firm-specific human capital and compensation organizational...

INTRODUCTION

Over the last decade, the transforma-tion to a knowledge-based economyhas increased the demand for pro-fessionals who are highly skilled inthe development and support of in-

formation technology and systems (IT). Al-though growth slowed in the high-tech in-dustry in the early 2000s, particularly in theUnited States, the demand for skilled IT pro-fessionals is rebounding and is projected toincrease later in the decade (Brown, 2004).According to the president of the Informa-

tion Technology Association of America(ITAA; 2002), “a sluggish job market todaycould turn off many prospective informa-tion systems and computer science students,resulting in rampant IT talent shortages afew years down the road.” Hecker (2005) es-timates that for the period 2002–2014, com-puter-related occupations are projected toadd more than 1.8 million jobs to the labormarket, causing IT to be one of the fastest-growing occupational sectors.

In an environment with significantfluctuations in the demand and supply ofIT professionals, compensation emerges as

FIRM-SPECIFIC HUMAN CAPITAL

AND COMPENSATION-

ORGANIZATIONAL TENURE

PROFILES: AN ARCHIVAL ANALYSIS

OF SALARY DATA FOR IT

PROFESSIONALS

S A N D R A A . S L A U G H T E R , S O O N A N G , A N D W A I F O N G B O H

We examine determinants of IT compensation using archival salary datafrom 2,251 IT professionals in Singapore. Consistent with human capital the-ory, we find that professionals in IT jobs requiring more firm-specific humancapital are paid more than those in jobs requiring less firm-specific humancapital. Moreover, compensation increases with organizational tenure at anincreasing rate for professionals in higher firm-specific human capital ITjobs, but at a decreasing rate for those in lower firm-specific human capitalIT jobs. Our results reveal firm-specific human capital as a primary determi-nant of compensation and a moderator of IT compensation-organizationaltenure profiles. © 2007 Wiley Periodicals, Inc.

Correspondence to: Sandra A. Slaughter, PhD, Professor of Information Technology Management, and Costley Chair,College of Management, Georgia Tech University, Atlanta, GA 30308, E-mail: [email protected]

Human Resource Management, Fall 2007, Vol. 46, No. 3, Pp. 373–394

© 2007 Wiley Periodicals, Inc.

Published online in Wiley InterScience (www.interscience.wiley.com).

DOI: 10.1002/hrm.20169

374 HUMAN RESOURCE MANAGEMENT, Fall 2007

an important lever for organizations in re-cruiting and retaining skilled IT employees(ITAA, 2004). In response to imbalances inthe IT labor market, the starting salary forIT professionals increased at unprece-dented rates, particularly during the late1990s. However, although we know thatnew hires in IT have experienced increas-ing compensation, less is known about thecompensation for IT job incumbents, par-ticularly those who have stayed with theircurrent organizations. Do incumbents re-ceive corresponding increases in compen-

sation for their tenure at a par-ticular organization?

It is important to explore therelationship between compensa-tion and organizational tenure inIT for several reasons. The rela-tionship between seniority andearnings is integral to under-standing compensation, employ-ment, and careers in a profession.Rising compensation with tenurecan reflect returns on joint train-ing investments by both the em-ployer and the employee or re-turns on maintaining long-termrelationships between workersand employers. On the otherhand, falling compensation withtenure implies that compensationgrowth over a career depends lesson worker-firm relationships andmore on workers’ own efforts to

continually develop marketable productivecompetencies and monitor opportunitiesavailable in the external labor market. Thus,examining the compensation-tenure rela-tionship addresses a fundamental questionabout the structure of earnings over careerswithin the IT profession. As highlighted bythe ITAA (2004, p. 3), it is important for ITprofessionals to be “stewards of their own ca-reers,” and “understanding the trends anddirections shaping the IT workforce is one ofthe best ways to launch or sustain an infor-mation technology career.” Hence, compen-sation-organizational tenure profiles haveimplications for both IT human resourcemanagers and IT professionals.

Classical human capital theory (Becker,1975; Mincer, 1970) suggests that compensa-tion rises with organizational tenure to re-flect increases in the worker’s competencefrom the accumulation of on-the-job experi-ence. However, given the rapid technologicaladvancements in IT, it is possible that ITskills can become obsolete very quickly if notcontinuously updated (Joseph & Ang, 2002;Pazy, 1996). A rapid erosion of skills suggeststhat compensation may not necessarily risewith organizational tenure in IT, as employ-ees who stay in a particular job for too longmay find their skills becoming obsolete. Onthe other hand, the business and managerialcompetencies needed to successfully imple-ment information systems and technologyin a particular organization are acquiredthrough on-the-job experience, which sug-gests that the value of firm-specific humancapital could increase with organizationaltenure for IT professionals.

This study examines whether compensa-tion increases with organizational tenure forIT professionals. Measuring the level of firm-specific human capital required for differentIT jobs, we evaluate how the firm specificityof different job types affects the salaries of2,251 IT professionals working in informa-tion systems development and infrastruc-ture jobs.

Theory and Hypotheses

Human capital theory offers a cogent expla-nation for the relationship between tenureand compensation (i.e., the compensation-tenure profile) (Topel, 1991). Human capitalembodies an individual’s productive compe-tencies that result from natural ability, edu-cation, training, and experience (Becker,1975). Accordingly, human capital theorysuggests that compensation is positively as-sociated with organizational tenure becausetenure reflects the level of human capital ac-cumulated over the period of employment(Becker & Lindsay, 1994; Parent, 2002;Topel, 1991).

Empirical studies on the compensation-tenure relationship have been conducted pri-marily in the field of labor economics. How-

…examining the

compensation-

tenure relationship

addresses a

fundamental

question about the

structure of

earnings over

careers within the

IT profession.

Human Resource Management DOI: 10.1002/hrm

An Archival Analysis of Salary Data for IT Professionals 375

ever, research to date has not shown consis-tent evidence of this relationship. Altonjiand Shakotko (1987) and Abraham and Far-ber (1987), for example, found the effect oftenure on compensation to be almost in-significant, whereas others (e.g., Goldsmith& Veum, 2002; Jacobsen & Levin, 2002;Topel, 1991) found strong, positive tenure ef-fects on compensation. It is thus recognizedthat compensation dynamics are more com-plicated and variegated than the simple pre-sumption that compensation monotonicallyincreases with job tenure (Munasinghe &O’Flaherty, 2005).

Researchers have attempted to examinethe complexity of compensation dynamicsby accounting for other factors such as theimpact of external labor market conditionsincluding the market distribution of externalwage offers (Munasinghe & O’Flaherty,2005), or exogenous worker displacementsfrom larger firms (Bingley & Westergaard-Nielsen, 2003). Some researchers have specu-lated that certain firms with more firm-spe-cific capital involved will adopt moreaggressive techniques for retaining theirworkers (Coleman, 1998; Farber, 1999). Inthis study, we extend previous research froma firm-level explanation to a job-level expla-nation for differential patterns in compensa-tion-tenure profiles. Specifically, we examinethe economic returns to tenure in IT bypositing that returns to tenure could differfor jobs requiring different levels of firm-spe-cific human capital.

Human Capital and Compensation-Tenure Profiles in IT

Human capital theory identifies two types ofhuman capital: general and specific (Becker,1975). General human capital increases theindividual’s productivity to many firms. Incontrast, firm-specific human capital in-creases the individual’s productivity exclu-sively to a particular firm. Human capitaltheory therefore suggests that the growth ofcompensation with organizational tenure(the duration of an individual’s employmentat a particular organization) is attributed tothe individual’s share of investments in firm-

specific skills (Parent, 2002). That is, individ-uals raise the value of their firm-specifichuman capital through on-the-job experi-ence over time, subsequently enhancingtheir productivity to their firms.

In IT, both general and specific humancapital are very important and necessary(Ives & Olson, 1981; D. M. S. Lee, Trauth, &Farwell, 1995; Nakayama & Sutcliffe, 2001).IT jobs are complex, requiring knowledge ofdifficult, abstract technical concepts such asdata modeling, database design, process en-gineering, information architectures, net-work design, and software designtheory (Feeny & Wilcocks, 1998;D. M. S. Lee et al., 1995; Todd,McKeen, & Gallupe, 1995). SuchIT knowledge and skills are gener-ally transferable from firm tofirm. However, due to the fast rateof change in IT, where new gener-ations of the technology canemerge every 18 months, thevalue of this human capital couldrapidly erode over time absentfurther investment in learningnew skills (Glass, 2000). This po-tential for erosion of technicalhuman capital with organiza-tional tenure is an interesting fea-ture of the IT profession.

