Worth Waiting For? Delayed Compensation, Training and Turnover in The United States and Japan
Summary (2 min read)
I. Introduction
- This paper tests two basic predictions of human capital theory: (1) establishments that provide high levels of training will have high returns to tenure; and (2) establishments with high returns to tenure will have low rates of turnover.
- These predictions are tested with a unique data set containing information on over 4000 employees of more than 80 manufacturing plants in the United States and Japan.
- The predictions of human capital theory are not supported in either the United States or Japan.
- Furthermore, establishments with high returns to tenure do not have low levels of turnover.
- At the individual level, workers who report that their jobs keep them learning new things have lower intentions to quit, as predicted by human capital theory.
II. Literature Review
- Several basic implications of the human capital model are summarized by Jacob Mincer and Yoshio Higuchi:.
- That greater volumes of job training imply steeper wage profiles on the job and over longer work experience is a theorem in human capital analysis.
- For firm-specific human capital, wages increase with tenure as a way to reduce turnover.
- They find that separations are low in industries with high returns to tenure (Mincer and Higuchi, 1988: 110) .
- Finally, Mincer and Higuchi use cross-sectional differences in the returns to tenure across industries; the current study examines differences in the returns to tenure across establishments, the unit of analysis relevant for establishment-specific human capital.
III. The Data
- The data are from 1982-83 surveys of manufacturing establishments in the Indianapolis area in the US and from the Atsugi region (outside of Tokyo) in Japan.
- Within this population, organizations were stratified by employment size and by industry, and randomly selected.
- In Japan the wage measure is the log of annual earnings, including the annual bonus and small bonuses based on family characteristics.
- Of the training measured by TRAIN may not be OJT, but may already have occurred before hiring.
- Three measures for turnover are available in this data set.
IV. Results Using Wage Equations
- Table 2 presents wage equations without OJT controls for the United States and for Japan, also known as Baseline wage equations.
- The magnitude of the training effect on wages increases their confidence that the OJT variables are valid measures of human capital.
- Furthermore, the industry effects (columns 1 and 3) and the firm effects (columns 2 and 4) remain virtually unchanged after controlling for OJT.
- In Japan, the coefficient is small and not significant.
- In the United States, establishments with high returns to tenure had statistically significantly more absences-the opposite of the predictions of Lazear's theory that high returns to tenure increase performance (Table 6 , column 3).
VI. Conclusions and Further Research
- Human capital theory is the foundation of contemporary neoclassical labor economics.
- While it has had numerous successes in predicting labor market phenomena, it has been less successful in the tests presented in this paper.
- Contradicting human capital theory, there is no evidence in either the United States or in Japan that OJT measures explain past wage anomalies.
- In neither country do plants with high returns to tenure provide above-average levels of training.
- Finally, there is no evidence that plants with high levels of average OJT enjoy lower average turnover.
Did you find this useful? Give us your feedback
Citations
200 citations
183 citations
Cites background or methods from "Worth Waiting For? Delayed Compensa..."
...0734-306X/99/1702-0002$02.50 298 / 9e14$$ap05 03-05-99 10:27:22 laeca UC: Labor Econ Barron, Black, and Loewenstein 1988; Brown 1989; Altonji and Spletzer 1991; Lynch 1991, 1992; Barron, Berger, and Black 1993; Levine 1993; Blanchflower and Lynch 1994; and Loewenstein and Spletzer 1995) ....
[...]
...As Levine (1993) recognizes, it may be that such data are too coarse to pick up that effect if it is there....
[...]
...Mincer (1988), Brown (1989), Barron, Black and Loewenstein (1989), Barron, Berger and Black (1993), and Levine (1993) have attempted to do so with training data that were of a qualitative and subjective nature while Lynch (1992) and Blanchflower and Lynch (1994) used a data set, the National Longitudinal Survey of Youth (NLSY), that has the actual number of weeks spent training....
[...]
153 citations
136 citations
[...]
135 citations
References
3,238 citations
1,699 citations
1,493 citations
1,142 citations