Is autocracy bad for growth?

Gordon Tullock frequently tells me there is no good economic argument for democracy, if we adjust properly for economic variables such as the absolute state of development.  After all, much of the European miracle occurred under non-democratic conditions, not to mention the golden ages of China or modern Singapore.  But now Kevin Grier and Mike Munger argue from the empirics that democracy is better for economic growth.  In particular:

New dictatorships grow very slowly, and very old dictatorships grow very slowly. But durable dictatorships in the middle years actually grow nearly as fast as democracies. A nonlinear specification fits almost exactly.

It is a tricky question which economic models of autocracy this is consistent with (try your hand at this in the comments).  Here is part of the paper’s abstract:

In this paper we study a large unbalanced annual panel of 134 countries covering the period 1950 – 2003. We show that autocracies grow almost one percentage point slower than non-autocracies, holding constant the effect of regime length on growth…

I usually tell Gordon that the costs of keeping out democracy are prohibitive for most contemporary societies; that alone should tip the balance in favor of democracy.  Sources of economic power and sources of political power have to stand in some sort of equilibrium relationship if stability is to persist.  Singapore is an exception because it is a) very small, b) disciplined by world markets to an extreme degree, and c) its citizens realize that its "democracy" would otherwise collapse into identity politics of the three major ethnic groups; they therefore do not demand so much democracy.

Gordon never agrees.  Here is the paper and further discussion.  More importantly, here is Kevin on stereo equipment and tubes.