indirectly and distorted in transmission. Given the high noise-to-signal ratio for this type of
information and the large gap between the bureaucratic rents and the consumption of an
ordinary worker, it is not surprising that the elasticity of response to the short-term
fluctuations in the latter is much higher. An important corollary to this finding is that future
empirical research on hierarchical regimes is not going to be significantly hampered by the
lack of access to the data that pertain to the opaque higher tiers of the ruling bureaucracies, as
long as economic variables corresponding to the negative incentives can be observed.
In part, these results reconfirm earlier findings by Schnytzer and Sustersic (1998) for
the former Yugoslavia. The essence of their findings is that the lower wages and employment,
the higher the supply of activists, as revealed in the party recruitment rates. In other words:
the worse for the country, the better for the bureaucracy. The negative treatment effect of the
post-1961 period indirectly contradicts the positive effect of repression on the support for the
regime identified by these authors, since it was the period of 1956-61 that was characterized
by more liberal policies in the Soviet Union. My results are also in general agreement with the
results of the cross-country studies by Feng and Zak (1999) and Barro (1999), although the
difference in theoretical premises and variable definitions makes a direct comparison difficult.
Both studies reveal positive correlation between low inequality and high level of education
with the probability of democratic transitions (the former) or propensity for democracy (the
latter). Both these findings and my results are consistent with the assumption of the rent-
maximizing ruling bureaucracy and the rational population, responsive to promotion
incentives. The analysis in this paper, however, shows clearly that both sides of the political
labor market influence the observed outcomes significantly. Economic incentives for both
38