Don’t Trust the Market’s Knee-Jerk Reaction to the Midterms

I'm hunkering down in my bomb shelter and waiting this one out.

Don’t Trust the Market’s Knee-Jerk Reaction to the Midterms

Politically driven trades can be justified, but they might backfire once postelection hype settles down

By James Mackintosh, WSJ

Nov. 8, 2026 5:30 am ET

Wall Street has good reason to worry in the wake of midterms. Jeenah Moon/Reuters

Take a breath. When it comes to trading politics, the truth is that no one knows anything. The scale of the selloff since Tuesday’s election shows just how widespread the belief is that the blue wave that washed Republicans out of Congress will also drown data-center development. Traders and investors think they know a lot.

The moves make sense on fundamental terms only if there’s going to be a near shutdown in data-center development. Shares in memory-chip maker Micron plunged 23% in two days, neocloud data-center operator CoreWeave was down 19% and chip makers Nvidia and Broadcom were both down 12%, amid an artificial-intelligence bloodbath.

There are good reasons for concern, although not because of Congress, which would need a lot of Republicans to vote with the Democrats to override the presidential veto. The problem for AI is new governors who promised moratoria on data centers, including some in states with a lot of construction planned, such as Texas and Ohio.

There are usually good reasons for politically driven trades. Yet they often prove catastrophic once the postelection hype settles down and political horse-trading begins—or even before, as with Friday’s big AI rally.

After an election, the consensus frequently swings toward an investment thesis so obvious that few could disagree—yet it ends up losing money, even if it eventually comes good.

Perhaps the most obvious trade ever, even in hindsight, was to buy defense stocks after Ronald Reagan was elected president. Not only was war raging again in the Middle East, where Iraq invaded revolutionary Iran two months before the election, but Reagan was an avowed military hawk who made higher military spending and the Soviet threat a large part of his campaign. On top of that, his promised supply-side reforms were exactly what the stock market wanted.

Sure enough, defense stocks soared, with the sector rising 10% in the week after the 1980 election to end the year up 24%. Despite, or perhaps because, it was so obvious, two months later defense shares had given back all their gains, and by March 1982 had almost halved from their high, as stocks more broadly were pummeled amid fallout from the U.S. recession.

To paraphrase H.L. Mencken, trading democracy works on the basis that investors know what they want and deserve to get it good and hard.

There’s an echo here of the “Trump trades” that became popular after the 2016 and 2024 elections. In 2016, the mistakes started with late-night traders driving futures limit-down before a placatory speech by Donald Trump led to a full reverse. Likewise, the S&P 500 was up on Wednesday after the midterms, before having its 15th-worst day since 2009 on Thursday as the mood shifted.

The clearest Trump trade was in 2024 on Trump Media & Technology Group, which has the trade right there in its name and its “DJT” ticker. It rose and fell almost exactly in line with the betting odds of Trump’s winning ahead of the election, and jumped 5% the day after. But it then crashed 20% in just over a week, and despite a valiant rally ahead of his taking office, halved in price within six months. It is now down 88% from the presidential election, having fallen almost a third in the two days after the midterms.

The next two most-obvious Trump trades were private prisons, set for a windfall in the form of millions of immigrants awaiting deportation, and the America First bet on U.S. stocks to beat Europe and the rest of the world.

Both trades followed the pattern of doing very well after the presidential election, only to give back their gains in the following months. Geo Group and CoreCivic, the leading private prison operators, rallied hard again recently but fell sharply after the midterms. At a vibe level it makes sense, given their unpopularity with Democrats, but a huge budget for immigration enforcement is already in place.

There are several ways the consensus could turn out to be wrong about data centers. Temporary building bans might be used as a negotiating tool to extract better terms from developers, compensating nearby residents but allowing construction to continue.

Data centers might just move to less-convenient locations farther from major cities, to states happy to take the business, to other countries—or, as Elon Musk tweeted after the election, perhaps even to space.

Data centers might be more expensive to build and operate, and so less profitable. But if demand is really there for AI use, they’ll surely be built somewhere.

The other potential mistake is to be right for the wrong reason. Perhaps the extreme selloff in AI stocks since the election is driven mainly by sentiment and the reverse of overcrowded positions, going much further than the politics justifies.

If AI has been a slowly deflating bubble since the summer, the election might just be the moment all investors recognized they had bet too much on the theme. They’re right to get out, but should have realized months ago.

I thought chip and data-center stocks ought to be a lot cheaper given the risks hanging over them. The new politics adds another threat—but is probably not as severe as the consensus trade suggests. When it comes to politics and the markets, no one knows anything.