Human Resource Management Review 24 (2014) 41–52

Contents lists available at ScienceDirect

Human Resource Management Review

j ourna l homepage: www.e lsev ie r .com/ locate /humres

Pay for (individual) performance: Issues, claims, evidence andthe role of sorting effects

Barry Gerhart⁎,a, Meiyu Fang b

a University of Wisconsin-Madison, 4194 Grainger Hall, Madison, WI 53706, United Statesb National Central University, Taiwan

a r t i c l e i n f o

⁎ Corresponding author. Tel.: +1 608 262 3895.E-mail address: bgerhart@bus.wisc.edu (B. Gerhar

1 More complete reviews, especially of earlierwork, aspay structure, benefits) are available elsewhere (GerharMcKenzie & Lee, 1998; Prendergast, 1999; Rynes et al., 2

1053-4822/$ – see front matter © 2013 Elsevier Inc. Ahttp://dx.doi.org/10.1016/j.hrmr.2013.08.010

a b s t r a c t

We address two broad questions: how much pay for individual performance (PFIP) is thereand what are the positive and negative effects of PFIP? We consider specific claims, includingthat PFIP does not motivate (or even de-motivates), that it is ineffective in teams, and that it isineffective in some national cultures. We demonstrate how incorporating sorting effects ofPFIP into conceptual treatments of PFIP can change how one views the likely effectiveness ofPFIP across contexts.

© 2013 Elsevier Inc. All rights reserved.

Keywords:CompensationMotivationPay for performancePerformanceIncentivesSorting

1. Introduction

Pay or employee compensation is a topic that generates considerable interest among scholars and practitioners. Over the years, theyhave asked questions like how important is money in motivation, to what degree do organizations pay for performance (especiallyindividual performance), does paying for performance contribute to higher performance or does paying for performance (again,especially individual performance) sometimes have undesired (and often unanticipated) consequences? The present paper will focuson pay for individual performance (PFIP) and will offer a selective survey that builds on classic work, but primarily highlights recentconceptual and empirical developments.1 In particular, the present paper will demonstrate how taking fuller account of sorting effects,the effect that PFIP has “on the composition (e.g., in terms of ability, personality) of an organization'sworkforce via attraction, selection,and attrition processes” (Gerhart & Rynes, 2003, p. 151), can change thinking regarding PFIP's effects in several arenas.

To begin, we classify pay for performance programs on three dimensions (Gerhart & Rynes, 2003): the emphasis on results-oriented (e.g., physical production output, sales, profits, total shareholder return) or behavior-oriented performance (e.g., supervisoror customer ratings of specific behaviors)measures, the emphasis on performancemeasured at the individual employee or aggregate(e.g., group, unit, organization) level, and the degree of incentive intensity. Virtually all private sector organizations in the UnitedStates (and in many other countries) use multiple pay for performance plans with the design (see the preceding three dimensions)varying as a function of factors such as job level, type of occupation, the way the work is organized, and the organization's strategy(Gerhart, 2000; Gomez-Mejia & Balkin, 1992; Gomez-Mejia, Berrone, & Franco-Santos, 2010; Milkovich, Newman, & Gerhart, 2013).For example, results-basedmeasures aremore likely to be used for sales occupations, executives, and other positionswhere objectiveperformance measures are available. Incentive intensity is typically stronger in higher level jobs and/or in jobs where results-basedperformancemeasures are available.Merit pay (a focus on subjective judgments of performance behaviors) is found in some form formost jobs, though it is less likely to be used for lower level positions, especially when a labor union is present. However, even in thiscontext, one is increasingly likely to find plant and/or firm level performance-based plans (e.g., gain sharing, profit sharing).

t).well as ofwork onother forms of pay for performance and on other dimensions of compensation (e.g., pay level,t & Milkovich, 1992; Gerhart & Rynes, 2003; Gerhart et al., 2009; Gibbons, 1998; Lawler, 1971; Lazear, 1995;005).

ll rights reserved.

42 B. Gerhart, M. Fang / Human Resource Management Review 24 (2014) 41–52

As noted, our primary focus here is on PFIP plans. Theories of compensation (e.g., reinforcement, expectancy, efficiency wage,agency) largely agree that incentives and reinforcement (central to any PFIP plan) are key drivers of important workplace behaviorssuch as employee performance and employee attraction/retention. Indeed, there is overwhelming (meta-analytic) evidence thatincentives drive behavior and that the effect is substantial (Locke, Feren, McCaleb, Shaw, & Denny, 1980; Gupta & Shaw, 1998; Guzzo,Jette, & Katzell, 1985; Jenkins, Mitra, Gupta & Shaw, 1998; Judiesch, 1994; Stajkovic & Luthans, 1997). (Gerhart & Rynes, 2003 andGerhart et al., 2009 provide reviews of the broader literature on pay for performance consequences.) In addition, PFIP decisionsmay beespecially strategic. For example, Locke et al. reported a 30% improvement in productivity due to the introduction of incentive plans.Gerhart and Milkovich (1990) found that organizations differentiate more with regard to how they pay (e.g., how much they rely onPFIP), somewhat less on how much they pay, and that organizations placing more emphasis on PFIP performed better on average.

