16
helps solve one of the major puzzles in political economy of the passage of workers
compensation [viz. the leadership taken by employer groups ] ... Many employers may
have supported the legislation in anticipation of passing a substantial portion of the costs
onto their workers in the form of lower wages.”
Nevertheless, Fishback and Kantor (1996a) also report that employers (while favoring
workers compensation) and unions feuded over the issue of benefit levels, noting that
where they were unable to reach a compromise the introduction of workers’
compensation was delayed for up to 15 years! One reason for this is the failure to observe
wage offsets in union settings, noted earlier. Employers of unionized labor may then have
had a strong incentive to minimize the size of injury benefits that they paid. Further, we
have seen that the cost of insurance was not fully shifted back on to (nonunion) workers
outside of coalmining. Nor was this pass back instantaneous. In this later study, therefore,
Fishback and Kantor focus attention on the determinants of their index of expected injury
benefits, 1910-30. The key regressors are an index of the risk of accidents in
manufacturing, an index of unionization in manufacturing, measures of the strength of
farm and manufacturing interests, the proportions of large and small firms, and
manufacturing value added per worker. In addition, as indicators of the ‘political climate’
the authors identify power shifts in the legislature, percent of the presidential vote for a
republican candidate and for a socialist candidate, and the presence of a workers’
compensation bureaucracy (instrumented). One of the most important findings is that
states with higher manufacturing risk had lower benefit levels. In turn, this suggests that
employers in the most dangerous industries had considerable strength in state legislatures,
and used this influence to keep their overall accident costs down. That said, greater union
density in a state and the presence of a bureaucratic agency to administer the law (the
alternative was through the court system) were each associated with higher expected
benefits, cet. par. There is also some indication in the authors’ data that political party
shifts in either one or both legislative chambers at state level were associated with higher
benefit levels. But political attitudes as indexed through votes in national elections were
unimportant in explaining benefits, which result the authors interpret as suggesting that
the views of state-level political parties did not necessarily match those of their national-