This paper analyzes the role of establishments in the upward trend in dispersion of earnings that has become a central topic in economic analysis and policy debate. It decomposes changes in the variance of log earnings among individuals into the part due to changes in earnings among establishments and the part due to changes in earnings within establishments. The main finding is that much of the 1970s–2010s increase in earnings inequality results from increased dispersion of the earnings among the establishments where individuals work. Our results direct attention to the role of establishment-level pay setting and economic adjustments in earnings inequality.

This paper examines unionism’s relationship to the size of the middle class and its
relationship to intergenerational mobility. Panel Study of Income Dynamics (PSID) 1985
and 2011 files are used to examine the change in the share of workers in a middle-income
group (defined by persons having incomes within 50 percent of the median) and use a
shift-share decomposition to explore how the decline of unionism contributes to the
shrinking middle class. The files are also used to investigate the correlation between
parents’ union status and the incomes of their children. Additionally, federal income tax
data is used to examine the geographical correlation between union density and
intergenerational mobility. Findings include that union workers are disproportionately in
the middle-income group or above, and some reach middle-income status due to the
union wage premium; the offspring of union parents have higher incomes than the
offspring of otherwise comparable non-union parents, especially when the parents are
low-skilled; and offspring from communities with higher union density have higher
average incomes relative to their parents compared to offspring from communities with
lower union density. These findings show a strong, though not necessarily causal, link
between unions, the middle class, and intergenerational mobility. 

We should worry less about the potential displacement of human labor by robots than about how to share fairly across society the prosperity that the robots produce.

This paper uses linked establishment-firm-employee data to examine the relationship between the scientists and engineers proportion (SEP) of employment, and productivity and labor earnings. We show that: (1) most scientists and engineers in industry are employed in establishments producing goods or services, and do not perform research and development (R&D); (2) productivity is higher in manufacturing establishments with higher SEP, and increases with increases in SEP; (3) employee earnings are higher in manufacturing establishments with higher SEP, and increase substantially for employees who move to establishments with higher SEP, but only modestly for employees within an establishment when SEP increases in the establishment. The results suggest that the work of scientists and engineers in goods and services producing establishments is an important pathway for increasing productivity and earnings, separate and distinct from the work of scientists and engineers who perform R&D.