The Calculus of Commitment

Eugene criticizes the overuse of the term ‘commit-phobic,’ since the suffix ‘-phobic’ implies an irrational fear, whereas for many men (the term is almost always applied to men) the fear of long-term romantic commitment is perfectly justified. Lots of interesting posts follow in the comments section; the award for best comment goes to Dan Simon.

Naturally, I tend to think about the subject in economic terms. The mating-and-dating process is an application of search theory. Just like someone searching for a job or a product, a person searching for a mate faces both (a) costs of searching and (b) a distribution of possible outcomes, some of which will be more satisfying than others. The optimal search strategy in such a context, it turns out, is to calculate a “reservation level” of satisfaction: a generous enough compensation package, a low enough product price, or in this case, a pleasing enough mate. Once you’ve calculated your reservation level, you keep searching until you find a mate who meets or exceeds it.

There are two ways you can botch the optimization. One is setting your reservation level too high. As a result, you’ll reject mates who should have been good enough, and you’ll likely spend too long on the market (thereby incurring unnecessarily large search costs). A person whose reservation level is definitely too high might justifiably be called a ‘commitment-phobe.’ But the commitment-phobe is easily confused with the perfectly rational searcher who simply has low costs of search or unusually high returns from finding a more compatible mate. Hence Eugene’s objection to the overuse of the term.

The other optimization error is setting your reservation level too low. As a result, you run the risk of accepting mates you should have rejected. Your search costs will be low because you exit the market quickly, but you’ll also suffer from a lower stream of returns from your low-quality match – not to mention the risk of getting divorced and having to re-enter the market again (with a “pre-owned” sticker attached). A person whose reservation level is definitely too low might be called a ‘commitment-phile,’ but for some reason this term hasn’t caught on. But given the divorce rate, it’s not hard to believe such people exist.

What causes optimization failures? The most obvious answer is mistaken beliefs about the parameters – e.g., underestimation of search costs, inflated notions of one’s own quality, and so on. But I suspect there’s also a more subtle source of error: poor understanding of probability theory. Suppose that, given one’s reservation level, the probability of finding a mate during one period of searching is p. It follows mathematically that your expected length of search is 1/p. Thus, if you stand a 2% chance of meeting an acceptable mate during each month of searching, on average it should take you 50 months, or just over 4 years, to find an acceptable mate. But there’s a wide distribution around the average. You might get lucky and find an acceptable mate in your first 6 months! And then you might think, “Wow, that was easy. Maybe my chances are better than I thought!” So you adjust your reservation level up; we might call this the “better shop around” effect. At the other extreme, you might get unlucky and find yourself still searching after 8 years. You might think to yourself, “This is harder than I thought; perhaps it’s time to lower the standards a bit.” So you adjust your reservation level down; this would be called the “settling” effect.

Here’s the rub: some adjustment to one’s reservation level is justified, to respond to learning and changed circumstances. But it’s difficult to sort out the signals. If you found an acceptable mate quickly, did you just get lucky, or had you defined acceptability too low relative to your prospects? If you’re taking too long to find a mate, have you been unlucky, or are you just being too damn picky? Given the incredibly small sample size – you’re the only you – it’s no simple matter to distinguish the good strategic adjustments from the bad.

Related posts here, here, and here.