that
"Including a trending variable such as SHARE,., as a separate regressor in these equations
results, because there is little trend in the rate of consumption growth.
does not affect the
48
includes only the
variable alone
is
SHARE1,*Aln
fifth
rows
in
rows, with SHARE2, replacing
to those for
variables,
SHARE 1
and
it
variable,
and shows that the explanatory power
greater than that of the lagged stock price variable alone.
The fourth and
third
P,.,
1
;
Table
1
SHARE1
the interaction term
show
6
is
,
in
84
results similar to those in the
the interaction terms.
the more important
in
The
of this
second and
findings are similar
the specification with both
has a positive effect on consumption while the lagged share price term alone
has a negative coefficient. Yet the collinearity problems remain
in
On
these specifications.
balance, the results do not support the view that changing patterns of stock ownership have
altered the link
between share
price fluctuations
and consumption. 85
4.6 The Forecast Power of Stock Returns, Dividend Movements, and Earnings
A
consider
final
is
aspect of the linkage between stock returns and consumption that
whether the source
of stock price fluctuations affects the predictive
stock returns for future consumption growth.
in
stock prices that are not associated with
transitory than similar price
differences
in
we now
power
Given the evidence that fluctuations
dividend-price ratio and earnings-price ratio predict future stock returns,
increases
Movements
rising
in
of
the
which implies that
dividends are more
likely to
be
changes backed by dividend fluctuations, one might expect
the predictive effects of different stock market shocks.
To explore whether
fluctuations
in
share prices, dividends, earnings,
or
some
84
A non-nested hypothesis test of the model with SHARE1,*Aln P,.,, against the model
with Aln P,., as the dependent variable, does not permit us to reject either hypothesis in favor
of the other.
85
0ne way to develop additional tests of whether stock ownership patterns affect the
wealth effect of stock prices on consumption would involve analyzing data from different
countries.
In Japan, for example, the fraction of shares held directly by individuals is
substantially lower than in the United States.
49
combination of these variables has the greatest predictive power for consumption spending,
we augment
payments
equation
for the
If
S&P
Aln c
(3)
(1), for
t
the simple case of a(L) =
Industrials (Aln
= a +
P
cr,* Aln
share price fluctuations predict the
source, then y should equal zero.
If
:
prices
move with D/P
with the lagged change
a,,
in
dividend
D,.,):
+ h # Aln D
t .,
same change
t .,
in
+
et
.
future
consumption regardless
stock prices only predict changes
in
of their
consumption when
constant, then y should be positive, and a, should be indistinguishable
:
from zero.
Table
growth
1
rates,
7 shows the results of estimating equation
and also lagged earning growth
suggest that changes
their source.
in
(3).
We report results with dividend
rates, in the specification.
stock prices predict similar changes
in
consumption, regardless of
The equations that include lagged dividend growth provide no support
view that dividend fluctuations can predict future consumption growth.
null
hypothesis that changes
in
dividends
real
real
dividends
in
real
share prices
for the
We cannot reject the
have no predictive power
consumption growth, and the estimated coefficient on
by including
The estimates
is virtually
for
future
unaffected
these equations. These results are insensitive to our choice of
consumption aggregate.
The
four broad
results with lagged earnings
consumption categories,
growth are more
total
difficult to interpret.
is
we estimated models
and four lagged values of
(not
in all
specifications.
shown) with four lagged values
real earnings
of the
marginally significant.
growth does not reduce the coefficient on the lagged change
stock prices, but actually raises this coefficient
results,
two
consumption and spending on durables, the lagged
earnings growth variable enters with a positive coefficient that
Controlling for earnings
For
As a
in real
further test of these
of real stock price changes,
growth. The results do not support the notion that
50
earnings fluctuations are an important determinant of consumption growth. The
coefficients on the four lagged stock return terms
different
is
positive
and
of the
statistically significantly
from zero. The sum of the coefficients on lagged earnings, however,
statistically insignificantly different
sum
is
negative ana
Thus, the evidence seems to suggest that
from zero.
share price increases have similar effects on consumption, regardless of their source.
We
rates,
also followed a separate strategy to identify the effects of shocks to discount
We
and shocks to expected cash flows, on consumption growth.
estimated a
first-
order autoregression for the dividend-price ratio, and defined the residuals from this equation
86
as estimates of the shock to discount rate expectations.
period
1947:2 to 1995:2, generates an estimated coefficient
value of D/P.
We
analogous to those
then include the value of
in
Table
1
2.
The
[(D/P),.,
-
This autoregression, for the
of .961 (.021)
.961 *(D/P),. 2 ]
results suggest that
it
is
in
regression equations
not possible to distinguish the
effects of discount rate shocks and cash flow shocks with the available data.
the estimation equation for total consumption
Aln c
t
= .004 + .013*Aln
(.001)
The imprecision
P,.,
(.016)
-
on the lagged
For example,
is:
.004 # [(D/P)
t .,
-
.961 *(D/P) t
.
R 2 = .0719
2]
(.003)
of the coefficient estimates
makes
it
difficult to
argue that shocks to
expected returns are more or less important than shocks to cash flow
in
predicting future
consumption growth.
5.
Conclusions and Future Directions
This paper documents substantial changes during the postwar period
86
in
the aggregate
This procedure follows the suggestion of Fama and French (1 988), who argue that "the
unexpected component of D/P can be interpreted as a (noisy) measure of the shock to
expected returns (p. 20)."
51
and cross-sectional patterns of corporate stock ownership.
directly
by households
in
the
1
950s and
early
1
While most shares were held
960s, there has been a gradual but significant
trend toward greater ownership of equity through mutual funds and through thrift plans (e.g.
401
(k)s,
ESOPs, and 403(b)s) and defined contribution pension
direct individual
but the rise of IRAs,
in
each
of these
cases
ownership has been replaced by indirect ownership through a financial
The incidence
intermediary.
plans. In
of stock
thrift plans,
ownership remained stable during the
and other related
institutions
1
960s and
1
970s,
has led to a substantial increase
stock ownership during the most recent decade.
We explore
the implications of growing intermediary ownership for the effect of stock
We find clear evidence that changes in share prices
price fluctuations
on consumer spending.
portend growth
consumer spending, with
durables.
The
first is
We
in
a particularly large effect for outlays
then try to distinguish between two alternative explanations for this finding.
the view that stock returns are a leading indicator, reflecting
a prospective increase in
second view
is
on consumer
news
that suggests
consumption before the spending change actually occurs.
The
the traditional wealth effect of asset market fluctuations, which suggests that
higher stock prices should
I'
ad to an increase
in
consumer spending.
We investigate the effect of stock returns on the share of consumption that is devoted
to luxury goods, and also study household level data from the Panel Survey of
Dynamics on the
that
correlation
between consumption growth and stock returns
own, and do not own, corporate stock.
price fluctuations
We do
for
Income
households
not find any pronounced effects of stock
on the mix of luxury and non-luxury consumption within the following year.
This evidence casts doubt on the short-run consumption consequences of wealth effects
52
We
87
associated with stock price movements.
do not
patterns of share ownership have altered the relationship
find
any evidence that changing
between stock
consumption, even though such effects might be expected
in
some
price fluctuations and
behavioral models of
saving and consumption determination.
These findings represent
consumption function that
found
is
in
many macroeconomic
budget constraints suggests that stock market
reflected either in higher
greater bequests.
It
is
model
a challenge to the traditional
of the
textbooks.
88
aggregate
The
that increase household wealth
rallies
logic of
must be
consumption during the lifetimes of current stock holders,
possible that consumption responds gradually to increases
in
or in
stock
market wealth, and that our focus on consumption fluctuations within a year of stock price
movements does not capture these
effects.
It
is
also possible that the effect of stock price
fluctuations on consumption operates through channels other than a direct wealth effect, for
example by alterning consumer confidence. More generally, our findings suggest the need to
develop better data, and possibly better models, for the determination of consumption
spending by high-wealth households.
