Outcome Bias

Outcome bias is the tendency to evaluate a decision by its result instead of by the quality of the reasoning and information behind it. Jonathan Baron and John Hershey documented it in a 1988 paper in the Journal of Personality and Social Psychology. Their undergraduate subjects, given full information about decisions others had made under uncertainty about medical treatments and monetary gambles, "rated the thinking as better, rated the decision maker as more competent, or indicated greater willingness to yield the decision when the outcome was favorable than when it was unfavorable." The finding held even though the subjects "felt that they should not consider outcomes in making these evaluations."

When outcomes are fair evidence, and when they are not. Baron and Hershey were careful about this. An outcome can legitimately tell an observer something when the decision maker knew facts the observer does not, and a run of outcomes from similar decisions can teach something about the odds. The bias is the case they isolated: the judge is given everything the decision maker knew, including the probabilities, and still grades the decision by how it came out. In their words, the question was "not whether people use outcome information but whether there are conditions under which they overuse it."

How it differs from hindsight bias. The paper draws the line itself. Hindsight bias, documented by Baruch Fischhoff in 1975, is "an effect of outcome information on the judged probability of an outcome": once the result is known, people believe it was more predictable than it was. Outcome bias is "a direct effect of outcome on evaluation of decisions, an effect that does not operate through an effect of outcome knowledge on a judge's assessed probabilities of outcomes." Baron and Hershey held the probabilities fixed to show the second effect on its own, while noting that in real life the two could work together.

Why it matters for money. Investment results combine skill and chance, and over a single year chance can dominate. A concentrated bet that happened to pay off can be rated a good decision, and a sensibly diversified portfolio that lagged can be rated a mistake. The same habit applied to a fund manager produces hiring and firing on recent returns, and applied to one's own decisions it rewards lucky risks and so encourages more of them. Judging the decision as it looked beforehand, with the information and the range of outcomes that were reasonable at the time, is the alternative.

The bias enters at the moment of review. Looking back at a decision, the known outcome is the most vivid fact, and it colors the judgment of how carefully the choice was made, how competent the decision maker was, and whether to trust them again. The defense is to separate the two questions and answer the first one using only what was knowable at the time.

For example, two investors each have $20,000. Kwame puts all of it into one small company's stock; Dana puts hers in a broad index fund. Suppose that, looking forward, Kwame's position had a 30 percent chance of doubling and a 70 percent chance of losing half its value, an expected ending value of 0.30 × $40,000 + 0.70 × $10,000 = $12,000 + $7,000 = $19,000, while Dana's fund was expected to grow modestly with far less risk. The stock doubles to $40,000 and the fund ends the year at $21,000. Judged by outcome, Kwame decided better, with a $19,000 lead. Judged by the decision, on those assumed odds he chose a bet with a lower expected value and a 70 percent chance of a $10,000 loss, and the good result was the less likely branch coming up. Reviewing it that way is what keeps a lucky result from becoming a strategy.

Pros

• Naming the bias separates two questions people blur: was the decision sound, and did it work out?
• It explains why a record of good results over a short period says less about skill than it seems to.
• The fix, a written record of reasons and expectations made at the time, is simple and costs little.

Cons

• Separating luck from judgment is hard in practice, because the probabilities behind real decisions are rarely known as precisely as in the experiments.
• Outcomes do carry information over many similar decisions, so ignoring them entirely would be its own mistake.
• The original evidence came from students rating hypothetical cases, so the size of the effect in real financial decisions is not established by that study.