David R. Henderson  

Sanctions and Game Theory

PRINT
Yes, Voters Will Sit on Their ... How Many Times Must We Bury Be...

Jacob Hornberger had an insightful blog post on Friday in which he applied basic game theory to the issue of sanctions. He didn't use the term "game theory," but he did the key thing that game theorists do: in thinking through various actions you might take to have an effect on B, ask yourself how you would respond to such actions if you were B. In other words, put yourself in the other person's shoes, not to get a touchy, feely, we're-all-in-this together Kumbaya moment, but to come up with your most-realistic analysis.

Hornberger wrote:

The Chinese government has threatened to impose sanctions on the United States if the U.S. government persists in its decision to sell weapons, including F-16s, to Taiwan. According to the New York Times, the threat was issued by a top Chinese military official, who did not specify what the sanctions would be. However, a possibility would be the wholesale dumping of U.S. government securities onto the international financial markets. Those instruments represent the enormous amounts of money that China has loaned the U.S. government to fund its enterprises in Iraq and Afghanistan.

Notwithstanding the U.S. government's steadfast insistence that its sanctions will induce Iranian government officials to submit to U.S. demands regarding its nuclear program, the U.S. government is steadfastly refusing to succumb to China's threat to impose sanctions on the United States.


Hornberger goes on to argue that if the U.S. government proudly ignores sanctions, why would Iran's government, whom no one has accused of having insufficient pride, act differently.

In a December 1996 talk I gave to a bunch of officials from the U.S. Department of Defense, I used such "put yourself in the other person's shoes" reasoning to explain how terrorists choose the countries to be terroristic against. Here's what happened next:

One person in the audience, noted game theory economist Martin Shubik, sarcastically accused me of advocating that "we all love one another." But he missed the point. A good game theorist puts himself in the shoes of the other person whether or not he loves him. Even if you hate your opponent, and especially if he hates you, it's good to know what motivates him and what pushes his button. Schelling would agree. Note that in the Schelling quote above, Schelling doesn't evince a lot of love for the burglar in his house: he just wants him to leave.

Comments and Sharing





COMMENTS (12 to date)
TobyBelch writes:

Would the dumping of securities really be considered sanctions?

But nonetheless, you don't need game theory to show that sanctions don't work; history is replete with examples of sanctions not working.

Snorri Godhi writes:

I am sorry to say that the 2005 article looks hopelessly naive to me. I realize that space constraints did not allow to develop the argument, but what about this:

"If you want to avoid acts of terrorism carried out against people in your country, avoid getting involved in the affairs of other countries."

Several questions come to mind:

1. In which way did the USA get involved in the affairs of Japan before Pearl Harbor?

2. To avoid getting involved in the affairs of other countries means to avoid all trade, foreign investment (either way), immigration, and emigration. Was that the suggestion?

3. Most important: the USA is already involved in the affairs of other countries. At this stage, the statement I quoted seems to imply: "the US response to terrorism should be to withdraw from the countries the terrorists come from." But such a policy would give an obvious incentive to terrorists. Is Dr. Henderson assuming that terrorists do not respond to incentives?

To put it another way, terrorists from country X do not attack country Y solely because country Y is involved in the affairs of country X: they attack because
a. country Y is involved in country X
AND
b. terrorists expect this involvement to decrease in response to their attack [NB: not necessarily as a direct consequence.]

After reading the 2005 article, one might be excused for thinking that Sun Tzu, Machiavelli, and the Sagas of Icelanders are a better guide than game theory. They are also more fun to read.

Redland Jack writes:

@Snorri

1) I would guess that they assumed that the U.S. was going to join the war on the side of the British (like we had the last time) and that the only hope the Axis powers had of winning was a preemptive strike. I think we were already tentatively supplying the U.K. with goods, though I could be mistaken.

2) Obviously not.

3) I'm not entirely sure that terrorists care about getting end 'Z'. I certainly don't know much about their true motivations, but it may be that they care more about getting and sustaining power, and that saying that they care about end 'Z' is the best method for doing this.

It seems likely to be true that terrorists are able to generate more followers when a foreign country is getting involved in their politics than when the foreign country is getting involved in trade.


Also, Game Theory is incredibly fun to read. Machiavelli and Sun Tzu are mediocre. (I haven't read the Icelanders)

Jim writes:

1. How about the oil and steel embargoes against Japan? Or the seizure of all Japanese commercial assets in the US?

2. It pretty clearly isn't. It'e more likely that David means to avoid governemnt interference in voluntary trade, investment, and migration across borders. The example he gives in the article, of relaxing immigration restrictions, strongly suggests this.

3. As you suggest, US withdrawal likely would produce such an incentive. So what? Why would incentivizing people to resist imperialism by foreign governments be a bad thing? If it is a bad thing would it not also be wrong to encourage Americans to resist foreign intervention or even invasion on US soil?

Snorri Godhi writes:

Redland Jack:

1) I would guess that they assumed that the U.S. was going to join the war on the side of the British (like we had the last time) and that the only hope the Axis powers had of winning was a preemptive strike. I think we were already tentatively supplying the U.K. with goods, though I could be mistaken.

