Performance Management Evaluation And Incentives And Harvard

Corporate Restructuring and Wage Inequality

John C. Dencker [email protected] 217-333-2383 Chichun Fang [email protected] 217-265-0953

Institute of Labor and Industrial Relations University of Illinois at Urbana-Champaign 504 East Armory Avenue Champaign, IL 61820

ABSTRACT We argue that corporate restructuring increased wage inequality in recent decades by eroding organizational and labor market structures in two key ways: through reductions in force—which increased the degree to which market forces influenced wages—and through the transformation of payment systems, practices, and policies—which reduced the degree to which pay depended on seniority. We assess this claim by analyzing twenty-five years of personnel files from a Fortune 500 energy sector firm that restructured multiple times in the 1980s and 1990s. Our findings indicate that much of the increase in wage inequality in the firm traced to a large decline in starting salaries for hourly employees beginning in the 1980s, coupled with an increase in wages for managers and professionals. Our results also reveal that wage inequality increased as a result of a decline in returns to job tenure, particularly in years following the change in the firm’s performance management system. These findings suggest that employee power over wage setting systems, practices, and policies plays an important role in wage inequality in recent decades, as do generational factors. Overall, the findings suggest that a critical mechanism leading to increased wage inequality in recent decades were the decisions by firms, and how they responded not only to pressures on wages from the external market, but also on pressures from shareholders to change the way employees were rewarded. We conclude by discussing the implications of our findings for research on stratification, labor markets, and organizations.

INTRODUCTION Scholars have bemoaned the lack of sociological research on the well-known but not fully understood increase in wage inequality in recent decades (cf. Morris and Western 1999; Myles 2003; Nielsen 2007). Given that firms are the main arena for matching wages to employees (cf.

1

Sørensen 1994), and the importance of firms in influencing stratification (cf. Baron 1984), we might expect the study of firms to play a prominent role in current sociological research on wage inequality. Yet, despite studies showing that organizations had a non-trivial influence on wage inequality in recent decades (Batt 2001; Fernandez 2001), there has been little assessment of the effects of structure and institutions on wages at the firm level, a somewhat surprising outcome since organizations were transformed substantially by restructuring initiatives from the early 1980s to the present (cf. Cappelli, Bassi, Katz, Knoke, Osterman and Useem 1997)—at the same time that wage inequality was increasing. In particular, despite important theoretical and empirical research on effects of restructuring on wage losses among laid off employees (cf. Baumol, Blinder and Wolff 2003; Farber 2003; McCall 2004; Sørensen 2000), there is little research on wage inequality among survivors of restructuring, and few studies that assess the mechanisms by which restructuring might lead to increased wage inequality. In this article, we argue that corporate restructuring eroded organizational and labor market structures protecting employees from layoffs, and develop a framework to specify the ways in which restructuring lead to increases in wage inequality. In particular, we maintain that two forms of restructuring had a key influence on temporal patterns in employee earnings: corporate reductions in force (RIF) that involved layoffs of previously protected groups such as managers in blue chip companies (Cappelli 1992; 2000); and the reorganization of performance management systems, practices, and policies, wherein institutional rules governing wage increases such as seniority-based pay increases were replaced with pay-for-performance systems (Cappelli et al. 1997). We analyze our claims using data from longitudinal personnel files of employees in a Fortune 500 manufacturing firm for the period 1969 to 1993.

2

ORGANIZATIONAL DETERMINANTS OF WAGE INEQUALITY Prior to the onset of corporate restructuring in the early 1980s, careers were influenced strongly by internal labor markets (ILMs)—sets of rules and processes whereby employment and wage decisions were made within firms rather than through a reliance on the external market (Doeringer and Piore, 1971). In ILMs, employees were buffered from market competition with employment separation decisions the right of the employee rather than a firm (Sørensen and Kalleberg, 1981). In ILMs, jobs existed independently of the persons that occupied them (White 1970), and were linked hierarchically, with employees typically beginning their careers at entry level portals, and progressing upwards through promotions over time (cf. Rosenfeld 1992). Wage growth in organizations prior to restructuring was dependent on an employee’s upward mobility rates, as well as on salary increases within jobs. In particular, human resource managers assessed jobs according to their worth and placed jobs of equal value into hierarchical grade levels, with each level having a salary range attached to it (cf. Gerhart and Rynes 2003). Pay decisions were structured as well, with pay increases dependent on seniority in a job. As a result, pay rose more rapidly with time in a grade level than performance did (cf. Medoff and Abraham 1980, 1981), perhaps as a result of disincentives to employees who were passed over for promotion multiple times (Gibbs 1995). In short, in ILMs, an employee’s wages were a function not only of his or her ability to move up the organizational job ladder, but also to earn salary increases within a given job. Corporate Restructuring and Wage Inequality. The widespread and ongoing corporate restructuring process began in the U.S. in the early 1980s, and has continued to the present. During this period, most firms engaged in some form of organizational change, with many doing so multiple times (Cascio, Morris, and Young 1997). Corporate restructuring