At the same time, the applica-tion of IT to a specific organiza-tion’s business problems involvesa high degree of tacit knowledgeabout organizational systems, structures,members, and procedures, and tacit skillsabout handling people and negotiating orga-nizational politics. This knowledge and asso-ciated skills mostly are firm-specific and arebest acquired through experience at the par-ticular organization (Nakayama & Sutcliffe,2001; Schenk, Vitalari, & Davis, 1998;Stinchcombe & Heimer, 1988). For example,to determine the IT strategies of an organiza-tion, or to lead IT projects involving manyparts of the organization, IT professionalsneed to have knowledge of their firm’s cor-porate strategy in relation to its industry, thedecision-making processes and biases of themanagement, the organizational structure,

…we examine the

economic returns to

tenure in IT by

positing that returns

to tenure could

differ for jobs

requiring different

levels of firm-

specific human

capital.



and other business processes and procedures(Levinson, 1988). To effectively manage ITprojects and to communicate with usersabout their requirements, IT professionalsalso require skills to manage interpersonalrelationships within the organization and toplan, organize and delegate tasks and proj-ects (C. K. Lee, 2005). The IT professionalalso must learn to build relationships withdifferent groups of individuals within thecorporation, including diverse users fromdifferent functional departments (e.g., fi-nance, marketing, accounting, production),

as well as with peers, superiors,and specific vendors that offer ITservices and products to the or-ganization. Such skills andknowledge are very much specificto a firm, are less transferablefrom firm to firm, and are ac-quired with work experience andorganizational tenure (Kim &Lerch, 1997).

Jobs as Embodiments ofFirm-Specific Capital

Different IT jobs require differentextents of firm-specific humancapital. An IT job such as an ap-plication programmer requiresless firm-specific human capitalas the knowledge and skills rele-vant to job performance (e.g.,proficiency in programming lan-

guages) are applicable across organizations.In contrast, an IT job such as a systems ana-lyst requires more firm-specific human capi-tal as the knowledge and skills relevant tojob performance (e.g., building and sustain-ing effective relationships with certain usersin functional units) are more relevant to aparticular organization (Ang & Slaughter,2000; D. M. S. Lee et al., 1995; Nakayama &Sutcliffe, 2001).

Compensation for different jobs in or-ganizations is determined by external labormarkets and internal managerial decisions,typically in the form of job evaluation(Campion & Berger, 1990). Job evaluation isa formal procedure that determines the

value or “worth” of a job within an organi-zation’s overall pay structure (Figart, 2001).Job evaluations typically are based on aquantitative evaluation system that consid-ers compensable factors. Four key categoriesof compensable factors are skills (e.g., yearsof education, degree of firm-specific skills re-quired); effort (primarily physical); responsi-bility for machinery, money, or people; andworking conditions (Figart, 2001). A key at-tribute of job evaluation is that it focuses ondetermining a rate for a job, rather than de-termining the compensation for a particularperson (Figart, 2001).

As skills often are a primary focus of jobevaluations, jobs that require higher levels offirm-specific skills are often more highlycompensated, controlling for job level (Cap-pelli & Cascio, 1991). This differentiation oc-curs due to the costs of replacing workerswhose jobs require a high level of firm-spe-cific human capital. According to humancapital theory, firms have to invest in devel-oping firm-specific human capital sincethere is no market for such firm-specificskills elsewhere, and job applicants cannotknow the job requirements or prepare forthem in advance (Becker, 1975). The poten-tial cost of replacing and retraining suchworkers with firm-specific skills if they doleave is thus very great; hence, firms offerhigher compensation to such employees toprevent them from leaving (Cappelli & Cas-cio, 1991).

Moreover, employees do not have an in-centive to invest in developing human capi-tal that is not transferable to other firms(Schoones & Bernhardt, 1998). To provideincentives for employees to develop firm-specific human capital such as that requiredto build effective relationships with users, afirm must provide inducements in the formof higher pay for jobs that require more firm-specific skills (Prendergast, 1993). Thus:

H1:Controlling for job level, individuals in ITjobs that require higher levels of firm-specifichuman capital are paid more, for a givenlevel of organizational tenure, than those inIT jobs requiring lower levels of firm-specifichuman capital.

As skills often are a

primary focus of job

evaluations, jobs

that require higher

levels of firm-

specific skills are

often more highly

compensated,

controlling for job

level.



Our next two hypotheses focus on therate of change in compensation with orga-nizational tenure for jobs varying in firmskills specificity. Human capital theory im-plies that compensation rises with organi-zational tenure to reflect increases in theworker’s human capital from on-the-job ex-perience (Topel, 1991). In a job requiring ahigh level of firm-specific skills, an IT pro-fessional’s firm-specific human capital willcontinue to increase over time with job ex-perience. With longer tenure, IT profes-sionals develop greater firm-specific knowl-edge and skills by deepening andimproving existing relationships and devel-oping new relationships with other organi-zational members or even with externalconstituents such as vendors and tradingpartners. IT professionals also becomemore knowledgeable about tactics for nego-tiating organizational politics and bureau-cracies, and this enables them to workmore effectively. Finally, as IT professionalswork for a longer period with the organiza-tion, they gain more knowledge about thebusiness processes and strategies of the or-ganization, which in turn facilitates theirwork, especially since their job typically re-quires a large extent of such knowledgeabout the organization.

Such development of firm-specific skills,of course, does not automatically come withtenure if IT professionals do not activelylearn from their experiences. Nevertheless,in a job that requires a high level of firm-spe-cific skills, opportunities exist for IT profes-sionals to develop such firm-specific skills onthe job, through learning by doing (Argote,1999). The learning from experience, there-fore, is expected to help IT professionals todevelop firm-specific skills.

The increase in firm-specific humancapital will enable IT professionals in a jobrequiring a high level of firm-specific skillsto be even more effective. Such IT profes-sionals thus become even more costly to re-place over time and, correspondingly, thefirm is more willing to continuously in-crease compensation for incumbents of ITjobs that require more firm-specific humancapital. Hence:

H2:Controlling for job level, compensation in-creases at an increasing rate with organiza-tional tenure for individuals in IT jobs withhigher levels of firm-specific human capital.

By contrast, in a job requiring less firm-specific human capital, an IT professional’shuman capital will increase initially becausethe professional needs time to adjust andadapt the skills and knowledge to a new en-vironment, which is consistent with classicalhuman capital theory. However,prior empirical research finds thathuman capital that is less firm-specific tends to erode with ageand tenure (Allen & de Grip,2005; Dalton & Thompson, 1971;Fossum & Arvey, 1986; Shearer &Steger, 1975) if professionals donot keep their skills up-to-date.

According to human capitaltheory, a firm is unwilling to in-vest in increasing the generalhuman capital of employees be-cause the firm cannot reap thebenefits from the investmentwhen the employees leave thefirm and transfer the generalhuman capital to another firm. InIT, organizations could be reluc-tant to train their IT professionalsfor fear that up-to-date humancapital will make these profes-sionals more attractive in the ex-ternal labor market.

In addition to firms’ reluc-tance to develop the generalhuman capital of employees,there are varied reasons why wewould expect to see the erosion ofhuman capital with age andtenure in IT. First, workers tend toexperience decreasing levels of motivation asthey grow older because of the difficulties inhaving to unlearn what they have special-ized in for a long time (Larwood, Rodkin, &Judson, 2001). Second, older workers are lesslikely to be recommended for training com-pared to equally qualified younger individu-als as a result of employers’ perceptions thatthey are difficult to train, or are not inter-

With longer tenure,

IT professionals

develop greater

firm-specific

knowledge and skills

by deepening and

improving existing

relationships and

developing new

relationships with

other organizational

members or even

with external

constituents such as

vendors and trading

partners.



ested in receiving training (Chiu, Chan,Snape, & Redman, 2001; Taylor & Walker,1998). Third, and more important, with in-creasing tenure and repetition of tasks in ajob, unused skills and knowledge are forgot-ten. Over time, older or longer-tenured indi-viduals are able to maintain a core of skill

competencies allowing them tofunction at an acceptable level solong as the job does not change(Larwood et al., 2001).

It is important to note thatgeneral human capital need notnecessarily erode with organiza-tional tenure, especially when ITprofessionals are continually up-dating and developing deepknowledge of a select set of skillsand when the organization stayscurrent with the latest technol-ogy releases. However, we ac-knowledge the persistent concernof IT professionals that deepknowledge acquired from work-ing on a specific technology orsystem (especially if it is legacyand proprietary in nature) doesnot necessarily translate readilyto marketable skills.

This observation is consistentwith Farmer and Campbell(1997), who found that IT profes-sionals often work within a verynarrow scope of technology andapplication systems domains.They state that due to overspecial-

ization, IT professionals can become frus-trated and concerned that they are compe-tent to conduct their specific job only. AsFarmer and Campbell (1997, p. 129) note:

This may be compounded by a feelingthat their competence even in that jobis slipping away as the pace of techno-logical change races inexorably on andthey perceive themselves as failing tokeep up.