While our primary focus
between the
for future
predictive
power
is
on testing
for
wealth effects,
of stock price changes,
in
in
dividends are transitory, yet
that consumption evolves differently after increases
lt is still
as suggested
98
and changes
also investigate the links
dividends and earnings,
consumption growth. Substantial evidence suggests that changes
that are not associated with changes
87
we
in
we
in
share prices
find find
no evidence
share prices associated with dividend
possible that cash realizations of past gains affect current consumption outlays,
in
Poterba (1991).
The findings in many ways parallel Morck, Shleifer, and Vishny's (1990) findings that
the stock market's role in predicting investment movements was largely due to its role as a
passive informant of future developments.
53
increases and changes associated with discount rate movements.
Our analysis focuses on a
prices,
and leaves many issues
relatively limited set of
household responses to higher stock
We
for further research.
have not considered the
example to
possibility
that higher share prices lead to
changes
those nearing retirement age.
Rising share values that lead to increases in household net
worth may also trigger changes
employment and
labor supply, for
in
occupation, for example
in
to strike out as an entrepreneur, and
Changing share prices
may
may
earlier retirement
decisions to leave
by
paid
also encourage retirement.
89
"consumer confidence," and influence spending
also affect
decisions through this poorly-understood channel.
Some
of the
most important issues that
arise
from changing stock ownership patterns,
and which our tests have not addressed, concern the effect of ownership structure on the
performance of asset markets.
If
the switch from dispersed individual ownership to
concentrated stock holdings alters the
stock returns, then
it
trader" models in financial
who may
money managers as
availability
economics
how
much
and market
new
information or to past
many aspects
volatility.
90
of
macroeconomic
Many discussions
implicitly portray individual investors
of "noise
as the poorly
be affected by fads or other investment trends, with professional
the arbitrageurs
requires information on
range of dimensions;
investors respond to
could have significant implications for
performance, including capital
informed traders
way
who
trade against this group.
institutional investors differ
of the core research
on
Exploring these issues
from individual investors along a
this issue
remains to be done.
"Holtz-Eakin, Joulfaian, and Rosen (1 994) present empirical evidence suggesting that the
become self-employed is sensitive to changes in net worth, in their case the
receipt of a bequest. Samwick (1995) summarizes the available evidence on the effects of
decision to
financial assets
90
on retirement decisions.
Friedman (1995).
54
REFERENCES
1963. "The 'Life Cycle' Hypothesis of Saving:
Ando, Albert, and Franco Modigliani.
Aggregate Implications and Tests." American Economic Review 53 (May), 55-84.
1994. "A Comparison of the Household Sector from the Flow of
Funds Accounts and the Survey of Consumer Finances." Washington: Board of
Governors of the Federal Reserve System, Division of Research and Statistics.
Antoniewicz, Rochelle.
1994. "Personal Saving in the United States." In J. Poterba, ed..
Comparisons of Household Saving Chicago: University of Chicago Press,
Attanasio, Orazio P.
International
.
57-124.
Avery, Robert B. and Gregory E. Elliehausen. 1 986. "Financial Characteristics of High-Income
Families." Federal Reserve Bulletin 72 (March), 163-77.
Avery, Robert B. and Gregory E. Elliehausen. 1988. 1983 Survey of Consumer Finances:
Washington: Board of Governors of the Federal
Technical Manual and Codebook
Reserve System.
.
Avery, Robert
Gregory
B.,
"1983 Survey
of
E.
Elliehausen, Glenn B. Canner,
and Thomas A. Gustafson.
Consumer Finances." Federal Reserve
Bulletin
1
984.
70 (September),
679-92.
Avery, Robert B., Gregory E. Elliehausen, and Arthur B. Kennickell. 1 988. "Measuring Wealth
with Survey Data: An Evaluation of the 1 983 Survey of Consumer Finances," Review
of Income and Wealth (December), 339-369.
Bagwell, Laurie, and John Shoven. 1989. "Cash Distributions to Shareholders." Journal of
Economic Perspectives 3, 129-140.
1990. "The Stock Market and Investment." Review of Financial Studies
115-131.
Barro, Robert.
3:
Bernheim, B. Douglas, and John B. Shoven. 1988. "Pension Funding and Saving." In Z.
Bodie, J. Shoven, and D. Wise, eds.. Pensions in the U.S. Economy
Chicago:
University of Chicago Press.
.
Bird, Laura.
1
995. "Tired of T-Shirts and No-Name Watches, Shoppers Return to Tiffany and
Chanel." Wall Street Journal 6 September, B1.
Laszlo
Birinyi,
and Keith
Jr.,
Miller.
"The Wealth Effect."
1987.
New
York: Salomon
Brothers.
Blinder,
Alan
S.,
Revisited."
Blume, Marshall
and Angus Deaton.
1985.
BPEA 2:1985 .465-512.
E.,
Jean Crockett, and Irwin
"The Time Series Consumption Function
Friend.
1974. "Stock Ownership
in
the United
55
States: Characteristics and Trends." Survey of Current Business (November), 16-40.
Blume, Marshall E. and Stephen P. Zeldes. 1993. "The Structure of Stock Ownership
United States," mimeo, Wharton School, University of Pennsylvania.
in
the
1991. "The Decline in Saving:
Bosworth, Barry, Gary Burtless, and John Sabelhaus.
Evidence from Household Surveys." Brookings Papers on Economic Activity 2.
Campbell, John Y. and Robert J. Shiller. 1988. "Stock Prices, Earnings, and Expected
Dividends." Journal of Finance 43: 661-676.
Campbell, John Y. and Robert J. Shiller. 1989. "The Dividend Price Ratio and Expectations
of Future Dividends and Discount Factors." Review of Financial Studies 1: 195-228.
Cochrane, John H. 1991. "Production Based Asset Pricing and the Line Between Stock
Returns and Economic Fluctuations." Journal of Finance 46, 209-237.
1994. "Permanent and Transitory Components of
Quarterly Journal of Economics 109, 241-265.
Cochrane, John H.
Prices."
GNP
and Stock
Thomas Juster, and James N. Morgan. 1989. "Survey Estimates of
An Assessment of Quality," in Richard Lipsey and Helen Tice, eds., The
Curtin, Richard T., F.
Wealth:
Measurement
of Savings.
Income, and Wealth
Dornbusch, Rudiger, and Stanley Fischer.
McGraw
Evidence."
994. Macroeconomics. Sixth Edition (New York:
A
1981. "Municipal Pension Funding:
141-178.
Theory and
Public Choice 37,
Fama, Eugene F. 1981. "Stock Returns, Real
Economic Review 71, 545-565.
Fama, Eugene
Chicago: University of Chicago Press.
Hill).
Epple, Dennis, and Katherine Shipper.
Some
1
.
and Kenneth
Activity, Inflation,
and Money."
American
1988. "Dividend Yields and Expected Stock
Returns." Journal of Financial Economics 22, 3-25.
F.,
R. French.
and James Poterba. 1983. "The Effective Tax
Rate and the Pretax Rate of Return." Journal of Public Economics 21 1 29-1 58.
Feldstein, Martin S., Louis Dicks-Mireaux,
,
and Lawrence H. Summers.
1:1977 211-227.
Feldstein, Martin S.,
1
977.
"Is
the Rate of Profit Falling?"
BPEA
.
and Robert Merton. 1 984. "Macroeconomics and Finance: The Role of the
Stock Market." Carnegie-Rochester Conference Series on Public Policy 21: 57-108.
Fischer, Stanley,
French, Kenneth R. and James M. Poterba. 1 991 "Were Japanese Stock Prices
Journal of Financial Economics 29, 337-363.
.
Too High?"
56
Friedman, Benjamin M.
of Portfolio
1
995. "Economic Implications of Changing Share Ownership." Journal
Forthcoming.
Management
.
