Why should we aggregate income comparisons by country (or the whole world) rather than by city? World history looks very different if we do the latter. Aren't most countries relatively recent inventions anyway? More generally, I would like a more disaggregated look at the data. Big chunks of the urbanized human population -- pre 1800 -- seem to have violated Malthusian precepts for centuries on end.
I think that this is a very important question, even though it gets more interesting for the later part of Clark's book (Cowen's post only covers the first part). When you look at country data, the process of economic development seems to be very sudden. One of Gregory Clark's arguments is that England industrialized within a very short span of historical time. Countries that developed later developed even faster.
My point is that when you look at the data from a country perspective, it appears that there are two modes of economic development: imperceptible improvement (what Clark calls the Malthusian trap); or rapid, accelerating gains in the standard of living.
Why do we observe this pattern? Possible explanations:
1) It could be an artifact of looking at country data. Perhaps if instead of grouping people by national borders you grouped them by, say, IQ cohorts, the pattern would look different.
2) It reflects "tipping points." Clark's "tipping point" explanation is that when enough people become sufficiently educated, disciplined, and nonviolent, you can have manufacturing.
3) Institutions matter. This is the view that Clark argues specifically against. However, it would seem to me that if the country is the relevant unit of observation, then institutional explanations are almost impossible to resist. One way to test this would be to compare development gaps between neighboring provinces within countries with gaps between neighboring provinces that are divided by national borders. If institutions do not matter, then the gaps should be no bigger across borders than within borders. I haven't done a formal study or considered examples from the 19th century, but the modern examples that stand out in my mind (North and South Korea, East and West Germany) seem to support an institutional view.
Just to repeat, if Clark does not believe that institutions matter, then why does he choose a country as his unit of analysis? Why not choose a unit that more closely maps to his thesis of population characteristics and tipping points?
Tyler also writes,
Isn't part of the historical inability to boost long-run living standards simply the result of recurring wars and depredations?
Part of the improvement in human character that Clark credits for industrial revolution is a reduction in the propensity for violence. But Clark believes that wars actually increase living standards in the short run. See below.
Tyler goes on to say,
Should I reject the Julian Simon model I grew up with? In that view there are increasing returns to scale within cities, where people usually don't starve. The countryside languishes in poor countries, in part because it is underpopulated. Rather than having a "one population model" with an aggregate "n," labor markets are local. The "plagued by diseconomies of scale rural folk" cannot sufficiently connect with the "economies of scale urban sophisticates," mostly because of bad institutions and backward infrastructure. And the cities prove unable to protect themselves against all ongoing predations. Doesn't that model fit the data too, and fit the disaggregated and shorter-run data better?
In Clark's view, there was more labor mobility in England prior to the Industrial Revolution than there was in other countries. This would be a positive factor in any model of development.
what is the short-run claim about the response of population to real wages
This question tripped me up for a long time while reading Clark's book. I always thought that the "Malthusian" model was one in which there was a strong response of population to the availability of food. So when times are good, people over-breed, until the food supply is no longer adequate. I had in mind a model in which shocks (weather-related?) to agricultural output caused cycles in population.
The way Clark uses the Malthusian model, the availability of food is essentially fixed in the short and medium run. The main shocks are to population--plagues, wars, and what have you. I use the metaphor of a meadow with humans grazing like a herd. When the herd is thinned by plagues or wars, the remaining humans get more to eat.
I think that to grok Clark, you have to shift from what I used to think of as Malthusianism to the metaphor of the meadow.