One of the broad issues that you
have been discussing today is the nature of financial risk. This
evening I will offer my perspective on the fundamental sources of
financial risk and the value added of banks and other financial
intermediaries. Then, from that perspective, I will delve into
some of the pitfalls inherent in risk-management models and the
challenges they pose for risk managers.
Risk, to state the obvious, is
inherent in all business and financial activity. Its evaluation is
a key element in all estimates of wealth. We are uncertain that
any particular nonfinancial asset will be productive. We're also
uncertain about the flow of returns that the asset might engender.
In the face of these uncertainties, we endeavor to estimate the
most likely long-term earnings path and the potential for actual
results to deviate from that path, that is, the asset's risk.
History suggests that day-to-day movements in asset values
primarily reflect asset-specific uncertainties, but, especially at
the portfolio level, changes in values are also driven by
perceptions of uncertainties relating to the economy as a whole
and to asset values generally. These perceptions of broad
uncertainties are embodied in the discount factors that convert
the expectations of future earnings to current present values, or
wealth.
In a market economy, all risks
derive from the risks of holding real assets or, equivalently,
unleveraged equity claims on those assets. All debt instruments
(and, indeed, equities too) are essentially combinations of long
and short positions in those real assets. The marvel of financial
intermediation is that, although it cannot alter the underlying
risk in holding direct claims on real assets, it can redistribute
risks in a manner that alters behavior. The redistribution of risk
induces more investment in real assets and hence engenders higher
standards of living.
This occurs because financial
intermediation facilitates diversification of risk and its
redistribution among people with different attitudes toward risk.
Any means that shifts risk from those who choose to withdraw from
it to those more willing to take it on permits increased
investment without significantly raising the perceived degree of
discomfort from risk that the population overall experiences.
Indeed, all value added from new
financial instruments derives from the service of reallocating
risk in a manner that makes risk more tolerable. Insurance, of
course, is the purest form of this service. All the new financial
products that have been created in recent years, financial
derivatives being in the forefront, contribute economic value by
unbundling risks and reallocating them in a highly calibrated
manner. The rising share of finance in the business output of the
United States and other countries is a measure of the economic
value added from its ability to enhance the process of wealth
creation.
But while financial
intermediation, through its impetus to diversification, can lower
the risks of holding claims on real assets, it cannot alter the
more deep-seated uncertainties inherent in the human evaluation
process. There is little in our historical annals that suggests
that human nature has changed much over the generations. But, as I
have noted previously, while time preference may appear to be
relatively stable over history, perceptions of risk and
uncertainty, which couple with time preference to create discount
factors, obviously vary widely, as does liquidity preference,
itself a function of uncertainty. These uncertainties are an
underlying source of risk that we too often have regarded as
background noise and generally have not endeavored to capture in
our risk models.
Almost always this has been the
right judgment. However, the decline in recent years in the equity
premium--the margin by which the implied rate of discount on
common stock exceeds the riskless rate of interest--should prompt
careful consideration of the robustness of our portfolio
risk-management models in the event this judgment proves wrong.
The key question is whether the
recent decline in equity premiums is permanent or temporary. If
the decline is permanent, portfolio risk managers need not spend
much time revisiting a history that is unlikely to repeat itself.
But if it proves temporary, portfolio risk managers could find
that they are underestimating the credit risk of individual loans
based on the market value of assets and overestimating the
benefits of portfolio diversification.
There can be little doubt that
the dramatic improvements in information technology in recent
years have altered our approach to risk. Some analysts perceive
that information technology has permanently lowered equity
premiums and, hence, permanently raised the prices of the
collateral that underlies all financial assets.
The reason, of course, is that
information is critical to the evaluation of risk. The less that
is known about the current state of a market or a venture, the
less the ability to project future outcomes and, hence, the more
those potential outcomes will be discounted.
The rise in the availability of
real-time information has reduced the uncertainties and thereby
lowered the variances that we employ to guide portfolio decisions.
At least part of the observed fall in equity premiums in our
economy and others over the past five years does not appear to be
the result of ephemeral changes in perceptions. It is presumably
the result of a permanent technology-driven increase in
information availability, which by definition reduces uncertainty
and therefore risk premiums. This decline is most evident in
equity risk premiums. It is less clear in the corporate bond
market, where relative supplies of corporate and Treasury bonds
and other factors we cannot easily identify have outweighed the
effects of more readily available information about borrowers.
The marked increase over this
decade in the projected slope of technology advance, of course,
has also augmented expectations of earnings growth, as evidenced
by the dramatic increase since 1995 in security analysts'
projections of long-term earnings. While it may be that the
expectations of higher earnings embodied in equity values have had
a spillover effect on discount factors, the latter remain
essentially independent of the earnings expectations themselves.
That equity premiums have
generally declined during the past decade is not in dispute. What
is at issue is how much of the decline reflects new, irreversible
technologies, and what part is a consequence of a prolonged
business expansion without a significant period of adjustment. The
business expansion is, of course, reversible, whereas the
technological advancements presumably are not.