3

represented a significant change to the nature of ILMs, although many structural features of these systems remained in large firms following restructuring. In particular, restructuring eroded key structural features of organizations in two main related ways: through corporate RIF, and through the transformation of performance and reward systems, practices, and policies. Firms engaging in RIF eliminated guarantees of protection against layoff, making continued employment a function of market rather than non-market factors (cf. Cappelli, 1992). Due to RIF, labor market competition was brought to bear on the employment relationship, as firms had greater flexibility in replacing a manager if they could find a more productive one at a given wage rate in the external market. By the same token, some surviving employees likely benefited from the increased incidence of market forces on wages. For instance, employees in high demand in the external market could command higher wages from their employers. In other words, the increase in market forces would reduce the effect of job structures on wages, and increase the influence of individual characteristics and demand for certain types of employees on wage inequality. Firms transforming their reward and appraisal systems also reduced the degree of structure inherent in ILMs. In particular, changes to reward systems increased the degree to which pay depended on short-term performance, thereby making pay much more variable than it was in ILMs (cf. Cappelli et al., 1997). That is, restructuring firms replaced their previous reliance on seniority in a firm and/or job in wage increase and promotion decisions with measures of employee performance. If corporate restructuring increased wage inequality, it should show up in a number of ways. First, employees whose jobs are less valued on the external market relative to their current salary should experience a decline in real wage over time, assuming they are not laid off during a RIF. That is, if similarly productive employees can be found in the external market at a cheaper

4

wage rate, there will be downward pressures brought to bear on wages of these employees. Second, employees who are in high demand in the external market may find substantial upward pressures on wages, with this effect magnified by changes in performance evaluation systems, as firms seek to retain highly demanded and highly productive employees. Third, one might expect to see a declining effect of starting salary on current salary over time. That is, the staying power of starting wages on current salary for a given cohort—holding for many years in firms characterized by ILMs (cf. Baker, Gibbs, and Holmstrom 1994)—likely dissipated as a result of restructuring. Finally, the effect of restructuring on wage inequality should show up in increasing wage dispersion within cohort groups, as pressures on wages for employees perceived to be more productive either in the external market and/or within firms experience significant wage increases, and as the elimination of seniority based wage increases results in a declining real wage for the least productive employees.

DATA, MEASURES, AND METHODS In order to assess whether corporate restructuring increased wage inequality, we analyze confidential personnel files obtained from a Fortune 500 manufacturing firm for the period 19691993. We also draw on information collected from corporate documents and semi-structured interviews conducted with several of the firm’s human resource managers at the firm’s main corporate office. The firm was a large, diversified energy sector firm. Like most other large firms, its ILM contained a salary grade level (SGL) system that consisted of jobs hierarchically ranked into grades to which salaries were attached. Non-exempt employees (e.g., secretarial, support staff) were in SGL 1 to 9. By contrast, exempt employees (e.g., managers and professionals) were in SGL 7 to 24. The exempt SGL system had common entry ports, with

