This preview shows page 1. Sign up to view the full content.
Unformatted text preview: 151 lectures Rising Wage Inequality Introduction Outline 1. 2. 3. Measurement issues: concepts, alternative summary statistics, data sources. Descriptive trends Underlying causes To discuss causes I will present an analytical framework (S, D, I). The main contenders on S, D are: skill-biased technical change (SBTC) and globalization (including trade and immigration). The main contenders on I are: (decline in college aid and inadequate expansion of higher ed., deregulation of product markets, union decline, lower minimum wage, changes in social norms for CEO pay). E-S text argues for SBTC. In these lectures I will present that case first, then discuss international comparisons and the institutional argument next time. I. A. 1. Measurement issues Concepts Earnings inequality: (per hour, per week, per year, over three to five years, over a career). Can also look at income from all sources or just at consumption inequality; consumption data sources are not very reliable, however. Is inequality growing because a) rich got richer, b) poor get poorer, or c) both of the above? We tend to prefer a > b > c. Data support c. 2. Recipient unit 1 Individual, family, household; adjust for economies of scale in consumption? 3. Mobility Issues here involve growth in instability versus permanent earnings, crosssectional snapshots versus mobility over a career, and intergenerational mobility. We do not have time for all of these issues. A one-sentence summary of a large literature is that mobility (and instability) studies using longitudinal data do not change the story of growing inequality. The U.S. has less career and intergenerational mobility than other countries and mobility has not increased over time. B. 1. Alternative summary statistics Variance: decomposable into within and between groups components Also—variance of log earnings (reduces weight of very high earners). Coefficient of variation: Standard deviation divided by the mean Shares of quintiles, deciles, percentiles. 2. 3. Requires knowing the overall income that goes in the denominator—usually by estimating the mean for the open-ended, or top-coded interval. 4. Ratios of levels: D9/D5 D5/D1 D9/D1 P80/P20 5. C. Lorenz curve and Gini ratio (see slides) Data sources 1. CPS March annual demographic supplement—basis for published annual earnings as used in E-S text. 2 2. CPS monthly outgoing rotation groups—better measures of hours worked and a sample that is three times larger. Gives different results for the 1990s than the March series. 3. To study trends within the top one percent or higher, need to use other data: best are the IRS files as used by Saez (2003). II. A. Descriptive trends Earnings See T. 14.1. For men, rich got richer and poor got poorer. For women, no trend. But note, this is for ages 25 plus only and uses only annual earnings concept. B. Occupations Increasingly bi-modal? See Table 14.2. E-S say high-paying occupations are growing and and lowest-paid are constant (as a percentage of total) So growth of high-paying occupations is at expense of middle-paying jobs. C. By experience and education—T. 14.3, for year-round full-time workers only Increases in returns to education but not to experience. Men: Caused by declines for high school grads rather than increases for college grads Women: some real increases for college grads. D. Hours T. 14.4 finds no differences in trends of hours by occupation using March CPS, but using CPS ORG data we do find some increasing polarization of hours. E. Within-group dispersion 3 Increase in inequality, most of it in the 1980s. I’d emphasize, none or very little in the 1990s, once compositional changes to higher average schooling levels are included. Summary to this point: increase in returns to education III. Underlying causes of rising inequality --Note that E-S focus only on causes of rising educational differential. See slides: fig. 14.3 and table 14.6. --Suppose cause is that supply curve shifts to the right. If increase is greater among HS graduates than among college graduates, predict increasing percentage of employment is HS, declining percentage are BAs. --T. 14.6 shows the opposite, that increasing proportion of employment is among those who were getting the biggest wage increases. This suggests that demand, not supply shifts primarily responsible. --This demand-driven hypothesis is supported also by studies of a) shifts in industrial composition to the sectors that use information technology (IT) more and, b) by Alan Krueger’s study with CPS microdata showing that workers who had computers at work were receiving a 20 percent wage premium compared to workers who did not. --E-S conclusion: technical change has been biased to increased demand of workers with more skills, especially computer skills. 4 ...
View Full Document
- Spring '08