ArticlePDF Available

Paying More to Get Less: Specific Skills, Incomplete Information and the Effects of External Hiring versus Internal Mobility

January 2011
Administrative Science Quarterly 56:369-407

Authors:

Matthew Bidwell

University of Pennsylvania

Summary Statistics for Key Variables *

…

. Ordered Logit Analyses of Performance.*

…

OLS Regressions of Determinants of Pay (N=14,515)*

…

Ordinary Least Squares Regression of Determinants of Compensation by Hiring Source

…

Figures - uploaded by Matthew Bidwell

Content may be subject to copyright.

Content uploaded by Matthew Bidwell

Content may be subject to copyright.

Paying More to Get Less: The Effects of External Hiring Versus Internal Mobility

Matthew Bidwell

University of Pennsylvania

ABSTRACT

Individuals often enter similar jobs via two different routes: internal mobility and external

hiring. I examine how the differences between these routes affect subsequent outcomes in those

jobs. Drawing on theories of specific skills and incomplete information, I propose that external

hires will initially perform worse than workers entering the job from inside the firm and have

higher exit rates, yet they will be paid more and have stronger observable indicators of ability as

measured by experience and education. I use the same theories to argue that the exact nature of

internal mobility (promotions, lateral transfers, or combined promotions and transfers) will also

affect workers’ outcomes. Analyses of personnel data from the U.S. investment banking arm of a

financial services company from 2003 to 2009 confirm strong effects on pay, performance and

mobility of how workers enter jobs. I find that workers promoted into jobs have significantly

better performance than workers hired into similar jobs for the first two years, and lower rates of

voluntary and involuntary exit. Nonetheless, the external hires are initially paid around 18 per

cent more than the promoted workers and have higher levels of experience and education. The

hires are also promoted faster. I further find that workers who are promoted and transferred at the

same time have worse performance than other internal movers.

Worker mobility has increased substantially in recent years as firms have moved away from an

employment model that focused on lifetime employment and internal mobility, and towards

greater reliance on the external labor market to staff positions (Cappelli, 1999; Farber, 2008).

An important consequence of these changes has been to alter the way that workers enter jobs.

Whereas higher-level jobs used to be almost entirely entered by promotions or internal transfers,

those jobs are often now entered by external hiring as well (Royal and Althauser, 2003).

Organizational and sociological approaches to employment have long observed that the

processes that match individuals with jobs affect employment outcomes (Granovetter, 1981).

Existing studies have taken two broad approaches to understanding such matching processes.

One literature has studied the processes that govern mobility in “internal labor markets” within

organizations (Doeringer and Piore, 1971; Althauser and Kalleberg, 1981; DiPrete, 1987;

Dencker, 2009), documenting the importance of formal job ladders, administrative rules, and

worker bargaining power in shaping movements within firms. A second literature has examined

the nature of hiring processes in external labor markets, demonstrating for example that different

job search methods are associated with different probabilities of receiving offers (Fernandez,

Castilla and Moore, 2000) and the importance of various signals of job quality in workers’

decisions to search for and take new jobs (Halaby, 1988; Greve and Fujiwara-Greve, 2003). As

workers increasingly use both internal and external mobility to access similar positions, though,

it is increasingly important to directly compare internal mobility and hiring, to understand

whether those different routes of entry into the job affect such subsequent outcomes as

performance, pay, and mobility, or even the kinds of workers ending up in those jobs.

Such a comparison of the effects of internal and external mobility is of both practical and

theoretical importance. If the route by which jobs are entered affects pay and performance, then

the quality of the human capital that employers are able to acquire and the amount that they pay

for that human capital is likely to depend on whether the workers involved are hired or

developed internally. A clear comparison of the effects of hiring and internal mobility also has

the potential to inform workers of the consequences of their career decisions. Comparing

internal and external labor market processes can also provide us with a richer theoretical

understanding of the trade-offs that characterize these pathways of worker mobility and career

development.

Although a number of studies have begun to explore differences between hiring and internal

mobility, they have stopped well short of developing the detailed comparison of internal and

external matching processes that would underpin a better understanding of these different

mobility processes. Research based on tournament theory has predicted and found that external

hires have faster subsequent promotion rates than workers promoted into the same job (Chan,

2006). A study of academic economists also indicated that lateral hires had stronger publication

records than those hired internally (Oyer, 2007). In their study of a large financial services firm,

Baker, Gibbs and Holmstrom,(1994a) showed that workers hired into higher organizational

levels had more work experience and education than those promoted into those positions, but the

authors made no attempt to control for characteristics of the job that might confound those

comparisons. Within the specific context of Chief Executive Officer (CEO) labor markets, Harris

and Helfat (1997) found that the pay of externally hired CEOs was higher than that of CEOs

promoted internally; however, their cross-firm design limited their ability to control for the

nature of the hiring organization. Yet other work has compared the pay and performance of

workers who move to another organization with those who stay within the same organization

(Brett and Stroh, 1997; Groysberg, Lee and Nanda, 2008). Such work offers great insight into the

prospects of different career strategies but cannot tell whether the gains and losses are due to the

types of jobs accessed or the nature of the matching process. With a few exceptions (Rosenbaum,

1979; Baker, Gibbs and Holmstrom, 1994a), there has also been little research on whether

different routes to jobs within firms, such as promotions or lateral transfers, affect subsequent

outcomes, despite reason to believe that those different routes will result in different outcomes.

To formulate a more comprehensive theory about the differences between external hiring and

internal mobility, it is useful to draw on two theories that are often used in the study of internal

and external labor markets: theories of specific skills and incomplete information. Research on

human resource systems and careers has noted that workers who move jobs within the firm

(“internal movers”) differ from external hires in their levels of firm-specific skills (Sonnenfeld

and Peiperl, 1988; Lepak and Snell, 1999), while internal labor market theory has explored the

consequences of these specific skills for performance and career paths within firms (Althauser,

1989). The kinds of jobs that workers are moving from – either from within the same firm or

from a different firm – should affect the specific skills that the workers bring to their new jobs

and hence their performance. Theories of information and labor market matching emphasize the

importance of information in finding a good match between the characteristics of the worker and

the demands of the job (Granovetter, 1981; Halaby, 1988). Internal labor market theory has long

observed that firms have better information about current employees than other workers,

allowing them to better assess this match (Doeringer and Piore, 1971). Access to such

information might affect both the characteristics of workers who enter jobs through hiring versus

internal mobility and how much they are paid.

Integrating these two theories, I predict that external hiring will have two disadvantages for firms

relative to internal mobility: external hires are likely to perform worse than internal movers but

be paid more. Theories of firm-specific skills predict that external hires should have lower initial

performance than internal movers. The effects of incomplete information can prevent employers

from balancing this lower performance with lower pay for external hires; because employers are

more uncertain about external hires’ abilities, they may compensate by hiring workers with

stronger observable indicators of ability than internal candidates. Because external hires know

less about their potential fit with the job than do internal candidates, they may also demand more

pay to make up for the risk of a poor fit. Although external hires should benefit from that higher

pay, their lack of firm-specific skills and uncertain fit should create costs for them too, in the

form of higher rates of voluntary and involuntary exit than internal movers.

I further develop these predictions by exploring the effects of different kinds of internal mobility.

First, workers who move to positions that are more different from their prior jobs, such as

workers who are simultaneously promoted and transferred laterally, should have fewer job-

specific skills and perform worse. Second, lateral transfers should have more externally valuable

experience and a worse potential fit than workers who are promoted, which may affect their pay.

I tested these predictions using seven years of personnel data from a U.S. financial services

institution, conducting a comprehensive comparison of outcomes for workers who enter similar

jobs from inside and outside the firm. I examined the effects of external hiring versus internal

mobility on workers’ pay, performance, indicators of ability, and subsequent mobility. I do not

address the costs incurred before the jobs are filled, such as the costs of training workers for

potential promotions or forecasting the demand for skilled workers, even though those costs may

also shape decisions about hiring versus internal mobility (Cappelli, 2008). Instead, I examine

the effects of different matching processes on what happens to workers after they enter their jobs.

EFFECTS OF EXTERNAL HIRING VERSUS INTERNAL MOBILITY

Workers can enter similar jobs – as defined by criteria such as functional roles, reporting

relationships, and hierarchical rank – through either internal mobility or external hiring. Internal

mobility itself can take a number of forms. Perhaps the most common type of internal mobility is

promotion, defined as moves within an organization to a job that is in a higher administrative

rank and usually associated with higher pay, status, responsibilities and skill requirements

(DiPrete and Soule, 1988; Spilerman and Lunde, 1991; Cohen, Broschak, and Haveman, 1998).

Although scholars initially argued that promotions were triggered by the need to fill a previously

defined vacancy (White, 1970; Stewman and Konda, 1983), more recent studies show that

promotions often occur when individuals are judged to have the skills needed for the higher rank,

regardless of whether there is a vacancy (Stewman and Yeh, 1991; Barnett and Miner, 1992).

Studies also emphasize that such rank mobility can occur without a single, discrete change in the

work that an individual is carrying out; although the promotion is defined by a discrete move

across ranks, jobs’ responsibilities often accrete more gradually over time.

Though a promotion may not entail an immediate change in work content, the different ranks

usually carry different overall responsibilities and require the demonstration of different skills to

enter them. Hence Pergamit and Veum’s (1999) analysis of the National Longitudinal Survey of

Youth found that over half of the workers who reported being promoted in the previous year

either described those promotions as “upgrade promotions” (commonly understood as a

reclassification of a job to a higher level (Barnett and Miner, 1992)) or said they continued to

perform basically the same duties as before. Yet those same workers who reported little

immediate change upon promotion had similar probabilities of receiving increased

responsibilities during the following year compared with those who said that their promotion led

into a higher-level job in a different section (Pergamit and Veum, 1999). Hence, although

promotions may not always involve a step change in the tasks that a worker performs, they

nonetheless represent entry into a different job.

A second form of internal mobility is lateral transfers, which occur when individuals remain

within the same vertical rank but move to a different organizational unit or a different kind of job

(Stewman, 1986). Although transfers also reflect organizational attempts to match workers to

appropriate jobs, the circumstances that lead to such transfers are likely more diverse than the

circumstances triggering promotions. In some cases, workers may be performing poorly in their

current jobs, and their new role may be expected to suit them better. In other cases, a transfer

may reflect a worker’s desire for a role with improved rewards or advancement prospects.

Transfers may also be used to broaden workers’ skills in preparation for future positions. A third

variant of internal mobility occurs when workers move vertically and horizontally at the same

time, by being promoted while also transferring to a different organizational unit or a job outside

the traditional line of progression. Stewman and Yeh (1991) defined such moves as “vacancy

promotions,” because transfers across units are more likely to be triggered by vacancies than

acquisition of new skills, but I do not adopt this terminology. Whether there is a vacancy is not

relevant to my theory development, which focuses on the change in the work the individual is

doing.

External hiring occurs when workers enter the organization for the first time. Unlike internal

mobility, external hiring is usually not classified based on the hierarchical job that hires came

from, because of the frequent absence of data on those prior jobs and the conceptual difficulty of

comparing hierarchical levels across firms. Hiring a worker for a higher level may happen when

there is a specific vacancy to be filled, but it can also be triggered by the availability of an

attractive candidate (Granovetter, 1974; Rosenbaum, 1990).

Because external and internal mobility often occur through different processes, they may not

directly compete with one another for every move (Rosenbaum et al., 1990). Promotions may

take place when a worker is believed to have reached a certain skill level without hiring being

considered as an alternative. Opportunistic hiring may sometimes take place without considering

whether to promote a current worker to the position. Instead, tradeoffs take place over time; in

particular, when the organization has promoted several workers in a particular area, it will have

less need to hire workers with the skills to take on similar responsibilities. Over time, both hiring

and promotion will be used as routes to fill similar jobs, even if they are not always considered

simultaneously. Despite being substitutes, though, these routes may lead to different subsequent

outcomes, due in part to differences in the skills that new hires and promoted workers will bring

with them.

Specific Skills and Performance

Every job requires specific knowledge about the formal processes and informal routines used to

do the work, about how to work with specific colleagues, the tools and technologies required on

the job, and so on (Quinones, Ford and Teachout, 1995). Although some of this knowledge may

be acquired through formal training, much of it comes from on-the-job learning. As a

consequence, workers who enter a position from a similar prior job are likely to have relevant

job-specific knowledge. The more different workers’ prior positions are from their new jobs, the

less job-specific knowledge they will have and the worse they are likely to perform as a result.

One form of transition that can require workers to learn a great deal of new knowledge is being

hired into a new firm (Lepak and Snell, 1999). Prior research has emphasized that firm-specific

skills - knowledge and abilities that can only be acquired and utilized while working in a specific

firm (Becker, 1962) – are an important form of job knowledge. Meta-analyses of studies across

many fields confirm strong effects of organizational tenure on job performance (Sturman, 2003),

although this could partly reflect the attrition of poor performers. Though the importance of

firm-specific skills may vary across different positions and organizations, research has found that

even professional jobs that demand high levels of general skills, such as securities research,

scientific research, and surgery, can require firm-specific skills (Allison and Long, 1990;

Huckman and Pisano, 2006; Groysberg, Lee, and Nanda, 2008). Although such work depends on

individual workers’ skills and knowledge, it can also require intense coordination with others in

the organization, coordination that is facilitated by mutual learning (Groysberg, 2010).

Because internal movers have longer experience within the firm, they are likely to have already

acquired important firm-specific skills that new hires will lack. There are good reasons to believe

that external hires may have stronger qualifications than those promoted, but firm-specific skills

may be important enough that new hires will still experience lower performance than internal

movers; in some cases, hires’ experience in other similar jobs may even be counter-productive

(Dokko, Wilk, and Rothband, 2009). Although those new hires will learn about the organization

over time, they are likely to suffer an initial performance disadvantage when firm-specific skills

are important:

Hypothesis 1a (H1a): External hires will have lower initial performance than internal movers.