5

individuals having bachelors’ degrees typically placed into level 7, and individuals with masters’ degrees placed into level 8. In addition, roughly 25 percent of the employees in the firm were paid on an hourly basis, with roughly one-third of these employees belonging to a union. Once hired, employees in the SGL system moved from lower to higher graded jobs through promotions that were based on relative performance, with a move to a job in a higher SGL representing a promotion, and a move to a job in the same level representing a transfer. Wages were structured as well, although there was some variation over time in the degree to which they were recurring. In particular, two years prior to the first RIF, the firm instituted an incentive pay program that allowed managers to award subordinates with non-recurring bonuses that depended on performance. The personnel files cover a long time period to provide a useful test for assessing how a large firm rewarded its employees in changing economic contexts. For instance, the records cover the period of the oil shocks in the early 1970s, high levels of employment growth in the late 1970s, recession in the early 1980s, and the restructuring period from the mid 1980s onward. During the restructuring period, the firm undertook two RIF. The first RIF occurred in the mid 1980s during a time of hostile takeover activity. The firm was not taken over in this period, nor was there any indication that it was a takeover target, although a number of the firm’s competitors were, a common pattern in mature industries (Cappelli et al., 1997: 33). The second RIF occurred in the early 1990s, subsequent to regulatory changes limiting takeovers, but during a period when institutional investors sought to exert their increasing power by pressuring firms to change the way they were managed (cf. Useem, 1996). In the interim between the two RIF, the firm transformed its performance management system. It sent senior managers to other firms to study the changes that they made, and hired

6

consultants to help design and implement the new performance management system. The firm transitioned from a seniority-based system to one in which pay was contingent on a manager’s performance, a broad action similar to that undertaken by other large firms in the time period. In doing so, the firm sought to make performance objectives measurable, attainable, and relevant. As part of the performance management change, performance records were eliminated soon after pay decisions were made. According to the firm, this decision was enacted in order to minimize potential bias in future performance rankings, in that prior performance in theory would be less likely to be taken into account in measuring current performance. For instance, by eliminating performance records, the firm sought to remove the problems that arise from labeling employees, which in turn would also ensure that wage increases were ‘re-earned” in each year. Data Set. The firm provided a 25% random sample of full career records of the firm’s U.S. employees tracing from 1967.1 For some employees, information on salary was missing in the earliest years of the sample. Because including these employees in the sample to be analyzed can lead to a survivorship bias (Petersen, 1995), we follow convention (cf. Petersen and Saporta, 2004) and study wage inequality only for managers whose careers could be traced from their initial entry in the firm beginning in 1969, resulting in a sample of 20,480 employees. Dependent Variable. Our dependent variable is the log of an employee’s salary, measured in 2007 constant dollars, with wage records updated each month based on the Consumer Price Index (CPI) Deflator. Independent Variables. In order to capture effects of restructuring on wages and wage inequality, we use several time varying variables. In particular, we examine wages in five equally-spaced time frames: 1969-1973, 1974-1978; 1979-1983, 1984-1988, and 1989-1993.

1

The firm also provided a data set containing information for all employees who were in the firm in 1967 or who entered at some point in time afterwards. However, this data set does not include career information.

7

The 1984-1988 time period contains the first RIF, whereas the 1989-1993 period contains both the second RIF and the transformation in the firm’s performance management system. We also provide graphs of real wages for each year from 1969 to 1993. We assess variation in wages for a number of different independent variables over time. We assess job type using a categorization of the firm’s SGL system, and with measures for hourly employees. Due in part to a lack of managers in a number of SGL in a given year, we grouped SGL that were similar on many dimensions. For exempt employees, we use the following scheme: levels 7 to 9 (entry managers); levels 10 to 12 (middle managers); levels 13 to 16 (upper middle managers); and levels 17 to 24 (upper level managers). Discussions with managers and an inspection of the data set helped me to create the salary grade level groupings. Main results were robust to models that included all salary grade levels. We grouped nonexempt grade levels as follows: levels 1 to 3 (entry level), levels 4 to 6, and levels 7 to 9. We also assess wage increasing with increasing time spent in a job level, with this measure updated monthly. We examine effects of starting salary on current salary with a measure of the log of real wages at time of hire. Lastly, because rates of departure can have an influence on wage inequality, we control for employment separation using a dummy measure capturing whether an employee departed the firm for any reason during a given time frame (coded one), or instead was censored (coded zero). Control Variables. We control for demographic, organizational, and human capital variables common in studies of wage inequality in large firms (cf. Petersen and Saporta, 2004): time spent in a job level, age, education, occupation, tenure in the firm, salary grade level, race, and sex. Age, and tenure are time varying and updated in each person-period of the sample. Education dummies were created to reflect whether or not an employee had a HS or HS