So, given the evidence that PFIP has positive effects, what is there to write about? Over the years, challenges to theconceptual underpinnings of PFIP have been raised, as have important and real-world challenges in successfully implementingPFIP. For example, work by Herzberg directly questioned whether money was an important motivator in the workplace,relegating money to the “hygiene” category of motivation-hygiene theory. Deci (1975) and Deci & Ryan (1985) went a stepfurther than simply arguing that money did not motivate. They argued that money could demotivate. Specifically, their modelmakes the case that under common circumstances, PFIP undermines intrinsic motivation and interest to perform a task or activity.This hypothesized undermining effect of PFIP, which is based on research conducted in non-work settings (often using childrenas subjects in an education setting), was subsequently picked up by others (e.g., Kohn, 1993; Pfeffer, 1998) and extrapolated toworkplace settings. Most recently, this undermining effect of PFIP has been popularized in a bestselling book, Drive (2009), by DanielPink (Pink was also the keynote speaker in 2011 at the World at Work, formerly known as the American Compensation Association,annual conference, thus having the maximum opportunity to communicate his message to compensation professionals).

Other concerns have to with PFIP's fit with different situational contexts. As one example, does PFIP fit the “new economy”where theuse of teams is thought to be greater,with greater importance placed on collaboration and cooperation (as opposed tonarrower individualperformance goals)? As another example, even if PFIP is seen as continuing to be a central part of managing people in a country like theUnited States, does PFIP fit other countries that differ on dimensions such as national culture? For example, China is now the world'ssecond largest economy (after the United States) and has a very different national culture profile (e.g., less individualistic; Hofstede, 1980,2001; House, Hanges, Javidan, Dorfman, & Gupta, 2004) than the United States. Does that mean PFIP will not work well in China?

Two final concerns stand in direct opposition to one another. One concern is that although organizations say they use PFIP,when one looks at the actual pay-performance relationship, it sometimes does not appear to be strong enough to be meaningful.In other words, the incentive intensity of PFIP in this case is too weak to be of any consequence. A second, opposite concern, is thatPFIP can be implemented with such strong incentive intensity (often using results-based performance measures) that the risk ofserious unintended negative consequences becomes greater (e.g., top executives manipulating profits and/or taking too littleor alternatively, excessive risks, Martin, Gomez-Mejia, & Wiseman, 2013; Sanders & Hambrick, 2007; employees focusing onquantity rather than quality or customer satisfaction and/or gaming the incentive system, rate cutting and goldbricking, Gerhart,2001; Roy, 1952; Whyte, 1955; employees paying attention only to performance objectives explicitly covered in the incentiveplan, Lawler, 1971; Milgrom & Roberts, 1992; Wright, George, Farnsworth, & McMahan, 1993). In this situation, agency theory(Berle & Means, 1932; Fama & Jensen, 1983; Holmstrom, 1979; Prendergast, 1999), with its focus on agency costs (adverseselection andmoral hazard) that arise from information asymmetry and goal incongruence is a standard tool for understandingwhatcan go wrong and also helps make apparent that when PFIP is designed to have strong incentive intensity, PFIP can be described as ahigh risk, high return strategy (Gerhart, 2004; Gerhart, Trevor, & Graham, 1996). Agency theory is also useful because of itsrecognition that while higher incentive intensity increases the magnitude of the incentive effect, it also increases the potential foragency costs and thus, unintended negative consequences. Higher incentive intensity also increases the amount of risk to be borne byemployees, which can easily translate into employee relations problems (especially when the incentive plan does not pay out).

Another concern is thatwhen onedigs into the evidence on PFIP inwork settings, there is less there than onemight have expected.For example, themeta-analytic evidence cited earlier, which shows that PFIP has a strong positive impact on performance pertains toa narrow set of contexts. Specifically, the evidence comes primarily from studies that use results-based (objective) measures ofperformance and from settingswherework is simple, easy tomeasure, and not interdependent.2 Gerhart andMilkovich (1992) notedthat these types of PFIP plans “are not applicable tomany jobs” (p. 523) and that there are often “administrative problems”with suchplans (p. 523), which include the cost of setting andmaintaining setting standards, both in financial terms and in employee relationsterms (e.g., due to gaming, rate cutting, and so forth). As such, theremay be a selection biaswhenwe study only the plans that survive,at least to the degree that there are many failed plans that do not make it into our sample (Gerhart et al., 1996).

Merit pay plans, with their broad applicability (i.e., to any job where a performance rating can be obtained) and behavior-orientedfocus (which allows one to consider whether results are obtained in a way judged to be in keeping with the organization's values andgoals) would seem to provide the solution to the potential limitations of results-based PFIP plans. However, merit pay plans have theirown challenges, both in terms of execution and in terms of how to design research to study their effects (Milkovich & Wigdor, 1991).Unfortunately, there is little rigorous research on the direct impact of merit pay plans because of challenges in doing such research(Heneman &Werner, 2000; Rynes, Gerhart, & Parks, 2005). On the execution side, as noted, there is often doubt as towhethermerit pay

2 Consider, for example, the meta-analysis of financial incentives conducted by Jenkins, Mitra, Gupta, and Shaw (1998). Of the 47 studies covering 3124employees they were able to find eight studies that were conducted in a field setting, covering 470 employees. The performance measures in the eight studieswere: number of trees planted (2), number of animals trapped (2), supervisory/behavioral ratings (2), exam completion time (1), number of items tested in amanufacturing setting (1). Exam completion time involved transforming a number by adding digits to it and looking it up in a catalog.

43B. Gerhart, M. Fang / Human Resource Management Review 24 (2014) 41–52

really exists in organizations. We return to this issue below. There is also often an execution challenge due to the sometimes suspectcredibility of the performance ratings, as they typically have lower than desired interrater reliability. (Meta-analytic evidence reports amean interrater reliability of 0.52, Viswesvaran, Ones, & Schmidt, 1996).

Finally, the interest in PFIP now goes well beyond the private sector. In the United States, PFIP is being considered (andincreasingly implemented) in health care and in the public sector (e.g., teachers). Theory and evidence on PFIP are usually closelyscrutinized in such situations,making it thatmuchmore important to clearly describe how PFIPworks and its potential pros and cons.

