Using a SLX Model to Examine the Impact of the Great Recession on Poverty in South Carolina

Abstract

ABSTRACT In this paper, we employ a model using spatially lagged explanatory variables (SLX) to model poverty in the rural South – a region with large pockets of counties with persistent and high poverty. The SLX model allows for separation of spillovers from non-metro counties versus those from metro counties, something missing from the existing literature. We employ various economic and demographic variables to model poverty rates pre- and post-Great Recession. We also estimate separate models for non-metro counties and metro counties. There were no significant impacts from the local metro county before the recession and no significant local impacts in non-metro counties after the recession. Before the recession, factors from surrounding metro counties contributed greater impact on both non-metro and metro poverty. After the recession, there were more significant factors from surrounding non-metro counties that displayed a greater impact. The results suggest that poverty is unique across space and time. Poverty reduction programs should not only adopt a discretionary strategy that accounts for geographical variation, but also consider the timing of business cycles. Our research suggests a universal, all-encompassing strategy toward poverty may be less than ideal in the rural south.