10.2Variables, Data and Descriptive Statistics

This section describes the variables and data as well as some descriptive statistics. Because of data availability concerns, not all the variables used by LRP are considered in this analysis. The purpose is to build a basic model of regional income inequality in order to introduce IT variables and study their effects.

The different variables used in the model presented above are defined in Table 10.1, and are described in Table 10.2. Data for the socio-demographic variables come from the decennial census of population for the year 1990 [U.S. Bureau of the Census (1994)]. Detailed explanations on the construction of the three IT variables appear in chapter 7. The data for the share of IT employment among total employment is obtained at the county level from the County Business Patterns survey of the Bureau of the Census (1990). These are aggregated at the state level. Finally, the densities of IT and traditional employment are calculated at the county level according to the model of Ciccone and Hall (1996). Data on county employment previously cited are used to compute these density indexes at the state level.

Table 10.1Definitions of Variables

The average Gini coefficient is 0.3943 with a standard deviation of 0.0226. Hence, income inequality differs across states, but not in high proportions. The share of employment that is of IT type is 30% on average, ranging from 20% to 43% across states. The density of traditional employment is, on average, higher than that of IT employment. However, the standard deviation of IT employment is higher than that of traditional employment. This fact supports the idea that the localization of IT employment varies across states more than the localization of traditional employment.

Finally, Table 10.3 reports values of the Gini coefficients, IT intensities and densities by state, as well as rankings. The table shows that New York is the most IT intensive and IT dense state and has the third greatest level of inequality. However, Massachusetts is the second state regarding IT variables, but ranks only 27th regarding its level of inequality. Thus it is hard to draw conclusions about the effects of IT on income inequality just from the gross data so a regression analysis is used.