Changes in the specific skills needed for different jobs may also lead to performance differences

within the group of internal movers. Just as people moving from one firm to another need to

learn about new technologies or build new relationships, so do people moving to new jobs within

the same firm (Quinones, Ford and Teachont, 1995; Gibbons and Waldman, 2006). When

workers receive rank promotions, they are expected to take on new responsibilities over time.

Transferred workers need to learn about their new work context. Though the effects of such

changes in job-specific skills may often be less than the effects of moving to a new firm, they

may nonetheless influence performance. I would therefore expect internal movers to have lower

performance when they take on jobs that are more different from their previous work.

There is little theoretical basis for predicting whether promotions or transfers involve a greater

change in job-specific skills, but mobility that entails both kinds of movement – promotions that

are combined with a lateral transfer – likely requires a greater change in job-specific skills than

would either move in isolation. Such a change in the job-specific skills demanded by their new

versus old roles is likely to impede the performance of workers who are simultaneously

transferred and promoted:

Hypothesis 1b (H1b): Workers who are promoted and transferred at the same time will have

lower initial performance than other internal movers.

Incomplete Information and Workers’ Characteristics

The second way in which workers’ routes into their jobs may affect their employment outcomes

is through those routes’ effects on the information used in the matching process. All of the

different processes that place individuals in jobs aim to create an effective match between the

characteristics of the job and the skills and needs of the workers taking those jobs. Firms look for

workers who will perform to an acceptable level in the job; workers seek jobs that will be a good

fit for their abilities and provide rewards that they value. A key determinant of the parties’ ability

to form an effective match is the amount of information they can use to assess this match

(Akerlof, 1970; Spence, 1973; Granovetter, 1981), but firms and workers often have highly

incomplete information about each other. Firms struggle to evaluate the true qualities of

applicants (Schmidt and Hunter, 1998), and workers struggle to know which of the jobs available

will best suit their preferences and abilities (Halaby, 1988). How the different routes into jobs

affect the information available to workers and firms may therefore influence subsequent

outcomes.

Problems of incomplete information shape the differences between internal and external mobility

because workers and employers expect to know more about each other during internal moves

than during external hiring (Waldman, 1984; Greenwald, 1986). Some correlates of performance,

such as a worker’s education and the jobs he or she has held, are “externally observable,” in that

they can be observed by all potential employers (Spence, 1973; Granovetter, 1981). Other

important correlates of performance are “externally unobservable,” though, in that they can only

be assessed by the current employer. Examples of such externally unobservable information

include details about how workers have performed in prior roles and how they would fit with the

idiosyncratic demands of the job or the organization (Chatman, 1991; Edwards, 1991; O'Reilly,

Chatman, and Caldwell, 1991).

Previous research has demonstrated the importance of externally unobservable information for

employment outcomes. Gibbons and Katz (1991) found that laid-off workers had better

employment prospects when their plant was shut down, suggesting that future employers were

concerned that plants were otherwise able to identify and lay off weaker employees. Hiring

studies have found that recruiters set great store on references from previous employers (Crain,

1984; Bills, 1999); given the difficulty of getting honest references (Miller and Rosenbaum,

1997), first-hand experience with the workers should be even more helpful. And though studies

of hiring emphasize the low predictive power of interviews and other common selection devices

(Arvey and Campion, 1982; Posthuma, Morgeson, and Campion, 2002), analyses of workers’

performance ratings within firms demonstrate a high cross-period correlation (Sturman,

Cheramie, and Cashen, 2005).

These differences in the information available during internal mobility versus external hiring are

likely to shape how firms screen candidates from these different routes. The firm’s challenge in

staffing higher-level positions is to find workers whose overall skills and ability allow them to

perform at an acceptable level. In choosing whether an internal candidate should be promoted to

a higher level, the firm can assess the worker based on what it knows about both his or her

externally observable and externally unobservable attributes. If the worker has the skills

necessary to be effective at the higher level, then he or she will be promoted. The organization is

similarly well placed to assess the externally unobservable attributes of potential transfers. We

would therefore expect the pool of chosen internal candidates to be above average on both

externally observable and unobservable attributes. In assessing external candidates, the firm

lacks information about externally unobservable attributes. The hiring managers must assume

that external candidates are, at best, average on those unobservable dimensions – although

adverse selection theories suggest that those candidates may be below average on those

unobservable dimensions, which is why they are seeking another job (Akerlof, 1970; Greenwald,

1986). We would therefore expect internal movers to have stronger externally unobservable

attributes than external hires.

It is likely that firms will compensate for these differences in externally unobservable attributes

by changing the way that they evaluate externally observable attributes. Given that many higher-

level jobs require a minimum level of performance, firms may be reluctant to hire workers who

are expected to fall below that performance threshold. One way to maintain performance above

that threshold is to require that external hires have stronger observable indicators of ability than

internal movers in similar positions. For example, a firm may be willing to promote an internal

candidate with poor observable attributes, because its externally unobservable knowledge

suggests that he or she she will be successful in that role. The firm would not be willing to hire

an external candidate with similarly low levels of externally unobservable attributes, because it

could not have externally unobservable knowledge that the new hire would succeed.

Just as firms know less about external hires than they do about internal movers, external hires

similarly know less about the firm. Given the difficulties of comparing jobs across organizations

(Baron and Bielby, 1986), external hires may sometimes apply for and accept jobs for which

they are overqualified. Though underqualified applicants are likely to be screened out by the

employer, overqualified candidates may end up being hired. Because internal candidates are less

likely to apply for jobs for which they are over qualified, such effects would also lead external

candidates to have stronger observable indicators of ability than internal movers. Based on both

of these arguments, I propose that:

Hypothesis 2a (H2a): External hires will have stronger externally observable indicators of

ability, such as experience or education, than internal movers.

There are also grounds to expect differences in externally observable characteristics among the

different kinds of internal movers. Specifically, the prior experience of lateral transfers should

provide them with stronger observable indicators of ability than other movers. One of the

externally observable indicators that employers most value is the record of specific jobs that an

individual has held (Bills, 1990; O'Mahony and Bechky, 2006). By definition, lateral transfers

are moving from a higher-ranked job than those promoted to similar jobs. The transfers therefore

have more experience in higher-ranked jobs than do those who are promoted, experience that is

valuable and externally observable.

In some cases, the processes by which transfers are selected may also shape their observable

attributes. Some organizations may use lateral transfers to redeploy workers who do not fit in

their current role. Decisions to redeploy poor performers could well reflect observable attributes:

a poor performer with weak observable attributes may be terminated; when the worker has

stronger observable attributes, the organization may be more likely to interpret poor performance

as a lack of fit with the current position and transfer that worker elsewhere. Based on both of

these arguments, I predict that:

Hypothesis 2b (H2b): Lateral transfers will have stronger observable indicators of ability,

notably work experience, than other internal movers.

Incomplete Information and Pay

Differences in the information available during internal mobility versus external hiring may also

affect how workers are paid. A simple view of labor markets suggests that workers entering the

job via internal mobility should be paid more than external hires if they have higher

performance. Yet considering the effects of incomplete information leads to the opposite

prediction: that external hires should be paid more than internal movers.

In part, pay differences between external hires and internal movers should reflect the

hypothesized differences in observable characteristics. External hires’ stronger observable

indicators of ability can help those workers find high-paying jobs in other organizations and will

be rewarded in the labor market. Unobservable attributes do not help workers find jobs in other

organizations, and employers will face less pressure to reward such attributes. If external hires

have stronger observable indicators of ability, we would expect them to be better paid than

internal movers.

The reduced information that external hires have about the firm and job may also affect their

pay. Although firms may seek to give prospective hires a clear description of their job, internal

movers are likely to have more direct knowledge about the nature of the job and how well they

will fit with it. Internal candidates should also have a clearer understanding of how they fit the

culture and values of the organization (Chatman, 1991), and should be a good fit for the

organization; otherwise, they would have left the organization already (Schneider, Goldstein, and

Smith, 1995). External hires lack this information and should assume that they will be, on

average, a worse fit for the job and organization than internal movers. Such poor fit can lead to

lower job satisfaction and higher turnover (Chatman, 1991; Edwards, 1991).

External hires’ worse expected fit with the job and organization should make the job less

attractive to them than to an internal mover. The pay that the employer needs to offer to persuade

an external hire to take a job is therefore higher than the pay demanded by an internal mover to

take a similar job. These considerations of fit and observable indicators of ability imply that:

Hypothesis 3a (H3a): External hires will be paid more than internal movers.

There may also be differences in the pay of different types of internal movers, based on these

same factors of observable human capital and the relative attractiveness of the job. As noted

above, transfers have valuable, externally observable experience in their current rank. That

valuable experience should raise what other employers are prepared to pay them over time and

consequently lead to pay raises from their current employer. People promoted into a position

lack that externally observable experience and associated pay rises. We would therefore expect

that transfers’ increased seniority within the higher rank will translate into higher pay.

The pay of transfers and promoted workers may also be shaped by the information that their

mobility provides about their fit with the job and organization. As I have noted above, transfers

can sometimes reflect mediocre prior job performance and an opportunity for a fresh start

elsewhere in the organization. In such cases, transfers’ expectations about their fit and potential

performance are likely to be lower than promoted workers who have performed well in their

prior role. Those transfers may therefore need to be paid more to stay in the job. At the same

time, the firm’s decision to transfer workers rather than terminating them indicates that the firm

continues to see some promise in them and may be willing to pay that premium to retain them.

Based on both of these factors, I expect that:

Hypothesis 3b (H3b): The pay of lateral transfers will be higher than the pay of other internal

movers.

Implications for Subsequent Mobility

The effects of specific skills and incomplete information are also likely to affect the subsequent

mobility of external hires and internal movers through promotions and exit. Promotions and exit

tend to occur disproportionately among workers near the tails of the performance distribution.

Promotions occur among those workers who are able to demonstrate the skills necessary to be

promoted to the next level. More rapid promotions occur among those who are performing well

ahead of expectations. By contrast, involuntary exit (both for cause and as part of downsizing)

tends to occur among workers performing well below the average.

The arguments developed above suggest that external hires may be disproportionately

represented in both tails of the performance distribution. Because employers are unable to screen

hires based on their externally unobservable attributes, hires are likely to have worse average

unobservable attributes than internal movers, but they will also have higher variance in those

attributes. Although many hires will turn out to have poor externally unobservable attributes,

others will have strong externally unobservable attributes. Because employers also require

stronger observable attributes from hires, the hires that turn out to have strong unobservable

attributes may be among the best performers, particularly as they begin to acquire firm-specific

skills.

If promotions are disproportionately drawn from the upper tail of the performance distribution,

then external hires may have higher overall promotion rates than internal movers. Chan (2006)

found evidence of such faster subsequent promotion among external hires, although he attributed

this effect to hires having higher overall ability, rather than their being over-represented among

higher performers. I predict that:

Hypothesis 4a (H4a): External hires will have a higher rate of subsequent promotion than

internal movers.

In contrast to promotions, involuntary exit (both for cause and as part of downsizing) tends to

occur among workers performing well below the average. External hires should also be well

represented among those lowest performers. Their lack of firm-specific skills likely reduces their

performance; the greater uncertainty about hires’ externally unobservable attributes also

increases the risk that they will turn out to be a poor match for their jobs (Halaby, 1988). I

predict that:

Hypothesis 5a (H5a): External hires will have a higher rate of involuntary exit than internal

movers.

Rates of promotion may also vary among the different kinds of internal movers. I argued that

workers who are promoted and transferred at the same time may suffer lower performance than

other internal movers because of their lack of job-specific skills. That poor performance should

leave those workers who are simultaneously promoted and transferred underrepresented at the

top of the performance distribution, and therefore less likely to be promoted:

Hypothesis 4b (H4b): Workers who are promoted and transferred at the same time will have

slower rates of subsequent promotion than other internal movers.

Similarly, the lower performance of workers who are promoted and transferred at the same time

should leave them well represented in the lower tail of performers. That low performance may

make them vulnerable to higher rates of involuntary exit:

Hypothesis 5b (H5b): Workers who are promoted and transferred at the same time will have a

higher rate of involuntary exit than other internal movers.

A further form of worker mobility is voluntary exit. Prior research suggests that such exit can be

affected by a poor match between workers and their job or organization, as well as by underlying

individual propensities to leave jobs (Blumen, Kogan, and McCarthy, 1995; Chatman, 1991;

Edwards, 1991; Farber, 1994). As noted above, external hires know less about the organization

than do internal candidates and are at greater risk of forming a poor match (Halaby, 1988). Given

that external hires have already left at least one other organization, they may also have a higher

average propensity to move to another firm, although such differences may be less marked in

fields in which interorganizational mobility is common. Together, these factors lead me to

predict:

Hypothesis 6 (H6): External hires will have a higher rate of voluntary exit than internal movers.

METHODS

I tested the hypotheses using personnel data from the U.S. investment banking arm of a financial

services institution, which I call “Croesus.” Investment banking represents an interesting context

in which to study the effects of internal versus external mobility. Organizational performance in

this industry often depends on the skills of the workforce, increasing the importance of personnel

decisions. Workers in banking are also notoriously mobile, making this a context in which

organizations regularly engage in external hiring at all levels. Finally, the hypotheses developed

above apply to jobs with high levels of firm-specific skills; Groysberg, Lee, and Nanda,(2008)

found that the performance of investment analysts who moved to a new firm fell substantially,

suggesting that firm-specific skills are important in investment banking.

Croesus is organized into four different business units, comprising one support unit and three

different revenue-producing units. Each business unit is then divided into divisions, with a

median size of 450 workers; divisions are further subdivided into departments, with a median

size of 100; and each department is itself composed of different groups, with a median size of 32.

