One chart that shows how worried the markets are about Donald Trump winning

Markets are worried about a Donald Trump presidency, writes economist Eric Zitzewitz. When news breaks that’s good for Donald Trump, the stock market falls; when he does poorly, markets rise.

One of the most dramatic examples of this happened on October 28, when FBI Director James Comey sent a letter to Congress saying the FBI was restarting its investigation of Hillary Clinton’s emails (that investigation wrapped up on Sunday). Within minutes, the stock market plunged:

The purple line shows the performance of the stock market on the day the news broke. And Zitzewitz, working with another economist, Justin Wolfers, did something clever: He superimposed the stock market chart with a chart showing how Hillary Clinton’s election chances changed on betting markets. As you can see, the two markets moved together: As bettors became more skeptical of Clinton’s election chances, stock market traders pushed stock prices downward.

This isn’t an isolated incident, Zitzewitz argues. You can see similar patterns the day of the first presidential debate — which Clinton was generally viewed as winning — and Comey’s Sunday announcement that he was wrapping up his investigation of Clinton’s email scandal. In both cases, this good news for Clinton was followed by stock market rallies.

And that’s not all; Zitzewitz and Wolfers have found that a number of other market indicators also saw significant shifts during significant events. The Mexican peso has been particularly volatile, presumably reflecting the expectation that President Trump would try to reduce trade between the United States and Mexico. That would likely hurt both economies, but as the smaller and poorer country, Mexico is likely to be hurt more.

By extrapolating from these relatively small changes, Zitzewitz estimates that the stock market under Hillary Clinton would be 12 to 16 percent higher than the stock market under Donald Trump. Markets currently expect a Clinton win, so when that is confirmed, we’re likely to see only a small bounce. In contrast, a surprise Trump victory could trigger a double-digit decline in stock prices.

However, there are some reasons to treat these estimates skeptically.

A big one is that it’s never possible to fully isolate the effect of any single event. At any given point in time, a lot of things are happening around the world that could affect stock prices. So the fact that stocks moved at the same time as a particular event happened doesn’t prove that the event caused the markets to move.

It’s particularly hard to measure stock market reactions to events that play out over a long period of time. For example, Trump’s improving poll numbers over the past week have probably been a drag on the stock market, but it would be a mistake to attribute the significant stock market declines we saw last week and the week before to Trump’s improving fortunes — there was just too much else going on during the same two-week period to be confident that Trump was the primary factor.

This is what makes big, unexpected announcements so valuable to economists. Because if a big announcement is followed by a big, immediate change in stock prices — and no other major news was announced at the same time — it’s reasonable to attribute the stock change primarily to the announcement.

Still, how much faith you have in this method depends on much faith you have in the stock market’s ability to accurately price risks in real time. The orthodox view in finance is that markets are “efficient,” meaning that prices adjust quickly and accurately to reflect new information. If that’s true, then we can extrapolate from these short-term movements to answer questions like, “How much would a Trump presidency reduce stock values?” In this case, Zitzewitz estimates that markets would be lower by double digits.

But skeptics argue that this gives markets too much credit. They often overreact to events, and so a big initial drop might reflect panic among traders instead of a rational assessment of the implications. That view doesn’t mean a Trump presidency wouldn’t tank the stock market; it just implies that we shouldn’t read too much into these kinds of short-term jumps.