Gentry, William. 1994. "Annuity Markets and Retirement Saving." In Proceedings of the
National Tax Institute-Tax Institute of America Columbus: National Tax Association,
.
178-183.
Haliassos, Michael, and Carol Bertaut.
Journal 105, 1110-1129.
Hall,
1995.
"Why Do So Few
993. "Industrial Research During the
Microeconomics BPEA 1993:2 289-244.
Bronwyn
H.
1
1
Economic
Hold Stocks?"
980s: Did the Rate of Return
Fall?"
.
Hall,
Robert E. 1978. "The Stochastic Implications of the Life Cycle-Permanent Income
Hypothesis: Theory and Evidence." Journal of Political Economy 86, 971-987.
Hall,
Robert
E.
1988.
"Intertemporal Substitution
in
Consumption."
Journal of Political
Economy 96, 339-357.
Heeringa, Steven G., Judith H. Connor, and R. Lou'se Woodburn. 1 994. "The 1 989 Surveys
of Consumer Finances Sample Design and Weighting Documentation." Ann Arbor, Ml:
Institute for Social Research, Working Paper.
1991. "Dividends and
Working Paper 3730.
Hines, James.
Profits:
Some
Unsubtle Foreign Influences."
NBER
Holtz-Eakin, Douglas, David Joulfaian, and Harvey Rosen.
and Liquidity Constraints."
Ibbotson Associates. 1994.
Ibbotson Associates.
RAND
1994. "Entrepreneurial Decisinos
Journal of Economics 25, 334-347.
Stocks Bonds
Bills
and
Inflation:
1
994 Yearbook
.
Chicago:
1992 Survey of Consumer Finances: Technical Manual and
Washington: Board of Governors of the Federal Reserve System.
Kennickell, Arthur B. 1995.
Codebook
.
and Janice Shack-Marquez. 1992. "Changes in Family Finances from
1983 to 1989: Evidence from the Survey of Consumer Finances." Federal Reserve
Bulletin 78 (January), 1-18.
Kennickell, Arthur B.
and Martha Starr-McCluer. 1 994. "Changes in Family Finances from
1989 to 1992: Evidence from the Survey of Consumer Finances." Federal Reserve
Bulletin 80 (October), 861-882.
Kennickell, Arthur B.
King,
Mervyn A. and Jonathan Leape. 1 984. "Wealth and
Evidence." NBER Working Paper 1468.
Portfolio Composition:
Theory and
Levy, Frank, and Richard Murnane. 1992. "U.S. Earnings Levels and Earnings Inequality: A
Review of Recent Trends and Proposed Explanations." Journal of Economic Literature
.
57
30, 1333-1381.
Mankiw, N. Gregory and Stephen
P.
Zeldes. 1991. "The Consumption of Stockholders and
Nonstockholders." Journal of Financial Economics 29: 97-1 12.
"The Stock Market and
1990.
Morck, Randall, Andrei Shleifer, and Robert Vishny.
Investment: Is the Market a Sideshow?" BPEA 2:1990 157-202.
.
1974. "The
Nordhaus, William.
988.
18. 56-59.
Pennar, Karen.
Poterba,
1
"It's
James M. 1991.
Share of
Falling
Almost As
BPEA 1:1974 169-208.
Profits."
.
Never Happened
If It
-
Almost." Business
Week
April
"Dividends, Capital Gains, and the Corporate Veil: Evidence from
Canada, and the United States." In B. Douglas Bernheim and John B. Shoven,
and Economic Performance Chicago: University of Chicago
Press, 49-71.
Britain,
eds.. National Savina
.
James M. 1995a. "A History
Department of Economics.
Poterba,
of Annuities in the United States."
Mimeo, MIT
James M. 1995b. "Corporate Stock Ownership in the United States." In New York
Stock Exchange, The Extent of Shareownership Among U.S. Households New York:
Poterba,
.
New
York Stock Exchange.
James M., Steven F. Venti, and David A. Wise. 1994. "Targetted Retirement
Saving and the Net Worth of Elderly Americans." American Economic Review 84
Poterba,
(May), 180-185.
Poterba,
James M., Steven
F.
Venti, and David A. Wise.
from Retirement Saving Plans: Receipt and
Projector,
Dorothy
S.,
"Lump Sum
NBER Working
1995.
Utilization."
Distributions
Paper.
and Gertrude S. Weiss. 1966. Survey of Financial Chaiacteristics of
of Governors of the Federal Reserve System).
Consumers (Washington: Board
Romer, Christina.
1990.
"The Great Crash and the Onset of the Great Depression."
Quarterly Journal of Economics 55, 597-624.
Samwick, Andrew. 1995. "The
of Retirement:
Schultz, Ellen E.
Some New
1995.
"Tidal
Joint Effect of Social Security and Pensions on the Timing
Evidence."
Wave
Mimeo, Dartmouth College.
of Retirement
Cash Anchors Mutual Funds."
Wall
Street Journal September 27, C1
Shleifer, Andrei,
and Lawrence Summers. 1990. "The Noise Trader Approach to Finance,"
Journal of Economic Perspectives
Shulman, David
S.,
4
(Spring), 19-33.
1995, "The Ghost of Keynes," Salomon Brothers Equity Strategy: July
58
(New York: Salomon
Brothers).
1987. "A Superior Measure of Consumption from the Panel Survey
Dynamics."
Economic Letters 23, 213-216.
Income
Skinner, Jonathan S.
Stock,
of
James H. and Mark W. Watson. 1990. "Business Cycle Properties of Selected U.S.
Economic Time Series, 1959-1988." Working Paper 3376. Cambridge, MA: National
Bureau of Economic Research.
Summers, Lawrence
H.
1
981
.
"Taxation and Corporate Investment:
A
Q-Theory Approach."
BPEA 1:1981 67-127.
.
1994. "Psychology and Savings
(May), 186-192.
Thaler, Richard.
Policies."
American Economic Review 84
Temin, Peter. 1976. Did Monetary Forces Cause the Great Depression?
United States.
New
York: Norton.
Task Force on Market Mechanisms. 1988. Report of the
Washington: U.S. Government
Task Force on Market Mechanisms
Presidential
Presidential
.
Printing Office.
United
1994.
States Bureau of Labor Statistics.
Washington: U.S. Department of Labor.
Weicher, John
C, 1995, "Changes
Consumer Expenditures
in
1993
.
the Distribution of Wealth: Increasing Inequality?,"
Federal Reserve Bank of St. Louis Review (January), 5-23.
Wigmore,
1985. The
Barrie A.
in
CRASH
and
Its
Aftermath
.
Westport, CT: Greenwood Press.
1994. "Trends in Household Wealth in the United States, 1963-83 and
1983-89." Review of Income and Wealth 40 (June). 143-174.
Wolff, Edward N.
Wolff, Edward N.
America
Zeldes,
.
Stephen
Top Heavy: A Study of the Increasing
New York: The Twentieth Century Fund Press.
1995.
P.
Investigation."
1989.
"Consumption and
Journal of Political
Liquidity
Inequality of
Constraints:
Economy 97, 305-346.