Given the issues identified above, we organize the present paper around three questions. First, to what degree do organizationsactually use and execute PFIP? Second, what are the hypothesized positive effects of PFIP (in cases where PFIP is actually executed)and what does the empirical evidence show? Third, what are the hypothesized potentially harmful effects of PFIP that have beenidentified and what is the evidence regarding these concerns? An important theme in addressing these questions will be how logicand conclusions regarding PFIP are challenged and changed when the sorting effects of PFIP are explicitly taken into account.

2. How much PFIP is there?

Although there is some skepticism, a good deal of evidence indicates that pay and individual performance are, in fact, stronglyrelated in organizations, especially among exempt employees. So, why then is there skepticism about the importance of PFIP? Amajor reason is that some have inferred based on various surveys of companies that PFIP is not widely used (e.g., Larkin, Pierce, &Gino, 2012). Another reason is that surveys of companies show that even when one particular PFIP program, merit pay, is used, itsincentive intensity (i.e., the difference in merit increases received by high and average performers) appears so weak as to raisequestions about its impact on motivation and behavior. We address both types of evidence below. We also show that definingPFIP only in terms of annual merit increases is too narrow and misses a major part of the pay–performance relationship.

To begin, it is important to recall, as noted earlier, that the use and incentive intensity of PFIP depends partly on contextualfactors such as the job. For example, as we will see, the majority of top executive compensation is in the form of PFIP, with theprimary component (long-term incentive/stock plans) being linked to total shareholder return and another major component(bonus) being (typically) linked to profit-related measures. On the other hand, hourly union employees, which now representless than 7% of the private sector workforce in the United States, are less likely to be covered by PFIP.

In looking at the economy overall and across employee groups, PFIP (contrary to some claims, e.g., Larkin et al., 2012) is pervasive.3

For example, a WorldatWork survey of 1001 of member companies reported that 92% of them used individual performance todetermine base salary increases (WorldatWork, 2012). In addition, 84% of companies in the same survey reported using variable pay(other than sales commissions) such as merit bonuses, individual incentives, performance sharing, and profit sharing. As such, thetypical company uses both merit pay (behavior-based) and variable pay (results-based) forms of PFIP programs.

What about the incentive intensity of PFIP when it is used? We begin with the traditional annual merit increase program.Towers Watson (2012) reports that for 2012, the highest performing (exempt) employee received a 4.7% increase, versus 3.2% foran average performer. Aon Hewitt reports that a high performer receives a 3.5% increase, versus 2.4% for an average performer(Abosch, 2012). Thus, the differential for high versus average performance is 1.1 to 1.5 percentage points. For an employeeearning $100,000 per year that translates into a difference of $1100 to $1500 per year, or about $21 to $29 per week, before taxes.Factoring in taxes (federal, state, sometimes city, social security), the marginal rate can easily hit 40%, meaning that the weeklypayoff to high performance falls to $13 to $18 per week. For an employee earning $50,000 per year, the difference is even smaller.Is such an annual merit increase difference enough of a payoff to motivate high performance? Perhaps not. (Of course, not allcompanies are average. In some, the payoff will be higher. In others, lower.)

However, the problem with the above scenario is that it greatly understates the career earnings payoff to high performance(and thus the potential motivational impact of pay for performance, if communicated and/or understood sufficiently) because itignores compounding effects and defines PFIP too narrowly (over the course of a career) for the following reasons. (See Exhibit 1.)

First, consistently high performers receive consistently higher annual merit increases over time, resulting in the cumulativerelationship between pay and performance being larger than observed in any single year, due both to the increased reliability ofcumulative performance ratings and to the compound interest effect (Gerhart & Milkovich, 1992).

Second, organizations increasingly use merit bonuses. Indeed, merit bonuses now appear to account for more of the pay–performance relationship than do the traditional and most often discussed form of PFIP, merit pay. Merit bonuses (unlike meritpay) do not become part of base salary, meaning that only consistently high performerwill receive significant bonuses during theircareers. Again, the cumulative effect should be that consistently high performers have higher career earnings, again strengtheningthe pay–performance relationship. Moreover, there is some evidence to suggest that recentmerit bonus pools, in companieswherethey are used, are substantially larger than in the past at 12% of payroll for exempt employees in 2012, up from about 3% in 1993

3 Larkin et al. state (p. 1196, footnote 2) that “pay-for-performance includes pay based on subjective measures of performance as well as objective ones.”However, they then rely (in part) on a survey of the Fortune 1000 to argue that “use of individual performance-based pay for nonexecutives” is infrequent. But,that survey (Lawler, 2003) did not ask about merit pay, which uses subjective measures of performance. It is also unclear whether the survey asked specificallyabout merit bonuses. In any case, even if it did, we know (Abosch, 2012) that the use of merit bonuses has increased substantially over time, making the surveyresults from 2003, which are a decade old, perhaps of limited relevance to what companies are doing today in the area of PFIP. It is accurate to say that individualincentive plans (which use physical output performance measures at the individual level) are infrequent (Lazear, 2000). But, as noted the use of merit pay andmerit bonuses is much higher, as is the use of other variable pay plans.

4 The data provided by Aon Hewitt (Abosch, 2012) pertain to “variable pay” plans, which are not limited to merit bonus plans. However, it is generally the casethat the higher the job level, the more likely it is that variable pay will be based primarily on individual performance.

