‘Really Bad Accidents’: Playing The Strategic Trade War Game

I don’t know if you’re one of the ~53 million people who follow the unofficial official (and in Trump world, that’s not an oxymoron) Twitter account of America’s fearless (and also feckless) orange leader, but suffice to say he’s been on what amounts to a week-long tweet binge and quite a bit of the “covfefe” has revolved around his ongoing efforts to piss off every single one of America’s major allies by rubbing his trade war in their face.

That continued on Friday morning ahead of the G-7 summit which he’s apparently going to attend for about an hour before flying off to Singapore to have dinner with a murderous, nuke-wielding child dictator.

The lead up to the G-7 has been nothing short of surreal as Macron and Merkel attempt to craft a “strategy” for dealing with Trump who has clearly decided to adopt the Navarro doctrine after the truce Mnuchin struck with China last month rubbed the isolationist faction at 1600 Penn. the wrong way. Of course Macron and Merkel will find themselves frustrated because as noted in a satirical post I wrote about chimpanzees and NORAD on Thursday, there’s no way to know what Trump’s reaction function is.

Here’s what Trump has come up with over the past 18 or so hours re: the G-7…

Obviously, that is all kinds of retarded for all kinds of reasons and the idea of tweeting all of that before you even get to the summit is straight up incomprehensible, but I guess he would say it’s all “strategery“.

On Friday morning, donning some of his signature bright orange face paint and sporting a blindingly yellow wig, Trump regaled reporters with his thoughts on damn near everything under the sun. Here are some of the G-7 related soundbites:

Actually, it’s not clear why Russia needs to be there, because after all, they’re already represented by the President of the United States.

As far as it not “mattering what you call it” and how everyone is free to “call it whatever you want”, BofAML “calls it” this:

Stumbling into a trade war

That’s the title of the global letter that accompanies the bank’s latest Global Econ Weekly note and the gist of it is that while market participants in general are still optimistic that the worst case scenario will be avoided, BofAML isn’t as confident.

Generally speaking, the bank notes that because the U.S. has no way of knowing, ahead of time, how far other countries can be pushed, the potential exists for miscalculations if everyone turns out to be more recalcitrant than everyone else anticipated. That, BofAML says, is how “accidents happen” and it’s precisely what I was trying to drive home earlier this week when I wrote the following in a post about China’s reported willingness to buy more U.S. goods:

Not to put too fine a point on it, but that ain’t gonna cut it. Or at least not in Trump’s mind.

He’s not a “meet in the middle” type of guy and I’m not at all sure the above constitutes the kind of clean relent he’s looking for when it comes to claiming “victory” in this increasingly quixotic crusade.

At a certain point, he’s going to push his luck too far. It seems like Xi wants to figure out a way to pacify this moron, but there’s only so much the Chinese are going to be willing to do — or at least willing to do publicly.

Remember, Trump isn’t the only one with domestic political concerns. Xi wants to present China as a more open economy and he’s anxious to step into any void that Trump leaves by pushing the U.S. towards more isolationism, but it’s probably safe to assume that there will come a tipping point when Beijing simply tells Washington to go fuck itself.

With respect to the China standoff, BofAML writes that the stalemate has the potential to result in “really bad accidents”.

There is a risk that the US and China will escalate protectionist measures, imposing substantial costs on the two largest economies in the world. In this case the rest of the global economy will also probably suffer. But we hope that both sides will see what is becoming increasingly evident, that they each have considerable capacity to retaliate, and so both will lose in a full-blown trade war.

Yeah, fingers crossed, right?

But the bulk of BofAML’s analysis revolves around a simple strategic framework for gaming out outcomes tied to escalations and de-escalations. This a fairly lengthy exercise and rather than take you through the whole thing, I’ll just excerpt the visual and the explanation of the various nodes. To wit:

Figure 1 illustrates our strategic framework. There are two countries, 1 and 2, which take turns at either de-escalating (D) or escalating (E) the negotiations. If either country deescalates, the negotiations end. For simplicity, we assume that the negotiations always end after at most four rounds (two for each country). Country 1 (e.g., the US) goes first. If it de-escalates immediately, the status quo is preserved and so the resulting payoffs are zero for both countries. If instead country 1 escalates, country 2 (e.g., a major trading partner of the US) is faced with the same decision. If it de-escalates, it is ceding to country 1’s demands. Therefore the resulting payoffs are 10 for country 1, which gains, and -10 for country 2, which loses. But if country 2 escalates in response to country 1’s escalation, we get a “trade skirmish.” This imposes a cost of c1 on country 1 and c2 on country 2.

StratFramework

The third and fourth rounds of the negotiation resemble the first two, with one key difference. If both countries escalate again, the negotiation ends in a “trade war”. The incremental cost of a trade war is much larger than that of the first-round skirmish: it is 3c1 for country 1 and 3c2 for country 2. This means that the total cost of getting into a trade war, including the initial cost of the skirmish, is 4c1 and 4c2 for the respective countries.

They go on to “solve” that framework and again, that’s a pretty tedious logical exercise that I’m reasonably sure most readers aren’t interested in, but this excerpt from the list of summary bullet points that accompanies their backward induction is worth noting:

Uncertainty in our framework creates a 2/3 probability of a trade skirmish and a 1/6 probability of a trade war. When c1 and c2 are mutually known, there is zero probability of a skirmish or a war.

Perhaps most important, BofAML reiterates what JPMorgan’s Marko Kolanovic has noted on multiple occasions this year (most recently here) – namely that this is all about the reaction in equities and the read-through from that for the midterms.

And on that note, I’ll leave you with one last quote from BofAML’s analysis:

If a market selloff or a Republican loss in the midterm elections reveals that the costs to the US from trade protectionism are large (i.e., c1 is large), the US might de-escalate after an initial skirmish. But if the markets remain sanguine and the Republicans perform well in November, the US’ trading partners might be forced to make concessions (because they would have learned that c1 is small).

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.

4 thoughts on “‘Really Bad Accidents’: Playing The Strategic Trade War Game

  1. Problems with BoFAML’s model:

    (recognized by Marko): blackboxing the U.S. isn’t useful. As you’ve written before, U.S. economic health really only matters as it impacts opinion of Trump, and the model doesn’t account for that.
    Per your chimpanzee piece, Trump’s decision-making calculus isn’t an ordinary sort of rational, so game theory probably isn’t the way to go.

NEWSROOM crewneck & prints