In sum, the value of human capital thatis less firm-specific diminishes with organi-zational tenure. Thus, we expect compensa-

tion to rise more slowly with organizationaltenure in IT jobs that require little firm-spe-cific human capital. That is:

H3:Controlling for job level, compensation in-creases with organizational tenure, but at adecreasing rate for individuals in IT jobs withlower levels of firm-specific human capital.

Method

Research Setting and Data

To evaluate our hypotheses, we examineannual salary data collected for 2,251 ITprofessionals across 43 organizations inSingapore. Singapore provides an ideal set-ting in which to study compensation-tenure profiles for IT professionals. Wellknown for its nationwide strategic plan,the “IT2000 Vision,” Singapore has lever-aged IT for national competitive advantage.As part of the vision, the Singapore govern-ment has the goal of becoming an IT leaderand has implemented significant economicincentives to attract major IT companies tothe country. These incentives have led toincreased demand for IT labor such thatthere are shortages of qualified IT profes-sionals (Ang, Slaughter, & Ng, 2002). Thus,Singapore has a highly competitive IT labormarket. However, because the country issmall and its population is not as diversi-fied as the United States or Europe, it is lessheterogeneous. Further, there is less re-liance on importing foreign workers in Sin-gapore, due to country-size limitations. Inaddition, organizations in Singapore arevery consistent in their definition of ITjobs. These factors result in a competitiveIT labor market but with better controlsthan in the United States or Europe, miti-gating the heterogeneity problems thathave plagued prior research on compensa-tion-tenure profiles.

The Information Technology Manage-ment Association (ITMA), the premier IT asso-ciation in Singapore, provided the salary data.Members of the ITMA include major IT usersrepresenting both public- and private-sectororganizations in industries such as financial

Singapore provides

an ideal setting in

which to study

compensation-

tenure profiles for

IT professionals.

Well known for its

nationwide

strategic plan, the

“IT2000 Vision,”

Singapore has

leveraged IT for

national

competitive

advantage.



services, manufacturing, retailing, and trans-portation. Data were obtained from the top ITexecutive or CIO (Chief Information Officer)as well as HR specialists of ITMA member or-ganizations. CIOs provided information oncompany background and demographic char-acteristics of the IT department. HR specialistsprovided salary information for all IT profes-sionals in each organization. Forty-three or-ganizations participated in the ITMA SalaryStudy, representing almost half of the ITMAmember organizations. A chi-square analysisbased on organization size and sector revealedno significant respondent bias, as the 43 or-ganizations are representative of the ITMApopulation in terms of size (χ2 = 2.041, p > .10)and sector (χ2 = 1.699, p > .10).

IT Job Families

We examined two basic job families for ITprofessionals: systems development and ITinfrastructure. An IT job family defines agroup of hierarchically interrelated jobs in IT(Chesebrough & Davis, 1983). In the systemsdevelopment job family, the primary job cat-egories include application developmentmanager, project manager, systems analyst,and application programmer. IT infrastruc-ture job categories include infrastructuremanager, technical specialist, systems admin-istrator, and computer operator. Based on thejob descriptions, we classified various jobsinto job categories that represent the primarytasks performed. For example, in the systemsdevelopment job family, Internet program-mer, Web developer, enterprise resource plan-ning (ERP) developer or engineer, and soft-ware engineer all are classified as applicationprogrammers since the primary task in thesejobs involves writing, testing, and imple-menting software code. An example in theinfrastructure job family is the technical spe-cialist, a category that includes database spe-cialist, data analyst or architect, systems ar-chitect, network specialist, wireless specialist,operating systems specialist, and security spe-cialist. These specialists all perform jobs thatinvolve designing and implementing particu-lar aspects of an organization’s technical ITinfrastructure. Table I provides detailed de-

scriptions of the jobs in the systems develop-ment and IT infrastructure jobfamilies, respectively.

Measures

Skills Specificity

To measure the level of skillsspecificity of each job category,we asked a panel of 15 experts torate the extent to which individu-als in each job category (1) re-quire knowledge about their com-pany and skills that are relativelyunique to their company; (2)must make a substantive invest-ment in acquiring company-spe-cific knowledge and in develop-ing skills that are tailored to theparticular needs of their com-pany; and (3) require extensiveknowledge and skills that are spe-cific to their company. Thesemeasures were adapted from Angand Straub (1998). The expertswho conducted the ratings in-cluded nine IT faculty memberswith postgraduate degrees in ITand an average of 11.63 years ofworking experience in the IT in-dustry and six IT industry profes-sionals with an average of 11.57years of working experience inthe IT industry. Fourteen out ofthe 15 experts had college educa-tion, and nine of the experts weremale. The mean age of the 15 ex-perts was 39.1 years.

To examine agreement in theexpert ratings of firm-specificknowledge requirements for eachjob category, we examined the Rwg

(interrater agreement of a groupof judges) (James, Demaree, &Wolf, 1984) and the Cronbach’salpha (reliability of the items) foreach job category. Table II showsthe results of this analysis. Rwg in-dicates convergence among raterson variables associated with a component

To measure the level

of skills specificity of

each job category,

we asked a panel of

15 experts to rate the

extent to which

individuals in each

job category (1)

require knowledge

about their company

and skills that are

relatively unique to

their company; (2)

must make a

substantive

investment in

acquiring company-

specific knowledge

and in developing

skills that are

tailored to the

particular needs of

their company; and

(3) require extensive

knowledge and skills

that are specific to

their company.



(where a value of 1 indicates complete agree-ment). All Rwg scores computed in this analy-sis resulted in positive scores greater than0.70 (ranging from 0.76 to 0.88), suggestinga high level of agreement between the ratersin their assessment of the level of firm speci-ficity for each job category (James, Demaree,& Wolf, 1993). The Cronbach’s alphas ofitem reliability range from 0.74 to 0.97, sug-gesting that the three items measuring thelevel of firm specificity for each job categoryhave a high level of reliability. Thus, we cre-ated a measure of the level of firm skillsspecificity for each job category (SKILLS

SPECIFICITY) by averaging the ratings of thethree items by the 15 expert raters for the re-spective job category. Table II also shows themeans and standard deviations for the firm-specificity ratings of each job.

In the systems development job family,the application development manager israted to require the highest level of firm-specific knowledge, followed by the systemsanalyst, and then the project manager. Theapplications programmer is rated to requirethe lowest level of firm-specific knowledge.In the IT infrastructure job family, the in-frastructure manager is rated to require the


Position Job Description

SYSTEMS DEVELOPMENT JOB FAMILY

Applications Plans and oversees multiple projects and project managers. Works with senior Development Manager management to determine systems development strategy and standards. (AM) Administers department budget and reviews project managers.

Project Manager Plans and oversees the development and support of one or more projects. (PM) Coordinates resources, schedules, and communications for applications develop-

ment projects. Develops project schedules and assigns tasks for applications devel-opment projects.

Systems Analyst (SA) Works directly with management and users to analyze, specify, and design busi-ness applications. Serves as contact with user groups and systems management.

Applications Develops detailed functional, system, and program specifications. Codes and Programmer (AP) maintains business applications on particular hardware/software platforms.

IT INFRASTRUCTURE JOB FAMILY

Infrastructure Plans, manages, and coordinates all matters related to IT hardware, software, Manager (IM) systems, data, network, and telecommunications infrastructure.

Technical Specialist (TS) Designs and manages specific aspects of the IT infrastructure such as hardware,operating systems, communication systems, Internet applications, database man-agement systems, or networks, etc.

Systems Administrator Installs operating systems, software, database management systems software, (AD) compilers, and utilities. Monitors and tunes systems software, peripherals, and net-

works. Installs new users, creates batch administration scripts, and runs systemsbackups. Resolves systems problems.

Computer Operator Performs tasks associated with operational computing and peripheral equipment in (CO) accordance with the workload, services, priorities, and deadlines as defined in the

daily production/testing schedule.

T A B L E I Job Category Descriptions


highest level of firm-specific knowledge(comparable to the firm-specific knowledgerequirements of the project manager in thesystems development job family). Whilethe systems administrator was rated to re-quire a higher level of firm-specific knowl-edge compared to the technical specialist,the ratings for these two job categories wereclose to that of the application programmer.The computer operator was rated to requirethe lowest level of firm-specific knowledgeamong all of the IT job categories.

