President Trump apparently believes that his tariff threats “succeeded” in getting the Mexican government to take more vigorous actions to block the flow of Central Americans looking for refuge status in the United States. The President also made it clear that he will continue to use economic threats in the future to win concessions on trade and immigration policy, two items that he believes are critical to his re-election. Can this work for him politically?

My answer is maybe. A political win does not necessarily require major policy achievements, although the latter can certainly facilitate the former. From a purely electoral perspective, a visible effort to deliver on campaign promises usually beats not trying, and winning even minor concessions is typically better than ending up with nothing at all.

To be sure, Trump’s crisis-generating MO does not seem like a carefully calculated strategy. It expresses his deep frustration with obstruction by the Democrats, judges, media, and even a few Republican lawmakers. Crisis creation is an addiction that the President does not want to quit. After all, it keeps him in the spotlight, satisfies his urge to act, distracts his opponents from other problems, and succeeds just often enough to encourage him to do more.

But given that practically everyone knows that this is Donald Trump’s MO and that he usually portrays any outcome, however modest, as a “beautiful” victory, why does his bluffing work? A poker player who bluffs too much usually loses the credibility necessary to convince other players to fold. Will this happen to Trump?

In the language of Game Theory 101, Trump has transformed trade and immigration issues from a cooperative into a zero sum game. Whereas President Obama and most economists believe that nations can do better by working together rather than pursuing narrowly defined national interests, President Trump sees economic rivalry as a game of chicken. It is all about winning the deal, not accommodating mutual interests. And when it comes to playing chicken, it can be perversely advantageous to seem irrational. When the two cars converge towards one another, the rational person swerves and lets the madman win.

In this latest crisis, the President made it clear that he was willing to sacrifice the economic interests of the U.S. auto industry, farmers, and others in order to force Mexico to take the Central American refugee crisis more seriously. In the earlier NAFTA negotiations, President Trump could at least argue that he was trying to get a better economic deal for his domestic supporters.

This time, however, he not only put the recently negotiated USMCA at risk, but he held some of his supporters economically hostage in order to get the immigration concessions he wanted. This sent a pretty clear signal to the Mexicans that he meant business, and that he was willing to absorb any domestic criticism that came with this strategy. If he could prevail in the negotiations, he might reasonably expect that most of his “hostages” would forgive him by the 2020 election.

But it is one thing to bluff the Mexicans and another to bluff the Chinese or even the North Koreans. The Chinese hold some strong economic cards, including a large amount of U.S. debt. And the North Koreans have their own “madman” who has shrewdly parlayed his country’s nuclear missile threat into political leverage. It will likely be much harder to get any face-saving concessions out of either.

So this gets us back to the electoral merits of trying to implement the MAGA platform versus not seeming to try at all. For voters who resent the coastal elites, who feel left behind by the economic recovery and high tech boon, or who fear that United States has been invaded by immigrants and sold out to the Chinese, the President, for all his flaws, is their guy. As for the rest of the Republican coalition, they go along to get along. Congressional Republicans and Wall Street may signal their anxiety with each new Trump crisis, but the market quickly recovers and the Members nod their heads in agreement when Trump declares victory.

Elections are about comparisons: The question in 2020 will be Trump compared to whom. If people believe that there are multiple crises, they may want someone strong to lead them out of the mess. Will the Democrats nominate anyone who seems stronger than Trump? Joe Biden’s retreat on abortion, which made sense in terms of the internal politics of the Democratic Party, did not help him seem stronger. And we shall see in the coming months how the others fare.

Then there are the election fundamentals: the state of the economy, involvement in unpopular wars, the relative balance of partisan strength, the length of time one party has been in office, and Presidential popularity. Many of these favor the Republicans. The economy has been a plus for the President so far. Tax cuts have extended the duration of the “Trump bump,” even as it bequeaths us a bigger deficit that will loom over us for many years after the 2020 election. Good for Trump, but bad for us.

For all his many flaws, the President is a pacifist at heart, so we are not currently as deeply mired in unpopular wars as we have been in recent years. The Democrats have an advantage in the generic vote, but the public tends to give first term Presidents the benefit of the doubt when they run for reelection.

The biggest negative against the President is his approval rating and highly flawed character.  Even so, the President’s job approval numbers have not varied much during his time in office: More people do not approve of him than do, but the margin is tight enough to shift. Assuming he cannot replicate the Electoral College victory without winning the popular vote, President Trump will need to win over more independent voters than his Democratic opponent.

If the economy turns dramatically down and the Chinese continue to call his bluff, the Trump bump could become a Trump slump leading to an electoral defeat. But if neither happens, and the Democrats cannot hold together, the President may finally get the popular vote victory he was denied in 2016.