In this paper we provide the exact formulas for the direct least squares estimation of statistical models that include both person and firm effects. We also provide an algorithm for determining the estimable functions of the person and firm effects (the identifiable effects). The computational techniques are also directly applicable to any linear two-factor analysis of covariance with two high-dimension non-orthogonal factors. We show that the application of the exact solution does not change the substantive conclusions about the relative importance of person and firm effects in the explanation of log real compensation; however, the correlation between person and firm effects is negative, not weakly positive, in the exact solution. We also provide guidance for using the methods developed in earlier work to obtain an accurate approximation.
-
Mixed-Effects Methods For Search and Matching Research
September 2023
Working Paper Number:
CES-23-43
We study mixed-effects methods for estimating equations containing person and firm effects. In economics such models are usually estimated using fixed-effects methods. Recent enhancements to those fixed-effects methods include corrections to the bias in estimating the covariance matrix of the person and firm effects, which we also consider.
View Full
Paper PDF
-
Unlocking the Information in Integrated Social Data
May 2002
Working Paper Number:
tp-2002-21
View Full
Paper PDF
-
A Formal Test of Assortative Matching in the Labor Market
November 2009
Working Paper Number:
CES-09-40
We estimate a structural model of job assignment in the presence of coordination frictions due to Shimer (2005). The coordination friction model places restrictions on the joint distribution of worker and firm effects from a linear decomposition of log labor earnings. These restrictions permit estimation of the unobservable ability and productivity differences between workers and their employers as well as the way workers sort into jobs on the basis of these unobservable factors. The estimation is performed on matched employer-employee data from the LEHD program of the U.S. Census Bureau. The estimated correlation between worker and firm effects from the earnings decomposition is close to zero, a finding that is often interpreted as evidence that there is no sorting by comparative advantage in the labor market. Our estimates suggest that his finding actually results from a lack of sufficient heterogeneity in the workforce and available jobs. Workers do sort into jobs on the basis of productive differences, but the effects of sorting are not visible because of the composition of workers and employers.
View Full
Paper PDF
-
Sorting Between and Within Industries: A Testable Model of Assortative Matching
January 2017
Working Paper Number:
CES-17-43
We test Shimer's (2005) theory of the sorting of workers between and within industrial sectors based on directed search with coordination frictions, deliberately maintaining its static general equilibrium framework. We fit the model to sector-specific wage, vacancy and output data, including publicly-available statistics that characterize the distribution of worker and employer wage heterogeneity across sectors. Our empirical method is general and can be applied to a broad class of assignment models. The results indicate that industries are the loci of sorting-more productive workers are employed in more productive industries. The evidence confirm that strong assortative matching can be present even when worker and employer components of wage heterogeneity are weakly correlated.
View Full
Paper PDF
-
Modeling Endogenous Mobility in Wage Determiniation
June 2015
Working Paper Number:
CES-15-18
We evaluate the bias from endogenous job mobility in fixed-effects estimates of worker- and
firm-specific earnings heterogeneity using longitudinally linked employer-employee data from the LEHD infrastructure file system of the U.S. Census Bureau. First, we propose two new residual diagnostic tests of the assumption that mobility is exogenous to unmodeled determinants of earnings. Both tests reject exogenous mobility. We relax the exogenous mobility assumptions by modeling the evolution of the matched data as an evolving bipartite graph using a Bayesian latent class framework. Our results suggest that endogenous mobility biases estimated firm effects toward zero. To assess validity, we match our estimates of the wage components to out-of-sample estimates of revenue per worker. The corrected estimates attribute much more of the variation in revenue per worker to variation in match quality and worker quality than the uncorrected estimates.
View Full
Paper PDF
-
The Measurement of Human Capital in the U.S. Economy
April 2002
Working Paper Number:
tp-2002-09
We develop a new approach to measuring human capital that permits the distinction of both observable
and unobservable dimensions of skill by associating human capital with the portable part
of an individual's wage rate. Using new large-scale, integrated employer-employee data containing
information on 68 million individuals and 3.6 million firms, we explain a very large proportion
(84%) of the total variation in wages rates and attribute substantial variation to both individual
and employer heterogeneity. While the wage distribution remained largely unchanged between
1992-1997, we document a pronounced right shift in the overall distribution of human capital.
Most workers entering our sample, while less experienced, were otherwise more highly skilled, a
difference which can be attributed almost exclusively to unobservables. Nevertheless, compared
to exiters and continuers, entrants exhibited a greater tendency to match to firms paying below
average internal wages. Firms reduced employment shares of low skilled workers and increased
employment shares of high skilled workers in virtually every industry. Our results strongly suggest
that the distribution of human capital will continue to shift to the right, implying a continuing
up-skilling of the employed labor force.
View Full
Paper PDF
-
Agent Heterogeneity and Learning: An Application to Labor Markets
October 2002
Working Paper Number:
tp-2002-20
I develop a matching model with heterogeneous workers, rms, and worker-firm
matches, and apply it to longitudinal linked data on employers and employees. Workers
vary in their marginal product when employed and their value of leisure when unemployed.