An
Wealth
in
Empirical
59
1: Real Value of the Stock Market,
Change
Real Value
in
Real
S&P
of S&P 500
Table
1955-1995
%
Year
1955
1956
1957
1958
1959
1960
253.4
251.9
212.6
-15.6
277.1
30.3
1962
1963
1964
1965
1966
1967
1968
1969
1970
25.5
-13.8
16.5
12.1
7.2
-14.3
420.8
449.0
361.8
338.7
361.2
413.9
1972
1973
1974
1975
1976
1977
1978
1979
1980
13.7
6.7
-19.4
-6.4
6.6
14.6
-25.8
-37.0
23.7
12.5
-16.0
307.1
193.5
239.3
269.2
226.2
212.5
210.4
231.5
1981
197.1
1982
1983
1984
1985
1986
1987
1988
1989
1990
213.8
242.9
233.8
283.9
336.8
312.6
343.5
413.8
367.8
421.8
459.6
478.4
455.2
539.4
1991
1992
1993
1994
1995*
Notes: Real
Entries for
8.5
-5.1
370.1
1971
-6.0
-1.0
10.1
-14.9
8.4
13.6
-3.7
21.4
18.6
-7.2
9.9
20.5
-11.1
14.7
9.0
4.1
-4.9
18.5
S&P 500 is computed
1995
relate to June;
Real Value of
Corporate Stock, FOF
-0.6
300.7
285.3
358.0
308.5
359.3
402.8
431.8
1961
500
using average
column 3
for
1,643.5
1,712.1
1,479.9
2,039.2
2,169.8
2,134.7
2,710.7
2,339.0
2,719.3
3,105.6
3,459.7
3,005.0
3,687.8
4,200.4
3,374.7
3,164.8
3,597.0
3,978.3
2,875.5
1,730.4
2,158.4
2,721.8
2,271.9
2,191.9
2,293.7
2,662.2
2,265.5
2,471.2
2,837.1
2,631.3
3,232.3
3,743.8
3,604.9
3,849.9
4,522.7
3,949.7
5,279.8
5,763.2
6,352.0
6,048.8
7,167.8
S&P index for each December and
1995
is
the authors' estimate.
the CPIU.
60
Table 2:
MV
Year
Relative
Shares/GDP
Measures
of Corporate Share Values,
P/E Ratio
9.5
1947
1948
1949
1950
6.6
7.2
7.2
9.7
1951
0.470
0.450
0.628
0.709
0.723
0.625
0.834
0.851
0.832
0.989
0.820
0.908
0.981
1.006
0.838
1.000
1.090
0.868
0.820
0.880
0.899
0.633
0.399
1952
1953
1954
1955
1956
1957
1958
1959
1960
1961
1962
1963
1964
1965
1966
1967
1968
1969
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
1980
11.1
9.9
13.0
12.6
13.7
11.9
19.1
17.7
17.8
22.4
17.2
18.7
18.6
17.8
14.5
18.1
18.0
15.9
18.0
17.9
18.4
12.0
7.7
0.481
11.3
10.8
8.7
7.8
7.3
9.2
8.0
0.578
0.460
0.421
0.456
0.542
0.459
0.504
0.541
0.478
0.570
0.637
0.593
0.614
0.713
0.631
0.839
0.885
0.955
0.877
1.039
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995'
11.1
11.8
10.1
14.5
16.7
14.1
11.7
15.5
15.5
26.2
22.8
21.3
15.0
16.3
1947-1995
P/D Ratio
18.0
16.0
14.8
13.8
16.7
18.3
17.2
22.5
24.1
23.6
21.6
30.0
31.5
29.3
35.1
29.4
32.0
32.8
32.8
27.9
32.4
Tobin's q
0.437
0.396
0.418
0.475
0.510
0.483
0.452
0.656
0.756
0.742
0.593
0.819
0.830
0.808
0.954
0.832
0.942
1.056
1.080
0.897
1.070
34.1
1.111
28.4
28.9
32.3
37.0
27.0
18.4
24.2
25.5
0.846
0.778
0.828
0.835
0.575
0.319
0.390
0.470
19.6
18.5
0.371
18.1
21.1
18.0
20.3
23.2
21.4
25.8
29.6
27.0
27.2
30.0
26.7
32.2
34.5
36.8
34.4
39.2
0.343
0.350
0.413
0.336
0.360
0.417
0.387
0.472
0.556
0.574
0.617
0.761
0.733
1.068
1.264
1.361
1.268
1.467
in the Flow of Funds accounts (year end value)
each year. Column 2 (Column 3) is the P/E ratio (reciprocal of the dividend
price ratio) for the Standard and Poor's Composite Stock Price Index averaged for the last month of each calendar
year. Column 3 is the market value of equity in nonfinancial corporations divided by the replacement cost of their
net assets, net of debt outstanding, from the Balance Sheets for the U.S. Economy. June 1 995 q is estimated by
Notes:
Column
1
is
the market value of corporate stock as estimated
divided by the fourth-quarter
the authors.
GDP
for
61
Table 3:
Year
1947
1948
1949
1950
1951
1952
1953
1954
1955
1956
1957
1958
1959
1960
Rate of Return to Nonfinancial Corporate Capital, 1947-1995
Unadjusted
11.0
12.4
10.8
12.6
12.8
10.9
10.3
9.6
11.9
10.3
9.4
Cycle-Adjusted
11.4
10.9
12.4
11.5
9.4
8.8
9.6
11.2
9.5
8.6
8.6
9.9
8.1
10.1
9.3
9.1
1961
9.3
9.8
1962
1963
1964
1965
1966
1967
1968
1969
1970
10.5
11.3
11.9
10.4
11.2
11.6
12.4
11.9
10.6
10.2
1971
1972
1973
1974
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
13.1
12.9
11.6
11.3
10.0
8.0
8.4
8.6
8.5
8.8
7.6
8.5
8.6
8.0
6.7
6.6
7.0
7.4
7.8
7.8
6.9
5.8
6.0
5.2
6.0
8.5
8.4
8.5
8.0
6.9
6.6
7.0
7.2
8.1
7.4
7.2
6.8
8.3
8.0
7.4
7.5
7.8
8.1
7.9
7.7
7.6
7.4
7.9
7.5
7.5
9.1
10.2
10.3
'
(est.)
7.9
8.8
9.7
10.3
10.2
Averages:
1950-59
1960-69
1970-79
1980-89
10.6
11.1
7.7
6.8
10.0
10.6
8.0
7.6
Source: Tabulations based on data from the National income and Product Accounts and Balance Sheets of the U.S.
Economy. The return is computed as the ratio of pretax profits for the nonfinancial corporate sector, with the capital
consumption adjustment and inventory valuation adjustment, plus net interest payments by NFCs, divided by an
estimate of the mid-year value of the tangible assets held by these corporations. The value for 1 995 is based on
quarters of profits and interest payouts, and an estimate of NFC tangible assets as of March 31, 1995.
two
62
Table 4: Dividends
Ratio of Total
Year
& Other Cash
Cash
Payouts to Cash Di/idends
Payouts, Nonfinancial Corporations, 1980-1995
Price/Divi-
Price/Total
dend Ratio
Cash Payouts
1980
0.993
21.1
21.2
1981
1.488
1.204
1.109
2.285
18.0
20.3
23.2
12.1
1982
1983
1984
1985
1986
1987
1988
1989
1990
2.334
2.409
2.238
2.712
2.312
1.602
1.169
1.080
1.154
1.462
1.206
1991
1992
1993
1994
1995
Column
16.8
20.9
9.4
11.0
12.3
12.0
10.0
13.0
16.7
27.5
31.9
31.9
23.5
32.5
21.4
25.8
29.6
27.0
27.2
30.0
26.7
32.2
34.5
36.8
34.4
39.2
the ratio of dividends plus gross share purchases by nonfinancial
in corporate control
transactions, to dividend payments by NFCs. Column 2 is the price-dividend ratio for the
S&P500; column three equals column 2 divided by column 1 Data on dividend payments for
NFCs are from NIPA Table 1.16; data on gross share purchases are from the Federal Reserve
Board. The value for 1995 is an estimate based on data for the first two quarters; the P/D
and P/Cash ratios are as of June 30, 1995.
Source:
1
is
corporations, including both share repurchases and shares bought
.
