JOHN G. MATSUSAKA
University of Southern California
The Endogeneity of the Initiative:
A Comment on Marschall and Ruhil
It is widely believed that direct democracy, in the form of the
initiative, brought about cuts in state taxes and spending over the last quarter
century. This belief is based on firsthand observation, case studies, and
more than a dozen statistical studies (Matsusaka 2004). Marschall and Ruhil
(2005) focus on one of the central issues in this literature: Did the initiative
cause this perceived reduction in taxes and spending, or was there some
unmeasured factor that led to adoption of the initiative as well as spending
and tax cuts?
Previous research addressed this problem of spurious correlation in a
variety of ways, ranging from detailed examinations of specific cases (for
example, Gerber et al. 2001) to explicit attempts to control for missing variables
(for example, Matsusaka 1995 and Merrifield 2000). Marschall and
Ruhil approach the problem using instrumental variables and arrive at a
conclusion diametrically opposite from the previous literature: the initiative
increased spending and taxes.
This comment explains why that surprising conclusion is probably mistaken.
Marschall and Ruhil are to be applauded for applying careful empirical
techniques to an important problem, and their article lays down valuable
methodological tracks for other researchers to follow. However, their empirical
results are fragile, and their model specification can be rejected on both
theoretical and empirical grounds. With a minor change in their specification
that can be justified theoretically and empirically, or the inclusion of another
instrument, their finding is reversed, coming into conformity with the rest
of the literature.
|
|