So I was right in point [2]: you do need to stop all trade to be sure that there won't be an attack on your country. (And thank you for reminding me that, in ww1, it was trade that led the US to war.)

2) Obviously not.

With all due respect: I asked if that was Dr. Henderson's suggestion, not yours. In any case, if that was not his suggestion, then he did not follow his own argument to its logical conclusion, see above.

3) I'm not entirely sure that terrorists care about getting end 'Z'.

In other words, you do not believe that terrorists respond to incentives. But then, game theory is inapplicable.

It seems likely to be true that terrorists are able to generate more followers when a foreign country is getting involved in their politics than when the foreign country is getting involved in trade.

That sounds like wishful thinking, and again ignores incentives.

Also, what do you mean by "involved in their politics"? internal or foreign politics? are trade agreements politics? if country X nationalizes all American-owned assets, should the US interfere in their politics? if country X arrests all Americans and sentences them to death on trumped-up charges (dropping hints that the sentences can be reversed after payment of adequate compensation), should the US interfere in their politics?

Snorri Godhi writes:

if country X arrests all Americans

Sorry, I meant: "all Americans in the country", of course.

KomoriKiri writes:

The bombing of Pearl Harbor is an act of terrorism rather than war now? Who's writing the history books this year?

Snorri Godhi writes:

Komorikiri:
The bombing of Pearl Harbor is an act of terrorism rather than war now?

I didn't say it was. It was, however, a murderous attack on American soil. Whether it was committed by a state agent or a supposedly non-state agent, seems irrelevant from the point of view of the victims.

Jim: for your points [1] and [2], I refer you to my reply to Redland Jack.

As for this:
3. As you suggest, US withdrawal likely would produce such an incentive. So what? Why would incentivizing people to resist imperialism by foreign governments be a bad thing? If it is a bad thing would it not also be wrong to encourage Americans to resist foreign intervention or even invasion on US soil?

These are ethical issues, and therefore cannot be answered by game theory alone. I do not intend to discuss ethics here: I am interested in the consequences of the actions that Dr. Henderson recommends; whether these consequences are desirable, is best left for another discussion.

Jim writes:

Snorri,

Your response to RJ regarding your point #1 gets it exactly wrong. It was not trade, but the forcible restraint of trade by the US government that provoked the Japanese attack.

Regarding point #2, as I said, it's clear that this was not David's suggestion. All you've done is construct a strawman based on falsely conflating the government of a country with the private citizens of a country.

Nowhere in his article does David make this (your) error. In the quoted passage, he is explicitly advising Defense Department officials not to get involved in the affairs of other countries. Nowhere does he advise private citizens not to trade, invest, or migrate between countries, or the US government to interfere with their doing so, in order to avoid terrorism.

David R. Henderson writes:

Thanks, Jim. What I would have replied was very close to what you said, although you probably said it better. I checked my desire to reply instantly because I like to see whether other commenters are paying attention. You obviously were.

Snorri Godhi writes:

Jim (and David, since Jim answers for both of you):
as I said, my main point is point [3]. I note that your answer does not address the issue of incentives, and Redland Jack denies that terrorists respond to incentives. Both your answers make game theory inapplicable.

But let me address the other points:
It was not trade, but the forcible restraint of trade by the US government that provoked the Japanese attack.

Few people would agree with your implication that the US should continue trading with a country that is engaging in genocide.

What about ww1? what should have been the US response to unrestricted submarine warfare? "forcible restraints on trade" are out, so should the US just tell shippers that they trade at their own risk?

Regarding point #2, as I said, it's clear that this was not David's suggestion.

Yes, I thought so; my point was that leaving citizens free to interact with foreigners leads to all the issues connected with unrestricted submarine warfare. See also my question to Redland Jack: what do you mean by "involved in their politics"? and following text.

Snorri Godhi writes:

Apologies for my going on about this.
From Hornberger's article:
Recall the brutal sanctions that the U.S. government enforced against Iraq for more than 10 years. Every year, they were causing the deaths of thousands of Iraqi children [...] Those deaths didn’t cause Saddam Hussein to leave office, which is what the U.S. government wanted.

(Let's leave aside whether it was in fact the sanctions that caused those deaths.)
Iraq was almost continuously at war, from the time Saddam assumed full powers in 1979 up to the end of the Gulf War. I find the correlation between sanctions and Saddam not going to war highly suggestive. Correlation does not imply causation, but at the very least, the US should have been aware that removing the sanctions might lead to war in the Gulf. That might be the reason why the sanctions were not removed.
(However, if the point was that the Iraq war was the least bad alternative to the sanctions, then I tend to agree.)

Also, from an article David Henderson linked to previously:
David Edelstein surveyed the record of postwar occupations from the time of Napoleon to the present day and found that two thirds of all occupations failed.

A glance at my historical atlas suggests that the failure rate was lower in previous centuries. I find the correlation between a high failure rate and the presence of the Pax Britannica or Pax Americana highly suggestive. Again, correlation does not imply causation, but it is difficult to construct this high failure rate as an argument for terminating the Pax Americana.

Comments for this entry have been closed
Return to top