Compensation was measured as the sumof base compensation (i.e., the 12-monthsalary of an individual) plus bonuses for theyear (COMPENSATION). Organizationaltenure is measured as a continuous variablethat assesses the number of years an IT pro-fessional has worked in the company (OR-GANIZATIONAL TENURE) as of the time ofthe study. Employees with less than one yearof tenure were scored as having zero years oftenure. Similarly, those whose tenure wasmore than one year but less than two yearswere coded as having one year of tenure,and so on.

Control Variables

We control for other variables that prior re-search has suggested are likely to influencecompensation, including level of education(GRADUATE), modeled as a dummy variablewith 1 representing college graduates and 0representing noncollege graduates; gender(GENDER), modeled as a dummy variablewhere 1 represents males and 0 represents fe-

males; organization sector (ORGANIZATIONSECTOR), modeled as a dummy variablewhere 1 represents for-profit sector firms and0 represents not-for-profit sector firms; andorganization size (ORGANIZATION SIZE),modeled as the natural logarithm of the totalnumber of employees in the firm to correctfor skewness in the variable’s distribution.We also control for job level (JOB LEVEL)using a variable to represent the implicit hi-erarchy of progression within each IT jobfamily. The levels of this variable correspondto the job categories shown and described inTable I. For the systems development jobfamily, application programmer is at level 1,followed by systems analyst at level 2, thenproject manager at level 3, and the applica-tion development manager at level 4. For theinfrastructure job family, the computer oper-ator is at level 1, followed by systems admin-istrator at level 2, then technical specialist atlevel 3, and the infrastructure manager atlevel 4.

Models and Analysis

While the systems development and IT infra-structure job families are related and consid-ered to be part of the same occupationalgroup, a different skill set is required in the ca-reer progression within each job family. Theranking of job levels is clear within a job fam-ily but is much less obvious across job fami-lies. For example, a technical specialist or sys-tem administrator is clearly ranked higher interms of job level than a computer operator inthe IT infrastructure job family. However, it is


Job Type Rwg Cronbach’s Alpha Mean Rating S.D. of Ratings

Applications Manager (AM) 0.81 0.88 5.289 1.326

Project Manager (PM) 0.82 0.83 4.644 1.294

Systems Analyst (SA) 0.81 0.89 5.111 1.333

Applications Programmer (AP) 0.82 0.96 3.533 1.290

Infrastructure Manager (IM) 0.81 0.76 4.956 1.331

Technical Specialist (TS) 0.80 0.94 3.156 1.357

System Administrator (AD) 0.76 0.85 3.733 1.429

Computer Operator (CO) 0.88 0.74 2.133 1.075

T A B L E I I Expert Ratings of Firm Skills Specificity for Each Job Category (n = 15 raters, J = 3 items)


not apparent how a technical specialist andsystem administrator should be ranked rela-

tive to application programmer,systems analyst, or project man-ager. It is therefore much more dif-ficult to control for job levels if wecombine both job families to-gether into one data set for analy-sis given that we cannot deter-mine how jobs within the ITinfrastructure family are rankedvis-à-vis those within the systemsdevelopment family. In addition,pooling the job families togetherrequires homogeneity of variancein the dependent variables andsimilar means and standard devia-tions of the independent variables;these requirements are not met inour data. Hence, we conductedseparate analyses for the systemsdevelopment and IT infrastructurejob families. We estimated the fol-lowing model for each job family:

ln (COMPENSATION) = β0 +β1GRADUATE + β2GENDER +β3ORGANIZATION SECTOR +β4ORGANIZATION SIZE +β5JOB LEVEL+ β6SKILLS SPECI-FICITY + β7ORGANIZATIONAL

TENURE + β8ORGANIZATIONALTENURE2 + β9SKILLS SPECIFICITY *ORGANIZATIONAL TENURE +β10SKILLS SPECIFICITY * ORGANIZA-TIONAL TENURE2 + ε

The normal distribution of residuals is acritical assumption for regression. In ouranalysis, the histogram and normal probabil-ity plots of the residuals indicated that theresiduals were highly skewed, a violation ofthe normality assumption (the assumptionof normally distributed residuals is rejectedby the Kolmogorov-Smirnov test of normal-ity for both systems development, Z = 3.342,p < 0.01, and IT infrastructure, Z = 2.502, p <0.01). We thus used the natural logarithm ofcompensation in our models and analysis toachieve a more symmetrical distribution ofthe error terms (Box & Cox, 1964). Residuals

were normally distributed after these trans-formations (Kolmogorov-Smirnov test ofnormality: Z = 1.599, p > 0.05 for systems de-velopment and Z = 0.868, p > 0.10 for IT in-frastructure). Also, every variable was stan-dardized to its Z-score before entering it intothe analysis to ease interpretation of esti-mated coefficients (as each variable is meas-ured using a different unit scale) and to mit-igate potential collinearity between the maineffects of firm skills specificity and organiza-tional tenure and their interactions (Aiken &West, 1991).

We expect a nonlinear compensation-tenure relationship because different IT jobshave varying levels of firm-specific humancapital and thus have different compensa-tion change rates with organizational tenure.Modeling the compensation-tenure relation-ship as nonlinear is also consistent with theliterature in labor economics (e.g., Mincer,1974). The natural logarithm of the depend-ent variable implies that the coefficient fororganizational tenure represents proportion-ate (percentage) changes in compensationwith changes in tenure; the quadratic termsfor organizational tenure capture the nonlin-earity effect; and the interaction terms forskills specificity and organizational tenurecapture the differences in average compensa-tion and in the change in compensationwith organizational tenure across differentlevels of firm-specific human capital for theIT jobs.

To analyze our models, we use hierarchi-cal moderated regression (Cohen & Cohen,1983), first estimating a model with the con-trol variables, then adding the main effects,and finally adding the interactive and quad-ratic effects.1 Although the ordinary-least-squares (OLS) assumption of normality ofresiduals is met in both models, White’s test(1980) rejects the null hypothesis of ho-moscedasticity. Thus, we employed weightedleast squares, using a general weighting pro-cedure to correct for heteroscedasticity in themodels (Greene, 2000). We found nocollinearity problems in the models, as thevariance inflation factors and condition in-dices were well below the commonly ac-cepted cutoff levels (condition index = 5.339

We expect a

nonlinear

compensation-

tenure relationship

because different IT

jobs have varying

levels of firm-

specific human

capital and thus

have different

compensation

change rates with

organizational

tenure.



and average VIF = 2.448 in systems develop-ment; condition index = 4.954 and averageVIF = 2.354 in IT infrastructure). The exis-tence of outliers was not relevant to the re-sults of the estimation. Application of theBelsely, Kuh, and Welsch (1980) proceduresdetected three outliers in the systems devel-opment model and two outliers in the IT in-frastructure model. However, the results didnot change materially when the analyseswere conducted without the outlying obser-vations; thus, we report the results with allobservations included.

Results

Tables III and IV display descriptive statisticsand the correlation matrices for the variablesin our models for systems development andIT infrastructure, respectively. The resultsfrom the hierarchical estimation of our mod-els are reported in Tables V and VI. As shownin these tables, our control variables aloneexplain significant variation in compensa-tion (58.4% for systems development and53.4% for IT infrastructure). In both models,adding the main effects for skills specificity(i.e., extent of firm-specific human capital inthe job) and organizational tenure explainssignificant incremental variation in compen-sation (R2 change = 9.3%, F = 210.04, p <0.001 for systems development and R2

change = 13.7%, F = 161.75, p < 0.001 for ITinfrastructure). Finally, adding the interac-tion and quadratic effects to the models ex-plains significant incremental variance incompensation (R2 change = 1.6%, F = 24.69,p < 0.001 for systems development and R2

change = 2.1%, F = 17.98, p < 0.001 for IT in-frastructure).

Our first hypothesis posited that thehigher the level of firm-specific human capi-tal required by the job, the greater the com-pensation of the individual in that job, allelse equal. We find that Hypothesis 1 is sup-ported: an individual in a job that requires ahigher level of firm skills specificity receivesa higher level of compensation for a givenlevel of organizational tenure (β6 = 0.032, z =3.556, p < 0.001 for systems developmentand β6 = 0.079, z = 4.158, p < 0.001 for IT in-frastructure).

