Decisions and outcomes are not necessarily related. One can make a good decision that results in a bad outcome, but this does not mean the decision itself was bad. This can be represented by a simple table:
Good Outcome | Bad Outcome | |
Good Decision | ||
Bad Decision |
Here are a few examples that come to mind:
Good Outcome | Bad Outcome | |
Good Decision | Eating a salad and not getting sick Planning for retirement and being able to retiring early | Eating a salad contaminated with E. coli Investing a lump sum in an index fund on the Friday before Black Monday |
Bad Decision | Buying a lotto ticket and winning big Driving drunk without incident | Speculating in BitCoin and losing Driving drunk and causing an accident |
We know that investing a lump sum now is better than dollar-cost averaging your way into stocks or timing the market by attempting to “buy the dip” (e.g., Williams & Bacon, 1993; Panyagometh & Zhu, 2016). Although lump-sum investing is the preferable decision, there is a nontrivial probability of an inferior outcome as compared to investing at a later time. If a bad outcome occurs, it is more salient than had a good outcome of equal magnitude occurred. However, this should basically be chalked up to bad luck. A bad outcome does not mean a bad decision was made.
Separating decisions from outcomes goes against our nature. It is contrary to human psychology. In her 2018 book, Thinking in Bets, poker champion Annie Duke calls the human prediction to judge decisions by the resultant outcomes “resulting.” Resulting is akin to confusing causation for correlation in science.
Making a bet where the odds are in your favor is a good decision, even if you lose. With more and more such bets, a result commensurate with the prudence of the decision approaches inevitability. In the stock market, you can think of each trading day as a bet, with these bets stacking up over time. Below, probabilities from Bloomberg data, compiled by Vanguard, show the probability of positive returns for S&P 500 investment time frames within the selected dates (1/04/1988 to 2/16/2018).
S&P 500 investment during 1/04/1988–2/16/2018 | Probability of positive return |
One day | .54 |
One week | .58 |
One month | .64 |
One year | .83 |
Ten years | .91 |
Although start and end points matter, the pattern has been shown to hold even over the duration of the stock market’s history, including the Great Depression. Above, we see the probabilities of positive returns averaged across all day, week, month, year, and 10-year periods within a 30-year range. A 54% chance of positive returns on any particular day increases to a 91% chance of positive returns during any particular 10-year period within the 30-year period sampled.
Of course, this data nevertheless shows a 9% chance of losing money in a 10-year span. However, if you are unlucky enough to have invested the bulk of your money at an unfortunate time, this does not mean your decision was bad—just that you happened to have a bad outcome. It takes longer than 10 years for the probability of positive returns to approach inevitability—more like 30 years. Time will tell whether the recent market peak on September 20, 2018 will require months, years, or more than a decade to overcome.
The financial industry is built on confounding decisions with outcomes. A hedge fund manager is said to be “hot,” endowed with stock-picking genius, if his speculations pay off in a given year. Even for investors who were lucky enough to pick him, their decision was certainly bad; picking a low-cost index-tracking mutual fund and sticking with it for many years is a better decision. The speculator’s success is based on chance and luck, not skill. The speculator’s decisions are always bad, although their outcomes may be good, for a time. Eventually, good luck will inevitably run out, leading to underperformance of the index-tracking mutual fund, or worse, a spectacular capital wipeout à la Enron or Bernie Madoff.
We must all take a step back to carefully consider whether a good outcome was actually the result of a good decision, and whether a bad outcome resulted from a bad decision, or from a good decision that should be repeated despite a bad outcome occurring this particular time. On the whole, as a series of good decisions lengthens, good outcomes become inevitable, and as a series of bad decisions lengthens, bad outcomes become inevitable. In making such determinations, our psychology and the limited information available may work against us.