5 Estimating the role of merit bonuses relative to merit increases economy-wide is something of a challenge. The U.S. Bureau of Labor Statistics does not providedetailed data on bonus versus salary payments. WorldatWork (2012) surveyed 2150 U.S. companies with annual revenues ranging from under $30 million toover $10 billion (mode = $1 billion to $3 billion). Eighty-two percent reported using “performance-based, lump-sum, short-term cash awards during the year(p. 38). For this 82% of companies, the following 2011 data on merit bonuses were reported by employee group (and we have added some computations):

Nonexempt hourly nonunion Nonexempt salaried Exempt salaried Officers/executives

Average percentage of employees eligible incompanies that use merit bonuses (A)

90% 94% 82% 95%

Average percentage of eligible employees receiving incompanies that use merit bonuses (B)

84% 88% 81% 92%

Average percentage of all employees receiving incompanies that use merit bonuses (A ∗ B = C)

76% 83% 66% 87%

Average percentage of employees receiving in allsurvey companies (82% ∗ C = D)

62% 68% 54% 72%

Average percentage of salary received as merit bonus incompanies using merit bonuses (E)

5.4% 6.0% 12.6% 37.0%

Average percentage of salary received as merit bonus across allsurvey companies (D ∗ E = F)

3.3% 4.1% 6.9% 26.5%

Therefore, we estimate that the average percentage of salary received in the form of a merit bonus is 3% to 7% for most employees (and substantially higher for officersexecutives). Thus,merit bonuses are as large as the traditionalmerit increase for nonexempt hourly nonunion employees and larger than the traditionalmerit increase for alother employee groups.Note that these data are not inconsistentwith thedata reportedbyAbosch (2012) ashis data (merit bonuses representing13%of salary in companiethat use them) pertain to exempt employees only, which would correspond to the two right-most columns in row E in the Table above. The above data exclude unionmembers, but note that only 6.9% of private sector employees were union members as of 2011, according to U.S. Bureau of Labor Statistics data.

6 Bebchuk et al. (Table 1) report that the CEO receives 35.7% of the total compensation paid to the top five executives, on average, in U.S. publicly traded firmincluded in the Compustat database. Thus, the other four executives receive (1 − .357) / 4 = 16.1%, on average. The ratio of 35.7%/16.1% = 2.2.

Annual Merit Increases

But Also:

Annual Merit Bonuses and/or other Variable Pay

Promotion

Immediate Pay Increase from Moving to Higher Pay Grade

Subsequent (Larger) Pay Increases Because of Lower Compa-ratio in New Grade

Greater Opportunities to Move to Alternative Employers to Receive Higher Pay

Compounding Effects of Higher Annual Pay Growth

Greater Likelihood of Keeping Job (and thus income)

Exhibit 1. How pay and performance are related over a career.

44 B. Gerhart, M. Fang / Human Resource Management Review 24 (2014) 41–52

(Abosch, 2012; original data from Hewitt/Aon Hewitt annual surveys).4 Even if the merit bonus for high versus low performersratio is no greater than reported above for merit increases (about 1.5), a bonus pool of 12% would thus imply a merit bonus ofroughly 18% for high performers and 12% for average performers, or about a 6 percentage point differential. We estimate thateconomy-wide (across companies and different employee groups), merit bonuses are somewhere between 3% and 7% of salaryeach year (and much higher for executives).5 That means that looking only at merit salary increases (which have recently beenaround 3% of salary) misses at least one-half of the amount of pay that is tied to performance in each year.

Third, PFIP is not only defined too narrowly by excluding bonuses and compounding effects, but also other important effects arealso usually overlooked. Merit/performance ratings influence not only annual salary increases (and annual bonuses), but alsopromotions (Gerhart &Milkovich, 1989). Promotion can be seen as one example of a broader and, until recently,mostly overlookedphenomenon, sorting, which will be a recurring theme in our paper. The average pay increase due to promotion (over 8%,WorldatWork, 2010) is larger than the typical within-grade merit increase (roughly 3% in recent years in the United States).Promotions are also more likely early in one's career, meaning that promotion-based pay increases will be compounded for manyyears. In addition, after being promoted, one typically begins relatively low in the new grade, which means (given the typicaldesign of a merit increase grid, which takes into account not only performance, but also compa-ratio) that high performance willresult in relatively large annual merit increases. As one moves higher, the promotion payoff to high merit ratings is larger, asdemonstrated, at the extreme for example, by the magnitude of the compensation differential (2.2×) between a chief executiveofficer and the average compensation of the next four highest paid executives (Bebchuck, Cremers, & Peyer, 2011).6

”

/ls

s

45B. Gerhart, M. Fang / Human Resource Management Review 24 (2014) 41–52

Fourth, another important type of sorting effect, in addition to promotion, relevant here is that people can change employersto increase their pay and doing so results in substantial pay increases (of 20 to 25% among those with graduate degrees, Dreher &Cox, 2000; Gomez-Mejia & Balkin, 1992; Milkovich et al., 2013). We also know that high performers are more likely to leave inresponse to weak PFIP (e.g., Trevor, Gerhart, & Boudreau, 1997), especially to the degree that there is strong pay communication(Shaw & Gupta, 2007). In summary, it seems likely that once this sorting process (based on individual performance) is taken intoaccount, the PFIP relationship will be stronger than often thought.

Finally, employees whose performance meets a minimum standard receive another payoff: they get to keep their jobs (and thecompensation that goes with holding their jobs). In the past (and perhaps still), GE has been known for asking about 10% of itsworkforce to leave each year. The SAS Institute, which is regularly at or near the top of the Fortune 100 Best Places to Work alsohas a significant (performance-based) involuntary turnover rate (4.2% as of several years ago, about twice as high as theirvoluntary turnover rate, Hansen, 2000).

