The Revenue Effects of the Kennedy and Reagan Tax Cuts: Some Time Series Estimates

Univariate time series models are estimated for sample periods ending with the enactment of major tax reductions in 1964 and 1981. These models are used to forecast government revenue for the period following the tax cut, and the pattern of forecast errors is examined. Unforecast revenue is negative and large relative to its standard error following the 1981 tax cuts but is close to zero following the 1964 cuts. This disparity occurs because national output behaved differently in the two cases, suggesting that short-run movements in output are dominated by factors other than tax rate changes.     

Reverse Regression,Collinearity, and Employment Discrimination

Apparently contradictory results between direct and reverse regression in employment-discrimination data analysis are a manifestation of collinearity in the data. An easily implemented guideline that alerts the analyst to the presence of contaminating collinearity is illustrated with employment data from Title VII litigation.     

Reverse Regression,Fairness, and Employment Discrimination

Possible salary discrimination can be studied by comparing mean salaries of, say, males and females, after statistical adjustment for differences in job qualifications. The adjustment is often made by regression, with salary as dependent variable, and job qualifications and sex as independent variables. One might also regress job qualifications on salary and sex, a procedure called reverse regression. Ideas about fairness as well as technical concepts are relevant to discrimination studies. There are two distinct aspects of fairness, one based on comparisons of salary and the other based on comparisons of qualifications. Both concepts are needed to evaluate fairness.     

A Note on Reverse Regression,Collinearity, and Employment Discrimination

This note refutes Whiteside and Narayanan's recent assertion that the conflict between direct and reverse regression for discrimination assessment (as pointed out by Conway and Roberts and others) is due to the collinearity in the data. Their mistake stems from misunderstanding of the elementary relationship between partial and multiple correlation coefficients.