The second and third hypotheses positthat compensation increases with organiza-tional tenure, and that the rate of increasedepends on the level of firm skills specificity.To examine these hypotheses, we need to es-tablish whether compensation does in factincrease with organizational tenure. To dothis, we compute the first derivative by dif-ferentiating the compensation equation withrespect to organizational tenure, yielding: {1}∂ ln Compensation / ∂YRCOY = β7 +2*β8YRCOY + β9SKILLS SPECIFICITY +


Mean S.D. 1 2 3 4 5 6 7

1. Compensation 3.86 0.40 -

2. Skills Specificity 4.04 0.78 .69** -

3. Organizational Tenure 5.54 5.96 .52** .41** -

4. Job Level 2.33 1.15 .75** .83** .42** -

5. Graduate 0.77 0.42 .18** .14** –.21** .15** -

6. Gender 0.51 0.50 –.02 .03 –.14** –.03 .14** -

7. Organization Sector 0.56 0.50 .08 .03 –.03 .02 –.12** –.03 -

8. Organization Size 0.22 0.42 –.03 –.14** –.07** –.02 .06* .04 –.19**

Notes: N = 1,471. *p < .05; **p < .01. Coding: Graduate: 1 = college graduate, 0 = nongraduate; Sector: 1 = private, 0 = public; Gender: 1 = male, 0 =female. Pearson correlations are reported between pairs of continuous variables, Spearman correlations between pairs of noncontinuous and contin-uous variables, and Phi correlations between pairs of noncontinuous variables. Means and standard deviations are provided for nonstandardizedvariables.

T A B L E I I I Descriptive Statistics and Correlation Matrix (Systems Development Jobs)


2*β10YRCOY*SKILLS SPECIFICITY. Sinceboth YRCOY and SKILLS SPECIFICITY are inexpression {1}, we chose different levels foreach of these variables—low (one standarddeviation below the mean) average, and high(one standard deviation above the mean)—and computed the result of expression {1} forthose different values. We found that the re-sult of expression {1} is positive for all valuesof YRCOY and SKILL SPECIFICITY, indicat-ing that compensation increases with orga-nizational tenure, at all levels of tenure andskill specificity.

Given that compensation does increasewith organizational tenure, our next step isto determine whether compensation is in-creasing at an increasing rate with organiza-tional tenure for high levels of skills speci-ficity as posited in Hypothesis 2, andwhether compensation is increasing at a de-creasing rate with organizational tenure forlow levels of skills specificity as posited inHypothesis 3. To do this, we obtain the sec-ond derivative by differentiating expression{1} again, with respect to organizationaltenure, yielding: {2} ∂2 ln Compensation /∂YRCOY2 = 2*β8 + 2*β10*SKILLS SPECI-FICITY.

Based on expression {2}, when skillsspecificity is high (one standard deviationabove the mean) we find that Hypothesis 2 issupported for systems development, where

the rate of change is positive = 0.023, but notfor IT infrastructure, where the rate ofchange is negative = –0.352. By contrast, wefind that Hypothesis 3 is supported for bothsystems development and IT infrastructure.The range of change is negative (–0.124 and–0.104) for systems development and IT in-frastructure, respectively, when skills speci-ficity is low (one standard deviation belowthe mean).

To give an intuitive view of our results,we graphed the compensation-tenure pro-files for the IT professionals in systems de-velopment and IT infrastructure (Figures 1and 2, respectively). For each job, we plottedfunctions for compensation against organi-zational tenure, holding the values for theother variables constant at their means, andusing the estimated coefficients and the datarange for organizational tenure from thedata sample.

As these graphs illustrate, compensationincreases with organizational tenure at a de-creasing rate for the jobs that require lowerlevels of firm-specific human capital and atan increasing rate for jobs that requirehigher levels of firm-specific human capital.The results shown in Tables V and VI, and il-lustrated graphically in Figures 1 and 2, sug-gest that while compensation increases withorganizational tenure for individuals in allIT jobs (both in systems development and in


Mean S.D. 1 2 3 4 5 6 7

1. Compensation 3.52 0.50 -

2. Skills Specificity 2.91 0.83 .58** -

3. Organizational Tenure 6.32 6.56 .28** –.09* -

4. Job Level 2.20 1.02 .69** .83** –.11** -

5. Graduate 0.36 0.48 .59** .46** –.18** .62** -

6. Gender 0.61 0.49 –.03 .11** –.22** .13** .09* -

7. Organization Sector 0.64 0.48 –.06 –.15** –.06 –.06 –.08* .15** -

8. Organization Size 0.18 0.39 –.06 .21** –.24** .14** –.01 .14** .12**

Notes: N =780. * p < .05; ** p < .01. Coding: Graduate: 1 = college graduate, 0 = nongraduate; Sector: 1 = private, 0 = public; Gender: 1 = male, 0 = fe-male. Pearson correlations are reported between pairs of continuous variables, Spearman correlations between pairs of noncontinuous and continu-ous variables, and Phi correlations between pairs of noncontinuous variables. Means and standard deviations are provided for nonstandardized vari-ables.

T A B L E I V Descriptive Statistics and Correlation Matrix (IT Infrastructure Jobs)


IT infrastructure), none of the individuals injobs in IT infrastructure, including the infra-structure manager, experience compensa-tion increases at an increasing rate. In simu-lating the compensation values for thegraphs, we found that the level of firm-spe-cific human capital needs to be higher thana certain threshold (a level of 5 out of 7 inour data) in order for the rate of compensa-tion to be increasing with organizationaltenure. None of the jobs in IT infrastructure,including infrastructure manager, have alevel of firm-specific human capital overthat threshold value.

However, in systems development, indi-viduals in two jobs with levels of firm-

specific human capital that are above thethreshold value (systems analyst and appli-cation development manager) experiencecompensation increases at an increasingrate. Finally, in both IT job families, individ-uals in jobs with lower levels of firm-specifichuman capital (application programmerand project manager in systems develop-ment and computer operator, systems ad-ministrator, technical specialist and infra-structure manager in IT infrastructure)experience compensation increases, but at adecreasing rate. This result implies that thelevel of firm-specific human capital requiredin an IT job needs to be sufficiently high be-fore IT professionals will experience an in-


Independent Variables Control Variables Only With Main Effects With Moderated Effects

Intercept (β0) 3.856** 3.855** 3.912**(0.006) (0.005) (0.007)

Graduate (β1) 0.034** 0.086** 0.093**(0.006) (0.005) (0.005)

Gender (β2) 0.000 0.010* 0.014**(0.006) (0.005) (0.005)

Organization Sector (β3) 0.019** 0.023** 0.021**(0.006) (0.005) (0.005)

Organization Size (β4) 0.017** 0.016** 0.014**(0.006) (0.005) (0.005)

Job Level (β5) 0.296** 0.177** 0.176**(0.006) (0.009) (0.009)

Skills Specificity (β6) 0.064** 0.032**(0.009) (0.009)

Organizational Tenure (β7) 0.151** 0.209**(0.006) (0.009)

Organizational Tenure2 (β8) –0.037**(0.005)

Skills Specificity * –0.060**Organizational Tenure (β9) (0.009)

Skills Specificity * 0.024**Organizational Tenure2 (β10) (0.004)

R2 0.584 0.677 0.692

R2 Change 0.093 0.016

F Change 210.04** 24.69**

Notes: N = 1,471. ** p <.01, * p < .05, + <.10.

T A B L E V Hierarchical Regression Results: Systems Development Jobs, Dependent Variable = ln (Compensation)


creasing rate of compensation with organi-zational tenure.

Conclusion

Drawing upon human capital theory, weposited that, all else equal, professionals in ITjobs requiring higher levels of firm-specifichuman capital would receive higher salariesthan those in IT jobs requiring less firm-spe-cific human capital. We found that for IT jobswith lower levels of firm-specific human cap-ital, the erosion of such human capital overtime led to progressively lower compensationincreases with tenure. By contrast, for IT jobsthat require more firm-specific human capi-

tal, we found that compensation increased atan increasing rate with organizational tenureto reflect the value of accumulating on-the-job experience in the firm.

Limitations

We acknowledge that our study has some lim-itations and that it raises several questions forfuture research. First, the salary data for thisstudy were collected at one point in time. Fu-ture research could examine the impact oftenure on compensation using panel datathat tracks the compensation, jobs, and or-ganizational tenure of individuals over time.Our study also focuses on permanent employ-


Independent Variables Control Variables Only With Main Effects With Moderated Effects

Intercept (β0) 3.502** 3.535** 3.607**(0.012) (0.010) (0.013)

Graduate (β1) 0.120** 0.159** 0.167**(0.017) (0.014) (0.013)

Gender (β2) –0.043** –0.014 0.001(0.012) (0.010) (0.010)

Organization Sector (β3) –0.014 –0.001 –0.006(0.012) (0.010) (0.010)

Organization Size (β4) –0.045** –0.013 –0.006(0.013) (0.011) (0.010)

Job Level (β5) 0.283** 0.199** 0.198**(0.017) (0.021) (0.020)

Skills Specificity (β6) 0.076** 0.079**(0.019) (0.019)

Organizational Tenure (β7) 0.216** 0.293**(0.012) (0.014)

Organizational Tenure2 (β8) –0.093**(0.012)

Skills Specificity * 0.031*Organizational Tenure (β9) (0.014)

Skills Specificity * –0.041**Organizational Tenure2 (β10) (0.012)

R2 0.534 0.672 0.693

R2 Change 0.137 0.021

F Change 161.75** 17.98**

Notes: N = 780. ** p <.01, * p < .05, + < .10.