63
Table
5:
Unadjusted Percentage of Flow of Funds Corporate Stock Holdings
Held by Major Investor Categories, 1952-1994
Pension
Year
1952
1953
1954
1955
1956
1957
1958
1959
1960
1961
1962
1963
1964
1965
1966
1967
1968
1969
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
Households
Funds
Mutual
Funds
Insurance
Foreign
Companies
89.7%
1.1%
3.1%
2.2%
3.4%
88.6
89.3
88.6
88.6
87.5
87.6
86.8
85.8
85.7
84.7
84.2
1.5
3.5
1.4
3.3
3.3
3.5
3.9
4.0
4.3
4.6
4.6
4.8
4.9
4.9
5.0
5.2
5.3
5.3
5.5
5.2
5.5
2.2
2.2
2.2
2.2
2.2
3.6
3.3
2.1
83.8
83.0
81.7
81.9
2.3
2.8
3.0
3.5
4.0
4.4
4.8
5.2
5.5
5.9
6.4
6.6
6.8
69.1
8.1
68.0
65.9
9.2
10.5
11.5
12.8
15.2
16.5
14.7
16.3
18.5
84.1
64.1
60.4
56.1
56.7
61.8
59.0
56.9
58.7
60.9
59.0
56.1
53.5
51.4
51.3
50.6
49.8
48.8
48.0
48.6
50.8
51.4
49.7
47.7
18.1
17.4
18.7
21.3
22.9
24.6
24.8
25.4
25.5
26.8
27.2
27.0
26.2
25.8
25.6
25.7
5.1
5.1
5.4
4.9
4.1
3.9
3.7
3.4
2.1
2.2
2.2
2.2
2.2
2.2
3.1
3.1
3.2
3.3
3.7
4.3
2.0
1.9
3.1
3.1
3.5
3.8
4.0
4.2
3.7
4.2
4.2
4.1
6.1
6.1
6.9
6.5
7.0
6.3
6.5
6.6
6.3
5.6
5.5
5.5
5.4
4.1
7.1
7.7
8.7
11.5
13.6
3.1
3.0
2.1
4.6
5.0
2.9
3.3
2.9
3.0
2.8
2.8
2.9
2.9
3.0
3.0
2.9
3.0
2.8
2.8
4.2
4.5
4.7
5.0
5.2
5.3
3.1
3.1
'
5.1
5.7
5.2
4.8
5.2
5.5
5.4
5.1
5.5
5.7
5.7
5.7
5.5
5.0
5.2
5.2
5.0
5.0
4.4
4.0
4.0
4.2
Source: Calculations based on Flow of Funds Accounts, July 1 995. Pensions include private
and government plans. Mutual funds include closed end as well as open end investment
companies. Households includes ownership by nonprofit institutions, insurance companies
includes both property/casualty and life insurance companies.
64
Table
6:
Adjusted Flow of Funds Data on Stock Ownership:
and Beneficial Ownership, 1952-1994
Individual Direct
Bank
FOF
Year Households
1952
1953
1954
1955
1956
1957
1958
1959
1960
1961
1962
1963
1964
1965
1966
1967
1968
1969
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
89.7
88.6
89.3
88.6
88.6
87.5
87.6
86.8
85.8
85.7
84.7
84.2
Non-
Personal
profits
Trusts
-14.1
68.0
65.9
-13.9
-14.0
-13.9
-13.9
-13.7
-13.8
-13.6
-13.5
-13.4
-13.3
-13.2
-13.2
-13.2
-13.0
-12.8
-12.9
-10.9
-10.7
-10.3
64.1
-10.1
60.4
-9.5
56.1
-8.8
56.7
61.8
59.0
56.9
58.7
60.9
59.0
-8.9
84.1
83.8
83.0
81.7
81.9
69.1
-9.7
-9.3
-8.9
-9.2
-9.6
-9.3
56.1
-8.8
53.5
51.4
51.3
50.6
49.8
48.8
48.0
48.6
50.8
51.4
49.7
47.7
-8.4
-8.1
-8.1
-7.9
-7.8
-7.7
-7.5
-7.6
-8.0
-8.1
-7.8
-7.5
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
10.5
10.4
10.7
11.0
12.0
12.7
11.5
10.3
10.6
10.4
9.6
8.8
8.8
8.3
7.9
7.9
7.3
5.9
5.6
5.5
5.4
b.4
4.8
4.0
2.9
2.7
'IndivVariable
Mutual
DC
Funds Pensions Annuities iduals"
0.3
0.4
0.4
0.6
0.6
0.8
0.8
3.1
3.5
3.3
3.3
3.5
3.9
4.0
4.3
1.0
4.6
4.6
4.7
4.8
4.8
4.8
5.0
5.2
1.1
1.2
1.3
1.4
1.5
1.6
1.7
1.7
5.1
1.7
5.3
5.0
5.2
4.9
4.7
4.9
4.4
3.6
3.4
3.3
3.0
2.6
2.6
2.8
3.6
4.0
4.5
5.5
2.0
2.3
2.5
2.7
3.0
3.5
3.8
3.4
3.7
4.3
4.2
6.1
7.1
5.7
7.5
7.7
4.1
4.3
5.0
5.2
6.2
6.2
6.4
6.1
6.0
6.5
7.2
9.3
11.0
Notes: Authors' calculations as described
7.1
7.1
7.5
7.8
7.7
in
the text.
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.1
0.1
0.1
0.1
0.2
0.2
0.2
0.2
0.3
0.3
0.3
0.4
0.4
0.5
0.7
0.7
0.8
0.9
0.8
0.9
1.2
2.0
79.0
78.6
79.0
78.6
78.8
78.4
78.7
78.4
78.0
78.0
77.4
77.2
77.1
77.1
76.7
75.8
75.8
76.1
75.0
74.0
72.6
70.8
68.4
67.7
69.6
67.7
66.2
66.5
67.0
65.8
63.7
62.2
61.7
61.6
61.0
61.5
60.6
60.5
60.5
61.9
63.0
63.2
63.7
65
Table
7:
Number
of Stockowners, 1962, 1983, and
Millions of
Stock Category
1962
Households
1983
1992
1992
Percentage of Households
1962
1983
1992
Households With Any Stock Holdings
Publicly
10.0
16.0
17.0
17.2
19.1
17.8
11.1
16.9
21.1
19.0
20.1
22.0
19.7
26.8
23.5
28.0
Plan
23.2
31.8
27.7
33.2
Plan
27.9
35.7
33.2
37.4
Traded
Plus Mutual
Fund
Plus IRA/Keogh
Plus
401
Plus All
(k)
DC
Households With Stock Holdings
Publicly
Traded
Plus Mutual Fund
> $2,000
7.3
11.5
12.4
12.6
13.7
12.9
8.3
12.3
16.3
14.3
14.6
17.0
Plus
IRA/Keogh
14.4
21.5
17.1
22.5
Plus
401
16.8
24.8
20.0
25.9
20.7
28.0
24.6
29.3
83.9
95.6
Plus All
Total
(k)
DC
Plan
Plan
57.9
Source: Authors' tabulations from 1962, 1983, and 1992 Surveys of Consumer
Finances. Further description may be found in the text.
CO
o
-1-
in
O
CN
CO
CD
CN
CN
CN
CO
CO
O
CO
CO
CN
CO
CM
r»
in
in
oo
co
d in
—
^
in
en
«-
«-
in
cm
^t
CM
CN
cm
X
CD
eu
X
o •
§ **
CJ
" co c
5
TJ
C
3
_
0_
CO
o
£*o
CO
LO
en
cb
o
in
**'
CN
CO
co
«r-
CM
CN
O
CO
CO
6
O
O
J2
»
13
co
CM
CO
co
C
D
O
CD
*-
tJ
"
If
*O
X
(0
iC
o o •
*
CD
a.
in
00
p^
en
CO
o
CO
CM
CM
CO
CO
CO
ao
en
"SO
CM
CO
*
in
CO
>
o O
CO
CO
o
o
ro
co
"tf
00
CO
^t
CM
i-
E
c
o
CO
P^
CO
CO
p^
in
en
O
CM
u
CM
CD
O)
o
<
3
TJ
Sf
e
Q.