2.1. PFIP and executive pay

Given the vast amount of work that examines executive pay, as well as the amount of attention and controversy generatedregarding whether there is PFIP in executive pay, it is important to say a few words regarding this specific employee group. Aclassic study by Jensen and Murphy (1990) reported a $3.25 gain in compensation for chief executive officers (CEOs) for each$1000 gain in shareholder wealth. In interpreting this finding, Jensen and Murphy famously concluded that “although theestimated pay–performance sensitivity… is statistically significant, the magnitude seems small in terms of the implied incentives”(1990:243), and went on to say, “We believe our results are inconsistent with the implications of formal agencymodels of optimalcontracting” (1990:227). The management literature picked up on this conclusion and meta-analytic evidence in the managementliterature likewise generally portrays a weak CEO pay–performance relationship (Tosi, Werner, Katz, & Gomez-Mejia, 2000; Dalton,Daily, Certo, & Roengpitya, 2003). For example, Hitt (2005), based on such research stated that “Research shows only a smallrelationship between executive compensation in any form and firm performance.”

However, there are limitations with the approaches used to estimate the CEO pay–performance relationship (Nyberg, Fulmer,Gerhart, & Carpenter, 2010).7 Questions have also been raised with Jensen and Murphy's interpretation of their own results asshowing a “small” CEO pay–performance relationship (e.g., Garen, 1994; Gerhart & Rynes, 2003; Hall & Liebman, 1998). Forexample, in 2012, the median market value of a company on the Fortune 100 list was roughly $50 billion. Thus, a 10% increase inmarket value would be $5 billion and the increase in executive pay would be $5 billion × $3.25 / $1000 = $16.25 million. Theaverage pay of a CEO on the list of 100 highest paid CEOs in 2012 according to Forbes was roughly $22 million. Thus, a 10%increase in market value would be associated with an increase of roughly $16.25 million / $22 million = 74%. In either dollarterms or percentage terms, the magnitude of the CEO pay–performance relationship is substantial. Further, subsequent empiricalstudies actually find a larger relationship (than found by Jensen & Murphy) between CEO pay and performance (Aggrawal &Samwick, 1999; Hall & Liebman, 1998; Nyberg et al., 2010).

3. What are the positive effects of PFIP? Incentive effects and sorting effects

As noted earlier, many theories (e.g., reinforcement, expectancy, equity, utility, agency, efficiency wage, attraction–selection–attrition, sorting, tournament) are relevant to examining and understanding the effects of pay in organizations. To greatlysimplify, pay influences performance via two different mechanisms (see Exhibit 2)—incentive effects and sorting effects (Gerhart &Milkovich, 1992; Gerhart & Rynes, 2003; Lazear, 1986). First, the incentive effect refers to the impact of PFP on current employees.For example, if an organization changes its pay strategy, does it find that employees present at the time of the change becomemore or less productive? As shown above, meta-analytic evidence on results-based performance measures (with the limitationsdiscussed above) strongly supports the positive incentive effect of PFIP on performance. In addition, a meta-analysis by Combs etal. (2006) of “high performance work” practices reported a corrected r = .15 for the relationship between the use of “incentivecompensation” (based on 31 studies and 8156 employees) and organizational performance.

Second, PFIP has another effect on performance that is also substantial, but until recently, has been largely ignored. The sortingeffect (Gerhart & Milkovich, 1992; Gerhart & Rynes, 2003; Lazear, 1986; 1999; Rynes, 1987) refers to a change in pay strategy thatimproves performance not by changing the behavior of current employees, but rather by changing who the current employees are(and what attributes they bring to the organization). Organizations have always, of course, known about the sorting effect of paylevel decisions. Less obvious to organizations (and perhaps scholars), however, is the sorting effect of PFIP.

The best illustration of the operation of incentive and sorting effects comes from a study by Lazear (2000). He observed a 44%increase in productivity when an automobile glass installation company switched from salaries to individual incentives. Of thisincrease, roughly one-half was due to workers present before and after the move to the new pay strategy increasing their

7 As one example, the meta-analyses just cited rely almost entirely on cash compensation (base salary + bonus) to measure executive pay. However, in mostyears, long-term incentive, or stock-related plans (e.g., stock options, stock grants) represent a larger component of total executive compensation. For example,based on Forbes executive compensation data for the 500 largest U.S. companies, over the 2000 to 2012 period, we found that the average “stock” compensationper was $5.6 million/year versus $3.3 million/year for base + bonus. Further, it is stock plans, not base salary + bonus plans, that are designed to align closelywith total shareholder return. (Base salary is probably more related to size. Bonus plans are typically designed to pay out based on accounting-relatedperformance measures, usually quarterly or annual.) Thus, excluding stock compensation from a study of executive pay for performance greatly reduces themagnitude of the observed relationship between executive pay and performance.

Incentive Effect(Impact on Current Employees)

Pay for Individual Performance (PFIP) Performance

Sorting Effect(Impact on Who Employees Are andtheir Performance-related Attributes)

Exhibit 2. Incentive and sorting effects.

46 B. Gerhart, M. Fang / Human Resource Management Review 24 (2014) 41–52

productivity (an incentive effect), while the other one-half of the 44% increase could not be explained by employees presentbefore and after the incentive plan implementation increasing their productivity. Instead, what Lazear found was that theremaining one-half of the effect was due to less productive workers quitting in response to the new incentive systems (with itsstronger incentive intensity) and being replaced by more productive workers over time (a sorting effect). His findings areconsistent with attraction–selection–attrition theory (Schneider, 1987), as well as other empirical evidence showing that paysystems that vary in their emphasis on individual performance likewise contribute to variations in the characteristics (includingperformance) of those they attract (Cable & Judge, 1994; Cadsby, Song, & Tapon, 2007; Trank, Rynes, & Bretz, 2002) and retain(Harrison, Virick, & William, 1996; Nyberg, 2010; Salamin & Hom, 2005; Shaw, Dineen, Fang, & Vellella, 2009; Trevor et al., 1997).Overall, these studies indicate that high individual performers are more likely to join and stay with organizations that morestrongly emphasize PFIP.

