Wednesday, February 28, 2018

How does fertility predict recession?

Posted:27 Feb 2018 07:56 AM PST (вчера обнаружил в фб, тут сокращённо)
There has been interest this week in the relationship between fertility behaviour and economic cycles, as picked up by the BBC and Financial Times, based on a NBER working paper by Buckles et al: Is Fertility a Leading Economic Indicator?

The answer essentially seems to be yes- for certain recessions like 2008 at least- there is evidence of a lagged effect of conception rates on economic slumps. Moreover, this is not an effect of increased rates of induced abortion- it does seem like this is an increase in the desire to reduce (or postpone) childbearing in response to economic turmoil.

Now, as full disclosure, this is somewhat inconvenient to me based on my previous arguments about our ability to use events like the Great Recession for causal claims. In particular, this tends to skewer the idea events like the recession were unanticipated- clearly if people are altering their behaviour prior to the event, there is some degree of anticipation.

So how do we explain this predictive effect? The basic model in Buckles et al. is essentially testing for the significance of lagged conception rates on GDP growth at an aggregate level. The most natural explanation would be through employment: spikes in unemployment are associated with a depressing effect on fertility on an aggregate level (although interestingly the effect is more mixed for females, both in Germany and the UK poor employment prospects for women tend to provoke motherhood). However, Buckles et al. find that falls in fertility generally tend to precede rising rates of unemployment- and that the large rises in unemployment are not manifested until after the recession hits. This does not therefore explain the relationship between fertility predicting recession

One interesting avenue is more general measures of economic confidence- Buckles et al. test this with measures of consumer confidence, purchases of consumer durable and housing stock purchases. In general, the anticipatory effects are weaker (although not eliminated) here, will fall in fertility coinciding with declines in these measures of confidence.
контрацепция и холодильники
Source: Buckles et al., pp. 36

This does seem to have some legs, and has somewhat been replicated in other research. Comolli (2017) for instance find an association between search terms for 'spread' (within the context of Italian Sovereign debt crises) and subsequent birth rates. That said, Buckles et al. do find that the decline in consumer confidence follows the fall in fertility, so we aren't quite at a full explanation.

To reiterate the points of some of my work on this area, even when we build a reasonable model of what we think should explain fertility behaviour from an economic perspective (including labour market status, wages, wealth stocks, local economic conditions) the persistent significance of crude dummies for pre- and post-recession birth rates indicates that there is something else at play which we can't quite explain: fertility animal spirits. Fertility within the context of macro-economic conditions remains a somewhat poorly understood phenomenon- it's not perhaps a surprise that we can't explain this predictive relationship if we don't have a fully fleshed out model in the first place.

No comments: