Personal distractions lessen a manager’s ability to produce alpha. This is true of both happy and unhappy distractions: that is, with both marriages and divorces. This seems intuitively likely at any rate, and it has just received empirical support in a new paper by Yan Lu of the University of Florida and two others.
Lu et al found that both marriage and divorce are “associated with significantly lower fund alpha.” The alpha lowering effect extends back from the event six months and forward two years. Relative to the pre-event returns, fund alpha falls by an annualized 8.5% due to a marriage and 7.39% due to a divorce.
Lu’s paper begins with a quote from Paul Tudor Jones II, of Tudor Investment Corp. Tudor made this statement at a panel discussion at the University of Virginia: “One of my No. 1 rules as an investor is as soon as … I find out that [a] manager is going through divorce, [I] redeem immediately. Because the emotional distraction that comes from divorce is so overwhelming.”
Not all distractions are alike, though. Lu et al find that marital events do more harm to return in the context of liquid and high-tempo strategies, such as managed futures and global macro. Strategies that turn on more illiquid assets and involve trading less frequently (such as distressed debt) can survive a marriage or divorce with fewer dings and dents.
“In the months surrounding marriage” the report says, “managers running high tempo strategies suffer an annualized 15.50 percent deterioration in risk adjusted performance while managers running low tempo strategies only suffer an annualized 7 percent drop….”
Not Just a Distraction
As the Tudor quote indicates, such a result is more intuitive for divorce than for marriage. Divorce is something worse than a mere distraction after all. For many people divorce is a kick in the gut. Marriage though? What exactly is the mechanism involved?
It is apparently a matter of inattention to business during the period of courtship and honeymoon. Inattentive managers, say these authors, “are less disciplined about cutting losses and more susceptible to the disposition effect.” The “disposition effect” is a behavioral-finance term for the asymmetry between the road up and the road down especially among equity traders. Traders are said to find it easier to take profits (that is, to sell a stock while it is on the way up) than they do to cut losses (selling while it is on the way down.) It is easier for single people, or for those who have settled into married life and put their sweet honeymoon days behind them, to resist the disposition effect than it is for love-birds to do so.
This study is in a sense a continuation of work by a Williams College professor with the apt name David Love. Love wrote The Effect of Marital Status and Children on Savings and Portfolio Choice in 2008. He was compiling evidence about household decisions there, not the alpha-generation of professional asset managers. Still, Love developed a “life-cycle model that feature[d] exogenous changes.” Lu et al are seeking to do much the same.
Yet it doesn’t follow that they reach the same conclusions as Love, even when the subject matter overlaps. Love’s model predicted, for example, that risk taking would increase after a divorce. Lu’s findings indicate that fund managers become more risk averse after marriage, but that divorce has no contrary effect. Indeed, divorce may cause another (slight) increase in risk aversion, says Lu, “although the effects are not reliably different from zero.”
One of the many questions that the Lu study leaves open involves the possibility that men may react differently from women to the marital events in question. We can’t really know, because their sample consisted almost entirely of men.
But for what it’s worth, the authors say in a footnote that “there are two female managers who tied the knot during our sample period. For these two female managers, the average impact of marriage on fund performance is consistent with that for the entire sample. In particular, their annualized fund alpha drops from 9.72 percent in the pre event period to 6.12 percent in the event period, and subsequently to -5.52 percent in the post event period.”