Abstract:
This paper investigates the robustness of recent findings on the effect of parental background on child health. We are particularly concerned with the extent to which their finding that income effects on child health are the result of spurious correlation rather than some causal mechanism. A similar argument can be made for the effect of education - if parental education and child health are correlated with some common unobservable (say, low parental time preference) then least squares estimates of the effect of parental education will be biased upwards. Moreover, it is very common for parental income data to be grouped, in which case income is measured with error and the coefficient on income will be biased towards zero and there are good reasons why the extent of bias may vary with child age. Fixed effect estimation is undermined by measurement error and here we adopt the traditional solution to both spurious correlation and measurement error and use an instrumental variables approach. Our results suggest that the income effects observed in the data are spurious.