I have data on all Croesus employees from the year 2003 to 2009. These employees include both

investment professionals, such as research analysts, advisors, and traders, as well as “back

office” staff such as lawyers, information technology workers, and administrators. The data

consist of annual records for each worker, giving details of his or her job and compensation in

that year, annual performance data, and demographic data.

Variables

Mobility Variables. Internal mobility at Croesus took place through promotion to a higher rank

or transfer to a different organizational unit. The employees at Croesus all occupied one of six

different ranks, which were used across the organization and were central determinants of pay

and responsibilities. Although titles vary across firms, the usual industry descriptions of these

ranks (which I use here to preserve the anonymity of the site) are analyst/non-officer, associate,

vice president (VP), director, managing director (MD), and senior executive.

Promotions across ranks usually occurred as individuals demonstrated the skills necessary to

operate in the rank above and almost always occurred once a year as part of the annual personnel

process. Nonetheless, workers in different ranks would usually be expected to carry out

significantly different tasks. In investment banking jobs, for example, industry experts report that

it is usual for analysts to work for associates in carrying out analyses. VPs provide direction for

those overall analyses and interact with the clients. Directors oversee deals and sell new work,

while MDs focus on managing client relationships and overall strategy. Within trading,

progression through the ranks involves becoming directly involved in trading, managing larger

portfolios of securities and more complex products, and ultimately managing a group and setting

strategy. I identified promotions by a change in rank from one year to the next.

Lateral transfers are defined as moves across organizational units (divisions, departments, or

groups) between one year and the next. Because different units generally focused on different

topic areas, such moves involve changes in the kinds of work that transfers do. I avoided

confusing transfers with reorganizations, in which an individual might be doing the same work

but in a differently labelled or constituted organizational unit, by restricting transfers to

transitions accompanied by fewer than 10 per cent of the source unit’s workers and by fewer than

10 per cent of the destination unit’s workers.

I used dummy variables to code internal entry into the job into three categories: simple

promotions (not accompanied by transfers), simple transfers (not accompanied by promotions),

and combined promotions and transfers (both promotion and transfer occur during the same

year). Entering a job by external hiring is the excluded category in the analyses. I verified that

those jobs were entered by hiring based on the hire date provided by Croesus. When external

hires then move into new jobs within the organization, they are recoded as entering those

subsequent jobs by internal mobility.

Indicators of observable ability. I measured indicators of observable ability with two variables

that have been extensively used in labor economics: education and experience. Work experience

provides workers with valuable work-related skills through on-the-job learning (Mincer, 1962).

Education is also a commonly used indicator of ability, both because it provides workers with

useful knowledge and because it may signal cognitive ability (Spence, 1973). Although

experience and education are weaker predictors of job performance than cognitive tests or

structured interviews (Schmidt and Hunter, 1998), they are valuable proxies for underlying

characteristics, are highly observable, and are well rewarded in the labor market as a

consequence (Mincer, 1970; Dustmann and Meghir, 2005).

The original dataset provided details of workers’ birthdates and their academic degrees.

Education data were missing for around 3,300 of the 15,000 workers originally in the data. I was

told that data were often missing due to failures by workers or administrators to fill in the field

for education; it may also reflect some individuals lacking degrees. I dropped workers who

lacked education data from the analysis.

I created dummy variables for the highest degree that a worker possessed (bachelor’s, master’s

or Ph.D.). I then calculated experience using workers’ ages. I assumed that workers had been

working since age 18 if they had no degree and added 4 years to this age for a bachelor’s degree,

2 years for a master’s, and 4 for a Ph.D. (in the very few cases in which this resulted in negative

experience, I reset it to zero).

Performance. I used three measures from Croesus’s annual performance evaluations.

Contribution was based on a 1-5 scale that assesses whether workers hit their performance

targets for the year: 1 = “Objectives significantly exceeded,” 3 = “Objectives met,” and 5 =

“Objectives not met.” Around 9 per cent of ratings are a 1, 56 per cent are a 2, 34 per cent are a

3, 2 per cent are a 4, and .03 per cent are a 5. Competence assesses a worker’s skills relative to

the requirements of the job. It was measured on a 1-5 scale, with 10 per cent of workers being

awarded a 1 (the highest rating), 50 per cent receiving a 2, 37 per cent receiving a 3, 2 per cent

receiving a 4, and .05 per cent receiving a 5. Performance rank is a forced ranking that is used to

determine compensation and other personnel decisions and is intended to reflect both

performance and overall value to the organization. Groups of managers rated many workers in

the same rank to develop these rankings. The ranking splits workers into the top 10 per cent, the

next 20 per cent, the next 60 per cent (in 2009, this segment was itself split into a higher 40 per

cent and a lower 20 per cent), and the bottom 10 per cent. I converted this ranking into a 6 point

scale, on which the 60 per cent category corresponded to a 4, and the divided category in 2009

represented a 3 and a 5 respectively.

I reversed the scales for each of the three performance measures so that higher values indicate

better performance. Although I could have combined these three measures to form a single

aggregate measure of performance, I chose to keep them separate, both because Cronbach’s

alpha for a combined scale was marginal at .69 and because comparing these performance

indicators can shed more light on the effects of mobility on performance.

Although subjective performance evaluations can be biased (Cascio, 1998: 65-66), many

researchers argue that such subjective evaluations are among the most valid measures of

performance. Subjective measures often correct for determinants of performance outside the

control of the individual (Campbell et al., 1993) and can encompass a wide variety of behaviors

and outputs relevant to the job (Medoff and Abraham, 1981). Meta-analyses have shown high

test-retest reliability in performance evaluations (Sturman, Cheramie, and Cashen, 2005), and

supervisory ratings are among the most common dependent variables in studies of performance

(Sturman, 2003).

Compensation. Croesus employees were paid a salary and an annual bonus that was based on

both individual and firm performance. Bonuses represented 38 per cent of pay, on average, and

rose as high as 98 per cent. I ran analyses on each pay component separately and on total

compensation. I used the logarithm of each of these components because they are highly skewed

(when individuals received zero bonus, I substituted a bonus of one dollar in order to calculate a

log value).

Exit. I coded exits into voluntary and involuntary exits based on a coding provided by Croesus.

From 2006 onward, all exits were coded as “Employee Initiated,” “Croesus Initiated,” or

“Neutral” (I restricted analyses to years after 2005 and omitted the 9 neutral exits). Of 10,952

person-years analyzed, 1,135 ended in involuntary exit, and 1,118 ended in voluntary exit.

Controls. The goal of the paper was to compare people being staffed in similar jobs through

different routes. I therefore controlled for as many aspects of those jobs as possible. I included

separate dummies for each of the six hierarchical ranks. I also used job titles and department and

group names to create dummies for 13 different functions: administrator; Human Resources

(HR); corporate management, marketing, legal, internal finance, operations, Information

Technology (IT), research; sales; trading; advisory; and other banking roles. I also controlled for

interactions between rank and function to allow the effect of rank to vary according to

occupation. I controlled for the city that the job was in (88 per cent of workers were in the

greater New York area). Finally, I included dummies for workers’ departments to provide even

more fine-grained controls for the kinds of work people were doing, but in some cases, analyses

would not converge when controlling for departments. In those cases, I controlled for divisions.

A particular concern is that unobserved factors that make the firm more likely to hire from

outside might be correlated with pay or performance. For example, we would expect more hiring

in those areas that have higher turnover or require new skills that the organization lacks.

Although highly detailed job-level dummies should control for fixed propensities of different

kinds of jobs to engage in more external hiring they may not capture over-time variation in

propensity to hire. I therefore also controlled for the proportion of new hires in the worker’s

group and the proportion of workers who left the worker’s group, as measured both in the year

that his or her job started and in the current year.

I included dummies for each year in the sample, reflecting changes in labor market conditions

between 2002 and 2009. I also controlled for how long an individual had been in the job,

measured in years. Although I had dates of entry for promotions and hires, I lacked these for

transfers. I assumed the transfers entered their jobs in the middle of the year. Finally, I controlled

for gender and ethnicity (Caucasian, Asian, and other).

Final Sample

I made a number of restrictions to the data. I dropped observations with missing education data,

and with missing performance data, which occurred when workers had been hired too recently to

assess or were terminated before the evaluation. I explored whether the missing education or

performance data might bias the results by rerunning analyses excluding these variables and

using all data. The results were very similar to those using the restricted data set. I also excluded

job spells in the lowest rank, which cannot be entered by promotion. I avoided sampling workers

based on the length of their job spell by limiting the sample to jobs that began during or after

2002 and for which I could observe the beginning date. I excluded 99 workers who met my other

criteria but appeared in the data set long after they were hired, either because they transferred

from elsewhere in the parent company or because they entered Croesus as part of a merger.

The final data set contains information on 5,260 workers in 7,129 job spells. The median length

of completed job spells was around two years for those that ended in promotion and around 1.5

years for those ending in transfers or exits.

RESULTS

Insert Table 1 about here

Table 1 provides means, standard deviations, and correlations for the main dependent and

independent variables. A year spent in a job is the unit of analysis. Of particular interest are the

means for the different modes of job entry. Simple promotion is by far the most common mode

of entry, representing 55 per cent of the observed job years. The next most common entry mode

is external hiring, with 32 per cent of observations. Simple transfers represent only 10 per cent of

observations, while combined transfers and promotions were very rare at the firm, providing the

means of entering only 3 per cent of job years (statistics using job spells as the unit of analysis

were very similar). Internal mobility therefore outweighs hiring at Croesus, but that mobility is

overwhelmingly vertical and within the same organizational unit.

Insert Table 2 about here

Performance

Table 2 presents analyses of each of the three performance measures. Because these are discrete,

ordinal measures, I used ordered logit analyses. The unit of analysis is the job-year. I clustered

the errors by individual to account for non-independence among the errors. The analyses did not

converge fully when I controlled for departments and cities, so instead I controlled for divisions

and whether the job was in the greater New York Area (ordinary least squares analyses produced

very similar results when controlling for either division or department). Performance was scaled

so that higher values always indicate better performance. I present models with and without

interactions between mobility type and time in job and include tests for differences between

internal movers.

Two broad patterns emerge from the models. First, the two most common forms of internal

mobility – simple promotion and simple transfers – lead to significantly higher initial

performance than external hiring. The main coefficients in the full models (2, 4, and 6) indicate

predicted performance on beginning the job, while the main coefficients in the reduced models

(1, 3, and 5) indicate average performance over the full period in the job. Workers who entered

the job through simple promotions perform better than external hires on all measures and in all

specifications. Simple transfers also outperform external hires initially in all specifications (main

effects in models 2, 4, and 6) and over the full duration of the job using contribution, which is the

performance measure most closely tied to objective results. The weaker effects for transfers in

models 1 and 3 appear to reflect the lower numbers of simple transfers compared with simple

promotions, rather than a rapid decay of their advantages. Initial differences between external

hires and workers who entered the job through combined promotions and transfers are much

weaker, only attaining a 10 per cent significance level in model 6.

Second, I find evidence that the performance of workers entering the job through combined

promotions and transfers is weaker than other internal movers’. I find significant differences

between workers entering their jobs through simple promotions versus combined promotions and

transfers in analyses of both ranked performance and contribution. It may be that workers who

entered the job through combined promotions and transfers are seen as having the core skills

required for their jobs but struggle to apply these skills in a new setting. I also find differences in

ranked performance between workers entering their jobs through simple transfers versus

combined promotions and transfers. A lack of significant differences between these groups for

the other measures may partly reflect a lack of statistical power, given the smaller number of

workers in each of these categories.

These results offer support for H1a and H1b, which argued that internal movers would perform

better than hires and that workers entering jobs through combined promotions and transfers

would perform worse than other internal movers. Though most internal movers do perform

significantly better than hires, the performance penalty suffered by workers entering the job

through combined promotions and transfers is so large that it often renders their performance

similar to external hires.

Interactions between entry modes and time in job suggest that the advantages received by

workers entering the job through simple promotions and simple transfers versus new hires

decline over time. The magnitudes of the interaction coefficients indicate that the performance of

external hires remains significantly weaker than that of workers entering the job through simple

promotions for more than two years for contribution and competence, and slightly less than two

years for ranked performance. In supplementary analyses, I created dummies for each year in the

job and interacted these with the mobility dummies (results available from the author). I found

that the performance of promoted workers was better than new hires for the first two years, after

which the performance of the two groups converged. The performance of new hires was never

significantly better than promoted workers. I found similar results when I also examined the

performance of the workers in their subsequent jobs.

This convergence in performance between new hires and internal movers most likely reflects the

acquisition of firm-specific skills by new hires. I confirmed that this performance convergence

was not due to attrition among the weaker hires by conducting individual fixed effects analyses

(not reported here). The fixed-effects analyses produced similar interaction effects to the cross-

sectional analysis, suggesting that hires’ performance improvement is due to within-individual

learning. A Heckman analysis, correcting for sample attrition using interactions between year

and function as an instrument, gave very similar results to OLS and was not able to disconfirm

the hypothesis that the selection and performance equations were independent.

Among the controls, I find higher performance ranks for men and white employees. Men do not

have higher competence or contribution ratings, though. I also find that workers with less

education and experience perform better. These results likely reflect selection effects: holding

the position constant, workers who reach that position at an earlier age and with less education

are likely to have higher innate ability, which shows up as higher performance.