4. What are the negative effects and limitations of PFIP?

A number of limitations and negative effects of pay for individual performance have been claimed over the years, including:(1) pay is not a motivator, (2) somewhat related, PFIP harms intrinsic interest/motivation in work, (3) PFIP (and the associatedwithin-grade dispersion in pay that results as PFIP is implemented more strongly) harms cooperation and teamwork where workis interdependent, (4) PFIP does not work/fit with many national cultures, (5) performance measures are often not reliable andvalid enough to use PFIP, and (6) PFIP can work “too well” in that the drive to achieve objectives (explicitly included in the PFIPplan) may be so relentless as to cause other important objective to be ignored and/or included objectives to be achieved throughgaming/undesirable methods and/or excessive risk-taking.

Not all of these claims can be covered here. So, we will focus on the claims (1 - 4) that are, in our opinion, most inconsistentwith empirical evidence: those having to do with money as a motivator, intrinsic motivation, fit with interdependent work, and fitwith other national cultures. We also choose to address these claims because we believe that sorting effects are especially relevantin these cases and may change the way we think about these claims. Claim (5), performance measures problems, is notcontroversial, but can be addressed (e.g., by using multiple raters/360 systems). Claim (6) is of major importance and is mostrelevant to situations where incentive intensity is high and/or agency problems (moral hazard, adverse selection) are especiallychallenging, typically situations where PFIP is primarily results-based. As noted earlier, agency problems have long beenrecognized (Berle & Means, 1932; Fama & Jensen, 1983; Holmstrom, 1979; Prendergast, 1999) as have ways to attempt to addresssuch problems (e.g., contract choice, use of balanced scorecards) in both the agency and compensation literatures (e.g., Gerhart,2001; Gerhart & Rynes, 2003; Gupta & Shaw, 1998; Lawler, 1971).

4.1. Claim: PFIP/pay does not motivate

Herzberg's (1968) conclusion that pay is not a motivator is generally seen as lacking empirical support. (see Gerhart & Rynes(2003) and Kanfer (1990) for summaries of key studies and reviews.) The work of Herzberg (and of Maslow, 1943) is probablymost useful in that it reminds us that there are many work rewards beyond pay. Measuring the importance of money as amotivator can be done by asking people directly how important money is or perhaps better, by observing people's behavior inresponse to money (Rynes, Gerhart, & Minette, 2004). In terms of pay's effects on behavior, we earlier presented evidence on itsincentive and sorting effects in influencing performance. We also know that organization level studies (e.g., Shaw, Delery, Jenkins,& Gupta, 1998) find that an increase in pay level of 10% is typically associated with a decrease in employee turnover of 10 to 35%(Milkovich et al., 2013). We further know, as discussed earlier, that when people change employers (voluntarily), their payincreases significantly.

When employees are asked which work rewards are most important to them, pay typically ranks high. For example, theSociety for Human Resource Management (SHRM) surveys 600 employees per year and has asked about the importance of 21rewards from 2004 to 2012. Across the years, 60.0% of the roughly 5400 employees said job security was “very important,”followed by benefits at 59.8%, and compensation/pay at 58.6% (SHRM, 2012). Thus, employees are most concerned about thelevel of their total compensation and that there compensation not be at risk. The importance of the other 21 rewards(e.g., opportunities to use skills/abilities, communication, autonomy and independence) was lower. Interestingly, in at least oneyear, SHRM (2006) asked not only employees, but also human resource (HR) professionals to assess the importance of differentrewards to employees. In that year, compensation/pay was ranked 1st in importance by employees, but was only ranked as 6thmost important to employees by HR managers. (Perhaps this evidence suggests that HR managers are more familiar with theargument that pay is not of central importance to employees.) Finally, SHRM (2012) reported the gap between the percentage of

47B. Gerhart, M. Fang / Human Resource Management Review 24 (2014) 41–52

employees saying a reward was very important and the percentage saying they were very satisfied with howmuch of that rewardthey received. The largest gap was for compensation/pay. By the logic of balance theories (e.g., equity theory), perhaps one mightthus argue that compensation/pay is the reward area on which employees are most likely to focus their attention.

Finally, sorting and attraction–selection–attrition processes mean that employees will gravitate to companies that have rewardsthatmatch their reward values/preferences. Therefore, data on what the average (U.S.) employee values is of limited use. Companiescan vary in terms of what rewards they emphasize and that will generate sorting effects. Whatever rewards employees in differentcompanies value, it seems likely that employees will expect that if they engage in certain behaviors (e.g., throwing themselves into aproject, remaining with the company), those rewards (monetary or nonmonetary) will be more likely to follow.

4.2. Claim: PFIP is a de-motivator: Undermining effect on intrinsic motivation and interest8

According to Kohn (1993), a problem with PFIP plans is that they undermine intrinsic motivation because “the more amanager stresses what an employee can earn for good work, the less interested that employee will be in the work itself” (p. 62).Pfeffer (1998) is more to the point: “extrinsic rewards diminish intrinsic motivation” and “large extrinsic rewards can actuallydecrease performance in tasks that require creativity and innovation” (p. 116). These concerns appear again and again (Kohn,1993; Pfeffer, 1998; Ordóñez, Schweitzer, Galinsky, and Bazerman, 2009; see Locke and Latham's reply, 2009 to Ordóñez et al.)and across disciplines (e.g., in accounting and economics, see Frey & Oberholzer-Gee, 1997; Frey & Jegen, 2001; Kreps, 1997; Kuntz &Pfaff, 2002; Prendergast, 2008). The undermining effect of PFIP is also highlighted in the practitioner literature. Fang and Gerhartalso observed that the book, Drive (Pink, 2009), which was recently the #1 selling book on Amazon.com in the categories ofmanagement/leadership and motivation, states that “pay-for-performance schemes…usually don't work and often do harm.”Among the claimed “seven deadly flaws” are that PFIP “can extinguish intrinsic motivation” and “can diminish performance.”