8

equivalency degree, a bachelors’ degree, or a post-graduate degree (e.g., MBA, PhD). Sex (female) was coded one for female managers and zero for male managers. Race (minority) was coded one for minority managers, and zero otherwise. We also controlled for union status. We also consider how a manager’s performance influences his or her chances of departure. As noted, the firm eliminated all records of these evaluations. We therefore consider several performance proxies. Following Gibbs (1995), we argue that the lowest performing manager in a level have the most experience in that level. Since this duration in a grade level measure does not perfectly reflect performance, we also examine another measure of performance, namely whether a manager had ever been demoted during his or her career. These performance proxies are time-varying, and updated in each month. Methods. For each employee in the firm, we generated a monthly event history file (cf. Allison 1982). These records were updated for each career change. For instance, if an employee was promoted, we updated his or her salary grade level change, as well as the associated wage change. As noted, nominal wages were updated to real wages in each month based on the CPI deflator. We use OLS regression to measure effects of our independent and control variables on log salary, and provide robust standard errors, clustered by employee.

RESULTS Table 1 provides descriptive statistics for employees at initial hire, and highlights a number of trends in entry-level wages. For non-exempt employees, real wages were generally stagnant or declining over time. For hourly employees, there was a sharp decline in starting wages beginning in the early 1980s, with both union and non-union entrants experiencing sharp reductions in wages throughout the 1980s and 1990s. In particular, real wages for unionized new

9

hires declined by more than $12,000 from 1984-1988 relative to the previous 5 year time frame, whereas real wages for non-union new hires declined by more than $16,000 in this period. For exempt entry level employees, starting real wages were roughly constant over time. However, for the few employees hired into middle management positions, real wages tended to increase over time. These patterns are shown graphically in Figure 1, which provides the mean starting salary (in 2007 US$) for hourly, exempt (SGL 7-24), and non-exempt (SGL 1-9) in each year from 1969 to 1993. In particular, Figure 1 shows that hourly employees starting real salaries increased slightly from 1969 to 1980, after which time it declined steadily for most of the 1980s. Non-exempt employees by contrast, experienced a slight increase in starting real wages over time, whereas exempt employees experienced a slight decrease in starting real wages throughout the 1970s, followed by an uptick in the early 1980s. ---Insert Table 1 and Figure 1 about here--Figure 2 provides information on average real wages for employees in different job groups over time (for new entrants to the firm from 1969 to 1993). Similar to the patterns in Figure 1, it shows that hourly employees experienced a decline in real wages, with the largest decrease coming from the early to mid 1980s to 1990. In addition, non-exempt employees experienced a slight constant increase in real wages over time. By contrast, exempt employees experienced a large steady increase in real wages from the late 1960s to the mid 1980s, with the biggest increase occurring following the recession of the early 1980s. In addition, Figure 3 shows that there were substantial differences in real wages of union and non-union hourly employees from the early 1980s onward, with both experiencing substantial declines, but with wages of non-union employees dropping at a much faster rate. ---Insert Figures 2 and 3 about here---

10

Table 2 provides results for regressions predicting the log of real wages for all new entrants to the firm from 1969-1993. Column 1 includes the controls measures, and shows that, as expected, wages are generally an increasing function of grade level, albeit with hourly employees paid at a much higher rate than entry level non-exempt SGL employees. Column 2 adds effects of entry year dummy measures, and shows that real wages upon entry are declining over time. Unreported supplemental analyses reveal that much of this pattern traces to the experience of hourly employees. ---Insert Table 2 about here--Column 3 of Table 2 adds the year dummy measures. It shows that real wages declined in increasing year, although, like the effect for entry year, much of the decline stemmed from the experience of hourly employees. Column 4 of Table 2 adds the measure for starting real wage. It shows that much of the effect of entry year on current wage was a function of the wage at hire, a pattern holding both for analyses of hourly and exempt/non-exempt sub-samples (results not reported). Nevertheless, the decline in real wages over time held even with the control for starting salary, particular among hourly employees. Column 5 of Table 2 adds the interaction between the departure measure and the year dummies to the variables in Column 4. Results reveal that the negative effect of departure on the log of real wages was strongest from the mid 1980s to 1993, during the period of restructuring. Column 6 of Table 2 adds the interaction between the tenure in job measure and the year dummies to the variables in Column 4. Findings indicate that there was a strong negative effect of time in a job on wages from the early 1980s to 1993, with the effect growing more negative over time. In other words, returns to seniority were declining substantially over time, with the strongest negative effect occurring during the period of corporate restructuring.

