In a recent post on the economics of immigration, Nick Rowe engages in the most astonishing example of “use the Cobb-Douglas model to assume away the very problem under discussion and don’t even realize you did it” that I’ve ever seen. Here’s Nick:
When economists disagree with public opinion, I normally agree with the economists. But we ought to think twice when this happens. Maybe, just maybe, public opinion is based not on ignorance but on something that our models leave out. When it comes to immigration policy, I think it is correct to say that elite opinion (which includes economists) in rich countries that are attractive to immigrants is generally more favourable to immigration than is public opinion. The elite finds public opinion a bit embarrassing, and tries to ignore it. They remind me of parents trying to get their kids to eat broccoli: “It’s good for you, and you will like it once you get used to it”.
The basic economics of immigration are quite simple. If we assume constant returns to scale, with output a function of labour and capital, and realise that increased labour creates increased capital (either immediately through capital inflows or slowly through immigrants’ saving), everything expands in proportion. Both supply and demand for labour expand, leaving real wages, per capita income, and the unemployment rate, all unchanged. [Bold added.]
Do you see that?! Marvel at its beauty. If people show up in a capital-rich country with nothing but their bare hands and work ethic, we can treat it the same as if they showed up with 8 tractors.
Later on, Nick starts relaxing assumptions, but he never says, “Oh yeah, it might take a generation or two for people from Africa to reach the level of per capita wealth that the US currently has, so that could depress wages and enrich capitalists only for 60 years tops. Maybe some people who work for wages don’t want to see a drop in their standard of living for 60 years–but they watch ‘Matlock’ too, the rubes.”
Also, I don’t think Nick even uses this approach consistently. Later in his post he writes: “So immigration probably doesn’t matter much for natives; but for immigrants it can sometimes matter a lot.”
But why would it? Wherever the home country of the immigrant is, they have the same Y=F(K,L) CRS production function. And Nick has told us that in the long run, K/L is mean-reverting. So real wages, per capita income, and the unemployment rate should be basically the same all over Earth, give or take. Why would any worker want to move to the US or Canada?