coefficient estimate is poorly determined). For its part, the specification with the total
number of exceptions in a state implies that each additional exception raises the
probability of legislative innovation by 5.2 percentage points. The coefficient estimates
of the other arguments are statistically insignificant. While concluding that laws may be
an efficient alternative to an attenuated at-will doctrine, Krueger has to conclude that the
threat to employers is not yet great enough to provoke sufficient support for legislation.
One reason of course may be the heightened use of temporary or atypical workers not
subject to the predations of the courts (see below).
Krueger’s view has attracted controversy, most obviously because there is no attempt to
model employer support. Another issue is whether the bills introduced into the nine state
legislatures accurately portray employer support for legislation. Thus, for example,
Stieber and Block (1992) argue that employers may have been reacting to other
legislation that they had no hand in shaping and which arguably was more coercive.
Krueger (1992, p. 797) counters that employers will be “less resistant” - maybe even
reluctantly favor - legislation if the common law has been modified. In turn, this
comment reveals the lingering imprecision of the argument. Why for example should the
public policy exception be the seemingly most important exception? How costly has it
proven in practice? And what were the differences between the three states in the sample
recognizing all three exceptions - Connecticut, Montana, and Nevada - that led just one
of them to introduce legislation?
The most interesting conclusion of Krueger’s (1991, p. 659) political compromise model
is that the “threat to employers under the common law is not yet great enough in most
states to provoke sufficient support for legislation.” As was hinted at earlier, one reason
for this may be the growth in employment forms not subject in practice to judicial review,
most notably temporary agency employment which has grown much faster than open-
ended employment in the last three decades. Autor (2003) has recently examined the
growth of the temporary help service (THS) industry between 1979 and 1995 and linked
this to the erosion of the common law at-will principle. Using data the Census Bureau’s
County Business Patterns files and the ORG files of the Current Population Survey, he