Go to Brad De Long's Home
Can a Market Be Too Efficient?
Brad De Long
J. Bradford De Long, an economic historian, is Associate Professor of
Economics at the University of Califomia at Berkeley. Until June 1995,
he was Deputy Assistant Secretary for Economic Policy of the United States
The nation-state came first. The industrial corporation came second.
The nation-state borrowed to finance wars of defense and of conquest. The
industrial corporation sought investors--equity owners to share risk and
debt holders to provide capital--to finance the plant and equipment to take
advantage of the modern production technologies discovered in the Industrial
Both nation-states and industrial corporations sought long-term capital.
Money spent building a factory and buying machine tools, or "invested"
in an army and a fleet, cannot be unspent and returned original lender.
Nation-states and corporations had to borrow long term. But investors
sought to lend short. Investors were eager to eam interest and dividends,
but not eager to commit their wealth irrevocably.
The solution was the liquid financial market, in which the long-term debt
and equity obligations of nation-states and industrial corporations are
bought and sold. Liquid financial markets permitted borrowers to borrow
long-term, and also allowed lenders to loan short-term. Liquid financial
markets gave investors the ability to rapidly turn their portfolios back
into spendable cash at little notice. Investors found such assets traded
on liquid financial markets highly attractive, and were willing to loan
and invest on terms nation-states and corporations found remarkably attractive.
And the more liquid the market, the lower the interest rate or dividend
yield required to support a debt or equity issue. Without liquid financial
markets, the cost of capital over the past century and a half would have
been much higher--the flow of capital into industry would have been much
lower--and we would be much poorer today.
How do modern financial markets work this magic? How do they allow borrowers
and issuers to borrow and issue for the long, fixed terms that they prefer,
while simultaneously allowing investors to invest only for the short, variable
terms that they prefer?
Financial markets work this magic by relying on statistics: the law of large
numbers. With many possible buyers and sellers, there will always be an
investor willing to a security that another investor wishes to sell. The
more "liquid" the market, the smaller the transaction costs, the
lower the risk that an investor wishing to sell will find no one willing
to buy at a price near that of the last transaction--and the better for
the investor, and ultimately for the issuing corporation as well.
However, there is a catch. The statistical law of large numbers applies
only if investors make separate decisions to buy or sell. If all,
or if a large chunk, of investors are responding to a single stimulus or
are participants in a wave of pessimism and rumor sweeping through the marketplace,
then there will be willing buyers at prices near that of the last transaction.
Prices will fall--the wave of pessimism and rumor will become a wave of
panic and despair--and the market will crash.
In a not-very-liquid financial market, the magnitude of the crash is likely
to be small. Only a fraction of those contaminated by the contagion of fear
will be able to get their sell orders executed. Some market makers and investors
will remain uncontaminated, and will view the incipient panic as a chance
to make money by buying assets at prices that are very attractive from a
long-run point of view. As long as those uncontaminated by pessimism and
rumor can mobilize enough assets quickly enough, relative to the wave of
panic selling coming through the marketplace, the market will experience
a panic but not a crisis.
A more liquid financial market may well increase the risk that rumor becomes
panic which turns into crisis. The more liquid are the institutional arrangements
of the market, the larger is the fraction of panicked and fearful investors
whose sell orders are executed. Liquidity-increasing institutional arrangements
are unlikley to multiply the ability of those uncontaminated by pessimism
and rumor to rapidly mobilize more capital and take larger positions in
a panic based on their assessment of long-run fundamentals.
With the potential for more destabilizing, panic sales by relatively uninformed
investors, and with little change in the amount of stabilizing speculation
by informed investors, there is a chance that markets with institutional
arrangements that provide more liquidity will be subject to more panics
A market crash has serious consequences not just for those investing in
financial markets but for the real economy as well: employment and growth
depend on the smooth working of the network of financial intermediation
that transforms individuals' savings into corporations' plant and equipment.
A market crash threatens the solvency of many financial institutions, and
stands a good chance of disrupting the network of financial intermediation.
What is the solution? How are we to reap the benefits of institutional transformations
that promise greater liquidity--thus lower interest rates for borrowers,
and better value for investors--without amplifying the risk that more "efficient"
financial institutions will create a greater vulnerability to panic, crisis,
and crash? That is a very hard question. "Circuit breakers" to
limit rapid price movements, suspend the use of automatic trading strategies,
and give potential buyers uncontaminated by the contagion of panic time
to come to the marketplace help--a little. Limits on the use of volatile
securities as collateral (because such use tends to make desired sales rise
as prices fall) help--a little. Regulating derivative transactions to make
it apparent exactly which marketplace participants are ultimately bearing
which risks help--a little. Larger pools of bank, investment bank, and mutual
fund capital devoted to seeking long-term value rather than short-term quarter-by-quarter
performance would help--perhaps more than a little.
But no one should think that financial market regulators have all the answers,
or have tools strong enough to handle all contingencies.
Go to Brad De Long's Home
Associate Professor of Economics Brad De
Long, 601 Evans
University of California at Berkeley; Berkeley, CA 94720-3880
(510) 643-4027 phone (510) 642-6615 fax