1

INTERNAL MARKETS, PERSONAL NETWORKS, AND HUMAN RESOURCE

ALLOCATION: HOW MANAGERS CREATE VALUE THROUGH INTERNAL HIRING

JR Keller
Wharton School of Business
Working Paper

ABSTRACT

Internal hiring matches current employees to new jobs within an organization and represents a

critical yet overlooked source of value creation, enabling managers to generate greater value from

their existing stock of human resources by creating complementary matches between people and

jobs. Yet despite the fact that more than half of all jobs are filled internally, our knowledge of how

workers are allocated to jobs within organizations remains grounded in work on bureaucratic

internal labor markets and intraorganizational careers describing internal labor markets that bear

little resemblance to their contemporary counterparts. In this paper, I describe the effects of two

processes that have emerged to replace the use of bureaucratic rules for facilitating internal

mobility – market-oriented posting and relationship-oriented sponsorship – on two sets of

outcomes which link directly to value creation and value capture – quality of hire and

compensation. Using data on over 11,000 internal hires made over a five year period within a large

US health insurance company, I find that market-oriented posting results in better hires but at a

higher cost. In addition to providing a more complete picture of hiring and mobility, this work

sheds light on the tradeoffs associated with the use of markets and networks for allocating

resources within firms.

 2

Amid conversations about boundaryless careers, free-agent workers and hired guns, it is

easy to forget that a substantial amount of mobility still takes place within organizations. In fact,

nearly half of all open jobs in large organizations - and substantially more at the executive levels -

are filled internally (e.g. Crispin & Mehler, 2013). Internal hiring is the primary process used to

allocate human resources within firms and creates opportunities for internal mobility, which serves

many useful functions. It facilitates the transfer of knowledge across internal boundaries (Argote

& Ingram, 2000), motivates employees by signaling opportunities for future advancement

(Bidwell & Keller, 2014), encourages the development of firm-specific skills (Campion,

Cheraskin, & Stevens, 1994), increases worker satisfaction, performance and productivity

(Jackson, 2013), and decreases dysfunctional turnover (Allen, Bryant, & Vardaman, 2010). With

human resources now representing the most important resource in most firms (Powell &

Snellman, 2004) and in light of recent research demonstrating the high costs and even higher

failure rates associated with external hires at all organizational levels (e.g. Bidwell, 2011;

Groysberg, Lee, & Nanda, 2008), the ability to find and create complementary matches between

people and jobs within the firm represents a key source of value creation in modern organizations

(Zenger, Felin, & Bigelow, 2011).

While scholars have explored how other key resources, such as financial capital (e.g. Stein,

1997) and managerial attention (e.g. Ocasio, 1997), are allocated within firms, we know

surprisingly little about the contemporary internal allocation of human resources. Recent work has

documented how job characteristics shape both whether a job is likely to be filled internally

(Bidwell & Keller, 2014) and the qualifications (e.g. experience) of the candidates likely to be

placed into the job (Drazin & Rao, 2002). Yet internal hiring processes – that is, the ways in which

managers search for, evaluate, and select among potential internal candidates – have received little

systematic attention. However, research has shown that the use of formal versus informal external
 3

hiring processes (e.g. the use of referrals versus job postings) shapes not only who is hired, but

also their pay, performance, and turnover (Burks, Cowgill, Hoffman, & Housman, 2013;

Fernandez, Castilla, & Moore, 2000; Seidel, Polzer, & Stewart, 2000), as has recent research

comparing internal versus external hiring (Agrawal, Knoeber, & Tsoulouhas, 2006; Bidwell,

2011). We therefore might expect any variations among internal hiring processes to have similarly

significant consequences for both workers and firms.

Studying internal hiring also promises to contribute important insights to ongoing

conversations about the changing nature of internal resource allocation more generally. As firm

have transitioned away from hierarchical structures characterized by centralized decision-making

and towards flatter, leaner structures characterized by decentralized decision-making, bureaucratic

internal labor markets have gradually disintegrated (Cappelli & Keller, 2014), internal markets and

social networks have emerged as the primary mechanisms through with current workers are

matched to new jobs, echoing broader changes in how resources are allocated within

organizations. Indeed, the failures of bureaucratic planning systems in contemporary firms have

received considerable attention (Cowen & Parker, 1997; McEvily, Soda, & Tortoriello, 2014;

Mintzberg, 1994), with recent work identifying internal markets and network forms of

coordination as the two primary substitutes for bureaucratic control over the allocation of key

resources. Work on internal markets has generally emphasized how the use of market pricing

within the firm may lead to improved managerial decision-making regarding resource allocation,

as all of the information regarding the resource being considered is reflected in its price (Ellig,

2001; Felin & Zenger, 2011). Other work on internal markets has explored how high-powered

incentives can be designed to reduce coordination costs by aligning the interests of managers

making allocative decisions with the interests of the firm (Zenger & Hesterly, 1997; Zenger,

2002). Work on network forms of coordination has emphasized how the social relationships
 4

among actors within the firm can lead to improved allocative decisions by providing managers

with opportunities to share information on resources that would be otherwise unavailable to

centralized, higher-level decision-makers (Nahapiet & Ghoshal, 1998; Tsai & Ghoshal, 1998).

Because internal markets and social networks have been presented as alternatives to bureaucratic

control, much of the current literature has either compared internal markets with bureaucratic

control (e.g. Stein, 1997) or compared social networks with bureaucratic control (e.g. Tsai, 2002).

Much less work has explored the tradeoffs associated with using internal markets or personal

networks to allocate internal resources, despite the fact that market mechanisms and social

networks are likely to operate simultaneously within firms.

Contemporary internal labor markets represent a particularly fruitful context for exploring

these tradeoffs. Much of our understanding of internal hiring and mobility is grounded in the

foundational research on traditional, hierarchal internal labor markets and a closely related

literature on intraorganizational careers, which drew sharp distinctions between the bureaucratic

processes for allocating human capital operating within the firm and the market processes

operating outside the firm. This work described how internal hiring decisions were centralized in

personnel offices and governed by strict bureaucratic rules used to maintain lines of progression

along clearly defined job ladders, with employees exerting little control over their careers within

the firm (Althauser & Kalleberg, 1981; Glaser, 1968; Diprete, 1987; Doeringer & Piore, 1971;

Rosenbaum, 1990). Contemporary internal labor markets look dramatically different: hiring

decisions having been largely decentralized, with authority over promotions, transfers and external

hiring delegated to individual managers; organizational delayering, broader job descriptions, and

the rise of project-based work have all but eliminated clear paths for advancement; and employees

have been tasked with taking control of their careers. The cumulative effect of these changes has

rendered the use of bureaucratic rules for allocating human resources obsolete (Piore, 2002).
 5

Absent bureaucratic rules, the allocation of workers to jobs now takes place through two

very different processes (Marsden & Gorman, 1999; Pinfield, 1995). Posting is a market-oriented

process. A manager creates an internal market for an open job by broadcasting information about

the position throughout the organization via an internal job board and inviting current employees

to apply. The use of internal job posting systems is widespread, with 95% of organizations posting

jobs internally (Taleo Research, 2005). Sponsorship is a more relational process. A manager

identifies a candidate through her personal network and appoints that candidate to the job without

others being formally considered. Not only are posting and sponsorship the two most commonly

used internal hiring processes, they often operate concurrently within firms, as managers have

been granted substantial discretion both over the hiring decision and the hiring process (Marsden

& Gorman, 1999; Pinfield, 1995). With little if any work examining internal hiring in

contemporary organizations, it is unclear whether a market-oriented or relationship-oriented

allocative process is more likely to generate more valuable internal matches and what tradeoffs, if

any, might be associated with the use of posting versus sponsorship.

In this paper, I shed light on these tradeoffs by developing theory to explain how key

differences between the market-oriented posting process and relationship-oriented sponsorship

process affect two outcomes with implications for value creation and value capture: quality of hire

– as revealed by job performance, turnover, and subsequent advancement – and compensation.

Though the informational benefits associated with social networks have received considerable

attention (Borgatti & Cross, 2003; Bradach & Eccles, 1989; Cross, Parker, Prusak, & Borgatti,

2001; Granovetter, 1985), I predict that the market-oriented posting process will improve

managerial decision-making relative to the more relational sponsorship process, creating more

value though higher quality internal hires. In doing so, I describe how posting brings two features

of markets into the firm that are largely absent in sponsorship – self-selection and formality. The
 6

behavioral theory of the firm highlights two key challenges facing boundedly rational decision

makers, identifying a set of alternatives and evaluating the consequences of those alternatives

(Cyert & March, 1963; Simon, 1955). I argue that in contrast to the active managerial search

required in sponsorship, allowing workers to self-select into the consideration set for an open job

reduces the likelihood an exceptional internal candidate will be overlooked. Moreover, when

compared to the relative informality of a relationship-oriented allocative process, the formality of

the market will encourage managers to seek out information that allows them to better evaluate the

fit between a candidate’s qualifications and the requirements of the job, while also limiting the use

of irrelevant information that may lead to poor hiring decisions.

However, I also predict that these same market features will lead to higher salaries though

their effect on salary negotiations, limiting the amount of value a firm is able to capture through

better internal hiring decisions. Self-selection and formality introduce open competition into the

internal hiring process. While competition often drives down prices in external markets, I argue

that it will serve to increase prices for human capital within the firm, as candidates who are hired

though a competitive process are more likely to both initiate and adopt a competitive approach to

salary negotiations.

Taken together, these predictions suggest that there are important tradeoffs associated with

allocating human capital formally though markets or informally through the use of a manager’s

social network, with posting resulting in better hires but at a higher cost. However, with the

performance and retention benefits associated with better internal hires likely to far exceed the

higher salary costs, posting is likely to allow firms to both create and capture substantially more

value. These arguments suggest that a market-oriented process will also benefit workers, who

similarly capture more value through higher salaries.

 7

I use multiple modeling strategies in testing these predictions, which are largely supported

using five years (2008-2012) of personnel records covering all employees of a large health

insurance provider, as well as data on more than 350,000 internal and external job applications.

While personnel records have been previously used to identify which employees move to which

jobs (Bidwell, 2011; Dencker, 2008), data on the processes by which workers move jobs is rare.

These data are distinctive in that they clearly identify the mechanism used to facilitate each of the

11,000+ internal hires made during this period, allowing me to conduct what is, to my knowledge,

the first detailed study of the differences in outcomes associated with these two very different

internal hiring processes.

In unpacking the processes used to allocate workers to jobs within contemporary

organizations, this study helps to provide a more complete understanding of labor markets and

mobility. While robust literatures are developing to explore outcomes associated with different

external hiring processes (e.g. Fernandez et al., 2000) as well as the cost and quality tradeoffs

associated with internal versus external hiring (e.g. Bidwell, 2011), this work has tended to treat

internal hiring as a homogeneous process, unintentionally obscuring potential variations in the

processes used to move workers to new jobs within firms. More broadly, this study contributes to

a growing body of literature exploring how resources are bundled and deployed within

contemporary organizations (Sirmon, Gove, & Hitt, 2008; Sirmon & Hitt, 2009) by highlighting

the tradeoffs associated with using markets versus network approaches for allocating internal

resources. This study extends recent work exploring the effects of introducing market mechanisms

into firms (Felin & Zenger, 2011) by showing how mechanisms other than prices and high-

powered incentives can be leveraged to improve managerial decision-making, while also adding to

a small but important collection of studies which highlight the potential limitations of relational

exchange (Casciaro & Lobo, 2008; Rogan & Sorenson, 2014; Sorenson & Waguespack, 2006).
 8

Moreover, in identifying the micro-level mechanisms (decision-making and negotiations) through

which these two distinct organizational processes shape outcomes which link directly to value

creation and value capture (quality of hire and compensation), this study contributes to a

burgeoning literature focused on identifying the micro-foundations of human-resource based

competitive advantage (Coff & Kryscynski, 2011; Ployhart & Hale, 2014).

THEORY & HYPOTHESES

Internal Hiring in Contemporary Organizations

Before developing theory, it is useful to both define internal hiring and describe key

features of the allocative processes at the center of this study – posting and sponsorship. Once the

province of centralized personnel offices, managers are now largely responsible for staffing the

jobs that fall under their direct supervision, having been granted substantial discretion over which

candidates are considered and who is ultimately hired (Jacoby, 2004). Internal hiring occurs when

a manager fills an open job by hiring a worker currently employed by the organization in a

different job, resulting in the reallocation of that worker to a new set of organizational activities. It

is important to note that internal hiring is conceptually distinct from corporate restructurings, in

which large groups of workers – often entire departments or lines of business – are redeployed, en

masse, to new products or markets (Capron, Dussauge, & Mitchell, 1998; London, 1996).

Internal job posting systems date back to at least the 1940s. They were originally instituted

at the insistence of unions, who saw them as a way to curtail managerial discretion over staffing

decisions and thereby limit the scope for discrimination, favoritism, and nepotism to affect

opportunities for internal advancement. Management saw these systems as requiring them to

relinquish a long held right, so they fought to limit their scope. As a result, early job posting

systems typically covered a limited set of jobs and placed significant restrictions on which

employees were allowed to apply for those jobs (Cappelli, 2008). For those jobs that were
 9

covered, detailed selection criteria were established and often privileged seniority over ability,

providing individual managers with little choice over whom to hire (Jacoby, 1985; Slichter, Healy,

& Livernash, 1960).

Contemporary job posting systems are much more encompassing. Managers post

information about open jobs located at all but the very top levels of the organization to an internal

job board and invite interested candidates located throughout the organization to apply and

compete for the job, mirroring the typical process used to hire from the outside1. While the

requirements of the job are likely to influence who applies, candidates who do not possess the

qualifications described in the job posting but nevertheless would like to be considered are still

able to submit an application. The hiring manager evaluates the applications and extends an offer

to their preferred candidate, who may or may not negotiate the terms before deciding whether to

accept or reject the job. Once an offer is accepted, it is typical for unsuccessful internal candidates

to be notified and given the opportunity to learn why they did not get the job (Miller, 1984;

Pinfield, 1995).

Posting is often the default internal hiring process. Upwards of 95% of firms now use

internal job boards (Taleo Research, 2005) and many have developed policies encouraging

managers to post open positions internally, in large part to protect themselves from allegations of

discrimination (Strum, 2001). Yet despite these policies and the widespread availability of internal

job boards, managers can and often do bypass the posting system. This is in part due to the fact

that no state or federal laws require firms to post jobs and in part because internal staffing policies
1
When posting a job, a manager has the option of restricting the competition to internal candidates or
opening the competition to both internal and external candidates. While it is hard to pin down an exact
figure, a reasonable estimate is that around half of jobs above entry-level (which are by definition filled
externally) are posted internally and externally at the same time (Jacoby, 2005). This means that the posting
process may or may not result in an internal hire, depending on who is selected. For the purposes of this
paper, I present posting as an internal allocative process, because my interest is in looking at outcomes
associated with posting when an internal hire is made. In supplemental empirical analyses, I control for the
presence of external candidates on the key dependent variables.
 10

are both rarely enforced and allow for flexible interpretations (Pinfield, 1995). In fact, the Society

of Human Resource Management, the leading professional association for human resource

professionals, has recommended to hiring managers that “even if you have an internal process for

posting available jobs, there may be times when you decide not to follow this process” (SHRM,

2000: 7). In particular, managers are likely to bypass the posting system when they already have a

qualified candidate in mind for the job.

When managers do bypass the posting process, they turn a relational process I refer to as

sponsorship2. Social networks are a central feature of organizational life, as interactions among

individuals inevitably lead workers to develop networks of personal relationships (Kanter, 1977;

McEvily et al., 2014; Podolny & Baron, 1997). A sponsoring manager uses her personal social

network to identify potential candidates and then appoints her preferred candidate to the job absent

open competition. Though it is possible for sponsoring managers to exhaust their personal

connections as they search for potential candidates, the vast majority of internal hires made

through sponsorship involve the consideration of a single candidate with a direct connection to the

hiring manager, typically a current or previous colleague (Pinfield, 1995).

An abundance of evidence suggests that posting and sponsorship operate side-by-side

within firms, as equally viable ways to fill jobs. Marsden and Gorman (1999) examined survey

data on a representative sample of US work establishments and found that posting and sponsorship

were widely used in combination for filling vacancies with internal candidates. Pinfield’s (1995)

study of internal hiring in a forest-products company revealed that more than half of all positions

were filled through sponsorship despite company policy that all jobs be posted. And though over

95 percent of firms report posting at least some open jobs internally (Taleo Research, 2005)

2
I adopt the term sponsorship from classic literatures on career mobility within organizations (Rosenbaum,
1979) and upward social mobility more generally (Turner, 1960), in which it is used to describe systems in
which individuals selected for advancement are shielded from competition.
 11

research continues to demonstrate the importance of relationships on internal advancement

(Bielby, 2000; Ibarra, 1993; Podolny & Baron, 1997).

It is important to note that while posting and sponsorship are conceptually distinct internal

allocative processes, sponsorship may sometimes masquerade as posting in practice. That is, it is

possible that a manager may post a job having already identified the candidate they are going to

select through their social network, a practice referred to as a “wired search” (Bielby, 2000).

However, the prevalence of “wired searches” and other practices that might artificially restrict the

openness of the posting process, such as shaping job requirements around a single candidate or

discouraging employees from applying for certain jobs, are largely mitigated by concerns about

allegations of discrimination (Strum, 2001) as well as the possibility that employees who feel they

were mistreated or misled will simply leave the organization (Billsberry, 2007; Cappelli, 2008).

From the descriptions above, it is clear that market-oriented posting and relationship-

oriented sponsorship differ in several ways. In developing theory to understand the tradeoffs

between these two allocative process, I highlight the fact that posting is characterized by two

market-like features, self-selection and formality, that are largely absent in sponsorship, which

instead involves active managerial search and using personnel connections to gather information.

Bounded Rationality and Quality of Hire

A key facet of behavioral theories of the firm is the presumption that decision-makers are

boundedly rational. Because decision-makers are cognitively limited and have limited time,

information, and resources at their disposal, neither the complete set of alternatives from which a

decision maker can choose is known ex ante, nor are the consequences involved in choosing

among the available alternatives (March & Simon, 1958). As a result, failures in generating and

evaluating alternatives have been identified as two of the chief reasons why managers fail to

optimally allocate available resources (Afuah & Tucci, 2012; Knudsen & Levinthal, 2007).
 12

Self-selection and generating alternatives. Because the complete set of alternatives is not

known ex ante, boundedly rational decision-makers must engage in search to generate alternatives.

Search not costless, however, and one of the ways decision-makers economize is by considering

only a small portion of available alternatives (Hauser & Wernerfelt, 1990; Shocker, Ben-Akiva,

Boccara, & Nedungadi, 1991). While considering more alternatives does not guarantee a better

decision (Iyengar & Kamenica, 2010), decision-making success does tend to improve when more

alternatives are considered (Alexander, 1979; Gemünden & Hauschildt, 1985; Nutt, 1998), in part

because it lessens the odds that a superior alternative will be left out of the consideration set.

Self-selection enables managers to generate more alternatives without incurring many of

the costs associated with a broader search (Afuah & Tucci, 2012). Markets facilitate resource

allocation by providing a common platform for widely dispersed buyers and sellers to exchange

information about their needs and preferences (Zenger et al., 2011). Self-selection refers to the

ability of market participants to choose which available opportunities to pursue based on this

information rather than have those matches determined by managerial authority.

One way managers (as buyers) are able to harness the power of self- selection is by

broadcasting information about an opportunity and allowing interested sellers to self-select into

the set of alternatives to be considered by the manager. Rather than the manager assuming the

responsibility for generating alternatives through active search, sellers search for opportunities that

match their preferences. Because sellers have more information on their preferences than

managers and are likely to seek out alternatives that meet those preferences, searches that may be

considered distant from the perspective of the manager may often be considered local from the

perspective of the seller. By essentially transforming local search into distant search (Afuah &

Tucci, 2012), self-selection should be expected to expand the number of alternatives considered by

a manager – and more alternatives reduces the risk that a quality alternative will be overlooked.
 13

The posting process enables managers to harness the power of self-selection by

broadcasting information about an open job and allowing internal candidates located throughout

the organization, including those in in more distant areas, such as workers located in a different

location, department or function, to self-select into a queue of candidates competing for the job. In

contrast, sponsorship provides no formal mechanism for employees to express their interest in an

open job. Rather than broadcast information about the opening and allow interested candidates to

self-select into the candidate pool, the manager actively searches for alternatives (Mintzberg,

Raisinghani, & Théorêt, 1976) through their personal network. The pool of potential candidates is

therefore restricted by a manager’s previous experience and connections. While information about

internal candidates residing outside a manager’s immediate network is likely accessible (e.g.

through human resource information systems, talking with HR, etc.), obtaining this information

takes time and effort and the likelihood of finding a superior alternative is uncertain. Moreover,

managers are more likely to place a higher value on information obtained through their social

network, further discouraging the search for candidates residing outside of it. As a result,

sponsoring managers typically only consider those candidates with whom they are already

familiar, and even the most well-connected managers in large organizations are unlikely to be

familiar will all potential internal candidates.

In sum, by enabling candidates located both within and outside a hiring manager’s social

network to self-select into the consideration set, posting should therefore be expected to generate a

larger set of candidates than would otherwise be accessible through a manager’s personal network,

reducing the likelihood that an exceptional internal candidate will be overlooked.

Formality and evaluating alternatives. In addition to the challenges associated with

generating alternatives, boundedly rational managers also face difficulties evaluating alternatives.

In particular, recent work has called attention to the problems associated with bounded awareness,
 14

which refers to the propensity of boundedly rational individuals to fail to seek out or incorporate

relevant and accessible information into their decision-making process, instead relying on less

relevant information. Bounded awareness is one symptom of intuitive thinking, which often fails

to allow for the possibility that evidence needed to make a good decision is missing. As a result,

decision-makers tend to make decisions based on a subset of available information (Brenner,

Koehler, & Tversky, 1996; Kahneman, 2011). This “misalignment between the information

needed for a good decision and information included in the decision-making process” (Bazerman

& Chugh, 2005: 10) can lead to costly errors. Such errors are problematic in the hiring context,

where managers are notorious for their “stubborn reliance on intuition and subjectivity”

(Highhouse, 2008: 333).

Research suggests that interpersonal networks serve as conduits for information exchange

within organizations, providing managers with ready access to information that is richer, more

complete, and perceived as more trustworthy than information obtained from other sources

(Bradach & Eccles, 1989; Granovetter, 1985; Podolny & Baron, 1997). While this information

might be expected to improve decision-making, access to information alone is not enough to

ensure a good decision; managers must still select which information to use and which

information to ignore. In fact, studies spanning different contexts and levels of analysis, from

those examining individual hiring decisions3 to firm-level decisions about selecting merger and

3
There is ample evidence that managers rely on irrelevant information when making hiring decisions. For
example, managers often give substantial weight to performance in a previous job despite the fact that is
often a poor predictor of future performance (Cascio & Aguinis, 2008) and frequently allow attributes such
as gender, race, attractiveness, and weight to influence hiring decisions (e.g. Agerström & Rooth, 2011;
Bertrand & Mullainathan, 2004; Marlowe, Schneider, & Nelson, 1996). There is also an equal amount of
evidence that hiring managers fail to seek out relevant information, even when it is easily accessible. For
example, managers routinely bypass proven selection aids in favor of unstructured interviews (Highhouse,
2008) and use interviews to confirm their first impressions of candidates at the expense of gathering job-
relevant information (Dougherty, Turban, & Callender, 1994).
 15

acquisition partners (Rogan & Sorenson, 2014), show that managers routinely struggle to

objectively evaluate alternatives with whom they have an existing connection.

A market-oriented allocative process may therefore actually be superior to an information-

rich relational process in helping managers to overcome problems associated with bounded

awareness by shaping the information used to evaluate alternatives. Markets are institutions

supported by a system of rules and conventions designed to facilitate exchange among buyers and

sellers (Casson, 1982; Menard, 1995; Polyani, 1957). These rules and conventions – which I refer

to as formality – are both impersonal and non-arbitrary, thereby providing a stable framework for

transactions to take place (Menard, 1995: 1967). These rules and conventions serve several

functions, two of which are establishing the terms of exchange and defining what constitutes

legitimate behaviors (Loasby, 2000). In establishing the terms of exchange, the market generates

at set of evaluation criteria; in defining what constitutes legitimate behaviors, the market imposes

accountability on buyers. The presence of formal evaluation criteria and a high level of

accountability discipline managerial decision-making in different ways.

Evaluation criteria and relevant information. For markets to function, there must be a

mechanism for managers (as buyers) to broadcast information to potential sellers about the good

or service they are looking to procure (Zenger et al., 2011). Though managers may have

difficulties fully articulating their needs (Nickerson & Zenger, 2004), this initial information helps

to establish the terms of exchange in that it serves as the initial criteria against which sellers

evaluate their interest in pursuing the opportunity and as the initial criteria against which the buyer

evaluate those sellers who self-select into the consideration set. As a result, sellers are likely to

provide, and buyers are likely to seek out, information that enables buyers to evaluate alternatives

against a set of established criteria.

 16

One way posting introduces the formality of markets is through the use of formal job

descriptions. The posting process requires a manager to create a formal job description, necessary

for broadcasting information about an open job to potential candidates. Though it can be difficult

to develop accurate, comprehensive job descriptions (Backhaus, 2004; Sanchez & Levine, 2012),

the requirements defined at this initial stage nevertheless serve as a set of formal criteria against

which potential candidates are evaluated. In contrast, sponsorship does not require the manager to

create a formal job description prior to evaluating candidates. A manager must possess a

reasonable understanding of the job requirements as well as desirable candidate attributes, but

because the manager does not have to broadcast information about the open job, a formal job

description is not necessary. This allows the manager to informally mold the job requirements

around their preferred candidate rather than evaluating the candidate against the requirements of

the job (Miner, 1987; Sanchez & Levine, 2012). When compared to sponsorship, the presence of

formal evaluation criteria in posting is therefore more likely to prompt managers to reconsider the

relevance of the information used to evaluate candidates, with managers more likely to recognize

and seek out information enabling them to evaluate a candidate’s ability to perform well in the job.

Accountability and irrelevant information. Mechanisms that impose responsibility and

accountability ensure the continued participation of market participants by instilling confidence

that future transactions will be completed in a fair, honest, and orderly manner. Of particular note,

perceptions of the process by which firms make allocative decisions in a market can effect

perceptions of fairness, with decisions that appear to be free of bias and based on objective criteria

perceived as more fair and legitimate (Bies, Tripp, & Neale, 1993; Williams, 1987). Market-based

accountability should therefore encourage managers to avoid using information that would lead

their decisions to be perceived as subjective.

 17

Posting embeds a market-like accountability in the hiring process through the custom of

requiring managers to explain to unsuccessful internal candidates why they were not selected. For

every successful internal candidate there are likely to be multiple unsuccessful candidates.

Because these unsuccessful internal candidates remain employees, it is important to clearly

communicate the reasons why they did not get the job in order to minimize any sense of unfairness

that may decrease motivation, performance, and potentially lead to dysfunctional turnover. By

ensuring that hiring decisions have to be communicated and defended, the custom of explaining to

employees why they were not selected embeds accountability into the internal hiring process

(Tetlock, 1992).

Sponsoring managers informally search for candidates through their personal network, so

workers are often unaware they are being considered (Pinfield, 1995). As a result, accountability is

more limited than it is in posting, though it is not entirely absent. Managers are required to

communicate their decision to a supervisor but because supervisors typically grant managers

substantial discretion over who is selected, those decisions do not have to defended to a broader

audience. The higher level of accountability generated though the competitive posting process

should therefore guard against the use of irrelevant information, as managers are more likely to

use objective criteria in justifying their hiring decisions to a broader audience.

As a whole, these arguments suggest that infusing self-selection and formality into the

internal hiring process will help boundedly rational managers overcome problems associated with

generating and evaluating alternatives. Self-selection is likely to be more effective than active

managerial search in reducing the likelihood that an exceptional candidate will be overlooked,

while the formality of the market is likely to be more effective at disciplining managers to avoid

costly errors associated with bounded awareness. As a result, I expect posting to create more value

than sponsorship by generating higher quality internal hires, as revealed by worker performance,
 18

turnover and subsequent advancement. Specifically, I predict that when compared to sponsored

internal hires:

H1: Workers hired through posting will have higher performance ratings in the new job.

H2: Workers hired through posting will be less likely to exit the firm.

H3: Workers hired through posting will be more likely to subsequently be promoted.

Competition and Compensation

Quality of hire is only part of story, as the value a firm is able to capture from even the

highest quality hire is largely contingent on how much they are paid. It is therefore important to

understand how posting and sponsorship shape compensation. To do so, it is useful to switch from

the manager’s perspective and think about how salary negotiations from the perspective of the

worker.

Markets are characterized by open competition, with sellers aware that they are competing

for buyers with other sellers (Menard, 1995). Self-selection and formality are two of the key

mechanisms supporting the competitive nature of markets. Self-selection allows sellers to pursue

the opportunities they are interested in and the formality facilitates the exchange of information

that ultimately allows buyers to compare information on widely dispersed alternatives and make a

selection (Zenger et al., 2011).

Posting is characterized by open competition. Interested candidates self-select into the

consideration set when they apply for an open job. In doing so, they form a labor queue – a set of

workers competing for a specific job at a specific time (Reskin & Roos, 1990). The formality of

the posting process underscores the competitive nature of posting, as the fact that employees have

to actively submit an application makes them aware that they are entering into a competition they

are not assured of winning. In contrast, to the extent there is competition in sponsorship, it lacks

structure and transparency. Because search is costly, managers routinely consider a small pool of
 19

internal candidates (often just a single candidate). Moreover, because managers gather information

on potential candidates informally, in those cases where multiple candidates are considered, those

who are not selected are often unaware of being considered (Pinfield, 1995).

While competition is often seen as a way to lower prices by pitting multiple suppliers

against one another, recent work exploring the social psychological aspects of negotiation

provides reason to expect the opposite in the internal hiring context. A key premise in this

literature is that situational factors prime individuals to place more or less emphasis on the

importance of dyadic relationships in negotiations. The more emphasis an individual places on the

dyadic relationship in a negotiation context, the more likely they are to adopt a relational

orientation to negotiation, and individuals adopting a relational orientation to negotiation are more

likely to forgo economic gain in an effort to develop relational capital (Curhan, Elfenbein, & Xu,

2006; Curhan, Neale, Ross, & Rosencranz-Engelmann, 2008: 193; Gelfand et al., 2006).

Sponsorship is much more likely to cue a worker to focus on their relationship with the

hiring manager. Because the hiring manager personally appointed the worker to the job absent any

formal competition, the relationship with the hiring manager is likely to be particularly salient and

highly valued at the time an initial job offer is presented. This is unlikely to be case in posting,

with the competitive nature of the process emphasizing the transactional nature of the employment

relationship. As a result, workers hired through sponsorship are more likely to adopt a relational

orientation approach when the time comes to negotiate compensation.

There are two reasons to expect that adopting a relational orientation towards salary

negotiations has important implications for how much a worker is likely to earn. First, a worker

adopting a relational orientation is less likely to initiate a salary negotiation. Focused on

developing relational capital, she will want to avoid appearing self-interested and feel

uncomfortable with the idea of asking for more money (Gelfand et al., 2006). She is also more
 20

likely to perceive the initial offer as fair; being more attuned to other party’s goals should reduce

the likelihood she will assume the other party’s interests are opposed to their own, a common error

in negotiations (Cross, Bacon, & Morris, 2000; Gelfand et al., 2006; Thompson & Deharpport,

1998). Second, an individual adopting a relational orientation is more likely to employ an

accommodative negotiation strategy whereas an individual adopting a transactional orientation is

more likely to negotiate competitively (Curhan et al., 2008), and accommodative approaches to

salary negotiation have been shown to result in lower raises than competitive approaches (Marks

& Harold, 2011).

This suggests that there is an important tradeoff between the quality and costs associated

with different internal hiring processes. Though posting is likely to create more value, the firm

does not capture all of that value, as workers capture a portion of that value through higher

compensation. Relative to sponsorship, the competitive nature of the posting process will increase

the likelihood that workers both initiate salary negotiations and adopt a more economically

beneficial approach when they do choose to negotiate, leading me to predict that:

H4: Internal candidates hired through posting will receive higher starting salaries than

sponsored employees entering equivalent jobs.

DATA & METHODS

I test my hypotheses using monthly personnel records covering the years 2008 to 2012

from the US operations of Fortune 100 health insurance company, which I call HealthCo. The

primary data for my analyses consist of more than 11,000 internal hires made during the

observation period, which are identified from a larger dataset consisting of 1,914,519 monthly

observations covering 56,811 worker employed by HealthCo between 2008 and 2012.

While using data from a single firm limits the confidence with which I can generalize my

results, these data are particularly well suited to test my hypotheses. The distinguishing feature of
 21

these data is that the way in which HealthCo has linked their various human resource information

systems allows me to clearly identify whether posting or sponsorship was used to facilitate each

and every internal hire. I am also able to link these personnel records to a companion dataset with

information on more than 350,000 internal and external job applications submitted during 2012,

allowing me to conduct several robustness checks. Using personnel data from a single

organization has several other advantages, including the fact that my performance measures are

standardized across jobs and that I am able to control for the effects of job content and the location

of different jobs (and therefore different attributes of moves between jobs), all of which would

pose substantial empirical difficulties in a multi-firm study. Moreover, obtaining this type of data

from even a single firm is quite challenging; many firms fail to systematically record any data

related to employee mobility (ERC, 2010; Oracle, 2012) and for those that do, the fear of

sanctions were internal analyses to reveal previously unrecognized patterns of illegal

discrimination has the perverse effect of discouraging firms from exploring these processes

(Strum, 2001). The setting itself reduces at least some concerns about generalizability, as

HealthCo mirrors other large contemporary organizations in several respects: hiring decisions are

delegated to individual managers, employees are explicitly encouraged to actively manage their

careers amid a lack of well-defined advancement paths, and there are substantial amounts of

lateral and vertical mobility across broadly defined jobs.

Identifying Internal Hires

An internal hire occurs when a manager fills an open job with a current employee4, as

indicated by a change in an employee’s job code, department, or both from one month to the next.

An employee who changes job codes takes on a new set of tasks and responsibilities. A move to a

4
HealthCo has a vacancy-driven hiring process, meaning that all hires – internal and external – are
preceded by an open job.
 22

new department represents a move to different area of the business, as departments are organized

around products, geographic markets, and customers (internal and external). All entry-level jobs at

HealthCo are filled through external hiring; internal hiring is used alongside external hiring to fill

jobs above entry-level.

It is important to emphasize that a change in job code does not simply represent a change

in title, with little change to the work actually performed (Miner, 1987). Jobs at HealthCo are

broadly defined by hierarchical level, function, and role. In an average year, approximately 34,000

employees were distributed across 462 job codes. Jobs are organized into nine different levels [(1)

Entry-level, (2) Team Lead, (3) Supervisor/Analyst, (4) Manager/Professional, (5)

Director/Technical Leader, (6) Vice President, (7) Senior Vice President, (8) Executive Vice

President, (9) CEO] and thirty functional areas, including those common to most large firms (e.g.

Sales, Finance/Accounting, HR, and Marketing) as well as several more specific to the health

insurance industry (e.g. Clinical Guidance, Provider Contracting). Roles indicate the specific

competencies needed to perform the job. For example, “Creative Developer” is a Level 3 role in

Marketing; “Recruiting Lead” is a Level 3 role in Human Resources; and “Architect”,

“Applications Consultant” and “Project Manager” are all Level 3 roles with IT, each linked to

different competencies. As a result, a change in job code reflects a meaningful change in the work

a person does.

Dependent Variables

Quality of hire and compensation are the two primary outcomes of interest in this study.

Researchers have used a wide variety of post-hire outcomes to assess quality of hire (e.g, see

Breaugh & Starke, 2000; Zottoli & Wanous, 2000). In an effort to provide a holistic accounting of

the quality of internal hires with HealthCo, I test my hypotheses using multiple indicators of

quality: performance ratings, relative performance, turnover, and subsequent advancement.

 23

Performance ratings and relative performance. I use five measures of job performance

from HealthCo’s annual performance evaluation as a first set of quality indicators.

Contribution score. A worker’s contribution score assesses their contribution to the

success of the organization. In jobs with a less direct impact on organization-level outcomes (e.g.

those at lower levels), the contribution score is typically used to assess their contribution to the

department or line of business. It is measured on a 0 to 4 scale [0=not contributing (0%), 1=low

contribution (3.2%), 2=moderate contribution (19%), 3=full contribution (66.8%), and

4=exemplary contribution (11%)].

Competency score. A worker’s competency score assesses their skills relative to what is

required for the job. Each worker receives a separate score for each of the eight competencies

assigned to their job code5. Each competency is measured on a 1 to 4 scale [1=Learning,

2=Exhibiting, 3=Demonstrating, 4=Modeling]. I average these individual scores to compute an

overall competency score [1-2 (11.4%), 2-3 (66.2%), 3-4 (12.4%)].

Relative performance. Managers also rank workers in similar jobs as part of an annual

calibration process (described below). However, workers are not simply ordered according to

their contribution and competency scores; rather, this is intended to be a measure of overall value

to the organization that takes into account both previous performance and future potential.

Although there are no formally established guidelines dictating how finely managers should

distinguish among workers, these calibration sessions typically create “buckets” of employees; a

group of 100 employees may not be ranked from 1 (highest) to 100 (lowest), but rather the top five

employees may receive a 1, the next ten a 2, the next twenty-five a 3, and so on. I use these

ranking to create three dichotomous measures of relative performance: whether a worker is ranked

5
Selected from an overall library of 124 competencies customized for HealthCo.
 24

in top the quartile (top 25%), bottom half (bottom 50%), or bottom quartile (bottom 25%) of her

cohort in her first year in the new job.

While all of these are subjective measures of performance, several researchers have argued

that subjective ratings are among the most valid measures of performance despite concerns about

the potential for managerial bias to affect ratings (Cascio, 1998). Subjective ratings enable

managers to take into consideration a variety of behaviors and outputs relevant to the job (Medoff

& Abraham, 1981) as well as account for factors affecting performance outside the control of the

individual worker (Campbell, McCloy, Oppler, & Sager, 1993). However, it is also important to

note that HealthCo’s annual evaluation process helps to reduce potential concerns about

supervisory bias affecting individual performance ratings. Managers who supervise workers in

similar jobs meet in person to review and discuss their ratings of individual workers – a process

known as calibration. These calibration discussions are intended to ensure that managers are

evaluating workers against a common standard and to identify and correct instances where

managers may have rated employees too harshly or leniently. Research has shown that calibration

tends to reduce subjectivity and bias in performance ratings both because ratings are likely to be

more consistent across employees when managers share a common view on rating standards

(Mclntyre, Smith, & Hassett, 1984; Pulakos, 1984) and because “fellow managers do not do not

usually let each other off easily if they believe an employee has been rated unfairly, creating peer

pressure that provides a powerful incentive to make accurate ratings” (Pulakos & O’Leary, 2011:

152–3).

Turnover. Turnover is a second indicator of match quality (Jovanovic, 1979; Mortensen,

1988). I create two dichotomous measures of turnover indicating whether a worker exits the

organization with the first 12 months (turnover12) or 24 months (turnover24) of moving to a new

job. I do not distinguish between voluntary and involuntary turnover for two reasons.
 25

Conceptually, both are indicators of poor matches, with the worker and firm both more likely to

terminate an ill-fitting employment relationship. Empirically, while the data do indicate whether

turnover was voluntary or involuntary, managers and HR staff at HealthCo both told me not to

trust these indicators as reflecting the actual reasons an employee left the organization. Consistent

with research showing that turnover rates are relatively low among internal hires in non-entry-

level jobs (Bidwell, 2011), approximately 5% of internal hires exited the firm within 12 months,

while 13% exited within 24 months.

Promotion. Subsequent promotion is a third indicator of quality. Promotions are internal

hires that result in the employee moving into a higher level job. Because time to promotion varies

across jobs, I create two dichotomous measures of subsequent advancement: whether a worker

was promoted within 24 months (prom24) or 36 months (prom36) of moving into a new job.

Approximately 14% of internal hires were subsequently promoted within 24 months, while 31%

were promoted within 36 months.

Starting salary. The starting salary is the natural logarithm of the salary a worker receives

in the first month in a new job. Salary accounts for the vast majority of compensation for most

workers at HealthCo. Sales workers represent the main exception and their bonuses, which are tied

to clearly defined sales targets, can account for a substantial portion of their total income.

However, the bonus amount these workers are expected to receive based on their targets are

factored into the annual salary figures recorded in HealthCo’s personnel records. For example, if a

salesperson is hired into a job with a base salary of $80,000 and expects to earn a bonus of

$40,000, for a total expected annual compensation of $120,000, her salary in the monthly

personnel record from which I pull this figure will be $120,000. This means that the salary figure I

use represents the total compensation she should expect to earn during the year, which is the figure

upon which she will be negotiating.

 26

Independent Variable

The independent variable is a dichotomous indicator of whether posting (1) or sponsorship

(0) was used to fill the job, identified through an unambiguous indicator in the first monthly

observation of a worker in her new job. 3,841 (43%) of internal hires were made through posting

and 5,458 (57%) through sponsorship.

Control Variables

An important concern when using non-experimental data is the potential for omitted

variables to create spurious correlations between the independent and dependent variables. Hiring

managers at Health Co. are able to choose whether to fill a job through posting or sponsorship. A

particular concern in this study is that there may be variables that both affect a manager’s decision

of which process to use as well as the outcomes of interest. In the absence of existing empirical

evidence, it seems likely that the choice to use posting or sponsorship might be affected by two

factors: attributes of job and the manager’s perception of the availability of qualified candidates. It

could be that there are certain jobs that are always posted and others that are always filled through

sponsorship. It could also be the case that managers only post jobs if they have been unable to or

assume they will be unable to identify qualified candidates through their personal network. I

leverage the level of detail in my data to address the first issue, which enable me to control for

many of the job characteristics that might be expected to have an effect on the choice between

posting and sponsorship. To address the second issue, I run a series of robustness checks using an

instrumental variable approach, which I describe in more detail later in the paper.

Job attributes. In order to compare the outcomes associated with different processes used

to staff similar jobs, I control for several job-level attributes. I include dummy variables to control

for hierarchical rank, function area and the state a job is located in (51% of job are located in the

headquarters state). In order to control for fixed propensities of different jobs to be filled by
 27

posting or sponsorship, I also include separate dummies for each of 266 job codes filled though

internal hiring during my observation period. Importantly, the data reveal no systematic

differences in the types of jobs that are filled through posting or sponsorship. Of all the job codes

filled internally from 2008-2012, 84% were filled by both posting and sponsorship, and moves

into those jobs accounted for 99% of all internal hires. Moreover, 75% of the job codes filled

exclusively through either posting or sponsorship were only filled once or twice, 88% only three

times, and none more than six times. It is therefore likely that even those job codes filled

exclusively through one process during this time periods are nevertheless open to being filled

through both processes, with the apparent exclusivity an artifact of so few hires having been made

into those job codes.

Job transition attributes. Another concern is that the outcomes may be affected by the

types of moves workers are making rather than how those moves are made. For example, we

might expect workers moving to more similar jobs to have higher levels of performance and that

sponsored workers are likely to come from more similar jobs because managers are more likely to

have personal connections to workers in similar jobs. To account for this possibility, I include

dummies for whether an internal hire resulted in a worker moving into a new job in the same

function or same department. Similarly, I include dummies indicating whether an internal hire

resulted in a promotion (a vertical advancement; n=4,843), an expansion (a lateral advancement

which results in an expansion of the worker’s competencies; n=3,594), or a transfer (a lateral

move to the same job in a different department; n=1,122), which may affect pay and/or

performance. I include the worker’s salary in the final month of their previous job in some models

to account for the possibility that internal salary adjustments may be based on an employee’s

previous salary even when moving across very different jobs.

 28

Individual attributes. I also include controls for a variety of candidate attributes.

Demographic characteristics include gender, age and age-squared, and ethnicity. Tenure and

tenure-squared are calculated as the number of months (squared) a worker has been employed by

the firm. I include a worker’s contribution and competency scores in their previous job as a rough

indicator of pre-hire quality. HealthCo does not include years of education or highest degree

completed in their personnel records.

Additional controls. I include dummies for each year in the sample to reflect changing

labor market conditions. In models predicting relative performance, I control for the size of a

worker’s performance rating cohort (e.g. the number of workers in their forced ranking group).

Sample Restrictions

The samples used for each of my analyses vary according to a number of restrictions I

placed on the data. In the analyses where performance ratings are used as dependent variables, I

dropped observations with missing performance data, which occurred when workers had been

hired to recently to assess or exited the firm before being assessed (n=2,484). In both the turnover

and advancement models, the samples are restricted to those workers who were hired early enough

for me to calculate the dependent variables of interest (e.g. to be included in the model predicting

turnover in 12 months, a worker either had to have exited the firm within 12 months of moving

into their new job or occupied the job for 12 months). Finally, in the compensation models, I

excluded observations with missing salary data (n=28).

ANALYSIS & RESULTS

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

independent variables, with each observation representing an internal hire. Of particular interest

are the correlations between the multiple measures of quality of hire. The correlation between the

contribution and competency scores (r = .67) which indicate that these two measures pick up
 29

different aspects of performance. The correlations between these two measures and the relative

ranking variables covary in the expected directions while also suggesting that they are also picking

up different aspects of performance, as they are positively correlated with a worker ending up in

the Top 25% (r = .43, .45) and negatively correlated with a worker ending up in the Bottom 50%

(r = -.46, -.50) and Bottom 25% (r = -.40, -.46). The low correlations among the performance

rating variables and the turnover and subsequent promotion variables (none exceed r = +/- .13)

suggest that each of these variables are capturing a different element of performance. It is also

important to note the low correlations between salary and each of the performance measures (none

exceed r = +/- .16), which underscore the notion that salary does not simply reflect a manager’s

expectations about performance in the new job, but is instead largely determined by a negotiation

process which occurs prior to a manager observing actual performance (Jovanovic, 1979). This is

important in interpreting the results, as I argue that while posting results in both higher quality

hires and higher salaries, the higher salaries are not a result of the manager expecting a higher

level of performance. In fact, the robustness checks suggest that, if anything, managers are likely

to expect sponsored hires to perform better.

------------------------------------
Insert Table 1 about here
------------------------------------
Performance

Table 2 presents analyses of each of the five performance measures. I use ordinary least

squares regressions in models using competency and contribution scores as the dependent

variables. Competency score is a continuous variable and though contribution score is a discrete,

ordinal measure, the OLS model is easier to interpret than an ordered logit model (and both

provide similar results). I use a logit specification for models where relative performance is

outcome of interest, as the dependent variables are binary. The unit of analysis in all models is an
 30

internal hire and the performance measures reflect performance in the first year on the job. I

cluster the errors by individuals to account for non-independence among the errors.

