Site Meter

Sunday, November 01, 2009

Commenting on Brad's Links to Save His Time

Brad DeLong reads very fast. He also writes fast, but he regularly provides links without comment to save time. His time is valuable. Mine less so (my serious work computer is not working -- it's not my fault -- all I can do today is blog until I get some serious work software installed (all with legit site license) on this computer Monday -- sigh).

Brad's links

Justin Fox discusses very special pleading by Jeremy Siegal.I have already noted my agreement with Fox in a comment over there.

Free Exchange: Here's your recovery A solid sensible article with a valid important theme "misguided deficit fears have probably made aggressive fiscal moves politically difficult," but concludes with 2 weird lines "a populist squeeze on high earners. That's a dangerous direction to travel."

Other aspects of the dangerous direction were listed before the squeeze on high earners, but what evidence is there that "a populist squeeze on high earners." is "a dangerous direction to travel" ? I know of none. The last time high earners were squeezed (in 1993) the economy did fine. Events following removal of the squeeze are evidence that not squeezing high earners is dangerous.

Perhaps The Economist's correspondent is convinced that a populist squeeze will have a different effect than a technocratic squeeze. I wonder how that would work ? Oh and populist anger was very strong in 1993 (During a debate Bush supporters pushed the "I agree" button when Clinton said that only the rich got tax cuts under Reagan and Bush -- this doesn't happen often). It seems to me that the journalists at The Economist just can't give up their non evidence based opposition to income redistribution.

Menzie Chin on Mulligan. Chin notes that Mulligan assumes that flash data are accurate and will not be revised (not a very strong criticism -- they are what we have and Mulligan should just have warned that they will be revised) *and* Mulligan assumes that the stimulus had no effect on GDP. Mulligan believes this, but he is claiming that new evidence supports his prediction. He can't assume that he is right when discussing which theories fit the evidence best. He simply contrasts what other people predicted would happen without a stimulus to what happened with a stimulus. The data may have refuted the predictions of the gloomiest Cassandra, but Mulligan presents no evidence that any actual human being made an innaccurate prediction.

Intellectual garbage pickup indeed.

The Economist on Casey Mulligan

REMEMBER when Casey Mulligan said that it was perhaps not a good thing, economically speaking, to give women in developing nations the ability to control family size, because larger populations increased the incidence of innovators, thereby boosting economic growth?

Well no I don't. I will note that there is a very strong negative correlation between fertility and per capita GDP growth in cross country regressions I have seen (t-stat of - 6 or so). This is not (pace The Economist) new evidence. It is also not relevant, because Mulligan is assuming that innovations are available to all countries and so only has a prediction of world GNP growth as a function of world population. He is probably looking at growth since 1,000,000 bc. He should know that for most of that time real per capita GNP probably was roughly fixed (Malthus was right about economic history up until the time of Malthus) and there was no formal schooling. Data dated BC is not relevant to the 21st century.

Mulligan is working from a model in which there is no formal schooling and in which innovation depends only on total output not per capita output. These are ultra stylized models and using them for policy advice is absurd. It is easy to write models in which higher fertility causes less innovation (you need to assume that R&D is human capital intensive and so relatively expensive if there is a low ratio of human capital to raw labor, plus you want the innovations to substitute for raw labor and so be less valuable (compared to GNP) if the marginal product of raw labor is low).

Actual observed innovations mainly come from rich countries. This is strong evidence that human capital per capita has a lot to do with the cost of R&D and that wages have a lot to do with the benefits. Profits from innovation don't have all that much to do with population or GNP in developing countries, given how much those countries import. Mulligan considers very simple models not designed to address the question to be worth as much as actual empirical evidence (as usual).

Brad admits that he was totally wrong about Ross Douthat and calls "uncle." I think he should ask for a Mulligan.

1 comment:

Hans Suter said...

Usually dead computers produce nothing which often is positive. Now here a dead computer produces something which is very positive.