Accurate political predictions are hard to make and rare to come by. There are a few predictive models that work, but even they have potentially fatal flaws. FiveThirtyEight predicted that Hillary Clinton would win the election 302 electoral college votes to 235. She lost 304 to 227. Alan Lichtman’s Keys to the White House model correctly predicted Donald Trump’s victory. But foreign election interference would shatter that model. Even professional political predictions can’t predict everything.
The Black Swan And Political Predictions
That’s one of many critical takeaways from Nassim Taleb’s book, The Black Swan. Taleb describes unpredictable events that have disproportionately negative effects.
For example, 9/11 was a Black Swan event. Terrorists using a commercial plane to attack the United States was previously unthinkable. But think about the monumental changes that came from the 9/11 attacks. The Iraq War, the Patriot Act, and the TSA screenings at airports were only some of the major changes that were part of the reaction to 9/11. It was a single event with a greater impact than any crime that law enforcement could predict.
In contrast, the 2008 financial crisis and the COVID-19 pandemic were not Black Swans. The 2008 financial crisis was the foreseeable consequence of widespread exposure to risky investments that were rated as far safer than they were. The pandemic was the foreseeable result of a globally interconnected world and continued pressure on the environment. They were both foreseeable disasters lying in wait rather than outlier events that were completely unaccounted for. (Taleb discusses these issues in more detail in one of his other books, Antifragile, a book worth blowing off work to read in one sitting.)
In The Black Swan, Taleb devotes one chapter each to five reasons we make ourselves vulnerable to unpredictable events:
- Confirmation Error
- Narrative Fallacy
- Ignoring Black Swans
- Silent Evidence
While these are not the only ways we blind ourselves to unpredictability, these are five fatal flaws that are still easy to watch for in political predictions.
Trap 1: Confirmation Error
Readers may be familiar with confirmation bias, the cognitive bias this error alludes to. Confirmation bias is when someone finds evidence that they’re right and ignores evidence that they’re wrong. It’s the prediction that sees the outraged liberals in a swing state but misses the low turnout from those same blindly confident voters.
Political bettors should watch for predictions that only provide evidence, but don’t say what would disprove their predictions. These predictions are vulnerable to being destroyed by Black Swan events and are potentially worth betting against.
Trap 2: Narrative Fallacy
Humans are storytellers. The narrative fallacy is the human tendency to tell great stories to avoid facts that we don’t like. Taleb opens his chapter on the narrative fallacy with an anecdote about it being applied to him.
An Italian researcher approached Taleb after his presentation on ideas related to the narrative fallacy. After agreeing with Taleb’s ideas, the researcher then committed the fallacy himself. The researcher attributed Taleb’s view of the outsized effects of unpredictable events to not growing up in a “Protestant society where people are told that efforts are linked to rewards and individual responsibility is emphasized.” Taleb defends his worldview with hard data, which makes his Lebanese heritage irrelevant to the researcher’s argument. It’s an untrue belief backed up by a reasonable-sounding (to the researcher) story.
When reading political predictions, don’t let the same columnist, organization, or predictor explain away past failures with anything but sincere introspection. If they didn’t see a major event coming, then they should say so. They shouldn’t make up intellectual stories about how they were half-right and saw parts of an unknown event coming.
Trap 3: Ignoring Black Swans In Political Predictions
Hope can be inspiring, but it can also be used to remain blind to unwelcome events. A 2016 opinion piece in the Guardian newspaper predicted that Trump’s rise could lead to a “more progressive era.” The Republican presidency, House, Senate, and three conservative Supreme Court justices did not usher in a new progressive era.
This is the narrative fallacy in reverse. It’s an attractive branch of history for the Guardian’s opinion readers. But it puts blinders up that prevent events that could doom that history from seeping in. Political bettors must watch out for those kinds of blinders when they consume political predictions.
Trap 4: Silent Evidence
Knowing how many times a political prediction has been correct isn’t enough. Political bettors also need to know how many times a political prediction has been wrong.
For example, Lichtman’s Keys to the White House has correctly predicted elections from 1984 to 2020, except for 2000. Since he developed the model in 1981, the Keys model has had a 90% success rate.
There are some limitations. 10 elections is a small sample size. It only seems to work in modern elections. So, a fundamental political shift could make this model obsolete. But it’s a vast improvement from over-relying on columnists who don’t have built-in scorecards.
Trap 5: Tunneling
Being on the other end of a Black Swan event can be traumatic. 9/11 remains a tragedy seared onto the American psyche. Adding TSA checkpoints in airports and locking cockpit doors on planes will certainly prevent another commercial airplane from being used as a weapon against the United States.
The tunneling trap is focusing so much on preventing a Black Swan that’s already happened that the next one is missed. A hypothetical example would be focusing so much on preventing another 9/11 that we miss the makeshift nuclear bomb being set off in a major city. (This is hypothetical because there are actually guidelines for a dirty bomb scenario, and it continues to receive attention.)
Political bettors shouldn’t get so focused on correcting specific mistakes that they miss other mistakes that they’re making. For example, bettors who got Trump’s election wrong shouldn’t underestimate him in the presidential primaries based solely on proxy battles in Georgia. Leading up to the primaries, in-fighting is expected. Donald Trump isn’t the only Republican who wants to be president.
Don’t Rely On Political Predictions, Profit From Them
By definition, no model can account for the world’s complexities. Even worse, they can’t predict outlier events that can turn a model on its head. It’s why opinion polls are junk, polling can go wrong, and why many political predictions fail. Since political predictions are often unreliable — especially months or years out from a race — a better way to look at political predictions is to try profiting from poor expectations.
PredictIt’s prediction markets are a good place to watch how market shocks create profitable opportunities. The Roe v. Wade leak made users wonder whether Democrats could secure the House or Senate in the midterms. The prices fluctuated and then returned near their original prices within a few days. The markets had a moment of panic, then returned to normal. These irrationalities create opportunities to make small amounts of profit. (PredictIt won’t allow a trade that makes a user more than $850.)
Political bettors may be disappointed that political predictions are so unreliable. But they should take heart. There’s an unquantifiable chance that the most interesting thing in this election cycle hasn’t happened yet.