11

Finally, (unreported) exploratory analyses of measures of wage inequality such as the variance in log wages (cf. Mouw and Kalleberg 2007), and quantile regression (cf. Morris, Bernhardt, and Handcock 1994) suggest that restructuring increased within-cohort inequality, due not only to the reduction in starting salaries for hourly employees, but also to the decline in effects of seniority in a job on wage growth along with increasing returns for employees who are promoted at a relatively more rapid rate.

DISCUSSION In this article, we developed a framework to explain the effects of corporate restructuring on wage inequality. Our findings indicate that corporate restructuring reduced effects of labor market structures on wages, such as those related to starting wages of hourly employees. These patterns help to explain why wage inequality began to increase in the early 1980s. In particular, our findings show a considerable decline in starting wages of hourly employees throughout the 1980s, a general increase in starting salaries for exempt employees for part of this time frame— and stagnant wages for non-exempt employees. Our findings have a number of important implications for research on inequality, organizations, and labor markets. For instance, they reveal that a non-trivial percentage of the increase in wage inequality from the 1980s onward was due to the elimination of labor market structures and institutions, especially those related to the barriers between employees and the external labor market—that is, an ILM—and transformations in rules governing wage setting. In other words, absent the erosion of these structures, wage inequality arguably would have been lower, as might be expected in firms in countries that preserved institutional features of labor markets, as was the case in Continental Europe (cf. DiPrete 2005).

12

Our findings also shed light on debates about efficiency wages and inter-industry wage differentials (cf. Krueger and Summers 1988). That is, the firm we studied reduced starting wages for hourly employees, while maintaining or increasing starting salaries for SGL employees. Moreover, results also suggest that the firm did not limit wage inequality within cohort and employee groups as in the past. Both of these outcomes are consistent with findings that industries previously characterized by high wages for all employee groups, as was the case in the industry of the firm we studied, experienced a decline in real wages. Thus, our findings suggest that notions of equity within organizations have disappeared in recent decades, and raise important concerns regarding generational and occupational differences in work outcomes. Our results are consistent with economic accounts such as the recent study by Lemieux, MacLeod, and Parent (2007) who find that pay-for-performance systems account for roughly one-fourth of the growth in variance in male wages from the late 1970s to the early 1990s. However, they counter recent claims of labor economists that cohort effects (i.e., the staying power of starting salary on current salary over periods of time) trace to task-specific human capital (Gibbons and Waldman 2006), rather than to institutional effects such as seniority based payment practices. Limitations and Directions for Future Research. The findings reported in this article stem from one large firm, raising questions regarding potential generalizability. Several factors reduce these concerns. First, results are largely similar to findings from other large firms in the 1970s (cf. Lazear 1992; Petersen and Saporta 2004). Second, like most large firms, the firm we studied restructured multiple times (cf. Cascio, Young, and Morris 1997), and relied on external advice from consultants on the design and implementation of restructuring initiatives, as well as from senior managers sent to other firms to examine best practices for restructuring.

13

Unfortunately, as noted, we do not have records of employee performance, which would provide additional insight into the determinants of wage inequality within firms over time. Thus, for instance, we can not say for certain that effects of performance on earnings would vary across racial and ethnic groups, as recent research suggests (Castilla, forthcoming), or of whether firms employing forced performance curves create a disjoint between wages and productivity. In addition, the lack of performance records somewhat preclude our ability to assess claims that unobserved ability is a key driver of wage inequality as economic accounts indicate (Juhn, Murphy and Pierce 1993), although they do lend support to the notion that sociologists must take such possibilities into account in their explanations for inequality (Nielsen 2007).

DISCUSSION In this article we responded to calls for a sociological approach to explain the well-known but little understood increase in wage inequality from the 1980s to present. Our firm-level framework highlights the importance of corporate restructuring in reducing effects of labor market structures and institutions on wages, which in turn leads to increased inequality among employees in restructuring firms. Our findings suggest that the firm we studied increasingly relied on market forces to govern employment outcomes, albeit with substantially different effects for different employee groups. Thus, although job and wage structures continue to play a key role in employee earnings in contemporary firms (cf. Goldthorpe 2000)—as suggested by differences in wages for union and non-union hourly employees—their ability to limit wage inequality has been substantially reduced, as perhaps suggested by the experience of firms in countries that preserved institutional features of labor markets in recent decades.