Using contribution and competency scores as the dependent variables, Models 1 and 2

provide support for Hypothesis 1, with the significant positive coefficient for posting in both

models indicate that internal hires made through posting outperform internal hires made through

sponsorship. The effect sizes are relatively small, with posting resulting in an increase of

approximately one-tenth of a standard deviation on each measure of performance, an issue I return

to in the robustness checks. Model 3 provides additional support for Hypothesis 1 in

demonstrating that internal hires made through posting are approximately 13% more likely to be

rated in the top quartile of their respective performance/potential distribution than sponsored

internal hires. Models 4 and 5 further reveal that internal hires made through posting are less likely

to have poor performance ratings in the new job. The significant negative coefficients for posting

indicate that internal hires made through posting are approximately 13% less likely to fall in the

bottom half of the ratings distribution and approximately 15% less likely to fall in the bottom

quartile. Taken together, these results strongly suggest that posting leads to better internal hiring

decisions, in part by helping managers avoiding hiring mistakes.

------------------------------------
Insert Table 2 about here
------------------------------------
Turnover

Models 7 and 8 in Table 3 use logit models to test Hypothesis 2, that internal hires made

through posting are less likely to exit the firm. Both models provide support for this hypothesis,

with the significant negative coefficients for posting indicating that internal hires made through

posting are around 20% less likely to exit the firm within 12 months (Model 7) and around 18%

less likely to exit the firm within 24 months (Model 8).

 31

------------------------------------
Insert Table 3 about here
------------------------------------
Subsequent Advancement

Models 9 and 10 in Table 3 also use logit models test Hypothesis 3, that internal hires

made through posting are more likely to be subsequently promoted. I find very limited support for

this argument. The coefficient for posting in not significantly different from zero in Model 9,

indicating that there is no difference the 24 month promotion rates of internal hires made through

posting or sponsorship. The coefficient for posting is positive but only marginally significant in

Model 10, provided some indication that posting may be more likely lead promotion over a 36

month period, but nothing conclusive.

Salary

Models 5 and 6 in Table 3 test Hypotheses 4, that internal hires made through posting will

receive higher starting salaries than sponsored hires entering equivalent jobs. The significant

positive coefficient for posting in Model 11 provides support for this hypothesis, revealing that

posted hires receives nearly 4% higher salaries, on average, than sponsored hires. The 4% posting

premium remains after controlling for performance in the previous job (Model 12).

Mobility Patterns and Within Department Hires

In theorizing about the effects of self-selection and formality on quality of hire, I made two

complementary arguments, one that suggests that self-selection improves quality by expanding the

pool of potential candidates and the other that formality helps to shape the information managers

use when evaluating candidates. The regression results above, however, do not allow me to

whether one or both of these mechanisms are driving the results. To examine this issue in more

detail, I first present descriptive statistics on internal source of hire (Table 4). These data reveal

that posting is significantly more likely to result in hires from different departments, functions,
 32

cities and even different buildings, consistent with my argument that self-selection lessens the

likelihood that manager will overlook an exception candidate by allowing managers to more

readily identify candidates widely dispersed throughout the organization. I then ran a series of

regressions restricting the analysis to internal hires made within departments (Table 5). Because

managers are likely to be aware of and have access to much more detailed information on

candidates located within their own department, this provides a stronger test of my argument that,

beyond providing more alternatives, posting improves decision making by disciplining what

information they use in evaluating candidates. Consistent with my theorizing, the results are nearly

identical to those presented in Tables 2 and 3. In fact, if we assume that managers not only have

ready access to, but already possess detailed information (both relevant and irrelevant to their

ability to do the job) on candidates within their department, we might conclude that markets are

particularly beneficial in helping managers to avoid the use of irrelevant information when

evaluating well-known alternatives, a commonly cited reason for hiring errors (Highhouse, 2008).

------------------------------------
Insert Tables 4 & 5 about here
------------------------------------
Supplementary Analyses and Robustness Checks

As noted above, a potential concern with my use of ordinary least squares and logit models

above is the potential endogeneity of posting. While I am able to control for the fixed propensities

for certain jobs to be filled by posting or sponsorship, I do not have a variable which would allow

me to control for the possibility that managers only post jobs if they are unable to first identify

qualified candidates through their personal network. However, it is important to note that if this

were indeed the case, we would expect the endogenous nature of this choice to affect the quality of

hire results in the opposite direction; if managers are most likely to use sponsorship when they are

able to personally identify an excellent candidate, sponsored hires should be expected to

 33

outperform hires made through posting. Empirically addressing this potential endogeneity

nevertheless seems prudent and I do so using an instrumental variable approach.

An instrumental variable should correlate strongly with endogenous variable (posting) but

not with the second stage error term (Hamilton & Nickerson, 2003; Semadeni, Withers, & Certo,

2014). A variable that appears to meet these two conditions is the percentage of similar jobs filled

by posting in other departments in the preceding two months. My discussions with both hiring

managers and HR staff at HealthCo confirmed that managers regularly, (a) ask other managers

how they recently filled similar jobs, or (b) contact HR and ask how similar positions have been

filled recently. While both mechanisms should be expected to influence the manager’s decision on

how to fill the job, how similar jobs are filled in other parts of the organization should have little

direct effect on post-hire outcomes or salary.

Table 6 compares the results for competency and contribution scores using OLS (Models 1

and 2) to those using both a traditional 2SLS approach (Models 3 and 4) and a 2SLS treatment

approach which accounts for the binary nature of my endogenous variable (Models 5 and 6). The

first stage estimates reveal that I do not have to worry about instrument weakness. As can be seen

by comparing the results of the different equations, the results are consistent with those obtained

using OLS. In fact, after I controlled for the endogeneity bias, the effect of posting of both

competency and contribution scores are substantially stronger, increasing from an increase of

around one-tenth of a standard deviation to nearly four-tenths of a standard deviation for

contribution and over half a standard deviation for competency.

------------------------------------
Insert Table 6 about here
------------------------------------
Table 7 presents the results for relative performance using instrumental variable probit

models. Aside from the change in coefficient magnitude, the main difference between these results
 34

and those from the logit models is that there is no significant relationship between how a worker

enters a job and their likelihood of landing in the top quartile of the performance distribution.

Taken together, the interpretation of these results remains unchanged – posting disciplines

managers to make better internal hiring decisions, in part by avoiding hiring mistakes.

------------------------------------
Insert Table 7 about here
------------------------------------
Table 8 presents the results for turnover and subsequent advancement using instrumental

variable probit models, as well as the result for salary using a traditional 2SLS approach. Again,

the results lead to similar interpretations as the non-instrumental variable models, with the main

difference being the magnitude of the coefficients. The one substantive difference is that the IV

Probit specification indicates that while internal hires are less likely to exit the firms with 12

months, they are not more or less likely to exit the firm within the longer 24 month time period.

------------------------------------
Insert Table 8 about here
------------------------------------
Wired Searches

While posting and sponsorship are conceptually distinct, it is possible for a manager to

post a job having already decided who they are going to select. Though these jobs are posted, they

are actually filled through sponsorship. If this was a common occurrence and/or these instances

were difficult to identify empirically, it would be difficult to interpret my results. These so-called

“wired searches” are likely to be most common in firms that require managers to post all jobs.

Because Health Co. has established formal systems for filling jobs through both posting and

sponsorship, managers are able to avoid posting jobs when they already have a candidate in mind

and thereby avoid the potential negative consequences that emerge from other employees having

felt they participated in an unfair selection process (Billsberry, 2007). I nevertheless conducted
 35

two additional robustness checks to rule of the possibility that my results are affected by the

inclusion of wired searches. First, I reviewed each of the 1,695 internal hires made between 2011

and 2012 for which I have detailed application data. Those instances where there was only a single

internal applicant (and no external applicants) could potentially be wired searches, with managers

either tailoring the job description around a particular candidate and/or discouraging others

employees from applying. Less than 5% of internal hires meet these criteria and of those, two-

thirds were open for more than a month, suggesting that the hiring manager was hoping to find

additional candidates. A total of 30 posted internal hires (1.4%) are potentially wired searches,

leading me to believe that such searches are not driving my results. Second, I identified all of the

postings during this period which were open a week or less (n = 45), another potential indication

that the hiring manager created the posting for a specific candidate and therefore wanted to limit

the number of other candidates who would apply. In both cases, dropping these observations from

the analyses does not substantively change the results.

External Candidates as an Alternative Explanation

I ran an additional series of robustness checks to ensure that the results for posting were

not driven by external market forces. When posting a job, a manager has the option of restricting

the competition to internal candidates by only posting the job internally, or opening the

competition to both internal and external candidates by also posting the job externally. An

alternative explanation for the performance findings might be that the inclusion of external

candidates allows managers to benchmark internal candidates against the market, leading to better

hires when an internal candidate is selected (Billsberry, 2007). Perhaps more importantly, an

alternative explanation for the higher starting salaries associated with posting might be that mere

exposure to the external market is responsible for driving up the starting salaries for posted jobs,
 36

with the results driven by those instances in which external candidates were considered but an

internal candidate was selected.

In order to rule out these alternative explanations, I use data from 2011 and 2012 that

allows me to identify the number of internal and external candidates that applied to every posted

job and how far they made it through the hiring process. This allows me to identify which job

postings were restricted to internal candidates and which were open to external candidates. Table 9

reveals the result of analyses including a dummy variable equal to one if external candidates were

considered for the position. Because this data begin in mid-2011, I am unable to calculate results

for turnover and subsequent advancement. Of the 869 internal hires for whom I have both

performance ratings and salary data, 578 (66%) were hired without considering an external

candidate; 291 (34%) internal hires competed against at least one external candidate. The results

indicate that opening the job to external candidates has no effect on performance (Models 1-5) or

starting salary (Models 6-7).

------------------------------------
Insert Table 9 about here
------------------------------------

DISCUSSION

This research advances our understanding of how human resources are allocated within

firms by shedding light on the processes used to facilitate internal hiring within contemporary

organizations. Despite the prevalence of internal hiring and its impact on the fortunes of firms and

workers, our current models of internal hiring are still largely based on literatures exploring

advancement with highly bureaucratic internal labor markets that bear little resemblance to their

more contemporary counterparts. This study provides a much-needed update to these earlier

models, identifying and describing the two most commonly used internal hiring processes –
 37

market-oriented posting and relationship-oriented sponsorship. I develop theory predicting the

relative effects of posting and sponsorship on quality of hire and compensation by highlighting

that posting is characterized by two market-like features, self-selection and formality, that are

largely absent in sponsorship, which instead involves active managerial search and a reliance on

personnel connections for gathering information.

I argued that introducing these two market features would improve decision-making by

helping managers to overcome challenges associated with generating and evaluating alternatives.

Self-selection should generate a larger pool of alternatives, reducing the likelihood and

exceptional alternative will be overlooked, while the formality of the market should encourage

managers to both incorporate relevant information and avoid irrelevant information as they

evaluate candidate qualifications against the requirements of the jobs. Consistent with these

arguments, I found that employees hired through posting outperform their counterparts hired

through sponsorship and are less likely to exit the firm, though there is no discernable difference

in their likelihood of subsequent advancement.

However, this improved decision-making comes at a price. I argued that employees hired

through the more competitive posting would be less likely to adopt a relational orientation to

salary negotiations therefore more likely to both initiate salary negotiations and adopt a more

effective, competitive approach when doing so. Though I am unable to test this mechanism

directly (I do not observe the difference between initial and final salary offers), the results are

consistent with this account, as employees hired through posting are paid ~4% more than

sponsored employees hired into equivalent jobs.

Despite this posting premium, these results suggest that posting enables firms to both

create and capture substantially more value than sponsorship. Unfortunately, my data do not allow

me to directly measure the value created by higher levels of individual performance or the savings
 38

associated with higher retention rates. However, with the costs of replacing an employee alone

running anywhere between 20% and 200% of an employee’s annual salary (Boushey & Glynn,

2012), the combined performance and retention benefits associated with better internal hires are

likely to far exceed the higher salary costs. Moreover, the posting premium suggests that workers

also share in the increased value creation, earning higher salaries for equivalent work. The market-

oriented posting process therefore appears to benefit both firms and workers.

As with all single-firm studies, some caution is necessary in generalizing these results.

Examining mobility within firms requires tradeoffs between depth and generalizability, requiring

detailed internal data that can be difficult to obtain from multiple sites. This is made even more

difficult by the fact that few firms actually capture data on the processes by which employees

move to new jobs. My conversations with HealthCo officials and with human resource leaders at

multiple additional organizations during the course of this study, however, have not revealed any

reason to believe that HealthCo’s internal hiring practices are different from other large

organizations.

In addition to not being able to directly measure value creation, this study does not account

for potential costs associated with creating an internal market for talent beyond the salary

premium. Additional costs might include the investments made to implement and maintain an

internal job board and the time involved in screening and evaluating job applications. Along these

lines, a particularly promising avenue for future research would be to explore whether posting

imposes additional costs by creating a visible set of employees who lost out in an open

competition, such as lower performance or increased turnover.

Despite these limitations, this work makes several important contributions. In highlighting

the ways in which self-selection and formality shape managerial decision-making and

negotiations, this study adds to the growing insights about the benefits and limitations of bringing
 39

market mechanisms within firms (Zenger et al., 2011). To date this work has focused

predominately on the ways in which the infusion of high powered incentives shapes individual

behavior (Ellig, 2001; Zenger & Hesterly, 1997; Zenger, 1992). More recently, scholars have

begun to explore the ways in which firms are able to harness the information aggregation powers

of the market to improve internal decision-making though the use of prediction and information

markets, with the idea being that the widely diffused information can be aggregated into

something akin to a price (Ellig, 2001; Felin & Zenger, 2011). This study complements and

extends this work by showing how two less explored features of markets – self-selection and

formality – improve decision-making in the absence of a price mechanism. Moreover, I show that,

contrary to what extant theory predicts about the role of competition in external markets, infusing

competition within the firm actually increases prices. Finally, in identifying the micro-level

mechanisms through which these macro-level allocative processes shape individual behaviors, this

work speaks to a growing body of literature interested in micro-foundations of strategic

organization (Felin & Foss, 2005; Ployhart & Hale, 2014) as well as a more specific literature on

the micro foundations of human capital-based competitive advantage (Coff & Kryscynski, 2011).
 40

REFERENCES

Afuah, A., & Tucci, C. L. 2012. Crowdsourcing as a solution to distant search. Academy of
Management Review, 37(3): 355–375.

Agerström, J., & Rooth, D.-O. 2011. The role of automatic obesity stereotypes in real hiring
discrimination. Journal of Applied Psychologysychology, 96(4): 790–805.

Agrawal, A., Knoeber, C. R., & Tsoulouhas, T. 2006. Are outsiders handicapped in CEO
successions? Journal of Corporate Finance, 12(3): 619–644.

Alexander, E. R. 1979. The design of alternatives in organizational contexts: A pilot study.

Administrative Science Quarterly, 24(3): 382–404.

Allen, D. G., Bryant, P. C., & Vardaman, J. M. 2010. Retaining talent: Replacing misconceptions
with evidence-based strategies. Academy of Management Perspectives, 24(2): 48–65.

Althauser, R. P., & Kalleberg, A. L. 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.

Argote, L., & Ingram, P. 2000. Knowledge transfer: A basis for competitive advantage in firms.
Organizational Behavior and Human Decision Processes, 82(1): 150–169.

Barney G. Glaser (Ed.). 1968. Organizational careers: A sourcebook for theory. Chicago: Aldine
Publishing.

Bazerman, M. H., & Chugh, D. 2005. Bounded awareness: Focusing failures in negotiation. In L.
Thompson (Ed.), Fronteirs of social psychology: Negotiation: 7–26. New York:
Psychological Press.

Bertrand, M., & Mullainathan, S. 2004. Are Emily and Greg More Employable Than Lakisha and
Jamal? A Field Experiment on Labor Market Discrimination. American Economic Review,
94(4): 991–1013.

Bidwell, M. J. 2011. Paying more to get less: The effects of external hiring versus internal
mobility. Administrative Science Quarterly, 56(3): 369–407.

Bidwell, M. J., & Keller, J. 2014. Within or without? How firms combine internal and external
labor markets to fill jobs. Academy of Management Journal, 57(4): 1035–1055.

Bielby, W. T. 2000. Minimizing workplace gender and racial bias. Comtemporary Sociology,
29(1): 120–129.

Bies, R. J., Tripp, T. M., & Neale, M. A. 1993. Procedural fairness and profit seeking: The
perceived legitimacy of market exploitation. Journal of Behavioral Decision Making,
6(June): 243–256.
 41

Billsberry, J. 2007. Experiencing recruitment and selection. West Sussex, England: John Wiley
& Sons.

Borgatti, S. P., & Cross, R. 2003. A relational view of information seeking and learning in social
networks. Management Science, 49(4): 432–445.

Bradach, J. L., & Eccles, R. G. 1989. Price, authority, and trust: From ideal types to plural forms.
Annual Review of Sociology, 15: 97–118.

Breaugh, J. A., & Starke, M. 2000. Research on employee recruitment: So many studies, so many
remaining questions. Journal of Management, 26(3): 405–434.

Brenner, L. A., Koehler, D. J., & Tversky, A. 1996. On the evaluation of one-sided evidence.
Journal of Behavioral Decision Making, 9: 59–70.

Burks, S., Cowgill, B., Hoffman, M., & Housman, M. 2013. The value of hiring through
referrals.

Campbell, J. P., McCloy, R. A., Oppler, S. H., & Sager, C. E. 1993. A theory of performance. In
N. Schmitt & W. C. Borman (Eds.), Personnel selection in organizations: 35–70. San
Fransisco, CA: Jossey-Bass.

Campion, M. A., Cheraskin, L., & Stevens, M. J. 1994. Career-related antecedents and outcomes
of job rotation. Academy of Management Journal, 37(6): 1518–1542.

Cappelli, P. 2008. Talent on demand: Managing talent in an age of uncertainty: 304.

Cambridge, MA: Harvard Business School Publishing.

Cappelli, P., & Keller, J. 2014. Talent management: Conceptual approaches and practical
challenges. Annual Review of Organizational Psychology and Organizational Behavior, 1:
305–331.

Capron, L., Dussauge, P., & Mitchell, W. 1998. Resource redeployment following horizontal
acquisitions in Europe and North America, 1988-1992. Strategic Management Journal,
19(7): 631–661.

Casciaro, T., & Lobo, M. S. 2008. When competence is irrelevant: The role of interpersonal affect
in task-related ties. Administrative Science Quarterly, 53(4): 655–684.

Cascio, W. F. 1998. Applied psychology in human resource management (5th ed.). Upper Saddle
River, NJ: Prentice Hall.

Cascio, W. F., & Aguinis, H. 2008. Staffing twenty-first century organizations. Academy of
Management Annals, 2(1): 133–165.

Casson, M. 1982. The entrepreneur: An economic theory. Oxford, UK: Martin Robertson.
 42

Coff, R. W., & Kryscynski, D. 2011. Drilling for micro-foundations of human capital-based
competitive advantages. Journal of Management, 37(5): 1429–1443.

Cowen, T., & Parker, D. 1997. Markets in the firm: A market-process approach to management.
London: The Institute of Economic Affairs.

Crispin, G., & Mehler, M. 2013. Sources of hire 2013: Perception is reality.

Cross, R., Parker, A., Prusak, L., & Borgatti, S. P. 2001. Knowing what we know: Supporting
knowledge creation and sharing in social networks. Organization Dynamics, 30(2): 100–120.

Cross, S. E., Bacon, P. L., & Morris, M. L. 2000. The relational-interdependent self-construal and
relationships. Journal of Personality and Social Psychology, 78(4): 791–808.

Curhan, J. R., Elfenbein, H. A., & Xu, H. 2006. What do people value when they negotiate?
Mapping the domain of subjective value in negotiation. Journal of Personality and Social
Psychology, 91(3): 493–512.

Curhan, J. R., Neale, M. A., Ross, L., & Rosencranz-Engelmann, J. 2008. Relational
accommodation in negotiation: Effects of egalitarianism and gender on economic efficiency
and relational capital. Organizational Behavior and Human Decision Processes, 107(2):
192–205.

Cyert, R. M., & March, J. G. 1963. A behavioral theory of the firm. Englewood Cliffs, N.J.:
Prentice-Hall.

Dencker, J. C. 2008. Corporate restructuring and sex differences in managerial promotion.

American Sociological Review, 73(3): 455–476.

Diprete, T. A. 1987. Horizontal and vertical mobility in organizations. Administrative Science

Quarterly, 32(3): 422–444.

Doeringer, P. B., & Piore, M. J. 1971. Internal labor markets and manpower analysis.
Lexington, MA: D.C. Heath and Company.

Dougherty, T. W., Turban, D. B., & Callender, J. C. 1994. Confirming first impressions in the
employment interview: A field study of interviewer Behavior. Journal of Applied
Psychology, 79(5): 659–665.

Drazin, R., & Rao, H. 2002. Harnessing managerial knowledge to implement product-line
extensions: how to mutual find families allocate portfolio managers to old and new funds?
Academy of Management Journal, 45(3): 609–619.

Ellig, J. 2001. Internal markets and the theory of the firm. Managerial and Decision Economics,
22(45): 227–237.

ERC. 2010. ERC staffing practices survey. Mayfield Village, OH.

 43

Felin, T., & Foss, N. J. 2005. Strategic organization: a field in search of micro-foundations.
Strategic Organization, 3(4): 441–455.

Felin, T., & Zenger, T. R. 2011. Information aggregation, matching and radical market-hierarchy
hybrids: Implications for the theory of the firm. Strategic Organization, 9(2): 163–173.

Fernandez, R. M., Castilla, E. J., & Moore, P. 2000. Social capital at work: Networks and
employment at a phone center. American Journal of Sociology, 105(5): 1288–1356.

Gelfand, M. J., Major, V. S., Raver, J. L., Nishii, L. H., Brien, O., Michele, J., et al. 2006.
Negotiating relationally: The dynamics of the relational self in negotiations. Academy of
Management Review, 31(2): 427–451.

Gemünden, H. G., & Hauschildt, J. 1985. Number of alternatives and efficiency in different types
of top-management decisions. European Journal of Operational Research, 22(2): 178–190.

Granovetter, M. S. 1985. Economic action and social structure: The problem of embeddedness.
American Journal of Sociology, 91(3): 481–510.

Groysberg, B., Lee, L.-E., & Nanda, A. 2008. Can they take it with them? The portability of star
knowledge workers’ performance. Management Science, 54(7): 1213–1230.

Hamilton, B. H., & Nickerson, J. A. 2003. Correcting for endogeneity in strategic management
research. Strategic Organization, 1(1): 51–78.

Hauser, J. R., & Wernerfelt, B. 1990. An evaluation cost model of consideration sets. Journal of
Consumer Research, 16(4): 393–408.

Highhouse, S. 2008. Stubborn reliance on intuition and subjectivity in employee selection.

Industrial and Organizational Psychology, 1(3): 333–342.

Ibarra, H. 1993. Personal networks of women and minorities in management: A conceptual

framework. Academy of Management Review, 18(1): 56–87.

Iyengar, S. S., & Kamenica, E. 2010. Choice proliferation, simplicity seeking, and asset allocation.
Journal of Public Economics, 94(7-8): 530–539.

Jackson, C. K. 2013. Match quality, worker productivity, and worker mobility: Direct evidence
from teachers. The Review of Economics and Statistics, 95(4): 1096–1116.

Jacoby, S. M. 1985. Employing bureaucracy: Managers, unions, and the transformation of

work in American industry. New York: Columbia University Press.

Jacoby, S. M. 2004. Employing bureaucracy: Managers, unions, and the transformation of

work in the 20th century. Mahwah, NJ: Lawrence Erlbaum.
 44

Jacoby, S. M. 2005. The embedded corporation: Corporate governance and employment

relations in Japan and the United States. Princeton, NJ: Princeton University Press.

Jovanovic, B. 1979. Job matching and the theory of turnover. Journal of Political Economy,
87(5): 972–990.

Kahneman, D. 2011. Thinking, fast and slow. New York: Farrar, Straus and Giroux.

Kanter, R. M. 1977. Men and women of the corporation. New York: Basic Books.

Knudsen, T., & Levinthal, D. A. 2007. Two faces of search: Alternative generation and alternative
evaluation. Organization Science, 18(1): 39–54.

Loasby, B. J. 2000. Market institutions and economic evolution. Journal of Evolutionary

Economics, 10(3): 297–309.

London, M. 1996. Redeployment and continuous learning in the 21st Century: Hard lessons and
positive examples from the downsizing era. Academy of Management Executive, 10(4): 67–
79.

March, J. G., & Simon, H. A. 1958. Organizations. New York: Wiley.

Marks, M., & Harold, C. 2011. Who asks and who receives in salary negotiation. Journal of
Organizational Behavior, 394: 371–394.

Marlowe, C. M., Schneider, S. L., & Nelson, C. E. 1996. Gender and attractiveness biases in
hiring decisions: Are more experienced managers less biased ? Journal of Applied
Psychology, 81(1): 11–21.

Marsden, P. V., & Gorman, E. H. 1999. Social capital in internal staffing practices. In R. T. A. J.
Leenders & S. M. Gabbay (Eds.), Corporate social capital and liability: 180–196. Boston,
MA: Kluwer Academic Publishers.

McEvily, B., Soda, G., & Tortoriello, M. 2014. More formally: Rediscovering the missing link
between formal organization and informal social structure. Academy of Management
Annals, 8(1): 299–345.

Mclntyre, R. M., Smith, D. E., & Hassett, C. E. 1984. Accuracy of performance ratings as affected
by rater training and perceived purpose of rating. Journal of Applied Psychology, 69(I): 147–
156.

Medoff, J. L., & Abraham, K. G. 1981. Are those paid more really more productive? The case of
experience. Journal of Human Resources, 16(2): 186–216.

Menard, C. 1995. Markets as institutions versus organizations as markets? Disentangling some

fundamental concepts. Journal of Economic Behavior and Organization, 28: 161–182.
 45

Miller, R. A. 1984. Job matching and occupational choice. Journal of Political Economy, 92(6):
1086–1120.

Miner, A. S. 1987. Idiosyncratic jobs in formalized organizations. Administrative Science

Quarterly, 32(3): 327–351.

Mintzberg, H. 1994. The rise and fall of strategic planning. New York: Free Press.

Mintzberg, H., Raisinghani, D., & Théorêt, A. 1976. The structure of “unstructured” decision
processes. Administrative Science Quarterly, 21(2): 246–275.

Mortensen, D. T. 1988. Wages, separations, and job tenure: On-the-job specific training or
matching? Journal of Labor Economics, 6(4): 445–471.

Nahapiet, J., & Ghoshal, S. 1998. Social capital, intellectual capital, and the organizational
advantage. Academy of Management Review, 23(2): 242–266.

Nickerson, J. A., & Zenger, T. R. 2004. A knowledge-based theory of the firm: The problem-
solving perspective. Organization Science, 15(6): 617–632.

Nutt, P. C. 1998. How decision makers evaluate alternatives and the influence of complexity.
Management Science, 44(8): 1148–1166.

Ocasio, W. 1997. Towards an attention-based view of the firm. Strategic, 18(Summer Special
Issue): 187–206.

Oracle. 2012. U.K. talent mobility.

Pinfield, L. T. 1995. The operation of internal labor markets: Staffing practices and vacancy
chains. New York: Plenum Press.

Piore, M. J. 2002. Thirty years later: Internal labor markets, flexibility and the new economy.
Journal of Management and Governance, 6(4): 271–279.

Ployhart, R. E., & Hale, D. 2014. The fascinating psychological microfoundations of strategy and
competitive advantage. Annual Review of Organizational Psychology and Organizational
Behavior, 1(1): 145–172.

Podolny, J. M., & Baron, J. N. 1997. Resources and Relationships: Social Networks and Mobility
in the Workplace. American Sociological Review, 62(5): 673.

Polyani, K. 1957. The great transformation. Boston: Beacon Press.

Powell, W. W., & Snellman, K. 2004. The knowledge economy. Annual Review of Sociology,
30(1): 199–220.
 46

Pulakos, E. D. 1984. A comparison of rater training programs: Error training and accuracy
training. Journal of Applied Psychology, 69(4): 581–588.

Pulakos, E. D., & O’Leary, R. S. 2011. Why is performance management broken? Industrial and
Organizational Psychology, 4: 146–164.

Reskin, B. F., & Roos, P. A. 1990. Job queues, gender queues: Explaining women’s inroads into
male occupations. Temple University Press.

Rogan, M., & Sorenson, O. 2014. Picking a (poor) partner: A relational perspective on
acquisitions. Administrative Science Quarterly, 59(2): 301–329.

Rosenbaum, J. E. 1979. Organizational career mobility: Promotion chances in a corporation

during periods of growth and contraction. American Journal of Sociology, 85(1): 21–48.

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.

Seidel, M.-D. L., Polzer, J. T., & Stewart, K. J. 2000. Friends in high places: The effects of social
networks on discrimination in salary negotiations. Administrative Science Quarterly, 45(1):
1–24.

Semadeni, M., Withers, M. C., & Certo, S. T. 2014. The perils of endogeneity and instrumental
variables in strategy research: Understanding through simulations. Strategic Management
Journal.

Shocker, A. D., Ben-Akiva, M., Boccara, B., & Nedungadi, P. 1991. Consideration set influences
on consumer decision-making and choice: Issues, models, and suggestions. Marketing
Letters, 2(3): 181–197.

Simon, H. A. 1955. A behavioral model of rational choice. Quarterly Journal of Economics, 69:
99–118.

Sirmon, D. G., Gove, S., & Hitt, M. A. 2008. Resource management in dyadic competitive rivalry:
The effects of resorce bundling and deployment. Academy of Management Journal, 51(5):
919–935.

Sirmon, D. G., & Hitt, M. A. 2009. Contingincies within dynamic managerial capabilities:
Interdependent effects of resource investment and depolyment on firm performance.
Strategic Management Journal, 1394: 1375–1394.

Slichter, S. H., Healy, J. J., & Livernash, E. R. 1960. The impact of collective bargaining of
management. Washington, D.C.: The Brookings Institution.

Sorenson, O., & Waguespack, D. M. 2006. Social structure and exchange: Self-confirming
dynamics in Hollywood. Administrative Science Quarterly, 51: 560–589.
 47

Stein, J. C. 1997. Internal capital markets and the competition for corporate resources. Journal of
Finance, 52(1): 111–133.

Strum, S. 2001. Second generation employement discrimination: A structural approach. Columbia

Law Review, 101: 458–568.

Taleo Research. 2005. Internal mobility. Redwood Shores, CA.

Tetlock, P. E. 1992. The impact of accountability on judgement and choice. In M. P. Zanna (Ed.),
Advance in experimental social psychology (25th ed.): 331–376. San Diego, CA: Academic
Press.

Thompson, L., & Deharpport, T. 1998. Relationships, goal incompatibility, and communal
orientation in negotiations. Basic and Applied Social Psychology, 20(1): 33–44.

Tsai, W. 2002. Social structure of “coopetition” within a multiunit organization: Coordination,

competition, and intraorganizational knowledge sharing. Organization Science, 12(2): 179–
190.

Tsai, W., & Ghoshal, S. 1998. Social capital and value creation: The role of intrafirm networks.
Academy of Management Journal, 41(4): 464–476.

Turner, R. H. 1960. Sponsored and contest mobility and the school system. American
Sociological Review, 25(6): 855–867.

Williams, P. F. 1987. The legitimate concern with fairness. Accounting, Organizations and
Society, 12(2): 169–189.

Zenger, T. R. 1992. Why do employers only reward extreme performance? Examining the
relationships among performance, pay, and turnover. Administrative Science Quarterly,
37(2): 198–219.

Zenger, T. R. 2002. Crafting internal hybrids: Complementarities, common change initiatives, and
the team-based organization. International Journal of the Economics of Business, 9(1): 79–
95.

Zenger, T. R., Felin, T., & Bigelow, L. 2011. Theories of the firm-market boundary. Academy of
Management Annals, 5(1): 89–133.

Zenger, T. R., & Hesterly, W. S. 1997. The disaggregation of corporations: Selective intervention,
high-powered incentives, and molecular units. Organization Science, 8(3): 209–222.

Zottoli, M. A., & Wanous, J. P. 2000. Recruitment source research: Current status and future
directions. Human Resource Management Review, 10(4): 353–382.
 48

TABLE 1
Descriptive Statistics & Correlationsa

Variable n Mean S.D. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

1 posting 11803 0.43 0.5
2 instrumental var. 11331 0.44 0.27 0.30
3 contribution score 9520 2.85 0.64 ‐0.05 ‐0.04
4 competency score 9590 2.55 0.53 ‐0.05 ‐0.03 0.67
5 top 25% 10115 0.29 0.45 ‐0.01 0.02 0.43 0.45
6 bottom 50% 10115 0.49 0.5 0.02 ‐0.02 ‐0.46 ‐0.50 ‐0.60
7 bottom 25% 10115 0.25 0.43 0.03 ‐0.01 ‐0.40 ‐0.46 ‐0.35 0.59
8 turnover (12) 8359 0.03 0.16 ‐0.01 0.01 ‐0.05 ‐0.03 ‐0.03 0.03 0.03
9 turnover (24) 5508 0.12 0.32 ‐0.03 ‐0.02 ‐0.12 ‐0.11 ‐0.09 0.08 0.07 0.56
10 promotion (24) 5675 0.13 0.33 0.02 0.06 0.06 0.06 0.06 ‐0.06 ‐0.06 ‐0.01 ‐0.09
11 promotion (36) 3739 0.24 0.43 0.07 0.10 0.06 0.06 0.05 ‐0.06 ‐0.06 ‐0.01 ‐0.08 0.87
12 salary (ln) 11776 11.02 0.4 ‐0.10 ‐0.18 0.08 0.13 0.03 ‐0.01 0.00 0.03 0.03 ‐0.15 ‐0.16
13 job level 11803 4.63 0.76 ‐0.15 ‐0.27 0.12 0.13 0.01 ‐0.01 ‐0.01 0.01 0.03 ‐0.32 ‐0.39 0.73
14 same dept 11803 0.58 0.49 ‐0.39 ‐0.06 0.17 0.17 0.11 ‐0.14 ‐0.14 ‐0.03 ‐0.11 ‐0.03 ‐0.10 ‐0.02 ‐0.01
15 same function 11803 0.68 0.47 ‐0.14 ‐0.02 0.03 0.08 0.06 ‐0.05 ‐0.05 ‐0.01 ‐0.02 ‐0.09 ‐0.12 0.07 0.03 0.21
16 promotion 11803 0.48 0.5 0.15 0.13 0.03 0.05 0.08 ‐0.05 ‐0.02 ‐0.04 ‐0.08 ‐0.07 ‐0.09 ‐0.30 ‐0.16 0.08 0.05
17 expansion 11803 0.39 0.49 ‐0.10 ‐0.14 0.00 ‐0.03 ‐0.06 0.03 0.01 0.02 0.04 0.03 0.03 0.23 0.19 0.03 ‐0.17
18 transfer 11803 0.13 0.33 ‐0.09 0.01 ‐0.04 ‐0.03 ‐0.04 0.03 0.01 0.04 0.06 0.06 0.09 0.12 ‐0.04 ‐0.16 0.17
19 female 11697 0.64 0.48 0.01 0.04 ‐0.05 ‐0.03 0.02 ‐0.02 ‐0.02 ‐0.03 ‐0.04 0.01 0.00 ‐0.22 ‐0.14 0.01 ‐0.03
20 minority 11697 0.24 0.42 0.03 0.05 ‐0.07 ‐0.06 ‐0.01 0.02 0.03 0.01 0.00 ‐0.03 ‐0.02 ‐0.11 ‐0.11 ‐0.03 0.02
21 white 11803 0.76 0.43 ‐0.02 ‐0.06 0.07 0.06 0.01 ‐0.02 ‐0.03 ‐0.01 0.00 0.02 0.01 0.12 0.11 0.03 ‐0.02
22 black 11803 0.1 0.3 0.04 0.06 ‐0.09 ‐0.10 ‐0.01 0.02 0.04 0.00 ‐0.01 ‐0.01 ‐0.01 ‐0.12 ‐0.13 ‐0.04 0.01
23 latino 11803 0.08 0.27 0.01 0.04 ‐0.03 0.02 0.00 0.01 ‐0.01 0.02 0.01 ‐0.03 ‐0.02 ‐0.10 ‐0.06 0.00 0.00
24 asian 11803 0.03 0.18 ‐0.03 ‐0.04 0.05 0.04 0.00 0.01 0.01 0.01 0.02 ‐0.02 ‐0.01 0.10 0.08 0.02 0.02
25 tenure 11803 5.4 5.22 ‐0.04 0.00 0.10 0.14 0.02 ‐0.03 ‐0.03 ‐0.03 ‐0.07 ‐0.06 ‐0.06 0.14 0.15 0.01 ‐0.05
26 tenure2 11803 56.33 122.2 ‐0.07 ‐0.02 0.05 0.08 0.01 ‐0.02 ‐0.01 ‐0.02 ‐0.05 ‐0.06 ‐0.07 0.12 0.12 0.04 ‐0.03
27 age 11697 38.59 9.93 ‐0.15 ‐0.12 ‐0.02 0.03 ‐0.02 0.01 0.00 0.01 0.04 ‐0.15 ‐0.20 0.31 0.26 0.03 0.03
28 age2 11697 1588.12 837.31 ‐0.15 ‐0.12 ‐0.03 0.02 ‐0.03 0.02 0.00 0.01 0.05 ‐0.14 ‐0.19 0.29 0.24 0.03 0.03
29 rating group size 10115 479.3 408.2 0.05 0.07 0.03 ‐0.01 0.02 ‐0.04 ‐0.06 0.01 ‐0.01 0.07 0.12 ‐0.18 ‐0.23 0.05 0.01
30 last salary (ln) 11787 10.93 0.42 ‐0.17 ‐0.23 0.09 0.13 0.02 ‐0.01 ‐0.02 0.03 0.04 ‐0.13 ‐0.14 0.94 0.72 0.00 0.07
31 last contribution 9866 2.89 0.61 ‐0.03 ‐0.05 0.52 0.42 0.24 ‐0.23 ‐0.19 ‐0.05 ‐0.09 0.04 0.08 0.09 0.12 0.10 0.00
32 last competency 9897 2.57 0.51 ‐0.02 ‐0.01 0.41 0.59 0.25 ‐0.25 ‐0.20 ‐0.05 ‐0.09 0.01 0.04 0.13 0.12 0.10 0.03
33 year 11803 2010.51 1.31 0.04 0.08 0.02 0.02 ‐0.02 0.01 ‐0.01 0.23 0.47 0.28 0.54 0.04 0.04 ‐0.19 ‐0.11
16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32
17 expansion ‐0.77
18 transfer ‐0.37 ‐0.30
19 female 0.01 ‐0.02 0.01
20 minority 0.07 ‐0.06 ‐0.02 0.02
21 white ‐0.07 0.06 0.02 ‐0.02 ‐1.00
22 black 0.06 ‐0.05 ‐0.03 0.07 0.62 ‐0.60
23 latino 0.04 ‐0.04 0.00 0.02 0.52 ‐0.51 ‐0.10
24 asian ‐0.01 0.00 0.01 ‐0.08 0.34 ‐0.33 ‐0.06 ‐0.05
25 tenure ‐0.09 0.11 ‐0.02 0.08 ‐0.06 0.07 ‐0.01 ‐0.02 ‐0.03
26 tenure2 ‐0.09 0.09 ‐0.01 0.08 ‐0.06 0.07 ‐0.02 ‐0.03 ‐0.04 0.92
27 age ‐0.25 0.18 0.11 0.06 ‐0.08 0.08 ‐0.05 ‐0.03 ‐0.03 0.28 0.29
28 age2 ‐0.23 0.16 0.11 0.06 ‐0.09 0.09 ‐0.06 ‐0.04 ‐0.03 0.26 0.28 0.99
29 rating group size 0.11 ‐0.11 0.00 ‐0.03 0.05 ‐0.06 0.01 0.04 0.06 0.03 0.01 ‐0.09 ‐0.09
30 last salary (ln) ‐0.40 0.29 0.18 ‐0.21 ‐0.13 0.13 ‐0.13 ‐0.12 0.11 0.18 0.16 0.36 0.34 ‐0.16
31 last contribution 0.05 ‐0.02 ‐0.05 ‐0.03 ‐0.06 0.06 ‐0.10 ‐0.01 0.04 0.14 0.07 ‐0.01 ‐0.02 0.01 0.09
32 last competency 0.08 ‐0.06 ‐0.03 0.00 ‐0.03 0.03 ‐0.10 0.05 0.02 0.20 0.11 0.05 0.04 ‐0.01 0.12 0.67
33 year ‐0.06 0.06 0.01 0.01 0.01 ‐0.02 0.00 0.02 ‐0.01 0.07 0.04 0.03 0.02 0.11 0.05 0.01 0.01

a
Unit of analysis is an internal hire
 49

TABLE 2
Performancea
Model 1 Model 2 Model 3 Model 4 Model 5
OLS OLS Logit Logit Logit
Variables Contribution Competency Top 25% Bottom 50% Bottom 25%
Posting 0.0542** 0.0621** 0.129* -0.133** -0.149*
[0.0155] [0.0123] [0.0548] [0.0516] [0.0634]
Level 4 -0.0406 -0.117* -0.34 0.275 0.14
[0.0862] [0.0551] [1.086] [1.001] [0.751]
Level 5 0.15 -0.395* -0.592 0.678 0.798*
[0.251] [0.166] [1.005] [0.916] [0.339]
Level 6 0.0491 -0.143 0.0583 0.128 0.566*
[0.260] [0.174] [0.200] [0.197] [0.271]
Same Department 0.208** 0.175** 0.392** -0.439** -0.495**
[0.0159] [0.0126] [0.0586] [0.0531] [0.0635]
Same Function 0.0316* 0.0751** 0.295** -0.155** -0.152*
[0.0149] [0.0125] [0.0591] [0.0534] [0.0644]
Promotion 0.00351 -0.0219 0.242** -0.0874 0.0152
[0.0216] [0.0168] [0.0820] [0.0718] [0.0857]
Female 0.0026 0.0233† 0.115* -0.108* -0.101
[0.0146] [0.0121] [0.0550] [0.0511] [0.0617]
Black -0.156** -0.135** -0.116 0.129 0.325**
[0.0240] [0.0193] [0.0863] [0.0790] [0.0888]
Latino -0.0497† 0.0601** -0.000748 0.0865 0.0346
[0.0272] [0.0232] [0.0951] [0.0897] [0.110]
Asian 0.000733 -0.0291 -0.0805 0.187 0.214
[0.0328] [0.0285] [0.140] [0.129] [0.146]
Other -0.0454 -0.0501 -0.0131 -0.0578 -0.0164
[0.0601] [0.0458] [0.193] [0.189] [0.229]
Tenure 0.0350** 0.0387** 0.0534** -0.0476** -0.0524**
[0.00379] [0.00345] [0.0148] [0.0129] [0.0153]
Tenure (sq) -0.00120** -0.00132** -0.00195** 0.00157** 0.00170**
[0.000155] [0.000148] [0.000663] [0.000548] [0.000613]
Age -0.00441** -0.00120† -0.00767** 0.00663* 0.00684*
[0.000828] [0.000676] [0.00291] [0.00266] [0.00324]
Rating Group Size 0.0000295 -0.000145* -0.000385**
[6.87e-05] [6.64e-05] [8.65e-05]
Constant 2.956@ 2.838@ -1.223! 0.0529 -0.973
[0.118] [0.0853] [0.524] [0.508] [0.599]
Observations 9300 9300 9289 9276 9262
R-squared 0.136 0.154
a
Standard errors, in brackets, are clustered by individual. All analyses also include dummies for job, function, state, and year.
† p < .10; ** p < .05; ** p < .01
 50

TABLE 3
Turnover, Advancement & Salarya

Model 6 Model 7 Model 8 Model 9 Model 10 Model 11

Logit Logit Logit Logit OLS OLS
Variables Turnover12 Turnover24 Promotion24 Promotion36 Salary (ln) Salary (ln)
Posting -0.202* -0.182* -0.0949 0.128† 0.0383** 0.0376**
[0.0986] [0.0761] [0.0645] [0.0733] [0.00256] [0.00272]
Level 4 0.754 0.058 -0.0866 -0.526 0.258** -0.0489
[0.861] [0.829] [0.222] [0.561] [0.0498] [0.0688]
Level 5 0.511 -0.546 -1.172** -1.506** 0.167** 0.0159
[0.467] [0.426] [0.148] [0.553] [0.0218] [0.0526]
Level 6 0.539† -0.365 -1.563** -2.166** 0.221** 0.0331
[0.311] [0.340] [0.361] [0.445] [0.0300] [0.0559]
Same Dept -0.0925 -0.259** 0.0402 0.0012 0.000872 0.00125
[0.0959] [0.0738] [0.0651] [0.0779] [0.00270] [0.00275]
Same Function -0.0336 0.0874 -0.217** -0.0758 0.0166** 0.0152**
[0.0987] [0.0671] [0.0610] [0.0756] [0.00264] [0.00270]
Promotion -0.302** 0.0791 -0.274** -0.279** 0.0960** 0.0865**
[0.116] [0.0954] [0.0777] [0.0892] [0.00874] [0.00890]
Female -0.236** -0.148* -0.121* -0.123† -0.0101** -0.00912**
[0.0871] [0.0648] [0.0577] [0.0666] [0.00230] [0.00235]
Black 0.0358 -0.0981 -0.269** -0.302** -0.00637* -0.00126
[0.131] [0.109] [0.0896] [0.0962] [0.00313] [0.00331]
Latino -0.116 -0.0186 -0.349** -0.395** 0.000767 0.00714
[0.165] [0.111] [0.101] [0.125] [0.00494] [0.00514]
Asian -0.0644 -0.0117 0.005 0.2 -0.00445 -0.00181
[0.214] [0.151] [0.159] [0.178] [0.00406] [0.00421]
Tenure -0.104** -0.0828** -0.00392 0.0333† -0.000281 -0.0009
[0.0223] [0.0159] [0.0152] [0.0187] [0.000559] [0.000588]
Tenure (sq) 0.00299** 0.00212** 0.0000181 -0.00150† 0.00000442 2.66E-05
[0.000912] [0.000625] [0.000653] [0.000882] [2.07e-05] [2.09e-05]
Age -0.00115 0.0114** -0.0144** -0.0153** -0.00153† -0.00163†
[0.00523] [0.00316] [0.00319] [0.00365] [0.000829] [0.000873]
Last salary (ln) 0.851** 0.863**
[0.00924] [0.00893]
Last contribution 0.00860**
[0.00256]
Last competency 0.00354
[0.00353]
Constant -6.111** -2.416** 0.061 0.596 1.561** 1.342**
[0.757] [0.594] [0.345] [0.633] [0.117] [0.101]
Observations 7451 5056 5178 3302 9292 8017
R-Squared 0.955 0.959
a
Standard errors, in brackets, are clustered by individual. All analyses also include dummies for job, function, state, and year.
† p < .10; ** p < .05; ** p < .01.
 51

TABLE 4
Source of Internal Hires by Posting and Sponsorship

Source of Hire Statistical

(Location of previous job within the organization) Posting Sponsorship Significance