What is the academic research basis for the above claims? It is primarily cognitive evaluation theory (CET; Deci, 1975; Deci,Koestner, & Ryan, 1999; Deci & Ryan, 1985; Ryan, Mims, & Koestner, 1983).9 CET theory, together with Maslow's need hierarchytheory and Herzberg'smotivation-hygiene theory, has provided the primary basis for skepticism and concern about the effectivenessof PFIP in the workplace (Rynes et al. (2005). To begin, we look to CET to define key terms. According to Ryan and Deci (2000, p. 55),“the most basic distinction is between intrinsic motivation, which refers to doing something because it is inherently interesting orenjoyable, and extrinsic motivation, which refers to doing something because it leads to a separable outcome.” Deci and Ryan (1985)further describe intrinsicmotivation as “based in the innate, organismic needs for competence and self-determination” (p. 33), whichis seenwhen “a person does an activity in the absence of a reward contingency or control” (p. 35). Intrinsic interest is a key to intrinsicmotivation because “When people are intrinsically motivated, they experience interest and enjoyment…” (Deci & Ryan, 1985, p. 34)and engage inwork primarily for its own sake, because thework itself is interesting, engaging, or in someway satisfying.” (Amabile etal., 1994, p. 950).

The effect of extrinsic rewards (e.g., PFIP) on these intrinsic outcomes depends on two characteristics under CET (Fang & Gerhart,2012). First, performance-contingent rewards are expected be experienced as controlling, which has a negative effect on intrinsicmotivation and interest (Ryan, Mims, & Koestner, 1983). Second, PFIP has an informational aspect, which is generally seen as likely tohave a positive effect (when performance is strong) because of the positive effect on perceived self-competence (Arnold, 1985;Eisenberger, Rhoades, & Cameron, 1999; Fisher, 1978; Gagné & Deci, 2005).

With some exceptions (see below), until recently, the potential for positive effects of extrinsic rewards has been mostlyoverlooked, while the potential negative effects on intrinsic outcomes have been the focus (Fang & Gerhart, 2012). Yet, Ryan et al.(1983) clearly recognized that if a reward providesmeaningful, positive information regarding self-competence in a context where aperson has discretion (self-determination) in choosing how and when to perform a task (i.e., the reward is not experienced ascontrolling), then the reward can actually increase intrinsic outcomes (Fang&Gerhart, 2012). The focus on the potentially controllingaspect of rewards has been a regular contributor over time to critiques of PFIP.

Nearly all research on the undermining effect of PFIP on intrinsic motivation/interest has been conducted in non-worksettings, often using children as subjects. Although the undermining effect of PFIP in non-workplace settings is settled science inthe opinion of some (Deci et al., 1999; Ryan & Deci, 2000) is not settled in the opinion of others (Eisenberger, Rhoades, &Cameron, 1999). Further, there have long been questions about whether CET theory and research have the same implications inworkplace settings (Gerhart & Milkovich, 1992; Gupta & Shaw, 1998; Kanfer, 1990).

First, an obvious issue is that motivation is not only intrinsic, but also extrinsic. Thus, even if PFIP decreases intrinsic motivation,what matters more is PFIP's effect on total (intrinsic + extrinsic) motivation and this basic question has received surprisingly littledirect attention (Gerhart and Milkovich (1992), p. 552; Gerhart & Rynes, 2003). If we return to themeta-analytic evidence reviewedearlier, we see that workplace studies and/or studies of adults clearly show that PFIP has a large, positive impact on performance,which implies that if there is an undermining effect on intrinsic motivation, it is usually dominated by the positive effect of PFIP onextrinsic motivation.

Second, when one looks at the design of the typical non-workplace setting test of CET, it quickly becomes clear how differentworkplace contexts are. For example, twomeasures of intrinsicmotivation are used in laboratory tests: self-reported intrinsic interest

8 This section draws freely on our recent study (Fang & Gerhart, 2012).9 See also the over-justification effect (Lepper, Greene, & Nisbett, 1973), based on attribution theory, which suggests that people who receive extrinsic rewards

for performing an interesting activity attribute the cause of their behavior to the extrinsic reward, thus discounting their interest in the activity as the cause oftheir behavior.

48 B. Gerhart, M. Fang / Human Resource Management Review 24 (2014) 41–52

and a free-choicemeasure. The latter is defined as “the degree towhichparticipants return to and persist at the target activity during afree-choice period subsequent to the experimental phase” (Deci et al., 1999, p. 635). The free-time measure of intrinsic motivationdoes not seem to make sense in a workplace setting. As Bartol and Locke (2000, p. 108) put it, “what people do during the time theyare not being paid is of no central importance” inwork settings. Another issue is the use of children inmany of the CET tests. EvenDeciet al. (1999) eventually acknowledged that extrinsic rewards “are more detrimental for children than for college students” (p. 656).One explanation they gave was that: “college students have greater cognitive capacity for separating the informational andcontrolling aspects of rewards and are also more accustomed to operating with performance-goal orientations, so they may be moreready to interpret rewards as indicators of their effective performance than as controllers of their behavior” (p.656). Onewould thinkthat the Deci et al. logic would apply even more strongly to adults in the workplace, thus reinforcing the idea that CET may applydifferently in the workplace.

