13
regulations introduced, and the inspection budget. The methodology for manufacturing is
a Weibull hazard specification with time-varying regressors. The goal is therefore to
address the timing of labor administration and factory inspectorate innovations. Apart
from the proxy for large firms (average employment size), the regressors are
manufacturing employment, either a union index (measuring the share of workers in
manufacturing relative to the national average) or the number of union chapters, and a
dummy for Southern states. It is found that the hazard ratios are greater than unity for
larger firms, consistent with earlier adoption. Specifically, a one standard deviation
increase in firm size is associated with a 31 (28) percent increase in the conditional
probability of adopting some form of labor administration (factory inspectorate).
Fishback argues that this finding is inconsistent with the view that large firms sought to
obstruct legislation. The union hazard ratios are both mixed and poorly determined so
that it is difficult to conclude that they either contributed to or were opposed to
legislation. One interpretation, and that favored by Fishback, is that unions were likely
more interested in building up their organizational strength to obtain influence than to
engage the polity.
Opposite results are obtained for coal mining in regressions estimated over a
reduced number of states (with bituminous coal production). That is, larger mines are not
associated with earlier adoption of coal safety legislation. This time, however, the union
‘effect’ exceeds unity and is statistically significant at conventional levels. Since the
former result might indicate that “larger mines were indifferent to coal regulations or that
they were unsuccessful in staving off the efforts of reformers” (p. 19), Fishback also
estimates OLS and state and year fixed effect models of the determinants of the size of
the inspection budget per coal worker and a coal mining law index capturing the reach of
legislation. The size of mine argument is negatively associated with each outcome
indicator. Union effects are measured by the share of the workforce in the United
Mineworkers and are weak throughout, a result that might hint at inadequate inspections
and (as before) induced self reliance to effect change. Fishback concludes that taken in
the round his results for mining indicate that larger employers were successful in limiting
the reach of legislation and in reducing inspection budgets.