Characteristics of External Hires versus Promoted Workers

Insert Table 3 about here

I tested for differences in workers’ observable indicators of ability using a multinomial logit

analysis, in which the dependent variable was mode of job entry and the unit of analysis was the

job spell. The results are shown in table 3. This analysis cannot distinguish the determinants of

selection into these jobs, because there is no information on the applicant pool from which the

workers were picked (Fernandez and Sosa, 2005). In the absence of data on the applicant pools,

however, these analyses can at least determine whether the characteristics of successful

candidates for hiring and internal moves are consistent with my theory. Because external hiring

is the excluded category, coefficients represent differences between the characteristics of internal

movers and external hires. To achieve convergence, I controlled only for division rather than

department. Chi-squared tests also indicated very poor fit for models including interactions

between rank and function, so I excluded these interactions from the models. I also excluded

controls for hiring and turnover at the current date, as I only included one observation per job

spell.

Consistent with H2a, external hires are significantly more experienced than promoted workers

(both simple and combined). External hires are also better educated than workers entering the job

through simple promotions, being more likely to have completed every level of education. The

large sizes of the standard errors on the education coefficients for simple transfers and combined

transfers and promotions prevent us from making inferences about these differences. As

predicted by H2b, transfers have more experience than workers entering the job through simple

promotions (p <.013). There are no significant differences between the experience of transfers

and external hires. I experimented with dropping workers from the second-lowest rank, in case

the results reflected recruitment directly out of graduate programs at that level and the results

were robust to this exclusion.

Among the other worker characteristics, men are more likely to enter the job by external hiring

than promotion, although the coefficient for simple promotions is only marginally significant

(p<10).

Pay

Insert Table 4 about here.

Table 4 analyzes the determinants of compensation. Because I was concerned that pay would be

more sensitive to the nature of the job than other outcomes, and because ordinary least squares

does not present convergence problems in the presence of multiple controls, I used additional

controls for all interactions between rank and division (I also experimented with controlling for

group, but this had little effect on the estimates). Because workers who had been at Croesus less

than a year would likely receive a lower bonus, I controlled for the number of days in that year

since the worker was hired (maximum value of 365) and whether he or she was terminated

during the year. I also controlled for performance and general human capital, although the results

were very similar when these controls were excluded. Table 4 examines the determinants of log

salary, log bonus, and log total compensation.

The results provide substantial support for H3a. All internal movers received significantly lower

salaries than external hires (models 1 and 2). Internal movers also received lower total

compensation than external hires when looking across all observations (model 5), although

simple transfers did not have lower total compensation in their first year (model 6). The lack of a

significant coefficient on simple transfers, however, may be a statistical artefact. Although the

interaction between simple transfers and time in job is non-significant, interpretation of the

coefficients suggests that simple transfers would have lower pay than external hires in all years

except their first one. Workers entering their jobs through both simple promotions and combined

promotions and transfers also receive lower bonuses than external hires.

The results also support H3b. Simple transfers have significantly higher salaries and total

compensation than other internal movers (models 1, 2, 5, and 6). They also have higher initial

bonuses (main effect in model 4).

The overall magnitudes of the effects are also large. Initially, workers entering their jobs through

simple promotions receive salaries and total compensation that are around 15 per cent lower than

external hires. The interaction coefficient with time in job suggests that the salaries of simple

promotions would only catch up with external hires after seven years – a longer time span than

almost anyone is present within the data – and that their total compensation would never

converge with external hires’.

Among the controls, measures of group hiring and turnover have positive effects on pay. Results

also show positive effects of indicators of ability (experience, Master’s and Ph.D.’s) on salary,

but negative effects on bonus, consistent with their effects on performance. When I dropped job-

level controls (not reported here), I found positive, significant effects of experience and

education on all components of pay.

I explored whether the higher pay of external hires might reflect labor market conditions. Baker,

Gibbs, and Holmstrom (1994b) found that the pay of continuing workers within a firm responded

less to market forces than the pay of new hires. Rapid increases in market pay during the period

of the study might then have affected the wage differential. The very high correlation between

year of entry and mode of entry prevented me from simply controlling for when workers entered

the firm. Instead, I compared trends in the pay given to new hires versus continuing workers

(results available from the author) and found that continuing workers’ compensation actually

varied more from year to year than did the compensation of new hires. This analysis suggests

that external labor market changes do not drive the pay premium.

Subsequent Mobility

I analyzed the determinants of subsequent promotions and exits at Croesus using Cox event

history models. The time that an individual is in a specific job is treated as a unique case, during

which the worker remains at risk of promotion or exit. I treated promotion, exit, and transfers as

competing risks; for each analysis, exits from the data for any cause other than the focal one

were classified as censored events (Allison, 1984). I also used slightly different samples for

analysis of terminations versus promotions. For the promotion analyses, I dropped observations

that were not at risk of promotion because they were in the top two ranks (I never observed

promotion out of those ranks) or in the final year of the data. For the exit analyses, I dropped

data from before 2006 because of problems with Croesus’s coding in prior years. The exit

analyses also used performance data from the year prior to termination, because workers did not

receive performance rankings in the year of their termination. I also included dummies for

division but not department in the exit analyses– analyses using department dummies gave

similar results but were less stable. The models are presented in table 5. For each outcome, I

present models with and without performance variables. Analyses that excluded salary and

demographic variables showed somewhat stronger effects of the mobility variables.

Insert Table 5 about here.

Consistent with H4a, I find that external hires are promoted more rapidly than workers who had

entered the job through simple promotions. I find no support for H4b, that workers entering the

job through combined promotions and transfers have even slower rates of promotion; instead,

their rates are slightly higher than those of workers entering through simple promotions. I also

find that simple transfers have the highest promotion rates. This result may reflect the experience

that such transfers have already accrued in their rank. If transfers can demonstrate performance

in their new job, their experience may allow them to be promoted more rapidly than others.

In developing H4a, I argued that new hires would have higher performance variance than

internal movers. I tested this assumption by exploring the distribution of performance scores for

hires versus internal movers. I found that hires were much more likely than internal movers to

receive below median performance ratings, but only slightly less likely to receive the highest

performance category relative to the median. These results confirmed that hires’ lower average

performance reflected a disproportionate likelihood of their being poor performers, rather than

uniformly lower performance across the distribution. I was not able to find evidence, though, that

hires were actually overrepresented at the top of the performance distribution. This may in part

reflect the relatively crude nature of the performance measures.

Results also show strong support for H5a and H6. External hires have higher involuntary and

voluntary exit rates than workers entering the job through either simple promotions or combined

promotions and transfers. External hires have around a 61 per cent higher hazard rate of

involuntary exit than workers entering through simple promotions, and a 21 per cent higher

hazard rate of voluntary exit (models 3 and 5). Performance explains a substantial amount of the

effect of job entry on all forms of exit and fully mediates the effects of simple promotion on

voluntary exit rates. This finding suggests that the increased voluntary exit of hires reflects

problems of firm-specific skills and fit, rather than a higher fixed propensity to turn over.

Results show no significant support for H5b; in fact, involuntary exit rates for workers entering

the job through combined promotions and transfers are significantly lower than those of other

internal movers. By contrast, involuntary exit rates for simple transfers are significantly higher

than those for workers entering the job by simple promotions (p<.000 in model 3) and higher

than external hires before performance controls. It is possible that this higher exit rate reflects the

process by which individuals became transfers. In supplementary analyses, I found that transfers

had previously had much lower performance evaluations than those remaining in their jobs. Such

a history of poor performance likely leaves transfers vulnerable to involuntary exits.

One concern is that some voluntary exits may reflect workers being counselled to leave. Croesus

coded the reasons for voluntary exits into 23 categories. Of the reasons given, those most likely

to reflect signals from management were the roughly 25 per cent of exits that occurred because

of dissatisfaction with career development or promotion opportunities. I conducted a robustness

check by dropping these exits and rerunning the analyses. I found the same results. I also reran

the analyses using a competing risk regression model, which assumes that the observations that

fail from a competing risk would never fail from the focal cause (Fine and Gray, 1999). The

results were somewhat weaker, reflecting the increased representation of external hires in all

classes of failure, but continued to support the hypotheses.

Supplementary Analysis: Source of Hires

Although this paper focuses on differences between external hires and internal movers, analyses

can shed further light on the underlying mechanisms by examining the effects of different

sources of hires. For example, I argued that new hires are paid more because they and the firm

have less information about each other. Workers hired through sources that can provide more

reliable information to each party should then receive lower pay than other hires. In particular,

we would expect that workers brought in through employee referrals would have the most

information about their potential match with the job, because internal contacts are more likely to

provide applicants with rich information about the proposed job, and firms and applicants may

be more likely to trust that information (Fernandez, Castilla, and Moore, 2000; Castilla, 2005). A

corollary of the above arguments is therefore that workers hired through employee referrals

should have lower pay. Looking at the source of hire can also shed light on the influence of

adverse selection. There is a concern that new hires may be disproportionately those workers

who were unsuccessful in their previous jobs and may even have been laid off. We might

therefore expect that workers hired through unsolicited applications are more likely to perform

poorly than those hired through a headhunter, who are more likely to be poached from current

jobs.

The data provided by Croesus included the referral source for the hire of each worker (although

these data were missing for 848 workers). Of the workers with source-of hire-data, 785 were

referred by an employee, 624 were intermediated hires, through an employment agency or

executive search firm, 407 were hired following unsolicited applications, 64 were hired through

an Internet application, 50 came from temporary jobs, and 18 were former employees. I dropped

the 33 workers who entered through mergers or mass hires of business and categories with fewer

than 15 workers to simplify the analysis. I analyzed determinants of performance and pay for

these external hires using the same specifications as the comparisons of external hires and

internal movers, but restricting the analyses to external hires for whom I had hiring source data.

Insert Table 6 about here.

Performance analyses are reported in table 6. The first model shows performance relative to

employee referrals, while the second model compares employee referrals and all other hires. The

results show that intermediated hires performed worse than either employee referrals or

unsolicited hires. This is the opposite of what adverse selection might predict, because

intermediated hires are among the least likely to have been already looking for a job due to poor

performance. Instead, the result raises the possibility that Croesus may place too much trust in its

intermediaries, underestimating the challenges that their candidates will face. Employee referrals

have significantly higher-ranked performance than other hires, but differences are not significant

for other measures. I also find that former employees are rated lower on competence than other

workers. This result suggests that firm-specific skills may decay during time spent away from the

organization.

Insert Table 7 about here.

Table 7 analyzes the effects of different hiring routes on how workers were paid. For each

measure, I again present results relative to employee referrals (the first two models) and relative

to all other categories (the third model). The results show that employee referrals received less

total compensation than other hires. The differences versus intermediated hires are particularly

significant. Results for salary and bonus are weaker, although intermediated hires receive

significantly more salary than employee referrals.

Although the findings are not consistent across all components of pay, the results for total

compensation are consistent with an effect of information on external hires’ pay premium. When

external hires have access to information through internal contacts at the organization, they

accept the job for less total compensation. This finding is also interesting in its own right: some

studies have suggested that workers hired through their personal networks receive higher pay

than others, although other work has questioned these results (see Mouw, 2003 for a detailed

discussion). I find that referrals actually receive lower pay, once the nature of the job is held

constant.

It is also interesting to note that the total pay for unsolicited hires is not significantly different

from intermediated hires (p < .18 in model 8), although there are differences in salary. I would

expect that unsolicited hires were less happy in their current jobs than intermediated hires, given

that unsolicited hires made the initial move to contact Croesus (Lee and Mitchell, 1994). It is

possible that some of those unsolicited hires were even laid off. Negotiation theory suggests that

workers’ pay is shaped in part by the value of their alternatives (Bazerman and Neale, 1992); we

would therefore expect that unsolicited hires would receive lower pay than intermediated hires,

but the observed differences are very small.

DISCUSSION

As firms increasingly use both external hiring and internal mobility to staff higher-level jobs, we

need to understand whether these different forms of worker mobility lead to different

employment outcomes. To do so, this paper bridges theories of internal and external labor market

matching, arguing that the routes by which workers enter jobs affects both the specific skills that

workers bring with them and the information available to match those workers to jobs. Results

show that external hires have worse performance than internal movers while being paid

substantially more. Compared with workers entering a job through simple promotion, the most

common form of internal mobility, external hires receive significantly lower performance

evaluations for their first two years in the job yet are initially paid around 18 per cent more.

Hires also have much higher rates of exit from the job but, if they stay, faster subsequent

promotion.

The results also showed that employment outcomes depend on how individuals move within the

organization, as the nature of the prior job shapes the resources that workers bring to their new

position. Workers who are simultaneously promoted and transferred have lower performance

than other kinds of internal movers, consistent with the gap between the specific skills that they

bring with them and those that they need for their new job. Lateral transfers have higher

experience and pay than other internal movers as well as higher rates of subsequent promotion

and exit. Those effects likely reflect the transfers’ accrued experience at a higher organizational

rank, as well as the selection of poorly performing workers for transfers. These results advance

our knowledge of how hiring competes with internal mobility, as well as the consequences of

different mobility paths within the labor market.

Performance, Skills, and Mobility

The results demonstrate how the way that workers enter their jobs affects their subsequent

performance. Both new hires and those who are simultaneously promoted and transferred

initially perform significantly worse than other workers. New hires’ performance converges with

that of internal movers over three years, suggesting that the new hires are not systematically less

able than those promoted to the job. Instead, the evidence suggests that hires and workers who

are simultaneously promoted and transferred must learn new specific skills before they can

perform as well as the other internal movers, who already have those skills.

The performance gap between hires and internal movers is particularly interesting given the

differences in their observable characteristics: hires have more experience and education yet still

perform worse than workers entering the job through simple promotions and simple transfers.

That lower initial performance underlines the importance of both firm-specific skills and

unobservable attributes in shaping performance, even in a setting in which interfirm mobility is

very common.

