Subscribe to Annie's Substack

Annie Duke for Gen@Medium: Why We Assume What Went Wrong in the Last Election Will Go Wrong Now

Why We Assume What Went Wrong in the Last Election Will Go Wrong Now

The Democrats’ obsession with the Rust Belt doesn’t make sense in 2020

Annie Duke – October 28 2020 · 6 min read

Read full article here.

On December 22, 2001, Richard Reid boarded American Airlines flight 63 from Paris to Miami. During the flight, Reid took a match to his shoes in an unsuccessful attempt to ignite the explosives hidden in them, earning him the infamous moniker, “the shoe bomber.” Reid’s attempt to bring down an airliner with his shoes may have earned him a life sentence, but it also earned the rest of us a lifetime of taking off our shoes every time we go through airport security.

That X-raying our shoes is still a thing 20 years after Reid boarded that flight offers a classic case of the last disaster bias, the tendency people have to overprepare for the last failure they faced. No matter how unlikely it might be that someone else would try to bring down a plane with a Birkenstock bomb or how small the chances are that the attempt would be successful, we will go to outsized lengths to avoid a repeat. The result? As a society, we’ve wasted a mind-boggling amount of time and money taking off our shoes and having them X-rayed before every flight.

The same dynamic is at play in the 2020 election, as Democrats fixate on what went wrong last time as they consider what is most likely to go wrong this time. For the Democrats who backed Clinton in 2016, Pennsylvania, Wisconsin, and Michigan are their shoe bomber. Those are the three states that successfully brought down the Clinton plane. That unexpected outcome has created palpable anxiety about a repeat loss in the Upper Midwest in 2020 that appears to be driving some campaign spending decisions that are otherwise hard to explain.

As of October 13, the Biden campaign and its allies had spent a combined total of $282.8 million on television advertising in Pennsylvania, Wisconsin, and Michigan. The $184.2 million of that allocated to Pennsylvania and Wisconsin doesn’t seem at all unreasonable given that Pennsylvania is the tipping point state and Wisconsin is second in line for that honor. Plus, they carry a combined electoral prize of 30 votes.

But it’s harder to make sense of the campaign’s massive spending in Michigan, where Biden and his allies have spent $98.6 million, a whopping number especially when you consider that they have allocated a mere $8.4 million to Ohio and $4 million in Georgia. The spending in Ohio marks a complete reversal from 2016 when Clinton spent $52.4 million trying to win the Buckeye State and only $3.3 million in Michigan and lost both.

You can’t explain that reversal in spending by looking at the public polling: according to FiveThirtyEight, Biden’s polling lead is currently over eight points, an advantage that has been steadily outside the margin of error since mid-May. Since that lead has been consistent since spring, we can’t say that it’s thanks to the Dems’ big ad spending. Meanwhile, in Ohio and Georgia, Trump and Biden are in a dead heat and have been since April. You can’t explain the difference as proportional to the electoral college votes up for grabs: Georgia and Michigan both carry 16 electoral college votes, while Ohio is the bigger electoral prize, with 18 electoral votes.

You can’t even explain the difference by invoking a desire of the Democrats to retake the Rust Belt since Ohio is, of course, a Rust Belt state. Add to the mix that spending in Ohio would also spill over into Western Pennsylvania and spending in Georgia would spill over into Northern Florida, both states that really matter, and the pattern gets even more perplexing.

Of course, campaign spending decisions are complex and there may be some non-obvious explanation for the Democrats’ spending just 12.5% of the resources in Ohio and Georgia than they’ve allocated to Michigan, a 180-degree departure from their 2016 allocation decisions. But one obvious explanation for the disparity is a simple one: In 2016, Clinton was expected to lose Ohio and Georgia by a lot and win Michigan pretty comfortably. The shocking, narrow loss in Michigan is the Democrats’ last disaster, and they may be irrationally over-allocating their resources to Michigan, to avoid a repeat performance, rather than to states like Ohio and Georgia that are less comfortably in Biden’s column.

The consensus explanation for why we overprepare for a recurrence of the last disaster is that we remember the pain of losing the last time so vividly, making us overestimate the chances of a repeat. This is related to a cognitive error called availability bias, where the easier something is to recall, the more common we think it is. Availability bias is why we overestimate the chances of dying in a terrorist attack and why if you go to the beach you are more worried about the dangers of a shark attack than death by a falling coconut, which is, in reality, 15 times more likely.

But the last disaster bias is not just a result of overestimating the frequency of infrequent things. It is also that we simply would feel stupid if we failed the same way twice. As the aphorism goes, “Fool me once, shame on you. Fool me twice, shame on me.”

The last disaster bias is a defense against the shame of being fooled twice.

The thing is Clinton’s stinging loss in Michigan in 2016 shouldn’t be driving spending decisions in 2020. Just because the Democrats lost Michigan last time doesn’t necessarily mean that Biden can’t count on the public polls in Michigan or that he, like Clinton in 2016, shouldn’t be swinging for Ohio or Georgia. The campaign’s goal should be to make decisions that win the election based on how it’s going this time, not based on a fear of looking foolish for making spending decisions that resemble Clinton’s losing bid.

Not every disaster is as likely to repeat as our brains make us think.

As with all cognitive biases, the last disaster bias doesn’t always steer us wrong. When an unexpected disaster occurs, sometimes it alerts us to both the possibility and the seriousness of an event that we hadn’t considered before. If our ancestors experienced an unusually long and harsh winter, they were likely to be better prepared to survive the next one. If the tribe got sick from eating a certain berry, they would be more likely to survive if they avoided those berries in the future.

Asia’s pandemic response offers a modern-day example of how the last disaster bias can steer us to a good outcome. As the pandemic swept the world, countries such as South Korea and Taiwan that took the brunt of the 2003 SARS outbreak were much better prepared for Covid-19 than those countries that largely avoided its devastation. The Asian countries hardest hit by SARS had more hospital beds per capita, a larger stockpile of PPE, and better testing and tracing capacity than most western countries. In the case of Covid, the last disaster bias saved a lot of lives in Asia.

Of course, not every disaster is as likely to repeat as our brains make us think. And that can make us look too much to the past and not enough toward the future in deciding where we want to spend our resources. The key is to figure out more objectively whether guarding against a recurrence of the last disaster actually makes sense. Even if there’s some chance that some other Richard Reid is going to try to take down an airliner with a shoe bomb, you have to make sure the resources you spend defending against a long shot wouldn’t be put to much better use elsewhere.

There is always going to be some chance that the polling is way off in Michigan; at least one GOP polling firm seems to think so, giving Trump a narrow lead. And it’s possible that Biden has information from internal polling that makes this all make sense. But from the outside looking in, the massive spending in Michigan like it would not be happening but for 2016. The Democrats’ last disaster may be causing them to play too much on the defense to avoid an unlikely path to a loss, creating lost opportunities in other states that might have opened more paths to a Biden win. While the Democrats are playing offense in the final stretch by going after Trump in OhioGeorgia, and even Texas, one has to wonder if the last disaster bias made them wait until too late in the game.