Different Function 40% 27% p < .01

Different Department 64% 26% p < .01

Different City 15% 4% p < .01

Different Building 42% 8% p < .01

Transfer (same job, different department) 10% 15% p < .01

Expansion (same level, different job) 32% 43% p < .01

Promotion (move up a level) 57% 42% p < .01

 52

TABLE 5
Performance, Turnover, Salary, and Advancement within Departmenta,b,c

Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 Model 7 Model 8 Model 9 Model 10
OLS OLS Logit Logit Logit OLS OLS Logit Logit OLS
Bottom Bottom
Variables Contrib. Comp. Top 25% Turn12 Turn24 Prom24 Prom36 Salary (ln)
50% 25%
Posting 0.0474** .0678** 0.160** -0.127* -0.165* -0.284* -0.151 -0.119 0.117 0.0468**
[0.0171] [0.0137] [0.0609] [0.0585] [0.0740] [0.123] [0.0988] [0.136] [0.150] [0.00325]
Level 4 0.0973 0.550** -1.277 -0.215 -0.418 0.442 0.897 0.328 -0.0378 0.231**
[0.0758] [0.0645] [1.505] [0.638] [0.890] [0.897] [0.763] [0.499] [0.524] [0.0637]
Level 5 0.308 0.281† -1.433 0.231 0.481 0.225 0.257 -2.087** -1.967** 0.156**
[0.239] [0.153] [1.444] [0.287] [0.394] [0.520] [0.229] [0.334] [0.388] [0.0259]
Level 6 0.237 0.537** 0.199 0.00687 0.271 0.367 0.329 -2.541** -2.888** 0.116**
[0.249] [0.162] [0.220] [0.231] [0.335] [0.381] [0.360] [0.845] [0.782] [0.0338]
Same Dept 0.152** 0.129** 0.319** -0.389** -.348** -0.0509 -0.0527 0.267† 0.127 0
[0.0190] [0.0152] [0.0688] [0.0624] [0.0765] [0.119] [0.0970] [0.149] [0.172] [0]
Same Funct 0.0196 0.0480** 0.261** -0.142* -0.138† 0.145 0.0614 -0.355** 0.054 0.0106**
[0.0175] [0.0150] [0.0696] [0.0638] [0.0807] [0.127] [0.0868] [0.133] [0.166] [0.00336]
Tenure 0.0396** 0.0441** 0.0662** -00569** -.0637** -0.105** -.0875** 0.00176 0.0842* -0.000655
[0.00413] [0.00376] [0.0162] [0.0146] [0.0174] [0.0250] [0.0184] [0.0348] [0.0418] [0.000650]
Tenure (sq) -.00138** -.00149** -.0023** .00183** .00198** .00315** .00239** -0.000855 -0.005* 0.00000997
[.000173] [.000163] [.000735] [.000623] [0.0007] [.000991] [0.0007] [0.00153] [0.00212] [2.42e-05]
Age -.00470** -.00153* -.0085** .00841** .0101** 0.000178 .0110** -.0311** -.033** -0.00174†
[0000887] [.000728] [0.00317] [0.00294] [0.00360] [0.00603] [0.00383] [0.00668] [0.00760] [0.00101]
Constant 2.830** 2.199** -1.486** 0.202 -0.723 -5.912** -3.468** 0.162 0.532 0.827**
[0.0877] [0.0738] [0.548] [0.528] [0.638] [1.115] [0.566] [0.746] [0.707] [0.134]

Obs 7568 7568 7559 7543 7505 5777 4044 4203 2720 7568
R-squared 0.138 0.161 0.966
a
Standard errors, in brackets, are clustered by individual. All analyses also include dummies for job, function, state, year, ethnicity, and gender.
b
Model 3, 4, and 5 contain additional controls for performance group size
c
Model 10 includes additional controls for last salary, last competency rating, and last contribution rating
† p < .10; ** p < .05; ** p < .01
 53

TABLE 6
Performance Instrumental Variables (Part 1)a
Model 1 Model 2 Model 3 Model4 Model5 Model 6
OLS OLS 2SLS 2SLS 1st stage 2SLS/Treatment 2SLS/Treatment 1st stage
Variables Contribution Competency Contribution Competency Contribution Competency
Posting 0.0542** 0.0621** 0.259** 0.271** 0.227** 0.199**
[0.0155] [0.0123] [0.0934] [0.0764] [0.0765] [0.0609]
Same Department 0.208** 0.175** 0.296** 0.262** -0.415** 0.283** 0.232** -1.265**
[0.0159] [0.0126] [0.0422] [0.0343] [0.01] [0.0353] [0.0281] [0.0344]
Same Function 0.0316* 0.0751** 0.0387* 0.0871** -0.068** 0.0366* 0.0824** -0.194**
[0.0149] [0.0125] [0.0164] [0.0136] [0.01] [0.0164] [0.0131] [0.0379]
Promotion 0.00351 -0.0219 -0.0445 -0.0711** 0.216** -0.0375 -0.0555** 0.636**
[0.0216] [0.0168] [0.0300] [0.0237] [0.016] [0.0264] [0.0211] [0.0483]
Female 0.0026 0.0233† -0.00083 0.0222† -0.00 -0.000927 0.0220† -0.00499
[0.0146] [0.0121] [0.0150] [0.0124] [0.01] [0.0143] [0.0114] [0.0342]
Black -0.156** -0.135** -0.148** -0.130** -0.040** -0.149** -0.133** -0.135**
[0.0240] [0.0193] [0.0245] [0.0197] [0.02] [0.0218] [0.0174] [0.0507]
Latino -0.0497† 0.0601** -0.0469 0.0657** -0.042* -0.0483† 0.0627** -0.141*
[0.0272] [0.0232] [0.0286] [0.0243] [0.02] [0.0247] [0.0196] [0.0579]
Tenure 0.0350** 0.0387** 0.0343** 0.0387** 0.00 0.0345** 0.0391** 0.0183*
[0.00379] [0.00345] [0.00387] [0.00340] [0.00] [0.00349] [0.00278] [0.00859]
Tenure (sq) -0.00120** -0.00132** -0.00115** -0.00128** -0.00** -0.00116** -0.00131** -0.00135**
[0.000155] [0.000148] [0.000160] [0.000144] [0.00] [0.000146] [0.000117] [0.000369]
Age -0.00441** -0.00120† -0.00406** -0.00062 -0.002** -0.00415** -0.000819 -0.00910**
[0.000828] [0.000676] [0.000878] [0.000720] [0.00] [0.000766] [0.000610] [0.00177]
IV: Other posting 0.307** 1.007**
[.02] [0.0670]
lambda -0.102* -0.0793*
[0.0455] [0.0363]
Constant 2.956** 2.838** 2.331** 2.061** 0.925*
[0.118] [0.0853] [0.194] [0.154] [0.451]
Observations 9300 9300 8929 8929 8929 8929 8929 8929
R-squared 0.136 0.154
a
Standard errors, in brackets, are clustered by individual. All analyses also include dummies for job, function, state, and year.
† p < .10; ** p < .05; ** p < .01
 54

TABLE 7
Performance Instrumental Variables (Part 2)a
Model 1 Model 2 Model 3 Model 4 Model 5 Model 6
Logit Logit Logit IV Probit IV Probit IV Probit
Variables Top 25% Bottom 50% Bottom 25% Top 25% Bottom 50% Bottom 25%
Posting 0.129* -0.133** -0.149* 0.305 -0.720** -0.572*
[0.0548] [0.0516] [0.0634] [0.209] [0.204] [0.227]
Same Department 0.392** -0.439** -0.495** 0.331** -0.540** -0.487**
[0.0586] [0.0531] [0.0635] [0.0938] [0.0914] [0.102]
Same Function 0.295** -0.155** -0.152* 0.192** -0.141** -0.120**
[0.0591] [0.0534] [0.0644] [0.0378] [0.0364] [0.0405]
Promotion 0.242** -0.0874 0.0152 0.088 0.083 0.108
[0.0820] [0.0718] [0.0857] [0.0656] [0.0627] [0.0694]
Female 0.115* -0.108* -0.101 0.0726* -0.0670* -0.0577†
[0.0550] [0.0511] [0.0617] [0.0321] [0.0311] [0.0348]
Black -0.116 0.129 0.325** -0.0611 0.0682 0.178**
[0.0863] [0.0790] [0.0888] [0.0490] [0.0475] [0.0519]
Latino -0.000748 0.0865 0.0346 0.0053 0.0266 -0.00066
[0.0951] [0.0897] [0.110] [0.0553] [0.0538] [0.0610]
Tenure 0.0534** -0.0476** -0.0524** 0.0319** -0.0271** -0.0300**
[0.0148] [0.0129] [0.0153] [0.00787] [0.00756] [0.00859]
Tenure (sq) -0.00195** 0.00157** 0.00170** -0.00113** 0.000815* 0.000941**
[0.000663] [0.000548] [0.000613] [0.000334] [0.000318] [0.000363]
Age -0.00767** 0.00663* 0.00684* -0.00379* 0.0023 0.00258
[0.00291] [0.00266] [0.00324] [0.00175] [0.00169] [0.00187]
pgsize 0.0000295 -0.000145* -0.000385** -0.00000173 -0.0000514 -0.000192**
[6.87e-05] [6.64e-05] [8.65e-05] [4.27e-05] [4.22e-05] [4.95e-05]
Constant -1.223* 0.0529 -0.973 -0.482 -0.0919 -0.838
[0.524] [0.508] [0.599] [0.426] [0.491] [0.586]
Observations 9289 9276 9262 8925 8916 8885
a
Standard errors, in brackets, are clustered by individual. All analyses also include dummies for job, function, state, and year.
† p < .10; ** p < .05; ** p < .01
 55

TABLE 8
Turnover, Advancement, and Salary using Instrumental Variablesa

Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 Model 7 Model 8 Model 9 Model 10 Model 11 Model 12
Logit Logit IV Probit IV Probit Logit Logit IV Probit IV Probit OLS OLS 2SLS 2SLS
Salary
Variables Turnover12 Turnover24 Turnover12 Turnover24 Prom24 Prom36 Prom24 Prom36 Salary (ln) Salary (ln) (ln) Salary (ln)
Posting -0.202* -0.182* -1.408* -0.459 -0.0949 0.128† -1.599† 3.060† 0.0383** 0.0376** 0.0443** 0.0430**
[0.0986] [0.0761] [0.637] [1.163] [0.0645] [0.0733] [0.833] [1.649] [0.00256] [0.00272] [0.0157] [0.0154]
Same Dept -0.0925 -0.259** -0.635* -0.37 0.0402 0.0012 -0.604† 1.067† 0.000872 0.00125 0.0039 0.00368
[0.0959] [0.0738] [0.291] [0.499] [0.0651] [0.0779] [0.358] [0.617] [0.00270] [0.00275] [0.00732] [0.00736]
Same Func -0.0336 0.0874 -0.0697 0.0811 -0.217** -0.0758 -0.291** 0.041 0.0166** 0.0152** 0.0169** 0.0155**
[0.0987] [0.0671] [0.111] [0.0855] [0.0610] [0.0756] [0.0767] [0.120] [0.00264] [0.00270] [0.00305] [0.00312]
Promotion -0.302** 0.0791 -0.0374 0.157 -0.274** -0.279** 0.095 -0.795* 0.0960** 0.0865** 0.0935** 0.0819**
[0.116] [0.0954] [0.191] [0.280] [0.0777] [0.0892] [0.207] [0.319] [0.00874] [0.00890] [0.0129] [0.0133]
Female -0.236** -0.148* -0.245* -0.157* -0.121* -0.123† -0.120† -0.114 -0.0101** -0.0912** -0.110** -0.00958**
[0.0871] [0.0648] [0.0975] [0.0664] [0.0577] [0.0666] [0.0620] [0.0836] [0.00230] [0.00235] [0.00235] [0.00241]
Tenure -0.104** -0.0828** -0.103** -0.0782** -0.00392 0.0333† 0.0000092 0.0188 -0.000281 -0.0009 -0.00042 -0.00100†
[0.0223] [0.0159] [0.0262] [0.0183] [0.0152] [0.0187] [0.0175] [0.0231] [0.000559] [0.000588] [0.000567] [0.000602]
Last salary 0.851** 0.863** -0.0181* 0.866**
[0.00924] [0.00893] [0.000844] [0.00908]
Last contr 0.00860** 0.858** 0.00948**
[0.00256] [0.00886] [0.00262]
Last comp 0.00354 0.00243
[0.00353] [0.00378]
Constant -6.111** -2.416** -0.664 -2.26 0.061 0.596 1.448 -2.657 1.561** 1.342** 1.406** 1.091**
[0.757] [0.594] [0.795] [1.420] [0.345] [0.633] [0.934] [1.996] [0.117] [0.101] [0.120] [0.104]
7451 5056 6415 4789 5178 3302 4989 3134 9292 8017 8923 7715
R-Squared 0.955 0.959 0.946 0.949
a
Standard errors, in brackets, are clustered by individual. All analyses also include dummies for job, function, state, year, age and ethnicity.
† p < .10; ** p < .05; ** p < .01
 56

TABLE 9
Effect of External Candidatesa

Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 Model 7

OLS OLS Logit Logit Logit OLS OLS
Contribution Competency Top 25% Bottom 50% Bottom 25% Salary (ln) Salary (ln)
External Candidate -0.039 -0.0535 -0.0588 -0.192 -0.186 -0.00683 -0.00537
[0.0441] [0.0377] [0.171] [0.170] [0.247] [0.00762] [0.00706]
Level 4 -0.972** -0.365† 1.005 -0.494 0.738 0.0533 0.242*
[0.344] [0.218] [1.252] [1.244] [1.841] [0.0377] [0.105]
Level 5 -0.406 0.239 0.7 -0.313 1.535 0.106* 0.271*
[0.452] [0.248] [1.131] [1.106] [1.573] [0.0523] [0.124]
Level 6 -2.000** -0.855** 0.369 -0.783 1.026 0.177** 0.257**
[0.360] [0.233] [0.716] [0.778] [1.074] [0.0559] [0.0704]
Same Department 0.137** 0.104** 0.577** -0.366† -0.278 0.0155* 0.0220**
[0.0493] [0.0393] [0.187] [0.188] [0.282] [0.00786] [0.00775]
Same Function 0.0165 0.0582 0.0209 0.0503 -0.420† 0.0226** 0.0161†
[0.0449] [0.0385] [0.189] [0.185] [0.250] [0.00855] [0.00878]
Advancement 0.0173 0.0394 0.00756 0.0149 0.117 0.0187 0.0311
[0.0686] [0.0567] [0.267] [0.278] [0.420] [0.0408] [0.0424]
Female 0.00303 0.0419 0.228 -0.195 -0.323 -0.0104 -0.00521
[0.0426] [0.0346] [0.169] [0.168] [0.241] [0.00800] [0.00832]
Black -0.0867 -0.064 -0.16 0.136 0.593† -0.0121 -0.0094
[0.0733] [0.0601] [0.268] [0.247] [0.326] [0.00876] [0.00887]
pgsize 0.000636** -0.000620** -0.00110**
[0.000200] [0.000197] [0.000316]
Constant 3.116** 2.265** -2.869* 2.485† 0.991 1.568** 1.809**
[0.301] [0.156] [1.385] [1.332] [1.934] [0.405] [0.522]
Observations 869 869 869 869 869 869 752
R-squared 0.168 0.178 0.955 0.96
a
Standard errors, in brackets, are clustered by individual. All analyses also include dummies for job, function, state, year, and age.
† p < .10; ** p < .05; ** p < .01
57
 Subscribe to DeepL Pro to translate larger documents.
Visit www.DeepL.com/pro for more information.
1

PASAR INTERNAL, JARINGAN PRIBADI, DAN ALOKASI SUMBER DAYA MANUSIA:

BAGAIMANA MANAJER MENCIPTAKAN NILAI MELALUI PEREKRUTAN
INTERNAL

JR Keller
Kertas Kerja Wharton School of
Business

ABSTRAK

Perekrutan internal mencocokkan karyawan yang ada saat ini dengan pekerjaan baru di dalam

sebuah organisasi dan merupakan sumber penciptaan nilai yang penting namun sering diabaikan,

sehingga memungkinkan para manajer untuk menghasilkan nilai yang lebih besar dari stok sumber

daya manusia yang ada dengan menciptakan kecocokan yang saling melengkapi antara orang dan

pekerjaan. Namun, terlepas dari kenyataan bahwa lebih dari separuh pekerjaan diisi secara

internal, pengetahuan kita tentang bagaimana pekerja dialokasikan ke pekerjaan di dalam

organisasi tetap didasarkan pada penelitian tentang pasar tenaga kerja internal yang birokratis dan

karier intraorganisasi yang menggambarkan pasar tenaga kerja internal yang tidak memiliki

banyak kemiripan dengan pasar tenaga kerja kontemporer. Dalam makalah ini, saya

menggambarkan efek dari dua proses yang muncul untuk menggantikan penggunaan aturan

birokrasi dalam memfasilitasi mobilitas internal - penempatan yang berorientasi pada pasar dan

sponsor yang berorientasi pada hubungan - pada dua set hasil yang berhubungan langsung dengan

penciptaan nilai dan penangkapan nilai - kualitas perekrutan dan kompensasi. Dengan

menggunakan data lebih dari 11.000 perekrutan internal yang dilakukan selama periode lima tahun

di sebuah perusahaan asuransi kesehatan besar di Amerika Serikat, saya menemukan bahwa

perekrutan yang berorientasi pada pasar menghasilkan perekrutan yang lebih baik, namun dengan

biaya yang lebih tinggi. Selain memberikan gambaran yang lebih lengkap mengenai perekrutan

dan mobilitas, penelitian ini menyoroti tradeoff yang terkait dengan penggunaan pasar dan

jaringan untuk mengalokasikan sumber daya di dalam perusahaan.

 2

Di tengah perbincangan mengenai karier tanpa batas, pekerja bebas, dan pekerja upahan,

mudah untuk melupakan bahwa sejumlah besar mobilitas masih terjadi di dalam organisasi.

Faktanya, hampir setengah dari semua pekerjaan yang terbuka di organisasi besar - dan lebih

banyak lagi di tingkat eksekutif - diisi secara internal (misalnya Crispin & Mehler, 2013).

Perekrutan internal adalah proses utama yang digunakan untuk mengalokasikan sumber daya

manusia di dalam perusahaan dan menciptakan peluang untuk mobilitas internal, yang memiliki

banyak fungsi yang berguna. Hal ini memfasilitasi transfer pengetahuan melintasi batas-batas

internal (Argote & Ingram, 2000), memotivasi karyawan dengan memberi sinyal peluang untuk

kemajuan di masa depan (Bidwell & Keller, 2014), mendorong pengembangan keterampilan

khusus perusahaan (Campion, Cheraskin, & Stevens, 1994), meningkatkan kepuasan pekerja,

kinerja dan produktivitas (Jackson, 2013), dan mengurangi perputaran disfungsional (Allen,

Bryant, & Vardaman, 2010). Dengan sumber daya manusia yang kini menjadi sumber daya

terpenting di sebagian besar perusahaan (Powell & Snellman, 2004) dan dengan adanya penelitian

terbaru yang menunjukkan tingginya biaya dan tingkat kegagalan yang lebih tinggi terkait dengan

perekrutan dari luar di semua tingkat organisasi (misalnya Bidwell, 2011; Groysberg, Lee, &

Nanda, 2008), maka kemampuan untuk menemukan dan menciptakan kecocokan yang saling

melengkapi antara orang dan pekerjaan di dalam perusahaan menjadi sumber utama dari

penciptaan nilai di dalam organisasi modern (Zenger, Felin, & Bigelow, 2011).

Meskipun para ahli telah mengeksplorasi bagaimana sumber daya utama lainnya, seperti

modal keuangan (misalnya Stein, 1997) dan perhatian manajerial (misalnya Ocasio, 1997),

dialokasikan di dalam perusahaan, secara mengejutkan kita hanya mengetahui sedikit sekali

mengenai alokasi sumber daya manusia secara internal. Penelitian terbaru telah

mendokumentasikan bagaimana karakteristik pekerjaan membentuk apakah suatu pekerjaan akan

diisi secara internal (Bidwell & Keller, 2014) dan kualifikasi (misalnya pengalaman) dari para

kandidat yang akan ditempatkan dalam pekerjaan tersebut (Drazin & Rao, 2002). Namun, proses

perekrutan internal - yaitu, cara-cara yang digunakan oleh para manajer untuk mencari,
 3

mengevaluasi, dan memilih di antara kandidat-kandidat internal yang potensial - belum banyak

mendapat perhatian sistematis. Namun, penelitian telah menunjukkan bahwa penggunaan proses

rekrutmen eksternal formal versus informal

 4

Proses perekrutan (misalnya, penggunaan referensi versus lowongan pekerjaan) tidak hanya

membentuk siapa yang dipekerjakan, tetapi juga gaji, kinerja, dan perputaran mereka (Burks,

Cowgill, Hoffman, & Housman, 2013; Fernandez, Castilla, & Moore, 2000; Seidel, Polzer, &

Stewart, 2000), seperti halnya penelitian baru-baru ini yang membandingkan perekrutan internal

dan eksternal (Agrawal, Knoeber, & Tsoulouhas, 2006; Bidwell, 2011). Oleh karena itu, kami

memperkirakan bahwa setiap variasi di antara proses perekrutan internal memiliki konsekuensi

yang sama pentingnya bagi pekerja dan perusahaan.

Mempelajari perekrutan internal juga menjanjikan untuk memberikan kontribusi wawasan

penting untuk percakapan yang sedang berlangsung tentang perubahan sifat alokasi sumber daya

internal secara umum. Ketika perusahaan telah beralih dari struktur hirarkis yang ditandai dengan

pengambilan keputusan terpusat dan menuju struktur yang lebih datar dan lebih ramping yang

ditandai dengan pengambilan keputusan yang terdesentralisasi, pasar tenaga kerja internal yang

birokratis secara bertahap hancur (Cappelli & Keller, 2014), pasar internal dan jaringan sosial

telah muncul sebagai mekanisme utama di mana para pekerja yang ada saat ini dicocokkan dengan

pekerjaan baru, yang menggemakan perubahan yang lebih luas tentang bagaimana sumber daya

dialokasikan di dalam organisasi. Memang, kegagalan sistem perencanaan birokrasi di

perusahaan-perusahaan kontemporer telah mendapat banyak perhatian (Cowen & Parker, 1997;

McEvily, Soda, & Tortoriello, 2014; Mintzberg, 1994), dengan penelitian terbaru yang

mengidentifikasi pasar internal dan bentuk-bentuk jaringan koordinasi sebagai dua pengganti

utama kontrol birokrasi atas alokasi sumber daya utama. Penelitian tentang pasar internal secara

umum menekankan bagaimana penggunaan harga pasar di dalam perusahaan dapat meningkatkan

pengambilan keputusan manajerial terkait alokasi sumber daya, karena semua informasi terkait

sumber daya yang dipertimbangkan tercermin dalam harganya (Ellig, 2001; Felin & Zenger,

2011). Penelitian lain tentang pasar internal telah mengeksplorasi bagaimana insentif berkekuatan

tinggi dapat dirancang untuk mengurangi biaya koordinasi dengan menyelaraskan kepentingan

manajer yang membuat keputusan alokatif dengan kepentingan perusahaan (Zenger & Hesterly,
 5

1997; Zenger, 2002). Penelitian tentang bentuk-bentuk jaringan koordinasi telah menekankan

bagaimana hubungan sosial

 6

Di antara para pelaku di dalam perusahaan dapat mengarah pada keputusan alokatif yang lebih

baik dengan memberikan kesempatan kepada para manajer untuk berbagi informasi mengenai

sumber daya yang tidak tersedia bagi para pengambil keputusan yang tersentralisasi dan berada di

tingkat yang lebih tinggi (Nahapiet & Ghoshal, 1998; Tsai & Ghoshal, 1998). Karena pasar

internal dan jaringan sosial telah disajikan sebagai alternatif dari kontrol birokrasi, banyak literatur

yang ada saat ini telah membandingkan pasar internal dengan kontrol birokrasi (misalnya Stein,

1997) atau membandingkan jaringan sosial dengan kontrol birokrasi (misalnya Tsai, 2002). Masih

sedikit penelitian yang mengeksplorasi tradeoff yang terkait dengan penggunaan pasar internal

atau jaringan pribadi untuk mengalokasikan sumber daya internal, terlepas dari kenyataan bahwa

mekanisme pasar dan jaringan sosial cenderung beroperasi secara simultan di dalam perusahaan.

Pasar tenaga kerja internal kontemporer merupakan konteks yang sangat bermanfaat untuk

mengeksplorasi pertukaran ini. Sebagian besar pemahaman kita tentang perekrutan dan mobilitas

internal didasarkan pada penelitian dasar tentang pasar tenaga kerja internal tradisional yang

hirarkis dan literatur yang terkait erat dengan karier intraorganisasi, yang menarik perbedaan

tajam antara proses birokrasi untuk mengalokasikan sumber daya manusia yang beroperasi di

dalam perusahaan dan proses pasar yang beroperasi di luar perusahaan. Penelitian ini

menggambarkan bagaimana keputusan perekrutan internal dipusatkan di kantor personalia dan

diatur oleh aturan birokrasi yang ketat yang digunakan untuk mempertahankan garis

perkembangan di sepanjang jenjang pekerjaan yang jelas, dengan karyawan yang hanya memiliki

sedikit kontrol atas karier mereka di dalam perusahaan (Althauser & Kalleberg, 1981; Glaser,

1968; Diprete, 1987; Doeringer & Piore, 1971; Rosenbaum, 1990). Pasar tenaga kerja internal saat

ini terlihat sangat berbeda: keputusan perekrutan sebagian besar telah didesentralisasi, dengan

otoritas atas promosi, transfer, dan perekrutan eksternal yang didelegasikan kepada masing-

masing manajer; penundaan organisasi, deskripsi pekerjaan yang lebih luas, dan munculnya

pekerjaan berbasis proyek telah menghilangkan jalur yang jelas untuk kemajuan; dan karyawan

telah ditugaskan untuk mengambil alih kendali atas karier mereka. Efek kumulatif dari perubahan-
 7

perubahan ini telah membuat penggunaan aturan birokrasi untuk mengalokasikan sumber daya

manusia menjadi usang (Piore, 2002).

 8

Tanpa adanya aturan birokrasi, alokasi pekerja untuk pekerjaan sekarang terjadi melalui

dua proses yang sangat berbeda (Marsden & Gorman, 1999; Pinfield, 1995). Penempatan adalah

proses yang berorientasi pada pasar. Seorang manajer menciptakan pasar internal untuk sebuah

pekerjaan yang terbuka dengan menyiarkan informasi tentang posisi tersebut ke seluruh organisasi

melalui papan lowongan kerja internal dan mengundang karyawan yang ada saat ini untuk

melamar. Penggunaan sistem lowongan kerja internal tersebar luas, dengan 95% organisasi

memposting lowongan kerja secara internal (Taleo Research, 2005). Sponsorship adalah proses

yang lebih bersifat relasional. Seorang manajer mengidentifikasi seorang kandidat melalui

jaringan pribadinya dan menunjuk kandidat tersebut untuk pekerjaan tersebut tanpa

mempertimbangkan kandidat lain secara formal. Tidak hanya posting dan sponsorship merupakan

dua proses perekrutan internal yang paling sering digunakan, keduanya sering kali berjalan

bersamaan di dalam perusahaan, karena para manajer diberikan keleluasaan yang cukup besar

dalam mengambil keputusan perekrutan dan proses perekrutan (Marsden & Gorman, 1999;

Pinfield, 1995). Dengan sedikitnya penelitian yang mengkaji perekrutan internal dalam organisasi

kontemporer, tidak jelas apakah proses alokatif yang berorientasi pada pasar atau berorientasi

pada hubungan lebih mungkin menghasilkan kecocokan internal yang lebih berharga dan apa saja

pengorbanannya, jika ada, yang mungkin terkait dengan penggunaan posting versus sponsor.

Dalam makalah ini, saya menjelaskan tentang pengorbanan ini dengan mengembangkan

teori untuk menjelaskan bagaimana perbedaan utama antara proses posting yang berorientasi pada

pasar dan proses sponsorship yang berorientasi pada hubungan mempengaruhi dua hasil yang

berimplikasi pada penciptaan nilai dan penangkapan nilai: kualitas perekrutan - seperti yang

terungkap dari kinerja pekerjaan, pergantian, dan kemajuan selanjutnya - dan kompensasi.

Meskipun manfaat informasi yang terkait dengan jejaring sosial telah mendapat perhatian yang

cukup besar (Borgatti & Cross, 2003; Bradach & Eccles, 1989; Cross, Parker, Prusak, & Borgatti,

2001; Granovetter, 1985), saya memperkirakan bahwa proses posting yang berorientasi pada pasar

akan meningkatkan pengambilan keputusan manajerial dibandingkan dengan proses sponsorship

 9

yang lebih bersifat relasional, sehingga menciptakan lebih banyak nilai melalui perekrutan internal

yang lebih berkualitas. Dengan demikian, saya menggambarkan bagaimana posting membawa dua

fitur pasar ke dalam perusahaan yang sebagian besar tidak ada dalam sponsorship - seleksi

mandiri dan formalitas. The

 10

Teori perilaku perusahaan menyoroti dua tantangan utama yang dihadapi para pengambil

keputusan yang rasional, yaitu mengidentifikasi serangkaian alternatif dan mengevaluasi

konsekuensi dari alternatif-alternatif tersebut (Cyert & March, 1963; Simon, 1955). Saya

berpendapat bahwa berbeda dengan pencarian manajerial aktif yang diperlukan dalam

sponsorship, mengizinkan pekerja untuk memilih sendiri ke dalam kumpulan pertimbangan untuk

pekerjaan terbuka mengurangi kemungkinan kandidat internal yang luar biasa akan terlewatkan.

Selain itu, jika dibandingkan dengan informalitas relatif dari proses alokatif yang berorientasi

pada hubungan, formalitas pasar akan mendorong para manajer untuk mencari informasi yang

memungkinkan mereka untuk mengevaluasi kecocokan antara kualifikasi kandidat dengan

persyaratan pekerjaan, sementara juga membatasi penggunaan informasi yang tidak relevan yang

dapat menyebabkan keputusan perekrutan yang buruk.

Namun, saya juga memprediksi bahwa fitur pasar yang sama ini akan mengarah pada gaji

yang lebih tinggi melalui pengaruhnya terhadap negosiasi gaji, sehingga membatasi jumlah nilai

yang dapat ditangkap oleh perusahaan melalui keputusan perekrutan internal yang lebih baik.

Seleksi mandiri dan formalitas memperkenalkan kompetisi terbuka ke dalam proses perekrutan

internal. Meskipun persaingan sering kali menurunkan harga di pasar eksternal, saya berpendapat

bahwa hal ini akan meningkatkan harga sumber daya manusia di dalam perusahaan, karena

kandidat yang dipekerjakan melalui proses yang kompetitif akan lebih mungkin untuk memulai

dan mengadopsi pendekatan yang kompetitif dalam negosiasi gaji.

Secara keseluruhan, prediksi ini menunjukkan bahwa ada pengorbanan penting yang

terkait dengan pengalokasian sumber daya manusia secara formal melalui pasar atau secara

informal melalui penggunaan jaringan sosial manajer, dengan posting yang menghasilkan

perekrutan yang lebih baik tetapi dengan biaya yang lebih tinggi. Namun, dengan manfaat kinerja

dan retensi yang terkait dengan perekrutan internal yang lebih baik yang mungkin jauh melebihi

biaya gaji yang lebih tinggi, posting cenderung memungkinkan perusahaan untuk menciptakan dan

menangkap nilai yang jauh lebih besar. Argumen ini menunjukkan bahwa proses yang berorientasi
 11

pada pasar juga akan menguntungkan pekerja, yang juga menangkap nilai lebih melalui gaji yang

lebih tinggi.
 12

Saya menggunakan beberapa strategi pemodelan dalam menguji prediksi ini, yang

sebagian besar didukung dengan menggunakan catatan personalia selama lima tahun (2008-2012)

yang mencakup semua karyawan dari penyedia asuransi kesehatan besar, serta data lebih dari

350.000 lamaran pekerjaan internal dan eksternal.

Meskipun catatan personalia sebelumnya telah digunakan untuk mengidentifikasi karyawan mana

yang berpindah ke pekerjaan mana (Bidwell, 2011; Dencker, 2008), data tentang proses

perpindahan karyawan masih jarang ditemukan. Data ini berbeda karena dengan jelas

mengidentifikasi mekanisme yang digunakan untuk memfasilitasi masing-masing dari 11.000+

perekrutan internal yang dilakukan selama periode ini, sehingga memungkinkan saya untuk

melakukan apa yang, sepengetahuan saya, merupakan studi terperinci pertama tentang perbedaan

hasil yang terkait dengan dua proses perekrutan internal yang sangat berbeda ini.

Dalam membongkar proses yang digunakan untuk mengalokasikan pekerja ke pekerjaan-

pekerjaan di dalam organisasi kontemporer, studi ini membantu memberikan pemahaman yang

lebih lengkap tentang pasar tenaga kerja dan mobilitas. Sementara literatur yang kuat berkembang

untuk mengeksplorasi hasil yang terkait dengan proses perekrutan eksternal yang berbeda

(misalnya Fernandez et al., 2000) serta pengorbanan biaya dan kualitas yang terkait dengan

perekrutan internal versus eksternal (misalnya Bidwell, 2011), penelitian ini cenderung

memperlakukan perekrutan internal sebagai proses yang homogen, yang secara tidak sengaja

mengaburkan potensi variasi dalam proses yang digunakan untuk memindahkan pekerja ke

pekerjaan baru di perusahaan. Secara lebih luas, penelitian ini berkontribusi pada literatur yang

terus berkembang yang mengeksplorasi bagaimana sumber daya digabungkan dan digunakan

dalam organisasi kontemporer (Sirmon, Gove, & Hitt, 2008; Sirmon & Hitt, 2009) dengan

menyoroti tradeoff yang terkait dengan penggunaan pendekatan pasar versus pendekatan jaringan

untuk mengalokasikan sumber daya internal. Studi ini memperluas penelitian terbaru yang

mengeksplorasi efek dari memperkenalkan mekanisme pasar ke dalam perusahaan (Felin &

Zenger, 2011) dengan menunjukkan bagaimana mekanisme selain harga dan insentif berdaya
 13

tinggi dapat dimanfaatkan untuk meningkatkan pengambilan keputusan manajerial, sementara

juga menambah koleksi studi kecil namun penting yang menyoroti keterbatasan potensial dari

pertukaran relasional (Casciaro & Lobo, 2008; Rogan & Sorenson, 2014; Sorenson &

Waguespack, 2006).
 14

Selain itu, dalam mengidentifikasi mekanisme tingkat mikro (pengambilan keputusan dan

negosiasi) yang melaluinya kedua proses organisasi yang berbeda ini membentuk hasil yang

terkait langsung dengan penciptaan nilai dan penangkapan nilai (kualitas perekrutan dan

kompensasi), penelitian ini berkontribusi pada literatur yang sedang berkembang yang berfokus

pada identifikasi dasar-dasar mikro keunggulan kompetitif berbasis sumber daya manusia (Coff &

Kryscynski, 2011; Ployhart & Hale, 2014).

TEORI & HIPOTESIS

Perekrutan Internal dalam Organisasi Kontemporer

Sebelum mengembangkan teori, ada baiknya kita mendefinisikan perekrutan internal dan

menjelaskan fitur-fitur utama dari proses alokasi yang menjadi fokus penelitian ini - penempatan

dan sponsor. Setelah sebelumnya menjadi bagian dari kantor personalia yang tersentralisasi, para

manajer kini sebagian besar bertanggung jawab untuk mengisi pekerjaan-pekerjaan yang berada

di bawah pengawasan langsung mereka, karena mereka telah diberi keleluasaan yang cukup besar

untuk menentukan kandidat mana yang akan dipertimbangkan dan siapa yang pada akhirnya akan

dipekerjakan (Jacoby, 2004). Perekrutan internal terjadi ketika seorang manajer mengisi pekerjaan

yang terbuka dengan mempekerjakan pekerja yang saat ini dipekerjakan oleh organisasi dalam

pekerjaan yang berbeda, yang mengakibatkan realokasi pekerja tersebut ke serangkaian kegiatan

organisasi yang baru. Penting untuk dicatat bahwa perekrutan internal secara konseptual berbeda

dengan restrukturisasi perusahaan, di mana kelompok besar pekerja - sering kali seluruh

departemen atau lini bisnis - dipindahkan secara massal ke produk atau pasar baru (Capron,

Dussauge, & Mitchell, 1998; London, 1996).

Sistem penempatan kerja internal sudah ada setidaknya sejak tahun 1940-an. Sistem ini

awalnya dilembagakan atas desakan serikat pekerja, yang melihatnya sebagai cara untuk

membatasi kebijaksanaan manajerial atas keputusan kepegawaian dan dengan demikian

membatasi ruang lingkup diskriminasi, pilih kasih, dan nepotisme yang memengaruhi peluang

untuk kemajuan internal. Manajemen melihat sistem ini mengharuskan mereka melepaskan hak
 15

yang sudah lama dipegang, sehingga mereka berjuang untuk membatasi cakupannya. Akibatnya,

sistem penempatan kerja awal biasanya mencakup serangkaian pekerjaan yang terbatas dan

menempatkan pembatasan yang signifikan pada karyawan yang diizinkan untuk melamar

pekerjaan tersebut (Cappelli, 2008). Untuk pekerjaan-pekerjaan yang

 16

tertutup, kriteria seleksi yang terperinci ditetapkan dan sering kali mengutamakan senioritas

daripada kemampuan, sehingga manajer individu tidak memiliki banyak pilihan tentang siapa

yang akan dipekerjakan (Jacoby, 1985; Slichter, Healy, & Livernash, 1960).

Sistem posting pekerjaan kontemporer jauh lebih luas. Manajer memposting informasi

tentang lowongan pekerjaan yang berada di semua tingkat organisasi kecuali tingkat paling atas ke

papan pekerjaan internal dan mengundang kandidat yang berminat yang berada di seluruh

organisasi untuk melamar dan bersaing untuk mendapatkan pekerjaan tersebut, yang

mencerminkan proses umum yang digunakan untuk merekrut dari luar1 . Meskipun persyaratan

pekerjaan cenderung memengaruhi siapa yang melamar, kandidat yang tidak memiliki kualifikasi

yang dijelaskan dalam lowongan pekerjaan namun ingin dipertimbangkan masih dapat

mengajukan lamaran. Manajer perekrutan akan mengevaluasi lamaran dan memberikan

penawaran kepada kandidat yang mereka sukai, yang mungkin akan menegosiasikan persyaratan

sebelum memutuskan untuk menerima atau menolak pekerjaan tersebut. Setelah tawaran diterima,

biasanya kandidat internal yang tidak berhasil akan diberitahu dan diberi kesempatan untuk

mengetahui mengapa mereka tidak mendapatkan pekerjaan tersebut (Miller, 1984; Pinfield, 1995).

Pengumuman sering kali merupakan proses perekrutan internal yang baku. Lebih dari 95%

perusahaan sekarang menggunakan papan lowongan kerja internal (Taleo Research, 2005) dan

banyak yang telah mengembangkan kebijakan yang mendorong para manajer untuk memposting

lowongan kerja secara internal, sebagian besar untuk melindungi diri mereka sendiri dari tuduhan

diskriminasi (Strum, 2001). Namun, terlepas dari kebijakan-kebijakan ini dan ketersediaan papan

lowongan kerja internal yang tersebar luas, para manajer dapat dan sering kali mengabaikan sistem

lowongan kerja. Hal ini sebagian disebabkan oleh fakta bahwa

bahwa tidak ada undang-undang negara bagian atau federal yang mewajibkan perusahaan untuk
memposting lowongan pekerjaan dan sebagian karena kebijakan kepegawaian internal

1Ketika memposting sebuah pekerjaan, seorang manajer memiliki pilihan untuk membatasi kompetisi
hanya untuk kandidat internal atau membuka kompetisi untuk kandidat internal dan eksternal. Meskipun
sulit untuk menentukan angka pastinya, perkiraan yang masuk akal adalah sekitar setengah dari pekerjaan
di atas level pemula (yang menurut definisi diisi secara eksternal) diposting secara internal dan eksternal
 17

pada saat yang bersamaan (Jacoby, 2005). Ini berarti bahwa proses posting bisa jadi menghasilkan
perekrutan internal atau tidak, tergantung siapa yang dipilih. Untuk tujuan makalah ini, saya menyajikan
posting sebagai proses alokatif internal, karena ketertarikan saya adalah untuk melihat hasil yang terkait
dengan posting ketika perekrutan internal dilakukan. Dalam analisis empiris tambahan, saya mengontrol
keberadaan kandidat eksternal pada variabel dependen utama.
 18

keduanya jarang ditegakkan dan memungkinkan adanya interpretasi yang fleksibel (Pinfield,

1995). Bahkan, Society of Human Resource Management, asosiasi profesional terkemuka bagi

para profesional sumber daya manusia, telah merekomendasikan kepada para manajer perekrutan

bahwa "meskipun Anda memiliki proses internal untuk mengumumkan lowongan pekerjaan yang

tersedia, mungkin ada kalanya Anda memutuskan untuk tidak mengikuti proses ini" (SHRM,

2000:7). Secara khusus, para manajer cenderung mengabaikan sistem posting ketika mereka

sudah memiliki kandidat yang memenuhi syarat untuk pekerjaan tersebut.

Ketika para manajer melewati proses posting, mereka mengubah proses relasional yang

saya sebut sebagai sponsorship2 . Jejaring sosial adalah fitur utama dalam kehidupan organisasi,

karena interaksi antar individu pasti mengarahkan pekerja untuk mengembangkan jaringan

hubungan pribadi (Kanter, 1977; McEvily et al., 2014; Podolny & Baron, 1997). Seorang manajer

yang mensponsori menggunakan jaringan sosial pribadinya untuk mengidentifikasi kandidat

potensial dan kemudian menunjuk kandidat yang ia sukai untuk pekerjaan tersebut jika tidak ada

kompetisi terbuka. Meskipun mungkin saja manajer yang mensponsori menguras koneksi pribadi

mereka saat mencari kandidat potensial, sebagian besar perekrutan internal yang dilakukan

melalui sponsorship melibatkan pertimbangan satu kandidat yang memiliki hubungan langsung

dengan manajer perekrutan, biasanya seorang rekan kerja saat ini atau sebelumnya (Pinfield,

1995).

Banyak bukti yang menunjukkan bahwa posting dan sponsorship berjalan berdampingan di

dalam perusahaan, sebagai cara yang sama baiknya untuk mengisi lowongan pekerjaan. Marsden

dan Gorman (1999) meneliti data survei pada sampel yang representatif dari perusahaan-

perusahaan di Amerika Serikat dan menemukan bahwa pengumuman dan sponsorship digunakan

secara luas dalam kombinasi untuk mengisi lowongan pekerjaan dengan kandidat internal. Studi

Pinfield (1995) tentang perekrutan internal di sebuah perusahaan produk hutan mengungkapkan

bahwa lebih dari separuh dari semua posisi diisi melalui sponsor meskipun ada kebijakan

perusahaan yang mengharuskan semua pekerjaan diumumkan. Dan meskipun lebih dari 95 persen
 19

perusahaan melaporkan bahwa mereka memposting setidaknya beberapa lowongan pekerjaan

secara internal (Taleo Research, 2005)

2 Saya mengadopsi istilah sponsorship dari literatur klasik tentang mobilitas karier dalam organisasi
(Rosenbaum, 1979) dan mobilitas sosial ke atas secara umum (Turner, 1960), di mana istilah ini
digunakan untuk menggambarkan sistem di mana individu yang dipilih untuk maju terlindung dari
persaingan.
 20

Penelitian terus menunjukkan pentingnya hubungan pada kemajuan internal (Bielby, 2000;

Ibarra, 1993; Podolny & Baron, 1997).

Penting untuk dicatat bahwa meskipun posting dan sponsorship merupakan proses alokasi

internal yang berbeda secara konseptual, namun dalam praktiknya, sponsorship terkadang

menyamar sebagai posting. Artinya, ada kemungkinan bahwa seorang manajer dapat memposting

sebuah pekerjaan setelah mengidentifikasi kandidat yang akan mereka pilih melalui jaringan sosial

mereka, sebuah praktik yang disebut sebagai "pencarian melalui kabel" (Bielby, 2000).

Namun, prevalensi "pencarian melalui kabel" dan praktik-praktik lain yang secara artifisial dapat

membatasi keterbukaan proses perekrutan, seperti membentuk persyaratan pekerjaan di sekitar

satu kandidat atau mencegah karyawan untuk melamar pekerjaan tertentu, sebagian besar

dimitigasi oleh kekhawatiran akan tuduhan diskriminasi (Strum, 2001) dan juga kemungkinan

bahwa karyawan yang merasa diperlakukan tidak adil atau disesatkan akan keluar dari organisasi

(Billsberry, 2007; Cappelli, 2008).

Dari uraian di atas, jelaslah bahwa sponsorship yang berorientasi pada pasar dan

sponsorship yang berorientasi pada hubungan berbeda dalam beberapa hal. Dalam

mengembangkan teori untuk memahami pertukaran antara kedua proses alokatif ini, saya

menyoroti fakta bahwa penempatan ditandai dengan dua fitur yang mirip pasar, yaitu seleksi

mandiri dan formalitas, yang sebagian besar tidak ada dalam sponsorship, yang justru

melibatkan pencarian manajerial yang aktif dan menggunakan hubungan personalia untuk

mengumpulkan informasi. Rasionalitas yang Terbatas dan Kualitas Perekrutan

Aspek utama dari teori perilaku perusahaan adalah anggapan bahwa pengambil keputusan

adalah rasional. Karena para pengambil keputusan memiliki keterbatasan kognitif dan memiliki

waktu, informasi, dan sumber daya yang terbatas, maka tidak ada satu pun dari sekian banyak

alternatif yang dapat dipilih oleh para pengambil keputusan yang dapat diketahui sebelumnya,

begitu pula dengan konsekuensi-konsekuensi yang terlibat dalam pemilihan di antara alternatif-

alternatif yang tersedia (March & Simon, 1958). Akibatnya, kegagalan dalam menghasilkan dan
 21

mengevaluasi alternatif telah diidentifikasi sebagai dua alasan utama mengapa manajer gagal

mengalokasikan sumber daya yang tersedia secara optimal (Afuah & Tucci, 2012; Knudsen &

Levinthal, 2007).
 22

Seleksi mandiri dan menghasilkan alternatif. Karena rangkaian alternatif yang lengkap

tidak diketahui secara ex ante, para pengambil keputusan yang rasional harus melakukan

pencarian untuk menghasilkan alternatif. Namun, pencarian tidak bebas biaya, dan salah satu cara

pengambil keputusan menghemat biaya adalah dengan mempertimbangkan hanya sebagian kecil

dari alternatif yang tersedia (Hauser & Wernerfelt, 1990; Shocker, Ben-Akiva, Boccara, &

Nedungadi, 1991). Meskipun mempertimbangkan lebih banyak alternatif tidak menjamin

keputusan yang lebih baik (Iyengar & Kamenica, 2010), keberhasilan pengambilan keputusan

cenderung meningkat ketika lebih banyak alternatif yang dipertimbangkan (Alexander, 1979;

Gemünden & Hauschildt, 1985; Nutt, 1998), sebagian karena hal tersebut mengurangi

kemungkinan bahwa alternatif yang lebih unggul akan ditinggalkan dalam pertimbangan.

