Association for Public Policy Analysis & Management

To what extent the size of the middle class in a region is related to other social and economic factor? This paper examines what the correlates of the degree of the middle-classness of the metropolitan area, measured by the proportion of household fall into the middle-class in a region, using Census 2010 data are. The concerns around income and wealth inequality in the United States have closely connected to spatial inequalities. However, most research that has included on the relationship between inequality and growth, findings are often limited to understand the two extremes of the income distribution. In this paper, the degree of middle-classness across metropolitan areas will be systematically reviewed using a variety of variables to characterize the middle class. To what extent do regions with large middle-classes differ from regions with small middle-class?

To this end, author attempt to group metropolitan areas into three categories: 1) regions with a small middle-class (more than one standard deviation below the mean), 2) regions with a medium-sized middle-class (within one standard deviation of the mean), and 3) regions with a large middle-class (more than one standard deviation above the mean). Then, the author compares the differences in means among the three groups for each characteristic. The working definition of the middle-class in this paper will be selected among various definitions by applying the selection criteria, such as the one with an income-boundary approach as opposed to a fixed-width slice of income-ranked households. Also, the operationalization for the middle class in this paper includes versions both before and after a regional cost-of-living adjustment based on Moody’s Analytic’s Cost-of-living Index. Using data on the size of the middle class by each metropolitan areas in 2010 (from author’s calculation), this paper attempts to identify what economic, social, and geographical factors – including metropolitan-level income inequality – are related to the size of the middle class in a region.

Part 1 examines the correlation between the degree of middle-classness and other socioeconomic variables. Pearson’s correlation coefficient will be examined to see whether the size of the middle class in a region is related to that area’s 1) inequality (the Gini coefficient), 2) demographic (such as percentage of household with extra income), 3) economic characteristics (such as area median household income), and 4) industry structure characteristics.  Part 2 will present a multivariate analysis, using the size of the middle class as the dependent variable in an ordinary least squares (OLS) analysis of all the inequality, demographic, economic, and industry-structure characteristics. This part is not intended to be a causal analysis, but it will point out the correlates of a middle-class size in a region while controlling for other correlates/characteristics. This research intends to help to understand the role of the size of the middle-class beyond city borders and the dualism of urban and suburban dialogue.