The Role of Regional Interaction in Regional Growth: Competition and Complementarity in the U.S. Regional System

The role of regional interaction in multi-regional growth is examined by estimating the parameters for a discrete non-linear model of relative dynamics with maximum likelihood methods. This model specification approaches the issues of spatial dependence in a different form from the methods that have been used to date. It provides evidence for the view of multi-regional growth as a zero-sum game, in which a mixed pattern of competition and complementarity exists among non-contiguous regions. The approach is illustrated with an aggregate set of regions for the U.S. economy and forecasts are made of the progress of regional convergence into the next century. It is suggested the methodology might prove to be a useful alternative to the usual method of incorporating exogenous changes into regional models.