The Election Forecasters' Investment Lesson

Mano a Mano In 2012, Nate Silver's FiveThirtyEight website and Sam Wang's Princeton Election Consortium (PEC) competed covering that year's Presidential campaign. It wasn't much of a contest. The two entities landed at essentially the same spot, meaning that they were both right. FiveThirtyEight called all 50 states correctly, while PEC missed only the tossup that was Florida, and predicted the popular vote almost perfectly. Win-win.

The 2016 election played out differently. As with the 2012 election, FiveThirtyEight and PEC held the same general view, which was the Democratic candidate led throughout the campaign in the polls, and thus had the better chances. Given that each prognosticator relied upon the same data sources--the national and state polls conducted by various outside firms--it would have been surprising if the two had pointed in opposite directions. They did not.

However, unlike four years ago, the two sites held vastly different views about the certainty of results. Whereas four years ago both had assigned consistently high probabilities to President Obama, dipping only briefly after the first Obama-Romney debate (and then rebounding after the second), their probabilities diverged sharply in 2016.

FiveThirtyEight treated the race as landing within a 50%-90% certainty band, with Donald Trump deadlocking the race twice: in early August, just after the Republican national convention, and in late September, just before the candidates' first debate. On Nov. 8, when the election occurred, Trump occupied the middle of his FiveThirtyEight range, having a roughly 30% chance of taking the electoral vote.

PEC told another story altogether. By its measure, the 2016 Presidential contest was "the most stable" in recent history, with Hillary Clinton leading from wire to wire. PEC consistently pegged Clinton's winning chances as being very high. (Often, they were listed at 99%-plus, and never less than 90%.) In a television interview, Wang promised to "eat a bug" if Donald Trump prevailed--which Wang defined not as winning the election, but instead as keeping things close, by collecting at least 240 electoral votes. (Gobble, gobble.)

Academic vs. Practical Considerations There were several official explanations for why PEC was more certain than FiveThirtyEight. The main ones were the treatment of undecided voters, how state polling errors might be correlated, and to what extent Trump could lose the popular vote but win the electoral college. While of interest to political junkies, those issues are immaterial to this column.

The subject is instead the forecasters' approaches. As Wang writes, although FiveThirtyEight is ostensibly for-profit and his site his academic, in reality the two roles are reversed. The ESPN-owned website run by the generalist is more academic in its mindset, while the blog written by the Ph.D. mathematician is geared for action.

Because FiveThirtyEight uses the more complex of the two models, its output is unstable. Writes Wang, "every parameter has an uncertainty attached to it. When all those parameters get put together to estimate the overall outcome, the resulting total carries greater risk of accumulating uncertainty that is hard to keep under control." FiveThirtyEight is for those who wish to muse, "mulling over the alternatives."

PEC, in contrast, was built for speed. (Relatively speaking, that is.) It does not seek discussions about questions that seem to have clear answers. If the Presidential contest looks to have an obvious result, as PEC's model indicated, then why waste the audience's time? Mark that topic as settled and move onto areas that are truly up for grabs, such as whether the Democrats will control the Senate.

In espousing his mindset, Wang specifically mentions the needs of professional investors:

Several weeks ago I visited a major investment company to talk about election forecasting. Many people there had strong backgrounds in math, computer science, and physics. They were highly engaged in the Princeton Election Consortium's math and were full of questions. I suddenly realized that we did the same thing: estimate the probability of real-world events, and find ways to beat the "market."

In the case of PEC, the "market" is conventional wisdom about whether a race is in doubt. If a race is a certain win or a certain loss, it is pointless to put in money and effort, assuming that the rest of the market is in the game. On the other hand, if a race is in doubt, then it may be moved by a little extra push. Think of it as "math for activism." This point of view heavily influences my calculations.

Less Was Less Oh, the irony. The model that was built for action urged a mistake. Meanwhile, the model that was built for rumination gave its users roughly what they needed. No, it did not predict the election's winner, but it asserted that the outcome was in reasonable doubt. Indeed, it most certainly was.

Simplicity sells. In politics, in election forecasting, and in investing. Successful politicians know not to delve into the details. As for election forecasters, this time around FiveThirtyEight was wildly criticized. It was too uncertain. It was "underconfident," even to the point of (according to The Huffington Post) rigging the polls in Donald Trump's favor. For their part, investment managers pound the table. They have "conviction" in their results. Such self-belief is almost always regarded as a good thing.

Whether the simpler message, arrived at through conviction, is the better approach for being right is a different matter. Complex approaches are not always superior to parsimonious methods in guiding the decisions of politicians, prediction models, or investment managers. Perhaps not even usually superior. However, as demonstrated by the case of FiveThirtyEight and PEC, in the Presidential election of 2016, at the very least they sometimes are better.

In short, I would not invest in the fund of a portfolio manager who was afraid to waffle. Better to equivocate, when required, than to show conviction when it is not warranted. The former might disturb prospective investors, for managers who are on the road looking to attract new business. But the latter will end up bothering existing shareholders--and that is the greater of the two sins.

John Rekenthaler has been researching the fund industry since 1988. He is now a columnist for Morningstar.com and a member of Morningstar's investment research department. John is quick to point out that while Morningstar typically agrees with the views of the Rekenthaler Report, his views are his own.