Laserfiche WebLink
<br /> <br />The Effects of a $15 Minimum Wage by 2019 in Santa Clara County and San Jose 35 <br /> <br />In general, the effect of automation on employment depends upon the elasticity of substitution of <br />capital for labor (sigma)—the change in the relative prices of capital and labor—and the share of <br />profits in revenue. The lower is sigma, the more difficult it is to substitute capital for labor. Robert <br />Chirinko, the leading economist specializing in estimates of sigma, finds an economy-wide sigma <br />of about 0.4 (Chirinko and Mallick 2016). While the estimates in this study are identified across <br />all economic sectors, most of the variation occurs among manufacturing industries. Lawrence <br />(Lawrence 2015) also finds that the economy-wide sigma is less than 1 and that it is lower still in <br />low-wage manufacturing industries than in high-wage manufacturing industries. <br />Alvarez-Cuadrado, Van Long and Poschke (2015) estimate substitution elasticities separately for <br />manufacturing and services using data on 16 countries. They find that service sector elasticities <br />are considerably lower than in manufacturing. However, their study does not examine low-wage <br />services separately. The results in these papers nonetheless suggest, as Autor et al. conjectured, <br />that automation possibilities are lower in low-service jobs. <br />Aaronson and Phelan (Aaronson and Phelan 2015) have carefully studied the short-run impact of <br />minimum wages on the automation of different kinds of low-wage jobs. Their study is the first to <br />examine automation within low-wage industry contexts. Aaronson and Phelan find that minimum <br />wage increases do reduce routinized low-wage jobs (such as cashiers) and increase the number <br />of less-routinized low-wage jobs (such as food preparation). As it turns out, the changes offset <br />each other almost equally, resulting in no net change in employment. Thus, Aaronson and Phelan <br />(2015) find that sigma is essentially zero in low-wage occupations. <br />We use a sigma of 0.2 in our calculations, half-way between Chirinko and Mallick and Aaronson <br />and Phelan. This conservative assumption may therefore result in an over-estimate of the <br />magnitude of the automation effect. <br />Aaronson and Phelan’s findings also suggest very little substitution of highly skilled workers for <br />lower skilled workers. Dube, Lester and Reich (2016) obtained a similar result. Consequently, we <br />do not include any effect of skilled labor being substituted for unskilled labor in our model. <br />Automation in practice <br />Machines that process automated transactions—at airports and in airplanes, banks, self- <br />checkout stations in retail stores, parking garages, and gasoline stations—have become <br />particularly widespread over the past 30 years. During this period, the price of computer-related <br />machines has rapidly declined. Labor-saving automation will occur even when wages do not rise, <br />insofar as the technological change continues to push down the price of equipment, making <br />investments in new equipment and software profitable. <br />The effects of a rising minimum wage on actual automation depend in part upon whether new <br />labor-saving technology that has not yet been adopted continues to become available. We <br />suggest that much of existing labor-saving technological change has already been embodied in <br />low-wage industries, in equipment and software such as smart electronic cash registers, remote <br />8.A. - Page 49