38th Congress of the European Regional Science Association: "Europe Quo Vadis? - Regional Questions at the Turn of the Century", 28 August - 1 September 1998, Vienna, Austria
The commonly used specification in regional economic research on labour force participation is the linear probability function. An important alternative recommended in the Handbook of Regional and Urban Economics in the contribution of Isserman et al. (1986) on `Regional Labor Market Analysis' is the logit probability function. Their argument for the logit probability function is as follows. Given that economic theory on labour force participation does not suggest to pick one functional form over another and that the parameters of the logit probability function are estimable by OLS under the usual assumptions about the error term, the benefit of the logit probability function is that any estimated value for L lies within the logical bounds [0,1]. This feature is particularly desirable in a forecasting context when out of sample data might otherwise potentially yield absurd labour force participation rates. In this note two counter-arguments are gathered against using the logit probability function which are lacking in the Handbook of Regional and Urban Economics. Furthermore, it is shown that the logit probability function in this discourse can be replaced by the probit probability function equally well. Keywords: logit, probit, labour force participation rate.