The Interdependence between Income and Education

Using single-equation approaches, previous studies concerned with human capital indicate that education expenditures have a large effect on income. Previous studies concerned with local government behavior indicate that income has a large effect on education expenditures. There is the possibility that essentially the same relationship is being measured, with dependent and independent variables merely interchanged. In this study, the equations are estimated simultaneously. To see why single- and simultaneous-equations estimates differ, expressions are developed for sources of single-equation bias.The major reason for a large single-equation bias in estimating the effect of education on income is found to be the relatively large reverse effect of income on education, which makes for high correlation between independent variable and residual in the equation for income. Single-equation bias in the equation explaining education expenditures is small for the same underlying reason, namely the reverse effect occurring in the equation is relatively small. A by-product is an analysis of demand for education allowing for effects of different forms of wealth, in contrast to the usual approach assuming one overall wealth or income effect. Nonhuman wealth is estimated to have twice as much effect on education expenditures as other forms of wealth.