Since the middle of the 1990’s, over 100 cities and counties in the U.S. have followed the example of Baltimore, Maryland, and adopted living wage legislation.  The spread of living wage legislation has been neither random nor accidental.  The purpose of this paper is to explore the factors that influence a local government’s decision to adopt living wage legislation.  An understanding of these factors is vital for those researchers who seek to measure the economic impacts of the legislation; the endogeneity of the legislation must be addressed in order to separate its effect(s) from its cause(s).
Assume that civic leaders support living wage legislation when the perceived net political benefits of the legislation (benefits in excess of costs) are positive.  A measurement of the perceived net political benefits (y*) is itself not directly observable.  However, assume that it is a stochastic function of a vector (X) of observable characteristics of the local community.  That is,
y* = Xβ + ε
where β is a vector of unknown slope coefficients and ε is a stochastic error term.  Then civic leaders support living wage legislation when y*>0; they do not support living wage legislation when y*<0.  Factors that could reasonably be included in X are measurements capturing the local economic conditions (such as poverty rate, unemployment rate, and median household income), the strength of certain pre-existing local interest groups, the local political attitudes, and the pre-existence of state wage policies.  
Though perceived net political benefits are not directly observable, the decision of community leaders to enact living wage legislation (y) is observable.  Let y=1 (living wage legislation has been enacted) where y*>0 (the net political benefits are positive), while y=0 (living wage legislation has not been enacted) where y*<0 (the net political benefits are not positive.)  Estimates of the parameters (β) in the specification for y* can be obtained by applying Probit (or Logit) analysis, if  is assumed to have a normal (or logistic) probability distribution.
The model is initially estimated for a sample of U.S. counties.  The analysis is repeated for a sample of U.S. cities with populations exceeding 25,000 in 1990. Information on the enactment of living wage legislation is available on the web. Measurements of observable characteristics of the local community are taken from the County and City Data Book (various years).
The principal finding is that, while the political character of the community plays a significant role, government action is also significantly influenced by local economic circumstances.  The analysis suggests that, ceteris paribus, the greater the poverty rate, the greater the likelihood that the legislation is adopted.  Clearly, economic researchers investigating the impact of living wage legislation on poverty rates must take care to control for the endogeneity of the legislation in their analyses, if they are to be successful in separating its effects from its causes.