Seleksi mandiri memungkinkan manajer untuk menghasilkan lebih banyak alternatif tanpa

mengeluarkan banyak biaya yang terkait dengan pencarian yang lebih luas (Afuah & Tucci,

2012). Pasar memfasilitasi alokasi sumber daya dengan menyediakan platform umum bagi

pembeli dan penjual yang tersebar luas untuk bertukar informasi mengenai kebutuhan dan

preferensi mereka (Zenger et al., 2011). Seleksi mandiri mengacu pada kemampuan pelaku pasar

untuk memilih peluang yang tersedia untuk dikejar berdasarkan informasi ini, bukannya

ditentukan oleh otoritas manajerial.

Salah satu cara manajer (sebagai pembeli) dapat memanfaatkan kekuatan seleksi mandiri

adalah dengan menyiarkan informasi tentang suatu peluang dan memungkinkan penjual yang

berminat untuk memilih sendiri ke d a l a m kumpulan alternatif yang akan dipertimbangkan oleh

manajer. Daripada manajer memikul tanggung jawab untuk menghasilkan alternatif melalui

pencarian aktif, penjual mencari peluang yang sesuai dengan preferensi mereka. Karena penjual

memiliki lebih banyak informasi tentang preferensi mereka daripada manajer dan cenderung

mencari alternatif yang memenuhi preferensi tersebut, pencarian yang mungkin dianggap jauh dari

sudut pandang manajer mungkin sering dianggap lokal dari sudut pandang penjual. Dengan

mengubah pencarian lokal menjadi pencarian jauh (Afuah & Tucci, 2012), seleksi mandiri
 23

diharapkan dapat memperluas jumlah alternatif yang dipertimbangkan oleh manajer - dan lebih

banyak alternatif mengurangi risiko bahwa alternatif yang berkualitas akan terlewatkan.
 24

Proses posting memungkinkan manajer untuk memanfaatkan kekuatan seleksi mandiri

dengan menyiarkan informasi tentang lowongan pekerjaan dan memungkinkan kandidat internal

yang berada di seluruh organisasi, termasuk mereka yang berada di daerah yang lebih jauh, seperti

pekerja yang berada di lokasi, departemen, atau fungsi yang berbeda, untuk memilih sendiri ke

dalam antrean kandidat yang bersaing untuk mendapatkan pekerjaan tersebut. Sebaliknya,

sponsorship tidak menyediakan mekanisme formal bagi karyawan untuk menyatakan minat

mereka terhadap suatu lowongan pekerjaan. Alih-alih menyiarkan informasi tentang lowongan

dan memungkinkan kandidat yang berminat untuk memilih sendiri ke dalam kumpulan kandidat,

manajer secara aktif mencari alternatif (Mintzberg, Raisinghani, & Théorêt, 1976) melalui

jaringan pribadi mereka. Oleh karena itu, kumpulan kandidat potensial dibatasi oleh pengalaman

dan koneksi manajer sebelumnya. Meskipun informasi tentang kandidat internal yang berada di

luar jaringan langsung manajer mungkin dapat diakses (misalnya melalui sistem informasi sumber

daya manusia, berbicara dengan SDM, dll.), namun untuk mendapatkan informasi ini memerlukan

waktu dan usaha dan kemungkinan untuk menemukan alternatif yang lebih unggul tidak pasti.

Selain itu, para manajer lebih cenderung memberikan nilai yang lebih tinggi pada informasi yang

diperoleh melalui jaringan sosial mereka, sehingga membuat mereka enggan untuk mencari

kandidat yang berada di luar jaringan tersebut. Akibatnya, manajer yang mensponsori biasanya

hanya mempertimbangkan kandidat yang sudah mereka kenal, dan bahkan manajer yang paling

terhubung dengan baik di organisasi besar pun tidak mungkin mengenal semua kandidat internal

yang potensial.

Singkatnya, dengan memungkinkan kandidat yang berada di dalam dan di luar jaringan

sosial manajer perekrutan untuk memilih sendiri ke dalam kumpulan pertimbangan, maka posting

diharapkan dapat menghasilkan kumpulan kandidat yang lebih besar daripada yang bisa diakses

melalui jaringan pribadi manajer, sehingga mengurangi kemungkinan kandidat internal yang luar

biasa terlewatkan.

Formalitas dan mengevaluasi alternatif. Selain tantangan yang terkait dengan

 25

menghasilkan alternatif, manajer yang rasional terbatas juga menghadapi kesulitan dalam

mengevaluasi alternatif. Secara khusus, penelitian terbaru telah memberikan perhatian pada

masalah yang terkait dengan kesadaran terbatas,

 26

yang mengacu pada kecenderungan individu yang rasional terbatas untuk gagal mencari atau

memasukkan informasi yang relevan dan dapat diakses ke dalam proses pengambilan keputusan

mereka, dan sebaliknya mengandalkan informasi yang kurang relevan. Kesadaran terbatas

adalah salah satu gejala dari pemikiran intuitif, yang sering kali gagal untuk mengizinkan

kemungkinan bahwa bukti yang diperlukan untuk membuat keputusan yang baik tidak ada.

Akibatnya, para pengambil keputusan cenderung mengambil keputusan berdasarkan sebagian

kecil informasi yang tersedia (Brenner, Koehler, & Tversky, 1996; Kahneman, 2011).

"Ketidaksesuaian antara informasi yang dibutuhkan untuk keputusan yang baik dan informasi

yang disertakan dalam proses pengambilan keputusan" (Bazerman & Chugh, 2005: 10) dapat

menyebabkan kesalahan yang merugikan. Kesalahan seperti ini menjadi masalah dalam konteks

perekrutan, di mana para manajer terkenal dengan "ketergantungannya yang kuat pada intuisi

dan subjektivitas" (Highhouse, 2008: 333).

Penelitian menunjukkan bahwa jaringan interpersonal berfungsi sebagai saluran untuk

pertukaran informasi dalam organisasi, memberikan manajer akses siap pakai ke informasi yang

lebih kaya, lebih lengkap, dan dianggap lebih dapat dipercaya daripada informasi yang diperoleh

dari sumber lain (Bradach & Eccles, 1989; Granovetter, 1985; Podolny & Baron, 1997).

Meskipun informasi ini diharapkan dapat meningkatkan pengambilan keputusan, akses terhadap

informasi saja tidak cukup untuk memastikan keputusan yang baik; manajer masih harus memilih

informasi mana yang akan digunakan dan informasi mana yang harus diabaikan. Faktanya,

penelitian-penelitian yang mencakup konteks dan tingkat analisis yang berbeda, mulai dari

penelitian yang meneliti keputusan perekrutan individu3 hingga keputusan di tingkat perusahaan

mengenai pemilihan merger dan akuisisi.

3 Ada banyak bukti bahwa para manajer mengandalkan informasi yang tidak relevan ketika membuat
keputusan perekrutan. Sebagai contoh, para manajer sering kali memberikan bobot yang cukup besar pada
kinerja di pekerjaan sebelumnya meskipun faktanya hal tersebut sering kali menjadi prediktor yang buruk
bagi kinerja di masa depan (Cascio & Aguinis, 2008) dan sering kali mengizinkan atribut-atribut seperti
jenis kelamin, ras, daya tarik, dan berat badan untuk memengaruhi keputusan perekrutan (misalnya
 27

Agerström & Rooth, 2011; Bertrand & Mullainathan, 2004; Marlowe, Schneider, & Nelson, 1996). Ada
juga bukti yang sama banyaknya bahwa para manajer perekrutan gagal mencari informasi yang relevan,
bahkan ketika informasi tersebut mudah diakses. Sebagai contoh, para manajer secara rutin mengabaikan
alat bantu seleksi yang telah terbukti dan memilih wawancara yang tidak terstruktur (Highhouse, 2008) dan
menggunakan wawancara untuk mengkonfirmasi kesan pertama mereka terhadap para kandidat dengan
mengorbankan pengumpulan informasi yang relevan dengan pekerjaan (Dougherty, Turban, & Callender,
1994).
 28

mitra akuisisi (Rogan & Sorenson, 2014), menunjukkan bahwa para manajer secara rutin berjuang

untuk secara obyektif mengevaluasi alternatif-alternatif yang memiliki hubungan dengan mereka.

Oleh karena itu, proses alokatif yang berorientasi pada pasar mungkin sebenarnya lebih

unggul daripada proses relasional yang kaya akan informasi dalam membantu para manajer

mengatasi masalah yang terkait dengan kesadaran terbatas dengan membentuk informasi yang

digunakan untuk mengevaluasi alternatif. Pasar adalah institusi yang didukung oleh sistem aturan

dan konvensi yang dirancang untuk memfasilitasi pertukaran di antara pembeli dan penjual

(Casson, 1982; Menard, 1995; Polyani, 1957). Aturan dan konvensi ini - yang saya sebut sebagai

formalitas - bersifat impersonal dan tidak arbitrer, sehingga memberikan kerangka kerja yang

stabil untuk terjadinya transaksi (Menard, 1995: 1967). Aturan dan konvensi ini memiliki

beberapa fungsi, dua di antaranya adalah menetapkan syarat-syarat pertukaran dan

mendefinisikan apa yang merupakan perilaku yang sah (Loasby, 2000). Dalam menetapkan

syarat-syarat pertukaran, pasar menghasilkan seperangkat kriteria evaluasi; dalam mendefinisikan

apa yang merupakan perilaku yang sah, pasar membebankan pertanggungjawaban pada pembeli.

Adanya kriteria evaluasi formal dan tingkat akuntabilitas yang tinggi mendisiplinkan

pengambilan keputusan manajerial dengan cara yang berbeda.

Kriteria evaluasi dan informasi yang relevan. Agar pasar dapat berfungsi, harus ada

mekanisme bagi para manajer (sebagai pembeli) untuk menyiarkan informasi kepada penjual

potensial tentang barang atau jasa yang ingin mereka beli (Zenger et al., 2011). Meskipun para

manajer mungkin mengalami kesulitan untuk mengartikulasikan kebutuhan mereka sepenuhnya

(Nickerson & Zenger, 2004), informasi awal ini membantu menetapkan persyaratan pertukaran

karena berfungsi sebagai kriteria awal yang digunakan penjual untuk mengevaluasi minat mereka

dalam mengejar peluang dan sebagai kriteria awal yang digunakan pembeli untuk mengevaluasi

penjual yang memilih sendiri ke dalam kumpulan pertimbangan. Hasilnya, penjual cenderung

memberikan, dan pembeli cenderung mencari, informasi yang memungkinkan pembeli untuk

mengevaluasi alternatif berdasarkan serangkaian kriteria yang telah ditetapkan.

 29

Salah satu cara posting memperkenalkan formalitas pasar adalah melalui penggunaan

deskripsi pekerjaan formal. Proses posting mengharuskan manajer untuk membuat deskripsi

pekerjaan formal, yang diperlukan untuk menyiarkan informasi tentang pekerjaan yang terbuka

kepada kandidat potensial. Meskipun sulit untuk mengembangkan deskripsi pekerjaan yang akurat

dan komprehensif (Backhaus, 2004; Sanchez & Levine, 2012), persyaratan yang ditetapkan pada

tahap awal ini tetap berfungsi sebagai seperangkat kriteria formal yang digunakan untuk

mengevaluasi kandidat potensial. Sebaliknya, sponsorship tidak mengharuskan manajer untuk

membuat deskripsi pekerjaan formal sebelum mengevaluasi kandidat. Seorang manajer harus

memiliki pemahaman yang wajar tentang persyaratan pekerjaan serta atribut kandidat yang

diinginkan, tetapi karena manajer tidak harus menyiarkan informasi tentang pekerjaan yang

terbuka, deskripsi pekerjaan formal tidak diperlukan. Hal ini memungkinkan manajer untuk secara

informal membentuk persyaratan pekerjaan di sekitar kandidat yang mereka sukai daripada

mengevaluasi kandidat berdasarkan persyaratan pekerjaan (Miner, 1987; Sanchez & Levine,

2012). Jika dibandingkan dengan sponsorship, kehadiran kriteria evaluasi formal dalam posting

lebih cenderung mendorong manajer untuk mempertimbangkan kembali relevansi informasi yang

digunakan untuk mengevaluasi kandidat, dengan manajer lebih mungkin untuk mengenali dan

mencari informasi yang memungkinkan mereka untuk mengevaluasi kemampuan kandidat untuk

bekerja dengan baik dalam pekerjaan.

Akuntabilitas dan informasi yang tidak relevan. Mekanisme yang menerapkan tanggung

jawab dan akuntabilitas memastikan partisipasi berkelanjutan dari para pelaku pasar dengan

menanamkan keyakinan bahwa transaksi di masa depan akan diselesaikan dengan cara yang adil,

jujur, dan teratur. Yang perlu diperhatikan adalah persepsi mengenai proses di mana perusahaan

mengambil keputusan alokatif di pasar dapat mempengaruhi persepsi mengenai keadilan, di mana

keputusan yang tampak bebas dari bias dan berdasarkan kriteria obyektif dianggap lebih adil dan

sah (Bies, Tripp, & Neale, 1993; Williams, 1987). Oleh karena itu, akuntabilitas berbasis pasar

seharusnya mendorong manajer untuk menghindari penggunaan informasi yang dapat membuat
 30

keputusan mereka dianggap subyektif.

 31

Posting menanamkan akuntabilitas seperti pasar dalam proses perekrutan melalui

kebiasaan yang mengharuskan manajer untuk menjelaskan kepada kandidat internal yang tidak

berhasil mengapa mereka tidak terpilih. Untuk setiap kandidat internal yang berhasil,

kemungkinan besar ada beberapa kandidat yang tidak berhasil.

Karena kandidat internal yang tidak berhasil ini tetap menjadi karyawan, maka penting untuk

mengkomunikasikan dengan jelas alasan mengapa mereka tidak mendapatkan pekerjaan tersebut

untuk meminimalisir rasa tidak adil yang dapat menurunkan motivasi, kinerja, dan berpotensi

menyebabkan pergantian disfungsional. Dengan memastikan bahwa keputusan perekrutan harus

dikomunikasikan dan dipertahankan, kebiasaan untuk menjelaskan kepada karyawan mengapa

mereka tidak terpilih akan menanamkan akuntabilitas ke dalam proses perekrutan internal

(Tetlock, 1992).

Manajer yang mensponsori mencari kandidat secara informal melalui jaringan pribadi

mereka, sehingga para pekerja seringkali tidak menyadari bahwa mereka sedang dipertimbangkan

(Pinfield, 1995). Akibatnya, akuntabilitas lebih terbatas dibandingkan dengan proses penempatan,

meskipun tidak sepenuhnya tidak ada. Manajer diharuskan mengkomunikasikan keputusan

mereka kepada atasan, namun karena atasan biasanya memberikan keleluasaan yang cukup besar

kepada manajer untuk menentukan siapa yang dipilih, maka keputusan-keputusan tersebut tidak

perlu dipertahankan kepada khalayak yang lebih luas. Tingkat akuntabilitas yang lebih tinggi yang

dihasilkan melalui proses penempatan yang kompetitif seharusnya dapat mencegah penggunaan

informasi yang tidak relevan, karena para manajer lebih cenderung menggunakan kriteria yang

obyektif dalam menjustifikasi keputusan perekrutan mereka kepada khalayak yang lebih luas.

Secara keseluruhan, argumen-argumen ini menunjukkan bahwa memasukkan seleksi

mandiri dan formalitas ke dalam proses perekrutan internal akan membantu para manajer yang

rasional untuk mengatasi masalah-masalah yang terkait dengan menghasilkan dan mengevaluasi

alternatif-alternatif. Seleksi mandiri cenderung lebih efektif daripada pencarian manajerial aktif

dalam mengurangi kemungkinan bahwa kandidat yang luar biasa akan terlewatkan, sementara
 32

formalitas pasar cenderung lebih efektif dalam mendisiplinkan para manajer untuk menghindari

kesalahan yang merugikan yang terkait dengan kesadaran terbatas. Sebagai hasilnya, saya

berharap posting dapat menciptakan nilai lebih dari sponsorship dengan menghasilkan karyawan

internal yang lebih berkualitas, seperti yang diungkapkan oleh kinerja pekerja,
 33

perputaran dan kemajuan selanjutnya. Secara khusus, saya memprediksi hal tersebut jika

dibandingkan dengan karyawan internal yang disponsori:

H1: Pekerja yang dipekerjakan melalui penempatan akan memiliki peringkat kinerja yang

lebih tinggi di pekerjaan baru. H2: Pekerja yang dipekerjakan melalui penempatan akan

lebih kecil kemungkinannya untuk keluar dari perusahaan.

H3: Pekerja yang dipekerjakan melalui posting akan lebih mungkin untuk dipromosikan.

Persaingan dan Kompensasi

Kualitas karyawan hanyalah sebagian dari cerita, karena nilai yang dapat diperoleh

perusahaan dari karyawan dengan kualitas terbaik sekalipun sebagian besar bergantung pada

seberapa banyak mereka dibayar. Oleh karena itu, penting untuk memahami bagaimana

penempatan dan sponsor membentuk kompensasi. Untuk melakukannya, ada baiknya kita beralih

dari sudut pandang manajer dan memikirkan bagaimana negosiasi gaji dari sudut pandang pekerja.

Pasar dicirikan oleh persaingan terbuka, dimana para penjual menyadari bahwa mereka

bersaing untuk mendapatkan pembeli dengan penjual lainnya (Menard, 1995). Seleksi mandiri

dan formalitas merupakan dua mekanisme kunci yang mendukung sifat kompetitif pasar. Seleksi

mandiri memungkinkan penjual untuk mengejar peluang yang mereka minati dan formalitas

memfasilitasi pertukaran informasi yang pada akhirnya memungkinkan pembeli untuk

membandingkan informasi tentang alternatif yang tersebar luas dan membuat pilihan (Zenger et

al., 2011).

Posting dicirikan oleh kompetisi terbuka. Kandidat yang berminat memilih sendiri ke

dalam kumpulan pertimbangan ketika mereka melamar pekerjaan yang terbuka. Dengan

demikian, mereka membentuk antrian tenaga kerja - sekumpulan pekerja yang bersaing untuk

mendapatkan pekerjaan tertentu pada waktu tertentu (Reskin & Roos, 1990). Formalitas dari

proses penempatan menggarisbawahi sifat kompetitif dari penempatan, karena fakta bahwa

karyawan harus secara aktif mengajukan lamaran membuat mereka sadar bahwa mereka masuk ke

dalam sebuah kompetisi yang belum tentu mereka menangkan. Sebaliknya, sejauh ada kompetisi
 34

dalam sponsorship, kompetisi ini tidak memiliki struktur dan transparansi. Karena pencarian itu

mahal, para manajer secara rutin mempertimbangkan sejumlah kecil

 35

kandidat internal (seringkali hanya satu kandidat). Selain itu, karena para manajer mengumpulkan

informasi mengenai kandidat potensial secara informal, dalam kasus-kasus di mana beberapa

kandidat dipertimbangkan, mereka yang tidak terpilih seringkali tidak menyadari bahwa mereka

sedang dipertimbangkan (Pinfield, 1995).

Meskipun persaingan sering dilihat sebagai cara untuk menurunkan harga dengan

mengadu domba beberapa pemasok satu sama lain, penelitian terbaru yang mengeksplorasi aspek

psikologis sosial dari negosiasi memberikan alasan untuk mengharapkan hal yang sebaliknya

dalam konteks perekrutan internal. Premis utama dalam literatur ini adalah bahwa faktor

situasional mendorong individu untuk memberikan penekanan yang lebih besar pada pentingnya

hubungan diadik dalam negosiasi. Semakin banyak penekanan yang diberikan seseorang pada

hubungan diadik dalam konteks negosiasi, semakin besar kemungkinan mereka untuk

mengadopsi orientasi relasional dalam negosiasi, dan individu yang mengadopsi orientasi

relasional dalam negosiasi lebih mungkin untuk mengabaikan keuntungan ekonomi dalam upaya

untuk mengembangkan modal relasional (Curhan, Elfenbein, & Xu, 2006; Curhan, Neale, Ross,

& Rosencranz-Engelmann, 2008: 193; Gelfand et al., 2006).

Sponsor jauh lebih mungkin untuk memberi isyarat kepada pekerja untuk fokus pada

hubungan mereka dengan manajer perekrutan. Karena manajer perekrutan secara pribadi

menunjuk pekerja untuk pekerjaan tersebut tanpa adanya persaingan formal, hubungan dengan

manajer perekrutan cenderung sangat menonjol dan sangat dihargai pada saat tawaran pekerjaan

awal diberikan. Hal ini tidak mungkin terjadi dalam penempatan, dengan sifat kompetitif dari

proses yang menekankan sifat transaksional dari hubungan kerja. Akibatnya, pekerja yang

dipekerjakan melalui sponsorship lebih cenderung mengadopsi pendekatan orientasi relasional

ketika tiba saatnya untuk menegosiasikan kompensasi.

Ada dua alasan untuk menduga bahwa mengadopsi orientasi relasional dalam negosiasi

gaji memiliki implikasi penting bagi seberapa besar kemungkinan seorang pekerja mendapatkan

gaji. Pertama, seorang pekerja yang mengadopsi orientasi relasional cenderung tidak akan
 36

memulai negosiasi gaji. Karena fokus pada pengembangan modal relasional, ia akan menghindari

kesan mementingkan diri sendiri dan merasa tidak nyaman dengan ide meminta lebih banyak uang

(Gelfand et al., 2006). Ia juga akan lebih

 37

lebih mungkin untuk melihat tawaran awal sebagai sesuatu yang adil; menjadi lebih selaras

dengan tujuan pihak lain akan mengurangi kemungkinan dia akan menganggap kepentingan pihak

lain berlawanan dengan kepentingan mereka, sebuah kesalahan umum dalam negosiasi (Cross,

Bacon, & Morris, 2000; Gelfand dkk., 2006; Thompson & Deharpport, 1998). Kedua, seseorang

yang mengadopsi orientasi relasional lebih cenderung menggunakan strategi negosiasi

akomodatif, sementara seseorang yang mengadopsi orientasi transaksional lebih cenderung

bernegosiasi secara kompetitif (Curhan et al., 2008), dan pendekatan akomodatif terhadap

negosiasi gaji terbukti menghasilkan kenaikan gaji yang lebih rendah dibandingkan dengan

pendekatan yang kompetitif (Marks & Harold, 2011).

Hal ini menunjukkan bahwa ada pertukaran yang penting antara kualitas dan biaya yang

terkait dengan proses perekrutan internal yang berbeda. Meskipun posting cenderung menciptakan

nilai lebih, perusahaan tidak menangkap semua nilai tersebut, karena pekerja menangkap sebagian

dari nilai tersebut melalui kompensasi yang lebih tinggi. Sehubungan dengan sponsorship, sifat

kompetitif dari proses penempatan akan meningkatkan kemungkinan pekerja untuk melakukan

negosiasi gaji dan mengadopsi pendekatan yang lebih menguntungkan secara ekonomi ketika

mereka memilih untuk bernegosiasi, sehingga saya memprediksi hal tersebut:

H4: Kandidat internal yang dipekerjakan melalui posting akan menerima gaji awal yang

lebih tinggi daripada karyawan yang disponsori yang memasuki pekerjaan yang setara.

DATA & METODE

Saya menguji hipotesis saya dengan menggunakan catatan personalia bulanan yang

mencakup tahun 2008 hingga 2012 dari operasi perusahaan asuransi kesehatan Fortune 100 di

Amerika Serikat, yang saya sebut sebagai HealthCo. Data utama untuk analisis saya terdiri dari

lebih dari 11.000 perekrutan internal yang dilakukan selama periode pengamatan, yang

diidentifikasi dari kumpulan data yang lebih besar yang terdiri dari 1.914.519 pengamatan

bulanan yang mencakup 56.811 pekerja yang dipekerjakan oleh HealthCo antara tahun 2008 dan

2012.
 38

Meskipun menggunakan data dari satu perusahaan membatasi kepercayaan diri saya untuk

menggeneralisasi hasil saya, data ini sangat cocok untuk menguji hipotesis saya. Fitur yang

membedakan dari
 39

Data-data ini adalah cara HealthCo menghubungkan berbagai sistem informasi sumber daya

manusia mereka yang memungkinkan saya untuk mengidentifikasi dengan jelas apakah posting

atau sponsor digunakan untuk memfasilitasi setiap perekrutan internal. Saya juga dapat

menautkan catatan personalia ini ke kumpulan data pendamping yang berisi informasi lebih dari

350.000 lamaran kerja internal dan eksternal yang diajukan selama tahun 2012, sehingga saya

dapat melakukan beberapa pemeriksaan yang lebih kuat. Menggunakan data personalia dari satu

organisasi memiliki beberapa keuntungan lain, termasuk fakta bahwa ukuran kinerja saya

terstandardisasi di seluruh pekerjaan dan bahwa saya dapat mengontrol efek dari konten pekerjaan

dan lokasi pekerjaan yang berbeda (dan karena itu atribut yang berbeda dari perpindahan antar

pekerjaan), yang kesemuanya akan menimbulkan kesulitan empiris yang substansial dalam studi

multi perusahaan. Selain itu, mendapatkan data jenis ini dari satu perusahaan saja sudah cukup

menantang; banyak perusahaan yang gagal mencatat secara sistematis data apa pun yang terkait

dengan mobilitas karyawan (ERC, 2010; Oracle, 2012) dan bagi mereka yang melakukannya,

ketakutan akan sanksi analisis internal untuk mengungkapkan pola diskriminasi ilegal yang

sebelumnya tidak dikenali memiliki dampak buruk yang membuat perusahaan enggan untuk

mengeksplorasi proses-proses ini (Strum, 2001). Pengaturan itu sendiri mengurangi setidaknya

beberapa kekhawatiran tentang generalisasi, karena HealthCo mencerminkan organisasi

kontemporer besar lainnya dalam beberapa hal: keputusan perekrutan didelegasikan kepada

manajer individu, karyawan secara eksplisit didorong untuk secara aktif mengelola karir mereka

di tengah kurangnya jalur kemajuan yang jelas, dan ada sejumlah besar mobilitas lateral dan

vertikal di seluruh pekerjaan yang didefinisikan secara luas.

Mengidentifikasi Karyawan Internal

Perekrutan internal terjadi ketika seorang manajer mengisi pekerjaan yang terbuka dengan

karyawan saat ini4 , seperti yang ditunjukkan oleh perubahan kode pekerjaan, departemen, atau

keduanya dari satu bulan ke bulan berikutnya. Seorang karyawan yang berganti kode jabatan akan

mengemban tugas dan tanggung jawab baru. Pindah ke

 40

4 HealthCo memiliki proses perekrutan berdasarkan lowongan, yang berarti bahwa semua perekrutan -
internal dan eksternal - didahului dengan lowongan pekerjaan.
 41

departemen baru merupakan perpindahan ke area bisnis yang berbeda, karena departemen diatur

berdasarkan produk, pasar geografis, dan pelanggan (internal dan eksternal). Semua pekerjaan

tingkat pemula di HealthCo diisi melalui perekrutan eksternal; perekrutan internal digunakan

bersamaan dengan perekrutan eksternal untuk mengisi pekerjaan di atas tingkat pemula.

Penting untuk ditekankan bahwa perubahan kode jabatan tidak hanya mewakili perubahan

jabatan, dengan sedikit perubahan pada pekerjaan yang dilakukan (Miner, 1987). Pekerjaan di

HealthCo didefinisikan secara luas berdasarkan tingkat hirarki, fungsi, dan peran. Dalam satu

tahun rata-rata, sekitar 34.000 karyawan didistribusikan di 462 kode pekerjaan. Pekerjaan diatur

ke dalam sembilan level yang berbeda [(1) Entry-level, (2) Team Lead, (3) Supervisor/Analis, (4)

Manajer/Profesional, (5) Direktur/Pemimpin Teknis, (6) Wakil Presiden, (7) Wakil Presiden

Senior, (8) Wakil Presiden Eksekutif, (9) CEO] dan tiga puluh area fungsional, termasuk yang

umum bagi sebagian besar perusahaan besar (mis. Penjualan, Keuangan/Akuntansi, SDM, dan

Pemasaran) serta beberapa yang lebih spesifik untuk industri asuransi kesehatan (mis. Bimbingan

Klinis, Kontraktor Penyedia). Peran menunjukkan kompetensi spesifik yang dibutuhkan untuk

melakukan pekerjaan. Sebagai contoh, "Pengembang Kreatif" adalah peran Level 3 di bidang

Pemasaran; "Merekrut Pimpinan" adalah peran Level 3 di bidang Sumber Daya Manusia; dan

"Arsitek", "Konsultan Aplikasi", dan "Manajer Proyek" adalah peran Level 3 di bidang TI, yang

masing-masing terkait dengan kompetensi yang berbeda. Akibatnya, perubahan kode jabatan

mencerminkan perubahan yang berarti dalam pekerjaan yang dilakukan seseorang.

Variabel Dependen

Kualitas perekrutan dan kompensasi adalah dua hasil utama yang menjadi perhatian utama
dalam penelitian ini.

Para peneliti telah menggunakan berbagai macam hasil pasca perekrutan untuk menilai kualitas

karyawan (misalnya, lihat Breaugh & Starke, 2000; Zottoli & Wanous, 2000). Dalam upaya untuk

memberikan akuntansi holistik tentang kualitas karyawan internal di HealthCo, saya menguji

hipotesis saya dengan menggunakan berbagai indikator kualitas: peringkat kinerja, kinerja relatif,
 42

perputaran, dan kemajuan selanjutnya.

 43

Peringkat kinerja dan kinerja relatif. Saya menggunakan lima ukuran kinerja pekerjaan

dari evaluasi kinerja tahunan HealthCo sebagai seperangkat indikator kualitas yang pertama.

Skor kontribusi. Skor kontribusi pekerja menilai kontribusi mereka terhadap keberhasilan

organisasi. Pada pekerjaan yang tidak terlalu berdampak langsung pada hasil di tingkat

organisasi (misalnya pada tingkat yang lebih rendah), skor kontribusi biasanya digunakan untuk

menilai kontribusi mereka pada departemen atau lini bisnis. Hal ini diukur pada skala 0 sampai 4

[0 = tidak berkontribusi (0%), 1 = kontribusi rendah (3,2%), 2 = kontribusi sedang (19%), 3 =

kontribusi penuh (66,8%), dan

4 = kontribusi yang patut dicontoh (11%)].

Skor kompetensi. Skor kompetensi seorang pekerja menilai keterampilan mereka relatif

terhadap apa yang dibutuhkan untuk pekerjaan tersebut. Setiap pekerja menerima skor terpisah

untuk masing-masing dari delapan kompetensi yang ditetapkan untuk kode pekerjaan mereka5 .

Setiap kompetensi diukur dengan skala 1 hingga 4 [1 = Belajar, 2 = Menunjukkan, 3 =

Mendemonstrasikan, 4 = Memodelkan]. Saya menghitung rata-rata nilai individu ini untuk

menghitung nilai kompetensi secara keseluruhan [1-2 (11,4%), 2-3 (66,2%), 3-4 (12,4%)].

Kinerja relatif. Manajer juga mengurutkan pekerja dalam pekerjaan yang serupa sebagai

bagian dari proses kalibrasi tahunan (dijelaskan di bawah). Namun, para pekerja tidak hanya

diurutkan berdasarkan kontribusi dan nilai kompetensi mereka; namun, hal ini dimaksudkan

sebagai ukuran nilai keseluruhan bagi organisasi yang memperhitungkan kinerja sebelumnya dan

potensi di masa depan.

Meskipun tidak ada pedoman yang ditetapkan secara formal yang menentukan seberapa halus

manajer harus membedakan pekerja, sesi kalibrasi ini biasanya menciptakan "ember" karyawan;

sekelompok 100 karyawan mungkin tidak diberi peringkat dari 1 (tertinggi) hingga 100

(terendah), tetapi lima karyawan teratas dapat menerima nilai 1, sepuluh karyawan berikutnya

mendapat nilai 2, 25 karyawan berikutnya mendapat nilai 3, dan seterusnya. Saya menggunakan

peringkat ini untuk membuat tiga ukuran dikotomis dari kinerja relatif: apakah seorang pekerja
 44

memiliki peringkat

5 Dipilih dari keseluruhan perpustakaan yang terdiri dari 124 kompetensi yang disesuaikan untuk HealthCo.
 45

berada di kuartil teratas (25% teratas), separuh terbawah (50% terbawah), atau kuartil terbawah

(25% terbawah) dari kelompoknya pada tahun pertama di pekerjaan barunya.

Meskipun semua ini merupakan ukuran kinerja yang subjektif, beberapa peneliti

berpendapat bahwa penilaian subjektif merupakan salah satu ukuran kinerja yang paling valid

meskipun ada kekhawatiran tentang potensi bias manajerial yang dapat mempengaruhi penilaian

(Cascio, 1998). Penilaian subyektif memungkinkan manajer untuk mempertimbangkan berbagai

perilaku dan hasil yang relevan dengan pekerjaan (Medoff & Abraham, 1981) serta

memperhitungkan faktor-faktor yang mempengaruhi kinerja di luar kendali individu pekerja

(Campbell, McCloy, Oppler, & Sager, 1993). Namun, penting juga untuk dicatat bahwa proses

evaluasi tahunan HealthCo membantu mengurangi potensi kekhawatiran tentang bias pengawasan

yang mempengaruhi peringkat kinerja individu. Para manajer yang mengawasi para pekerja dalam

pekerjaan yang sama bertemu secara langsung untuk meninjau dan mendiskusikan penilaian

mereka terhadap masing-masing pekerja - sebuah proses yang dikenal sebagai kalibrasi. Diskusi

kalibrasi ini dimaksudkan untuk memastikan bahwa para manajer mengevaluasi para pekerja

dengan standar yang sama dan untuk mengidentifikasi dan memperbaiki contoh-contoh di mana

para manajer mungkin menilai karyawan terlalu keras atau terlalu lunak. Penelitian telah

menunjukkan bahwa kalibrasi cenderung mengurangi subjektivitas dan bias dalam penilaian

kinerja, baik karena penilaian cenderung lebih konsisten di antara para karyawan ketika para

manajer memiliki pandangan yang sama mengenai standar penilaian (Mclntyre, Smith, & Hassett,

1984; Pulakos, 1984) maupun karena "sesama manajer biasanya tidak akan mudah membiarkan

satu sama lain jika mereka percaya bahwa seorang karyawan telah dinilai secara tidak adil,

sehingga menciptakan tekanan dari sesama manajer yang memberikan insentif yang kuat untuk

memberikan penilaian yang akurat" (Pulakos & O'Leary, 2011: 152-3).

Pergantian pemain. Pergantian karyawan adalah indikator kedua dari kualitas

pertandingan (Jovanovic, 1979; Mortensen, 1988). Saya membuat dua ukuran dikotomis dari

turnover yang menunjukkan apakah seorang pekerja keluar dari organisasi dalam 12 bulan
 46

pertama (turnover12) atau 24 bulan (turnover24) setelah pindah ke pekerjaan baru. Saya tidak

membedakan antara perputaran sukarela dan tidak sukarela karena dua alasan.
 47

Secara konseptual, keduanya merupakan indikator ketidakcocokan, di mana pekerja dan

perusahaan cenderung mengakhiri hubungan kerja yang tidak cocok. Secara empiris, meskipun

data memang menunjukkan apakah pergantian karyawan dilakukan secara sukarela atau tidak

sukarela, para manajer dan staf SDM di HealthCo mengatakan kepada saya bahwa mereka tidak

mempercayai indikator-indikator ini sebagai cerminan alasan sebenarnya seorang karyawan

meninggalkan organisasi. Konsisten dengan penelitian yang menunjukkan bahwa tingkat turnover

relatif rendah di antara karyawan internal pada pekerjaan non-entry level (Bidwell, 2011), sekitar

5% karyawan internal keluar dari perusahaan dalam waktu 12 bulan, sementara 13% keluar dalam

waktu 24 bulan.

Promosi. Promosi berikutnya adalah indikator kualitas yang ketiga. Promosi adalah

perekrutan internal yang mengakibatkan karyawan pindah ke pekerjaan yang lebih tinggi. Karena

waktu untuk promosi bervariasi di setiap pekerjaan, saya membuat dua ukuran dikotomis untuk

kemajuan berikutnya: apakah seorang pekerja dipromosikan dalam waktu 24 bulan (prom24) atau

36 bulan (prom36) setelah pindah ke pekerjaan baru.

Sekitar 14% dari karyawan internal kemudian dipromosikan dalam waktu 24 bulan, sementara

31% dipromosikan dalam waktu 36 bulan.

Gaji awal. Gaji awal adalah logaritma natural dari gaji yang diterima pekerja pada bulan

pertama di pekerjaan baru. Gaji merupakan sebagian besar kompensasi bagi sebagian besar

pekerja di HealthCo. Pekerja bagian penjualan merupakan pengecualian utama dan bonus mereka,

yang terkait dengan target penjualan yang jelas, dapat menyumbang sebagian besar dari total

pendapatan mereka.

Namun, jumlah bonus yang diharapkan akan diterima oleh para pekerja ini berdasarkan target

mereka diperhitungkan ke dalam angka gaji tahunan yang tercatat dalam catatan personalia

HealthCo. Sebagai contoh, jika seorang tenaga penjualan dipekerjakan dalam suatu pekerjaan

dengan gaji pokok sebesar $80.000 dan berharap mendapatkan bonus sebesar

$40.000, dengan total kompensasi tahunan yang diharapkan sebesar $120.000, gajinya dalam
 48

catatan personalia bulanan yang saya ambil adalah $120.000. Ini berarti angka gaji yang saya

gunakan mewakili total kompensasi yang seharusnya ia dapatkan selama setahun, yang merupakan

angka yang akan dinegosiasikan.

 49

Variabel Independen

Variabel independen adalah indikator dikotomis apakah posting (1) atau sponsorship

(0) digunakan untuk mengisi pekerjaan tersebut, yang diidentifikasi melalui indikator yang tidak

ambigu dalam pengamatan bulanan pertama pekerja di pekerjaan barunya. 3.841 (43%) dari

perekrutan internal dilakukan melalui pengumuman dan 5.458 (57%) melalui sponsor.

Variabel Kontrol

Perhatian penting ketika menggunakan data non-eksperimental adalah potensi variabel

yang dihilangkan untuk menciptakan korelasi palsu antara variabel independen dan dependen.

Manajer perekrutan di Health Co. dapat memilih apakah akan mengisi sebuah pekerjaan melalui

posting atau sponsor. Perhatian khusus dalam penelitian ini adalah bahwa mungkin ada variabel

yang mempengaruhi keputusan manajer tentang proses mana yang akan digunakan dan juga hasil

yang diinginkan. Dengan tidak adanya bukti empiris yang ada, tampaknya pilihan untuk

menggunakan posting atau sponsorship mungkin dipengaruhi oleh dua faktor: atribut pekerjaan

dan persepsi manajer tentang ketersediaan kandidat yang memenuhi syarat. Bisa jadi ada

pekerjaan tertentu yang selalu di-posting dan pekerjaan lain yang selalu diisi melalui sponsor. Bisa

juga manajer hanya memposting pekerjaan jika mereka tidak dapat atau menganggap mereka tidak

akan dapat mengidentifikasi kandidat yang memenuhi syarat melalui jaringan pribadi mereka.

Saya memanfaatkan tingkat detail dalam data saya untuk mengatasi masalah pertama, yang

memungkinkan saya untuk mengontrol banyak karakteristik pekerjaan yang mungkin diharapkan

memiliki efek pada pilihan antara memposting dan sponsor. Untuk mengatasi masalah kedua, saya

melakukan serangkaian pemeriksaan ketahanan dengan menggunakan pendekatan variabel

instrumental, yang akan saya jelaskan secara lebih rinci dalam makalah ini.

Atribut pekerjaan. Untuk membandingkan hasil yang terkait dengan proses yang berbeda

yang digunakan untuk pekerjaan yang serupa, saya mengontrol beberapa atribut tingkat pekerjaan.

Saya memasukkan variabel dummy untuk mengontrol peringkat hierarki, area fungsi dan negara

bagian tempat suatu pekerjaan berada (51% pekerjaan berada di negara bagian kantor pusat).
 50

Untuk mengontrol kecenderungan tetap dari pekerjaan yang berbeda untuk diisi oleh
 51

Selain itu, saya juga menyertakan dummy terpisah untuk setiap 266 kode pekerjaan yang diisi

melalui perekrutan internal selama periode pengamatan saya. Yang penting, data menunjukkan

tidak ada perbedaan sistematis dalam jenis pekerjaan yang diisi melalui posting atau sponsor.

Dari semua kode pekerjaan yang diisi secara internal dari tahun 2008-2012, 84% diisi melalui

penempatan dan sponsor, dan perpindahan ke pekerjaan tersebut mencapai 99% dari semua

perekrutan internal. Selain itu, 75% dari kode pekerjaan yang diisi secara eksklusif melalui

posting atau sponsor hanya diisi sekali atau dua kali, 88% hanya tiga kali, dan tidak ada yang

lebih dari enam kali. Oleh karena itu, ada kemungkinan bahwa bahkan kode pekerjaan yang diisi

secara eksklusif melalui satu proses selama periode waktu ini tetap terbuka untuk diisi melalui

kedua proses, dengan eksklusivitas yang tampak sebagai artefak dari begitu sedikitnya perekrutan

yang dilakukan untuk kode pekerjaan tersebut.

Atribut transisi pekerjaan. Kekhawatiran lainnya adalah bahwa hasil yang diperoleh dapat

dipengaruhi oleh jenis perpindahan yang dilakukan pekerja, bukan oleh bagaimana perpindahan

tersebut dilakukan. Sebagai contoh, kita mungkin mengharapkan pekerja yang pindah ke

pekerjaan yang lebih mirip memiliki tingkat kinerja yang lebih tinggi dan bahwa pekerja yang

disponsori cenderung berasal dari pekerjaan yang lebih mirip karena manajer lebih cenderung

memiliki hubungan pribadi dengan pekerja dalam pekerjaan yang mirip. Untuk memperhitungkan

kemungkinan ini, saya menyertakan dummy untuk mengetahui apakah perekrutan internal

menghasilkan pekerja yang pindah ke pekerjaan baru dalam fungsi yang sama atau departemen

yang sama. Demikian pula, saya menyertakan dummy yang mengindikasikan apakah perekrutan

internal menghasilkan promosi (kenaikan jabatan secara vertikal; n=4.843), ekspansi (kenaikan

jabatan secara lateral yang menghasilkan perluasan kompetensi pekerja; n=3.594), atau transfer

(perpindahan jabatan secara lateral ke pekerjaan yang sama di departemen yang berbeda;

n=1.122), yang dapat memengaruhi gaji dan/atau kinerja. Saya menyertakan gaji pekerja di bulan

terakhir pekerjaan mereka sebelumnya dalam beberapa model untuk memperhitungkan

kemungkinan bahwa penyesuaian gaji internal dapat didasarkan pada gaji karyawan sebelumnya,
 52

bahkan ketika berpindah ke pekerjaan yang sangat berbeda.

 53

Atribut individu. Saya juga menyertakan kontrol untuk berbagai atribut kandidat.

Karakteristik demografis mencakup jenis kelamin, usia dan usia-kuadrat, dan etnis. Masa kerja

dan masa kerja-kuadrat dihitung sebagai jumlah bulan (kuadrat) seorang pekerja telah

dipekerjakan oleh perusahaan. Saya menyertakan kontribusi dan nilai kompetensi pekerja dalam

pekerjaan mereka sebelumnya sebagai indikator kasar dari kualitas sebelum perekrutan.

HealthCo tidak menyertakan tahun pendidikan atau gelar tertinggi yang diselesaikan dalam

catatan personalia mereka.

Kontrol tambahan. Saya memasukkan boneka untuk setiap tahun dalam sampel untuk

mencerminkan perubahan kondisi pasar tenaga kerja. Dalam model yang memprediksi kinerja

relatif, saya mengontrol ukuran kelompok peringkat kinerja pekerja (misalnya jumlah pekerja

dalam kelompok peringkat paksa). Pembatasan Sampel

Sampel yang digunakan untuk setiap analisis saya bervariasi sesuai dengan sejumlah

batasan yang saya terapkan pada data. Dalam analisis di mana peringkat kinerja digunakan

sebagai variabel dependen, saya menghilangkan pengamatan dengan data kinerja yang hilang,

yang terjadi ketika pekerja baru saja dipekerjakan untuk dinilai atau keluar dari perusahaan

sebelum dinilai (n=2.484). Pada model perputaran dan model kemajuan, sampel dibatasi pada

pekerja yang dipekerjakan cukup awal sehingga saya dapat menghitung variabel dependen yang

diminati (misalnya, untuk dapat dimasukkan ke dalam model yang memprediksi perputaran dalam

12 bulan, pekerja harus keluar dari perusahaan dalam waktu 12 bulan setelah pindah ke pekerjaan

baru mereka atau menempati pekerjaan tersebut selama 12 bulan). Terakhir, dalam model

kompensasi, saya mengecualikan pengamatan dengan data gaji yang hilang (n=28).

ANALISIS & HASIL

Tabel 1 menyajikan rata-rata, standar deviasi, dan korelasi untuk variabel dependen dan

independen utama, dengan setiap pengamatan mewakili perekrutan internal. Yang menarik

adalah korelasi antara berbagai ukuran kualitas perekrutan. Korelasi antara skor kontribusi dan

kompetensi (r = .67) yang mengindikasikan bahwa kedua ukuran ini mengambil

 Subscribe to DeepL Pro to translate larger documents.
Visit www.DeepL.com/pro for more information.
29

aspek kinerja yang berbeda. Korelasi antara kedua ukuran ini dan variabel peringkat relatif

bervariasi ke arah yang diharapkan sekaligus menunjukkan bahwa mereka juga mengambil aspek-

aspek kinerja yang berbeda, karena keduanya berkorelasi positif dengan pekerja yang masuk

dalam 25% teratas (r = .43, .45) dan berkorelasi negatif dengan pekerja yang masuk dalam 50%

terbawah (r = -.46, -.50) dan 25% terbawah (r = -.40, -.46). Korelasi yang rendah antara variabel

peringkat kinerja dan variabel pergantian dan promosi berikutnya (tidak ada yang melebihi r = +/-

.13) menunjukkan bahwa masing-masing variabel ini menangkap elemen kinerja yang berbeda.

Penting juga untuk mencatat korelasi yang rendah antara gaji dan masing-masing ukuran kinerja

(tidak ada yang melebihi r = +/- .16), yang menggarisbawahi gagasan bahwa gaji tidak hanya

mencerminkan ekspektasi manajer tentang kinerja dalam pekerjaan baru, tetapi lebih banyak

ditentukan oleh proses negosiasi yang terjadi sebelum manajer mengamati kinerja yang

sebenarnya (Jovanovic, 1979). Hal ini penting dalam menginterpretasikan hasil penelitian, karena

saya berpendapat bahwa meskipun penempatan karyawan yang lebih berkualitas dan gaji yang

lebih tinggi, gaji yang lebih tinggi bukanlah hasil dari ekspektasi manajer terhadap tingkat kinerja

yang lebih tinggi. Faktanya, pemeriksaan ketahanan menunjukkan bahwa, jika ada, manajer

cenderung mengharapkan karyawan yang disponsori untuk berkinerja lebih baik.