Some analysts have offered an
entirely different interpretation of the drop in equity premiums.
They assert that a long history of a rate of return on equity
persistently exceeding the riskless rate of interest is bound to
induce a learning-curve response that will eventually close the
gap. According to this argument, much, possibly all, of the
decline in equity premiums over the past five years reflects this
learning response. It would be a mistake to dismiss such notions
out of hand. We have learned to no longer cower at an
eclipse of the sun or to run for cover at the sight of a
newfangled automobile.
But are we really observing in
today's low equity premiums a permanent move up the learning curve
in response to decades of data? Or are other factors at play? Some
analysts have suggested several problems with the learning curve
argument. One is the persistence of an equity premium in the face
of the history of "excess" equity returns.
Is it possible that responses
toward risk are more akin to claustrophobia than to a learning
response? No matter how many times one emerges unscathed from a
claustrophobic experience, the sensitivity remains. In that case,
there is no learning experience.
Whichever case applies, what is
certain is that the question of the permanence of the decline in
equity premiums is of critical importance to risk managers. They
cannot be agnostic on this question because any abrupt rise in
equity premiums must inevitably produce declines in the values of
most private financial obligations. Thus, however clearly they may
be able to evaluate asset-specific risk, they must be careful not
to overlook the possibilities of macro risk that could undermine
the value of even a seemingly well-diversified portfolio.
I have called attention to this
risk-management challenge in a different context when discussing
the roots of the international financial crises of the past two
and a half years. My focus has been on the perils of risk
management when periodic crises--read sharply rising risk
premiums--undermine risk-management structures that fail to
address them.
During a financial crisis, risk
aversion rises dramatically, and deliberate trading strategies are
replaced by rising fear-induced disengagement. Yield spreads on
relatively risky assets widen dramatically. In the more extreme
manifestation, the inability to differentiate among degrees of
risk drives trading strategies to ever-more-liquid instruments
that permit investors to immediately reverse decisions at minimum
cost should that be required. As a consequence, even among riskless
assets, such as U.S. Treasury securities, liquidity premiums rise
sharply as investors seek the heavily traded
"on-the-run" issues--a behavior that was so evident last
fall.
As I have indicated on previous
occasions, history tells us that sharp reversals in confidence
occur abruptly, most often with little advance notice. These
reversals can be self-reinforcing processes that can compress
sizable adjustments into a very short period. Panic reactions in
the market are characterized by dramatic shifts in behavior that
are intended to minimize short-term losses. Claims on far-distant
future values are discounted to insignificance. What is so
intriguing, as I noted earlier, is that this type of behavior has
characterized human interaction with little appreciable change
over the generations. Whether Dutch tulip bulbs or Russian
equities, the market price patterns remain much the same.
We can readily describe this
process, but, to date, economists have been unable to anticipate
sharp reversals in confidence. Collapsing confidence is generally
described as a bursting bubble, an event incontrovertibly evident
only in retrospect. To anticipate a bubble about to burst requires
the forecast of a plunge in the prices of assets previously set by
the judgments of millions of investors, many of whom are highly
knowledgeable about the prospects for the specific investments
that make up our broad price indexes of stocks and other assets.
Nevertheless, if episodic
recurrences of ruptured confidence are integral to the way our
economy and our financial markets work now and in the future, the
implications for risk measurement and risk management are
significant.
Probability distributions
estimated largely, or exclusively, over cycles that do not include
periods of panic will underestimate the likelihood of extreme
price movements because they fail to capture a secondary peak at
the extreme negative tail that reflects the probability of
occurrence of a panic. Furthermore, joint distributions estimated
over periods that do not include panics will underestimate
correlations between asset returns during panics. Under these
circumstances, fear and disengagement on the part of investors
holding net long positions often lead to simultaneous declines in
the values of private obligations, as investors no longer
realistically differentiate among degrees of risk and liquidity,
and to increases in the values of riskless government securities.
Consequently, the benefits of portfolio diversification will tend
to be overestimated when the rare panic periods are not taken into
account.
The uncertainties inherent in
valuations of assets and the potential for abrupt changes in
perceptions of those uncertainties clearly must be adjudged by
risk managers at banks and other financial intermediaries. At a
minimum, risk managers need to stress test the assumptions
underlying their models and set aside somewhat higher contingency
resources--reserves or capital--to cover the losses that will
inevitably emerge from time to time when investors suffer a loss
of confidence. These reserves will appear almost all the time to
be a suboptimal use of capital. So do fire insurance premiums.
More important, boards of
directors, senior managers, and supervisory authorities need to
balance emphasis on risk models that essentially have only dimly
perceived sampling characteristics with emphasis on the skills,
experience, and judgment of the people who have to apply those
models. Being able to judge which structural model best describes
the forces driving asset pricing in any particular period is
itself priceless. To paraphrase my former colleague Jerry
Corrigan, the advent of sophisticated risk models has not made
people with grey hair, or none, wholly obsolete.