In turn, the rate of growth of the labor force, L?/L, is higher under dictatorships. The
observed values are 2.28 percent per annum under dictatorships and 1.59 under
democracies. Since labor force grows at a slower rate in wealthier countries, one might
suspect again that this difference is due to the distribution of regimes by income. Some of it
is, but not enough to eradicate the effect of regimes. Even when the regimes are matched for
their income, their colonial heritage, and the frequencies of Catholics, Protestants, and
Moslems, even if they are matched for demand, or for the lagged rate of population growth,
or for country-specific effects, labor force grows faster under dictatorships. The selectioncorrected difference is again small in countries with incomes under $3,000, but quite large
in those over $3,000.
To examine the effect of regimes on the efficiency with which resources are used in
production, we need to compare the selection-corrected coefficients of the respective
production functions. The constant measures total factor productivity, while the coefficients
on capital and labor represent the
elasticities of output with regard to these factors. As shown in the first two columns of
Table 3, which provide the results for the entire sample, total factor productivity is higher in
democracies. The elasticity with regard to capital is slightly higher in dictatorships, while
elasticity with regard to labor is slightly higher in democracies. Hence, one is led to
conclude that democracies benefit more from technical progress and use labor somewhat
more effectively, while dictatorships employ somewhat more efficiently the physical capital
stock.
What we learned, therefore, is that regimes have no impact on the rate of growth of capital
stock (or the investment share), while labor force grows faster under dictatorships. In turn,
democracies enjoy faster technical progress and use labor somewhat more effectively, while
dictatorships exploit somewhat better their capital stock. As a result, the rate of growth of
total output is almost identical under the two regimes.
Another way to test the effect of regimes is to focus on the countries that experienced
regime changes. Here again, however, one should proceed prudently. Countries in which
regimes are unstable may be different from those where the same regime persisted during
the entire period. Yet the observed average rate of growth is the same in those countries
which did not experience any regime transitions and those which underwent through one or
more regime changes: the rate of growth for the former was 4.23 percent (N=2813) and for
the latter 4.25 percent (N=1313). Stable dictatorships grew at the rate of 4.38 percent
(N=1709), while dictatorships in the countries that also experienced democracy grew at the
rate of 4.51 percent (N=772). Stable democracies grew at the rate of 3.98 percent (N=1104),
while democracies that rose from or gave room to dictatorships grew at 3.88 percent
(N=541). Hence, there is no reason to think that growth in the countries where regimes were
stable was different from that in countries where regimes changed.