Sisipkan Tabel 1 di sini

Kinerja

Tabel 2 menyajikan analisis dari masing-masing lima ukuran kinerja. Saya menggunakan

regresi kuadrat terkecil biasa dalam model yang menggunakan skor kompetensi dan kontribusi

sebagai variabel dependen. Skor kompetensi adalah variabel kontinu dan meskipun skor

kontribusi adalah ukuran ordinal yang bersifat diskrit, model OLS lebih mudah diinterpretasikan

daripada model logit yang dipesan (dan keduanya memberikan hasil yang serupa). Saya

menggunakan spesifikasi logit untuk model-model yang menggunakan kinerja relatif sebagai

hasil yang diinginkan, karena variabel-variabel dependennya adalah biner. Unit analisis dalam
 30
semua model adalah sebuah
 31

perekrutan internal dan ukuran kinerja mencerminkan kinerja pada tahun pertama bekerja. Saya

mengelompokkan kesalahan berdasarkan individu untuk memperhitungkan ketidakindependenan

di antara kesalahan-kesalahan tersebut.

Dengan menggunakan skor kontribusi dan kompetensi sebagai variabel dependen, Model 1

dan 2 memberikan dukungan untuk Hipotesis 1, dengan koefisien positif yang signifikan untuk

posting di kedua model tersebut menunjukkan bahwa perekrutan internal yang dilakukan melalui

posting lebih baik daripada perekrutan internal yang dilakukan melalui sponsor. Ukuran efeknya

relatif kecil, dengan posting menghasilkan peningkatan sekitar sepersepuluh dari standar deviasi

pada setiap ukuran kinerja, sebuah isu yang saya bahas kembali dalam uji ketahanan. Model 3

memberikan dukungan tambahan untuk Hipotesis 1 dalam menunjukkan bahwa perekrutan

internal yang dilakukan melalui penempatan memiliki kemungkinan sekitar 13% lebih besar untuk

berada di kuartil teratas dari distribusi kinerja/potensi mereka masing-masing daripada perekrutan

internal yang disponsori. Model 4 dan 5 lebih lanjut mengungkapkan bahwa karyawan internal

yang direkrut melalui posting lebih kecil kemungkinannya untuk memiliki peringkat kinerja yang

buruk di pekerjaan baru. Koefisien negatif yang signifikan untuk posting menunjukkan bahwa

karyawan internal yang dibuat melalui posting memiliki kemungkinan sekitar 13% lebih kecil

untuk berada di bagian bawah distribusi peringkat dan sekitar 15% lebih kecil kemungkinannya

untuk berada di kuartil terbawah. Secara keseluruhan, hasil ini menunjukkan bahwa posting

mengarah pada keputusan perekrutan internal yang lebih baik, sebagian dengan membantu para

manajer menghindari kesalahan perekrutan.

Sisipkan Tabel 2 di sini

Perputaran

Model 7 dan 8 pada Tabel 3 menggunakan model logit untuk menguji Hipotesis 2,

bahwa karyawan internal yang direkrut melalui posting lebih kecil kemungkinannya untuk

keluar dari perusahaan. Kedua model memberikan dukungan untuk hipotesis ini, dengan
 32
koefisien negatif yang signifikan untuk posting yang menunjukkan bahwa karyawan internal

yang dibuat melalui posting memiliki kemungkinan 20% lebih kecil untuk keluar dari

perusahaan dalam waktu 12 bulan (Model 7) dan sekitar 18% lebih kecil kemungkinannya

untuk keluar dari perusahaan dalam waktu 24 bulan (Model 8).

 33

Sisipkan Tabel 3 di sini

Kemajuan Selanjutnya

Model 9 dan 10 pada Tabel 3 juga menggunakan model logit untuk menguji Hipotesis 3,

yaitu bahwa karyawan internal yang direkrut melalui penempatan lebih mungkin untuk

dipromosikan. Saya menemukan dukungan yang sangat terbatas untuk argumen ini. Koefisien

untuk posting tidak berbeda secara signifikan dari nol pada Model 9, yang mengindikasikan

bahwa tidak ada perbedaan dalam tingkat promosi selama 24 bulan untuk karyawan internal yang

dipromosikan melalui posting atau sponsor. Koefisien untuk posting adalah positif tetapi hanya

sedikit signifikan dalam Model 10, memberikan beberapa indikasi bahwa posting mungkin lebih

mungkin memimpin promosi selama periode 36 bulan, tetapi tidak ada yang konklusif.

Gaji

Model 5 dan 6 pada Tabel 3 menguji Hipotesis 4, bahwa karyawan internal yang direkrut

melalui penempatan akan menerima gaji awal yang lebih tinggi daripada karyawan sponsor yang

memasuki pekerjaan yang setara. Koefisien positif yang signifikan untuk posting dalam Model

11 memberikan dukungan untuk hipotesis ini, yang mengungkapkan bahwa karyawan yang

diposting menerima gaji hampir 4% lebih tinggi, rata-rata, daripada karyawan yang disponsori.

Premi penempatan sebesar 4% tetap ada setelah mengendalikan kinerja di pekerjaan sebelumnya

(Model 12).

Pola Mobilitas dan Perekrutan di Dalam Departemen

Dalam berteori tentang dampak seleksi mandiri dan formalitas terhadap kualitas

perekrutan, saya membuat dua argumen yang saling melengkapi, satu yang menyatakan bahwa

seleksi mandiri meningkatkan kualitas dengan memperluas kumpulan kandidat potensial dan yang

lainnya menyatakan bahwa formalitas membantu membentuk informasi yang digunakan para

manajer ketika mengevaluasi kandidat. Namun, hasil regresi di atas tidak memungkinkan saya

untuk mengetahui apakah salah satu atau kedua mekanisme ini yang mendorong hasilnya. Untuk
 34
memeriksa masalah ini secara lebih rinci, saya pertama-tama menyajikan statistik deskriptif

mengenai sumber perekrutan internal (Tabel 4). Data ini menunjukkan bahwa posting secara

signifikan lebih mungkin menghasilkan perekrutan dari departemen dan fungsi yang berbeda,
 35

kota dan bahkan gedung yang berbeda, konsisten dengan argumen saya bahwa seleksi mandiri

mengurangi kemungkinan manajer akan mengabaikan kandidat pengecualian dengan

memungkinkan manajer untuk lebih mudah mengidentifikasi kandidat yang tersebar luas di

seluruh organisasi. Saya kemudian melakukan serangkaian regresi yang membatasi analisis pada

perekrutan internal yang dilakukan di dalam departemen (Tabel 5). Karena para manajer

cenderung menyadari dan memiliki akses ke informasi yang jauh lebih rinci tentang kandidat yang

berada di dalam departemen mereka sendiri, hal ini memberikan pengujian yang lebih kuat

terhadap argumen saya bahwa, selain memberikan lebih banyak alternatif, posting meningkatkan

pengambilan keputusan dengan mendisiplinkan informasi apa yang mereka gunakan dalam

mengevaluasi kandidat. Konsisten dengan teori saya, hasilnya hampir sama dengan yang disajikan

pada Tabel 2 dan 3. Bahkan, jika kita mengasumsikan bahwa para manajer tidak hanya memiliki

akses yang siap pakai, tetapi juga sudah memiliki informasi yang rinci (baik yang relevan maupun

yang tidak relevan dengan kemampuan mereka untuk melakukan pekerjaan itu) mengenai para

kandidat di dalam departemen mereka, kita dapat menyimpulkan bahwa pasar sangat bermanfaat

untuk membantu para manajer menghindari penggunaan informasi yang tidak relevan saat

mengevaluasi alternatif-alternatif yang sudah dikenal, sebuah alasan yang sering disebut-sebut

sebagai penyebab terjadinya kesalahan dalam perekrutan (Highhouse, 2008).

Sisipkan Tabel 4 & 5 di sini

Analisis Tambahan dan Pemeriksaan Kekokohan

Seperti yang telah disebutkan di atas, kekhawatiran potensial dengan penggunaan kuadrat

terkecil biasa dan model logit di atas adalah potensi endogenitas posting. Meskipun saya dapat

mengontrol kecenderungan tetap untuk pekerjaan tertentu yang akan diisi oleh posting atau

sponsor, saya tidak memiliki variabel yang memungkinkan saya untuk mengontrol kemungkinan

bahwa manajer hanya memposting pekerjaan jika mereka tidak dapat terlebih dahulu

mengidentifikasi kandidat yang memenuhi syarat melalui jaringan pribadi mereka. Namun,
 36
penting untuk dicatat bahwa jika ini memang benar terjadi, kita akan mengharapkan sifat endogen

dari pilihan ini untuk mempengaruhi kualitas hasil perekrutan dalam arah yang berlawanan; jika

para manajer kemungkinan besar akan menggunakan sponsorship ketika mereka dapat secara

pribadi mengidentifikasi kandidat yang sangat baik, perekrutan yang disponsori harus diharapkan

untuk
 37

mengungguli perekrutan yang dilakukan melalui posting. Namun demikian, secara empiris,

mengatasi potensi endogenitas ini tampaknya bijaksana dan saya melakukannya dengan

menggunakan pendekatan variabel instrumental.

Sebuah variabel instrumental harus berkorelasi kuat dengan variabel endogen

(penempatan) tetapi tidak dengan error term tahap kedua (Hamilton & Nickerson, 2003;

Semadeni, Withers, & Certo, 2014). Variabel yang tampaknya memenuhi dua kondisi ini adalah

persentase pekerjaan serupa yang diisi oleh posting di departemen lain dalam dua bulan

sebelumnya. Diskusi saya dengan manajer perekrutan dan staf SDM di HealthCo mengkonfirmasi

bahwa para manajer secara teratur, (a) bertanya kepada manajer lain bagaimana mereka baru-baru

ini mengisi pekerjaan yang serupa, atau (b) menghubungi bagian SDM dan bertanya bagaimana

posisi serupa telah diisi baru-baru ini. Meskipun kedua mekanisme tersebut diharapkan dapat

mempengaruhi keputusan manajer tentang bagaimana mengisi pekerjaan tersebut, namun

bagaimana pekerjaan serupa diisi di bagian lain dalam organisasi seharusnya tidak terlalu

berpengaruh langsung terhadap hasil atau gaji setelah perekrutan.

Tabel 6 membandingkan hasil untuk skor kompetensi dan kontribusi menggunakan OLS

(Model 1 dan 2) dengan mereka yang menggunakan pendekatan 2SLS tradisional (Model 3 dan 4)

dan pendekatan perlakuan 2SLS yang memperhitungkan sifat biner dari variabel endogen saya

(Model 5 dan 6). Estimasi tahap pertama menunjukkan bahwa saya tidak perlu khawatir tentang

kelemahan instrumen. Seperti yang dapat dilihat dengan membandingkan hasil dari persamaan

yang berbeda, hasilnya konsisten dengan hasil yang diperoleh dengan menggunakan OLS.

Bahkan, setelah saya mengendalikan bias endogenitas, efek dari posting skor kompetensi dan

kontribusi secara substansial lebih kuat, meningkat dari peningkatan sekitar sepersepuluh standar

deviasi menjadi hampir empat persepuluh standar deviasi untuk kontribusi dan lebih dari setengah

standar deviasi untuk kompetensi.

Sisipkan Tabel 6 di sini

 38
Tabel 7 menyajikan hasil untuk kinerja relatif dengan menggunakan model probit variabel

instrumental. Selain dari perubahan besaran koefisien, perbedaan utama antara hasil-hasil ini
 39

dan dari model logit adalah bahwa tidak ada hubungan yang signifikan antara bagaimana seorang

pekerja memasuki pekerjaan dan kemungkinan mereka berada di kuartil teratas dari distribusi

kinerja.

Secara keseluruhan, interpretasi dari hasil-hasil ini tetap tidak berubah - posting mendisiplinkan

para manajer untuk membuat keputusan perekrutan internal yang lebih baik, sebagian dengan

menghindari kesalahan perekrutan.

Sisipkan Tabel 7 di sini

Tabel 8 menyajikan hasil untuk perputaran dan kenaikan jabatan dengan menggunakan

model probit variabel instrumental, serta hasil untuk gaji dengan menggunakan pendekatan 2SLS

tradisional. Sekali lagi, hasilnya mengarah pada interpretasi yang sama dengan model variabel

non-instrumental, dengan perbedaan utama adalah besarnya koefisien. Satu perbedaan substantif

adalah bahwa spesifikasi Probit IV menunjukkan bahwa meskipun karyawan internal lebih kecil

kemungkinannya untuk keluar dari perusahaan dalam waktu 12 bulan, mereka tidak lebih atau

kurang mungkin untuk keluar dari perusahaan dalam periode waktu 24 bulan yang lebih lama.

Sisipkan Tabel 8 di sini

Pencarian Kabel

Meskipun posting dan sponsorship secara konseptual berbeda, seorang manajer bisa saja

memposting pekerjaan setelah memutuskan siapa yang akan mereka pilih. Meskipun pekerjaan

ini diposting, namun sebenarnya pekerjaan tersebut diisi melalui sponsor. Jika hal ini merupakan

kejadian yang umum terjadi dan/atau contoh-contoh ini sulit diidentifikasi secara empiris, maka

akan sulit untuk menginterpretasikan hasil penelitian saya. Apa yang disebut "pencarian melalui

kabel" ini kemungkinan besar paling sering terjadi di perusahaan yang mengharuskan manajer

untuk memposting semua pekerjaan.

Karena Health Co. telah menetapkan sistem formal untuk mengisi pekerjaan baik melalui posting
 40
maupun sponsor, para manajer dapat menghindari posting pekerjaan ketika mereka telah

memiliki kandidat dan dengan demikian menghindari potensi konsekuensi negatif yang muncul

dari karyawan lain yang merasa bahwa mereka berpartisipasi dalam proses seleksi yang tidak adil

(Billsberry, 2007). Namun demikian, saya melakukan

 41

dua pemeriksaan ketahanan tambahan untuk mengesampingkan kemungkinan bahwa hasil saya

terpengaruh oleh penyertaan pencarian melalui kabel. Pertama, saya meninjau masing-masing dari

1.695 perekrutan internal yang dilakukan antara tahun 2011 dan 2012 di mana saya memiliki data

lamaran yang terperinci. Kejadian-kejadian di mana hanya ada satu pelamar internal (dan tidak

ada pelamar eksternal) yang berpotensi menjadi pencarian melalui kabel, di mana para manajer

menyesuaikan deskripsi pekerjaan di sekitar kandidat tertentu dan/atau menghalangi karyawan

lain untuk melamar. Kurang dari 5% dari perekrutan internal yang memenuhi kriteria ini dan dari

jumlah tersebut, dua pertiganya dibuka selama lebih dari satu bulan, menunjukkan bahwa manajer

perekrutan berharap untuk menemukan kandidat tambahan. Sebanyak 30 lowongan internal yang

diposting (1,4%) berpotensi menjadi pencarian kabel, yang membuat saya yakin bahwa pencarian

semacam itu tidak mendorong hasil saya. Kedua, saya mengidentifikasi semua lowongan selama

periode ini yang dibuka seminggu atau kurang (n = 45), sebuah indikasi potensial lainnya bahwa

manajer perekrutan membuat lowongan untuk kandidat tertentu dan oleh karena itu ingin

membatasi jumlah kandidat lain yang akan melamar. Dalam kedua kasus tersebut, menghilangkan

pengamatan ini dari analisis tidak mengubah hasil secara substantif.

Kandidat Eksternal sebagai Penjelasan Alternatif

Saya melakukan serangkaian pemeriksaan ketahanan tambahan untuk memastikan bahwa

hasil posting tidak didorong oleh kekuatan pasar eksternal. Ketika memposting pekerjaan,

manajer memiliki pilihan untuk membatasi kompetisi untuk kandidat internal dengan hanya

memposting pekerjaan tersebut secara internal, atau membuka kompetisi untuk kandidat internal

dan eksternal dengan memposting pekerjaan tersebut secara eksternal. Sebuah penjelasan

alternatif untuk temuan kinerja mungkin adalah bahwa penyertaan kandidat eksternal

memungkinkan manajer untuk membandingkan kandidat internal dengan pasar, yang mengarah

pada perekrutan yang lebih baik ketika kandidat internal dipilih (Billsberry, 2007). Mungkin yang

lebih penting lagi, penjelasan alternatif untuk gaji awal yang lebih tinggi yang terkait dengan

penempatan mungkin adalah bahwa paparan ke pasar eksternal bertanggung jawab untuk
 42
menaikkan gaji awal untuk pekerjaan yang ditempatkan,
 43

dengan hasil yang didorong oleh kasus-kasus di mana kandidat eksternal dipertimbangkan tetapi

kandidat internal yang terpilih.

Untuk mengesampingkan penjelasan alternatif ini, saya menggunakan data dari tahun 2011

dan 2012 yang memungkinkan saya untuk mengidentifikasi jumlah kandidat internal dan eksternal

yang melamar untuk setiap lowongan pekerjaan yang diposting dan sejauh mana mereka berhasil

melewati proses perekrutan. Hal ini memungkinkan saya untuk mengidentifikasi lowongan

pekerjaan mana yang dibatasi untuk kandidat internal dan mana yang terbuka untuk kandidat

eksternal. Tabel 9 menunjukkan hasil analisis termasuk variabel dummy yang bernilai satu jika

kandidat eksternal dipertimbangkan untuk posisi tersebut. Karena data ini dimulai pada

pertengahan tahun 2011, saya tidak dapat menghitung hasil untuk perputaran dan kenaikan

jabatan. Dari 869 karyawan internal yang saya miliki peringkat kinerja dan data gajinya, 578

(66%) dipekerjakan tanpa mempertimbangkan kandidat eksternal; 291 (34%) karyawan internal

bersaing dengan setidaknya satu kandidat eksternal. Hasil penelitian menunjukkan bahwa

membuka lowongan untuk kandidat eksternal tidak berpengaruh terhadap kinerja (Model 1-5) atau

gaji awal (Model 6-7).

Sisipkan Tabel 9 di sini

DISKUSI

Penelitian ini memajukan pemahaman kita tentang bagaimana sumber daya manusia

dialokasikan di dalam perusahaan dengan menyoroti proses yang digunakan untuk memfasilitasi

perekrutan internal di dalam organisasi kontemporer. Terlepas dari prevalensi perekrutan internal

dan dampaknya terhadap nasib perusahaan dan pekerja, model perekrutan internal saat ini

sebagian besar masih didasarkan pada literatur yang mengeksplorasi kemajuan pasar tenaga kerja

internal yang sangat birokratis dan tidak memiliki banyak kemiripan dengan model yang lebih

kontemporer. Studi ini memberikan pembaruan yang sangat dibutuhkan untuk model-model
 44
sebelumnya, mengidentifikasi dan menggambarkan dua proses perekrutan internal yang paling

umum digunakan -
 45

penempatan yang berorientasi pada pasar dan sponsorship yang berorientasi pada hubungan. Saya

mengembangkan teori yang memprediksi efek relatif dari penempatan dan sponsorship terhadap

kualitas rekrutmen dan kompensasi dengan menyoroti bahwa penempatan dicirikan oleh dua fitur

yang mirip pasar, yaitu seleksi mandiri dan formalitas, yang sebagian besar tidak ada pada

sponsorship, yang justru melibatkan pencarian manajerial yang aktif dan ketergantungan terhadap

koneksi personalia untuk mengumpulkan informasi.

Saya berpendapat bahwa memperkenalkan dua fitur pasar ini akan meningkatkan

pengambilan keputusan dengan membantu para manajer mengatasi tantangan yang terkait

dengan menghasilkan dan mengevaluasi alternatif. Seleksi mandiri seharusnya menghasilkan

kumpulan alternatif yang lebih besar, mengurangi kemungkinan dan alternatif yang luar biasa

akan terlewatkan, sementara formalitas pasar seharusnya mendorong para manajer untuk

memasukkan informasi yang relevan dan menghindari informasi yang tidak relevan ketika

mereka mengevaluasi kualifikasi kandidat terhadap persyaratan pekerjaan. Konsisten dengan

argumen-argumen ini, saya menemukan bahwa karyawan yang dipekerjakan melalui posting

mengungguli rekan-rekan mereka yang dipekerjakan melalui sponsor dan lebih kecil

kemungkinannya untuk keluar dari perusahaan, meskipun tidak ada perbedaan yang nyata dalam

hal kemungkinan mereka untuk naik jabatan.

Namun, pengambilan keputusan yang lebih baik ini ada harganya. Saya berpendapat

bahwa karyawan yang dipekerjakan melalui penempatan yang lebih kompetitif akan lebih kecil

kemungkinannya untuk mengadopsi orientasi relasional dalam negosiasi gaji, sehingga mereka

akan lebih mungkin untuk memulai negosiasi gaji dan mengadopsi pendekatan yang lebih efektif

dan kompetitif ketika melakukannya. Meskipun saya tidak dapat menguji mekanisme ini secara

langsung (saya tidak mengamati perbedaan antara penawaran gaji awal dan akhir), hasilnya

konsisten dengan penjelasan ini, karena karyawan yang dipekerjakan melalui posting dibayar ~

4% lebih tinggi daripada karyawan yang disponsori yang dipekerjakan dalam pekerjaan yang

setara.
 46
Terlepas dari adanya premium posting, hasil ini menunjukkan bahwa posting

memungkinkan perusahaan untuk menciptakan dan mendapatkan nilai yang jauh lebih besar

daripada sponsorship. Sayangnya, data saya tidak memungkinkan saya untuk secara langsung

mengukur nilai yang diciptakan oleh tingkat kinerja individu yang lebih tinggi atau penghematan
 47

terkait dengan tingkat retensi yang lebih tinggi. Namun, dengan biaya penggantian karyawan yang

berkisar antara 20% hingga 200% dari gaji tahunan karyawan (Boushey & Glynn, 2012),

gabungan manfaat kinerja dan retensi yang terkait dengan perekrutan internal yang lebih baik

kemungkinan besar akan jauh melebihi biaya gaji yang lebih tinggi. Selain itu, premi penempatan

menunjukkan bahwa pekerja juga ikut serta dalam peningkatan penciptaan nilai, dengan

mendapatkan gaji yang lebih tinggi untuk pekerjaan yang setara. Oleh karena itu, proses

penempatan yang berorientasi pada pasar tampaknya menguntungkan perusahaan dan pekerja.

Seperti halnya semua studi perusahaan tunggal, diperlukan kehati-hatian dalam menggeneralisasi hasil
ini.

Meneliti mobilitas di dalam perusahaan membutuhkan tradeoff antara kedalaman dan

generalisasi, yang membutuhkan data internal yang terperinci yang mungkin sulit diperoleh dari

beberapa lokasi. Hal ini menjadi semakin sulit karena hanya sedikit perusahaan yang benar-benar

mengumpulkan data tentang proses perpindahan karyawan ke pekerjaan baru. Percakapan saya

dengan para pejabat HealthCo dan para pemimpin sumber daya manusia di beberapa organisasi

lain selama penelitian ini, bagaimanapun juga, tidak menemukan alasan untuk meyakini bahwa

praktik perekrutan internal HealthCo berbeda dengan organisasi besar lainnya.

Selain tidak dapat secara langsung mengukur penciptaan nilai, studi ini tidak

memperhitungkan biaya potensial yang terkait dengan penciptaan pasar internal untuk talenta di

luar premi gaji. Biaya tambahan mungkin termasuk investasi yang dilakukan untuk

mengimplementasikan dan memelihara papan lowongan kerja internal dan waktu yang dibutuhkan

untuk menyaring dan mengevaluasi lamaran pekerjaan. Sejalan dengan hal ini, jalan yang sangat

menjanjikan untuk penelitian di masa depan adalah mengeksplorasi apakah penempatan karyawan

membebankan biaya tambahan dengan menciptakan sekumpulan karyawan yang kalah dalam

kompetisi terbuka, seperti kinerja yang lebih rendah atau peningkatan perputaran.

Terlepas dari keterbatasan ini, penelitian ini memberikan beberapa kontribusi penting.

Dengan menyoroti cara-cara di mana pemilihan sendiri dan formalitas membentuk pengambilan
 48
keputusan dan negosiasi manajerial, penelitian ini menambah wawasan yang berkembang tentang

manfaat dan keterbatasan membawa

 49

mekanisme pasar di dalam perusahaan (Zenger et al., 2011). Sampai saat ini, penelitian ini lebih

banyak berfokus pada cara-cara pemberian insentif berkekuatan tinggi yang membentuk perilaku

individu (Ellig, 2001; Zenger & Hesterly, 1997; Zenger, 1992). Baru-baru ini, para peneliti mulai

mengeksplorasi cara-cara di mana perusahaan dapat memanfaatkan kekuatan agregasi informasi

dari pasar untuk meningkatkan pengambilan keputusan internal melalui penggunaan prediksi dan

pasar informasi, dengan gagasan bahwa informasi yang tersebar luas dapat diagregasi menjadi

sesuatu yang mirip dengan harga (Ellig, 2001; Felin & Zenger, 2011). Studi ini melengkapi dan

memperluas penelitian ini dengan menunjukkan bagaimana dua fitur pasar yang kurang

dieksplorasi - seleksi mandiri dan formalitas - meningkatkan pengambilan keputusan tanpa

adanya mekanisme harga. Selain itu, saya menunjukkan bahwa, berlawanan dengan apa yang

diprediksi oleh teori yang ada mengenai peran persaingan di pasar eksternal, menanamkan

persaingan di dalam perusahaan justru meningkatkan harga. Akhirnya, dalam mengidentifikasi

mekanisme tingkat mikro yang melaluinya proses alokatif tingkat makro ini membentuk perilaku

individu, penelitian ini berbicara pada literatur yang semakin banyak yang tertarik pada dasar-

dasar mikro organisasi strategis (Felin & Foss, 2005; Ployhart & Hale, 2014) serta literatur yang

lebih spesifik tentang dasar-dasar mikro keunggulan kompetitif berbasis sumber daya manusia

(Coff & Kryscynski, 2011).

 50

REFERENSI

Afuah, A., & Tucci, C. L. 2012. Crowdsourcing sebagai solusi untuk pencarian jarak jauh.
Academy of Management Review, 37(3): 355-375.

Agerström, J., & Rooth, D.-O. 2011. Peran stereotip obesitas otomatis dalam diskriminasi
perekrutan nyata. Jurnal Psikologysikologi Terapan, 96(4): 790-805.

Agrawal, A., Knoeber, CR, & Tsoulouhas, T. 2006. Apakah orang luar menghambat suksesi
CEO? Jurnal Keuangan Perusahaan, 12(3): 619-644.

Alexander, E. R. 1979. Desain alternatif dalam konteks organisasi: Sebuah studi percontohan.
Administrative Science Quarterly, 24(3): 382-404.

Allen, D. G., Bryant, P. C., & Vardaman, J. M. 2010. Mempertahankan bakat: Mengganti
kesalahpahaman dengan strategi berbasis bukti. Academy of Management Perspectives,
24(2): 48-65.

Althauser, R. P., & Kalleberg, A. L. 1981. Perusahaan, pekerjaan, dan struktur pasar tenaga kerja:
Sebuah analisis konseptual. Dalam I. Berg (Ed.), Perspektif Sosiologis tentang pasar
tenaga kerja: 119-
149. New York: Academic Press.

Argote, L., & Ingram, P. 2000. Transfer pengetahuan: Sebuah dasar untuk keunggulan kompetitif di
perusahaan.
Perilaku Organisasi dan Proses Keputusan Manusia, 82(1): 150-169.

Barney G. Glaser (Ed.). 1968. Karier organisasi: Sebuah buku sumber untuk teori. Chicago:
Aldine Publishing.

Bazerman, MH, & Chugh, D. 2005. Kesadaran yang terikat: Memfokuskan kegagalan dalam
negosiasi. Dalam L. Thompson (Ed.), Garis depan psikologi sosial: Negosiasi: 7-26. New
York: Psychological Press.

Bertrand, M., & Mullainathan, S. 2004. Apakah Emily dan Greg Lebih Mudah Dipekerjakan
Dibanding Lakisha dan Jamal? Sebuah Eksperimen Lapangan tentang Diskriminasi Pasar
Tenaga Kerja. American Economic Review, 94(4): 991-1013.

Bidwell, M. J. 2011. Membayar lebih banyak untuk mendapatkan lebih sedikit: Dampak
perekrutan eksternal versus mobilitas internal. Administrative Science Quarterly, 56(3):
369-407.

Bidwell, M. J., & Keller, J. 2014. Di dalam atau di luar? Bagaimana perusahaan menggabungkan
pasar tenaga kerja internal dan eksternal untuk mengisi pekerjaan. Academy of
Management Journal, 57(4): 1035-1055.

Bielby, W. T. 2000. Meminimalkan bias gender dan ras di tempat kerja. Comtemporary
Sociology, 29(1): 120-129.

Bies, R. J., Tripp, T. M., & Neale, M. A. 1993. Keadilan prosedural dan pencarian keuntungan:
Legitimasi yang dirasakan dari eksploitasi pasar. Jurnal Pengambilan Keputusan Perilaku,
 51
6(Juni): 243-256.
 52

Billsberry, J. 2007. Mengalami rekrutmen dan seleksi. West Sussex, Inggris: John Wiley &
Sons.

Borgatti, SP, & Cross, R. 2003. Pandangan relasional tentang pencarian informasi dan
pembelajaran dalam jaringan sosial. Ilmu Manajemen, 49(4): 432-445.

Bradach, J. L., & Eccles, R. G. 1989. Harga, otoritas, dan kepercayaan: Dari tipe-tipe ideal hingga bentuk-
bentuk jamak.
Tinjauan Tahunan Sosiologi, 15: 97-118.

Breaugh, J. A., & Starke, M. 2000. Penelitian tentang perekrutan karyawan: Begitu banyak
penelitian, begitu banyak pertanyaan yang tersisa. Jurnal Manajemen, 26(3): 405-434.

Brenner, L. A., Koehler, D. J., & Tversky, A. 1996. Tentang evaluasi bukti satu sisi.
Jurnal Pengambilan Keputusan Perilaku, 9: 59-70.

Burks, S., Cowgill, B., Hoffman, M., & Housman, M. 2013. Nilai perekrutan melalui rujukan.

Campbell, J. P., McCloy, R. A., Oppler, S. H., & Sager, C. E. 1993. Sebuah teori tentang kinerja. Dalam
N. Schmitt & W. C. Borman (Eds.), Seleksi personalia dalam organisasi: 35-70. San
Fransisco, CA: Jossey-Bass.

Campion, MA, Cheraskin, L., & Stevens, MJ. 1994. Anteseden yang berhubungan dengan karir
dan hasil dari rotasi pekerjaan. Academy of Management Journal, 37(6): 1518-1542.

Cappelli, P. 2008. Bakat sesuai permintaan: Mengelola talenta di era ketidakpastian: 304.
Cambridge, MA: Harvard Business School Publishing.

Cappelli, P., & Keller, J. 2014. Manajemen talenta: Pendekatan konseptual dan tantangan praktis.
Tinjauan Tahunan Psikologi Organisasi dan Perilaku Organisasi, 1: 305-331.

Capron, L., Dussauge, P., & Mitchell, W. 1998. Penempatan kembali sumber daya setelah
akuisisi horizontal di Eropa dan Amerika Utara, 1988-1992. Strategic Management
Journal, 19(7): 631-661.

Casciaro, T., & Lobo, M. S. 2008. Ketika kompetensi tidak relevan: Peran pengaruh interpersonal
dalam ikatan yang berhubungan dengan tugas. Administrative Science Quarterly, 53(4):
655-684.

Cascio, W. F. 1998. Psikologi terapan dalam manajemen sumber daya manusia (5th ed.). Upper
Saddle River, NJ: Prentice Hall.

Cascio, W. F., & Aguinis, H. 2008. Kepegawaian organisasi abad kedua puluh satu. Academy of
Management Annals, 2(1): 133-165.

Casson, M. 1982. Pengusaha: Sebuah teori ekonomi. Oxford, UK: Martin Robertson.
 53

Coff, R. W., & Kryscynski, D. 2011. Menggali fondasi mikro dari keunggulan kompetitif
berbasis sumber daya manusia. Journal of Management, 37(5): 1429-1443.

Cowen, T., & Parker, D. 1997. Pasar dalam perusahaan: Pendekatan proses pasar terhadap manajemen.
London: Institute of Economic Affairs.

Crispin, G., & Mehler, M. 2013. Sumber-sumber perekrutan 2013: Persepsi adalah kenyataan.

Cross, R., Parker, A., Prusak, L., & Borgatti, S. P. 2001. Mengetahui apa yang kita ketahui:
Mendukung penciptaan dan berbagi pengetahuan dalam jejaring sosial. Organization
Dynamics, 30(2): 100-120.

Cross, SE, Bacon, PL, & Morris, ML. 2000. Konseptual diri dan hubungan yang saling
bergantung secara relasional. Jurnal Psikologi Kepribadian dan Sosial, 78(4): 791-808.

Curhan, JR, Elfenbein, HA, & Xu, H. 2006. Apa yang dihargai orang ketika mereka bernegosiasi?
Memetakan domain nilai subjektif dalam negosiasi. Jurnal Psikologi Kepribadian dan
Sosial, 91(3): 493-512.

Curhan, JR, Neale, MA, Ross, L., & Rosencranz-Engelmann, J. 2008. Akomodasi relasional
dalam negosiasi: Pengaruh egalitarianisme dan gender terhadap efisiensi ekonomi dan
modal relasional. Perilaku Organisasi dan Proses Keputusan Manusia, 107(2): 192-205.

Cyert, R. M., & March, J. G. 1963. Teori perilaku perusahaan. Englewood Cliffs, N.J.:
Prentice-Hall.

Dencker, J. C. 2008. Restrukturisasi perusahaan dan perbedaan jenis kelamin dalam promosi manajerial.
American Sociological Review, 73(3): 455-476.

Diprete, T. A. 1987. Mobilitas horizontal dan vertikal dalam organisasi. Administrative Science
Quarterly, 32(3): 422-444.

Doeringer, P. B., & Piore, M. J. 1971. Pasar tenaga kerja internal dan analisis tenaga kerja.
Lexington, MA: D.C. Heath and Company.

Dougherty, T. W., Turban, D. B., & Callender, J. C. 1994. Mengkonfirmasi kesan pertama dalam
wawancara kerja: Sebuah studi lapangan tentang perilaku pewawancara. Jurnal Psikologi
Terapan, 79(5): 659-665.

Drazin, R., & Rao, H. 2002. Memanfaatkan pengetahuan manajerial untuk mengimplementasikan
perluasan lini produk: bagaimana cara saling menemukan keluarga mengalokasikan manajer
portofolio ke dana lama dan baru? Academy of Management Journal, 45(3): 609-619.

Ellig, J. 2001. Pasar internal dan teori perusahaan. Ekonomi Manajerial dan Keputusan, 22(45):
227-237.

ERC. 2010. Survei praktik kepegawaian ERC. Desa Mayfield, OH.

 54

Felin, T., & Foss, N. J. 2005. Organisasi strategis: sebuah bidang untuk mencari landasan mikro.
Organisasi Strategis, 3(4): 441-455.

Felin, T., & Zenger, T. R. 2011. Agregasi informasi, pencocokan, dan hibrida hirarki pasar yang
radikal: Implikasi untuk teori perusahaan. Organisasi Strategis, 9(2): 163-173.

Fernandez, R. M., Castilla, E. J., & Moore, P. 2000. Modal sosial di tempat kerja: Jaringan dan
pekerjaan di pusat telepon. American Journal of Sociology, 105(5): 1288-1356.

Gelfand, M. J., Major, V. S., Raver, J. L., Nishii, L. H., Brien, O., Michele, J., dkk. 2006.
Bernegosiasi secara relasional: Dinamika diri relasional dalam negosiasi. Academy of
Management Review, 31(2): 427-451.

Gemünden, H. G., & Hauschildt, J. 1985. Jumlah alternatif dan efisiensi dalam berbagai jenis
keputusan manajemen puncak. European Journal of Operational Research, 22(2): 178-190.

Granovetter, M. S. 1985. Tindakan ekonomi dan struktur sosial: Masalah keterlekatan.

Jurnal Sosiologi Amerika, 91(3): 481-510.

Groysberg, B., Lee, L.-E., & Nanda, A. 2008. Dapatkah mereka membawanya? Portabilitas
kinerja pekerja pengetahuan bintang. Management Science, 54(7): 1213-1230.

Hamilton, B. H., & Nickerson, J. A. 2003. Mengoreksi endogenitas dalam penelitian manajemen
strategis. Organisasi Strategis, 1(1): 51-78.

Hauser, JR, & Wernerfelt, B. 1990. Sebuah model biaya evaluasi dari kumpulan pertimbangan.
Jurnal Penelitian Konsumen, 16(4): 393-408.

Highhouse, S. 2008. Ketergantungan pada intuisi dan subjektivitas dalam seleksi karyawan.
Psikologi Industri dan Organisasi, 1(3): 333-342.

Ibarra, H. 1993. Jaringan pribadi perempuan dan minoritas dalam manajemen: Sebuah kerangka
kerja konseptual. Academy of Management Review, 18(1): 56-87.

Iyengar, S. S., & Kamenica, E. 2010. Proliferasi pilihan, pencarian kesederhanaan, dan alokasi aset.
Jurnal Ekonomi Publik, 94(7-8): 530-539.

Jackson, C. K. 2013. Kualitas kecocokan, produktivitas pekerja, dan mobilitas pekerja: Bukti
langsung dari guru. The Review of Economics and Statistics, 95(4): 1096-1116.

Jacoby, S. M. 1985. Mempekerjakan birokrasi: Manajer, serikat pekerja, dan transformasi

pekerjaan di industri Amerika. New York: Columbia University Press.

Jacoby, S. M. 2004. Mempekerjakan birokrasi: Manajer, serikat pekerja, dan transformasi

pekerjaan di abad ke-20. Mahwah, NJ: Lawrence Erlbaum.
 55

Jacoby, S. M. 2005. Korporasi yang tertanam (embedded corporation): Tata kelola perusahaan
dan hubungan kerja di Jepang dan Amerika Serikat. Princeton, NJ: Princeton University
Press.

Jovanovic, B. 1979. Pencocokan pekerjaan dan teori perputaran. Jurnal Ekonomi Politik, 87(5):
972-990.

Kahneman, D. 2011. Berpikir, cepat dan lambat. New York: Farrar, Straus dan Giroux.

Kanter, R. M. 1977. Laki-laki dan perempuan dalam korporasi. New York: Basic Books.

Knudsen, T., & Levinthal, D. A. 2007. Dua wajah pencarian: Generasi alternatif dan evaluasi
alternatif. Organization Science, 18(1): 39-54.

Loasby, B. J. 2000. Institusi pasar dan evolusi ekonomi. Jurnal Ekonomi Evolusioner, 10(3):
297-309.

London, M. 1996. Penempatan kembali dan pembelajaran berkelanjutan di abad ke-21: Pelajaran
keras dan contoh positif dari era perampingan. Academy of Management Executive, 10(4):
67- 79.

March, J. G., & Simon, H. A. 1958. Organisasi. New York: Wiley.

Marks, M., & Harold, C. 2011. Siapa yang meminta dan siapa yang menerima dalam negosiasi
gaji. Jurnal Perilaku Organisasi, 394: 371-394.

Marlowe, C. M., Schneider, S. L., & Nelson, C. E. 1996. Bias gender dan daya tarik dalam
keputusan perekrutan: Apakah manajer yang lebih berpengalaman lebih tidak bias? Jurnal
Psikologi Terapan, 81(1): 11-21.

Marsden, P. V., & Gorman, E. H. 1999. Modal sosial dalam praktik kepegawaian internal. Dalam
R. T. A. J. Leenders & S. M. Gabbay (Eds.), Modal sosial dan tanggung jawab
perusahaan: 180-196. Boston, MA: Kluwer Academic Publishers.

McEvily, B., Soda, G., & Tortoriello, M. 2014. Lebih formal: Menemukan kembali mata rantai
yang hilang antara organisasi formal dan struktur sosial informal. Academy of Management
Annals, 8(1): 299-345.

Mclntyre, R. M., Smith, D. E., & Hassett, C. E. 1984. Keakuratan penilaian kinerja yang
dipengaruhi oleh pelatihan penilai dan tujuan penilaian yang dirasakan. Jurnal Psikologi
Terapan, 69(I): 147- 156.

Medoff, J. L., & Abraham, K. G. 1981. Apakah mereka yang dibayar lebih banyak benar-benar
lebih produktif? Kasus pengalaman. Jurnal Sumber Daya Manusia, 16(2): 186-216.

Menard, C. 1995. Pasar sebagai institusi versus organisasi sebagai pasar? Menguraikan beberapa
konsep fundamental. Jurnal Perilaku Ekonomi dan Organisasi, 28: 161-182.
 56

Miller, R. A. 1984. Pencocokan pekerjaan dan pilihan pekerjaan. Jurnal Ekonomi Politik, 92(6):
1086-1120.

Miner, A. S. 1987. Pekerjaan-pekerjaan istimewa dalam organisasi-organisasi formal.

Administrative Science Quarterly, 32(3): 327-351.

Mintzberg, H. 1994. Kebangkitan dan kejatuhan perencanaan strategis. New York: Free Press.

Mintzberg, H., Raisinghani, D., & Théorêt, A. 1976. Struktur proses keputusan yang "tidak
terstruktur". Administrative Science Quarterly, 21(2): 246-275.

Mortensen, D. T. 1988. Upah, pemisahan, dan masa kerja: Pelatihan khusus di tempat kerja atau
pencocokan? Jurnal Ekonomi Ketenagakerjaan, 6(4): 445-471.

Nahapiet, J., & Ghoshal, S. 1998. Modal sosial, modal intelektual, dan keunggulan organisasi.
Academy of Management Review, 23(2): 242-266.

Nickerson, J. A., & Zenger, T. R. 2004. Teori berbasis pengetahuan tentang perusahaan:
Perspektif pemecahan masalah. Organization Science, 15(6): 617-632.

Nutt, P. C. 1998. Bagaimana para pengambil keputusan mengevaluasi alternatif dan pengaruh kompleksitas.
Ilmu Manajemen, 44(8): 1148-1166.

Ocasio, W. 1997. Menuju pandangan berbasis perhatian terhadap perusahaan. Strategic, 18 (Edisi
Khusus Musim Panas): 187-206.

Oracle. 2012. Mobilitas bakat di Inggris.

Pinfield, L. T. 1995. Operasi pasar tenaga kerja internal: Praktik-praktik kepegawaian dan
rantai lowongan kerja. New York: Plenum Press.

Piore, M. J. 2002. Tiga puluh tahun kemudian: Pasar tenaga kerja internal, fleksibilitas dan ekonomi baru.
Jurnal Manajemen dan Tata Kelola, 6(4): 271-279.

Ployhart, R. E., & Hale, D. 2014. Dasar-dasar mikro psikologis yang menarik dari strategi dan
keunggulan kompetitif. Tinjauan Tahunan Psikologi Organisasi dan Perilaku Organisasi,
1(1): 145-172.

Podolny, J. M., & Baron, J. N. 1997. Sumber Daya dan Hubungan: Jaringan Sosial dan Mobilitas
di Tempat Kerja. American Sociological Review, 62(5): 673.

Polyani, K. 1957. Transformasi besar. Boston: Beacon Press.

Powell, W. W., & Snellman, K. 2004. Ekonomi pengetahuan. Tinjauan Tahunan Sosiologi,
30(1): 199-220.
 57

Pulakos, E. D. 1984. Perbandingan program pelatihan rater: Pelatihan kesalahan dan pelatihan
akurasi. Jurnal Psikologi Terapan, 69(4): 581-588.

Pulakos, E. D., & O'Leary, R. S. 2011. Mengapa manajemen kinerja rusak? Psikologi Industri dan
Organisasi, 4: 146-164.

Reskin, B. F., & Roos, P. A. 1990. Antrian pekerjaan, antrian gender: Menjelaskan terobosan
perempuan ke dalam pekerjaan laki-laki. Temple University Press.

Rogan, M., & Sorenson, O. 2014. Memilih mitra (yang buruk): Perspektif relasional tentang
akuisisi. Administrative Science Quarterly, 59(2): 301-329.

Rosenbaum, J. E. 1979. Mobilitas karir organisasi: Kesempatan promosi dalam perusahaan

selama periode pertumbuhan dan kontraksi. American Journal of Sociology, 85(1): 21-48.

Rosenbaum, J. E. 1990. Model-model struktural karir organisasi: Sebuah tinjauan kritis dan arah
baru. Dalam R. L. Breiger (Ed.), Mobilitas Sosial dan Struktur Sosial: 272-307.
Cambridge: Cambridge University Press.

Seidel, M.-D. L., Polzer, J. T., & Stewart, K. J. 2000. Teman-teman di tempat tinggi: Pengaruh
jaringan sosial terhadap diskriminasi dalam negosiasi gaji. Administrative Science
Quarterly, 45(1): 1-24.

Semadeni, M., Withers, M. C., & Certo, S. T. 2014. Bahaya endogenitas dan variabel
instrumental dalam penelitian strategi: Pemahaman melalui simulasi. Jurnal Manajemen
Strategis.

Shocker, A. D., Ben-Akiva, M., Boccara, B., & Nedungadi, P. 1991. Pengaruh rangkaian
pertimbangan pada pengambilan keputusan dan pilihan konsumen: Isu, model, dan saran.
Marketing Letters, 2(3): 181-197.

Simon, HA. 1955. Sebuah model perilaku pilihan rasional. Quarterly Journal of Economics, 69:
99-118.

Sirmon, D. G., Gove, S., & Hitt, M. A. 2008. Manajemen sumber daya dalam persaingan
kompetitif diadik: Efek dari bundling dan penyebaran sumber daya. Academy of
Management Journal, 51(5): 919-935.

Sirmon, D. G., & Hitt, M. A. 2009. Keterkaitan dalam kapabilitas manajerial yang dinamis: Efek
saling ketergantungan dari investasi sumber daya dan depolitisasi pada kinerja perusahaan.
Strategic Management Journal, 1394: 1375-1394.

Slichter, SH, Healy, JJ, & Livernash, ER. 1960. Dampak dari tawar-menawar kolektif
manajemen. Washington, D.C.: The Brookings Institution.

Sorenson, O., & Waguespack, D. M. 2006. Struktur dan pertukaran sosial: Dinamika konfirmasi
diri di Hollywood. Administrative Science Quarterly, 51: 560-589.
 58

Stein, J. C. 1997. Pasar modal internal dan kompetisi untuk sumber daya perusahaan. Journal of
Finance, 52(1): 111-133.

Strum, S. 2001. Diskriminasi ketenagakerjaan generasi kedua: Sebuah pendekatan struktural.

Columbia Law Review, 101: 458-568.

Taleo Research. 2005. Mobilitas internal. Redwood Shores, CA.

Tetlock, P. E. 1992. Dampak akuntabilitas pada penilaian dan pilihan. Dalam M. P. Zanna (Ed.),
Kemajuan dalam psikologi sosial eksperimental (25th ed.): 331-376. San Diego, CA:
Academic Press.

Thompson, L., & Deharpport, T. 1998. Hubungan, ketidakcocokan tujuan, dan orientasi komunal
dalam negosiasi. Psikologi Sosial Dasar dan Terapan, 20(1): 33-44.

Tsai, W. 2002. Struktur sosial "kooperasi" dalam organisasi multi unit: Koordinasi, kompetisi,
dan berbagi pengetahuan dalam organisasi. Organization Science, 12(2): 179- 190.

Tsai, W., & Ghoshal, S. 1998. Modal sosial dan penciptaan nilai: Peran jaringan antar perusahaan.
Academy of Management Journal, 41(4): 464-476.

Turner, RH. 1960. Mobilitas yang disponsori dan kontes dan sistem sekolah. American
Sociological Review, 25(6): 855-867.

