Page 1 of 8
Archives of Business Research – Vol. 10, No. 1
Publication Date: January 25, 2022
DOI:10.14738/abr.101.11500. Tannen, M. (2022). Starting and Subsequent Earnings of Young Workers in Different Sized Firms. Archives of Business Research,
10(01). 43-50.
Services for Science and Education – United Kingdom
Starting and Subsequent Earnings of Young Workers in Different
Sized Firms
Michael Tannen
School of Business and Public Administration
University of the District of Columbia
4200 Connecticut Ave NW, Washington DC 20008
ABSTRACT
Based upon the highly innovative Project TALENT Longitudinal Database, this study
tracks the starting earnings and subsequent early earnings growth of young males
who began their work careers at either a smaller (<100 employees) or larger
private firm more than a generation ago. Prior evidence based upon less rich
databases found that earnings were systematically higher in larger firms but did not
have access to many other variables that could affect projected earnings which are
available in the TALENT database. Earnings regressions are estimated here
including not only usual explanatory variables of years of schooling and labor
market experience, but also adding other variables pertaining to prior job
experience, military service, IQ, socioeconomic background and some other factors.
The findings indicate that while starting earnings of those in this database were
indeed higher in larger firms, the gap evaporated fairly quickly with projected
earnings of those in smaller firms featuring a small earnings premium. The results
here suggest guidance based upon the body of prior evidence may have been less
reliable than thought, and that evidence itself may not provide as useful a baseline
as desired for subsequent research addressing whether this pattern continues for
recent cohorts.
Keywords: Young Worker Pay, Firm Size and Earnings, Early Career Advancement
Should a youthful job candidate be concerned about differing pay prospects according to the
size of firm he/she is considering working for? A seeming wealth of evidence from a generation
ago [5], [8], [9], [10], [11], [14], [15] indicated ‘yes’ because earnings were apparently higher in
larger firms. Much of this evidence, though, indicated neither whether starting earnings were
higher in larger firms, grew more rapidly there, or both. Studies also relied upon databases
containing limited information on productive characteristics of individual workers, which
might have differed by employer size. Guidance inferred from this body of evidence may have
been less reliable than thought, and the evidence itself may not provide as useful a baseline as
desired for subsequent research addressing whether this pattern continues for recent cohorts.
In this paper Project TALENT, an early rich but underutilized (in business and economics)
national longitudinal database of high school graduates is used to examine the relationship
between firm size and early earnings profiles, correcting for limitations noted above.
i
Surveys
of the same cohort of individuals, one conducted shortly after high school graduation and a
followup five years later contained a question whether employed individuals worked for a firm
of greater than 100 employees, a representation of a large firm at the time. ii Using this
information, starting and subsequent earnings for young men in larger and smaller firms can
Page 2 of 8
44
Archives of Business Research (ABR) Vol. 10, Issue 1, January-2022
Services for Science and Education – United Kingdom
be compared using a ‘human capital’ earnings specification to provide a better indication than
is presently available of how early career prospects are intertwined with size of firm.
THE MODEL
The analysis begins with a semilogarithmic earnings specification below developed by Mincer
(1974):
(1) Ln Y = b0 + b1S + b2X +b3X2 + u
where Y represents individual earnings, S is the number of school years completed, and X is
number of years of labor market experience. The b’s are parameters to be estimated, while u
serves as the residual term.
Focusing on the behavior of earnings while employed by the present firm, experience is
decomposed into length of service with the current employer (TENURE) and labor market
experience acquired prior to that time (PRIOR), giving us:
(2) Ln Y = b0 + b1S + b2 PRIOR +b3 PRIOR2 + b4 TENURE + b5 TENURE2 + u
To then illustrate the relationship between earnings profiles and size of firm, TENURE is
separated into variables that apply if an individual is employed by a larger firm or by a smaller
one, yielding:
(3) Ln Y = b0 + b1 S + b2 PRIOR +b3 PRIOR2 + b4 TENLRG + b5 TENLRG2 + b6 TENSM
+ b7 TENSM2 +B8 LARGEMP + u
where TENLRG denotes years of experience with the current large firm and TENSM indicates
experience with the current small one. LARGEMP is a binary variable (coded unity for a large
firm, zero for a small one) added to capture a premium or deficit (intercept value) in starting
pay in a large firm. When equation (3) is estimated the appropriate combination of employer- related coefficients can be used to infer the starting and subsequent shapes of earnings profiles
in different sized firms.
