View original document

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



April 29, 1988

October Postmortem
The causes and effects of the October 19, 1987
decline in world equity prices will be debated
for many years. Already, a study by a presidential commission and reports by various agencies
and industry groups have been completed, and a
multitude of academicstudies are underway. To
a large extent, the considerable interest in the
1987 crash reflects the remarkable similarity
between the pattern of price movements in 1929
and 1987 and the worry that subsequent events
will be similarly severe.
This Letter reviews some of the explanations for
the October crash. No single change in market
fundamentals or flaw in the operation or regulation of financial markets explains the events of
last October. This may favor the view that the
October crash was the bursting of a speculative
bubble. Proponents of this explanation argue
that asset markets can and do produce
"rational" price bubbles and breaks.

The events
The price break of October 1987, followed an
historically rapid increase in U.s. and world
equity prices. From january 1986 until its peak
in August 1987, the Dow jones Industrial Average (DjIA) increased about 200 percent. On
October 1, 1986, the index stood at 1782 and
barely a year later at 2600. The price-earnings
ratio for the S&P 500 averaged 22 in October,
compared to the 30-year average of 14.
Equity prices on most of the world's exchanges
moved upward during this period as well. The
most dramatic increases were on the Tokyo and
Singapore exchanges, where share prices
increased 220 and 250 percent, respectively,
from the beginning of 1986 until the peak in
1987. Others were only slightly less dramatic: in
the same period, London's index increased by
171 percent; Toronto's by 148 percent; and
Switzerland's Swiss Bank Corp index by 136
percent. None of the world's major exchanges
experienced a downturn During this period.
The price break in the United States began on
Tuesday, October 12, with the DjlA declining by
ten and one half percent by the end of the week.

Individual day declines were large, but not record-breaking. Following the weekend of October 17-18, the U.S. stock indexes declinE;d
sharply, with the DjlA falling by 23 percentto
1738 on Monday, October 19. Sharp declines
followed on all other world exchanges, with the
Australia and Singapore exchanges falling most
abruptly (by nearly 50 percent). The high-flying
japanese market fell modestly in comparison,
declining less than 15 percent.
Chart 1 shows the magnitude of the price break
in the DjIA, as well as the bumpy recovery of
share prices since then. The DjlA now stands at
around 2050.

Causes of the crash
The abrupt decline in equity prices has been
attributed to numerous "fundamental" economic and "technical" factors. Fundamental
factors are those that could be expected to alter
the present discounted value of the earnings
prospects of corporations and, hence, the value
of their shares of equity. Such fundamental factors might include rising uncertainty about the
outlook for the dollar and the trade balance,
uncertainty about whether the Federal budget
deficit would be resolved, and legislative attacks
on leveraged takeovers.
All of these factors do, of course, have the
potential to influence U.S. corporate share
values. A sluggish recovery in net exports bodes
ill for future corporate earnings and a sharply
weakening dollar raises the possibility that u.s.
interest rates would rise. Continued failure to
resolve the federal budget deficit raises the prospect of future tax increases and slower economic growth. Finally, the prospect of adverse
tax treatment of leveraged takeovers proposed
by Congress in early October could make firms
that were takeover candidates less attractive.
It is difficult to square the abrupt movements in
the stock market indexes with these fundamental
factors, however. All of these considerations
had been elements in the behavior of the market
for several years. It seems unlikely that these factors could have been reassessed within a space

of a few days in October. Moreover, these fundamental forces were advantageous to some
overseas economies and thus the sharp declines
in domestic equity markets elsewhere in the
world should not have been ubiquitous.

Technical factors
A number of technical factors also have been
examined for their role in the October 19
decline. Technical factors are those that affect
the price-setti ng processes of the market, but not
necessarily the assessment of the underlying
value of individual firms. One such factor frequently mentioned is the behavior of portfolio
insurers. Portfolio insurers provide institutional
investors with "insurance" against the value of
their portfolio dipping below a specified minimum. The insurers provide this protection by
determining an appropriate balance between
stock and cash in the portfolio. Carrying out
portfolio insurance strategies generally involves
selling equity futures into falling markets.
As the market began to fall in October, portfolio
insurers sold equity futures heavily, as their strategy dictated. The Presidential Task Force on
Market Mechanisms (the "Brady Commission")
argued that this behavior of portfolio insurers
was a key factor in the market's decline, pointing to the close coincidence in the timing of
insurer sales and the intraday downticks in the
market. Another practice, index arbitrage, which
exploits pricing disparities by entering into
opposite transactions in the future and cash markets, then transmitted the downward pressure on
stock futures prices to the stock market itself.
Portfolio insurers, however, argue that their strategy is no different in concept than the stop-loss
orders that have been employed routinely in the
securities markets for a century. (A stop-loss
order is one that instructs the broker to sell the
specified shares whenever their price falls below
a specified minimum.) In addition, they argue
that their sales - roughly $10 billion during the
market decline - were trivial compared to the
total value of shares outstanding (roughly $3 trillion). Moreover, it is unlikely that these sales
made the decline worse than it otherwise might
have been since other investors would have
been induced to buy if.such sales had been
"irrationally" mechanical.