One concern in interpreting the performance results is that the measures could be affected by

supervisory bias. Although I lack objective performance data, there is variation in how the

performance measures reflect objective outcomes. Although competence measures are largely

subjective, contribution tracks the more concrete achievement of objectives. As table 2 revealed,

results were actually stronger for the more objective contribution measure than for the

competence or performance rank measures. Similarly, performance in some jobs is based on

clearly measurable outcomes; in particular, traders and salespeople have clearly measurable

performance in terms of profit and revenue. Supplementary analyses showed that performance

differences between workers entering jobs through promotions and external hires were slightly

greater for traders and salespeople, not weaker as we would expect if differences were driven by

subjective biases. These results suggest that supervisory biases may actually minimize

differences between internal movers and external hires. Those biases may reflect an escalation of

commitment following a manager’s decision to hire a worker (Schoorman, 1988) and are

consistent with evidence that objective performance measures are more sensitive to

organizational tenure than subjective measures (Sturman, 2003).

The findings showing that external hires perform worse than internal movers run contrary to

predictions based on tournament theory. Chan (1996) suggested that firms’ reluctance to weaken

internal incentives by substituting hiring for promotion would lead them to hire only workers

who were expected to perform substantially better than the workers they would promote. The

higher observable human capital and faster rates of promotion of external hires observed in other

studies has been used to support this theory (Baker, Gibbs, and Holmstrom, 1994a; Chan, 2006;

Oyer, 2007). Though I replicated the prior results about promotion rates and observable human

capital, I found that hires have a significantly lower average performance than workers promoted

into the job. This result suggests that the stronger observable human capital and faster promotion

of external hires reflect the need to compensate for their lower and less certain unobservable

skills, not their higher overall ability.

The results also show that the performance effects of internal mobility depend on the nature of

that mobility. The overwhelming majority of internal moves in my sample, both simple

promotions and simple transfers, resulted in higher initial performance than hiring. The superior

performance of simple transfers compared with external hires suggests that the performance

advantages of workers who entered jobs through simple promotions are not a consequence of

their remaining in exactly the same job: simple transfers have performance similar to promoted

workers, despite moving to a different part of the organization. This strong performance of

simple transfers is particularly noteworthy, given that those transfers usually followed poor

performance in a prior role.

I did find, though, that the 5 per cent of internal moves that combined promotions and transfers

led to performance that was little better than external hires. One interpretation of the

performance similarities between hires and workers entering the job by combined promotions

and transfers is that the kinds of skills that we think of as firm-specific are highly job-specific

too. Radical job moves inside the organization may be as disruptive to working relationships and

critical job knowledge as moves to another firm. An alternative interpretation is that internal

moves to very different jobs require workers to develop different skills from external hires. If

new hires usually come from similar roles in other organizations, they may experience less

change in much of their work content than do the workers entering jobs by combined promotions

and transfers. It may be that the need to learn the content of a new kind of job has similar

consequences for performance as the loss of firm-specific skills.

The two hypotheses that were not supported were H4b and H5b, which suggested that workers

entering their jobs through combined promotions and transfers would have lower rates of

promotion and higher rates of involuntary exit than other internal movers. The lack of support for

those hypotheses may reflect managers’ recognition of the difficulties involved in making such

complex moves and their forebearance during personnel decisions.

Incomplete Information, Worker Characteristics and Pay

The results also demonstrate how differences between hiring and internal mobility can affect the

characteristics of workers entering jobs through those routes, and their subsequent pay. The

findings are consistent with two pathways through which incomplete information affects hires’

pay: external hires have stronger observable indicators of ability, which should raise the wages

that they command; there is also evidence that external hires’ poorer expected fit with their jobs

may increase their pay. Farber and Gibbons (1996) argued that pay differentials due to indicators

of human capital should be constant over time, as those indicators reflect differences in workers’

value to employers. That external hires’ salary premium declines over time therefore suggests

that this premium may not be due solely to their higher externally observable indicators of

ability. External hires’ higher turnover and termination rates also provide strong evidence of their

poorer fit with their jobs, suggesting that they should demand higher wages than those who were

promoted. The fact that employees’ referrals received lower total compensation than other hires

is also consistent with workers demanding less pay when they know more about the job,

although results for salary and bonus were not significant.

These findings on hires’ compensation extend Harris and Helfat’s (1997) research on CEO labor

markets. Their cross-firm study found that externally hired CEOs were paid more than internally

promoted CEOs. I show that this phenomenon holds for workers being hired into similar jobs

within the same organization and that it occurs despite a large performance gap between the

external hires and promoted workers. I also shed new light on the mechanism behind this effect.

While Harris and Helfat suggest that the pay premium for external CEOs reflects demands for

scarce specialist skills and the need to compensate hires for a loss of firm-specific skills, I

provide evidence that the pay premium also reflects a need for hires to have stronger general

observable attributes and the risk that they may be a poor fit for their new job.

I also found that levels of observable characteristics and pay varied systematically across

different kinds of internal movers. In particular, transfers had significantly more experience than

did those promoted, as well as higher pay, promotion, and involuntary exit rates. These

differences between transfers and promoted workers likely reflect the increased time that

transfers have spent in the higher levels of the firm, experience that translates into higher pay and

improved prospects for promotion. It is possible that some of these results could also reflect the

specific way that transfers were used in this organization: supplementary analyses confirmed that

lateral transfers usually followed poor performance in a prior role, suggesting that the transfers

provided a second chance to those who were struggling in their existing positions. Such use of

transfers to find improved fit would explain the increased rate of both promotions and exits for

transfers: when transfers do well in a new position, they appear to be promoted rapidly; when

they struggle, they are swiftly terminated. These findings raise the question, though, of why the

organization gives transfers a second chance, given the costs of doing so. My other findings

suggest an answer: though transfers were more expensive than promoted workers, they had both

higher performance and lower costs than the external hires who were their potential

replacements.

Implications for Theory and Future Research

The findings of this paper could be developed in a number of ways. First, future research should

explore how the results from this single site generalize elsewhere. One important influence on

the generalizability of the results is likely to be the demands for firm-specific skills in different

jobs and organizations, especially because an untested corollary of my arguments is that

performance differences between hires and promoted workers will increase as demands for firm-

specific skills increase. Although workers in investment banking are notoriously mobile, prior

work has indicated that firm-specific skills are important in this setting (Groysberg, et Lee, and

Nanda, 2008). Performance differences between hires and promoted workers should be lower in

settings with fewer demands for firm-specific skills.

Similarly, the effects of promotion versus hiring should depend in part on the nature of

promotion. At Croesus, almost all promotions involve some measure of continuity with the prior

job. Evidence from other studies suggests that such rank promotions are the norm in the broader

labor market (Pergamit and Veum, 1999). The results of this study should therefore generalize to

other settings, particularly those involving professional work in which responsibilities accrete

gradually. Performance differences between hires and promoted workers should be lower in

settings in which promotions involve greater changes in task content.

I conducted a partial exploration of how the findings might generalize by exploring their

applicability across occupations and organizations. I carried out separate analyses on the

investment professionals (traders, salespeople, research analysts, and investment bankers) versus

support staff at Croesus to check that the findings were not being driven by a particular

occupational group. For the core findings of pay and performance I found consistent, significant

results in each of these two groups, with the one exception of total compensation for simple

transfers, which was higher among support staff. Second, I replicated the analyses in two other

firms: an investment bank and a publishing company, although the less detailed data provided by

these firms did not enable me to conduct all of the analyses here. In both cases, I found the same

effects of paying more for external hires while providing them with lower performance ratings.

Future research should address when firms choose to fill jobs through hiring versus internal

mobility. In part, such research could address concerns about the endogeneity of staffing

decisions, which might bias the results of this paper if factors that correlated with decisions to

hire also correlate with pay or performance. I dealt with this issue in part by using very detailed

controls for different jobs, including the levels of hiring and turnover in each group. These

controls allowed me to hold constant a wide variety of measurable characteristics of the job,

comparing workers who are hired and promoted into very similar jobs. A natural experiment or

valid instrumental variable would provide a more robust approach to assessing causality.

Although I was not able to find such an instrument in this study, the use of one would contribute

to this research in the future.

A better understanding of the reasons for hiring would also be particularly interesting, given the

high costs and poor performance of new hires documented here. It is possible that hiring reflects

the difficulties of staffing all jobs by promotion (Cappelli, 2008), perhaps because the work

requires relatively low ratios of junior workers to senior workers or because demands for

particular kinds of workers are growing very rapidly. Alternatively, the firm might be prepared to

accept poorer average performance from external hires if they receive disproportionate benefits

from those few hires that turn out to have very high performance. Evidence from supplementary

analyses (not reported here) suggests that firms are not hiring to find a few “stars,” though, as

hiring was less likely in positions with more variation in individual performance, which are also

the positions for which stars should be most valuable.

It would also be valuable to gather data on the jobs that external hires had come from, to

establish whether their jobs had been at the same rank as the one that they entered at Croesus and

whether they had been working in the same area as their new job. The similar experience of

transfers and new hires suggests that external hires may be coming from similar levels in other

organizations. My interviews also suggested that the organization was more likely to hire from a

parallel level: experience at a similar level was an important indicator of ability to the firm; while

Croesus might be willing to let an internal candidate learn the skills needed to operate in a new

rank or area, external hires would be expected to have demonstrated relevant experience already.

A further important area for future study is the application of these ideas to gender and racial

differences within organizations. I explored how problems of incomplete information led to

differences between hires and internal movers in the achieved characteristics of experience and

education. Much other research has explored how ascriptive characteristics such as race and

gender are also used as indicators in hiring (Phelps, 1972; Heilman, 1980; Petersen and Saporta,

2004). Gender in particular may play an important role in employers’ inferences when jobs

correspond to gender stereotypes (Cejka and Eagly, 1999), as is the case in investment banking

(Groysberg, 2010). Gender and race differ substantially from experience and education, in that

they do not relate directly to workers’ productivity. Nonetheless, when employers use such

characteristics as indicators of ability, we might expect to see gender and racial differences in

how workers enter jobs. I found some evidence that women are more likely to enter jobs through

promotion rather than external hiring, consistent with the findings of Petersen and Saporta (2004)

(but see Gorman and Kmec, 2009). Given the absence of data on both internal and external

applicant pools (Fernandez and Sosa, 2005), however these results should be interpreted with

caution. To the extent that women are less likely to reach higher-level jobs through external

hiring, though, the results of this study have implications for overall gender inequality, because

hires are better paid than internal movers. Further exploration of the way that indicators such as

race and gender are used in promotion versus hiring decisions could therefore form a valuable

extension to this research.

The findings of this study contribute to internal labor market theory by showing how the

allocation of jobs and rewards in organizations is affected by hiring into higher levels. A central

focus of internal labor market theory has always been to explain the patterns of pay and mobility

in organizations (Stewman, 1986; Osterman, 1987; Althauser, 1989), yet this literature’s focus

on internal mobility has precluded detailed analysis of how external labor markets penetrate

these structures. I contribute to this literature by showing that outcomes and mobility in internal

labor markets can vary substantially based on whether workers were hired, promoted, or

transferred into their current jobs. Such a finding is particularly important as firms become more

open to external hiring.

The study also has implications for research on interorganizational careers (Arthur and

Rousseau, 1996; Brett and Stroh, 1997). Building on Granovetter (1981), I show that how

workers arrive in a position has important consequences for their careers, driving both their pay

and future prospects. The results emphasize that cross-firm moves represent a double-edged

sword for workers: holding the nature of the job constant, those who enter a job from outside

receive higher pay but face a higher risk of termination. The results also provide a window into

when people choose to build inter- or intraorganizational careers by showing how workers with

stronger indicators of observable ability are better able to reach higher-level jobs externally.

Perhaps most importantly, the paper provides unique evidence on the value to firms of internal

labor market structures. Results show that internal mobility allows the firm to staff higher-level

jobs with workers who have better performance but are paid less. These results provide further

insight into the effects of turnover on organizational performance (Glebbeek and Bax, 2004;

Shaw, Gupta, and Delery, 2005) by specifying some of the costs of external hiring. They also

contribute to debates on the functions of internal labor markets (Doeringer and Piore, 1971;

Jacoby, 1985; Althauser, 1989; Cappelli, 2000) by providing unique evidence that workers

promoted into jobs via the internal labor market do in fact have higher levels of firm-specific

skills. The paper also identifies a novel benefit of internal labor markets - lowering wage costs

by reducing the uncertainty that firms and workers face in the matching process. By detailing the

strong advantages of internal mobility over external hires, these findings help to explain the

continued resilience of internal labor markets in the face of pressures for worker mobility.

Acknowledgements

I am very grateful to Heski Bar-Isaac, William Barnett, Forrest Briscoe, Diane Burton, Peter

Cappelli, Olivier Chatain, Isabel Fernandez-Mateo, Adam Grant, Martine Haas, Rahul Kapoor,

Katherine Klein, John Paul MacDuffie, Ethan Mollick, Felipe Monteiro, Evan Rawley, Nancy

Rothbard, Jesper Sørensen, Valery Yakubovich, Henrich Greve, three anonymous reviewers and

participants at the Stanford Organizational Behavior Seminar for their helpful comments and

suggestions. I would also like to thank the various people at Croesus, who helped me access and

understand the data.

REFERENCES

Akerlof, G. A.

1970 "The Market for ’Lemons’: Quality uncertainty and the market mechanism." Quarterly

Journal of Economics, 84: 488-500.

Allison, P.D.

1984 Event History Analysis : Regression for Longitudinal Event Data. Beverly Hills, Calif.:

Sage Publications.

Allison, P. D., and J. S. Long

1990 "Departmental effects on scientific productivity." American Sociological Review, 55:

469-478.

Althauser, R. P.

1989 "Internal labor markets." Annual Review of Sociology, 15: 143-161.

Althauser, R. P., and A. L. Kalleberg

1981 "Firms, occupations and the structure of labor markets: A conceptual analysis." In I.

Berg (ed.), Sociological Perspectives on Labor Markets: 119-149. New York: Academic Press.