14

REFERENCES Allison, Paul D. 1982. “Discrete-Time Methods for the Analysis of Event Histories.” Pp. 61-98 in Sociological Methodology, Vol. 12, edited by S. Leinhardt. San Francisco: Jossey-Bass. Baker, George F, Michael Gibbs, and Bengt Holmstrom. 1994. “The Wage Policy of a Firm." Quarterly Journal of Economics, 109, 921-55. Baron, James N. 1984. “Organizational Perspectives on Stratification.” Annual Review of Sociology, 10-37-69. Batt, Rosemary. 2001. “Explaining Wage Inequality in Telecommunications Services: Customer Segmentation, Human Resource Practices, and Union Decline.” Industrial and Labor Relations Review 54: 425-449. Baumol, W., A. Blinder, and E. Wolff. 2003. Downsizing in America: Reality, Causes, and Consequences. New York: Russell Sage Foundation. Cappelli, Peter. 1992. “Examining Managerial Displacement.” Academy of Management Journal, 35, 203-217. _____. 2000. “Examining the incidence of downsizing and its effect on establishment performance.” NBER Working Paper # 7742. Cappelli, Peter, Laurie Bassi, Harry Katz, David Knoke, Paul Osterman and Michael Useem, Eds. 1997. Change at Work. New York: Oxford University Press. Cascio, Wayne. F., Clifford Young, and James R. Morris. 1997. “Financial Consequences of Employment Change Decisions in Major US Corporations.” Academy of Management Journal 40: 1175-89. Castilla, Emilio J. Forthcoming. “Gender, Race, and Meritocracy in Organizational Careers.” American Journal of Sociology DiPrete, Thomas A. 2005. “Labor Markets, Inequality, and Change?: A European Perspective.” Work and Occupations 32: 119-139. Doeringer, Peter, and Michael Piore. 1971. Internal Labor Markets and Manpower Analysis. Lexington, MA: Heath. Farber, Henry. 2003. “Has the Rate of Job Loss Increased in the Nineties.” Working Paper No. 394. Princeton University, Industrial Relations Section. Fernandez, Roberto. 2001. “Skill-Biased Technological Change and Wage Inequality: Evidence from a Plant Retooling.” American Journal of Sociology, 107: 273-320. Gerhart, Barry, and Sarah Rynes. 2003. Compensation: Theory, Evidence and Strategic Implications. Thousand Oaks, CA: Sage Publications. Gibbons, Robert, and Michael Waldman. 2006. “Enriching a Theory of Wage and Promotion Dynamics inside Firms.” Journal of Labor Economics 24: 59-107. Gibbs, Michael. 2005 “Incentive Compensation in a Corporate Hierarchy." Journal of Accounting and Economics 19: 247-77. Goldthorpe, John H. 2000. “Rent, Class Conflict, and Class Structure: A Commentary on Sørensen.” American Journal of Sociology 105: 1572-1582. Juhn, Chinhui, Kevin M. Murphy, and Brooks Pierce. 1993. “Wage Inequality and the Rise in Returns to Skill.” Journal of Political Economy 101: 410-442. Krueger, A. and L. Summers. 1988. “Efficiency Wages and the Interindustry Wage Structure.” Econometrica, 56: 259-93. Lazear, Edward. 1992. “The Job as a Concept.” In William J. Bruns, Jr. (Ed.), Performance Management, Evaluation, and Incentives. Boston: Harvard Business School Press.

