On the 1st January, 2007, Germany made a significant change to the payments it made upon the birth of a child. Prior to that, they had a means tested payment of between 5,400 and 7,200 Euros paid over 12-24 months. After that date, a payment of 67% of the mother’s pay would be made for 12 months with additional bits if the father took time off too. Of course, it was capped and so the payment was between 300 and 1800 Euros per month for 12 months. Some would actually lose under the scheme but working mothers, especially those in higher paid jobs, would gain.
So there likely existed families who would gain 25,200 Euros if they had a child in the January rather than December. Not surprisingly, German news reported lots of attempts to delay births in the last week of 2006. This is, of course, familiar to Australians where similar things occurred for just a $3,000 marginal payment. Now, the hard data is in and a new paper by Marcus Tamm looks at it using the same methodology as I did with Andrew Leigh.
He finds that about 8 percent of births were shifted as a result of the policy change and he has more data on the characteristics of the babies and mothers and so could observe shifts of 13 percent for working mothers. Interestingly, that is less than the 16% shift we observed in Australia but the policy change and its application were far more transparent. In addition, the German payment varies with time off work, etc., so while for some the benefits would be salient that isn’t otherwise the case. Nonetheless, as in the Australian case, there is statistical evidence that some births were delayed over 2 weeks from mid-December.
More interesting are the findings on health outcomes. Tamm confirms that birth delay appears to cause higher birth weight just as it did in Australia. But he also examined the impact on the probability of still births (something we could not do due to a lack of sufficient data). The proportion of still births in the first week of January were 50 percent lower than expected (a statistically significant result). So delay was associated with an improved health outcome on that front. I expect that the numbers on still births are pretty low but this was enough for Tamm to report it but the mechanism for this is unclear and let’s face it, controversial as the whole thing here was a deviation from ordinary medical practice. This confirms what I have long suspected that, whatever else these policy timing issues might be, they are an opportunity to study what happens to health outcomes in the face of non-medical variation in timing and in this case, duration of pregnancy.
Are you sure the still-births finding is a health outcome? It might just reflect a change in reporting, as the definition of still-birth can vary depending on when exactly the baby died. Might there be some incentives in the new German scheme that increase payments for babies who are born live but die, as opposed to babies who are stillborn? (I note that the Australian baby bonus was paid to all birth outcomes, presumably to avoid nasty media stories about governments profiting from tragedy.)