T A B L E V I Hierarchical Regression Results: IT Infrastructure Jobs, Dependent Variable = ln (Compensation)



FIGURE 1. Compensation-Organizational Tenure Extrapolation by Level of Firm Skills Specificity forSystems Development Jobs

FIGURE 2. Compensation-Organizational Tenure Extrapolation by Level of Firm Skills Specificity for ITInfrastructure Jobs


ees. With the increasing prevalence of out-sourcing and the use of contractors (Ho, Ang,& Straub, 2003; Koh, Ang, & Straub, 2004), fu-ture research could compare the longitudinal

unfolding of the compensation ofIT contractors vis-à-vis permanentemployees. Longitudinal analysiswould also be instructive to assessinteractions between the externaland internal labor market overtime in influencing the compensa-tion-tenure relationship in IT.

Second, our results are basedon data collected from Singapore,thereby opening up the possibilitythat the findings are specific to theSingapore context. There are, how-ever, several positive aspects tostudying the IT labor market inSingapore that can mitigate thislimitation. First, prior research hashighlighted that “although muchresearch has been performed onmanaging IT employees, most ofthese studies have been restrictedto examining North America” (Tan& Igbaria, 1994, p. 220). With in-creasing interest among re-searchers to study information sys-tems topics in an internationalcontext, researchers have called onothers to examine the human re-source aspects of IT in differentcountries (e.g., Hunter & Beck,2000; Tan & Igbaria, 1994). Sec-ond, Singapore is a culturally di-verse nation founded as a parlia-mentary republic and operatesbased on the strategic pragmatismof a free market economy (seeSchein, 1996). The latest WorldEconomic Forum 2005–6 ranks thecountry sixth in its Global Com-petitiveness Report. Hence, Singa-

pore operates in a free market economy, andthe laws of demand and supply work in the ITlabor market in Singapore, just as they wouldoperate in the United States and Europe(Lopez-Claros, Porter, & Schwab, 2006). Also,as we have noted earlier, Singapore has ahighly competitive IT labor market, akin to

other nations in other parts of the world. Nev-ertheless, we acknowledge that although theIT labor market context in Singapore may besimilar to other settings, the institutional con-text could be different. This is especially trueregarding the country’s size. Singapore is verysmall and not as geographically dispersed asthe United States or countries in Europe. Thus,studies of compensation in Singapore may notbe constrained by heterogeneity across regionsor regional mobility issues otherwise presentin nations with regional differences. Obvi-ously, our findings must be validated in othersettings, and extension of our ideas to otherinstitutional contexts is essential.

Third, we did not control for job per-formance in our analysis, and job perform-ance has been shown to be a major determi-nant of individuals’ pay. Nevertheless, we donot see this as a major issue, as we collecteddata for all IT professionals in 43 companies,in different jobs. As there is a distribution ofgood and poor performers in each job and ineach company, the potential for introducingbiases by not including a control for job per-formance is mitigated.

Contributions to Research

Notwithstanding the potential limitationshighlighted above, our study makes severalcontributions to research, and has several im-plications for both HR managers and IT pro-fessionals. Our study contributes to humancapital theory by examining not only the re-lationship between compensation andtenure, but also, and more important, themoderating influence of firm skills specificityon the rate of change in compensation as itrelates to tenure. Our results confirm thetenets of classical human theory within theIT field, in that compensation does rise withorganizational tenure, all else being equal.More significantly, our results show that thecompensation change rate could differ basedon the level of firm-specific human capitalembodied in IT jobs. Specifically, the changerate of compensation increases with organi-zational tenure for IT professionals in jobsthat require more firm-specific human capi-tal, reflecting the increasing value of this

More significantly,

our results show that

the compensation

change rate could

differ based on the

level of firm-specific

human capital

embodied in IT jobs.

Specifically, the

change rate of

compensation

increases with

organizational tenure

for IT professionals

in jobs that require

more firm-specific

human capital,

reflecting the

increasing value of

this firm-specific

human capital.



firm-specific human capital. Our resultsimply that IT professionals in jobs requiringmore firm-specific human capital (such as ap-plication development manager) receive pro-portionately greater compensation increaseswith organizational tenure to reflect returnson their investments in a long-term relation-ship with a firm. On the other hand, IT pro-fessionals in jobs requiring less firm-specifichuman capital (such as application program-mer or technical specialist) may need to con-tinually invest in developing marketable andproductive technical competencies to experi-ence compensation increases.

Much of the prior empirical research oncompensation has focused on a single jobcategory, primarily on executives and man-agers (e.g., Gerhart & Milkovich, 1990;Stroth, Brett, Baumann, & Reilly, 1996; Tosi& Gomez-Mejia, 1989). Other research hasexamined compensation across a wide arrayof jobs in many different professions but hasnot been able to examine compensation-tenure profiles within specific professionsdue to insufficient sample size (e.g., Topel,1991). As such, none of the prior studies hasexamined the relationship between compen-sation and organizational tenure within oneoccupation for jobs that require different lev-els of firm-specific human capital. Our studydemonstrates that different jobs within anoccupation could require different compe-tencies, some of which are more general andcan be applied in many firms and others thatare more specific to a particular firm. Thus,to understand whether compensation poli-cies are influenced by the specificity of thehuman capital embodied in different jobs aspredicted by human capital theory, it is nec-essary to measure the firm skills specificity ofthe jobs and to examine how the level offirm skills specificity relates to the compen-sation-tenure profiles for the jobs.

Implications for Practice

Prior research has highlighted the significantrole of compensation in individuals’ deci-sions to enter and remain within an occupa-tion (McLean, 1996). In terms of implica-tions for practice, this study highlights

several key insights for managers and indi-vidual IT professionals.

Our results show that organizations valueand are willing to pay more for jobs requiringhigher levels of firm skills specificity. Consis-tent with the findings of prior research (e.g., D.M. S. Lee et al., 1995; Vitalari,1985), we found that jobs requir-ing higher levels of firm-specificskills tend to be the managerial po-sitions, whereas jobs requiringhigher levels of general skills tendto be the technical IT jobs. The re-wards to staying in an organizationare significantly higher for IT pro-fessionals in managerial positionsthat require more firm-specifichuman capital, compared to ITprofessionals in positions that re-quire more technical generalhuman capital. This suggests thatthere may be a path of upward mo-bility across several jobs within ajob family (e.g., from an applica-tions programmer to a systems an-alyst, or from a project manager toan applications development man-ager). Due to the path of upwardmobility across jobs within a jobfamily, part of the tenure of an ap-plications development managerin a company may be time spent ina position like a systems analyst. Hence, ITprofessionals can start accumulating firm-spe-cific capital even when they are in jobs that re-quire more technical general human capital.However, it is likely that individuals obtain themost opportunities to accumulate firm-spe-cific capital when they are in jobs requiringhigh levels of firm skills specificity.

There are thus several actions that HRmanagers can take based on our results. As afirst step, HR managers could formally verbal-ize to IT employees that building firm-specificcapital should be a professional goal, especiallyif individuals wish to progress to managerialjobs that require such firm-specific skills. It isimportant that HR managers explain specifi-cally how these efforts will be rewarded interms of compensation and advancement. Sec-ond, HR managers may wish to design pro-

As a first step, HR

managers could

formally verbalize to

IT employees that

building firm-specific

capital should be a

professional goal,

especially if

individuals wish to

progress to

managerial jobs that

require such firm-

specific skills.



grams to help IT professionals develop firm-specific knowledge. This may not be throughformal classroom programs, but ratherthrough programs like mentoring schemes,

where newer employees can bementored by employees with moretenure in the firm (Hobday, 2000).Finally, the fact that the pay forjobs requiring higher levels of firm-specific human capital increases atan increasing rate suggests that or-ganizations should be proactive inmanaging turnover of personnel inthose IT jobs so as to retain thefirm-specific human capital of thejob incumbents.

Our study also has implica-tions for HR managers in helpingthem to determine effective com-pensation strategies for IT profes-sionals. As highlighted by prior re-search, variance occurs in the wagepractices adopted for different jobswithin firms (Cappelli & Cascio,1991). Our study suggests that jobsrequiring a higher level of firm skillspecificity typically operate withinthe internal labor market, whereasjobs requiring a higher level ofgeneral human capital typicallyoperate within the external mar-ket. This result implies that man-agers should weight compensablefactors that reflect firm-specificskills higher in job evaluations,thus placing jobs that require morefirm-specific capital higher in thepay structure. When designingcompensation for jobs requiringtechnical general human capitallike applications programmer andtechnical specialists, HR managersmay need to focus more on analyz-ing the prevailing compensationrates in the external labor market.

