Failures of Marginal Coercion

One bit of economic theory that doesn't get talked about is marginalism.  That is, that the value of things and how they affect our behavior is usually based on how much adding or subtracting a little bit from our current circumstances changes them.  The classic examples is that we put a much higher price on diamonds than on water even though the total value we get from water is much higher than from diamonds since we'd quickly die with no water.  But we have lots of water and few diamonds so we get more value from one more diamond than one more cup of water, and thus we put a higher prices on the diamonds despite water having a higher total and even average value.

The same can apply when people are trying to incentivize different behaviors.  A while ago my state was considering a carbon tax which would raise the price of gasoline among other things.  But all the money it raised would be distributed evenly back to the state's tax payers.  My coworker was confused as to how this could change people's behavior, if everybody ends up with the same amount of money after they pay the tax and then get the money back, on average, how could it have a large effect on behavior?  The answer is thinking on the margin.  If everybody drives every day and you are the one person who decides to start biking to work you'll get almost the full rebate but now avoid paying the tax.  If you're the last person still driving you're paying the full tax but barely getting any rebate.  In every case regardless of the averages switching your behavior will have a financial incentive attached to it.  The real world has a lot of complications in it around how people respond but none that mean they aren't mostly thinking on the margin when changing their behavior.

This applies to threats of violence too.  If a mafia enforcer comes by and demands that you pay your debts or he'll break your knee you have to contend with two marginal factors, that you like money but you like having intact knees even more.  But weighing the value of money versus the value of knees evenly requires trust in the threat.  If you think there's a chance that the thug might not follow through than that increases the temptation to not pay up.  And if you think the thug is a psycho who will break your knee anyways then you might decide keep it and recuperate in the hospital with the money rather than without.

This sometimes leads to a principal agent problem.  If one group of soldiers is fighting with another and the former decide to throw down their weapons and surrender sometimes their opponents are tempted to avenge their fallen comrades and just gun them down.  But that leads to fewer surrenders in the future, maybe to different groups of soldiers.  So for reasons of both wisdom and morality the soldiers' superiors will usually punish them for this.

In loosely knit or underground groups this can be a problem.  If a terrorist group is trying to achieve some political concession through coercive attacks then its easy for them to convince the government they're targeting of their will to violence: they just bomb things.  But they also have to convince the government that they can and will stop bombing for a compromise instead of total victory.  People on the terrorist side might be acting more from blood lust or anger than calculation and might not want to stop.  For this reason terrorist groups that actually achieve something often act like the IRA and periodically enact ceasefires or truces to show that they can actually stop the violence when they get a deal.

Not all terrorists are trying to coerce a settlement.  Some, like the Tamil Tigers are trying to provoke reprisals in a cycle of violence that leads to a civil war they assume they will win (note: the Tigers lost).  They aren't trying to apply coercion at all and so don't care about the ability to accept victory and stop the violence.  And if operation security means adopting a cell structure where nobody has the authority to stop the violence, they can do that like Al Qaeda did and not lose anything they care about.

But strange need for trust if coercion is to be effective bears on the current third Gulf War in a couple of ways.  The first is that by killing Ali Khamenei in the opening stage of the war the US made Iran much less able to stop shooting back.  Before the war Iran had let it be known that if their leadership was killed decisions about firing missiles would devolve to local Revolutionary Guard commanders.  By holding up the prospect of unstoppable violence Iran hoped to dissuade their opponents from attacking their leadership, a strategy that seems to have failed.  At this stage some central authority in Iran seems to have enough control to let certain oil shipment through the straits without being attacked so maybe they have the ability to accept a peace deal.  But with lesser standing Mojtaba Khamenei, the new leader, is going to be less able to force his people to accept a peace when they're angry than his predecessor was and so will make getting to that peace harder.

And on the other side is Donald Trump, and the question of whether Iran can trust him enough to make it worth their while to accept a peace deal.  Starting this war by surprise in the middle of negotiations will tend to make the Iranians more skeptical that any future negotiations are in good faith.  But more importantly it doesn't seem to be clear to anyone what exactly the US's goals are in this conflict.  From the Iranian perspective they might worry that they could agree to, say, giving up all their uranium for peace but then later have Trump decide he wants to overthrow the government and re-start the war once he's received the uranium.  And on the other hand they might give up their uranium when if they had held out for a few more weeks Trump might have gotten frustrated and given up.  Just like an unreliable mafia thug might tempt a debtor to keep his money by weakening the link between payment and intact knees, we've made our bombing less coercive without decreasing the amount of pain it inflicts.