Williams, P. F. 1987. Kepedulian yang sah terhadap keadilan. Akuntansi, Organisasi dan
Masyarakat, 12(2): 169-189.

Zenger, T. R. 1992. Mengapa perusahaan hanya memberi penghargaan pada kinerja yang
ekstrim? Meneliti hubungan antara kinerja, upah, dan perputaran. Administrative Science
Quarterly, 37(2): 198-219.

Zenger, T. R. 2002. Membuat hibrida internal: Saling melengkapi, inisiatif perubahan bersama,
dan organisasi berbasis tim. Jurnal Internasional Ekonomi Bisnis, 9(1): 79- 95.

Zenger, T. R., Felin, T., & Bigelow, L. 2011. Teori-teori tentang batas perusahaan-pasar.
Academy of Management Annals, 5(1): 89-133.

Zenger, T. R., & Hesterly, W. S. 1997. Pemilahan perusahaan: Intervensi selektif, insentif
berkekuatan tinggi, dan unit-unit molekuler. Ilmu Organisasi, 8(3): 209-222.

Zottoli, M. A., & Wanous, J. P. 2000. Penelitian sumber rekrutmen: Status saat ini dan arah masa
depan. Human Resource Management Review, 10(4): 353-382.
 59

TABEL 1
Statistik Deskriptif & Korelasia

Variabel n Berarti S.D. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

1 posting 11803 0.43 0.5
2 var instrumental. 11331 0.44 0.27 0.30
3 skor kontribusi 9520 2.85 0.64 -0.05 -0.04
4 skor kompetensi 9590 2.55 0.53 -0.05 -0.03 0.67
5 25% teratas 10115 0.29 0.45 -0.01 0.02 0.43 0.45
6 bagian bawah 50 10115 0.49 0.5 0.02 -0.02 -0.46 -0.50 -0.60
7 25% terbawah 10115 0.25 0.43 0.03 -0.01 -0.40 -0.46 -0.35 0.59
8 omset (12) 8359 0.03 0.16 -0.01 0.01 -0.05 -0.03 -0.03 0.03 0.03
9 omset (24) 5508 0.12 0.32 -0.03 -0.02 -0.12 -0.11 -0.09 0.08 0.07 0.56
10 promosi (24) 5675 0.13 0.33 0.02 0.06 0.06 0.06 0.06 -0.06 -0.06 -0.01 -0.09
11 promosi (36) 3739 0.24 0.43 0.07 0.10 0.06 0.06 0.05 -0.06 -0.06 -0.01 -0.08 0.87
12 gaji (ln) 11776 11.02 0.4 -0.10 -0.18 0.08 0.13 0.03 -0.01 0.00 0.03 0.03 -0.15 -0.16
13 tingkat pekerjaan 11803 4.63 0.76 -0.15 -0.27 0.12 0.13 0.01 -0.01 -0.01 0.01 0.03 -0.32 -0.39 0.73
14 departemen yang 11803 0.58 0.49 -0.39 -0.06 0.17 0.17 0.11 -0.14 -0.14 -0.03 -0.11 -0.03 -0.10 -0.02 -0.01
sama
15 fungsi yang sama 11803 0.68 0.47 -0.14 -0.02 0.03 0.08 0.06 -0.05 -0.05 -0.01 -0.02 -0.09 -0.12 0.07 0.03 0.21
16 promosi 11803 0.48 0.5 0.15 0.13 0.03 0.05 0.08 -0.05 -0.02 -0.04 -0.08 -0.07 -0.09 -0.30 -0.16 0.08 0.05
17 ekspansi 11803 0.39 0.49 -0.10 -0.14 0.00 -0.03 -0.06 0.03 0.01 0.02 0.04 0.03 0.03 0.23 0.19 0.03 -0.17
18 transfer 11803 0.13 0.33 -0.09 0.01 -0.04 -0.03 -0.04 0.03 0.01 0.04 0.06 0.06 0.09 0.12 -0.04 -0.16 0.17
19 perempuan 11697 0.64 0.48 0.01 0.04 -0.05 -0.03 0.02 -0.02 -0.02 -0.03 -0.04 0.01 0.00 -0.22 -0.14 0.01 -0.03
20 minoritas 11697 0.24 0.42 0.03 0.05 -0.07 -0.06 -0.01 0.02 0.03 0.01 0.00 -0.03 -0.02 -0.11 -0.11 -0.03 0.02
21 putih 11803 0.76 0.43 -0.02 -0.06 0.07 0.06 0.01 -0.02 -0.03 -0.01 0.00 0.02 0.01 0.12 0.11 0.03 -0.02
22 hitam 11803 0.1 0.3 0.04 0.06 -0.09 -0.10 -0.01 0.02 0.04 0.00 -0.01 -0.01 -0.01 -0.12 -0.13 -0.04 0.01
23 Latin 11803 0.08 0.27 0.01 0.04 -0.03 0.02 0.00 0.01 -0.01 0.02 0.01 -0.03 -0.02 -0.10 -0.06 0.00 0.00
24 asia 11803 0.03 0.18 -0.03 -0.04 0.05 0.04 0.00 0.01 0.01 0.01 0.02 -0.02 -0.01 0.10 0.08 0.02 0.02
25 masa jabatan 11803 5.4 5.22 -0.04 0.00 0.10 0.14 0.02 -0.03 -0.03 -0.03 -0.07 -0.06 -0.06 0.14 0.15 0.01 -0.05
26 masa jabatan2 11803 56.33 122.2 -0.07 -0.02 0.05 0.08 0.01 -0.02 -0.01 -0.02 -0.05 -0.06 -0.07 0.12 0.12 0.04 -0.03
27 usia 11697 38.59 9.93 -0.15 -0.12 -0.02 0.03 -0.02 0.01 0.00 0.01 0.04 -0.15 -0.20 0.31 0.26 0.03 0.03
28 usia2 11697 1588.12 837.31 -0.15 -0.12 -0.03 0.02 -0.03 0.02 0.00 0.01 0.05 -0.14 -0.19 0.29 0.24 0.03 0.03
29 ukuran kelompok 10115 479.3 408.2 0.05 0.07 0.03 -0.01 0.02 -0.04 -0.06 0.01 -0.01 0.07 0.12 -0.18 -0.23 0.05 0.01
peringkat
30 gaji terakhir (ln) 11787 10.93 0.42 -0.17 -0.23 0.09 0.13 0.02 -0.01 -0.02 0.03 0.04 -0.13 -0.14 0.94 0.72 0.00 0.07
31 kontribusi terakhir 9866 2.89 0.61 -0.03 -0.05 0.52 0.42 0.24 -0.23 -0.19 -0.05 -0.09 0.04 0.08 0.09 0.12 0.10 0.00
32 kompetensi 9897 2.57 0.51 -0.02 -0.01 0.41 0.59 0.25 -0.25 -0.20 -0.05 -0.09 0.01 0.04 0.13 0.12 0.10 0.03
terakhir
33 tahun 11803 2010.51 1.31 0.04 0.08 0.02 0.02 -0.02 0.01 -0.01 0.23 0.47 0.28 0.54 0.04 0.04 -0.19 -0.11
16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32
17 ekspansi -0.77
18 transfer -0.37 -0.30
19 perempuan 0.01 -0.02 0.01
20 minoritas 0.07 -0.06 -0.02 0.02
21 putih -0.07 0.06 0.02 -0.02 -1.00
22 hitam 0.06 -0.05 -0.03 0.07 0.62 -0.60
23 Latin 0.04 -0.04 0.00 0.02 0.52 -0.51 -0.10
24 asia -0.01 0.00 0.01 -0.08 0.34 -0.33 -0.06 -0.05
25 masa jabatan -0.09 0.11 -0.02 0.08 -0.06 0.07 -0.01 -0.02 -0.03
26 masa jabatan2 -0.09 0.09 -0.01 0.08 -0.06 0.07 -0.02 -0.03 -0.04 0.92
27 usia -0.25 0.18 0.11 0.06 -0.08 0.08 -0.05 -0.03 -0.03 0.28 0.29
28 usia2 -0.23 0.16 0.11 0.06 -0.09 0.09 -0.06 -0.04 -0.03 0.26 0.28 0.99
29 ukuran kelompok 0.11 -0.11 0.00 -0.03 0.05 -0.06 0.01 0.04 0.06 0.03 0.01 -0.09 -0.09
peringkat
30 gaji terakhir (ln) -0.40 0.29 0.18 -0.21 -0.13 0.13 -0.13 -0.12 0.11 0.18 0.16 0.36 0.34 -0.16
31 kontribusi terakhir 0.05 -0.02 -0.05 -0.03 -0.06 0.06 -0.10 -0.01 0.04 0.14 0.07 -0.01 -0.02 0.01 0.09
32 kompetensi 0.08 -0.06 -0.03 0.00 -0.03 0.03 -0.10 0.05 0.02 0.20 0.11 0.05 0.04 -0.01 0.12 0.67
terakhir
33 tahun -0.06 0.06 0.01 0.01 0.01 -0.02 0.00 0.02 -0.01 0.07 0.04 0.03 0.02 0.11 0.05 0.01 0.01

a Unit analisis adalah karyawan internal

 60

TABEL 2
Kinerjaa
Model 1 Model 2 Model 3 Model 4 Model 5
OLS OLS Logit Logit Logit
Variabel Kontribusi Kompetensi 25% teratas Bawah 50% 25% terbawah
Posting 0.0542** 0.0621** 0.129* -0.133** -0.149*
[0.0155] [0.0123] [0.0548] [0.0516] [0.0634]
Level 4 -0.0406 -0.117* -0.34 0.275 0.14
[0.0862] [0.0551] [1.086] [1.001] [0.751]
Level 5 0.15 -0.395* -0.592 0.678 0.798*
[0.251] [0.166] [1.005] [0.916] [0.339]
Level 6 0.0491 -0.143 0.0583 0.128 0.566*
[0.260] [0.174] [0.200] [0.197] [0.271]
Departemen yang 0.208** 0.175** 0.392** -0.439** -0.495**
sama
[0.0159] [0.0126] [0.0586] [0.0531] [0.0635]
Fungsi yang sama 0.0316* 0.0751** 0.295** -0.155** -0.152*
[0.0149] [0.0125] [0.0591] [0.0534] [0.0644]
Promosi 0.00351 -0.0219 0.242** -0.0874 0.0152
[0.0216] [0.0168] [0.0820] [0.0718] [0.0857]
Perempuan 0.0026 0.0233† 0.115* -0.108* -0.101
[0.0146] [0.0121] [0.0550] [0.0511] [0.0617]
Hitam -0.156** -0.135** -0.116 0.129 0.325**
[0.0240] [0.0193] [0.0863] [0.0790] [0.0888]
Latino -0.0497† 0.0601** -0.000748 0.0865 0.0346
[0.0272] [0.0232] [0.0951] [0.0897] [0.110]
Asia 0.000733 -0.0291 -0.0805 0.187 0.214
[0.0328] [0.0285] [0.140] [0.129] [0.146]
Lainnya -0.0454 -0.0501 -0.0131 -0.0578 -0.0164
[0.0601] [0.0458] [0.193] [0.189] [0.229]
Masa kerja 0.0350** 0.0387** 0.0534** -0.0476** -0.0524**
[0.00379] [0.00345] [0.0148] [0.0129] [0.0153]
Kepemilikan (meter -0.00120** -0.00132** -0.00195** 0.00157** 0.00170**
persegi)
[0.000155] [0.000148] [0.000663] [0.000548] [0.000613]
Usia -0.00441** -0.00120† -0.00767** 0.00663* 0.00684*
[0.000828] [0.000676] [0.00291] [0.00266] [0.00324]
Ukuran Kelompok 0.0000295 -0.000145* -0.000385**
Peringkat
[6.87e-05] [6.64e-05] [8.65e-05]
Konstan 2.956@ 2.838@ -1.223! 0.0529 -0.973
[0.118] [0.0853] [0.524] [0.508] [0.599]
Pengamatan 9300 9300 9289 9276 9262
R-kuadrat 0.136 0.154
aKesalahan standar, dalam tanda kurung, dikelompokkan berdasarkan individu. Semua analisis juga menyertakan boneka untuk pekerjaan,
fungsi, negara bagian, dan tahun.
 61
† p <.10; ** p <.05; ** p <.01
 62

TABEL 3
Perputaran, Kemajuan & Gajia

Model 6 Model 7 Model 8 Model 9 Model 10 Model 11

Logit Logit Logit Logit OLS OLS
Variabel Perputaran12 Perputaran24 Promosi24 Promosi36 Gaji (ln) Gaji (ln)
Posting -0.202* -0.182* -0.0949 0.128† 0.0383** 0.0376**
[0.0986] [0.0761] [0.0645] [0.0733] [0.00256] [0.00272]
Level 4 0.754 0.058 -0.0866 -0.526 0.258** -0.0489
[0.861] [0.829] [0.222] [0.561] [0.0498] [0.0688]
Level 5 0.511 -0.546 -1.172** -1.506** 0.167** 0.0159
[0.467] [0.426] [0.148] [0.553] [0.0218] [0.0526]
Level 6 0.539† -0.365 -1.563** -2.166** 0.221** 0.0331
[0.311] [0.340] [0.361] [0.445] [0.0300] [0.0559]
Dept yang sama -0.0925 -0.259** 0.0402 0.0012 0.000872 0.00125
[0.0959] [0.0738] [0.0651] [0.0779] [0.00270] [0.00275]
Fungsi yang sama -0.0336 0.0874 -0.217** -0.0758 0.0166** 0.0152**
[0.0987] [0.0671] [0.0610] [0.0756] [0.00264] [0.00270]
Promosi -0.302** 0.0791 -0.274** -0.279** 0.0960** 0.0865**
[0.116] [0.0954] [0.0777] [0.0892] [0.00874] [0.00890]
Perempuan -0.236** -0.148* -0.121* -0.123† -0.0101** -0.00912**
[0.0871] [0.0648] [0.0577] [0.0666] [0.00230] [0.00235]
Hitam 0.0358 -0.0981 -0.269** -0.302** -0.00637* -0.00126
[0.131] [0.109] [0.0896] [0.0962] [0.00313] [0.00331]
Latino -0.116 -0.0186 -0.349** -0.395** 0.000767 0.00714
[0.165] [0.111] [0.101] [0.125] [0.00494] [0.00514]
Asia -0.0644 -0.0117 0.005 0.2 -0.00445 -0.00181
[0.214] [0.151] [0.159] [0.178] [0.00406] [0.00421]
Masa kerja -0.104** -0.0828** -0.00392 0.0333† -0.000281 -0.0009
[0.0223] [0.0159] [0.0152] [0.0187] [0.000559] [0.000588]
Kepemilikan 0.00299** 0.00212** 0.0000181 -0.00150† 0.00000442 2.66E-05
(meter persegi)
[0.000912] [0.000625] [0.000653] [0.000882] [2.07e-05] [2.09e-05]
Usia -0.00115 0.0114** -0.0144** -0.0153** -0.00153† -0.00163†
[0.00523] [0.00316] [0.00319] [0.00365] [0.000829] [0.000873]
Gaji terakhir (ln) 0.851** 0.863**
[0.00924] [0.00893]
Kontribusi terakhir 0.00860**
[0.00256]
Kompetensi 0.00354
terakhir
[0.00353]
Konstan -6.111** -2.416** 0.061 0.596 1.561** 1.342**
[0.757] [0.594] [0.345] [0.633] [0.117] [0.101]
Pengamatan 7451 5056 5178 3302 9292 8017
R-Squared 0.955 0.959
aKesalahan standar, dalam tanda kurung, dikelompokkan berdasarkan individu. Semua analisis juga menyertakan boneka untuk pekerjaan,
fungsi, negara bagian, dan tahun.
 63
† p <.10; ** p <.05; ** p <.01.
 64

TABEL 4
Sumber Perekrutan Internal melalui Posting dan Sponsor

Sumber Penyewaan Signifikansi

(Lokasi pekerjaan sebelumnya dalam organisasi) Posting Sponsorship Statistik

Fungsi yang berbeda 40% 27% p < .01

Departemen yang berbeda 64% 26% p < .01

Kota yang berbeda 15% 4% p < .01

Bangunan yang berbeda 42% 8% p < .01

Pemindahan (pekerjaan yang sama, 10% 15% p < .01

departemen yang berbeda)
Perluasan (tingkat yang sama, pekerjaan 32% 43% p < .01
yang berbeda)
Promosi (naik tingkat) 57% 42% p < .01
 65

TABEL 5
Kinerja, Perputaran, Gaji, dan Kemajuan dalam Departemena,b,c

Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 Model 7 Model 8 Model 9 Model 10
OLS OLS Logit Logit Logit OLS OLS Logit Logit OLS
Bawah 25%
Variabel Contrib. Komp. 25% teratas Belok12 Turn24 Prom24 Prom36 Gaji (ln)
50% terba
wah
Posting 0.0474** .0678** 0.160** -0.127* -0.165* -0.284* -0.151 -0.119 0.117 0.0468**
[0.0171] [0.0137] [0.0609] [0.0585] [0.0740] [0.123] [0.0988] [0.136] [0.150] [0.00325]
Level 4 0.0973 0.550** -1.277 -0.215 -0.418 0.442 0.897 0.328 -0.0378 0.231**
[0.0758] [0.0645] [1.505] [0.638] [0.890] [0.897] [0.763] [0.499] [0.524] [0.0637]
Level 5 0.308 0.281† -1.433 0.231 0.481 0.225 0.257 -2.087** -1.967** 0.156**
[0.239] [0.153] [1.444] [0.287] [0.394] [0.520] [0.229] [0.334] [0.388] [0.0259]
Level 6 0.237 0.537** 0.199 0.00687 0.271 0.367 0.329 -2.541** -2.888** 0.116**
[0.249] [0.162] [0.220] [0.231] [0.335] [0.381] [0.360] [0.845] [0.782] [0.0338]
Dept yang sama 0.152** 0.129** 0.319** -0.389** -.348** -0.0509 -0.0527 0.267† 0.127 0
[0.0190] [0.0152] [0.0688] [0.0624] [0.0765] [0.119] [0.0970] [0.149] [0.172] [0]
Fungsi yang 0.0196 0.0480** 0.261** -0.142* -0.138† 0.145 0.0614 -0.355** 0.054 0.0106**
sama
[0.0175] [0.0150] [0.0696] [0.0638] [0.0807] [0.127] [0.0868] [0.133] [0.166] [0.00336]
Masa kerja 0.0396** 0.0441** 0.0662** -00569** -.0637** -0.105** -.0875** 0.00176 0.0842* -0.000655
[0.00413] [0.00376] [0.0162] [0.0146] [0.0174] [0.0250] [0.0184] [0.0348] [0.0418] [0.000650]
Kepemilikan -.00138** -.00149** -.0023** .00183** .00198** .00315** .00239** -0.000855 -0.005* 0.00000997
(meter persegi)
[.000173] [.000163] [.000735] [.000623] [0.0007] [.000991] [0.0007] [0.00153] [0.00212] [2.42e-05]
Usia -.00470** -.00153* -.0085** .00841** .0101** 0.000178 .0110** -.0311** -.033** -0.00174†
[0000887] [.000728] [0.00317] [0.00294] [0.00360] [0.00603] [0.00383] [0.00668] [0.00760] [0.00101]
Konstan 2.830** 2.199** -1.486** 0.202 -0.723 -5.912** -3.468** 0.162 0.532 0.827**
[0.0877] [0.0738] [0.548] [0.528] [0.638] [1.115] [0.566] [0.746] [0.707] [0.134]

Obs 7568 7568 7559 7543 7505 5777 4044 4203 2720 7568
R-kuadrat 0.138 0.161 0.966
a Kesalahan standar, dalam tanda kurung, dikelompokkan berdasarkan individu. Semua analisis juga menyertakan boneka untuk pekerjaan, fungsi, negara bagian, tahun, etnis,
dan jenis kelamin.
b Model 3, 4, dan 5 berisi kontrol tambahan untuk ukuran kelompok kinerja
 66
cModel 10 mencakup kontrol tambahan untuk gaji terakhir, peringkat kompetensi terakhir, dan peringkat kontribusi terakhir
† p <.10; ** p <.05; ** p <.01
 67

TABEL 6
Variabel Instrumental Kinerja (Bagian 1)a
Model 1 Model 2 Model 3 Model4 Model5 Model 6
OLS OLS 2SLS 2SLS Tahap pertama 2SLS / Perawatan 2SLS / Perawatan Tahap pertama
Variabel Kontribusi Kompetensi Kontribusi Kompetensi Kontribusi Kompetensi
Posting 0.0542** 0.0621** 0.259** 0.271** 0.227** 0.199**
[0.0155] [0.0123] [0.0934] [0.0764] [0.0765] [0.0609]
Departemen yang sama 0.208** 0.175** 0.296** 0.262** -0.415** 0.283** 0.232** -1.265**
[0.0159] [0.0126] [0.0422] [0.0343] [0.01] [0.0353] [0.0281] [0.0344]
Fungsi yang sama 0.0316* 0.0751** 0.0387* 0.0871** -0.068** 0.0366* 0.0824** -0.194**
[0.0149] [0.0125] [0.0164] [0.0136] [0.01] [0.0164] [0.0131] [0.0379]
Promosi 0.00351 -0.0219 -0.0445 -0.0711** 0.216** -0.0375 -0.0555** 0.636**
[0.0216] [0.0168] [0.0300] [0.0237] [0.016] [0.0264] [0.0211] [0.0483]
Perempuan 0.0026 0.0233† -0.00083 0.0222† -0.00 -0.000927 0.0220† -0.00499
[0.0146] [0.0121] [0.0150] [0.0124] [0.01] [0.0143] [0.0114] [0.0342]
Hitam -0.156** -0.135** -0.148** -0.130** -0.040** -0.149** -0.133** -0.135**
[0.0240] [0.0193] [0.0245] [0.0197] [0.02] [0.0218] [0.0174] [0.0507]
Latino -0.0497† 0.0601** -0.0469 0.0657** -0.042* -0.0483† 0.0627** -0.141*
[0.0272] [0.0232] [0.0286] [0.0243] [0.02] [0.0247] [0.0196] [0.0579]
Masa kerja 0.0350** 0.0387** 0.0343** 0.0387** 0.00 0.0345** 0.0391** 0.0183*
[0.00379] [0.00345] [0.00387] [0.00340] [0.00] [0.00349] [0.00278] [0.00859]
Kepemilikan (meter -0.00120** -0.00132** -0.00115** -0.00128** -0.00** -0.00116** -0.00131** -0.00135**
persegi)
[0.000155] [0.000148] [0.000160] [0.000144] [0.00] [0.000146] [0.000117] [0.000369]
Usia -0.00441** -0.00120† -0.00406** -0.00062 -0.002** -0.00415** -0.000819 -0.00910**
[0.000828] [0.000676] [0.000878] [0.000720] [0.00] [0.000766] [0.000610] [0.00177]
IV: Posting lainnya 0.307** 1.007**
[.02] [0.0670]
lambda -0.102* -0.0793*
[0.0455] [0.0363]
Konstan 2.956** 2.838** 2.331** 2.061** 0.925*
[0.118] [0.0853] [0.194] [0.154] [0.451]
Pengamatan 9300 9300 8929 8929 8929 8929 8929 8929
R-kuadrat 0.136 0.154
aKesalahan standar, dalam tanda kurung, dikelompokkan berdasarkan individu. Semua analisis juga menyertakan boneka untuk pekerjaan, fungsi, negara bagian, dan tahun.
† p <.10; ** p <.05; ** p <.01
 68

TABEL 7
Variabel Instrumental Kinerja (Bagian 2)a
Model 1 Model 2 Model 3 Model 4 Model 5 Model 6
Logit Logit Logit IV Probit IV Probit IV Probit
Variabel 25% teratas Bawah 50% 25% terbawah 25% teratas Bawah 50% 25% terbawah
Posting 0.129* -0.133** -0.149* 0.305 -0.720** -0.572*
[0.0548] [0.0516] [0.0634] [0.209] [0.204] [0.227]
Departemen yang sama 0.392** -0.439** -0.495** 0.331** -0.540** -0.487**
[0.0586] [0.0531] [0.0635] [0.0938] [0.0914] [0.102]
Fungsi yang sama 0.295** -0.155** -0.152* 0.192** -0.141** -0.120**
[0.0591] [0.0534] [0.0644] [0.0378] [0.0364] [0.0405]
Promosi 0.242** -0.0874 0.0152 0.088 0.083 0.108
[0.0820] [0.0718] [0.0857] [0.0656] [0.0627] [0.0694]
Perempuan 0.115* -0.108* -0.101 0.0726* -0.0670* -0.0577†
[0.0550] [0.0511] [0.0617] [0.0321] [0.0311] [0.0348]
Hitam -0.116 0.129 0.325** -0.0611 0.0682 0.178**
[0.0863] [0.0790] [0.0888] [0.0490] [0.0475] [0.0519]
Latino -0.000748 0.0865 0.0346 0.0053 0.0266 -0.00066
[0.0951] [0.0897] [0.110] [0.0553] [0.0538] [0.0610]
Masa kerja 0.0534** -0.0476** -0.0524** 0.0319** -0.0271** -0.0300**
[0.0148] [0.0129] [0.0153] [0.00787] [0.00756] [0.00859]
Kepemilikan (meter -0.00195** 0.00157** 0.00170** -0.00113** 0.000815* 0.000941**
persegi)
[0.000663] [0.000548] [0.000613] [0.000334] [0.000318] [0.000363]
Usia -0.00767** 0.00663* 0.00684* -0.00379* 0.0023 0.00258
[0.00291] [0.00266] [0.00324] [0.00175] [0.00169] [0.00187]
pgsize 0.0000295 -0.000145* -0.000385** -0.00000173 -0.0000514 -0.000192**
[6.87e-05] [6.64e-05] [8.65e-05] [4.27e-05] [4.22e-05] [4.95e-05]
Konstan -1.223* 0.0529 -0.973 -0.482 -0.0919 -0.838
[0.524] [0.508] [0.599] [0.426] [0.491] [0.586]
Pengamatan 9289 9276 9262 8925 8916 8885
a Kesalahan standar, dalam tanda kurung, dikelompokkan berdasarkan individu. Semua analisis juga menyertakan boneka untuk pekerjaan, fungsi, negara bagian, dan tahun.

† p <.10; ** p <.05; ** p <.01

 69

TABEL 8
Perputaran, Kemajuan, dan Gaji menggunakan Variabel Instrumentala

Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 Model 7 Model 8 Model 9 Model 10 Model 11 Model 12
Logit Logit IV Probit IV Probit Logit Logit IV Probit IV Probit OLS OLS 2SLS 2SLS
Gaji
Variabel Perputaran12 Perputaran24 Perputaran12 Perputaran24 Prom24 Prom36 Prom24 Prom36 Gaji (ln) Gaji (ln) (ln) Gaji (ln)
Posting -0.202* -0.182* -1.408* -0.459 -0.0949 0.128† -1.599† 3.060† 0.0383** 0.0376** 0.0443** 0.0430**
[0.0986] [0.0761] [0.637] [1.163] [0.0645] [0.0733] [0.833] [1.649] [0.00256] [0.00272] [0.0157] [0.0154]
Dept yang -0.0925 -0.259** -0.635* -0.37 0.0402 0.0012 -0.604† 1.067† 0.000872 0.00125 0.0039 0.00368
sama
[0.0959] [0.0738] [0.291] [0.499] [0.0651] [0.0779] [0.358] [0.617] [0.00270] [0.00275] [0.00732] [0.00736]
Fungsi yang -0.0336 0.0874 -0.0697 0.0811 -0.217** -0.0758 -0.291** 0.041 0.0166** 0.0152** 0.0169** 0.0155**
sama
[0.0987] [0.0671] [0.111] [0.0855] [0.0610] [0.0756] [0.0767] [0.120] [0.00264] [0.00270] [0.00305] [0.00312]
Promosi -0.302** 0.0791 -0.0374 0.157 -0.274** -0.279** 0.095 -0.795* 0.0960** 0.0865** 0.0935** 0.0819**
[0.116] [0.0954] [0.191] [0.280] [0.0777] [0.0892] [0.207] [0.319] [0.00874] [0.00890] [0.0129] [0.0133]
Perempuan -0.236** -0.148* -0.245* -0.157* -0.121* -0.123† -0.120† -0.114 -0.0101** -0.0912** -0.110** -0.00958**
[0.0871] [0.0648] [0.0975] [0.0664] [0.0577] [0.0666] [0.0620] [0.0836] [0.00230] [0.00235] [0.00235] [0.00241]
Masa kerja -0.104** -0.0828** -0.103** -0.0782** -0.00392 0.0333† 0.0000092 0.0188 -0.000281 -0.0009 -0.00042 -0.00100†
[0.0223] [0.0159] [0.0262] [0.0183] [0.0152] [0.0187] [0.0175] [0.0231] [0.000559] [0.000588] [0.000567] [0.000602]
Gaji terakhir 0.851** 0.863** -0.0181* 0.866**
[0.00924] [0.00893] [0.000844] [0.00908]
Kontras 0.00860** 0.858** 0.00948**
terakhir
[0.00256] [0.00886] [0.00262]
Komputasi 0.00354 0.00243
terakhir
[0.00353] [0.00378]
Konstan -6.111** -2.416** -0.664 -2.26 0.061 0.596 1.448 -2.657 1.561** 1.342** 1.406** 1.091**
[0.757] [0.594] [0.795] [1.420] [0.345] [0.633] [0.934] [1.996] [0.117] [0.101] [0.120] [0.104]
7451 5056 6415 4789 5178 3302 4989 3134 9292 8017 8923 7715
R-Squared 0.955 0.959 0.946 0.949
aKesalahan standar, dalam tanda kurung, dikelompokkan berdasarkan individu. Semua analisis juga menyertakan boneka untuk pekerjaan, fungsi, negara bagian, tahun, usia, dan
etnis.
† p <.10; ** p <.05; ** p <.01
 70

TABEL 9
Pengaruh Kandidat Eksternala

Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 Model 7

OLS OLS Logit Logit Logit OLS OLS
Kontribusi Kompetensi 25% teratas Bawah 50% 25% terbawah Gaji (ln) Gaji (ln)
Kandidat Eksternal -0.039 -0.0535 -0.0588 -0.192 -0.186 -0.00683 -0.00537
[0.0441] [0.0377] [0.171] [0.170] [0.247] [0.00762] [0.00706]
Level 4 -0.972** -0.365† 1.005 -0.494 0.738 0.0533 0.242*
[0.344] [0.218] [1.252] [1.244] [1.841] [0.0377] [0.105]
Level 5 -0.406 0.239 0.7 -0.313 1.535 0.106* 0.271*
[0.452] [0.248] [1.131] [1.106] [1.573] [0.0523] [0.124]
Level 6 -2.000** -0.855** 0.369 -0.783 1.026 0.177** 0.257**
[0.360] [0.233] [0.716] [0.778] [1.074] [0.0559] [0.0704]
Departemen yang sama 0.137** 0.104** 0.577** -0.366† -0.278 0.0155* 0.0220**
[0.0493] [0.0393] [0.187] [0.188] [0.282] [0.00786] [0.00775]
Fungsi yang sama 0.0165 0.0582 0.0209 0.0503 -0.420† 0.0226** 0.0161†
[0.0449] [0.0385] [0.189] [0.185] [0.250] [0.00855] [0.00878]
Kemajuan 0.0173 0.0394 0.00756 0.0149 0.117 0.0187 0.0311
[0.0686] [0.0567] [0.267] [0.278] [0.420] [0.0408] [0.0424]
Perempuan 0.00303 0.0419 0.228 -0.195 -0.323 -0.0104 -0.00521
[0.0426] [0.0346] [0.169] [0.168] [0.241] [0.00800] [0.00832]
Hitam -0.0867 -0.064 -0.16 0.136 0.593† -0.0121 -0.0094
[0.0733] [0.0601] [0.268] [0.247] [0.326] [0.00876] [0.00887]
pgsize 0.000636** -0.000620** -0.00110**
[0.000200] [0.000197] [0.000316]
Konstan 3.116** 2.265** -2.869* 2.485† 0.991 1.568** 1.809**
[0.301] [0.156] [1.385] [1.332] [1.934] [0.405] [0.522]
Pengamatan 869 869 869 869 869 869 752
R-kuadrat 0.168 0.178 0.955 0.96
aKesalahan standar, dalam tanda kurung, dikelompokkan berdasarkan individu. Semua analisis juga menyertakan boneka untuk pekerjaan, fungsi, negara bagian, tahun, dan
usia.
† p <.10; ** p <.05; ** p <.01
57
 1

INTERNAL MARKETS, PERSONAL NETWORKS, AND HUMAN RESOURCE

ALLOCATION: HOW MANAGERS CREATE VALUE THROUGH INTERNAL HIRING

JR Keller
Wharton School of Business
Working Paper

ABSTRACT

Internal hiring matches current employees to new jobs within an organization and represents a

critical yet overlooked source of value creation, enabling managers to generate greater value from

their existing stock of human resources by creating complementary matches between people and

jobs. Yet despite the fact that more than half of all jobs are filled internally, our knowledge of how

workers are allocated to jobs within organizations remains grounded in work on bureaucratic

internal labor markets and intraorganizational careers describing internal labor markets that bear

little resemblance to their contemporary counterparts. In this paper, I describe the effects of two

processes that have emerged to replace the use of bureaucratic rules for facilitating internal

mobility – market-oriented posting and relationship-oriented sponsorship – on two sets of

outcomes which link directly to value creation and value capture – quality of hire and

compensation. Using data on over 11,000 internal hires made over a five year period within a large

US health insurance company, I find that market-oriented posting results in better hires but at a

higher cost. In addition to providing a more complete picture of hiring and mobility, this work

sheds light on the tradeoffs associated with the use of markets and networks for allocating

resources within firms.

 2

Amid conversations about boundaryless careers, free-agent workers and hired guns, it is

easy to forget that a substantial amount of mobility still takes place within organizations. In fact,

nearly half of all open jobs in large organizations - and substantially more at the executive levels -

are filled internally (e.g. Crispin & Mehler, 2013). Internal hiring is the primary process used to

allocate human resources within firms and creates opportunities for internal mobility, which serves

many useful functions. It facilitates the transfer of knowledge across internal boundaries (Argote

& Ingram, 2000), motivates employees by signaling opportunities for future advancement

(Bidwell & Keller, 2014), encourages the development of firm-specific skills (Campion,

Cheraskin, & Stevens, 1994), increases worker satisfaction, performance and productivity

(Jackson, 2013), and decreases dysfunctional turnover (Allen, Bryant, & Vardaman, 2010). With

human resources now representing the most important resource in most firms (Powell &

Snellman, 2004) and in light of recent research demonstrating the high costs and even higher

failure rates associated with external hires at all organizational levels (e.g. Bidwell, 2011;

Groysberg, Lee, & Nanda, 2008), the ability to find and create complementary matches between

people and jobs within the firm represents a key source of value creation in modern organizations

(Zenger, Felin, & Bigelow, 2011).

While scholars have explored how other key resources, such as financial capital (e.g. Stein,

1997) and managerial attention (e.g. Ocasio, 1997), are allocated within firms, we know

surprisingly little about the contemporary internal allocation of human resources. Recent work has

documented how job characteristics shape both whether a job is likely to be filled internally

(Bidwell & Keller, 2014) and the qualifications (e.g. experience) of the candidates likely to be

placed into the job (Drazin & Rao, 2002). Yet internal hiring processes – that is, the ways in which

managers search for, evaluate, and select among potential internal candidates – have received little

systematic attention. However, research has shown that the use of formal versus informal external
 3

hiring processes (e.g. the use of referrals versus job postings) shapes not only who is hired, but

also their pay, performance, and turnover (Burks, Cowgill, Hoffman, & Housman, 2013;

Fernandez, Castilla, & Moore, 2000; Seidel, Polzer, & Stewart, 2000), as has recent research

comparing internal versus external hiring (Agrawal, Knoeber, & Tsoulouhas, 2006; Bidwell,

2011). We therefore might expect any variations among internal hiring processes to have similarly

significant consequences for both workers and firms.

Studying internal hiring also promises to contribute important insights to ongoing

conversations about the changing nature of internal resource allocation more generally. As firm

have transitioned away from hierarchical structures characterized by centralized decision-making

and towards flatter, leaner structures characterized by decentralized decision-making, bureaucratic

internal labor markets have gradually disintegrated (Cappelli & Keller, 2014), internal markets and

social networks have emerged as the primary mechanisms through with current workers are

matched to new jobs, echoing broader changes in how resources are allocated within

organizations. Indeed, the failures of bureaucratic planning systems in contemporary firms have

received considerable attention (Cowen & Parker, 1997; McEvily, Soda, & Tortoriello, 2014;

Mintzberg, 1994), with recent work identifying internal markets and network forms of

coordination as the two primary substitutes for bureaucratic control over the allocation of key

resources. Work on internal markets has generally emphasized how the use of market pricing

within the firm may lead to improved managerial decision-making regarding resource allocation,

as all of the information regarding the resource being considered is reflected in its price (Ellig,

2001; Felin & Zenger, 2011). Other work on internal markets has explored how high-powered

incentives can be designed to reduce coordination costs by aligning the interests of managers

making allocative decisions with the interests of the firm (Zenger & Hesterly, 1997; Zenger,

2002). Work on network forms of coordination has emphasized how the social relationships
 4

among actors within the firm can lead to improved allocative decisions by providing managers

with opportunities to share information on resources that would be otherwise unavailable to

centralized, higher-level decision-makers (Nahapiet & Ghoshal, 1998; Tsai & Ghoshal, 1998).

Because internal markets and social networks have been presented as alternatives to bureaucratic

control, much of the current literature has either compared internal markets with bureaucratic

control (e.g. Stein, 1997) or compared social networks with bureaucratic control (e.g. Tsai, 2002).

Much less work has explored the tradeoffs associated with using internal markets or personal

networks to allocate internal resources, despite the fact that market mechanisms and social

networks are likely to operate simultaneously within firms.

Contemporary internal labor markets represent a particularly fruitful context for exploring

these tradeoffs. Much of our understanding of internal hiring and mobility is grounded in the

foundational research on traditional, hierarchal internal labor markets and a closely related

literature on intraorganizational careers, which drew sharp distinctions between the bureaucratic

processes for allocating human capital operating within the firm and the market processes

operating outside the firm. This work described how internal hiring decisions were centralized in

personnel offices and governed by strict bureaucratic rules used to maintain lines of progression

along clearly defined job ladders, with employees exerting little control over their careers within

the firm (Althauser & Kalleberg, 1981; Glaser, 1968; Diprete, 1987; Doeringer & Piore, 1971;

Rosenbaum, 1990). Contemporary internal labor markets look dramatically different: hiring

decisions having been largely decentralized, with authority over promotions, transfers and external

hiring delegated to individual managers; organizational delayering, broader job descriptions, and

the rise of project-based work have all but eliminated clear paths for advancement; and employees

have been tasked with taking control of their careers. The cumulative effect of these changes has

rendered the use of bureaucratic rules for allocating human resources obsolete (Piore, 2002).
 5

Absent bureaucratic rules, the allocation of workers to jobs now takes place through two

very different processes (Marsden & Gorman, 1999; Pinfield, 1995). Posting is a market-oriented

process. A manager creates an internal market for an open job by broadcasting information about

the position throughout the organization via an internal job board and inviting current employees

to apply. The use of internal job posting systems is widespread, with 95% of organizations posting

jobs internally (Taleo Research, 2005). Sponsorship is a more relational process. A manager

identifies a candidate through her personal network and appoints that candidate to the job without

others being formally considered. Not only are posting and sponsorship the two most commonly

used internal hiring processes, they often operate concurrently within firms, as managers have

been granted substantial discretion both over the hiring decision and the hiring process (Marsden

& Gorman, 1999; Pinfield, 1995). With little if any work examining internal hiring in

contemporary organizations, it is unclear whether a market-oriented or relationship-oriented

allocative process is more likely to generate more valuable internal matches and what tradeoffs, if

any, might be associated with the use of posting versus sponsorship.

In this paper, I shed light on these tradeoffs by developing theory to explain how key

differences between the market-oriented posting process and relationship-oriented sponsorship

process affect two outcomes with implications for value creation and value capture: quality of hire

– as revealed by job performance, turnover, and subsequent advancement – and compensation.

Though the informational benefits associated with social networks have received considerable

attention (Borgatti & Cross, 2003; Bradach & Eccles, 1989; Cross, Parker, Prusak, & Borgatti,

2001; Granovetter, 1985), I predict that the market-oriented posting process will improve

managerial decision-making relative to the more relational sponsorship process, creating more

value though higher quality internal hires. In doing so, I describe how posting brings two features

of markets into the firm that are largely absent in sponsorship – self-selection and formality. The
 6

behavioral theory of the firm highlights two key challenges facing boundedly rational decision

makers, identifying a set of alternatives and evaluating the consequences of those alternatives

(Cyert & March, 1963; Simon, 1955). I argue that in contrast to the active managerial search

required in sponsorship, allowing workers to self-select into the consideration set for an open job

reduces the likelihood an exceptional internal candidate will be overlooked. Moreover, when

compared to the relative informality of a relationship-oriented allocative process, the formality of

the market will encourage managers to seek out information that allows them to better evaluate the

fit between a candidate’s qualifications and the requirements of the job, while also limiting the use

of irrelevant information that may lead to poor hiring decisions.

However, I also predict that these same market features will lead to higher salaries though

their effect on salary negotiations, limiting the amount of value a firm is able to capture through

better internal hiring decisions. Self-selection and formality introduce open competition into the

internal hiring process. While competition often drives down prices in external markets, I argue

that it will serve to increase prices for human capital within the firm, as candidates who are hired

though a competitive process are more likely to both initiate and adopt a competitive approach to

salary negotiations.

Taken together, these predictions suggest that there are important tradeoffs associated with

allocating human capital formally though markets or informally through the use of a manager’s

social network, with posting resulting in better hires but at a higher cost. However, with the

performance and retention benefits associated with better internal hires likely to far exceed the

higher salary costs, posting is likely to allow firms to both create and capture substantially more

value. These arguments suggest that a market-oriented process will also benefit workers, who

similarly capture more value through higher salaries.

 7

I use multiple modeling strategies in testing these predictions, which are largely supported

using five years (2008-2012) of personnel records covering all employees of a large health

insurance provider, as well as data on more than 350,000 internal and external job applications.

While personnel records have been previously used to identify which employees move to which

jobs (Bidwell, 2011; Dencker, 2008), data on the processes by which workers move jobs is rare.

These data are distinctive in that they clearly identify the mechanism used to facilitate each of the

11,000+ internal hires made during this period, allowing me to conduct what is, to my knowledge,

the first detailed study of the differences in outcomes associated with these two very different

internal hiring processes.

In unpacking the processes used to allocate workers to jobs within contemporary

organizations, this study helps to provide a more complete understanding of labor markets and

mobility. While robust literatures are developing to explore outcomes associated with different

external hiring processes (e.g. Fernandez et al., 2000) as well as the cost and quality tradeoffs

associated with internal versus external hiring (e.g. Bidwell, 2011), this work has tended to treat

internal hiring as a homogeneous process, unintentionally obscuring potential variations in the

processes used to move workers to new jobs within firms. More broadly, this study contributes to

a growing body of literature exploring how resources are bundled and deployed within

contemporary organizations (Sirmon, Gove, & Hitt, 2008; Sirmon & Hitt, 2009) by highlighting

the tradeoffs associated with using markets versus network approaches for allocating internal

resources. This study extends recent work exploring the effects of introducing market mechanisms

into firms (Felin & Zenger, 2011) by showing how mechanisms other than prices and high-

powered incentives can be leveraged to improve managerial decision-making, while also adding to

a small but important collection of studies which highlight the potential limitations of relational

exchange (Casciaro & Lobo, 2008; Rogan & Sorenson, 2014; Sorenson & Waguespack, 2006).
 8

Moreover, in identifying the micro-level mechanisms (decision-making and negotiations) through

which these two distinct organizational processes shape outcomes which link directly to value

creation and value capture (quality of hire and compensation), this study contributes to a

burgeoning literature focused on identifying the micro-foundations of human-resource based

competitive advantage (Coff & Kryscynski, 2011; Ployhart & Hale, 2014).

THEORY & HYPOTHESES

Internal Hiring in Contemporary Organizations

Before developing theory, it is useful to both define internal hiring and describe key

features of the allocative processes at the center of this study – posting and sponsorship. Once the

province of centralized personnel offices, managers are now largely responsible for staffing the

jobs that fall under their direct supervision, having been granted substantial discretion over which

candidates are considered and who is ultimately hired (Jacoby, 2004). Internal hiring occurs when

a manager fills an open job by hiring a worker currently employed by the organization in a

different job, resulting in the reallocation of that worker to a new set of organizational activities. It

is important to note that internal hiring is conceptually distinct from corporate restructurings, in

which large groups of workers – often entire departments or lines of business – are redeployed, en

masse, to new products or markets (Capron, Dussauge, & Mitchell, 1998; London, 1996).

Internal job posting systems date back to at least the 1940s. They were originally instituted

at the insistence of unions, who saw them as a way to curtail managerial discretion over staffing

decisions and thereby limit the scope for discrimination, favoritism, and nepotism to affect

opportunities for internal advancement. Management saw these systems as requiring them to

relinquish a long held right, so they fought to limit their scope. As a result, early job posting

systems typically covered a limited set of jobs and placed significant restrictions on which

employees were allowed to apply for those jobs (Cappelli, 2008). For those jobs that were
 9

covered, detailed selection criteria were established and often privileged seniority over ability,

providing individual managers with little choice over whom to hire (Jacoby, 1985; Slichter, Healy,

& Livernash, 1960).

Contemporary job posting systems are much more encompassing. Managers post

information about open jobs located at all but the very top levels of the organization to an internal

job board and invite interested candidates located throughout the organization to apply and

compete for the job, mirroring the typical process used to hire from the outside1. While the

requirements of the job are likely to influence who applies, candidates who do not possess the

qualifications described in the job posting but nevertheless would like to be considered are still

able to submit an application. The hiring manager evaluates the applications and extends an offer

to their preferred candidate, who may or may not negotiate the terms before deciding whether to

accept or reject the job. Once an offer is accepted, it is typical for unsuccessful internal candidates

to be notified and given the opportunity to learn why they did not get the job (Miller, 1984;

Pinfield, 1995).

Posting is often the default internal hiring process. Upwards of 95% of firms now use

internal job boards (Taleo Research, 2005) and many have developed policies encouraging

managers to post open positions internally, in large part to protect themselves from allegations of

discrimination (Strum, 2001). Yet despite these policies and the widespread availability of internal

job boards, managers can and often do bypass the posting system. This is in part due to the fact

that no state or federal laws require firms to post jobs and in part because internal staffing policies
1
When posting a job, a manager has the option of restricting the competition to internal candidates or
opening the competition to both internal and external candidates. While it is hard to pin down an exact
figure, a reasonable estimate is that around half of jobs above entry-level (which are by definition filled
externally) are posted internally and externally at the same time (Jacoby, 2005). This means that the posting
process may or may not result in an internal hire, depending on who is selected. For the purposes of this
paper, I present posting as an internal allocative process, because my interest is in looking at outcomes
associated with posting when an internal hire is made. In supplemental empirical analyses, I control for the
presence of external candidates on the key dependent variables.
 10

are both rarely enforced and allow for flexible interpretations (Pinfield, 1995). In fact, the Society

of Human Resource Management, the leading professional association for human resource

professionals, has recommended to hiring managers that “even if you have an internal process for

posting available jobs, there may be times when you decide not to follow this process” (SHRM,

2000: 7). In particular, managers are likely to bypass the posting system when they already have a

qualified candidate in mind for the job.