C
go
en
CO
c-g
>
2
> O* yj
CO
to
CO
o
en
CM
co
CO
CO
CM
*
O
CN
O
m o
o
O en
co
CO
CO
CN
co
(0
CD
TJ
in
CM
CO
3
CO
00
(7)
TJ
c o "O
o ®
• **
c
o
C
to
on
<
o
^ a
3 o
c
Sep
o
o
o
J*
"2
<
S£
en
co
co
CM
CN
CO
CO
CM
c o "
O CD
CD
CD
o.
^
o
^ m m
CM
CM
CO
>
CO
CM
r»
«•o o
o
CO
en
~ in
CM
CO
00
en
co
•-
CN
"SO
o
CM
00
en
en
eo
«-
go
cj
XI
CD
3
X
'C
CO
CO
CM
CO
p^
«-
cm
en
in
oo
»-
cm
en
o io
o
CN
CM
c-g
5 o
€
>O
r/5
CM
lO
CO
0.
Q.
o
en
co
o
in
p".
CM
co
co
CO
in
CN
i-
p^
CM
co
co
en
i-
en
00
CO
00
00
in
CM
CM
in
en
CO
CO
cm
en
eo
CM
p^
o
CM
00
CM
pv
en
o
CO
CM
,
!
CD
00
to
E
3
ca
C
o
u
>
CO
CN
CO
_£
XI
c
c
to
to
c
g
'^
3
Q
O
CD
lO
o
o
CM
(A
CD
CD
CO
**
CO
en
en
CN
en
XI
CO
CO
co
* ~
tj
c
3
CD
m
CO
CJ
J*
o
o
t£
*O £C
"so
in
O
d
00
CM
•tf
o o
£ »
CO
CN
in
CN
co
00
CO
®
CO
CO
CO
in
CM
CD
>
3
(O
CM
en
en
TJ
CM
cn
C
CO
CO
00
en
C/)
tj
CD
TJ
n
k.
I-
CO™
O
CD
CD
CD
|
>
e£
o-
_>•
CO
CO
"SO
o in
o
m CO
CM
in
CO
in
m
in
eo
oo
en
^
oo
CM
t-
in
eo
00
«-
o
CO
CM
CM
CO
o
x3
a.
E
o
c
o
C O C *
co 2 ; o
" > * 2
• > ° co
Q.
*
in
CM
a a
< 2cs
O
00
CO
CM
CM
CO
O
o
CM
CM
CM
CM
CO
in
w
CD
tj
c
Z)
i-'
o
CO
CO
00
m
CO
in
CO
in
in
in
eo
i
m
CM
CD
in
CO
>
O
in
en
en
in
CM
a>
TJ
C
3
co
JJ
r-
00
CO
O
CM
CM
CM
co
CM
o
^1-
CO
'^
JO
en
»CM
in
CO
in
eo
in
m o>
CO
xo
3
<
w>
I
CD
in
CM
X3
in
co
in
o
u.
3
O
cn
'
CO
C
o
m
00
CN
cn
CN
CO
in
m o
oi
cn
<*
in
CO
co
oo
en
en
en
CN
CN
O
CN
CN
«-
CN
oo
en
cn
CN
en
i-
co
CN
cn
£
i
a>
01
O
CM
CD
>
"^ .2
co
=
7\
*•*
LU
fe
CO
o
iT
CN
o
CO
CO
in
00
in
CO
CO
CN
*"
CM
m
en
CO
o
6
en
in
en
*
\S
\-o
CN
a> ic
CN
O
I
o
is
l>
CM
cn
cn
ICD
— >
c
TJ
r»
C
^
to
LL
en
C
00
en
m
o
CO
*t
en
CO
CO
in
en
^
in
en
CN
CO
CO
t•
*
O
^
en
^
«in
CN
en
*—
^
o
co
® S
°.
^ S
o
^ 2
CO
ci
en
in
en
in
in
IE
i
c
^Lo
en i^
in 1.9-
ic
IU
CO
00
1(0
en
I®
CN
CO
cn
CO
<*
3 5
1
d
1?
o
en
in
in
CN
<*
o
in
in
o
in
en
CO
CN
CN
co
CN
^
CN
O
en
en
en
cn
CN
co
00
en
eo
cn
en
£
!£»
co
co
en
CN
»-
It
uc
nr
i««
a>
«o
lo
ic
<
ICO
«o
co
CO
3
.2
*c
CO
CN
>
C
o
CO
O
**^
>
^
.2
«o
UJ
J=
en
(0
rr
CD
O
c
o
CN
O)
c
a>
u
c
o
en
en
**'
CN
i^
en
co
CN
CN
r«i
co
°°.
CN
CM
CN
*~
CO
CN
i-
en
«-
CO
CN
en
cn
CN
co
o
in
CN
in
in
*
o
cn
in
en
en
CN
cn
en
i-
*-
CN
ci
CN
1
13
CN
CO
en
en
CO
cn
CO
en
s!
lu
q q
00
CN
r^
ICO
IS
lo
CN
CO
IS
^
«-
1ST
ICO
\i
>-T3
*t
U
=
T3
XI
£1
(0
u
o o
CN
CN
en
en
05
CO
en
co
co
in
00
CO
m
t«-
00
CN
CO
*'
CN
o
o
6 6
00
00
«cn
to
to
r^
ao
cn
—
«-
f*
«-
o o
o
" ^ 6
oo
in
cn
r*
«n
CN
oo
^n
to
i*-
o
d
CN
in
&
OI
£
o
r^
oil
en
icn
c
5
itn
O
.2*
'5
«
cn
»
!
Iff
en
en
CN
in
en
o
00
oo
CN
co
cn
o
t-
en
o
cn
cn
co
6
«-
00
CN
en
in
d
CN
00
O
00
CO
CN
CN
CO
in
en
CN
cn
CO
^'
00
q o
CN
6
cn
r-
en
<*
«-;
en
in
oi
*-
CN
CN
CN
o
en
co
^-
«-
o
o
6
IE
IE
IO
1.2
co
_®
'£»
c
a>
o
cp
a.
cn
en
cn
CN
cn
cn
2 ^ -
m in o O
in
6 6
2
O
m
«
CP
s
®
z Z
CO
o o
in
o
rin
CO
in
E
§
o
CD
co
CM
CO
in
CN
CO
*
CN
o
t-
X
CP
4-*
X
CP
CP
S
CP
s
i-^
m q q q o
d oo
in
*t
d 6
a.
q
E
o
V
o
z Z Z Z m
m o q
in
d d ^ in
a.
o
l-
Er*
o
x
cp
!ip
cp
ti
i
co
!
CO
r^
CM
CD
CM
00
CO
CO
in
in
o
en
in
in
en
q
*^
CD
CM
cn
o
CM
CM
CO
i
CO
co
CO
i
CM
00
CO
CO
CO
q
in
CD
i
i
O
j
x
CO
1*^
a o »
"O
i
u.
_
iS
o.
!
!
°-
O
d
CO
CO
en
<*
CD
CO
°°.
too I
CM
o
cn
CO
CD
CM
o
o
d
CM
en
CO
CM
CM
CM
CM
cn
00
*"
CM
in
CO
CM
CM
CO
q
CO
CO
en
CM
CO
CM
00
CO
CO
CN
00
CM
r-'
en
00
CM
CDJ.E
2i§
3o
o
w
(N
If
05
73
co
<
3
Q-
!
r>-'
"SO
go
£?
<D
a.
en
g 8 •
0)
>
co
c-g
1
* 2
>
>Om
i
jm
o
°
!
Q.
>
^
CM
CO
CO
CO
CO
CO
CM
cn
in
in
CO
CO
in
CO
CN
*t
**
t
CD
en
00
O
r>-
CO
CO
in
CM
CM
en
CM
in
r*
en
r^
in
CO
CO
r^
CO
CM
00
CM
CM
in
cn
cn
CO
q q
r*"
00
CM
It
eni|
is.
d
it
lO
he
IO
ha
in
CN
O
^~
r^
'-
in r°
r^
OJ
CO
in
r^
«- CO
<n ^c
m
13
E
1
X>
o
o
c
c
• O
re
1
3
E
!
re
u.