Our findings also have impli-cations for IT professionals who want to bet-ter manage their career paths. IT researchersgenerally distinguish the career paths of ITprofessionals into two types: technical ormanagerial (Crepeau, Crook, Goslar, & Mc-

Murtrey, 1992; Ginzberg & Baroudi, 1988).Technical career paths occur when IT profes-sionals hold a series of IT technical jobs dur-ing their careers. Managerial career pathsoccur when IT professionals begin their ca-reers in IT technical jobs and then move intoIT supervisory or managerial positions. Ourstudy shows that, controlling for organiza-tional tenure, average compensation ishigher for professionals in IT jobs that re-quire higher levels of firm-specific humancapital, and that these jobs tend to be mana-gerial in orientation. This suggests that indi-viduals who choose to go into such jobs in ITtend to achieve greater career success, interms of average compensation, than indi-viduals who choose to stay in jobs withlower levels of firm-specific human capital.

These results appear to dispel the roman-ticism of a dual-career path for the IT profes-sional. IT professionals who stay in technicalpositions and choose not to progress to man-agerial positions may find that their pay-rategrowth decelerates with increased organiza-tional tenure. On the other hand, their peerswho take the managerial path may find theirpay-rate growth accelerating with organiza-tional tenure. Individuals who wish to stayon a technical career path may do better bykeeping their technical general skills updatedand marketable, and operate in the externallabor market by moving to organizationsthat require and value their technical generalskills. There may be a real threat of obsoles-cence for IT professionals with general ITcompetencies if they remain within a singlefirm and do not upgrade their IT skills andgain exposure to technical jobs in other or-ganizations. The “state of the practice” incomputing is continually changing, andmost practitioners have knowledge of only anarrow set of technologies (Glass, 2000).However, given the reluctance of organiza-tions to invest in general training of their ITstaff, the onus lies on IT professionals to con-stantly upgrade and develop their general ITcompetencies in order to stay relevant to theorganization and to the profession. Thus, ITprofessionals may need to invest in reskillingthemselves in state-of-the-art IT competen-cies in order to sustain their market value.

Given the reluctance

of organizations to

invest in general

training of their IT

staff, the onus lies

on IT professionals

to constantly

upgrade and

develop their

general IT

competencies in

order to stay

relevant to the

organization and to

the profession. Thus,

IT professionals may

need to invest in

reskilling themselves

in state-of-the-art IT

competencies in

order to sustain

their market value.



Note

1. As a robustness check, we estimated our modelsusing generalized least squares (Greene, 2000) to

accommodate the possibility that residuals maynot be independent, but rather correlated acrossindividuals within an organization. The results areconsistent with those from the OLS estimation.


SANDRA A. SLAUGHTER (PhD, Minnesota) is Professor of Information Technology Man-agement and Costley Chair in the College of Management at Georgia Tech University.Before joining the College’s faculty in 2007, she was a member of the information tech-nology faculty in the Tepper School of Business at Carnegie Mellon University, whereshe was an Associate Professor. Prior to her academic career, she worked for Hewlett-Packard, Rockwell International, and Square D Company as an information systems an-alyst and project leader. She currently serves as a senior editor for Information SystemsResearch and Production & Operations Management, as an associate editor for Man-agement Science, and on the editorial board for Organization Science. Her research hasbeen published in Management Science, Information Systems Research, OrganizationScience, MIS Quarterly, Communications of the ACM, IEEE Transactions on Software En-gineering, and other journals. She has received best paper awards at the InternationalConference on Information Systems (ICIS), Academy of Management, Hawaii Interna-tional Conference on System Sciences (HICSS), the Americas Conference on Informa-tion Systems (AMCIS), and the Association for Computing Machinery Special InterestGroup on Management Information Systems (ACM SIGMIS) conferences, and her the-sis won first place in the doctoral dissertation competition held by the ICIS in 1995. Herresearch has been funded by grants from the National Science Foundation and SloanIndustry Center.

SOON ANG (PhD, Minnesota) is the Distinguished Goh Tjoei Kok Chair Professor in Man-agement & IT at the Nanyang Business School, Nanyang Technological University in Sin-gapore. She specializes in IT talent management, outsourcing, cultural intelligence, andglobal leadership. She has published extensively in the Academy of Management Jour-nal, the Journal of Applied Psychology, Management Science, Information Systems Re-search, the Journal of Organizational Behavior, MIS Quarterly, Organization Science, So-cial Forces, and others and coauthored two pioneer books on cultural intelligence(Stanford University Press). She is a winner of numerous academic awards includingBest Paper at the Academy of Management, Best Reviewer for Human Relations, and the2006 Distinguished Leadership Award for International Alumni from the University ofMinnesota. She is senior editor of Information Systems Research, associate editor ofManagement Science and Decision Sciences, and on editorial boards of other manage-ment and IT journals.

WAI FONG BOH (PhD, Carnegie Mellon) is an assistant professor at the Nanyang Busi-ness School, Nanyang Technological University in Singapore. Her research interestsare in the areas of knowledge management, management of human capital, and ITstandards management. Her work has been published or is forthcoming in Manage-ment Science, Organization Science, the Journal of Management Information Sys-tems, Communications of the ACM, and Information & Organization. She won anaward for ICIS Runner-Up to Best Theme-Related Paper Award in 2002. Her disserta-tion won the 2005 Organizational Communication and Information Systems Best Dis-sertation Award and was a Newman Award nominee for the Academy of Managementmeetings.


References

Abraham, K., & Farber, H. S. (1987). Job duration, sen-iority and earnings. American Economic Review,77(3), 278–297.

Aiken, L. S., & West, S. G. (1991). Multiple regression:Testing and interpreting interactions. ThousandOaks, CA: Sage.

Allen, J., & de Grip, A. (2005). Skill obsolescence, life-long learning and labor market participation. Paperpresented at the EALE/SOLE World Conference,San Francisco, CA.

Altonji, J. G., & Shakotko, R. A. (1987). Do wages risewith job seniority? Review of Economic Studies, 54(179), 437–459.

Ang, S., & Slaughter, S. (2000). The missing context ofinformation technology personnel: A review andfuture directions for research. In R. Zmud (Ed.),Framing the domains of IT management: Project-ing the future through the past (pp. 305–328).Cincinnati, OH: Pinnaflex Education Resources.

Ang, S., Slaughter, S., & Ng, K. (2002). Human capitaland institutional determinants of pay for informa-tion technology professionals: Modeling multi-level and cross-level interactions. ManagementScience, 48, 1427–1445.

Ang, S., & Straub, D. (1998). Production and transac-tion economies and IS outsourcing: A study of theU.S. banking industry. MIS Quarterly, 22, 535–552.

Argote, L. (1999). Organizational learning: Creating,retaining and transferring knowledge. Norwell,MA: Kluwer.

Becker, E. (1975). Human capital. Chicago: Universityof Chicago Press.

Becker, E., & Lindsay, C. M. (1994). Sex differences intenure profiles: Effects of shared firm-specific in-vestment. Journal of Labor Economics, 12, 98–117.

Belsely, D. A., Kuh, E., & Welsch, R. E. (1980). Regres-sion diagnostics: Identifying influential data andsources of collinearity. New York: Wiley.

Bingley, P., & Westergaard-Nielsen, N. (2003). Returnsto tenure, firm-specific human capital and workerheterogeneity. International Journal of Manpower,24, 774–788.

Box, G. E. P., & Cox, D. R. (1964). An analysis of trans-formations (with discussion). Journal of the RoyalStatistical Society, 26, 211–252.

Brown, D. (2004). Trouble with assessing IT skills, badnews as demand heats up. Canadian HR Reporter,17, 1–2.

Campion, M. A., & Berger, C. J. (1990). Conceptual in-

tegration and empirical test of job design andcompensation relationships. Personnel Psychol-ogy, 43, 525–554.

Cappelli, P., & Cascio, W. F. (1991). Why some jobscommand wage premiums: A test of career tourna-ment and internal labor-market hypotheses. Acad-emy of Management Journal, 34, 848–868.

Chesebrough, P. H., & Davis, G. B. (1983). Planning acareer path in information systems. Journal ofSystems Management, 34, 6–13.

Chiu, W. C. K., Chan, A. W., Snape, E., & Redman, T.(2001). Age stereotypes and discriminatory atti-tudes towards older workers: An East-West com-parison. Human Relations, 54, 629–661.

Cohen, J., & Cohen, P. (1983). Applied multiple regres-sion/correlation analysis for the behavioral sci-ences (2nd ed.). Hillsdale, NJ: Erlbaum.

