revenue to the sum of total intergovernmental and own-source revenue (strevrat), which is a
measure of fiscal burden, does not achieve statistical significance at the p < 0.10 level. Moreover,
the results show that fiscal stress at the local-level has no apparent effect on the extent of local
government contracting for services. We included two indicators of local fiscal burden in the
model: ratio of local property tax revenue to sum of intergovernmental revenue and property tax
revenue (fburden) and per capita total short-term debt outstanding (capsdebt). We had
hypothesized that as fiscal burden rises, the demand for smaller government would increase,
resulting in more contracting in order to reduce the cost of service delivery. Both variables,
however, fail to achieve statistical significance at the p < 0.10 level.
An ICMA survey indicator of
perceived fiscal stress (x1_fs) also fails to achieve statistical significance in the model (p < 0.54)
(see also Hefetz and Warner, 2004). Finally, however, per capita personal income (pincome0),
which is an indicator of state fiscal capacity, is in the opposite direction than what was expected.
Per capita personal income exerts a positive impact on local government contracting (p < 0.07).
It
appears that citizens with higher income levels generally prefer smaller government and more
efficient service delivery.
As indicated in the literature review, a variety of political and ideological factors can
influence the extent to which local governments rely on contracting out. Three population
subgroups have been hypothesized to favor a larger role for government and, thus, to oppose
privatization of service delivery: blacks, the elderly, and individuals with low income. The results
show, however, that the percentage of the population that is black (blackpop), the percentage of the
population that is 65 years and older (pop65), and the percentage of households receiving public
assistance income (pasincpop) all have coefficients that fail to achieve statistical significance at the
p < 0.10 level. This finding is contrary to some previous research (e.g., Thompson and Elling,
2000; Miranda, 1992). A more direct measure of opposition to privatization, citizen opposition to
privatization (q5a_1), measured using an ICMA survey indicator, does achieve statistical
19
Convention