When managers do bypass the posting process, they turn a relational process I refer to as

sponsorship2. Social networks are a central feature of organizational life, as interactions among

individuals inevitably lead workers to develop networks of personal relationships (Kanter, 1977;

McEvily et al., 2014; Podolny & Baron, 1997). A sponsoring manager uses her personal social

network to identify potential candidates and then appoints her preferred candidate to the job absent

open competition. Though it is possible for sponsoring managers to exhaust their personal

connections as they search for potential candidates, the vast majority of internal hires made

through sponsorship involve the consideration of a single candidate with a direct connection to the

hiring manager, typically a current or previous colleague (Pinfield, 1995).

An abundance of evidence suggests that posting and sponsorship operate side-by-side

within firms, as equally viable ways to fill jobs. Marsden and Gorman (1999) examined survey

data on a representative sample of US work establishments and found that posting and sponsorship

were widely used in combination for filling vacancies with internal candidates. Pinfield’s (1995)

study of internal hiring in a forest-products company revealed that more than half of all positions

were filled through sponsorship despite company policy that all jobs be posted. And though over

95 percent of firms report posting at least some open jobs internally (Taleo Research, 2005)

2
I adopt the term sponsorship from classic literatures on career mobility within organizations (Rosenbaum,
1979) and upward social mobility more generally (Turner, 1960), in which it is used to describe systems in
which individuals selected for advancement are shielded from competition.
 11

research continues to demonstrate the importance of relationships on internal advancement

(Bielby, 2000; Ibarra, 1993; Podolny & Baron, 1997).

It is important to note that while posting and sponsorship are conceptually distinct internal

allocative processes, sponsorship may sometimes masquerade as posting in practice. That is, it is

possible that a manager may post a job having already identified the candidate they are going to

select through their social network, a practice referred to as a “wired search” (Bielby, 2000).

However, the prevalence of “wired searches” and other practices that might artificially restrict the

openness of the posting process, such as shaping job requirements around a single candidate or

discouraging employees from applying for certain jobs, are largely mitigated by concerns about

allegations of discrimination (Strum, 2001) as well as the possibility that employees who feel they

were mistreated or misled will simply leave the organization (Billsberry, 2007; Cappelli, 2008).

From the descriptions above, it is clear that market-oriented posting and relationship-

oriented sponsorship differ in several ways. In developing theory to understand the tradeoffs

between these two allocative process, I highlight the fact that posting is characterized by two

market-like features, self-selection and formality, that are largely absent in sponsorship, which

instead involves active managerial search and using personnel connections to gather information.

Bounded Rationality and Quality of Hire

A key facet of behavioral theories of the firm is the presumption that decision-makers are

boundedly rational. Because decision-makers are cognitively limited and have limited time,

information, and resources at their disposal, neither the complete set of alternatives from which a

decision maker can choose is known ex ante, nor are the consequences involved in choosing

among the available alternatives (March & Simon, 1958). As a result, failures in generating and

evaluating alternatives have been identified as two of the chief reasons why managers fail to

optimally allocate available resources (Afuah & Tucci, 2012; Knudsen & Levinthal, 2007).
 12

Self-selection and generating alternatives. Because the complete set of alternatives is not

known ex ante, boundedly rational decision-makers must engage in search to generate alternatives.

Search not costless, however, and one of the ways decision-makers economize is by considering

only a small portion of available alternatives (Hauser & Wernerfelt, 1990; Shocker, Ben-Akiva,

Boccara, & Nedungadi, 1991). While considering more alternatives does not guarantee a better

decision (Iyengar & Kamenica, 2010), decision-making success does tend to improve when more

alternatives are considered (Alexander, 1979; Gemünden & Hauschildt, 1985; Nutt, 1998), in part

because it lessens the odds that a superior alternative will be left out of the consideration set.

Self-selection enables managers to generate more alternatives without incurring many of

the costs associated with a broader search (Afuah & Tucci, 2012). Markets facilitate resource

allocation by providing a common platform for widely dispersed buyers and sellers to exchange

information about their needs and preferences (Zenger et al., 2011). Self-selection refers to the

ability of market participants to choose which available opportunities to pursue based on this

information rather than have those matches determined by managerial authority.

One way managers (as buyers) are able to harness the power of self- selection is by

broadcasting information about an opportunity and allowing interested sellers to self-select into

the set of alternatives to be considered by the manager. Rather than the manager assuming the

responsibility for generating alternatives through active search, sellers search for opportunities that

match their preferences. Because sellers have more information on their preferences than

managers and are likely to seek out alternatives that meet those preferences, searches that may be

considered distant from the perspective of the manager may often be considered local from the

perspective of the seller. By essentially transforming local search into distant search (Afuah &

Tucci, 2012), self-selection should be expected to expand the number of alternatives considered by

a manager – and more alternatives reduces the risk that a quality alternative will be overlooked.
 13

The posting process enables managers to harness the power of self-selection by

broadcasting information about an open job and allowing internal candidates located throughout

the organization, including those in in more distant areas, such as workers located in a different

location, department or function, to self-select into a queue of candidates competing for the job. In

contrast, sponsorship provides no formal mechanism for employees to express their interest in an

open job. Rather than broadcast information about the opening and allow interested candidates to

self-select into the candidate pool, the manager actively searches for alternatives (Mintzberg,

Raisinghani, & Théorêt, 1976) through their personal network. The pool of potential candidates is

therefore restricted by a manager’s previous experience and connections. While information about

internal candidates residing outside a manager’s immediate network is likely accessible (e.g.

through human resource information systems, talking with HR, etc.), obtaining this information

takes time and effort and the likelihood of finding a superior alternative is uncertain. Moreover,

managers are more likely to place a higher value on information obtained through their social

network, further discouraging the search for candidates residing outside of it. As a result,

sponsoring managers typically only consider those candidates with whom they are already

familiar, and even the most well-connected managers in large organizations are unlikely to be

familiar will all potential internal candidates.

In sum, by enabling candidates located both within and outside a hiring manager’s social

network to self-select into the consideration set, posting should therefore be expected to generate a

larger set of candidates than would otherwise be accessible through a manager’s personal network,

reducing the likelihood that an exceptional internal candidate will be overlooked.

Formality and evaluating alternatives. In addition to the challenges associated with

generating alternatives, boundedly rational managers also face difficulties evaluating alternatives.

In particular, recent work has called attention to the problems associated with bounded awareness,
 14

which refers to the propensity of boundedly rational individuals to fail to seek out or incorporate

relevant and accessible information into their decision-making process, instead relying on less

relevant information. Bounded awareness is one symptom of intuitive thinking, which often fails

to allow for the possibility that evidence needed to make a good decision is missing. As a result,

decision-makers tend to make decisions based on a subset of available information (Brenner,

Koehler, & Tversky, 1996; Kahneman, 2011). This “misalignment between the information

needed for a good decision and information included in the decision-making process” (Bazerman

& Chugh, 2005: 10) can lead to costly errors. Such errors are problematic in the hiring context,

where managers are notorious for their “stubborn reliance on intuition and subjectivity”

(Highhouse, 2008: 333).

Research suggests that interpersonal networks serve as conduits for information exchange

within organizations, providing managers with ready access to information that is richer, more

complete, and perceived as more trustworthy than information obtained from other sources

(Bradach & Eccles, 1989; Granovetter, 1985; Podolny & Baron, 1997). While this information

might be expected to improve decision-making, access to information alone is not enough to

ensure a good decision; managers must still select which information to use and which

information to ignore. In fact, studies spanning different contexts and levels of analysis, from

those examining individual hiring decisions3 to firm-level decisions about selecting merger and

3
There is ample evidence that managers rely on irrelevant information when making hiring decisions. For
example, managers often give substantial weight to performance in a previous job despite the fact that is
often a poor predictor of future performance (Cascio & Aguinis, 2008) and frequently allow attributes such
as gender, race, attractiveness, and weight to influence hiring decisions (e.g. Agerström & Rooth, 2011;
Bertrand & Mullainathan, 2004; Marlowe, Schneider, & Nelson, 1996). There is also an equal amount of
evidence that hiring managers fail to seek out relevant information, even when it is easily accessible. For
example, managers routinely bypass proven selection aids in favor of unstructured interviews (Highhouse,
2008) and use interviews to confirm their first impressions of candidates at the expense of gathering job-
relevant information (Dougherty, Turban, & Callender, 1994).
 15

acquisition partners (Rogan & Sorenson, 2014), show that managers routinely struggle to

objectively evaluate alternatives with whom they have an existing connection.

A market-oriented allocative process may therefore actually be superior to an information-

rich relational process in helping managers to overcome problems associated with bounded

awareness by shaping the information used to evaluate alternatives. Markets are institutions

supported by a system of rules and conventions designed to facilitate exchange among buyers and

sellers (Casson, 1982; Menard, 1995; Polyani, 1957). These rules and conventions – which I refer

to as formality – are both impersonal and non-arbitrary, thereby providing a stable framework for

transactions to take place (Menard, 1995: 1967). These rules and conventions serve several

functions, two of which are establishing the terms of exchange and defining what constitutes

legitimate behaviors (Loasby, 2000). In establishing the terms of exchange, the market generates

at set of evaluation criteria; in defining what constitutes legitimate behaviors, the market imposes

accountability on buyers. The presence of formal evaluation criteria and a high level of

accountability discipline managerial decision-making in different ways.

Evaluation criteria and relevant information. For markets to function, there must be a

mechanism for managers (as buyers) to broadcast information to potential sellers about the good

or service they are looking to procure (Zenger et al., 2011). Though managers may have

difficulties fully articulating their needs (Nickerson & Zenger, 2004), this initial information helps

to establish the terms of exchange in that it serves as the initial criteria against which sellers

evaluate their interest in pursuing the opportunity and as the initial criteria against which the buyer

evaluate those sellers who self-select into the consideration set. As a result, sellers are likely to

provide, and buyers are likely to seek out, information that enables buyers to evaluate alternatives

against a set of established criteria.

 16

One way posting introduces the formality of markets is through the use of formal job

descriptions. The posting process requires a manager to create a formal job description, necessary

for broadcasting information about an open job to potential candidates. Though it can be difficult

to develop accurate, comprehensive job descriptions (Backhaus, 2004; Sanchez & Levine, 2012),

the requirements defined at this initial stage nevertheless serve as a set of formal criteria against

which potential candidates are evaluated. In contrast, sponsorship does not require the manager to

create a formal job description prior to evaluating candidates. A manager must possess a

reasonable understanding of the job requirements as well as desirable candidate attributes, but

because the manager does not have to broadcast information about the open job, a formal job

description is not necessary. This allows the manager to informally mold the job requirements

around their preferred candidate rather than evaluating the candidate against the requirements of

the job (Miner, 1987; Sanchez & Levine, 2012). When compared to sponsorship, the presence of

formal evaluation criteria in posting is therefore more likely to prompt managers to reconsider the

relevance of the information used to evaluate candidates, with managers more likely to recognize

and seek out information enabling them to evaluate a candidate’s ability to perform well in the job.

Accountability and irrelevant information. Mechanisms that impose responsibility and

accountability ensure the continued participation of market participants by instilling confidence

that future transactions will be completed in a fair, honest, and orderly manner. Of particular note,

perceptions of the process by which firms make allocative decisions in a market can effect

perceptions of fairness, with decisions that appear to be free of bias and based on objective criteria

perceived as more fair and legitimate (Bies, Tripp, & Neale, 1993; Williams, 1987). Market-based

accountability should therefore encourage managers to avoid using information that would lead

their decisions to be perceived as subjective.

 17

Posting embeds a market-like accountability in the hiring process through the custom of

requiring managers to explain to unsuccessful internal candidates why they were not selected. For

every successful internal candidate there are likely to be multiple unsuccessful candidates.

Because these unsuccessful internal candidates remain employees, it is important to clearly

communicate the reasons why they did not get the job in order to minimize any sense of unfairness

that may decrease motivation, performance, and potentially lead to dysfunctional turnover. By

ensuring that hiring decisions have to be communicated and defended, the custom of explaining to

employees why they were not selected embeds accountability into the internal hiring process

(Tetlock, 1992).

Sponsoring managers informally search for candidates through their personal network, so

workers are often unaware they are being considered (Pinfield, 1995). As a result, accountability is

more limited than it is in posting, though it is not entirely absent. Managers are required to

communicate their decision to a supervisor but because supervisors typically grant managers

substantial discretion over who is selected, those decisions do not have to defended to a broader

audience. The higher level of accountability generated though the competitive posting process

should therefore guard against the use of irrelevant information, as managers are more likely to

use objective criteria in justifying their hiring decisions to a broader audience.

As a whole, these arguments suggest that infusing self-selection and formality into the

internal hiring process will help boundedly rational managers overcome problems associated with

generating and evaluating alternatives. Self-selection is likely to be more effective than active

managerial search in reducing the likelihood that an exceptional candidate will be overlooked,

while the formality of the market is likely to be more effective at disciplining managers to avoid

costly errors associated with bounded awareness. As a result, I expect posting to create more value

than sponsorship by generating higher quality internal hires, as revealed by worker performance,
 18

turnover and subsequent advancement. Specifically, I predict that when compared to sponsored

internal hires:

H1: Workers hired through posting will have higher performance ratings in the new job.

H2: Workers hired through posting will be less likely to exit the firm.

H3: Workers hired through posting will be more likely to subsequently be promoted.

Competition and Compensation

Quality of hire is only part of story, as the value a firm is able to capture from even the

highest quality hire is largely contingent on how much they are paid. It is therefore important to

understand how posting and sponsorship shape compensation. To do so, it is useful to switch from

the manager’s perspective and think about how salary negotiations from the perspective of the

worker.

Markets are characterized by open competition, with sellers aware that they are competing

for buyers with other sellers (Menard, 1995). Self-selection and formality are two of the key

mechanisms supporting the competitive nature of markets. Self-selection allows sellers to pursue

the opportunities they are interested in and the formality facilitates the exchange of information

that ultimately allows buyers to compare information on widely dispersed alternatives and make a

selection (Zenger et al., 2011).

Posting is characterized by open competition. Interested candidates self-select into the

consideration set when they apply for an open job. In doing so, they form a labor queue – a set of

workers competing for a specific job at a specific time (Reskin & Roos, 1990). The formality of

the posting process underscores the competitive nature of posting, as the fact that employees have

to actively submit an application makes them aware that they are entering into a competition they

are not assured of winning. In contrast, to the extent there is competition in sponsorship, it lacks

structure and transparency. Because search is costly, managers routinely consider a small pool of
 19

internal candidates (often just a single candidate). Moreover, because managers gather information

on potential candidates informally, in those cases where multiple candidates are considered, those

who are not selected are often unaware of being considered (Pinfield, 1995).

While competition is often seen as a way to lower prices by pitting multiple suppliers

against one another, recent work exploring the social psychological aspects of negotiation

provides reason to expect the opposite in the internal hiring context. A key premise in this

literature is that situational factors prime individuals to place more or less emphasis on the

importance of dyadic relationships in negotiations. The more emphasis an individual places on the

dyadic relationship in a negotiation context, the more likely they are to adopt a relational

orientation to negotiation, and individuals adopting a relational orientation to negotiation are more

likely to forgo economic gain in an effort to develop relational capital (Curhan, Elfenbein, & Xu,

2006; Curhan, Neale, Ross, & Rosencranz-Engelmann, 2008: 193; Gelfand et al., 2006).

Sponsorship is much more likely to cue a worker to focus on their relationship with the

hiring manager. Because the hiring manager personally appointed the worker to the job absent any

formal competition, the relationship with the hiring manager is likely to be particularly salient and

highly valued at the time an initial job offer is presented. This is unlikely to be case in posting,

with the competitive nature of the process emphasizing the transactional nature of the employment

relationship. As a result, workers hired through sponsorship are more likely to adopt a relational

orientation approach when the time comes to negotiate compensation.

There are two reasons to expect that adopting a relational orientation towards salary

negotiations has important implications for how much a worker is likely to earn. First, a worker

adopting a relational orientation is less likely to initiate a salary negotiation. Focused on

developing relational capital, she will want to avoid appearing self-interested and feel

uncomfortable with the idea of asking for more money (Gelfand et al., 2006). She is also more
 20

likely to perceive the initial offer as fair; being more attuned to other party’s goals should reduce

the likelihood she will assume the other party’s interests are opposed to their own, a common error

in negotiations (Cross, Bacon, & Morris, 2000; Gelfand et al., 2006; Thompson & Deharpport,

1998). Second, an individual adopting a relational orientation is more likely to employ an

accommodative negotiation strategy whereas an individual adopting a transactional orientation is

more likely to negotiate competitively (Curhan et al., 2008), and accommodative approaches to

salary negotiation have been shown to result in lower raises than competitive approaches (Marks

& Harold, 2011).

This suggests that there is an important tradeoff between the quality and costs associated

with different internal hiring processes. Though posting is likely to create more value, the firm

does not capture all of that value, as workers capture a portion of that value through higher

compensation. Relative to sponsorship, the competitive nature of the posting process will increase

the likelihood that workers both initiate salary negotiations and adopt a more economically

beneficial approach when they do choose to negotiate, leading me to predict that:

H4: Internal candidates hired through posting will receive higher starting salaries than

sponsored employees entering equivalent jobs.

DATA & METHODS

I test my hypotheses using monthly personnel records covering the years 2008 to 2012

from the US operations of Fortune 100 health insurance company, which I call HealthCo. The

primary data for my analyses consist of more than 11,000 internal hires made during the

observation period, which are identified from a larger dataset consisting of 1,914,519 monthly

observations covering 56,811 worker employed by HealthCo between 2008 and 2012.

While using data from a single firm limits the confidence with which I can generalize my

results, these data are particularly well suited to test my hypotheses. The distinguishing feature of
 21

these data is that the way in which HealthCo has linked their various human resource information

systems allows me to clearly identify whether posting or sponsorship was used to facilitate each

and every internal hire. I am also able to link these personnel records to a companion dataset with

information on more than 350,000 internal and external job applications submitted during 2012,

allowing me to conduct several robustness checks. Using personnel data from a single

organization has several other advantages, including the fact that my performance measures are

standardized across jobs and that I am able to control for the effects of job content and the location

of different jobs (and therefore different attributes of moves between jobs), all of which would

pose substantial empirical difficulties in a multi-firm study. Moreover, obtaining this type of data

from even a single firm is quite challenging; many firms fail to systematically record any data

related to employee mobility (ERC, 2010; Oracle, 2012) and for those that do, the fear of

sanctions were internal analyses to reveal previously unrecognized patterns of illegal

discrimination has the perverse effect of discouraging firms from exploring these processes

(Strum, 2001). The setting itself reduces at least some concerns about generalizability, as

HealthCo mirrors other large contemporary organizations in several respects: hiring decisions are

delegated to individual managers, employees are explicitly encouraged to actively manage their

careers amid a lack of well-defined advancement paths, and there are substantial amounts of

lateral and vertical mobility across broadly defined jobs.

Identifying Internal Hires

An internal hire occurs when a manager fills an open job with a current employee4, as

indicated by a change in an employee’s job code, department, or both from one month to the next.

An employee who changes job codes takes on a new set of tasks and responsibilities. A move to a

4
HealthCo has a vacancy-driven hiring process, meaning that all hires – internal and external – are
preceded by an open job.
 22

new department represents a move to different area of the business, as departments are organized

around products, geographic markets, and customers (internal and external). All entry-level jobs at

HealthCo are filled through external hiring; internal hiring is used alongside external hiring to fill

jobs above entry-level.

It is important to emphasize that a change in job code does not simply represent a change

in title, with little change to the work actually performed (Miner, 1987). Jobs at HealthCo are

broadly defined by hierarchical level, function, and role. In an average year, approximately 34,000

employees were distributed across 462 job codes. Jobs are organized into nine different levels [(1)

Entry-level, (2) Team Lead, (3) Supervisor/Analyst, (4) Manager/Professional, (5)

Director/Technical Leader, (6) Vice President, (7) Senior Vice President, (8) Executive Vice

President, (9) CEO] and thirty functional areas, including those common to most large firms (e.g.

Sales, Finance/Accounting, HR, and Marketing) as well as several more specific to the health

insurance industry (e.g. Clinical Guidance, Provider Contracting). Roles indicate the specific

competencies needed to perform the job. For example, “Creative Developer” is a Level 3 role in

Marketing; “Recruiting Lead” is a Level 3 role in Human Resources; and “Architect”,

“Applications Consultant” and “Project Manager” are all Level 3 roles with IT, each linked to

different competencies. As a result, a change in job code reflects a meaningful change in the work

a person does.

Dependent Variables

Quality of hire and compensation are the two primary outcomes of interest in this study.

Researchers have used a wide variety of post-hire outcomes to assess quality of hire (e.g, see

Breaugh & Starke, 2000; Zottoli & Wanous, 2000). In an effort to provide a holistic accounting of

the quality of internal hires with HealthCo, I test my hypotheses using multiple indicators of

quality: performance ratings, relative performance, turnover, and subsequent advancement.

 23

Performance ratings and relative performance. I use five measures of job performance

from HealthCo’s annual performance evaluation as a first set of quality indicators.

Contribution score. A worker’s contribution score assesses their contribution to the

success of the organization. In jobs with a less direct impact on organization-level outcomes (e.g.

those at lower levels), the contribution score is typically used to assess their contribution to the

department or line of business. It is measured on a 0 to 4 scale [0=not contributing (0%), 1=low

contribution (3.2%), 2=moderate contribution (19%), 3=full contribution (66.8%), and

4=exemplary contribution (11%)].

Competency score. A worker’s competency score assesses their skills relative to what is

required for the job. Each worker receives a separate score for each of the eight competencies

assigned to their job code5. Each competency is measured on a 1 to 4 scale [1=Learning,

2=Exhibiting, 3=Demonstrating, 4=Modeling]. I average these individual scores to compute an

overall competency score [1-2 (11.4%), 2-3 (66.2%), 3-4 (12.4%)].

Relative performance. Managers also rank workers in similar jobs as part of an annual

calibration process (described below). However, workers are not simply ordered according to

their contribution and competency scores; rather, this is intended to be a measure of overall value

to the organization that takes into account both previous performance and future potential.

Although there are no formally established guidelines dictating how finely managers should

distinguish among workers, these calibration sessions typically create “buckets” of employees; a

group of 100 employees may not be ranked from 1 (highest) to 100 (lowest), but rather the top five

employees may receive a 1, the next ten a 2, the next twenty-five a 3, and so on. I use these

ranking to create three dichotomous measures of relative performance: whether a worker is ranked

5
Selected from an overall library of 124 competencies customized for HealthCo.
 24

in top the quartile (top 25%), bottom half (bottom 50%), or bottom quartile (bottom 25%) of her

cohort in her first year in the new job.

While all of these are subjective measures of performance, several researchers have argued

that subjective ratings are among the most valid measures of performance despite concerns about

the potential for managerial bias to affect ratings (Cascio, 1998). Subjective ratings enable

managers to take into consideration a variety of behaviors and outputs relevant to the job (Medoff

& Abraham, 1981) as well as account for factors affecting performance outside the control of the

individual worker (Campbell, McCloy, Oppler, & Sager, 1993). However, it is also important to

note that HealthCo’s annual evaluation process helps to reduce potential concerns about

supervisory bias affecting individual performance ratings. Managers who supervise workers in

similar jobs meet in person to review and discuss their ratings of individual workers – a process

known as calibration. These calibration discussions are intended to ensure that managers are

evaluating workers against a common standard and to identify and correct instances where

managers may have rated employees too harshly or leniently. Research has shown that calibration

tends to reduce subjectivity and bias in performance ratings both because ratings are likely to be

more consistent across employees when managers share a common view on rating standards

(Mclntyre, Smith, & Hassett, 1984; Pulakos, 1984) and because “fellow managers do not do not

usually let each other off easily if they believe an employee has been rated unfairly, creating peer

pressure that provides a powerful incentive to make accurate ratings” (Pulakos & O’Leary, 2011:

152–3).

Turnover. Turnover is a second indicator of match quality (Jovanovic, 1979; Mortensen,

1988). I create two dichotomous measures of turnover indicating whether a worker exits the

organization with the first 12 months (turnover12) or 24 months (turnover24) of moving to a new

job. I do not distinguish between voluntary and involuntary turnover for two reasons.
 25

Conceptually, both are indicators of poor matches, with the worker and firm both more likely to

terminate an ill-fitting employment relationship. Empirically, while the data do indicate whether

turnover was voluntary or involuntary, managers and HR staff at HealthCo both told me not to

trust these indicators as reflecting the actual reasons an employee left the organization. Consistent

with research showing that turnover rates are relatively low among internal hires in non-entry-

level jobs (Bidwell, 2011), approximately 5% of internal hires exited the firm within 12 months,

while 13% exited within 24 months.

Promotion. Subsequent promotion is a third indicator of quality. Promotions are internal

hires that result in the employee moving into a higher level job. Because time to promotion varies

across jobs, I create two dichotomous measures of subsequent advancement: whether a worker

was promoted within 24 months (prom24) or 36 months (prom36) of moving into a new job.

Approximately 14% of internal hires were subsequently promoted within 24 months, while 31%

were promoted within 36 months.

Starting salary. The starting salary is the natural logarithm of the salary a worker receives

in the first month in a new job. Salary accounts for the vast majority of compensation for most

workers at HealthCo. Sales workers represent the main exception and their bonuses, which are tied

to clearly defined sales targets, can account for a substantial portion of their total income.

However, the bonus amount these workers are expected to receive based on their targets are

factored into the annual salary figures recorded in HealthCo’s personnel records. For example, if a

salesperson is hired into a job with a base salary of $80,000 and expects to earn a bonus of

$40,000, for a total expected annual compensation of $120,000, her salary in the monthly

personnel record from which I pull this figure will be $120,000. This means that the salary figure I

use represents the total compensation she should expect to earn during the year, which is the figure

upon which she will be negotiating.

 26

Independent Variable

The independent variable is a dichotomous indicator of whether posting (1) or sponsorship

(0) was used to fill the job, identified through an unambiguous indicator in the first monthly

observation of a worker in her new job. 3,841 (43%) of internal hires were made through posting

and 5,458 (57%) through sponsorship.

Control Variables

An important concern when using non-experimental data is the potential for omitted

variables to create spurious correlations between the independent and dependent variables. Hiring

managers at Health Co. are able to choose whether to fill a job through posting or sponsorship. A

particular concern in this study is that there may be variables that both affect a manager’s decision

of which process to use as well as the outcomes of interest. In the absence of existing empirical

evidence, it seems likely that the choice to use posting or sponsorship might be affected by two

factors: attributes of job and the manager’s perception of the availability of qualified candidates. It

could be that there are certain jobs that are always posted and others that are always filled through

sponsorship. It could also be the case that managers only post jobs if they have been unable to or

assume they will be unable to identify qualified candidates through their personal network. I

leverage the level of detail in my data to address the first issue, which enable me to control for

many of the job characteristics that might be expected to have an effect on the choice between

posting and sponsorship. To address the second issue, I run a series of robustness checks using an

instrumental variable approach, which I describe in more detail later in the paper.

Job attributes. In order to compare the outcomes associated with different processes used

to staff similar jobs, I control for several job-level attributes. I include dummy variables to control

for hierarchical rank, function area and the state a job is located in (51% of job are located in the

headquarters state). In order to control for fixed propensities of different jobs to be filled by
 27

posting or sponsorship, I also include separate dummies for each of 266 job codes filled though

internal hiring during my observation period. Importantly, the data reveal no systematic

differences in the types of jobs that are filled through posting or sponsorship. Of all the job codes

filled internally from 2008-2012, 84% were filled by both posting and sponsorship, and moves

into those jobs accounted for 99% of all internal hires. Moreover, 75% of the job codes filled

exclusively through either posting or sponsorship were only filled once or twice, 88% only three

times, and none more than six times. It is therefore likely that even those job codes filled

exclusively through one process during this time periods are nevertheless open to being filled

through both processes, with the apparent exclusivity an artifact of so few hires having been made

into those job codes.

Job transition attributes. Another concern is that the outcomes may be affected by the

types of moves workers are making rather than how those moves are made. For example, we

might expect workers moving to more similar jobs to have higher levels of performance and that

sponsored workers are likely to come from more similar jobs because managers are more likely to

have personal connections to workers in similar jobs. To account for this possibility, I include

dummies for whether an internal hire resulted in a worker moving into a new job in the same

function or same department. Similarly, I include dummies indicating whether an internal hire

resulted in a promotion (a vertical advancement; n=4,843), an expansion (a lateral advancement

which results in an expansion of the worker’s competencies; n=3,594), or a transfer (a lateral

move to the same job in a different department; n=1,122), which may affect pay and/or

performance. I include the worker’s salary in the final month of their previous job in some models

to account for the possibility that internal salary adjustments may be based on an employee’s

previous salary even when moving across very different jobs.

 28

Individual attributes. I also include controls for a variety of candidate attributes.

Demographic characteristics include gender, age and age-squared, and ethnicity. Tenure and

tenure-squared are calculated as the number of months (squared) a worker has been employed by

the firm. I include a worker’s contribution and competency scores in their previous job as a rough

indicator of pre-hire quality. HealthCo does not include years of education or highest degree

completed in their personnel records.

Additional controls. I include dummies for each year in the sample to reflect changing

labor market conditions. In models predicting relative performance, I control for the size of a

worker’s performance rating cohort (e.g. the number of workers in their forced ranking group).

Sample Restrictions

The samples used for each of my analyses vary according to a number of restrictions I

placed on the data. In the analyses where performance ratings are used as dependent variables, I

dropped observations with missing performance data, which occurred when workers had been

hired to recently to assess or exited the firm before being assessed (n=2,484). In both the turnover

and advancement models, the samples are restricted to those workers who were hired early enough

for me to calculate the dependent variables of interest (e.g. to be included in the model predicting

turnover in 12 months, a worker either had to have exited the firm within 12 months of moving

into their new job or occupied the job for 12 months). Finally, in the compensation models, I

excluded observations with missing salary data (n=28).

ANALYSIS & RESULTS

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

independent variables, with each observation representing an internal hire. Of particular interest

are the correlations between the multiple measures of quality of hire. The correlation between the

contribution and competency scores (r = .67) which indicate that these two measures pick up
 29

different aspects of performance. The correlations between these two measures and the relative

ranking variables covary in the expected directions while also suggesting that they are also picking

up different aspects of performance, as they are positively correlated with a worker ending up in

the Top 25% (r = .43, .45) and negatively correlated with a worker ending up in the Bottom 50%

(r = -.46, -.50) and Bottom 25% (r = -.40, -.46). The low correlations among the performance

rating variables and the turnover and subsequent promotion variables (none exceed r = +/- .13)

suggest that each of these variables are capturing a different element of performance. It is also

important to note the low correlations between salary and each of the performance measures (none

exceed r = +/- .16), which underscore the notion that salary does not simply reflect a manager’s

expectations about performance in the new job, but is instead largely determined by a negotiation

process which occurs prior to a manager observing actual performance (Jovanovic, 1979). This is

important in interpreting the results, as I argue that while posting results in both higher quality

hires and higher salaries, the higher salaries are not a result of the manager expecting a higher

level of performance. In fact, the robustness checks suggest that, if anything, managers are likely

to expect sponsored hires to perform better.

------------------------------------
Insert Table 1 about here
------------------------------------
Performance

Table 2 presents analyses of each of the five performance measures. I use ordinary least

squares regressions in models using competency and contribution scores as the dependent

variables. Competency score is a continuous variable and though contribution score is a discrete,

ordinal measure, the OLS model is easier to interpret than an ordered logit model (and both

provide similar results). I use a logit specification for models where relative performance is

outcome of interest, as the dependent variables are binary. The unit of analysis in all models is an
 30

internal hire and the performance measures reflect performance in the first year on the job. I

cluster the errors by individuals to account for non-independence among the errors.

Using contribution and competency scores as the dependent variables, Models 1 and 2

provide support for Hypothesis 1, with the significant positive coefficient for posting in both

models indicate that internal hires made through posting outperform internal hires made through

sponsorship. The effect sizes are relatively small, with posting resulting in an increase of

approximately one-tenth of a standard deviation on each measure of performance, an issue I return

to in the robustness checks. Model 3 provides additional support for Hypothesis 1 in

demonstrating that internal hires made through posting are approximately 13% more likely to be

rated in the top quartile of their respective performance/potential distribution than sponsored

internal hires. Models 4 and 5 further reveal that internal hires made through posting are less likely

to have poor performance ratings in the new job. The significant negative coefficients for posting

indicate that internal hires made through posting are approximately 13% less likely to fall in the

bottom half of the ratings distribution and approximately 15% less likely to fall in the bottom

quartile. Taken together, these results strongly suggest that posting leads to better internal hiring

decisions, in part by helping managers avoiding hiring mistakes.

------------------------------------
Insert Table 2 about here
------------------------------------
Turnover

Models 7 and 8 in Table 3 use logit models to test Hypothesis 2, that internal hires made

through posting are less likely to exit the firm. Both models provide support for this hypothesis,

with the significant negative coefficients for posting indicating that internal hires made through

posting are around 20% less likely to exit the firm within 12 months (Model 7) and around 18%

less likely to exit the firm within 24 months (Model 8).

 31

------------------------------------
Insert Table 3 about here
------------------------------------
Subsequent Advancement

Models 9 and 10 in Table 3 also use logit models test Hypothesis 3, that internal hires

made through posting are more likely to be subsequently promoted. I find very limited support for

this argument. The coefficient for posting in not significantly different from zero in Model 9,

indicating that there is no difference the 24 month promotion rates of internal hires made through

posting or sponsorship. The coefficient for posting is positive but only marginally significant in

Model 10, provided some indication that posting may be more likely lead promotion over a 36

month period, but nothing conclusive.

Salary

Models 5 and 6 in Table 3 test Hypotheses 4, that internal hires made through posting will

receive higher starting salaries than sponsored hires entering equivalent jobs. The significant

positive coefficient for posting in Model 11 provides support for this hypothesis, revealing that

posted hires receives nearly 4% higher salaries, on average, than sponsored hires. The 4% posting

premium remains after controlling for performance in the previous job (Model 12).

Mobility Patterns and Within Department Hires

In theorizing about the effects of self-selection and formality on quality of hire, I made two

complementary arguments, one that suggests that self-selection improves quality by expanding the

pool of potential candidates and the other that formality helps to shape the information managers

use when evaluating candidates. The regression results above, however, do not allow me to

whether one or both of these mechanisms are driving the results. To examine this issue in more

detail, I first present descriptive statistics on internal source of hire (Table 4). These data reveal

that posting is significantly more likely to result in hires from different departments, functions,
 32

cities and even different buildings, consistent with my argument that self-selection lessens the

likelihood that manager will overlook an exception candidate by allowing managers to more

readily identify candidates widely dispersed throughout the organization. I then ran a series of

regressions restricting the analysis to internal hires made within departments (Table 5). Because

managers are likely to be aware of and have access to much more detailed information on

candidates located within their own department, this provides a stronger test of my argument that,

beyond providing more alternatives, posting improves decision making by disciplining what

information they use in evaluating candidates. Consistent with my theorizing, the results are nearly

identical to those presented in Tables 2 and 3. In fact, if we assume that managers not only have

ready access to, but already possess detailed information (both relevant and irrelevant to their

ability to do the job) on candidates within their department, we might conclude that markets are

particularly beneficial in helping managers to avoid the use of irrelevant information when

evaluating well-known alternatives, a commonly cited reason for hiring errors (Highhouse, 2008).

------------------------------------
Insert Tables 4 & 5 about here
------------------------------------
Supplementary Analyses and Robustness Checks

As noted above, a potential concern with my use of ordinary least squares and logit models

above is the potential endogeneity of posting. While I am able to control for the fixed propensities

for certain jobs to be filled by posting or sponsorship, I do not have a variable which would allow

me to control for the possibility that managers only post jobs if they are unable to first identify

qualified candidates through their personal network. However, it is important to note that if this

were indeed the case, we would expect the endogenous nature of this choice to affect the quality of

hire results in the opposite direction; if managers are most likely to use sponsorship when they are

able to personally identify an excellent candidate, sponsored hires should be expected to

 33

outperform hires made through posting. Empirically addressing this potential endogeneity

nevertheless seems prudent and I do so using an instrumental variable approach.

An instrumental variable should correlate strongly with endogenous variable (posting) but

not with the second stage error term (Hamilton & Nickerson, 2003; Semadeni, Withers, & Certo,

2014). A variable that appears to meet these two conditions is the percentage of similar jobs filled

by posting in other departments in the preceding two months. My discussions with both hiring

managers and HR staff at HealthCo confirmed that managers regularly, (a) ask other managers

how they recently filled similar jobs, or (b) contact HR and ask how similar positions have been

filled recently. While both mechanisms should be expected to influence the manager’s decision on

how to fill the job, how similar jobs are filled in other parts of the organization should have little

direct effect on post-hire outcomes or salary.

Table 6 compares the results for competency and contribution scores using OLS (Models 1

and 2) to those using both a traditional 2SLS approach (Models 3 and 4) and a 2SLS treatment

approach which accounts for the binary nature of my endogenous variable (Models 5 and 6). The

first stage estimates reveal that I do not have to worry about instrument weakness. As can be seen

by comparing the results of the different equations, the results are consistent with those obtained

using OLS. In fact, after I controlled for the endogeneity bias, the effect of posting of both

competency and contribution scores are substantially stronger, increasing from an increase of

around one-tenth of a standard deviation to nearly four-tenths of a standard deviation for

contribution and over half a standard deviation for competency.

------------------------------------
Insert Table 6 about here
------------------------------------
Table 7 presents the results for relative performance using instrumental variable probit

models. Aside from the change in coefficient magnitude, the main difference between these results
 34

and those from the logit models is that there is no significant relationship between how a worker

enters a job and their likelihood of landing in the top quartile of the performance distribution.

Taken together, the interpretation of these results remains unchanged – posting disciplines

managers to make better internal hiring decisions, in part by avoiding hiring mistakes.

------------------------------------
Insert Table 7 about here
------------------------------------
Table 8 presents the results for turnover and subsequent advancement using instrumental

variable probit models, as well as the result for salary using a traditional 2SLS approach. Again,

the results lead to similar interpretations as the non-instrumental variable models, with the main

difference being the magnitude of the coefficients. The one substantive difference is that the IV

Probit specification indicates that while internal hires are less likely to exit the firms with 12

months, they are not more or less likely to exit the firm within the longer 24 month time period.

------------------------------------
Insert Table 8 about here
------------------------------------
Wired Searches

While posting and sponsorship are conceptually distinct, it is possible for a manager to

post a job having already decided who they are going to select. Though these jobs are posted, they

are actually filled through sponsorship. If this was a common occurrence and/or these instances

were difficult to identify empirically, it would be difficult to interpret my results. These so-called

“wired searches” are likely to be most common in firms that require managers to post all jobs.

Because Health Co. has established formal systems for filling jobs through both posting and

sponsorship, managers are able to avoid posting jobs when they already have a candidate in mind

and thereby avoid the potential negative consequences that emerge from other employees having

felt they participated in an unfair selection process (Billsberry, 2007). I nevertheless conducted
 35

two additional robustness checks to rule of the possibility that my results are affected by the

inclusion of wired searches. First, I reviewed each of the 1,695 internal hires made between 2011

and 2012 for which I have detailed application data. Those instances where there was only a single

internal applicant (and no external applicants) could potentially be wired searches, with managers

either tailoring the job description around a particular candidate and/or discouraging others

employees from applying. Less than 5% of internal hires meet these criteria and of those, two-

thirds were open for more than a month, suggesting that the hiring manager was hoping to find

additional candidates. A total of 30 posted internal hires (1.4%) are potentially wired searches,

leading me to believe that such searches are not driving my results. Second, I identified all of the

postings during this period which were open a week or less (n = 45), another potential indication

that the hiring manager created the posting for a specific candidate and therefore wanted to limit

the number of other candidates who would apply. In both cases, dropping these observations from

the analyses does not substantively change the results.

External Candidates as an Alternative Explanation

I ran an additional series of robustness checks to ensure that the results for posting were

not driven by external market forces. When posting a job, a manager has the option of restricting

the competition to internal candidates by only posting the job internally, or opening the

competition to both internal and external candidates by also posting the job externally. An

alternative explanation for the performance findings might be that the inclusion of external

candidates allows managers to benchmark internal candidates against the market, leading to better

hires when an internal candidate is selected (Billsberry, 2007). Perhaps more importantly, an

alternative explanation for the higher starting salaries associated with posting might be that mere

exposure to the external market is responsible for driving up the starting salaries for posted jobs,
 36

with the results driven by those instances in which external candidates were considered but an

internal candidate was selected.

In order to rule out these alternative explanations, I use data from 2011 and 2012 that

allows me to identify the number of internal and external candidates that applied to every posted

job and how far they made it through the hiring process. This allows me to identify which job

postings were restricted to internal candidates and which were open to external candidates. Table 9

reveals the result of analyses including a dummy variable equal to one if external candidates were

considered for the position. Because this data begin in mid-2011, I am unable to calculate results

for turnover and subsequent advancement. Of the 869 internal hires for whom I have both

performance ratings and salary data, 578 (66%) were hired without considering an external

candidate; 291 (34%) internal hires competed against at least one external candidate. The results

indicate that opening the job to external candidates has no effect on performance (Models 1-5) or

starting salary (Models 6-7).

------------------------------------
Insert Table 9 about here
------------------------------------

DISCUSSION

This research advances our understanding of how human resources are allocated within

firms by shedding light on the processes used to facilitate internal hiring within contemporary

organizations. Despite the prevalence of internal hiring and its impact on the fortunes of firms and

workers, our current models of internal hiring are still largely based on literatures exploring

advancement with highly bureaucratic internal labor markets that bear little resemblance to their

more contemporary counterparts. This study provides a much-needed update to these earlier

models, identifying and describing the two most commonly used internal hiring processes –
 37

market-oriented posting and relationship-oriented sponsorship. I develop theory predicting the

relative effects of posting and sponsorship on quality of hire and compensation by highlighting

that posting is characterized by two market-like features, self-selection and formality, that are

largely absent in sponsorship, which instead involves active managerial search and a reliance on

personnel connections for gathering information.

I argued that introducing these two market features would improve decision-making by

helping managers to overcome challenges associated with generating and evaluating alternatives.

Self-selection should generate a larger pool of alternatives, reducing the likelihood and

exceptional alternative will be overlooked, while the formality of the market should encourage

managers to both incorporate relevant information and avoid irrelevant information as they

evaluate candidate qualifications against the requirements of the jobs. Consistent with these

arguments, I found that employees hired through posting outperform their counterparts hired

through sponsorship and are less likely to exit the firm, though there is no discernable difference

in their likelihood of subsequent advancement.

However, this improved decision-making comes at a price. I argued that employees hired

through the more competitive posting would be less likely to adopt a relational orientation to

salary negotiations therefore more likely to both initiate salary negotiations and adopt a more

effective, competitive approach when doing so. Though I am unable to test this mechanism

directly (I do not observe the difference between initial and final salary offers), the results are

consistent with this account, as employees hired through posting are paid ~4% more than

sponsored employees hired into equivalent jobs.

Despite this posting premium, these results suggest that posting enables firms to both

create and capture substantially more value than sponsorship. Unfortunately, my data do not allow

me to directly measure the value created by higher levels of individual performance or the savings
 38

associated with higher retention rates. However, with the costs of replacing an employee alone

running anywhere between 20% and 200% of an employee’s annual salary (Boushey & Glynn,

2012), the combined performance and retention benefits associated with better internal hires are

likely to far exceed the higher salary costs. Moreover, the posting premium suggests that workers

also share in the increased value creation, earning higher salaries for equivalent work. The market-

oriented posting process therefore appears to benefit both firms and workers.

As with all single-firm studies, some caution is necessary in generalizing these results.

Examining mobility within firms requires tradeoffs between depth and generalizability, requiring

detailed internal data that can be difficult to obtain from multiple sites. This is made even more

difficult by the fact that few firms actually capture data on the processes by which employees

move to new jobs. My conversations with HealthCo officials and with human resource leaders at

multiple additional organizations during the course of this study, however, have not revealed any

reason to believe that HealthCo’s internal hiring practices are different from other large

organizations.

In addition to not being able to directly measure value creation, this study does not account

for potential costs associated with creating an internal market for talent beyond the salary

premium. Additional costs might include the investments made to implement and maintain an

internal job board and the time involved in screening and evaluating job applications. Along these

lines, a particularly promising avenue for future research would be to explore whether posting

imposes additional costs by creating a visible set of employees who lost out in an open

competition, such as lower performance or increased turnover.

Despite these limitations, this work makes several important contributions. In highlighting

the ways in which self-selection and formality shape managerial decision-making and

negotiations, this study adds to the growing insights about the benefits and limitations of bringing
 39

market mechanisms within firms (Zenger et al., 2011). To date this work has focused

predominately on the ways in which the infusion of high powered incentives shapes individual

behavior (Ellig, 2001; Zenger & Hesterly, 1997; Zenger, 1992). More recently, scholars have

begun to explore the ways in which firms are able to harness the information aggregation powers

of the market to improve internal decision-making though the use of prediction and information

markets, with the idea being that the widely diffused information can be aggregated into

something akin to a price (Ellig, 2001; Felin & Zenger, 2011). This study complements and

extends this work by showing how two less explored features of markets – self-selection and

formality – improve decision-making in the absence of a price mechanism. Moreover, I show that,

contrary to what extant theory predicts about the role of competition in external markets, infusing

competition within the firm actually increases prices. Finally, in identifying the micro-level

mechanisms through which these macro-level allocative processes shape individual behaviors, this

work speaks to a growing body of literature interested in micro-foundations of strategic

organization (Felin & Foss, 2005; Ployhart & Hale, 2014) as well as a more specific literature on

the micro foundations of human capital-based competitive advantage (Coff & Kryscynski, 2011).
 40

REFERENCES

Afuah, A., & Tucci, C. L. 2012. Crowdsourcing as a solution to distant search. Academy of
Management Review, 37(3): 355–375.

Agerström, J., & Rooth, D.-O. 2011. The role of automatic obesity stereotypes in real hiring
discrimination. Journal of Applied Psychologysychology, 96(4): 790–805.

Agrawal, A., Knoeber, C. R., & Tsoulouhas, T. 2006. Are outsiders handicapped in CEO
successions? Journal of Corporate Finance, 12(3): 619–644.

Alexander, E. R. 1979. The design of alternatives in organizational contexts: A pilot study.

Administrative Science Quarterly, 24(3): 382–404.

Allen, D. G., Bryant, P. C., & Vardaman, J. M. 2010. Retaining talent: Replacing misconceptions
with evidence-based strategies. Academy of Management Perspectives, 24(2): 48–65.

Althauser, R. P., & Kalleberg, A. L. 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.

Argote, L., & Ingram, P. 2000. Knowledge transfer: A basis for competitive advantage in firms.
Organizational Behavior and Human Decision Processes, 82(1): 150–169.

Barney G. Glaser (Ed.). 1968. Organizational careers: A sourcebook for theory. Chicago: Aldine
Publishing.