1
>i
J3
00
CD
3-=
3 O
!
i
Sn £
C/5
3
c
o
3
CO
in
"~
il
CM
CM
o
CO
CO
d
in
CM
CO
CM
CO
CO
en
Q. "5
cn
CO
O
I
CM
i
p>;
a.
0.
in
(N
CD
CO
r^
C r. c -*
co
2 5 O
" > * 2
>
O yj
9
o
in
-N
CM
en
CM
**
CD
q r'o
d g Ire
d d d CM
* IC
P
IE
o ®
CO
c
CO
cn
!|
co
o
u
o
^
*
<S
Dc
cn
CM
CD
CD
CO
*~
CO
CO
00
CO
CO
d
CM
CM
CM
in
"~
CM
CO
en
CO
CM
*~
CD
CM
<—
CO
**
en
CM
CM
CO
CM
O
m
CO
in
CO
in
in
CD
in
r^
in
O
CD
CD
CD
CO
t
00
<*
r*
m
CO
s
d ?
-!c5
CO
CO
CO
CO
CO
CM
eri
cn
CO
CO
CO
CN
00
r^
^
i—
**
00
in
CO
CO
>
in
t
* 1
>
icn
P
5
So
82
Q
6
u
o
el
r"2
'
§
"SO
CO
i
in
!
**
CO
CO
CM
CO
CM
cn
CO
in
CO
CO
CO
1^
d
CM
^
in
00
in
CM
00
en
CM
o
CM
CO
CM
en
CO
en
CN
in
CO
*
CM
i
re
°!co
*>
is
icn
»*-
00
-O
(0
;
iS
Eo 2
o 82?
a>
i
in
(0
|n!
a-
o
o
"So
in
en
CM
CO
m
O
CM
en
CM
en
CO
CO
m O
t'
in
CO
CO
in
CO
CO
CM
<6
CO
CO
CO
CO
CM
CO
q
ai
9|E
«"
CM 1°
jnt
l«0
Li
3
a.
g
y
oc-g
€ > o
00
0-
!
CO
°
CO
cn
u>
ai
o
i
r*.
in
**
CO
00
"~
r>.
CM
O
O
O
cn
in
in
«—
CO
o
t t CO
CO
00
o
CM
CO
o
O
in
00
*t
CO
"~
cn
"~
CM
CO
CO
*t
CO
cn
r^
m
co
'tZ
'iS
m
ns
CO fa,
w
O
a,
2"
E 2
o g>
c ra
—U
I
i
i
':
^ w o m o o
o
in
in
o CM
rv
CM
in
CM
DC m in o in o >
CM
in
o o
3
k.
co
I
I
1
r*.
JZ
in
\m
CO
XI
c
in
CM
1
o
in
i
in
r*
o °|3
o
o in
CM
CM Pt
I
m CM
m o
m
in
o
o o
1
3,
;1
CM
5|
IT)
CD
O
IX)
CM
O
O
00
CO
o
in
o
CO
r^
CO
o o
CM
CO
*~
CM
CM
CO
CO
io
CD
CO
-0
co
lo?
u
=
_
CM
CD
re
0_
*"
o
o CM
o t
oo
c
£•50
CO
CD
*•
CO
**
CD
CO
00
en
CM
o
CD
CO
CM
^-
IO
to
CM
CO
oo
«-'
io
CD
.
O
«-
<D
CO
«-
«O
io
o
o
d
CM
O)
i-
CD
i-
CO
CO
00
cr>
CM
CO
CD
CM
IO
CM
oo
CO
o
CD
CO 'C
IO 3
«-
no
5o
co
"D
C
as
TJ
co
oo
en
"
If
O
<
o
O)
»
is*
re
0.
3
O
CO
CM
g 8 |
Q-
o
o
o
CO
**
CO
CO
CM
* CO .5 io o CO
^
N
d
° "
*CM
CM
«*-
00
CD
«-;
CO
CO
in
co
*-
«-
o
O
">
O)
LO
re
E
c
o
r*5
c
en
p|l|
>O
®
*
en
oo
CM
O
**
co
a.
00
LO
00
IO
CM
IO
en
00
CO
CO
CM
CO
IO
CO
en
en
CM
IO
o
00
00
CM
CO
IO
CO
r^
_ CD n
5 *
CO
IO
.
en
«-
io
CO
»CM
o
CO
CM
en
*-
r-
IO
— £
CO
CO
d IO
00
00
CD
CO
CO
CO
en
\%
u
05
9
o
o
CO
CO
CD ro
1^
IO
CD
CO jx
IO
«
CO
CO
<
c
So®
re
"re
'o
c
re
c
00
o
d
TJ
c
3
;
O)
CO
o
**
CO
OS
3 O
,
!
I
e
o
"8
CM
oo
oo
CO
CO
CM
*-
IO
CD
©
CO
CM
CM
o
CM
CD
CD
§=
^
<->
*t
IO
C O
CO
CO
CO
IO
OO
CD
CD
CO
IO
CO
CO
CD
O
d
>-
(DO®
S«
|
JE
< Q-® —O O
o
O
O
IO
CD
CD
IO
J-
m
CO
r»
oo
tt
oo
r>»
CM
CO
IO
IO
O
CM
o
LO
*-
CO
1^
II
e|!
CO
•-
gSE
>
re
i*
a>
re
c
§
o
-O
re
3
ft
2 5 o
£O
o
IO
CD
CM
IO
CM
CO
CM
(73
CD
CO
IO
en
CO
O
io
CD
t
CO
CD
CO
o
en
co
CM
CO
CD
CO
CO
CM
CM
CD
i-
O
O
CO
CM
CO
CO
CO
CO
CD
IO
a.
iCD
u
o
(/)
So*
c
o
CM
CO
**
c
o
3
.*
U
O
CO
T3
(0
a>
o
b
5
-O
re
re
t
£•50
IO
CO
*fr
00
o
*
CO
CO
00
00
CD
co
io
io
CM
r^
^
,
co
00
o
o O
O i-
CO
«-~
CM
CO
CO
CM
CD
CO
iCD
CO
CO
d
^
oolyj
ire
CD
** te
|00
p
o 2
E
9 o e
£5) I
d
CO
00
IO
"*
CM
CO
Q£
CO
CO
CO
CD
CO
^
K '
00
I®
IOt
tCM
00
CM
en
CD'
V
CO*
»
d
CO
CD
^*
CD
CO
o CO
o o
CM
00
CD
LO
o
CO
CM
cd!^
"!|
^ CO
io
1*=
in
o
Li
i
3
a.
c o e -x
• 2 ; o
IO
o
CM
CD
f»
o
CO
CO
IO
00
*
r*.
^
*""
T—
^
>
,
CM
o
*•
CM
CO
IO
Tf
CO
t
00
«~
en
co
IO
IO
CM
CM
r«
o o
£ CM
CO
IO
CO
CM
LO
00
CO
LO
en
o
io
§1
en
co
|re
J**
r»
<
o
o
re
O
IO
o
1
N o
CM
i
IO
o
o o
o
O
IO
o IO
IO
CM
CM
•
IO
CM
1
k.
o >
o
o o
IO
re
LO
N
LO
O
o
io
IO
CM
LO
CM
O
«-
IO
CM
o
O
o
LO
II
Si!
cm r*5 S
°?