Another frequently cited technical factor was the
failure of the order and market-making mechanisms at the stock exchanges to keep pace with
transactions volume. Economist Larry Harris has
pointed out that the congestion in the cash market meant that index arbitrageurs could not be
certain that they could sell stock (while buying
the futures). This both reduced the buyers in the
futures markets and increased the sellers, as
traders moved activity from the cash to the
futures market. The futures market declined in
an exaggerated fashion as a consequence, as
reflected in spreads of as much as 20 percent
between cash and futures prices. In addition,
because the futures market is a bellwether of the
cash market, "further downward pressure was
then brought to bear on the stock (cash) markets

Speculative bubbles
The concept of speculative bubbles offers a
third, though controversial, way of thinking
about the causes of the October crash. A speculative bubble is a movement in the price of an
asset that is unjustified by changes in fundamental or technicalfactors. Some economists argue
that speculative bubbles can exist in markets
that are otherwise "efficient." That is, speculative bubbles can exist even if individuals'
behavior and expectations about the future are
rational and markets clear without arbitrage
A bubble may form, it is argued, when individuals find it difficult to evaluate fundamental factors. Thus, they may hold the asset at a high
price simply because it has enjoyed substantial
capital gains recently. Successively higher asset
prices yield higher capital gains expectations
and higher prices. Market participants may suspect that such a pricing process has some risk of
crashing. However, unless it is certain that the
price will crash in anyone period, it may be
rational for individuals to hold the asset, and for
the price to rise until it crashes.
Such "capital gains bubbles" would seem less
likely to occur in assets with easy to define fundamentals. The bond market, for example, will
not exhibit speculative capital gains bubbles if
investors are rational. This is because a bond's
value at maturity is fixed and known. It would

Chart 1
The DOW Jones Industrial Average:
400 1929 and 1987



Chart 2
Leptokurtosis in the
DOW Jones Industrial Average













Daily Growth Rate of DJIA




150 --'----

175 Trading Days



120 Trading Days






....~----- 10/28/29 ----l.~

be inconsistent for investors to hold bonds in the
expectation that the.price could rise without
constraint, as is implied by the "capital gains
bubble" process. (This is not to say, of course,
that bond prices cannot change abruptly if "fundamental" interest rate or default risk expectations were to change.)

Was October 19th a bubble?
The fundamental value of corporate equities is
much more difficult to define and - because
they are non-par value, perpetual instrumentsstocks are more susceptible to the processes that
start capital gains bubbles. To identify possible
"bubbles" in asset price data, economists use
two simple types of tests. One test compares the
volatility of actual asset prices with the volatility
that would be expected from a "fundamental'!
model of these asset prices. Robert Shiller and
other economists have argued that when actual
volatility is "too high" based on these models,
these asset prices must exhibit non-fundamental
influences. An obvious weakness of such a test
is that it is only as good as the underlying
model; if the wrong "fundamental" model is
used, the inferences drawn from price volatility
will be incorrect.
A second type of test simply looks at the statistical behavior of asset prices independent of an
underlying model. Some analysts argue that
price movements in bubble-prone assets are
characterized by long periods of stable rates of
change punctuated by periods of very large positive or negative rates of change. Such a pattern
in data is called "Ieptokurtosis."

Economists by no means agree on the usefulness
of these tests to identify bubbles. They have
been used, however, to argue that bubbles can
exist in stock prices. Specifically, fundamental
models of stock prices that use volatility in dividend flows to infer the "proper" volatility in
stock prices fail to explain the high volatility of
stock prices. Additionally, as Chart 2 shows,
share price changes do tend to be distributed in
a leptokurtic fashion.

The possibility that the stock price break of
October 19, 1987, represented a bursting speculative bubble makes the events of that period
much easier to interpret, and, and in many
ways, much less ominous. Capital gains bubbles
burst when the probability of their persisting
declines; the severity of the October crash then
could be reconciled with modest fundamental
changes which affected this probability.
Similarly, the more severe movements in the
prices of takeover shares also squares with the
notion that they are more susceptible to bubbles
because their fundamentals are more difficult to
evaluate. The uncanny relationship between the
October 1987 price movements and those of
October 1929 then simply may be the bursting
of two bubbles and may say very little about the
similarities of the fundamental economic

Randall Johnston Pozdena

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




!!omoH O!UJOdllO)




JO ~uo8
aAJaSa~ IOJapa:::l
~uaw~Jodaa lpJOaSa~
The table entitled, "Selected Assets and Liabilities of Large Commercial Banks in the Twelfth
Federal Reserve District," will no longer be published in conjunction with the Weekly Letter.
For those in need of these data, a more timely publication entitled, "Weekly Consolidated
Condition Report of Large Commercial Banks and Domestic Subsidiaries" (F .R. 2416x), is
avai lable from the Statistical and Data Services Department of this Bank.