Bazerman, M. H., & Chugh, D. 2005. Bounded awareness: Focusing failures in negotiation. In L.
Thompson (Ed.), Fronteirs of social psychology: Negotiation: 7–26. New York:
Psychological Press.

Bertrand, M., & Mullainathan, S. 2004. Are Emily and Greg More Employable Than Lakisha and
Jamal? A Field Experiment on Labor Market Discrimination. American Economic Review,
94(4): 991–1013.

Bidwell, M. J. 2011. Paying more to get less: The effects of external hiring versus internal
mobility. Administrative Science Quarterly, 56(3): 369–407.

Bidwell, M. J., & Keller, J. 2014. Within or without? How firms combine internal and external
labor markets to fill jobs. Academy of Management Journal, 57(4): 1035–1055.

Bielby, W. T. 2000. Minimizing workplace gender and racial bias. Comtemporary Sociology,
29(1): 120–129.

Bies, R. J., Tripp, T. M., & Neale, M. A. 1993. Procedural fairness and profit seeking: The
perceived legitimacy of market exploitation. Journal of Behavioral Decision Making,
6(June): 243–256.
 41

Billsberry, J. 2007. Experiencing recruitment and selection. West Sussex, England: John Wiley
& Sons.

Borgatti, S. P., & Cross, R. 2003. A relational view of information seeking and learning in social
networks. Management Science, 49(4): 432–445.

Bradach, J. L., & Eccles, R. G. 1989. Price, authority, and trust: From ideal types to plural forms.
Annual Review of Sociology, 15: 97–118.

Breaugh, J. A., & Starke, M. 2000. Research on employee recruitment: So many studies, so many
remaining questions. Journal of Management, 26(3): 405–434.

Brenner, L. A., Koehler, D. J., & Tversky, A. 1996. On the evaluation of one-sided evidence.
Journal of Behavioral Decision Making, 9: 59–70.

Burks, S., Cowgill, B., Hoffman, M., & Housman, M. 2013. The value of hiring through
referrals.

Campbell, J. P., McCloy, R. A., Oppler, S. H., & Sager, C. E. 1993. A theory of performance. In
N. Schmitt & W. C. Borman (Eds.), Personnel selection in organizations: 35–70. San
Fransisco, CA: Jossey-Bass.

Campion, M. A., Cheraskin, L., & Stevens, M. J. 1994. Career-related antecedents and outcomes
of job rotation. Academy of Management Journal, 37(6): 1518–1542.

Cappelli, P. 2008. Talent on demand: Managing talent in an age of uncertainty: 304.

Cambridge, MA: Harvard Business School Publishing.

Cappelli, P., & Keller, J. 2014. Talent management: Conceptual approaches and practical
challenges. Annual Review of Organizational Psychology and Organizational Behavior, 1:
305–331.

Capron, L., Dussauge, P., & Mitchell, W. 1998. Resource redeployment following horizontal
acquisitions in Europe and North America, 1988-1992. Strategic Management Journal,
19(7): 631–661.

Casciaro, T., & Lobo, M. S. 2008. When competence is irrelevant: The role of interpersonal affect
in task-related ties. Administrative Science Quarterly, 53(4): 655–684.

Cascio, W. F. 1998. Applied psychology in human resource management (5th ed.). Upper Saddle
River, NJ: Prentice Hall.

Cascio, W. F., & Aguinis, H. 2008. Staffing twenty-first century organizations. Academy of
Management Annals, 2(1): 133–165.

Casson, M. 1982. The entrepreneur: An economic theory. Oxford, UK: Martin Robertson.
 42

Coff, R. W., & Kryscynski, D. 2011. Drilling for micro-foundations of human capital-based
competitive advantages. Journal of Management, 37(5): 1429–1443.

Cowen, T., & Parker, D. 1997. Markets in the firm: A market-process approach to management.
London: The Institute of Economic Affairs.

Crispin, G., & Mehler, M. 2013. Sources of hire 2013: Perception is reality.

Cross, R., Parker, A., Prusak, L., & Borgatti, S. P. 2001. Knowing what we know: Supporting
knowledge creation and sharing in social networks. Organization Dynamics, 30(2): 100–120.

Cross, S. E., Bacon, P. L., & Morris, M. L. 2000. The relational-interdependent self-construal and
relationships. Journal of Personality and Social Psychology, 78(4): 791–808.

Curhan, J. R., Elfenbein, H. A., & Xu, H. 2006. What do people value when they negotiate?
Mapping the domain of subjective value in negotiation. Journal of Personality and Social
Psychology, 91(3): 493–512.

Curhan, J. R., Neale, M. A., Ross, L., & Rosencranz-Engelmann, J. 2008. Relational
accommodation in negotiation: Effects of egalitarianism and gender on economic efficiency
and relational capital. Organizational Behavior and Human Decision Processes, 107(2):
192–205.

Cyert, R. M., & March, J. G. 1963. A behavioral theory of the firm. Englewood Cliffs, N.J.:
Prentice-Hall.

Dencker, J. C. 2008. Corporate restructuring and sex differences in managerial promotion.

American Sociological Review, 73(3): 455–476.

Diprete, T. A. 1987. Horizontal and vertical mobility in organizations. Administrative Science

Quarterly, 32(3): 422–444.

Doeringer, P. B., & Piore, M. J. 1971. Internal labor markets and manpower analysis.
Lexington, MA: D.C. Heath and Company.

Dougherty, T. W., Turban, D. B., & Callender, J. C. 1994. Confirming first impressions in the
employment interview: A field study of interviewer Behavior. Journal of Applied
Psychology, 79(5): 659–665.

Drazin, R., & Rao, H. 2002. Harnessing managerial knowledge to implement product-line
extensions: how to mutual find families allocate portfolio managers to old and new funds?
Academy of Management Journal, 45(3): 609–619.

Ellig, J. 2001. Internal markets and the theory of the firm. Managerial and Decision Economics,
22(45): 227–237.

ERC. 2010. ERC staffing practices survey. Mayfield Village, OH.

 43

Felin, T., & Foss, N. J. 2005. Strategic organization: a field in search of micro-foundations.
Strategic Organization, 3(4): 441–455.

Felin, T., & Zenger, T. R. 2011. Information aggregation, matching and radical market-hierarchy
hybrids: Implications for the theory of the firm. Strategic Organization, 9(2): 163–173.

Fernandez, R. M., Castilla, E. J., & Moore, P. 2000. Social capital at work: Networks and
employment at a phone center. American Journal of Sociology, 105(5): 1288–1356.

Gelfand, M. J., Major, V. S., Raver, J. L., Nishii, L. H., Brien, O., Michele, J., et al. 2006.
Negotiating relationally: The dynamics of the relational self in negotiations. Academy of
Management Review, 31(2): 427–451.

Gemünden, H. G., & Hauschildt, J. 1985. Number of alternatives and efficiency in different types
of top-management decisions. European Journal of Operational Research, 22(2): 178–190.

Granovetter, M. S. 1985. Economic action and social structure: The problem of embeddedness.
American Journal of Sociology, 91(3): 481–510.

Groysberg, B., Lee, L.-E., & Nanda, A. 2008. Can they take it with them? The portability of star
knowledge workers’ performance. Management Science, 54(7): 1213–1230.

Hamilton, B. H., & Nickerson, J. A. 2003. Correcting for endogeneity in strategic management
research. Strategic Organization, 1(1): 51–78.

Hauser, J. R., & Wernerfelt, B. 1990. An evaluation cost model of consideration sets. Journal of
Consumer Research, 16(4): 393–408.

Highhouse, S. 2008. Stubborn reliance on intuition and subjectivity in employee selection.

Industrial and Organizational Psychology, 1(3): 333–342.

Ibarra, H. 1993. Personal networks of women and minorities in management: A conceptual

framework. Academy of Management Review, 18(1): 56–87.

Iyengar, S. S., & Kamenica, E. 2010. Choice proliferation, simplicity seeking, and asset allocation.
Journal of Public Economics, 94(7-8): 530–539.

Jackson, C. K. 2013. Match quality, worker productivity, and worker mobility: Direct evidence
from teachers. The Review of Economics and Statistics, 95(4): 1096–1116.

Jacoby, S. M. 1985. Employing bureaucracy: Managers, unions, and the transformation of

work in American industry. New York: Columbia University Press.

Jacoby, S. M. 2004. Employing bureaucracy: Managers, unions, and the transformation of

work in the 20th century. Mahwah, NJ: Lawrence Erlbaum.
 44

Jacoby, S. M. 2005. The embedded corporation: Corporate governance and employment

relations in Japan and the United States. Princeton, NJ: Princeton University Press.

Jovanovic, B. 1979. Job matching and the theory of turnover. Journal of Political Economy,
87(5): 972–990.

Kahneman, D. 2011. Thinking, fast and slow. New York: Farrar, Straus and Giroux.

Kanter, R. M. 1977. Men and women of the corporation. New York: Basic Books.

Knudsen, T., & Levinthal, D. A. 2007. Two faces of search: Alternative generation and alternative
evaluation. Organization Science, 18(1): 39–54.

Loasby, B. J. 2000. Market institutions and economic evolution. Journal of Evolutionary

Economics, 10(3): 297–309.

London, M. 1996. Redeployment and continuous learning in the 21st Century: Hard lessons and
positive examples from the downsizing era. Academy of Management Executive, 10(4): 67–
79.

March, J. G., & Simon, H. A. 1958. Organizations. New York: Wiley.

Marks, M., & Harold, C. 2011. Who asks and who receives in salary negotiation. Journal of
Organizational Behavior, 394: 371–394.

Marlowe, C. M., Schneider, S. L., & Nelson, C. E. 1996. Gender and attractiveness biases in
hiring decisions: Are more experienced managers less biased ? Journal of Applied
Psychology, 81(1): 11–21.

Marsden, P. V., & Gorman, E. H. 1999. Social capital in internal staffing practices. In R. T. A. J.
Leenders & S. M. Gabbay (Eds.), Corporate social capital and liability: 180–196. Boston,
MA: Kluwer Academic Publishers.

McEvily, B., Soda, G., & Tortoriello, M. 2014. More formally: Rediscovering the missing link
between formal organization and informal social structure. Academy of Management
Annals, 8(1): 299–345.

Mclntyre, R. M., Smith, D. E., & Hassett, C. E. 1984. Accuracy of performance ratings as affected
by rater training and perceived purpose of rating. Journal of Applied Psychology, 69(I): 147–
156.

Medoff, J. L., & Abraham, K. G. 1981. Are those paid more really more productive? The case of
experience. Journal of Human Resources, 16(2): 186–216.

Menard, C. 1995. Markets as institutions versus organizations as markets? Disentangling some

fundamental concepts. Journal of Economic Behavior and Organization, 28: 161–182.
 45

Miller, R. A. 1984. Job matching and occupational choice. Journal of Political Economy, 92(6):
1086–1120.

Miner, A. S. 1987. Idiosyncratic jobs in formalized organizations. Administrative Science

Quarterly, 32(3): 327–351.

Mintzberg, H. 1994. The rise and fall of strategic planning. New York: Free Press.

Mintzberg, H., Raisinghani, D., & Théorêt, A. 1976. The structure of “unstructured” decision
processes. Administrative Science Quarterly, 21(2): 246–275.

Mortensen, D. T. 1988. Wages, separations, and job tenure: On-the-job specific training or
matching? Journal of Labor Economics, 6(4): 445–471.

Nahapiet, J., & Ghoshal, S. 1998. Social capital, intellectual capital, and the organizational
advantage. Academy of Management Review, 23(2): 242–266.

Nickerson, J. A., & Zenger, T. R. 2004. A knowledge-based theory of the firm: The problem-
solving perspective. Organization Science, 15(6): 617–632.

Nutt, P. C. 1998. How decision makers evaluate alternatives and the influence of complexity.
Management Science, 44(8): 1148–1166.

Ocasio, W. 1997. Towards an attention-based view of the firm. Strategic, 18(Summer Special
Issue): 187–206.

Oracle. 2012. U.K. talent mobility.

Pinfield, L. T. 1995. The operation of internal labor markets: Staffing practices and vacancy
chains. New York: Plenum Press.

Piore, M. J. 2002. Thirty years later: Internal labor markets, flexibility and the new economy.
Journal of Management and Governance, 6(4): 271–279.

Ployhart, R. E., & Hale, D. 2014. The fascinating psychological microfoundations of strategy and
competitive advantage. Annual Review of Organizational Psychology and Organizational
Behavior, 1(1): 145–172.

Podolny, J. M., & Baron, J. N. 1997. Resources and Relationships: Social Networks and Mobility
in the Workplace. American Sociological Review, 62(5): 673.

Polyani, K. 1957. The great transformation. Boston: Beacon Press.

Powell, W. W., & Snellman, K. 2004. The knowledge economy. Annual Review of Sociology,
30(1): 199–220.
 46

Pulakos, E. D. 1984. A comparison of rater training programs: Error training and accuracy
training. Journal of Applied Psychology, 69(4): 581–588.

Pulakos, E. D., & O’Leary, R. S. 2011. Why is performance management broken? Industrial and
Organizational Psychology, 4: 146–164.

Reskin, B. F., & Roos, P. A. 1990. Job queues, gender queues: Explaining women’s inroads into
male occupations. Temple University Press.

Rogan, M., & Sorenson, O. 2014. Picking a (poor) partner: A relational perspective on
acquisitions. Administrative Science Quarterly, 59(2): 301–329.

Rosenbaum, J. E. 1979. Organizational career mobility: Promotion chances in a corporation

during periods of growth and contraction. American Journal of Sociology, 85(1): 21–48.

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.

Seidel, M.-D. L., Polzer, J. T., & Stewart, K. J. 2000. Friends in high places: The effects of social
networks on discrimination in salary negotiations. Administrative Science Quarterly, 45(1):
1–24.

Semadeni, M., Withers, M. C., & Certo, S. T. 2014. The perils of endogeneity and instrumental
variables in strategy research: Understanding through simulations. Strategic Management
Journal.

Shocker, A. D., Ben-Akiva, M., Boccara, B., & Nedungadi, P. 1991. Consideration set influences
on consumer decision-making and choice: Issues, models, and suggestions. Marketing
Letters, 2(3): 181–197.

Simon, H. A. 1955. A behavioral model of rational choice. Quarterly Journal of Economics, 69:
99–118.

Sirmon, D. G., Gove, S., & Hitt, M. A. 2008. Resource management in dyadic competitive rivalry:
The effects of resorce bundling and deployment. Academy of Management Journal, 51(5):
919–935.

Sirmon, D. G., & Hitt, M. A. 2009. Contingincies within dynamic managerial capabilities:
Interdependent effects of resource investment and depolyment on firm performance.
Strategic Management Journal, 1394: 1375–1394.

Slichter, S. H., Healy, J. J., & Livernash, E. R. 1960. The impact of collective bargaining of
management. Washington, D.C.: The Brookings Institution.

Sorenson, O., & Waguespack, D. M. 2006. Social structure and exchange: Self-confirming
dynamics in Hollywood. Administrative Science Quarterly, 51: 560–589.
 47

Stein, J. C. 1997. Internal capital markets and the competition for corporate resources. Journal of
Finance, 52(1): 111–133.

Strum, S. 2001. Second generation employement discrimination: A structural approach. Columbia

Law Review, 101: 458–568.

Taleo Research. 2005. Internal mobility. Redwood Shores, CA.

Tetlock, P. E. 1992. The impact of accountability on judgement and choice. In M. P. Zanna (Ed.),
Advance in experimental social psychology (25th ed.): 331–376. San Diego, CA: Academic
Press.

Thompson, L., & Deharpport, T. 1998. Relationships, goal incompatibility, and communal
orientation in negotiations. Basic and Applied Social Psychology, 20(1): 33–44.

Tsai, W. 2002. Social structure of “coopetition” within a multiunit organization: Coordination,

competition, and intraorganizational knowledge sharing. Organization Science, 12(2): 179–
190.

Tsai, W., & Ghoshal, S. 1998. Social capital and value creation: The role of intrafirm networks.
Academy of Management Journal, 41(4): 464–476.

Turner, R. H. 1960. Sponsored and contest mobility and the school system. American
Sociological Review, 25(6): 855–867.

Williams, P. F. 1987. The legitimate concern with fairness. Accounting, Organizations and
Society, 12(2): 169–189.

Zenger, T. R. 1992. Why do employers only reward extreme performance? Examining the
relationships among performance, pay, and turnover. Administrative Science Quarterly,
37(2): 198–219.

Zenger, T. R. 2002. Crafting internal hybrids: Complementarities, common change initiatives, and
the team-based organization. International Journal of the Economics of Business, 9(1): 79–
95.

Zenger, T. R., Felin, T., & Bigelow, L. 2011. Theories of the firm-market boundary. Academy of
Management Annals, 5(1): 89–133.

Zenger, T. R., & Hesterly, W. S. 1997. The disaggregation of corporations: Selective intervention,
high-powered incentives, and molecular units. Organization Science, 8(3): 209–222.

Zottoli, M. A., & Wanous, J. P. 2000. Recruitment source research: Current status and future
directions. Human Resource Management Review, 10(4): 353–382.
 48

TABLE 1
Descriptive Statistics & Correlationsa

Variable n Mean S.D. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

1 posting 11803 0.43 0.5
2 instrumental var. 11331 0.44 0.27 0.30
3 contribution score 9520 2.85 0.64 ‐0.05 ‐0.04
4 competency score 9590 2.55 0.53 ‐0.05 ‐0.03 0.67
5 top 25% 10115 0.29 0.45 ‐0.01 0.02 0.43 0.45
6 bottom 50% 10115 0.49 0.5 0.02 ‐0.02 ‐0.46 ‐0.50 ‐0.60
7 bottom 25% 10115 0.25 0.43 0.03 ‐0.01 ‐0.40 ‐0.46 ‐0.35 0.59
8 turnover (12) 8359 0.03 0.16 ‐0.01 0.01 ‐0.05 ‐0.03 ‐0.03 0.03 0.03
9 turnover (24) 5508 0.12 0.32 ‐0.03 ‐0.02 ‐0.12 ‐0.11 ‐0.09 0.08 0.07 0.56
10 promotion (24) 5675 0.13 0.33 0.02 0.06 0.06 0.06 0.06 ‐0.06 ‐0.06 ‐0.01 ‐0.09
11 promotion (36) 3739 0.24 0.43 0.07 0.10 0.06 0.06 0.05 ‐0.06 ‐0.06 ‐0.01 ‐0.08 0.87
12 salary (ln) 11776 11.02 0.4 ‐0.10 ‐0.18 0.08 0.13 0.03 ‐0.01 0.00 0.03 0.03 ‐0.15 ‐0.16
13 job level 11803 4.63 0.76 ‐0.15 ‐0.27 0.12 0.13 0.01 ‐0.01 ‐0.01 0.01 0.03 ‐0.32 ‐0.39 0.73
14 same dept 11803 0.58 0.49 ‐0.39 ‐0.06 0.17 0.17 0.11 ‐0.14 ‐0.14 ‐0.03 ‐0.11 ‐0.03 ‐0.10 ‐0.02 ‐0.01
15 same function 11803 0.68 0.47 ‐0.14 ‐0.02 0.03 0.08 0.06 ‐0.05 ‐0.05 ‐0.01 ‐0.02 ‐0.09 ‐0.12 0.07 0.03 0.21
16 promotion 11803 0.48 0.5 0.15 0.13 0.03 0.05 0.08 ‐0.05 ‐0.02 ‐0.04 ‐0.08 ‐0.07 ‐0.09 ‐0.30 ‐0.16 0.08 0.05
17 expansion 11803 0.39 0.49 ‐0.10 ‐0.14 0.00 ‐0.03 ‐0.06 0.03 0.01 0.02 0.04 0.03 0.03 0.23 0.19 0.03 ‐0.17
18 transfer 11803 0.13 0.33 ‐0.09 0.01 ‐0.04 ‐0.03 ‐0.04 0.03 0.01 0.04 0.06 0.06 0.09 0.12 ‐0.04 ‐0.16 0.17
19 female 11697 0.64 0.48 0.01 0.04 ‐0.05 ‐0.03 0.02 ‐0.02 ‐0.02 ‐0.03 ‐0.04 0.01 0.00 ‐0.22 ‐0.14 0.01 ‐0.03
20 minority 11697 0.24 0.42 0.03 0.05 ‐0.07 ‐0.06 ‐0.01 0.02 0.03 0.01 0.00 ‐0.03 ‐0.02 ‐0.11 ‐0.11 ‐0.03 0.02
21 white 11803 0.76 0.43 ‐0.02 ‐0.06 0.07 0.06 0.01 ‐0.02 ‐0.03 ‐0.01 0.00 0.02 0.01 0.12 0.11 0.03 ‐0.02
22 black 11803 0.1 0.3 0.04 0.06 ‐0.09 ‐0.10 ‐0.01 0.02 0.04 0.00 ‐0.01 ‐0.01 ‐0.01 ‐0.12 ‐0.13 ‐0.04 0.01
23 latino 11803 0.08 0.27 0.01 0.04 ‐0.03 0.02 0.00 0.01 ‐0.01 0.02 0.01 ‐0.03 ‐0.02 ‐0.10 ‐0.06 0.00 0.00
24 asian 11803 0.03 0.18 ‐0.03 ‐0.04 0.05 0.04 0.00 0.01 0.01 0.01 0.02 ‐0.02 ‐0.01 0.10 0.08 0.02 0.02
25 tenure 11803 5.4 5.22 ‐0.04 0.00 0.10 0.14 0.02 ‐0.03 ‐0.03 ‐0.03 ‐0.07 ‐0.06 ‐0.06 0.14 0.15 0.01 ‐0.05
26 tenure2 11803 56.33 122.2 ‐0.07 ‐0.02 0.05 0.08 0.01 ‐0.02 ‐0.01 ‐0.02 ‐0.05 ‐0.06 ‐0.07 0.12 0.12 0.04 ‐0.03
27 age 11697 38.59 9.93 ‐0.15 ‐0.12 ‐0.02 0.03 ‐0.02 0.01 0.00 0.01 0.04 ‐0.15 ‐0.20 0.31 0.26 0.03 0.03
28 age2 11697 1588.12 837.31 ‐0.15 ‐0.12 ‐0.03 0.02 ‐0.03 0.02 0.00 0.01 0.05 ‐0.14 ‐0.19 0.29 0.24 0.03 0.03
29 rating group size 10115 479.3 408.2 0.05 0.07 0.03 ‐0.01 0.02 ‐0.04 ‐0.06 0.01 ‐0.01 0.07 0.12 ‐0.18 ‐0.23 0.05 0.01
30 last salary (ln) 11787 10.93 0.42 ‐0.17 ‐0.23 0.09 0.13 0.02 ‐0.01 ‐0.02 0.03 0.04 ‐0.13 ‐0.14 0.94 0.72 0.00 0.07
31 last contribution 9866 2.89 0.61 ‐0.03 ‐0.05 0.52 0.42 0.24 ‐0.23 ‐0.19 ‐0.05 ‐0.09 0.04 0.08 0.09 0.12 0.10 0.00
32 last competency 9897 2.57 0.51 ‐0.02 ‐0.01 0.41 0.59 0.25 ‐0.25 ‐0.20 ‐0.05 ‐0.09 0.01 0.04 0.13 0.12 0.10 0.03
33 year 11803 2010.51 1.31 0.04 0.08 0.02 0.02 ‐0.02 0.01 ‐0.01 0.23 0.47 0.28 0.54 0.04 0.04 ‐0.19 ‐0.11
16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32
17 expansion ‐0.77
18 transfer ‐0.37 ‐0.30
19 female 0.01 ‐0.02 0.01
20 minority 0.07 ‐0.06 ‐0.02 0.02
21 white ‐0.07 0.06 0.02 ‐0.02 ‐1.00
22 black 0.06 ‐0.05 ‐0.03 0.07 0.62 ‐0.60
23 latino 0.04 ‐0.04 0.00 0.02 0.52 ‐0.51 ‐0.10
24 asian ‐0.01 0.00 0.01 ‐0.08 0.34 ‐0.33 ‐0.06 ‐0.05
25 tenure ‐0.09 0.11 ‐0.02 0.08 ‐0.06 0.07 ‐0.01 ‐0.02 ‐0.03
26 tenure2 ‐0.09 0.09 ‐0.01 0.08 ‐0.06 0.07 ‐0.02 ‐0.03 ‐0.04 0.92
27 age ‐0.25 0.18 0.11 0.06 ‐0.08 0.08 ‐0.05 ‐0.03 ‐0.03 0.28 0.29
28 age2 ‐0.23 0.16 0.11 0.06 ‐0.09 0.09 ‐0.06 ‐0.04 ‐0.03 0.26 0.28 0.99
29 rating group size 0.11 ‐0.11 0.00 ‐0.03 0.05 ‐0.06 0.01 0.04 0.06 0.03 0.01 ‐0.09 ‐0.09
30 last salary (ln) ‐0.40 0.29 0.18 ‐0.21 ‐0.13 0.13 ‐0.13 ‐0.12 0.11 0.18 0.16 0.36 0.34 ‐0.16
31 last contribution 0.05 ‐0.02 ‐0.05 ‐0.03 ‐0.06 0.06 ‐0.10 ‐0.01 0.04 0.14 0.07 ‐0.01 ‐0.02 0.01 0.09
32 last competency 0.08 ‐0.06 ‐0.03 0.00 ‐0.03 0.03 ‐0.10 0.05 0.02 0.20 0.11 0.05 0.04 ‐0.01 0.12 0.67
33 year ‐0.06 0.06 0.01 0.01 0.01 ‐0.02 0.00 0.02 ‐0.01 0.07 0.04 0.03 0.02 0.11 0.05 0.01 0.01

a
Unit of analysis is an internal hire
 49

TABLE 2
Performancea
Model 1 Model 2 Model 3 Model 4 Model 5
OLS OLS Logit Logit Logit
Variables Contribution Competency Top 25% Bottom 50% Bottom 25%
Posting 0.0542** 0.0621** 0.129* -0.133** -0.149*
[0.0155] [0.0123] [0.0548] [0.0516] [0.0634]
Level 4 -0.0406 -0.117* -0.34 0.275 0.14
[0.0862] [0.0551] [1.086] [1.001] [0.751]
Level 5 0.15 -0.395* -0.592 0.678 0.798*
[0.251] [0.166] [1.005] [0.916] [0.339]
Level 6 0.0491 -0.143 0.0583 0.128 0.566*
[0.260] [0.174] [0.200] [0.197] [0.271]
Same Department 0.208** 0.175** 0.392** -0.439** -0.495**
[0.0159] [0.0126] [0.0586] [0.0531] [0.0635]
Same Function 0.0316* 0.0751** 0.295** -0.155** -0.152*
[0.0149] [0.0125] [0.0591] [0.0534] [0.0644]
Promotion 0.00351 -0.0219 0.242** -0.0874 0.0152
[0.0216] [0.0168] [0.0820] [0.0718] [0.0857]
Female 0.0026 0.0233† 0.115* -0.108* -0.101
[0.0146] [0.0121] [0.0550] [0.0511] [0.0617]
Black -0.156** -0.135** -0.116 0.129 0.325**
[0.0240] [0.0193] [0.0863] [0.0790] [0.0888]
Latino -0.0497† 0.0601** -0.000748 0.0865 0.0346
[0.0272] [0.0232] [0.0951] [0.0897] [0.110]
Asian 0.000733 -0.0291 -0.0805 0.187 0.214
[0.0328] [0.0285] [0.140] [0.129] [0.146]
Other -0.0454 -0.0501 -0.0131 -0.0578 -0.0164
[0.0601] [0.0458] [0.193] [0.189] [0.229]
Tenure 0.0350** 0.0387** 0.0534** -0.0476** -0.0524**
[0.00379] [0.00345] [0.0148] [0.0129] [0.0153]
Tenure (sq) -0.00120** -0.00132** -0.00195** 0.00157** 0.00170**
[0.000155] [0.000148] [0.000663] [0.000548] [0.000613]
Age -0.00441** -0.00120† -0.00767** 0.00663* 0.00684*
[0.000828] [0.000676] [0.00291] [0.00266] [0.00324]
Rating Group Size 0.0000295 -0.000145* -0.000385**
[6.87e-05] [6.64e-05] [8.65e-05]
Constant 2.956@ 2.838@ -1.223! 0.0529 -0.973
[0.118] [0.0853] [0.524] [0.508] [0.599]
Observations 9300 9300 9289 9276 9262
R-squared 0.136 0.154
a
Standard errors, in brackets, are clustered by individual. All analyses also include dummies for job, function, state, and year.
† p < .10; ** p < .05; ** p < .01
 50

TABLE 3
Turnover, Advancement & Salarya

Model 6 Model 7 Model 8 Model 9 Model 10 Model 11

Logit Logit Logit Logit OLS OLS
Variables Turnover12 Turnover24 Promotion24 Promotion36 Salary (ln) Salary (ln)
Posting -0.202* -0.182* -0.0949 0.128† 0.0383** 0.0376**
[0.0986] [0.0761] [0.0645] [0.0733] [0.00256] [0.00272]
Level 4 0.754 0.058 -0.0866 -0.526 0.258** -0.0489
[0.861] [0.829] [0.222] [0.561] [0.0498] [0.0688]
Level 5 0.511 -0.546 -1.172** -1.506** 0.167** 0.0159
[0.467] [0.426] [0.148] [0.553] [0.0218] [0.0526]
Level 6 0.539† -0.365 -1.563** -2.166** 0.221** 0.0331
[0.311] [0.340] [0.361] [0.445] [0.0300] [0.0559]
Same Dept -0.0925 -0.259** 0.0402 0.0012 0.000872 0.00125
[0.0959] [0.0738] [0.0651] [0.0779] [0.00270] [0.00275]
Same Function -0.0336 0.0874 -0.217** -0.0758 0.0166** 0.0152**
[0.0987] [0.0671] [0.0610] [0.0756] [0.00264] [0.00270]
Promotion -0.302** 0.0791 -0.274** -0.279** 0.0960** 0.0865**
[0.116] [0.0954] [0.0777] [0.0892] [0.00874] [0.00890]
Female -0.236** -0.148* -0.121* -0.123† -0.0101** -0.00912**
[0.0871] [0.0648] [0.0577] [0.0666] [0.00230] [0.00235]
Black 0.0358 -0.0981 -0.269** -0.302** -0.00637* -0.00126
[0.131] [0.109] [0.0896] [0.0962] [0.00313] [0.00331]
Latino -0.116 -0.0186 -0.349** -0.395** 0.000767 0.00714
[0.165] [0.111] [0.101] [0.125] [0.00494] [0.00514]
Asian -0.0644 -0.0117 0.005 0.2 -0.00445 -0.00181
[0.214] [0.151] [0.159] [0.178] [0.00406] [0.00421]
Tenure -0.104** -0.0828** -0.00392 0.0333† -0.000281 -0.0009
[0.0223] [0.0159] [0.0152] [0.0187] [0.000559] [0.000588]
Tenure (sq) 0.00299** 0.00212** 0.0000181 -0.00150† 0.00000442 2.66E-05
[0.000912] [0.000625] [0.000653] [0.000882] [2.07e-05] [2.09e-05]
Age -0.00115 0.0114** -0.0144** -0.0153** -0.00153† -0.00163†
[0.00523] [0.00316] [0.00319] [0.00365] [0.000829] [0.000873]
Last salary (ln) 0.851** 0.863**
[0.00924] [0.00893]
Last contribution 0.00860**
[0.00256]
Last competency 0.00354
[0.00353]
Constant -6.111** -2.416** 0.061 0.596 1.561** 1.342**
[0.757] [0.594] [0.345] [0.633] [0.117] [0.101]
Observations 7451 5056 5178 3302 9292 8017
R-Squared 0.955 0.959
a
Standard errors, in brackets, are clustered by individual. All analyses also include dummies for job, function, state, and year.
† p < .10; ** p < .05; ** p < .01.
 51

TABLE 4
Source of Internal Hires by Posting and Sponsorship

Source of Hire Statistical

(Location of previous job within the organization) Posting Sponsorship Significance

Different Function 40% 27% p < .01

Different Department 64% 26% p < .01

Different City 15% 4% p < .01

Different Building 42% 8% p < .01

Transfer (same job, different department) 10% 15% p < .01

Expansion (same level, different job) 32% 43% p < .01

Promotion (move up a level) 57% 42% p < .01

 52

TABLE 5
Performance, Turnover, Salary, and Advancement within Departmenta,b,c

Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 Model 7 Model 8 Model 9 Model 10
OLS OLS Logit Logit Logit OLS OLS Logit Logit OLS
Bottom Bottom
Variables Contrib. Comp. Top 25% Turn12 Turn24 Prom24 Prom36 Salary (ln)
50% 25%
Posting 0.0474** .0678** 0.160** -0.127* -0.165* -0.284* -0.151 -0.119 0.117 0.0468**
[0.0171] [0.0137] [0.0609] [0.0585] [0.0740] [0.123] [0.0988] [0.136] [0.150] [0.00325]
Level 4 0.0973 0.550** -1.277 -0.215 -0.418 0.442 0.897 0.328 -0.0378 0.231**
[0.0758] [0.0645] [1.505] [0.638] [0.890] [0.897] [0.763] [0.499] [0.524] [0.0637]
Level 5 0.308 0.281† -1.433 0.231 0.481 0.225 0.257 -2.087** -1.967** 0.156**
[0.239] [0.153] [1.444] [0.287] [0.394] [0.520] [0.229] [0.334] [0.388] [0.0259]
Level 6 0.237 0.537** 0.199 0.00687 0.271 0.367 0.329 -2.541** -2.888** 0.116**
[0.249] [0.162] [0.220] [0.231] [0.335] [0.381] [0.360] [0.845] [0.782] [0.0338]
Same Dept 0.152** 0.129** 0.319** -0.389** -.348** -0.0509 -0.0527 0.267† 0.127 0
[0.0190] [0.0152] [0.0688] [0.0624] [0.0765] [0.119] [0.0970] [0.149] [0.172] [0]
Same Funct 0.0196 0.0480** 0.261** -0.142* -0.138† 0.145 0.0614 -0.355** 0.054 0.0106**
[0.0175] [0.0150] [0.0696] [0.0638] [0.0807] [0.127] [0.0868] [0.133] [0.166] [0.00336]
Tenure 0.0396** 0.0441** 0.0662** -00569** -.0637** -0.105** -.0875** 0.00176 0.0842* -0.000655
[0.00413] [0.00376] [0.0162] [0.0146] [0.0174] [0.0250] [0.0184] [0.0348] [0.0418] [0.000650]
Tenure (sq) -.00138** -.00149** -.0023** .00183** .00198** .00315** .00239** -0.000855 -0.005* 0.00000997
[.000173] [.000163] [.000735] [.000623] [0.0007] [.000991] [0.0007] [0.00153] [0.00212] [2.42e-05]
Age -.00470** -.00153* -.0085** .00841** .0101** 0.000178 .0110** -.0311** -.033** -0.00174†
[0000887] [.000728] [0.00317] [0.00294] [0.00360] [0.00603] [0.00383] [0.00668] [0.00760] [0.00101]
Constant 2.830** 2.199** -1.486** 0.202 -0.723 -5.912** -3.468** 0.162 0.532 0.827**
[0.0877] [0.0738] [0.548] [0.528] [0.638] [1.115] [0.566] [0.746] [0.707] [0.134]

Obs 7568 7568 7559 7543 7505 5777 4044 4203 2720 7568
R-squared 0.138 0.161 0.966
a
Standard errors, in brackets, are clustered by individual. All analyses also include dummies for job, function, state, year, ethnicity, and gender.
b
Model 3, 4, and 5 contain additional controls for performance group size
c
Model 10 includes additional controls for last salary, last competency rating, and last contribution rating
† p < .10; ** p < .05; ** p < .01
 53

TABLE 6
Performance Instrumental Variables (Part 1)a
Model 1 Model 2 Model 3 Model4 Model5 Model 6
OLS OLS 2SLS 2SLS 1st stage 2SLS/Treatment 2SLS/Treatment 1st stage
Variables Contribution Competency Contribution Competency Contribution Competency
Posting 0.0542** 0.0621** 0.259** 0.271** 0.227** 0.199**
[0.0155] [0.0123] [0.0934] [0.0764] [0.0765] [0.0609]
Same Department 0.208** 0.175** 0.296** 0.262** -0.415** 0.283** 0.232** -1.265**
[0.0159] [0.0126] [0.0422] [0.0343] [0.01] [0.0353] [0.0281] [0.0344]
Same Function 0.0316* 0.0751** 0.0387* 0.0871** -0.068** 0.0366* 0.0824** -0.194**
[0.0149] [0.0125] [0.0164] [0.0136] [0.01] [0.0164] [0.0131] [0.0379]
Promotion 0.00351 -0.0219 -0.0445 -0.0711** 0.216** -0.0375 -0.0555** 0.636**
[0.0216] [0.0168] [0.0300] [0.0237] [0.016] [0.0264] [0.0211] [0.0483]
Female 0.0026 0.0233† -0.00083 0.0222† -0.00 -0.000927 0.0220† -0.00499
[0.0146] [0.0121] [0.0150] [0.0124] [0.01] [0.0143] [0.0114] [0.0342]
Black -0.156** -0.135** -0.148** -0.130** -0.040** -0.149** -0.133** -0.135**
[0.0240] [0.0193] [0.0245] [0.0197] [0.02] [0.0218] [0.0174] [0.0507]
Latino -0.0497† 0.0601** -0.0469 0.0657** -0.042* -0.0483† 0.0627** -0.141*
[0.0272] [0.0232] [0.0286] [0.0243] [0.02] [0.0247] [0.0196] [0.0579]
Tenure 0.0350** 0.0387** 0.0343** 0.0387** 0.00 0.0345** 0.0391** 0.0183*
[0.00379] [0.00345] [0.00387] [0.00340] [0.00] [0.00349] [0.00278] [0.00859]
Tenure (sq) -0.00120** -0.00132** -0.00115** -0.00128** -0.00** -0.00116** -0.00131** -0.00135**
[0.000155] [0.000148] [0.000160] [0.000144] [0.00] [0.000146] [0.000117] [0.000369]
Age -0.00441** -0.00120† -0.00406** -0.00062 -0.002** -0.00415** -0.000819 -0.00910**
[0.000828] [0.000676] [0.000878] [0.000720] [0.00] [0.000766] [0.000610] [0.00177]
IV: Other posting 0.307** 1.007**
[.02] [0.0670]
lambda -0.102* -0.0793*
[0.0455] [0.0363]
Constant 2.956** 2.838** 2.331** 2.061** 0.925*
[0.118] [0.0853] [0.194] [0.154] [0.451]
Observations 9300 9300 8929 8929 8929 8929 8929 8929
R-squared 0.136 0.154
a
Standard errors, in brackets, are clustered by individual. All analyses also include dummies for job, function, state, and year.
† p < .10; ** p < .05; ** p < .01
 54

TABLE 7
Performance Instrumental Variables (Part 2)a
Model 1 Model 2 Model 3 Model 4 Model 5 Model 6
Logit Logit Logit IV Probit IV Probit IV Probit
Variables Top 25% Bottom 50% Bottom 25% Top 25% Bottom 50% Bottom 25%
Posting 0.129* -0.133** -0.149* 0.305 -0.720** -0.572*
[0.0548] [0.0516] [0.0634] [0.209] [0.204] [0.227]
Same Department 0.392** -0.439** -0.495** 0.331** -0.540** -0.487**
[0.0586] [0.0531] [0.0635] [0.0938] [0.0914] [0.102]
Same Function 0.295** -0.155** -0.152* 0.192** -0.141** -0.120**
[0.0591] [0.0534] [0.0644] [0.0378] [0.0364] [0.0405]
Promotion 0.242** -0.0874 0.0152 0.088 0.083 0.108
[0.0820] [0.0718] [0.0857] [0.0656] [0.0627] [0.0694]
Female 0.115* -0.108* -0.101 0.0726* -0.0670* -0.0577†
[0.0550] [0.0511] [0.0617] [0.0321] [0.0311] [0.0348]
Black -0.116 0.129 0.325** -0.0611 0.0682 0.178**
[0.0863] [0.0790] [0.0888] [0.0490] [0.0475] [0.0519]
Latino -0.000748 0.0865 0.0346 0.0053 0.0266 -0.00066
[0.0951] [0.0897] [0.110] [0.0553] [0.0538] [0.0610]
Tenure 0.0534** -0.0476** -0.0524** 0.0319** -0.0271** -0.0300**
[0.0148] [0.0129] [0.0153] [0.00787] [0.00756] [0.00859]
Tenure (sq) -0.00195** 0.00157** 0.00170** -0.00113** 0.000815* 0.000941**
[0.000663] [0.000548] [0.000613] [0.000334] [0.000318] [0.000363]
Age -0.00767** 0.00663* 0.00684* -0.00379* 0.0023 0.00258
[0.00291] [0.00266] [0.00324] [0.00175] [0.00169] [0.00187]
pgsize 0.0000295 -0.000145* -0.000385** -0.00000173 -0.0000514 -0.000192**
[6.87e-05] [6.64e-05] [8.65e-05] [4.27e-05] [4.22e-05] [4.95e-05]
Constant -1.223* 0.0529 -0.973 -0.482 -0.0919 -0.838
[0.524] [0.508] [0.599] [0.426] [0.491] [0.586]
Observations 9289 9276 9262 8925 8916 8885
a
Standard errors, in brackets, are clustered by individual. All analyses also include dummies for job, function, state, and year.
† p < .10; ** p < .05; ** p < .01
 55

TABLE 8
Turnover, Advancement, and Salary using Instrumental Variablesa

Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 Model 7 Model 8 Model 9 Model 10 Model 11 Model 12
Logit Logit IV Probit IV Probit Logit Logit IV Probit IV Probit OLS OLS 2SLS 2SLS
Salary
Variables Turnover12 Turnover24 Turnover12 Turnover24 Prom24 Prom36 Prom24 Prom36 Salary (ln) Salary (ln) (ln) Salary (ln)
Posting -0.202* -0.182* -1.408* -0.459 -0.0949 0.128† -1.599† 3.060† 0.0383** 0.0376** 0.0443** 0.0430**
[0.0986] [0.0761] [0.637] [1.163] [0.0645] [0.0733] [0.833] [1.649] [0.00256] [0.00272] [0.0157] [0.0154]
Same Dept -0.0925 -0.259** -0.635* -0.37 0.0402 0.0012 -0.604† 1.067† 0.000872 0.00125 0.0039 0.00368
[0.0959] [0.0738] [0.291] [0.499] [0.0651] [0.0779] [0.358] [0.617] [0.00270] [0.00275] [0.00732] [0.00736]
Same Func -0.0336 0.0874 -0.0697 0.0811 -0.217** -0.0758 -0.291** 0.041 0.0166** 0.0152** 0.0169** 0.0155**
[0.0987] [0.0671] [0.111] [0.0855] [0.0610] [0.0756] [0.0767] [0.120] [0.00264] [0.00270] [0.00305] [0.00312]
Promotion -0.302** 0.0791 -0.0374 0.157 -0.274** -0.279** 0.095 -0.795* 0.0960** 0.0865** 0.0935** 0.0819**
[0.116] [0.0954] [0.191] [0.280] [0.0777] [0.0892] [0.207] [0.319] [0.00874] [0.00890] [0.0129] [0.0133]
Female -0.236** -0.148* -0.245* -0.157* -0.121* -0.123† -0.120† -0.114 -0.0101** -0.0912** -0.110** -0.00958**
[0.0871] [0.0648] [0.0975] [0.0664] [0.0577] [0.0666] [0.0620] [0.0836] [0.00230] [0.00235] [0.00235] [0.00241]
Tenure -0.104** -0.0828** -0.103** -0.0782** -0.00392 0.0333† 0.0000092 0.0188 -0.000281 -0.0009 -0.00042 -0.00100†
[0.0223] [0.0159] [0.0262] [0.0183] [0.0152] [0.0187] [0.0175] [0.0231] [0.000559] [0.000588] [0.000567] [0.000602]
Last salary 0.851** 0.863** -0.0181* 0.866**
[0.00924] [0.00893] [0.000844] [0.00908]
Last contr 0.00860** 0.858** 0.00948**
[0.00256] [0.00886] [0.00262]
Last comp 0.00354 0.00243
[0.00353] [0.00378]
Constant -6.111** -2.416** -0.664 -2.26 0.061 0.596 1.448 -2.657 1.561** 1.342** 1.406** 1.091**
[0.757] [0.594] [0.795] [1.420] [0.345] [0.633] [0.934] [1.996] [0.117] [0.101] [0.120] [0.104]
7451 5056 6415 4789 5178 3302 4989 3134 9292 8017 8923 7715
R-Squared 0.955 0.959 0.946 0.949
a
Standard errors, in brackets, are clustered by individual. All analyses also include dummies for job, function, state, year, age and ethnicity.
† p < .10; ** p < .05; ** p < .01
 56

TABLE 9
Effect of External Candidatesa

Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 Model 7

OLS OLS Logit Logit Logit OLS OLS
Contribution Competency Top 25% Bottom 50% Bottom 25% Salary (ln) Salary (ln)
External Candidate -0.039 -0.0535 -0.0588 -0.192 -0.186 -0.00683 -0.00537
[0.0441] [0.0377] [0.171] [0.170] [0.247] [0.00762] [0.00706]
Level 4 -0.972** -0.365† 1.005 -0.494 0.738 0.0533 0.242*
[0.344] [0.218] [1.252] [1.244] [1.841] [0.0377] [0.105]
Level 5 -0.406 0.239 0.7 -0.313 1.535 0.106* 0.271*
[0.452] [0.248] [1.131] [1.106] [1.573] [0.0523] [0.124]
Level 6 -2.000** -0.855** 0.369 -0.783 1.026 0.177** 0.257**
[0.360] [0.233] [0.716] [0.778] [1.074] [0.0559] [0.0704]
Same Department 0.137** 0.104** 0.577** -0.366† -0.278 0.0155* 0.0220**
[0.0493] [0.0393] [0.187] [0.188] [0.282] [0.00786] [0.00775]
Same Function 0.0165 0.0582 0.0209 0.0503 -0.420† 0.0226** 0.0161†
[0.0449] [0.0385] [0.189] [0.185] [0.250] [0.00855] [0.00878]
Advancement 0.0173 0.0394 0.00756 0.0149 0.117 0.0187 0.0311
[0.0686] [0.0567] [0.267] [0.278] [0.420] [0.0408] [0.0424]
Female 0.00303 0.0419 0.228 -0.195 -0.323 -0.0104 -0.00521
[0.0426] [0.0346] [0.169] [0.168] [0.241] [0.00800] [0.00832]
Black -0.0867 -0.064 -0.16 0.136 0.593† -0.0121 -0.0094
[0.0733] [0.0601] [0.268] [0.247] [0.326] [0.00876] [0.00887]
pgsize 0.000636** -0.000620** -0.00110**
[0.000200] [0.000197] [0.000316]
Constant 3.116** 2.265** -2.869* 2.485† 0.991 1.568** 1.809**
[0.301] [0.156] [1.385] [1.332] [1.934] [0.405] [0.522]
Observations 869 869 869 869 869 869 752
R-squared 0.168 0.178 0.955 0.96
a
Standard errors, in brackets, are clustered by individual. All analyses also include dummies for job, function, state, year, and age.
† p < .10; ** p < .05; ** p < .01
57