We present a measure of the correlation between the education levels of spouses based on a bivariate ordered probit model. The change in this correlation over time can be measured while controlling for the large changes in the educational attainment levels. The model is estimated with data from...

We use data on the National Hockey League Entry Draft and subsequent performances of the players to predict performance as a function of the player's position, year, and overall rank in the draft. This is done by inequality-constrained least squares using a variation of the...

Suppose an agency awards a …xed number of prizes to applicants in different categories such that the applicant-to-winner ratio is constant by category. It is demonstrated in a simple theoretical model that the number of awards in a category will typically be positively related to the degree of...

The O(n-1) bias and O(n-2) MSE of OLS are derived for iid samples. An approach is suggested for handling nonexistent finite sample moments. Bias corrections based on plug-in, weighting, jackknife and pairs bootstrap methods are equal to Op(n-3/2). Sometimes they are effective at lowering bias...

We use Canadian Surveys of Consumer Finances 1971-1993 to study the wages of full-time, full-year male and female workers. Median real wages of 24-year-old males without a university degree fell by 25% between 1978 and 1993. For 24-year-old females the decline was more modest and reversed in...

There is a gap between the predictions of capital tax competition models and the reality they purport to describe. In a standard capital-tax model, with head taxes, capital-importing regions tax capital and capital-exporting regions subsidize capital. In the real-world, competing regions appear...