15

Lemieux, Thomas, W. Bentley MacLeod, and Daniel Parent. 2007. “Performance Pay and Wage Inequality.” NBER Working Paper #13128. McCall, Leslie. 2004. “The Impact of Organizational Changes on Aggregate Inequality: The Case of Downsizing.” Working Paper. Medoff, J. and K. Abraham. 1980. “Experience, Performance and Earnings.” Quarterly Journal of Economics 95: 703-736. _____. 1981. “Are Those Paid Really More Productive?” The Journal of Human Resources 16:186-216. Morris, Martina, Annette Bernhardt, and Mark Handcock. 1994. “Economic Inequality: New Methods for New Trends.” American Sociological Review 59: 205-219. Morris, Martina, and Bruce Western. 1999. “Inequality in Earnings at the Close of the Twentieth Century.” Annual Review of Sociology 25: 623-657. Mouw, Ted, and Arne Kalleberg. 2007. “Occupations and the Structure of Wage Inequality in the United States, 1980s-2000s.” Working Paper. Myles, John. 2003. “Where Have All the Sociologists Gone?: Explaining Economic Inequality.” Canadian Journal of Sociology 28: 553-561. Nielsen, François. 2007. “Economic Inequality, Pareto, and Sociology: The Route Not Taken.” American Behavioral Scientist 50: 619-638. Petersen, Trond. 1995. “Analysis of event histories.” Pp. 453-517 in Handbook of Statistical Modeling for the Social and Behavioral Sciences, edited by G. Arminger, C. C. Clogg, and M. E. Sobel. New York: Plenum Press. Petersen, Trond, and Ishak Saporta. 2004. “The Opportunity Structure for Discrimination.” American Journal of Sociology 109: 852-901. Petersen, Trond, Seymour Spilerman and Sven-Age Dahl. 1989. “The Structure of Employment and Terminations among Clerical Employees in a Large Bureaucracy.” Acta Sociologica 32:319-38. Rosenfeld, Rachel. 1992. “Job Mobility and Career Processes.” Annual Review of Sociology 18:39-61. Sørensen, Aage B. 1994. “Firms, Wages, and Incentives.” Pp. 504-28 in The Handbook of Economic Sociology, edited by N. Smelser and R. Swedberg. Princeton: Russell Sage. _____. 2000. “Toward a Sounder Basis for Class Analysis.” American Journal of Sociology 105: 1523-71. Sørensen, Aage B., and Arne Kalleberg. 1981. “An outline of a theory for the matching of persons to jobs.” Pp. 49-74 in Sociological Perspectives on Labor Markets, edited by Ivar Berg. New York: Academic Press. White, Harrison C. 1970. Chains of Opportunity: System Models of Mobility in Organization. Cambridge, MA: Harvard University Press.

16

FIGURE 1 80000

Mean Starting Wage by Year in 2007 Dollars

60000 Real Income

40000

20000

1970

1975

1980

1985

1990

1995

Year Hourly Exempt

Non-Exempt

FIGURE 2 100000

Mean Real Wage for New Entrants from 19691993

80000 Real Wage 60000

40000

20000

1970

1975

1980

1985

1990

1995

Year Hourly Exempt

Non-Exempt

17

40000

45000

50000

Mean Real Wage for Union and Non Union Members

35000

Real Wage

55000

FIGURE 3

1970

1975

1980

1985

1990

1995

Year Union Hourly

Non-Union Hourly

18

TABLE 1. Characteristics of Employees at Initial Hire in a Large U.S. Manufacturing Firm Variable Total Number of New Hires % SGL 1-3 Non-Exempt Average Real Wage ($2007) % SGL 4-6 Non-Exempt Average Real Wage ($2007) % SGL 7-9 Non-Exempt Average Real Wage ($2007) % SGL 7-9 Exempt Average Real Wage ($2007) % SGL 10-12 Exempt Average Real Wage ($2007) % SGL 13-16 Exempt Average Real Wage ($2007) % SGL 17-24 Exempt Average Real Wage ($2007) % Hourly Average Real Wage All Hourly ($2007) Average Real Wage Union Hourly ($2007) % Main Corporate Office % Human Resource Function Age % with bachelor’s Degree % with master’s or Ph.D. Degree % Female % Minority

1969-1973 3,352 26.49% $26,598 33.05% $35,801 0.92% $49,074 13.93% $59,563 3.85% $87,679 0.69% $138,746 0.06% $278,435 21.00% $45,619 $46,989 10.05% 0.57% 27.44 9.76% 4.39% 34.13% 22.34%

1974-1978 4,242 18.69% $24,761 21.64% $35,557 1.74% $52,941 20.89% $61,170 5.23% $91,877 0.64% $133,368 0.05% $269,529 31.12% $48,480 $51,401 16.50% 0.64% 27.51 17.47% 8.13% 32.15% 18.39%

1979-1983 5,658 16.31% $24,731 21.14% $33,619 3.98% $49,704 27.24% $59,108 5.83% $93,929 0.51% $147,005 0.04% $225,840 24.96% $44,232 $49,721 15.59% 0.67% 28.30 24.44% 9.12% 37.31% 14.92%

1984-1988 3,666 13.07% $23,035 17.27% $35,429 3.27% $47,925 20.84% $58,108 6.93% $95,086 0.87% $145,343 0.08% $263,606 37.67% $27,736 $37,393 21.14% 0.92% 29.42 20.29% 7.09% 47.71% 17.73%