!co
70
Table 12: Aggregate Consumption Spending and Stock Market Fluctuations
Consumption
Concept
Total
Total (Four Quarters)
Nondurables
Nondurables (Four
Quarters)
Services
Services (Four Quarters)
Durables
Durables (Four Quarters)
Lagged Change
Constant
.0037
.031
(.0011)
(.008)
.0036
.064
(.0011)
(.014)
.0021
.032
(.0011)
(.008)
.0023
.054
(.0011)
(.015)
.0055
.007
(.0008)
(.006)
.0056
.025
(.0008)
(.010)
.0017
.139
(.0057)
(.041)
-.0011
(.0057)
in
Real
InlStock Price)
.290
.068
.101
.065
.070
.009
.015
.058
.091
(.076)
Notes: Each row reports estimates of the coefficient a u or the sum of the coefficients a from an
equation of the form
Aln c, = a + a(L)*Aln P,., + e,.
All equations are estimated from 1947:3 to 1995:2. Standard errors are shown in parentheses. All
equations include quarterly indicator variables.
it
71
Table 13:
Luxury and "Anti-Luxury" Consumption
Spending Ratio
CEX Share of
$70K+ Households
Consumption
Subcategory
All
for
70K + Households/
20-30K Households
NIPA
Consumption
Share
Consumption
.236
1.85
1.000
.314
3.10
0.023
.308
3.24
0.023
.369
3.69
0.003
.337
3.44
0.004
.296
2.76
0.003
Rented Dwellings
.045
0.25
0.037
Tobacco Products
.099
0.54
0.008
Expenditure
"Luxuries":
New
Automobiles
Education
"Other Lodging"
(Hotels and Motels)
Entertainment
-
Fees
and Admissions
Household Operations
Personal Services
-
"Anti-Luxuries":
in the first two columns are based on data from the 1991-1993 Consumer
Expenditure Surveys, as reported in U.S. Bureau of Labor Statistics (1994). Households are
allocated to income groups based on income before taxes. Entries in the last column are from the
National Income and Product Accounts data for 1994. "Education" is the NIPA entry for "private
education and research," and "Entertainment-Fees and Admissions" corresponds to the sum of
NIPA categories "motion picture admissions," "legitimate theatre," and "spectator sports."
Notes: Tabulations
72
Table 14:
Luxury Consumption and Stock Price Fluctuations
One Quarter Lagged Change
Consumption
Concept
in
Four Quarter
Changes
ln(Stock Price)
in
Sum
of
Lagged
InfStock Price)
"Luxuries":
New
0.250
0.627
(0.103)
(0.193)
-0.022
-0.041
(0.012)
(0.024)
Automobiles
Education
-0.007
0.011
(0.041)
(0.082)
-0.086
-0.128
and Admissions
(0.056)
(0.109)
Household Operations
-0.000
0.019
Personal Services
(0.033)
(0.065)
-0.026
-0.051
(0.007)
(0.015)
-0.018
-0.071
(0.024)
(0.046)
"Other Lodging"
(Hotels and Motels)
Entertainment
-
Fees
"Anti-Luxuries":
Rented Dwellings
Tobacco Products
Notes: All equations are estimated from 1959:1 to 1995:2, with the exception of the equation for
new automobiles, which is estimated from 1947:2 to 1995:2. The estimating equation is
log(L,/C,)
where
L,
+ o,*log(Lt 1 /C l
.
.1 )
+ a2 *AlnP,., +
e,
denotes luxury (or anti-luxury) consumption and C, denotes aggregate consumption
Standard errors are shown in parentheses. All equations include seasonal dummy
spending.
variables.
= a
73
Table 15: Correlation of Consumption Growth
Any Stock
Growth Rate
& Stock
Returns, Stockholders vs. Non-Stockholders
Stock > $1,000
of Per Capita
Stock > $10,000
Food Consumption
.120
.076
.078
Without Thrift Plan
.058
.077
.063
With
Thrift Plan
.350
.049
.226
Without Pension Plan
.059
.071
.044
With Pension Plan
.143
.047
.080
.125
.214
.286
Without Thrift Plan
.190
.133
.226
With
Thrift Plan
.043
.261
.314
Without Pension Plan
.148
.134
.339
With Pension Plan
.089
.205
.232
Non-Stockholders
Stockholders
Growth Rate
of Per Capita
Sk inner Consumption Index
-.125
-.146
-.126
-.208
-.178
-.182
.218
-.010
.048
Without Pension Plan
-.217
-.191
-.202
With Pension Plan
-.004
-.090
-.052
.011
.068
.219
.081
.042
.269
Thrift Plan
-.095
.094
.097
Without Pension Plan
.055
.026
.226
-.012
.092
.187
Non-Stockholders
Without Thrift Plan
With Thrift Plan
Stockholders
Without Thrift Plan
With
With Pension Plan
Each entry reports the correlation of a measure of consumption growth with excess return
on the stock market for PSID households in each category, as described in the text. The sample
period is 1970-1992 excluding 1973, 1974, 1988, 1989, and 1990. Stockholders versus nonstockholders are defined based on the criteria at the column head. The correlation of excess stock
market returns with food (Skinner index) consumption growth for all households is .129 (-.070).
The standard error of each correlation may be computed as [(1-p 2 )/16] B where 16 denotes the
degrees of freedom; these standard errors are approximately .25 for each entry above.
Notes:
,
MIT LIBRARIES
3 9080 00931 0084
74
Table
1
6:
Stock Ownership
&
the Stock Price Fluctuations-Consumption Linkage
Lagged Change
Consumption
Constant
Category
Real ln(Stock Price)
Panel A:
Total
Ownership Share*
Lagged Change in
in
Real ln(Stock Price)
No Ownership Measure
.0047
.030
(.0005)
(.007)
.098
Panel B: Direct Individual Stock Ownership Measure
Total
Total
.0047
-.003
.060
(.0005)
(.031)
(.055)
.0047
.056
(.0005)
(.012)
.098
.104
Panel C: Expanded Individual Stock Ownership Measure
Total
Total
Notes:
-.064
.136
(.069)
(.099)
.0047
.045
(.0005)
(.010)
.102
.103
Each row corresponds to an estimate of an equation of the form:
Aln
All
.0047
(.0005)
c,
= a + o,*Aln
estimates correspond to
t..e
SHARE1 and SHARE2, which
P,.,
+ ^•SHARE1,.
,
1
Aln
P,.,
+
c
.
period 1947:2-1995:2,
replaces
SHARE1
in
193 quarterly observations. The values of
the foregoing equation, are defined based on
Flow of Funds data for the period 1952:4-1994:4. They are extrapolated
Standard errors are shown in parentheses.
of the sample.
t
at the beginning
and end
75
Table 17: Consumption Spending, Stock Market Fluctuations, Dividends, and Earnings
Constant
Total
Total
Nondurables
Nondurables
Services
Services
Durables
Durables
Notes:
Lagged
Lagged
Consumption
Concept
.0045
.030
.002
(.0005)
(.007)
(.021)
c,
Lagged
.071
.0044
.032
.015
(.007)
(.010)
.0022
.032
.013
(.0006)
(.008)
(.021)
.0021
.033
.010
(.008)
(.011)
.0059
.008
.001
(.006)
(.016)
.0058
.008
.005
(.006)
(.008)
.0070
.137
-.008
(.041)
(.113)
.0063
.146
.100
(.041)
(.056)
= a
+
a, * Aln
P,.,
+
y,
* Aln D,.,
.001
.046
(.0029)
in
.076
-.001
(.0004)
(.0030)
.081
.074
(.0006)
(.0004)
R2
Aln (Earnings)
(.0005)
Each row reports estimates of the coefficients
Aln
Aln (Dividends)
AlnlStock Price)
.062
the equation
+
e,
same equation with E,., replacing D,.,. All estimates correspond to the period 1947:31995:2, 192 quarterly obsarvations. Dividends denotes the real value of dividend payments to
shares in the S&P 500 Index, Earnings the analogous measure of earnings for these firms.
Standard errors are shown in parentheses.
or the
Date Due
NOV 1 5
OCT
2
199£
1
2003
Lib-26-67
MIT LIBRARIES
I
I
II
3 9080 02165 799'