It would be beneficial to decompose
risk management into independent
activities that could be aggregated.
distributions that go into it. We disagree, and believe that
simulating a financial plan before investing in it is like shaking a ladder before you climb on it. Even though the distribution of “shaking” forces is not the same as the distribution
of “climbing” forces, we will continue to recommend ladder
shaking as an effective risk management technique.
In this light, the DIST is not a magic bullet. We do not
believe that new technologies, on their own, can necessarily provide us with the “right answers.” Instead, successful
breakthroughs, like the spreadsheet itself, are those that make
it easier to ask the “right question.” Accordingly, our sample
model is not based on a single “right” scenario library, but
allows the user to toggle between two sets of assumptions.
Business unit managers will still be rewarded on actual
performance, not merely for writing plans that look good under the DISTs. The CPO might think the chance of a new
product succeeding is only 1%, but that judgment does not
constrain the actions of the product manager, who might
launch it in the belief success is much more likely.
Perhaps the DIST library’s most important role is to raise
awareness of imprecisely understood market relationships.
Indeed, some of the new simulation packages are so fast that
the probability of success of a product, the likelihood of an
adverse market event or the impact of a Yak butter surplus
on the S&P 500 may be varied by the user in real time, to
gauge the robustness of a strategic position.
Long-Term Implications
Risk transparency, which means access to both probabilistic
information and the tools to manipulate it, leads to better
decisions from general staff to the front line. “The DIST
format provides a standard that allows the top of the organization to pass down global risk inputs to the divisions and
then consolidate the divisional risks in a coherent way,” says
Professor Scholtes.
At headquarters, executives, regulators and investors can
run standardized DIST libraries through models of their divisions, industry member organizations or investments, and
then aggregate the results to identify concentrations of risk.
They can ensure their corporate resources are adequate to
support the probabilistic needs of their business units.
At the front, risk takers will have additional signals for
evaluating bets. Risk decisions will be debated before the fact
with shared information, rather than being political issues to
be spun after the fact, with great advantage to those who can
hide or forge key data.
Adopting an additive approach to transmitting, viewing
and interacting with probability distributions will not accomplish all these goals, any more than Arabic numerals solved
all quantitative problems. But standardization and ease-of-use, in writing numbers and discussing risk, is an essential
step, and a large one. DISTs can help people make better
decisions immediately, and in the longer-term, change the
way they think about uncertainty.
FOOTNOTES
1. C. Seife. Zero: The Biography of a Dangerous Idea. Penguin Books,
New York, 2000.
2. S. Craighead and M. Tenney. “Economic Scenario Genera-
tor for Insurance and Pension Rational Decision Making Under
Uncertainty,” Actuarial Research Clearing House, 1997 (Vol 1).
3. A. Cairns, T. Kleinow, S. Sahin, , A. D. Wilkie. “Revisiting
the Wilke Investment Model.” June 12, 2008. See http://www.
actuaries.org/AFIR/Colloquia/Rome2/Cairns_Kleinow_Sa-
hin_Wilkie.pdf
4. http://www.soa.org/professional-interests/technology/tech-
scenario-file-format.aspx
5. S. Savage, S. Scholtes and D. Zweidler. “Probability Manage-
ment,” OR/MS Today, February 2006, Volume 33, Number 1.
6. S. Savage. The Flaw of Averages: Why We Underestimate Risk in the
Face of Uncertainty. John Wiley & Sons, Hoboken, 2009.
7. XML is a generalization of HTML, the string format for creat-
ing web pages.
8. J. Tukey. Exploratory Data Analysis. Addison-Wesley, 1977.
9. www.ProbabilityManagement.org.
Sam L. Savage is the chairman of ProbabilityManagement.org, a fellow of the
Judge Business School at Cambridge University and a consulting professor of Management Science and Engineering at Stanford University. He describes the evolution
of the DIST data type in his recently published book, The Flaw of Averages:
Why We Underestimate Risk in the Face of Uncertainty (Wiley, 2009).
Aaron Brown is the risk officer at AQR Capital Management and the author of A
World of Chance (Cambridge University Press, 2008) and The Poker Face
of Wall Street (Wiley, 2006). He serves on Risk Professional’s editorial board
and can be reached at Aaron.Brown@aqr.com.