Coleman, J. S. (1998). The effect of tenure on earningsfor males: A public/private analysis. Applied Eco-nomics Letters, 5(11), 707–710.

Crepeau, R. G., Crook, C. W., Goslar, M. D., & Mc-Murtrey, M. E. (1992). Career anchors of informa-tion systems personnel. Journal of ManagementInformation Systems, 9(2), 145–160.

Dalton, G. W., & Thompson, P. H. (1971). Acceleratingobsolescence of older engineers. Harvard BusinessReview, 49(5), 57–67.

Farber, H. (1999). Mobility and stability: The dynamicsof job change in labor markets. In O. Ashenfelter &D. Card (Eds.), Handbook of labor economics (Vol.3, pp. 2439–2483). Amsterdam: North-Holland.

Farmer, J., & Campbell, F. (1997). Information profes-sionals, CPD and transferable skills. Library Man-agement, 18(3), 129–134.

Feeny, D. F., & Wilcocks, L. P. (1998). Core IS capabili-ties for exploiting information technology. SloanManagement Review, 39(3), 9–21.

Figart, D. M. (2001). Wage-setting under Fordism: Therise of job evaluation and the ideology of equalpay. Review of Political Economy, 13, 405–425.

Fossum, J. A., & Arvey, R. D. (1986). Modeling theskills obsolescence process: A psychological/eco-nomic integration. Academy of Management Re-view, 11, 362–374.

Gerhart, B., & Milkovich, G. (1990). Organizational dif-ferences in managerial compensation and financialperformance. Academy of Management Journal,33, 663–691.

Ginzberg, M. J., & Baroudi, J. J. (1988). MIS careers—A theoretical perspective. Communications of theACM, 31, 586–594.



Glass, R. (2000). On personal technical obsolescence.Communications of the ACM, 43(7), 15–17.

Goldsmith, A. H., & Veum, J. R. (2002). Wages and thecomposition of experience. Southern EconomicJournal, 69(2), 429–443.

Greene, W. H. (2000). Econometric analysis. UpperSaddle River, NJ: Prentice Hall.

Hecker, D. E. (2005). Occupational employment pro-jections to 2014. Monthly Labor Review, 128(11),70–101.

Ho, V., Ang, S., & Straub, D. W. (2003). When subordi-nates become IT contractors. Information SystemsResearch, 14(1), 66–86.

Hobday, M. (2000). The project-based organization: Anideal form for managing complex products andsystems? Research Policy, 29, 871–893.

Hunter, M. G., & Beck, J. E. (2000). Using repertorygrids to conduct cross-cultural information sys-tems research. Information Systems Research,11(1), 93–101.

Information Technology Association of the Americas(ITAA). (2002). Bouncing back: Jobs, skills and thecontinuing demand for IT workers. Arlington, VA:Author.

Information Technology Association of the Americas(ITAA). (2004). Adding value . . . growing careers,the employment outlook in today’s increasinglycompetitive IT job market. Arlington, VA: Author.

Ives, B., & Olson, M. H. (1981). Manager or techni-cian? The nature of information systems manager’sjob. MIS Quarterly, 5(4), 49–63.

Jacobsen, J. P., & Levin, L. M. (2002). Calculation ofreturns to job tenure revisited. Applied EconomicsLetters, 9, 473–477.

James, L. R., Demaree, R. G., & Wolf, G. (1984). Esti-mating within-group interrater reliability with andwithout response bias. Journal of Applied Psychol-ogy, 69, 85–98.

James, L. R., Demaree, R. G., & Wolf, G. (1993). Rwg:An assessment of within group interrater agree-ment. Journal of Applied Psychology, 78,306–309.

Joseph, D., & Ang, S. (2002). Threat of professionalobsolescence and how IT professionals cope withit. Paper presented at the Thirty-Fifth Hawaii Inter-national Conference On System Sciences (HICSS-35), Hawaii.

Kim, J., & Lerch, J. (1997). Why is programming(sometimes) so difficult? Programming as scientificdiscovery in multiple problem spaces. InformationSystems Research, 8(1), 25–50.

Koh, C., Ang, S., & Straub, D. W. (2004). IT outsourc-ing success: Psychological contract perspective. In-formation Systems Research, 15(4), 356–373.

Larwood, L., Rodkin, S., & Judson, D. (2001). Retrain-ing and the technological productivity paradox. In-ternational Journal of Organizational Theory andBehavior, 3, 201–224.

Lee, C. K. (2005). Transferability of skills over the ITcareer path. Paper presented at the SIGMIS Com-puter Personnel Research Conference, Atlanta, GA.

Lee, D. M. S., Trauth, E. M., & Farwell, D. (1995). Criti-cal skills and knowledge requirements of IS profes-sionals: A joint academic/industry investigation.MIS Quarterly, 19, 311–340.

Levinson, N. S. (1988). Repositioning the informationsystems management function: Implications for in-formation systems personnel. Paper presented atthe SIGMIS Computer Personnel Research Confer-ence, College Park, MD.

Lopez-Claros, A., Porter, M. E., & Schwab, K. (2006).Global competitiveness report 2005–2006. NewYork: Palgrave Macmillan.

McLean, E. R. (1996). The importance of salary on joband career attitudes of information systems pro-fessionals. Information & Management, 30,291–299.

Mincer, J. (1970). The distribution of labor incomes: Asurvey. Journal of Economic Literature, 8(1), 1–26.

Mincer, J. (1974). Schooling, experience and earnings.New York: Columbia University Press.

Munasinghe, L., & O’Flaherty, B. (2005). Specific train-ing sometimes cuts wages and always cutsturnover. Journal of Labor Economics, 23(2),213–233.

Nakayama, M., & Sutcliffe, N. G. (2001). IT skills port-folio research in SIGCPR proceedings: Analysis,Synthesis and Proposals. Paper presented at theSIGMIS Computer Personnel Research Conference,San Diego, CA.

Parent, D. (2002). Matching, human capital, and thecovariance structure of earnings. Labor Economics,9, 375–404.

Pazy, A. (1996). Concept and career-stage differentia-tion in obsolescence research. Journal of Organiza-tional Behavior, 17, 59–78.

Prendergast, C. (1993). The role of promotion in induc-ing specific human capital acquisition. QuarterlyJournal of Economics, 108, 523–534.

Schein, E. H. (1996). Career anchors revisited: Implica-tions for career development in the 21st century.Academy of Management Executive, 10, 80–88.



Schenk, K., Vitalari, N., & Davis, K. (1998). Differencesbetween novice and expert systems analysts: Whatdo we know and what do we do? Journal of Man-agement Information Systems, 15(1), 9–50.

Schoones, D., & Bernhardt, D. (1998). Promotion,turnover, and discretionary human capital acquisi-tion. Journal of Labor Economics, 16(1), 122–141.

Shearer, R. L., & Steger, J. A. (1975). Manpower obso-lescence: A new definition and empirical investiga-tion of personal variables. Academy of Manage-ment Journal, 18, 263–275.

Stinchcombe, A. L., & Heimer, C. A. (1988). Inter-orga-nizational relations and careers in computer soft-ware firms. In S. B. Bacharach (Ed.), Research inthe sociology of work (Vol. 4, pp. 179–204). Green-wich, CT: JAI Press.

Stroth, L. K., Brett, J. M., Baumann, J. P., & Reilly, A.H. (1996). Agency theory and variable pay compen-sation strategies. Academy of Management Jour-nal, 39, 751–767.

Tan, M., & Igbaria, M. (1994). Turnover and remunera-tion of information technology professionals in

Singapore. Information & Management, 26(4),219–229.

Taylor, P., & Walker, A. (1998). Employers and olderworkers: Attitudes and employment practices. Age-ing & Society, 18, 641–658.

Todd, P. A., McKeen, J. D., & Gallupe, R. B. (1995). Theevolution of IS job skills: A content analysis of ISjob advertisements from 1970 to 1990. MIS Quar-terly, 19(1), 1–27.

Topel, R. (1991). Specific capital, mobility, and wages:Wages rise with job seniority. Journal of PoliticalEconomy, 99(1), 145–176.

Tosi, H. L., & Gomez-Mejia, L. R. (1989). The decou-pling of CEO pay and performance: An agency the-ory perspective. Administrative Science Quarterly,34, 169–189.

Vitalari, N. (1985). Knowledge as a basis for expertisein systems analysis: An empirical study. MIS Quar-terly, 9, 221–241.

White, H. (1980). A heteroscedasticity-consistent co-variance matrix estimator and a direct test for het-eroscedasticity. Econometrica, 48, 817–838.


firm-specific human capital and compensation organizational...

Documents