Arthur, M. B., and D. M. Rousseau

1996 The Boundaryless Career : A New Employment Principle for a New Organizational Era.

New York: Oxford University Press.

Arvey, R. D., and J. E. Campion

1982 "The employment interview: A summary and review of recent research." Personnel

Psychology, 35: 281-322.

Baker, G., M. Gibbs, and B. Holmstrom

1994a "The internal economics of the firm: Evidence from personnel data." Quarterly Journal of

Economics, 109: 881-919.

Baker, G., M. Gibbs, and B. Holmstrom

1994b "The wage policy of a firm." The Quarterly Journal of Economics, 109: 921-955.

Barnett, W. P., and A. S. Miner

1992 "Standing on the shoulders of others: Career interdependence in job mobility."

Administrative Science Quarterly, 37: 262-281.

Baron, J. N., and W. T. Bielby

1986 "The proliferation of job titles in organizations." Administrative Science Quarterly, 31:

561-586.

Bazerman, M. H., and M. A. Neale

1992 Negotiating Rationally. New York, and Toronto: Free Press and Maxwell Macmillan

Canada.

Becker, G. S.

1962 "Investment in human-capital: A theoretical analysis." Journal of Political Economy, 70:

9-49.

Bills, D. B.

1990 "Employers use of job history data for making hiring decisions: A fuller specification of

job assignment and status attainment." Sociological Quarterly, 31: 23-35.

Bills, D. B.

1999 "Labor market information and selection in a local restaurant industry: The tenuous

balance between rewards, commitments and costs." Sociological Forum, 14: 583-607.

Blumen, I., M. Kogan, and P. J. McCarthy

1955 The Industrial Mobility of Labor as a Probability Process. Ithaca, NY: Cornell

University.

Brett, J. M., and L. K. Stroh

1997 "Jumping ship: Who benefits from an external labor market career strategy?" Journal of

Applied Psychology, 82: 331-341.

Campbell, J. P., R. A. McCloy, S. H. Oppler, and C. E. Sager

1993 "A theory of performance." In N. Schmitt and W. C. Borman (eds.), Personnel Selection

in Organizations: 35-70. San Francisco: Jossey Bass.

Cappelli, P.

1999 The New Deal at Work: Managing the Market-driven Workforce. Boston: Harvard

Business School Press.

Cappelli, P.

2000 "Market mediated employment: The historical context." In M. Blair and T. A. Kochan

(eds.), The New Employment Relationship pp 66-101: Brookings New York Institution

Cappelli, P.

2008 Talent on Demand: Managing Talent in an Age of Uncertainty. Boston: Harvard Business

Press.

Cascio, W. F.

1998 Applied Psychology in Human Resource Management, 5th ed. Upper Saddle River, NJ:

Prentice Hall.

Castilla, E. J.

2005 "Social networks and employee performance in a call center."American Journal of

Sociology, 110: 1243-1283.

Cejka, M. A., and A. H. Eagly

1999 "Gender-stereotypic images of occupations correspond to the sex segregation of

employment." Personality and Social Psychology Bulletin, 25: 413-423.

Chan, W.

1996 "External recruitment versus internal promotion." Journal of Labor Economics, 14: 555-

570.

Chan, W.

2006 "External recruitment and intrafirm mobility." Economic Inquiry, 44: 169-184.

Chatman, J. A.

1991 "Matching people and organizations: Selection and socialization in public accounting

firms." Administrative Science Quarterly, 36: 459-484.

Cohen, L. E., J. P. Broschak, and H. A. Haveman

1998 "And then there were more? The effect of organizational sex composition on the hiring

and promotion of managers." American Sociological Review, 63: 711-727.

Crain, R. L.

1984 "The quality of American high school graduates: What personnel officers say and do

about it." John Hopkins University, Center for Social Organization of Schools.

Dencker, J. C.

2009 "Relative bargaining power, corporate restructuring, and managerial incentives."

Administrative Science Quarterly, 54: 453-485.

DiPrete, T. A.

1987 "Horizontal and vertical mobility in organizations." Administrative Science Quarterly,

32: 422-444.

DiPrete, T. A., and W. T. Soule

1988 "Gender and promotion in segmented job ladder systems." American Sociological

Review, 53: 26-40.

Doeringer, P. B., and M. J. Piore

1971 Internal labor markets and manpower analysis. Lexington, MA: D.C. Heath.

Dokko, G., S. L. Wilk, and N. P. Rothbard

2009 "Unpacking prior experience: How career history affects job performance." Organization

Science, 20: 51-68.

Dustmann, C., and C. Meghir

2005 "Wages, experience and seniority." Review of Economic Studies, 72: 77-108.

Edwards, J. R.

1991 "Person-job fit: A conceptual integration, literature review and methodological critique."

In C. L. Cooper and I. T. Robertson (eds.), International Review of Industrial and Organizational

Psychology: 283-357: Wiley

Farber, H. S.

1994 "The analysis of interfirm worker mobility." Journal of Labor Economics, 12: 554-593.

Farber, H. S.

2008 "Employment insecurity: The decline in worker-firm attachment in the United States."

Princeton University Industrial Relations Section Working Paper.

Farber, H. S., and R. Gibbons

1996 "Learning and wage dynamics." Quarterly Journal of Economics, 111: 1007-1047.

Fernandez, R. M., E. J. Castilla, and P. Moore

2000 "Social capital at work: Networks and employment at a phone center." American Journal

of Sociology, 105: 1288-1356.

Fernandez, R. M., and M. L. Sosa

2005 "Gendering the job: Networks and recruitment at a call center." American Journal of

Sociology, 111: 859-904.

Fine, J. P., and R. J. Gray

1999 "A proportional hazards model for the subdistribution of a competing risk." Journal of the

American Statistical Association, 94: 496-509.

Gibbons, R., and L. F. Katz

1991 "Layoffs and lemons." Journal of Labor Economics, 9: 351-380.

Gibbons, R., and M.Waldman

2006 "Enriching a theory of wage and promotion dynamics inside firms." Journal of Labor

Economics, 24: 59-107.

Glebbeek, A. C., and E. H. Bax

2004 "Is high employee turnover really harmful? An empirical test using company records. "

Academy of Management Journal, 47: 277-286. Academy of Management.

Gorman, E. H., and J. A. Kmec

2009 "Hierarchical rank and women's organizational mobility: Glass ceilings in corporate law

firms." American Journal of Sociology, 114: 1428-1474.

Granovetter, M. S.

1974 Getting a Job: A Study of Contacts and Careers. Cambridge, MA: Harvard University

Press.

Granovetter, M. S.

1981 "Toward a sociological theory of income differences." In I. Berg (ed.), Sociological

Perspectives on Labor Markets: 11-47. New York: Academic Press.

Greenwald, B. C.

1986 "Adverse selection in the labour market." Review of Economic Studies, 53: 325-347.

Greve, H. R., and T. Fujiwara-Greve

2003 "Job search with organizational size as a signal." Social Forces, 82: 643-669.

Groysberg, B.

2010 Chasing Stars: The Myth of Talent and the Portability of Performance. Princeton, NJ:

Princeton University Press.

Groysberg, B., L.-E. Lee, and A. Nanda

2008 "Can they take it with them? The portability of star knowledge workers' performance."

Management Science, 54: 1213-1230.

Halaby, C. N.

1988 "Action and information in the job mobility process: The search decision." American

Sociological Review, 53: 9-25.

Harris, D., and C. Helfat

1997 "Specificity of CEO human capital and compensation." Strategic Management Journal,

18: 895-920.

Heilman, M. E.

1980 "The impact of situational factors on personnel decisions concerning women: Varying the

sex composition of the applicant pool." Organizational Behavior and Human Performance, 26:

386-395.

Huckman, R. S., and G. P. Pisano

2006 "The firm specificity of individual performance: Evidence from cardiac surgery."

Management Science, 52: 473-488.

Jacoby, S. M.

1985 Employing Bureaucracy: Managers, Unions, and the Ttransformation of Work in

American Industry, 1900-1945. New York: Columbia University Press.

Lee, T. W., and T. R. Mitchell

1994 "An alternative approach: The unfolding model of voluntary employee turnover."

Academy of Management Review, 19: 51-89.

Lepak, D. P., and S. A. Snell

1999 "The human resource architecture: Toward a theory of human capital allocation and

development." The Academy of Management Review, 24: 31-48.

Medoff, J. L., and K. G. Abraham

1981 "Are those paid more really more productive? The case of experience." Journal of Human

Resources, 16: 186-216.

Miller, S. R., and J. E. Rosenbaum

1997 "Hiring in a Hobbesian Worldworld: Social infrastructure and employers' use of

information." Work and Occupations, 24: 498-523.

Mincer, J.

1962 "On-the-job training: Costs, returns, and some implications." Journal of Political

Economy, 70: 50-79.

Mincer, J.

1970 "The distribution of labor incomes: A survey with special references to the human capital

approach." Journal of Economic Literature, 8: 1-26.

Mouw, T.

2003 "Social capital and finding a job: Do contacts matter?" American Sociological Review,

68: 868-898.

O'Mahony, S., and B. A. Bechky

2006 "Stretchwork: Managing the career progression paradox in external labor markets."

Academy of Management Journal, 49: 918-941.

O'Reilly, C. A. III, J. Chatman, and D. F. Caldwell

1991 "People and organizational culture: A profile comparison approach to assessing person-

organization fit." Academy of Management Journal, 34: 487-516.

Osterman, P.

1987 "Choice of employment systems in internal labor markets." Industrial Relations, 26: 46-

67.

Oyer, P.

2007 "Is there an insider advantage in getting tenure?" American Economic Review, 97: 501-

505.

Pergamit, M. R., and J. R. Veum

1999 "What is a promotion?" Industrial and Labor Relations Review, 52: 581-601.

Petersen, T., and I. Saporta

2004 "The opportunity structure for discrimination." American Journal of Sociology, 109: 852-

901.

Phelps, E. S.

1972 "The statistical theory of racism and sexism." American Economic Review, 62: 659-661.

Posthuma, R. A., F. P. Morgeson, and M. A. Campion

2002 "Beyond employment interview validity: A comprehensive narrative review of recent

research and trends over time." Personnel Psychology, 55: 1-80.

Quinones, M. A., J. K. Ford, and M. S. Teachout

1995 "The relationship between work experience and job performance: A conceptual and meta-

analytic review." Personnel Psychology, 48: 887-910: Wiley-Blackwell.

Rosenbaum, J. E.

1979 "Tournament mobility: Career patterns in a corporation." Administrative Science

Quarterly, 24: 220-241.

Rosenbaum, J. E.

1990 "Structural models of organizational careers: A critical review and new directions." In R.

L. Breiger (ed.), Social Mobility and Social Structure: 272-307. Cambridge: Cambridge

University Press.

Rosenbaum, J. E., T. Kariya, R. Settersten, and T. Maier

1990 "Market and network theories of the transition from highschool to work: Their

application to industrialized societies." Annual Review of Sociology, 16: 263-299.

Royal, C., and R. P. Althauser

2003 "The labor markets of knowledge workers: Investment bankers' careers in the wake of

corporate restructuring." Work and Occupations, 30: 214-233.

Schmidt, F. L., and J. E. Hunter

1998 "The validity and utility of selection methods in personnel psychology: Practical and

theoretical." Psychological Bulletin, 124: 262-274.

Schneider, B., H. W. Goldstein, and D. B. Smith

1995 "The ASA framework: An update." Personnel Psychology, 48: 747-773:

Schoorman, F. D.

1988 "Escalation bias in performance appraisals: An unintended consequence of supervisor

participation in hiring decisions." Journal of Applied Psychology, 73: 58-62.

Shaw, J. D., N. Gupta, and J. E. Delery

2005 "Alternative conceptualizations of the relationship between voluntary turnover and

organizational performance." Academy of Management Journal, 48: 50-68. Academy of

Management.

Sonnenfeld, J. A., and M. A. Peiperl

1988 "Staffing policy as a strategic response: A typology of career systems." Academy of

Management Review, 13: 588-600.

Spence, M.

1973 "Job market signaling." Quarterly Journal of Economics, 87: 355-374.

Spilerman, S., and T. Lunde

1991 "Features of educational attainment and job promotion prospects." American Journal of

Sociology, 97: 689-720.

Stewman, S.

1986 "Demographic models of internal labor markets." Administrative Science Quarterly, 31:

212-247.

Stewman, S., and S L. Konda

1983 "Careers and organizational labor markets: Demographic models of organizational

behavior." American Journal of Sociology, 88: 637-685.

Stewman, S., and K. S. Yeh

1991 "Structural pathways and switching mechanisms for individual careers." In R. Althauser

and M. Wallace eds. Research in Social Stratification and Mobility, 10: 133-168. JAI Press,

Greenwich Conn.

Sturman, M. C.

2003 "Searching for the inverted U-shaped relationship between time and performance: Meta-

analyses of the experience/performance, tenure/performance, and age/performance

relationships." Journal of Management, 29: 609-640.

Sturman, M. C., R. A. Cheramie, and L. H. Cashen

2005 "The impact of job complexity and performance measurement on the temporal

consistency, stability, and test-retest reliability of employee job performance ratings." Journal of

Applied Psychology, 90: 269-283.

Waldman, M.

1984 "Job assignments, signaling, and efficiency." RAND Journal of Economics, 15: 255-267.

White, H. C.

1970 Chains of Opportunity: System Models of Mobility in Organizations. Cambridge, MA:

Harvard University Press.

Table 1. Summary Statistics for Key Variables*

Variable Mean S.D. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18

Simple promotion

.55

.50

Promotion/transfer

.03

.18

-.20

Simple transfer

.10

.30

-.37

-.06

Experience

14.82

6.90

-.12

-.07

.13

Highest bachelor’s

.53

.50

.08

.03

.02

-.08

Highest Master’s

.37

.48

-.04

-.03

-.02

.02

-.81

Highest Ph.D.

