Wednesday, July 9, 2014

Tragedy of the Commons: Fertility Edition

At Quartz, Sarah Perry claims that having kids has become a Tragedy of the Commons. Because the value of children now accrues more to the general public instead of to the family that begat them, would-be parents have less incentive to procreate.


The mark of a good claim is that it can be tested empirically. In that sense, this is a good claim. I am in the habit of being a bit leery of monocausal explanations for complex phenomena, as I am sure you are, but any relative price story has at least one treatment effect. Here, there's an embarrassment of econometric riches. We have lots of natural experiments with transfer payment legislation, varying both geographically and over time (for panel regressions), we have big treatments like the 1996 welfare reforms (for discontinuity regressions), and we have plenty of historical data (for time series econometrics). There's lots to choose from.

But behind all that are McCloskey's ol' admonitions: why is this question important? Are the results meaningful (i.e. are the coefficients substantial, and not merely statistically significant)? Who is the relevant population? What's the relevant counterfactual?

It's always that last one that needles me when I hear claims about changes in rates (of any kind). I strain to imagine why it should be that historical rates were somehow closer to a Platonic ideal than what we've got today. I keep a bottle of aspirin handy for the headaches that inevitably develop from all the eye-rolling I do every time I hear some DC politician wringing his hands over "cuts" to this program or that, as if last year's budget was some magical cloud farted from the candy-coated buttocks of a majestic unicorn in flight aboard a helicopter made of chocolate and powered by pixie dust.

That expressly is not what Perry is doing, however. I won't claim that Caldwell is or isn't making this mistake here, but I will note that there's a big gulf between positive statistical analysis and public opinion and that stealth assumptions have a nasty way of sneaking into your empirical claims and drinking all the milk in your fridge, so to speak. Perry is making a straightforward relative price claim that's perfectly consistent with ordinary plain-jane workhorse Economics 101. The tragedy is when someone reads that and goes all Helen Lovejoy, setting about to rush hasty legislation through hot on the heels of an imagined crisis: "won't someone think of the (unborn) children?"

The point is, if I were inclined towards the economics of the family (and I'm not, since competing in the same arena as Justin Wolfers means my contributions would be rapidly and summarily overshadowed), I'd want to very carefully establish the following:
  1. Am I measuring the right thing? Should I care about fertility per se or about, say, a combination of fertility and mortality? Should I care about utility-adjusted life-years or something? Total aggregate discounted lifetime productivity maybe? How should any of this be measured? What controls should I include? Why?
  2. Are historical comparisons appropriate? Why or why not? 
  3. How strong is the treatment effect of state-provided goods (including in this case retirement benefits, education, direct transfer payments, and medical subsidies)? How does this effect compare to reasonable alternative hypotheses, like widely-available contraception, household composition effects, selection effects, et al?
  4. Finally, what are the public choice implications? 
Sadly, many authors behave as if there is no four. There's quite frequently a throwaway closing paragraph in most empirical papers outlining what the appropriate policy response should be, but these tend to be laughably naïve, invoking the mighty powers of the Sikorskean unicorn referenced above.

Tradeoffs abound, including in the maternity ward. I'm very much a member of the Julian Simon club: people strongly tend to be net assets rather than net liabilities, so I'm naturally sympathetic to the idea of encouraging fertility. Similarly, my priors lean towards private provision of private goods (including most education), with perhaps some public funding where appropriate. But careful analysis must account for the public costs of policy alterations.

I have a hunch that thorough research would produce some of the same results. Proving definitively that a tragedy of the commons actually exists is a grueling task, but it could be an interesting exercise. If nothing else, it'd be a nice way to needle both the left and the right at the same time, and that's gotta be worth something to someone, right?

Kids aren't euvoluntary because regret. Consider very very very carefully what this might imply for public policy. Consider also historical context and the undeniable fact that people get mighty squirrely when confronting questions of reproduction.

h/t The Peej

No comments:

Post a Comment

Do you have suggestions on where we could find more examples of this phenomenon?