Pharmacy today is a highly remunerated female-majority profession with a small gender earnings gap and low earnings dispersion. Using extensive surveys of pharmacists, as well as the US Census, American Community Surveys, and Current Population Surveys, we explore the gender earnings gap, penalty to part-time work, demographics of pharmacists relative to other college graduates, and evolution of the profession during the last half-century. Technological changes increasing substitutability among pharmacists, growth of pharmacy employment in retail chains and hospitals, and related decline of independent pharmacies reduced the penalty to part-time work and contribute to the narrow gender earnings gap in pharmacy.

Recent technological advances and a widening of the wage structure have led many to conclude that technology and human capital are relative complements. The possibility that such a relationship exists today has prompted a widely held conjecture that technology and skill have always been relative complements. According to this view, technological advance always serves to widen the wage structure, and only large injections of education slow its relentless course. A related literature demonstrates that capital and skill are relative complements today and in the recent past (Zvi Griliches, 1969). Thus capital deepening appears also to have increased the relative demand for the educated, serving further to stretch the wage structure. Physical capital and technology are now regarded as the relative complements of human capital, but have they been so for the past two centuries? Some answers have already been provided. A literature has emerged on the bias to technological change across history that challenges the view that physical capital and human capital have always been relative complements. Many of the major technological advances of the 19th century substituted physical capital, raw materials, and unskilled labor for highly skilled artisans (John A. James and Jonathan S. Skinner, 1985). But if physical capital and human skill were not always relative complements, when did they become so, and when did new technology become skilled labor's complement? We argue that capital-skill complementarity was manifested in the aggregate economy as particular technologies spread, specifically batch and continuous-process methods of production. Across the past two centuries, manufacturing shifted first from artisanal to mechanized and nonmechanized factory production, then from simple factories to assembly lines, and finally from assembly lines to continuous and batch processes. Although few products were manufactured by more than two of the technologies mentioned, manufacturing, as a whole, progressed in the fashion described. In considering our argument it is useful to envision manufacturing as having two distinct stages: (i) a machine-installation and machine-maintenance segment and (ii) a production or assembly portion. Capital and educated (skilled) labor, we will argue, are always complements in the machine-maintenance segment of manufacturing, regardless of the technology. Machinists, for example, are needed to install machinery and make it run. The workable capital created by skilled labor plus raw capital is then used by unskilled labor to create the final product in the production or assembly segment of manufacturing. How the adoption of a technology alters the relative demand for skilled workers will depend on whether the machinemaintenance demand for skilled labor is offset by the production-process demand for unskilled labor.

We study employers' perceptions of the value of postsecondary degrees using a field experiment. We randomly assign the sector and selectivity of institutions to fictitious resumes and apply to real vacancy postings for business and health jobs on a large online job board. We find that a business bachelor's degree from a for-profit online institution is 22 percent less likely to receive a callback than one from a nonselective public institution. In applications to health jobs, we find that for-profit credentials receive fewer callbacks unless the job requires an external quality indicator such as an occupational license.