William J. Bernstein
A Mean-Variance Society?
Over the past decade, my forays into finance have brought me into contact with a lot of fascinating folks, most of whom are generally pretty pleasant, quantitatively gifted, and often possessed of unique insights into human nature and the world around them. All of which has made for stimulating emails and enjoyable restaurant dining.
One other thing. They are, almost to a man and to a woman, politically conservative: card-carrying Democrats are as rare on the floor of the New York Stock Exchange or the meeting rooms of a CFA Institute convention as infidels in Mecca.
It is interesting, then, that one of the most basic tools of finance offers unique insights into what makes the different species of homo politicus tick. I’m talking, of course, about mean-variance analysis—the "shot heard round the world," fired by Harry Markowitz in the Journal of Finance 1952.
As every visitor to these pages knows, Markowitz famously posited that both low risk and high return are of utility. Thus, at a given level of risk, defined as variance of returns, investors always preferred more return, and that at a given level of return, investors always preferred less risk. The obvious implication: the rational investor (if there is indeed such a beast) should always be willing to trade off return and risk. Just how much return the investor is willing to exchange for a given unit of risk reduction depends upon his risk aversion; for those with low risk aversion, the answer is very little, and for those with high risk aversion, quite a lot.
At the extremes, one finds only a few rara avises: investors so risk-tolerant that they happily hold leveraged portfolios of risky assets, and those so risk-averse that they hold only T-bills. The vast majority choose to hold a prudent mix of risky and riskless assets.
What does this have to do with politics? Everything. For "return" substitute "economic growth," and for "variance," substitute "income/wealth dispersion." Just as increased return provides utility, so does increased economic growth. And just as decreased return variance provides utility, so does reduced income and wealth dispersion. To paraphrase Mencken, an unhappy man is one whose salary is less than that of his wife’s brother-in-law. The greater the gap, the greater the misery.
Consider, for example, the increasing wealth dispersions seen in this nation, as demonstrated by the proportion of national income gathered by the top one percent of wage earners in a landmark study by economists Thomas Piketty and Emmanuel Saez:
And, for good measure, take a gander at this plot of year-to-year wage variance among individuals, calculated by political scientist Jacob Hacker:
What do you see in these pictures? Perhaps your mind’s eye envisions a vibrant economy generously rewarding the heroes who create the jobs and prosperity for everyone else. Then you, as the Governor of California might say, are a Republican. Contrariwise, perhaps you see an increasingly vicious winner-take-all society in which per-capita GDP rises at the expense of millions getting their dinners from the dumpster, and tens of millions without adequate medical coverage. In which case, you’re probably not.
By now, perhaps, you anticipate the punch line: like the risk-tolerant investor who cares only about return and gives not a damn about risk, the libertarian concentrates only on mean societal wealth—raw, unvarnished per-capita GDP—and not the damage done by the variance. And just like the risk-adverse T-bill owner, those on the left care only about the variance, and ignore the damage done by overly vigorous leveling to the mean. (Another tip-off to where someone lies in the political spectrum: Is the mean or the median the preferred summary descriptor?)
But here, unfortunately, the similarity ends. Whereas both financial species usually have some insight into the nature of their portfolio choices, those at either end of the political spectrum are about as self-aware as Donald Trump. Liberals seem blissfully unaware of the potential dangers of too-generous public spending and over-regulation, while libertarians are congenitally unable to recognize the highly corrosive nature of increasing wealth and income inequalities. Further, the tails at both political extremes are much fatter than those at the financial extremes.
I can excuse those on the left, most of whom never got within, shall we say, shouting distance of an economics classroom. My libertarian friends in finance, however, I find hard to fathom. While they have little trouble understanding the application of mean-variance to money management, when they approach politics and social policy, they become strangely dyslexic.
The quality of political discourse in this nation would improve greatly if our national-policy debates emulated the deliberations of a collegial investment committee consisting of individuals with varying degrees of risk tolerance as it gropes towards an agreement on the risk/reward characteristics of its investment portfolio.
There is much progress to be made on both sides. Those on the left need a better understanding of the mean; those on the right, the variance, and both need to clearly grasp the tradeoffs therein.
Copyright © 2006, William J. Bernstein. All rights reserved.
The right to download, store and/or output any material
on this Web site is granted for viewing use only. Material may not
be reproduced in any form without the express written permission of William
J. Bernstein. Reproduction or editing by any means, mechanical or electronic,
in whole or in part, without the express written permission of William J.
Bernstein is strictly prohibited. Please read the disclaimer.
The right to download, store and/or output any material on this Web site is granted for viewing use only. Material may not be reproduced in any form without the express written permission of William J. Bernstein. Reproduction or editing by any means, mechanical or electronic, in whole or in part, without the express written permission of William J. Bernstein is strictly prohibited. Please read the disclaimer.