ABSTRACT
This chapter shows that there is a much greater degree of factor mobility among the units of observation, obviating the need for measures of market openness. Otherwise, variable selection and inclusion is guided by prior attempts to test neoclassical and endogenous models, to build in sectoral differentiation, and to incorporate institutional differences such as the role of government. The chapter determines the unit of observation for the standard metropolitan statistical area the US Department of Commerce. While data are relatively available for most of these metropolitan statistical area a number were deleted because consistent information was unavailable over the entire time span. A corollary difficulty with the use of the level form of data is that it is likely to posses a unit root problem. A unit root problem is said to exist when there is a violation of the standard assumption that the means and variances of the variables are well defined constants and independent of time.
