The Power of Certainty: How Behavioral Economics Helps Us Understand the "Trump Tape"


By now, pretty much everyone in America knows about the tape of Donald Trump saying lewd things about women while on an Access Hollywood bus in 2005. The recording has had a seismic effect on the presidential campaign, and has triggered an important conversation about sexual assault. The backlash is, on the one hand, completely understandable. But some are wondering – why are we so surprised? Based on his public statements, isn’t it clear that Donald Trump is capable of saying things like this in private? Perhaps behavioral economics can help explain why we were so collectively shocked.

Let’s think about this from a “rational” perspective. In a "rational" model, people assess uncertain situations based on the information that is available to them. In the past few months, voters have wondered – is Donald Trump a sexist? Based on his public statements, is he the “kind of guy” who would say offensive things about women? In assessing this, people were making what’s known as a probabilistic judgement, whether they realized it or not. Based on what Mr. Trump has said publicly about women, a rational voter might easily have assigned a pretty high probability to the chances that Mr. Trump would, in private conversation, say something like what he said on that bus.

So let’s suppose that a rational voter thought, subconsciously, that the odds that Trump had said vulgar things like that in private at some point was, say 60%. Given that, how might this hypothetical rational person have responded to the tape? Well, the tape certainly increased those odds – right up to 100%. And this might lead some rational people to change their views about Trump. But should they have changed so dramatically? An increase from 60% to 100% is a big difference, but it’s hardly a major shock.

I would argue that the world is not full of rational voters – it is made up of people swayed by “behavioral” forces, which go beyond rational calculations of probabilities and likelihoods. First and foremost, psychologists and behavioral economists have long studied the idea of salience. That is, things hit us harder when they are right in front of our faces. In the days following the tape’s release, you could hardly watch a typical news broadcast for more than a few minutes without hearing audio clips from Trump on the bus. So the sheer ubiquity of the tape in the media upped its salience to the average citizen – and drove our strong response to it.

There is another behavioral principle that may be subtly influencing us as well, and it is one of the centerpieces of behavioral economics – prospect theory. Prospect theory deals with how we assess uncertain situations. One often-overlooked aspect of this is "probability weighting," whereby people are thought to (among other things) underweight high probabilities when assessing situations of uncertainty. For example, most newlyweds underestimate their odds of getting divorced (unfortunately a rather high probability event). When we get the probabilities wrong, it can lead us to act in less than rational ways.

Figure 1. The probability weighting function (in red) shows how when we make decisions we tend to overweight low probability events and underweight high probability (but not certain) events.

Figure 1. The probability weighting function (in red) shows how when we make decisions we tend to overweight low probability events and underweight high probability (but not certain) events.

There is, however, one exception to our skewed assessments of probability – 100% certainty, which people do tend to assess accurately. In fact, relative to nearly-certain outcomes, we seem to give an extra “boost” to things that are certain.

Take a famous example from Harvard’s Richard Zeckhauser (who was my PhD advisor, as it so happens!). Suppose you are playing Russian Roulette, using a gun with six bullet chambers. You know there is one bullet in the gun’s six chambers, but you don’t know where it is. You hold the gun to your head… but I stop you and say, “how much would you pay me to take one bullet out of that gun before you pull the trigger?” A lot! But suppose the gun actually had four bullets in it, instead of just one. Now how much would you pay to take a single bullet out? Not as much, right? This may seem reasonable, but it’s not rational. Either way, I’m offering you a chance to pay for a 16.67% increase in your chance of survival. But the first situation feels like a better offer. Why? Because I’m not just offering you an increased chance of surviving – I'm also offering you certainty along with it. And we humans love certainty.

So how does this apply to the Trump tape? Let’s think back to our pre-tape world, where we thought the odds that Trump would say vulgar things about women in private stood at 60%. If we were subject to prospect theory’s probability weighting here, we may have been underweighting this 60% probability. So even though we knew it was pretty likely he was a private sexist, we weren’t really sure, and so it was hard for us to react rationally to the (high) likelihood that he would say such things. In other words, before the tape, voters might have thought - "Is he a sexist ‘kind of guy?’ Well maybe he is, and maybe he isn't... 60%? 30%? Who knows?" As with the Russian Roulette gun loaded with 4 bullets, voters might not have been thinking deeply about the precise underlying probabilities. But when the tape came out, it raised a mere probability to a certainty, and that packed a powerful behavioral punch, beyond what it would in a world full of rational voters. And while we won’t know for a couple of weeks, it seems increasingly certain that the tape dealt Trump’s chances of winning a mortal blow.