Firms vary in their marginal product and cost of maintaining a vacancy. The
marginal product of a worker-firm match also depends on a match-specific interaction
between worker and rm that I call match quality. Agents have complete information
about worker and rm heterogeneity, and symmetric but incomplete information about
match quality. They learn its value slowly by observing production outcomes. There
are two key results. First, under a Nash bargain, the equilibrium wage is linear in a
person-specific component, a firm-specific component, and the posterior mean of beliefs
about match quality. Second, in each period the separation decision depends only on
the posterior mean of beliefs and person and rm characteristics. These results have
several implications for an empirical model of earnings with person and rm eects.
The rst implies that residuals within a worker-firm match are a martingale; the second
implies the distribution of earnings is truncated.
I test predictions from the matching model using data from the Longitudinal
Employer-Household Dynamics (LEHD) Program at the US Census Bureau. I present
both xed and mixed model specifications of the equilibrium wage function, taking
account of structural aspects implied by the learning process. In the most general
specification, earnings residuals have a completely unstructured covariance within a
worker-firm match. I estimate and test a variety of more parsimonious error structures,
including the martingale structure implied by the learning process. I nd considerable
support for the matching model in these data.
View Full
Paper PDF
-
NOISE INFUSION AS A CONFIDENTIALITY PROTECTION MEASURE FOR GRAPH-BASED STATISTICS
September 2014
Working Paper Number:
CES-14-30
We use the bipartite graph representation of longitudinally linked em-ployer-employee data, and the associated projections onto the employer and em-ployee nodes, respectively, to characterize the set of potential statistical summar-ies that the trusted custodian might produce. We consider noise infusion as the primary confidentiality protection method. We show that a relatively straightfor-ward extension of the dynamic noise-infusion method used in the U.S. Census Bureau's Quarterly Workforce Indicators can be adapted to provide the same confidentiality guarantees for the graph-based statistics: all inputs have been modified by a minimum percentage deviation (i.e., no actual respondent data are used) and, as the number of entities contributing to a particular statistic increases, the accuracy of that statistic approaches the unprotected value. Our method also ensures that the protected statistics will be identical in all releases based on the same inputs.
View Full
Paper PDF
-
Wage Dispersion, Compensation Policy and the Role of Firms
November 2005
Working Paper Number:
tp-2005-04
Empirical work in economics stresses the importance of unobserved firm- and person-level characteristics
in the determination of wages, finding that these unobserved components account for the overwhelming
majority of variation in wages. However, little is known about the mechanisms sustaining these wage di'er-
entials. This paper attempts to demystify the firm-side of the puzzle by developing a statistical model that
enriches the role that firms play in wage determination, allowing firms to influence both average wages as
well as the returns to observable worker characteristics.
I exploit the hierarchical nature of a unique employer-employee linked dataset for the United States,
estimating a multilevel statistical model of earnings that accounts for firm-specific deviations in average
wages as well as the returns to components of human capital - race, gender, education, and experience -
while also controlling for person-level heterogeneity in earnings. These idiosyncratic prices reflect one aspect
of firm compensation policy; another, and more novel aspect, is the unstructured characterization of the
covariance of these prices across firms.
I estimate the model's variance parameters using Restricted (or Residual) Maximum Likelihood tech-
niques. Results suggest that there is significant variation in the returns to worker characteristics across
firms. First, estimates of the parameters of the covariance matrix of firm-specific returns are statistically
significant. Firms that tend to pay higher average wages also tend to pay higher than average returns to
worker characteristics; firms that tend to reward highly the human capital of men also highly reward the
human capital of women. For instance, the correlation between the firm-specific returns to education for
men and women is 0.57. Second, the firm-specific returns account for roughly 9% of the variation in wages
- approximately 50% of the variation in wages explained by firm-specific intercepts alone. The inclusion of
firm-specific returns ties variation in wages, otherwise attributable to firm-specific intercepts, to observable
components of human capital.
View Full
Paper PDF
-
Total Error and Variability Measures for the Quarterly Workforce Indicators and LEHD Origin Destination Employment Statistics in OnTheMap
September 2020
Working Paper Number:
CES-20-30
We report results from the first comprehensive total quality evaluation of five major indicators in the U.S. Census Bureau's Longitudinal Employer-Household Dynamics (LEHD) Program Quarterly Workforce Indicators (QWI): total flow-employment, beginning-of-quarter employment, full quarter employment, average monthly earnings of full-quarter employees, and total quarterly payroll. Beginning-of-quarter employment is also the main tabulation variable in the LEHD Origin-Destination Employment Statistics (LODES) workplace reports as displayed in On-TheMap (OTM), including OnTheMap for Emergency Management. We account for errors due to coverage; record-level non response; edit and imputation of item missing data; and statistical disclosure limitation. The analysis reveals that the five publication variables under study are estimated very accurately for tabulations involving at least 10 jobs. Tabulations involving three to nine jobs are a transition zone, where cells may be fit for use with caution. Tabulations involving one or two jobs, which are generally suppressed on fitness-for-use criteria in the QWI and synthesized in LODES, have substantial total variability but can still be used to estimate statistics for untabulated aggregates as long as the job count in the aggregate is more than 10.
View Full
Paper PDF