Spatial and Geographical Effects in Regional Multiplier Analysis

In this paper I extend regional multiplier analysis by incorporating both spatial (neighbor) and geographical (place-contained) effects. Multipliers, taken as elasticities, are estimated by regression analyses of the associations between retail trade earnings by place of work and total income by place of residence (from earnings, property, and transfers) across New England's counties in 2002. The multiplier analyses are conducted in two basic forms, both estimated using a simultaneous autoregressive specification and both incorporating a spatial neighbor income variable. In one, geographical effects are treated as dummy variables alone, so their impact on a particular income multiplier coefficient is read only indirectly. In the other, geographical effects are allowed to interact with property and transfer income so that their multiplier coefficients in the region vary from county to county depending upon particular geographical characteristics. The spatial effect of neighboring county income is consistent across the two forms, and suggests a competitive effect in the region's retail sector. Further results indicate that the region's property income and transfer receipts tend to offset each other in geographical context in their impact on retail-trade income.