.08

.27

-.08

-.01

.02

-.31

-.23

Ranked

performance

3.74

1.35

.08

-.01

-.04

-.15

.03

-.03

.01

Contribution

3.72

.65

.12

-.03

-.08

.01

-.01

.56

10.

Competence

3.71

.70

.07

.01

-.04

-.13

-.01

.01

.52

.54

11.

Log total

compensation 12.56 .88 .03 -.04 -.05 .12 -.13 .11 .06 .18 .22 .24

12. Log bonus 10.74 3.60 .05 0 -.04 -.04 -.06 .05 .05 .16 .2 .2 .63

13. Log salary 11.72 .26 -.19 -.07 .05 .43 -.19 .12 .15 .06 .05 .02 .62 .21

14. Time in job 1.76 1.14 .13 -.01 -.15 .24 0 .01 -.01 .09 .08 .05 .09 -.05 .19

15. Year 2007 1.76 -.02 .04 .08 .11 .02 -.01 -.02 .04 .01 -.04 -.17 -.33 .18 .26

16. Rank 3.02 .98 .06 -.04 .01 .41 -.12 .09 .09 .03 .08 .04 .68 .27 .81 .14 .03

17. Male .77 .42 -.02 -.01 -.01 .01 -.04 .05 -.02 .02 0 .03 .22 .08 .17 -.01 -.05 .14

18. White .78 .41 .06 0 .02 .14 .09 -.09 -.02 .04 .04 .04 .13 .1 .11 .07 -.07 .2 .04

19. Asian .16 .37 -.05 0 -.03 -.14 -.1 .1 .02 -.03 -.02 -.02 -.11 -.08 -.07 -.06 .07 -.17 -.04 -.84

*Unit of analysis is a year spent in a job. Each observation contains pay in that year, and most recent performance evaluation.

Table 2. Ordered Logit Analyses of Performance.*

Dependent Variable

Ranked performance

Competence

Contribution

Model 1

Model 2

Model 3 Model 4 Model 5 Model 6

Time in job

.232

.387

0.163

0.233

0.198

0.408

(.020)

(.038)

(0.019)

(0.036)

(0.019)

(0.038)

Group turnover proportion at hire

.019

-.087

-0.171

-0.22

-0.211

-0.355

(.22)

(.214)

(0.205)

(0.206)

(0.237)

(0.234)

Group new hire proportion at hire

.325

.243

0.480

0.440

0.497

0.384

(.154)

(.156)

(0.158)

(0.159)

(0.168)

(0.169)

Group new hire proportion

-.176

.036

-0.281

-0.181

-0.344

-0.058

(.126)

(.132)

(0.132)

(0.139)

(0.135)

(0.140)

Group turnover proportion

-.346

-.297

0.007

0.032

0.087

0.165

(.176) (.175) (.170) (.170) (.184) (.182)

Experience

-.078

-.076

-.046

-.051

(.004)

Male

.109

.034

.005

.003

(.052)

(.055)

Asian

.054

.057

.195

.196

.208

.215

(.104)

(.105)

(.110)

(.115)

(.116)

White

.267

.272

.304

.306

.305

.315

(.096)

(.097)

(.099)

(.105)

(.106)

Highest Bachelor’s

-.165

-.18

-.318

-.325

-.352

-.374

(.154)

(.144)

(.140)

Highest Master’s

-.257

-.271

-.417

-.423

-.438

-.457

(.156)

(.147)

(.148)

(.143)

Highest Ph.D.

-.187

-.194

-.17

-.174

-.397

-.410

(.176)

(.168)

(.169)

Simple promotion

.168

.509

.148

.308

.427

.893

(.054)

(.072)

(.054)

(.077)

(.055)

(.077)

Promotion and transfer

-.122

.073

.071

.138

.182

.343

(.119)

(.156)

(.120)

(.188)

(.130)

(.180)

Simple transfer

.115

.386

.146

.258

.238

.612

(.073) (.096) (.074) (.100) (.074) (.099)

Simple prom x Time in job

-.218

-.101

-.297

(.042)

(.041)

(.042)

Prom. and transfer x Time in job

-.125

-.041

-.094

(.096)

(.117)

Simple transfer x Time in job

-.191

-.076

-.265

(.064) (.063) (.062)

Cut 1

-2.236

-2.011

-8.244

-8.145

-7.282

-6.993

(.405)

(.412)

(.606)

(.610)

(.731)

(.735)

Cut 2

-1.771

-1.546

-4.249

-4.150

-2.887

-2.599

(.405)

(.412)

(.445)

(.449)

(.445)

(.451)

Cut 3

.839

1.070

-.706

-.605

.762

1.064

(.405)

(.412)

(.441)

(.446)

(.442)

(.448)

Cut 4

1.184

1.416

2.049

2.150

3.906

4.217

(.405)

(.413)

(.441)

(.446)

(.448)

(.455)

Cut 5

2.709

2.945

(.407)

(.414)

Observations

14,515

Log pseudo-likelihood

-20254.1

-20231.1

-14255

-14250

-13436

-13398

Chi-squared

13335

13269

8909

8901

7934

7930

Degrees of Freedom 123 126 122 125 122 125

Probability

.0000

Prob vs. prior model

.0000

.0979

.0000

P(simple prom. vs. prom. & transfer)

.01

.004

.50

.35

.05

.002

P(simple transfer vs. prom. & transfer)

.06

.05

.55

.53

.68

.14

Standard errors, in parentheses, are clustered by individual. All analyses also include

dummies for rank, function, interactions between each rank and function, division, year, and

greater New York area.

Table 3. Multinominial Logit Analysis of the Determinants of Mode of Job Entry (N= 7346)*

Model 1

Model 2

Variable

Simple promotion

Simple transfer

Promotion and

transfer

Simple promotion

Simple transfer

Promotion and

transfer

Group turnover proportion at job

entry

9.863

12.00

9.395

9.659

12.03

9.127

(.371)

(1.529)

(.589)

(.373)

(1.586)

(.595)

Group new hire proportion at job

entry

-7.158

.715

-3.826

-7.050

.987

-3.742

(.281)

(1.180)

(.549)

(.283)

(1.201)

(.546)

White

.143

-.257

.281

(.134)

(.804)

(.290)

Asian

-.014

-.572

-.0743

(.148)

(.920)

(.320)

Male

-.134

-.111

-.362

(.079)

(.484)

(.160)

Experience

-.0540

.034

-.079

(.0058)

(.030)

(.014)

Highest Bachelor’s

-.745

-.614

.077

(.246)

(1.157)

(.660)

Highest Master’s

-1.247

-.858

-.368

(.249)

(1.177)

(.666)

Highest Ph.D.

-1.689

-.692

-.905

(.273)

(1.226)

(.724)

Log likelihood

-4678

-4338

Chi-squared

2364

3316

Degrees of freedom

324

345

Probability

.0000

Probability versus prior model

(likelihood ratio)

.0000

*Standard errors in parentheses. All analyses include dummies for rank, function, division, year, and city. Base case is entry by hiring.

Table 4: OLS Regressions of Determinants of Pay (N=14,515)*

Dependent Variable

Log Salary

Log Bonus

Log Total Compensation

Model 1

Model 2

Model 3

Model 4

Model 5

Model 6

Full time

1.042

1.048

1.181

1.234

1.065

1.076

(.145)

(.621)

(.624)

(.182)

(.183)

Terminated during year

.0118

.0106

-9.680

-9.685

-.367

-.368

(.0171)

(.237)

(.236)

(.0611)

(.0609)

Days since hiring (max 365)

-.0002

-.00005

.0009

.0014

.0001

.0002

(.00003)

(.0006)

(.0007)

(.0001)

Group turnover proportion at entry

-.0128

-.0053

-.0317

-.0001

-.0153

-.0077

(.0130)

(.0129)

(.241)

(.0471)

(.0466)

Group hiring proportion at entry

.0166

.0188

-.152

-.144

.0303

.032

(.0101)

(.0100)

(.167)

(.0326)

(.0325)

Group hiring proportion

.0227

.013

.428

.396

.0709

.0634

(.0084)

(.0083)

(.197)

(.199)

(.032)

(.0326)

Group turnover proportion

.0282

.0265

-.339

-.344

-.0957

-.0971

(.0097)

(.0096)

(.258)

(.257)

(.0412)

(.0410)

Time in job

.0162

-.001

-.0387

-.0839

.0358

.0243

(.0011)

(.0022)

(.0239)

(.0552)

(.00455)

(.0113)

Experience

.0027

-.0095

-.0096

-.0024

(.0003)

(.0038)

(.0008)

Male

.0176

.0175

-.0391

-.0403

.0711

.0708

(.0037)

(.0436)

(.01)

Asian

.0076

.0154

.0166

-.0102

-.0099

(.0064)

(.0908)

(.0906)

(.0183)

(.0181)

White

-.0026

-.0028

.0324

.0332

-.0360

-.0358

(.0059)

(.0808)

(.0806)

(.0171)

(.0170)

Highest Bachelor’s

.0090

.0096

-.217

-.219

-.0521

-.0527

(.0116)

(.0117)

(.132)

(.0273)

(.0274)

Highest Master’s

.0259

.0262

-.197

-.201

-.0427

-.0436

(.0116)

(.134)

(.0280)

(.0281)

Highest Ph.D.

.0577

.0573

-.114

-.122

.0051

.0034

(.0127)

(.149)

(.0317)

(.0318)

Contribution

-.0009

-.0005

.233

.234

.0756

.0758

(.002)

(.0397)

(.0396)

(.0077)

Competence

.007

.0066

.114

.113

.0484

.0482

(.0017)

(.0360)

(.0064)

(.0063)

Ranked performance

.0091

.0092

.356

.357

.0783

.0785

(.001)

(.0181)

(.0034)

Simple promotion

-.113

-.161

-.137

-.294

-.129

-.167

(.0035)

(.0057)

(.0533)

(.121)

(.0116)

(.0234)

Promotion and transfer

-.113

-.171

-.108

-.423

-.176

-.264

(.0073)

(.0115)

(.103)

(.215)

(.0214)

(.0403)

Simple transfer

-.0376

-.0619

-.0488

.0493

-.0552

-.0384

(.0049)

(.0071)

(.075)

(.143)

(.0148)

(.0267)

Simple promotion x Time in job

.0233

.0769

.0184

(.0025)

(.0605)

(.0123)

Promotion and transfer x Time in job

.0305

.178

.0498

(.0049)

(.111)

(.0192)

Simple transfer x Time in job

.0079

-.108

-.0211

(.0035)

(.085)

(.0147)

R-squared

.863

.865

.69

.691

.845

Probability vs. prior model (Wald)

.0000

.0213

.0002

P (simple prom. vs. simple trans.)

.0000

.21

.0018

.0000

P (prom./trans. vs. simple trans.

.0000

.60

.022

.0000

*Standard errors, in parentheses, are clustered by individual. Includes controls for city, year,

rank, function, department, all interactions between rank and function, and all interactions

between rank and division.

Table 5. Cox Analysis of Hazard Rates of Mobility.*

Promotion Involuntary Exit Voluntary Exit

Variable

Model 1

Model 2

Model 3 Model 4 Model 5 Model 6

Full time

-.794

-1.104

-.097

-.332

-1.344

-1.302

(1.421)

(1.482)

(.290)

(.288)

(.235)

(.230)

Group turnover proportion

-.133

.33

3.609

3.424

2.973

2.853

(.356)

(.362)

(.133)

(.136)

(.156)

Group turnover proportion at entry

-1.615

-1.502

1.303

1.447

.580

.682

(.318)

(.345)

(.199)

(.214)

(.213)

Group hiring proportion

.402

.389

-1.937

-1.904

-.459

-.463

(.291)

(.300)

(.400)

(.399)

(.298)

Group hiring proportion at entry

-.274

-.154

-.922

-1.132

.286

.27

(.242)

(.252)

(.262)

(.269)

(.232)

(.233)

Experience

-.089

-.042

.0365

.0222

-.037

-.051

(.007)

(.0051)

(.005)

(.007)

Male

.233

.159

-.011

.008

.179

(.078)

(.08)

(.077)

(.078)

(.081)

Asian

.149

.309

.065

.062

.012

.042

(.148)

(.151)

(.141)

(.139)

(.143)

White

.322

.283

-.251

-.247

-.13

-.076

(.134)

(.138)

(.126)

(.131)

Highest Bachelor’s

-.796

-.661

.055

.067

.202

.164

(.224)

(.233)

(.202)

(.204)

(.302)

Highest Master’s

-.726

-.489

.222

.174

.064

-.026

(.225)

(.233)

(.206)

(.208)

(.304)

Highest Ph.D.

-.669

-.414

.254

.188

-.038

-.139

(.249)

(.256)

(.238)

(.240)

(.328)

(.329)

Log salary

3.729

2.296

-.574

-.017

1.263

1.644

(.316)

(.327)

(.296)

(.308)

(.306)

(.313)

Simple promotion

-.459

-.653

-.464

-.181

-.190

-.0631

(.077)

(.08)

(.088)

(.091)

(.087)

(.088)

Promotion and transfer

-.297

-.330

-.846

-.533

-.462

-.326

(.175)

(.179)

(.205)

(.206)

(.174)

Simple transfer

1.463

1.464

.284

.173

-.108

-.233

(.112)

(.116)

(.104)

(.107)

(.123)

(.126)

Ranked performance

.584

-.179

-.209

(.034)

(.033)

Contribution

-.041

-.481

-.01

(.065)

(.063)

Competence

.269

-.341

-.132

(.058)

(.056)

Observations

10,751

10,952

Log likelihood

-10226

-9857

-7610

-7435

-8001

-7953

Chi-squared

1728

2466

1850

2200

1391

1488

Degrees of freedom

638

641

148

151

148

151

Probability vs. prior model (LR test)

.0000

P(simple prom. vs. Prom. & trans.)