1989-1993 3,562 8.87% $24,397 16.56% $35,698 5.87% $43,309 24.65% $56,565 5.95% $96,494 1.12% $154,532 0.14% $215,040 36.83% $25,353 $36,708 24.20% 1.54% 28.90 24.99% 6.43% 49.32% 27.09%

All Years 20,480 16.60% $24,956 21.70% $35,098 3.22% $47,686 22.15% $58,897 5.60% $93,560 0.74% $144,950 0.07% $243,830 29.93% $37,548 $46,077 17.36% 0.84% 28.30 19.95% 7.31% 39.67% 19.47%

19

TABLE 2 OLS Models Predicting Log Real Wages for New Entrants in a Large US Manufacturing Firm, 1969-1993 Variable SGL 1-3 Non-Exempt SGL 4-6 Non-Exempt SGL 7-9 Non-Exempt SGL 7-9 Exempt SGL 10-12 Exempt SGL 13-16 Exempt SGL 17-24 Exempt Departure Job Tenure in Months*12 Log Starting Wage Entry Year 1974-1978 Entry Year 1979-1983 Entry Year 1984-1988 Entry Year 1989-1993 Year 1974-1978 Year 1979-1983 Year 1984-1988 Year 1989-1993 Year 1974-1978*Depart Year 1979-1983*Depart Year 1984-1988*Depart Year 1989-1993*Depart Year 1974-1978*Tenure Year 1979-1983*Tenure Year 1984-1988*Tenure Year 1989-1993*Tenure

Model 1 -.33*** (.00) -.09*** (.00) .16*** (.00) .34*** (.00) .63*** (.00) .94*** (.00) 1.35*** (.00) -.28*** (.00) .005*** (.00)

Constant Df Adj R-squared

10.48*** 27 .78

(.00)

Model 2 -.34*** (.00) -.09*** (.00) .17*** (.00) .34*** (.00) .63*** (.00) .94*** (.00) 1.36*** (.00) -.24*** (.00) .005*** (.00)

Model 3 -.34*** (.00) -.09*** (.00) .17*** (.00) .34*** (.00) .63*** (.00) .94*** (.00) 1.37*** (.00) -.24*** (.00) .005*** (.00)

-.004*** -.046*** -.135*** -.175***

.03*** .02*** -.04*** -.05*** -.01*** -.07*** -.09*** -.13***

10.49*** 31 .79

(.00) (.00) (.00) (.00)

(.00)

10.47*** 35 .79

(.00) (.00) (.00) (.00) (.00) (.00) (.00) (.00)

(.00)

Model 4 -.15*** (.00) -.02*** (.00) .10*** (.00) .19*** (.00) .35*** (.00) .53*** (.00) .79*** (.00) -.14*** (.00) -.014*** (.00) .61*** (.00) .02*** (.00) .03*** (.00) .01*** (.00) .01*** (.00) .00 (.00) -.05*** (.00) -.06*** (.00) -.10*** (.00)

4.06*** 36 .90

(.00)

Model 5 -.15*** (.00) -.02*** (.00) .10*** (.00) .19*** (.00) .35*** (.00) .53*** (.00) .79*** (.00) -.04*** (.00) -.014*** (.00) .62*** (.00) .02*** (.00) .03*** (.00) .01*** (.00) .01*** (.00) .00 (.00) -.05*** (.00) -.06*** (.00) -.10*** (.00) -.06*** (.00) -.07*** (.00) -.15*** (.00) -.14*** (.00)

4.06*** 40 .90

(.00)

Model 6 -.14*** (.00) -.02*** (.00) .10*** (.00) .19*** (.00) .35*** (.00) .54*** (.00) .79*** (.00) -.13*** (.00) .008*** (.00) .61*** (.00) .03*** (.00) .03*** (.00) .00† (.00) -.01*** (.00) -.01*** (.00) -.03*** (.00) -.03*** (.00) -.06*** (.00)

-.003*** -.019*** -.021*** -.025***

(.00) (.00) (.00) (.00)

4.06*** 40 .90

(.00)

Note: Hourly employees are the omitted job-type reference group. Controls for age, seniority in firm, education, promotion, demotion, gender, ethnicity, division, and union status included in all models (results not reported). Sample size in all models is 20,480 employees (1,197,360 employee months). Robust (Huber/White) standard errors in parentheses. †p<.10; *p<.05; **p<.01; ***p<.001 (two-tailed tests).

20