The nature of union endogeneity is examined in the light of recently proposed estimators for union wage differentials. The instrumental variable (IV) approach adopted by Duncan and Leigh is shown to yield little information on the precise nature of the endogenous process. In particular it cannot be used to sign the direction of selection into the union and nonunion sectors since by construction the IV estimates impose opposite signs on the selection terms.

Measures of four basic skills, constructed from the Dictionary of Occupational Titles, are used to examine the source of human capital specificity. The measures are used to characterize the skill portfolio of each job and to construct distance measures between jobs. Wage losses in the Displaced Worker Surveys are shown to be more closely associated with switching skill portfolios than switching industry or occupation code per se. These switches represent large decreases in the skill portfolio in the postdisplacement job. The recent evidence for industry‐specific capital is reexamined. The results suggest a difference between fluid and crystallized skills.

The paper attempts to integrate new approaches to estimating union wage effects with the analysis of public-private sector wage differentials. Estimates of the union differential in both public and private sectors, allowing for the endogeneity of union status, are presented. The hypothesis that the recently measured rents to public sector employment primarily reflect the recent increase in unionization in that sector is examined, and receives considerable empirical support. There was evidence of positive selection into the union sector, especially for private sector workers. Union status appears to be strongly influenced by the expected wage gain from joining the union sector.

This article explores a model in which a union confronts many competitive workers, firms, and consumers. Under "monopoly" unionism, union coverage may be incomplete; then, union wages and employment are insensitive to product demand variation. Under "efficient" unionism, coverage can never be incomplete; some union variables necessarily vary with product demand. Preliminary evidence on the demand independence under incomplete coverage hypothesis is presented. Also, more structure is imposed and further hypotheses are derived, and the manner in which the model can be enriched to allow for a variety of union-related issues within a consistent framework is set out.

A frequent practice in empirical work is to "preanalyze" the data via various sample inclusion rules. Truncation of "outliers" is common. These procedures are a form of sample censoring imposed by the investigator. Such censoring produces effects familiar from the sample selection literature. This paper investigates the question why an investigator might want to censor a sample and what the costs are. In an empirical example, using a variance components model of a wage equation, potential inconsistency problems are highlighted. The results indicate that while the slope coefficients, <tex-math>\hatβ</tex-math>, may typically be less sensitive to censoring than the variance components, some common forms of censoring also markedly affect <tex-math>\hatβ</tex-math>. Finally, a Bayesian estimator that incorporates prior information in a flexible way was developed. The usual Bayesian procedure was reversed, by using the Bayesian estimator to recover the prior beliefs that an investigator imposes by any proposed truncation of outliers. Especially in large samples, extremely dogmatic prior beliefs may be imposed when outliers are eliminated. Prior distributions of the type developed in the paper may be used by the investigator to clarify the nature of his prior beliefs revealed by a willingness to truncate data points and to assess whether or not any proposed truncation accurately reflects thoes beliefs.

A substantial literature has developed to estimate the “true” cyclicality of real wages, that is, composition bias free. Two major issues are addressed in this article: aggregation of heterogeneous workers and potential bias in the measurement of the labor input. A general analysis of the biases is presented, and alternative approaches in the literature are nested in a single framework. Estimates based on an efficiency units concept that avoids the usual aggregation problems are presented. Composition bias underestimates the usual parameters of interest unless both the price and the quantity of the labor input are adjusted appropriately.

In the last three decades, Canada and the United States showed different paths in per capita gross domestic product growth, skill premiums, and inequality. Worker quality and price differences both play a role but are difficult to distinguish. Human capital prices and quantities are estimated using methods we developed previously. In the United States, there was faster growth and a much more rapid rise in skill premia and inequality. This was primarily due to different paths for the relative price paid to rent high-skilled human capital in the two countries, rather than differences in relative quantities.