.34

.06

.08

.09

.11

P(simple trans. vs. Prom./ & rans.)

.000

.001

.48

.16

nly contain years 2006-2009. Promotions and transfers are treated as censored events. All

analyses also include controls for city, year, rank, function, all interactions between rank and

function. Promotion analyses control for department; exit analyses control for division.

Table 6. Ordered Logit Analyses of Determinants of Performance by Hiring Source (N =

3,792).

Ranked Performance

Competence

Contribution

Variable

Model 1

Model 2

Model 3

Model 4

Model 5

Model 6

Experience

-.078

-.076

-.039

-.038

-.052

-.049

(.008)

Male -.036 -.026 -.132 -.117 -.145 -.119

(.097) (.097) (.101) (.1) (.101) (.101)

Asian

.124

.119

.259

.241

.449

.432

(.149)

(.151)

(.174)

(.172)

(.176)

White .458 .466 .490 .477 .496 .499

(.137) (.139) (.156) (.154) (.156) (.156)

Highest Bachelor’s

-.285

-.283

-.306

-.315

-.953

-.979

(.255)

(.252)

(.308)

(.310)

(.294)

(.300)

Highest Master’s -.382 -.387 -.395 -.417 -1.002 -1.041

(.259) (.257) (.311) (.314) (.297) (.303)

Highest Ph.D.

-.558

-.555

-.251

-.266

-.959

-.994

(.290)

(.288)

(.339)

(.341)

(.335)

(.341)

Time in job .401 .393 .279 .263 .454 .431

(.041) (.041) (.042) (.041) (.043) (.0419)

Former employee

-.575

-.962

-.675

(.663)

(.485)

(.552)

Internet application -.219 -.445 -.101

(.221) (.209) (.201)

Temporary

.274

-.02

.411

(.306)

(.24)

(.338)

Unsolicited -.011 .095 .298

(.113) (.117) (.116)

Intermediated

-.340

-.204

-.274

(.092)

(.097)

(.099)

Employee referral

.200 .122

.052

(.081) (.085)

(.086)

Cut 1

-.559

-.561

-8.441

-7.829

-6.405

-6.373

(.865)

(.905)

(1.033)

(1.051)

(1.463)

(1.562)

Cut 2 -.162 -.165 -4.913 -4.302 -2.091 -2.061

(.865) (.907) (.866) (.884) (1.055) (1.189)

Cut3

2.565

2.551

-1.246

-.645

1.829

1.836

(.868)

(.908)

(.863)

(.883)

(1.054)

(1.187)

Cut 4 2.941 2.927 1.534 2.126 5.236 5.219

(.868) (.907) (.864) (.885) (1.060) (1.192)

Cut 5

4.497

4.480

(.871)

(.910)

Log pseudo likelihood -5121 -5131 -3646 -3655 -3307 -3327

Chi-squared 3616 3725 2831 2811 2539 2518

Degrees of Freedom

105

101

103

100

103

100

Probability

.0000

* Standard errors, clustered by individual, in parentheses. All analyses include dummies for rank,

function, interactions between rank and function, division, year, and greater New York area.

Table 7: Ordinary Least Squares Regression of Determinants of Compensation by Hiring Source

(N=3,792)*

Log Salary Log Bonus Log Total Compensation

Variable

Model 1

Model 2

Model 3

Model 4

Model 5

Model 6

Model 7

Model 8

Model 9

Experience

.002

.003

-.041

-.013

-.0132

-.009

-.0006

-.0009

(.0004)

(.008)

(.00166)

(.001)

(.0014)

Male

.013

.012

-.16

-.17

.05

.062

.06

(.006)

(.1)

(.09)

(.094)

(.02)

(.016)

Asian

.002

-.00008

-.00007

-.04

-.11

-.10

.005

-.017

-.018

(.01)

(.18)

(.17)

(.03)

(.026)

White

.006

.002

.004

-.17

.01

-.02

-.022

(.009)

(.16)

(.15)

(.03)

(.02)

(.023)

Time in job

.001

-.0008

-.0003

-.006

-.11

-.10

.03

.003

.004

(.002)

(.07)

(.01)

(.013)

(.012)

Group turnover

proportion at entry .03 .027 .03 -.027 -.18 -.23 -.038 -.13 -.14

(.03)

(.028)

(.03)

(.62)

(.54)

(.14)

(.13)

Group hire proportion at

entry -.03 -.029 -.03 -.14 -.05 .015 .047 .07 .074

(.02)

(.018)

(.40)

(.36)

(.088)

(.09)

(.091)

Group hire proportion

.015

.014

.012

.80

.78

.80

.049

.08

.081

(.013) (.013) (.013) (.35) (.34) (.34) (.053) (.05) (.052)

Group turnover

proportion .036 .035 .035 .77 .68 .69 -.034 -.16 -.16

(.018)

(.53)

(.52)

(.077)

(.085)

Full time

.79

.77

1.35

.74

.75

.81

.80

.81

(.24)

(.23)

(1.48)

(1.35)

(.20)

Terminated during year

-.061

-.017

.037

-14.53

-12.85

-13.13

-2.39

-2.08

-2.17

(.028)

(.027)

(.013)

(.49)

(.67)

(.55)

(.093)

(.14)

Days since hiring (max

-.00002

-.00006

.003

.0015

.0008

.0004

365)

(.00004)

(.0009)

(.0002)

Contribution

.002

.0013

.13

.12

.056

.054

(.003)

(.089)

(.09)

(.016)

Competence

.01

.0098

.15

.14

.049

.050

(.003)

(.07)

(.012)

Ranked performance

.009 .009

.46 .46

.087 .087

(.0016)

(.002)

(.04)

(.007)

Highest Bachelor’s

-.02

-.017

-.016

-.3

-.18

-.20

-.14

(.016)

(.015)

(.4)

(.36)

(.05)

(.037)

(.038)

Highest Master’s

-.004

.0005

.001

-.40

-.22

-.18

-.11

(.016)

(.015)

(.40)

(.36)

(.05)

(.038)

(.04)

Highest Ph.D. .02 .028 .028 -.48 -.25 -.25 -.13 -.062 -.063

(.017)

(.43)

(.38)

(.05)

(.046)

Former employee

.06

.064

-.19

-.16

-.12

-.037

(.026)

(.025)

(.46)

(.40)

(.06)

(.05)

Internet

-.001

.0017

.28

.37

-.021

.04

(.012)

(.24)

(.036)

(.03)

Temporary

-.0009

-.001

.20

.15

-.008

-.026

(.017)

(.016)

(.21)

(.18)

(.037)

(.043)

Unsolicited

.001

.0005

.053

.047

.036

.03

(.006)

(.13)

(.12)

(.028)

(.02)

Intermediated

.009

.011

.023

.14

.05

.062

(.005)

(.11)

(.098)

(.02)

(.021)

Employee referral

-.0077

-.11

-.044

(.0045)

(.08)

(.018)

R-squared .89 .895 .894 .701 .73 .729 .796 .87 .87

* Standard errors, clustered by individual, are in parentheses. Includes controls for city, year, all

interactions between rank and function, all interactions between rank and department, and division.

The Employment Relationship and Inequality: How and Why Changes in Employment Practices are Reshaping Rewards in Organizations

Article

Jun 2013

Stepping Sideways to Step up: Lateral Mobility and Career Advancement Inside Organizations

Article

Full-text available

Apr 2024

Although internal labor market theory emphasizes promotions as the main form of mobility within organizations, many internal job moves take people sideways into jobs that are at the same hierarchical level as the one that was left. Despite the prevalence of these lateral moves, however, we have little evidence on what role they play in workers’ careers. We argue that lateral mobility can facilitate subsequent career advancement by allowing for the development and demonstration of new skills and can, therefore, help those who would struggle to be promoted from their current job to develop their careers further. We establish empirical evidence on the implications of lateral mobility using eight years of personnel data from a large U.S. healthcare company. Our analyses show that those employees who move laterally are more likely to be subsequently promoted and achieve substantially higher pay growth than a matched sample of employees. We also find that lateral moves are more likely to be undertaken by those who have spent longer in the job but have lower performance than those who are promoted. This pattern of results suggests that lateral mobility provides an important avenue for career growth as people who step sideways in organizations are more likely to subsequently step up. This paper was accepted by Lamar Pierce, organizations. Supplemental Material: The online appendix and data files are available at https://doi.org/10.1287/mnsc.2021.03746 .

FFECT OF INTERNATIONAL FINANCIAL REPORTING STANDARDS (IFRSS) ADOPTION AND ORGANIZATIONAL MOBILITY OF PROFESSIONAL ACCOUNTANTS IN NIGERIA

Article

Full-text available

Feb 2024

The Role of Spiritual Leadership in Employee Loyalty in the General Company for Communications and Informatics in Iraq

Article

Jan 2024

The research aims to clarify the role that spiritual leadership exercises in the loyalty of workers in the General Company for Communications and Informatics in Iraq. The research was conducted on a deliberate sample of the company’s employees, numbering (100) respondents, based on a main hypothesis that states (strengthening spiritual leadership and loyalty of workers in the General Company for Communications and Informatics In Iraq) the research followed the descriptive analytical approach, basing its data analysis on the structural modeling equation (SEM). The research was able to prove its main hypothesis based on the results that showed a significant impact of spiritual leadership on the loyalty of employees in the General Company for Communications and Informatics in Iraq. The research recommended the need to establish Dimensions of spiritual leadership represented by (vision, belonging, faith, and altruism) among employees of the General Company for Communications and Informatics in Iraq.

Brokered Careers: The Role of Search Firms in Managerial Career Mobility

Article

Full-text available

May 2022

The authors explore how career paths are shaped by the involvement of search firms in hiring. Drawing on theories of market intermediation, they argue that search firms constrain horizontal moves across functions and industries by favoring workers from within the same function and industry as the role being filled. Using survey data on 1,342 job moves undertaken by 816 MBA alumni, the authors find that individuals who move jobs through a search firm experience lower horizontal mobility than those who move through other means. Findings also suggest that these results are not driven by firms’ decisions to use a search firm to fill the job. Supplementary analyses show no evidence that the job matches that are formed using search firms result in a better fit between workers and employers. Overall, the findings point to the significant institutional role search firms play in managerial careers.

Do Startup Employees Earn More in the Long Run?

Article

Full-text available

Jan 2021

Evaluating the attractiveness of startup employment requires an understanding of both what startups pay and the implications of these jobs for earnings trajectories. Analyzing Danish registry data, we find that employees hired by startups earn roughly 17% less over the next 10 years than those hired by large, established firms. About half of this earnings differential stems from sorting—from the fact that startup employees have less human capital. Long-term earnings also vary depending on when individuals are hired. Although the earliest employees of startups suffer an earnings penalty, those hired by already-successful startups earn a small premium. Two factors appear to account for the earnings penalties for the early employees: Startups fail at high rates, creating costly spells of unemployment for their (former) employees. Job-mobility patterns also diverge: After being employed by a small startup, individuals rarely return to the large employers that pay more.

No Vacancies? Building Theory on How Organizations Move People across Jobs

Chapter

Full-text available

Jul 2020

Matthew Bidwell

Training, Recruitment, and Outplacement as Endogenous Adverse Selection

Article

Aug 2022

This paper presents a model of a competitive labour market where workers vary in firm‐specific and general skills, and firms choose the kind of information to disclose. We show that provision of general training, recruitment, and outplacement are linked. Disclosing general human capital information on bad matches but revealing nothing about good matches (information disclosure that resembles outplacement support in professional service firms) leads to an efficient allocation of workers and creates adverse selection that enables workers to pay for general training. It also implies that wages of released workers can be higher than wages of those who are retained.

Human Capital Development Practices and Career Success: The moderating role of country development and income inequality

Article

Jan 2021

Employees can enhance their human capital through participation in organizationally‐sponsored development activities. However, there is little research on the extent to which the effects of such practices vary depending on national context. Adopting a human capital theory perspective, we hypothesized a positive relationship between human capital development practices experienced in one’s career and objective and subjective career success (salary level and perceived financial success respectively), and tested two country‐level institutional factors (country development and income inequality) as moderators. Results from our multilevel analyses of a large‐scale sample of over 8,800 managers and professionals from 28 countries showed that, as expected, experiencing a larger number of different human capital development practices was associated with higher salary level and greater perceived financial success. The relationship between development practices and salary level was stronger in the case of developed countries and weakly negatively affected for countries with higher income inequality. The relationship between development practices and perceived financial success was weaker for developed countries and unaffected by income inequality. Our research thus identifies boundary conditions to the application of human capital theory in different contexts and contributes to the comparative careers literature by showing that institutional factors affect the outcome of organizationally‐sponsored development activities.

In Search of Inspiration: External Hiring, Internal Mobility, and Creative Production

Article

Jan 2019

The Validity and Utility of Selection Methods in Personnel Psychology: Practical and Theoretical Implications of 85 Years of Research Findings

Article

Full-text available

Sep 1998

This article summarizes the practical and theoretical implications of 85 years of research in personnel selection. On the basis of meta-analytic findings, this article presents the validity of 19 selection procedures for predicting job performance and training performance and the validity of paired combinations of general mental ability (GMA) and the 18 other selection procedures. Overall, the 3 combinations with the highest multivariate validity and utility for job performance were GMA plus a work sample test (mean validity of .63), GMA plus an integrity test (mean validity of .65), and GMA plus a structured interview (mean validity of .63). A further advantage of the latter 2 combinations is that they can be used for both entry level selection and selection of experienced employees. The practical utility implications of these summary findings are substantial. The implications of these research findings for the development of theories of job performance are discussed.

The Boundaryless Career: A New Employment Principle for a New Organizational Era

Article

Full-text available

Jan 1998