Now by setting each of the tenure variables in equation (3) equal to zero, the following
relationship describing starting earnings appears:
(4) Ln Yt = b0 + b1 S + b2 PRIOR + b3 PRIOR2 + b8 LARGEMP + u’
Where t denotes the time period in which employment with the current firm began. The
coefficient on the LARGEMP variable provides the desired estimate of the effect of firm size
upon starting earnings.
Subsequent earnings growth can then be described by subtracting equation (4) from equation
(3). This yields:
(5) Ln Y – Ln Yt = b4 TENLRG + b5 TENLRG2 + b6 TENSM + b7 TENSM2 + u”
which describes earnings behavior in terms of typical ‘human capital’ explanatory variables,
but does not yet take into account additional factors or characteristics that might systematically
affect earnings.
Equations (4) and (5) form the basis of the empirical investigation reported on below.
Page 3 of 8
45
Tannen, M. (2022). Starting and Subsequent Earnings of Young Workers in Different Sized Firms. Archives of Business Research, 10(01). 43-50.
URL: http://dx.doi.org/10.14738/abr.101.11500
DATA AND NECESSARY MODIFICATIONS
To estimate this equation, detailed and nationally representative information on individual
earnings, firm size, experience on the present and prior jobs and school attainment (years) is
required. Additional information on worker quality and institutional characteristics (as
proxied by personal attributes ̧ union membership and various occupational practices) is also
desired to remove the contaminating influence of these intervening factors. Project TALENT
has been selected for this study because it contains most of the information necessary to
support such an investigation.
A nationwide (U.S.) probability sample of 375,000 high school students were involved in a two- day aptitude and achievement testing program. iii The tests were supplemented by a
questionnaire which inquired about high school performance and family situation. Followup
surveys were later mailed to these individuals. In the one, administered five years after
scheduled high school graduation, individuals were asked to provide information concerning
subsequent education, occupation and employment. Some questions addressed how long they
have been with their current employer and what were there present and starting earnings
there. They were also asked to classify their employer according to designated type. Those
who worked in the private sector needed to identify whether they were self-employed, worked
for a family business, a large firm (100 or more employees) or a smaller one. Only those who
worked for a for-profit employer but were neither self-employed nor worked for a family
business are included in the sample used here.
The richness of the database allows for the expansion of the model in equation (5) to include
additional factors that can affect starting pay and earnings growth. Variables have been added
to improve the efficiency of the estimates and remove some potential sources of bias. These
emanate from three sources: (a) the influence of unions; (b) systematic differences in the
qualifications of workers hired by small and large firms; and (c) data deficiencies associated
with the prior experience measure.
By limiting access to certain occupations, some unions can raise starting and subsequent
earnings. This effect might be substantially greater in large firms, where union influence can
pervade a broader and longer occupational ladder. Perhaps the most pronounced effect of this
type, however, is likely to occur in apprenticeship programs. A dummy (binary) variable
corresponding to whether a worker had participated in an apprenticeship program has been
added to the explanatory variables in equation (5).
The union influence on earnings, though, can influence earnings more broadly. To represent
other possible effects, a second variable representing the degree of unionization in particular
occupations is also added. This is the percent of male workers aged 25 or less in each
occupation who are union members. Actual information on union membership is not available
in the Project TALENT database and were inferred using U.S. Census Current Population Sample
(CPS) data.
Turning now to systematic differences in worker qualifications firms which plan to invest more
in their employees (usually perceived as occurring in larger firms) have more to lose by not
selecting carefully among applicants. They are more likely to place greater emphasis on