Indeed, in a major development, key architect of CET, Edward Deci, now seems to be moving toward explicitly acknowledgingthis fact (Fang & Gerhart, 2012). Specifically, in 2005, Gagné and Deci (2005) published an article that included a discussion of “Theproblems with CET as a theory of work (emphasis added) motivation.” Much of their (revised) framework is consistent witharguments made by Kanfer (1990) that PFIP in the workplace is more likely to have a positive impact on both perceivedself-competence and perceived autonomy, rather than perceived control (Fang & Gerhart, 2012). As we noted earlier, CET alwaysreadily acknowledged that PFIP could have a positive effect on self-competence. So, the major change is its recognition that PFIP, atleast in workplace settings, may also have a positive effect on perceived autonomy. This evolution is consistent with Kanfer'sargument that “the controlling features of evaluative contingencies are likely to be less salient due to widespread beliefs about theappropriateness of such evaluations in theworkplace” (p. 89) and also consistentwith recent empirical findings indicating that whenemployees perceive a stronger link between pay and performance, they perceive they have more, not less, autonomy (Eisenberger,Rhoades, & Cameron, 1999; Fang&Gerhart, 2012). Gagné andDecinowdescribe CET as being part of self-determination theory (SDT),which, importantly, “explains how extrinsically motivated behavior can become autonomous.” They go on to say that under SDT,“when rewards are administered in an autonomy-supportive climate, they are less likely to undermine intrinsic motivation and, insome cases, can enhance intrinsic motivation” (p. 354, emphasis added).

Third, Fang and Gerhart (2012) demonstrate the importance of sorting effects in understanding employee reactions (includingeffects on their intrinsic motivation/interest) in an employment setting. In laboratory experiments, subjects are randomly assigned toconditions (e.g., PFIP versus no PFIP). In contrast, the process by which employees join and leave work organizations is not random.The difference in processes is important because the nonrandom sorting (or attraction–selection–attrition, Schneider, 1987) process“means that there is less probability of amismatch betweenworker preferences for PFIP and the actual PFIP system that covers them.”As such, even if PFIP has a detrimental incentive effect (as Deci et al. (2009) argue) in the laboratory and/or with children as subjects,in the workplace:

The matching/fit process that occurs…while not perfect, should weaken, eliminate, or even reverse the sign of any suchdetrimental effect of PFIP on intrinsic interest in the work place. The reason is that those employees most likely to reactpositively to PFIP are the higher performers (who do better under PFIP), and it is these employees who are most likely to havehigh perceived competence andwho are alsomost likely to remainwith the organization. In contrast, those employees (lowerperformance, lower perceived competence) most likely to react negatively to PFIP are less likely to remain in an organizationthat emphasizes PFIP (Fang & Gerhart, 2012, p. 1181).

Consistent with this sorting argument, Fang and Gerhart (2012) found that in companies having stronger PFIP, employees haddifferent motivation-related traits. Specifically, the higher the PFIP, the higher the levels of both extrinsic motivation orientation andinternalwork locus of control in the company and these traitsmediated part of the positive relationship (not negative, contrary to CETas traditionally interpreted) between company PFIP strength and intrinsic interest. Fang and Gerhart also found no evidencethat intrinsic and extrinsic motivation orientation were substitutes for one another, instead observing a small, positive correlation(r = .14) between them.

Finally, CET (and SDT) theory has also been used to argue that extrinsic incentives may be especially harmful to one particularaspect of performance: creative performance. As noted, such arguments stem from a primary focus on the potential controllingaspect of rewards. However, another theory, learned industriousness (Eisenberger, 1992), “focuses on the informational aspectsof rewards…that [guide] goal-directed behavior and, thus, increase creative performance” (Byron & Khazanchi, 2012, p. 809).Indeed, the meta-analytic review by the latter authors found that in 34 experimental studies, the use of extrinsic incentivescontingent on creative performance resulted in a positive effect on creative performance (Hedges g = .62, meaning that creativeperformance was .62 standard deviations higher when extrinsic incentives were used, a large effect size using conventionalbenchmarks). Byron and Khazanchi also summarized evidence from eight nonexperimental studies. The effect size was againpositive, but much smaller (Hedges g = .07). This difference in effect sizes reinforces the challenge of translating results from thelaboratory to the field.

4.3. Claim: PFIP does not fit with teams

Equity theory (Adams, 1963) tells us that the effect of differences in pay between employees (i.e., unequal pay, pay dispersion)on employees' behaviors depends on whether the pay dispersion is seen as equitable. That equity perception depends on(a) whether differences in inputs like performance are perceived to exist and (b) whether they are perceived as providing the

49B. Gerhart, M. Fang / Human Resource Management Review 24 (2014) 41–52

basis for differences in outcomes like pay. Whether work is independent or interdependent (e.g., in a team setting), employeereactions (including performance level) to pay dispersion, according to equity theory, depend on whether the dispersion isperceived as being credibly explained by performance-related inputs or not. Empirical evidence from settings where work is lowon interdependence (i.e., cooperation/collaboration needs are low) supports the equity theory prediction that pay dispersion, ifbased on performance differences, can have positive effects on performance (Kepes, Delery, & Gupta, 2008; Shaw, Gupta, &Delery, 2002).

However, conventional wisdom has been that pay dispersion, even if based on employee differences in performance, is not agood idea where work is interdependent (Bloom, 1999; Pfeffer & Langton, 1993; Shaw et al., 2002). These authors generallyrecognize advantages (e.g., individual motivation) and disadvantages (e.g., jealousy, decreased cooperation, even sabotage) of paydispersion based on individual performance, but conclude that the disadvantages generally dominate in team settings. Work inthe field is still at an early stage in understanding pay dispersion's effects where work is interdependent, but one likelycontingency variable, consistent with equity theory, is the degree to which there is a performance basis for pay dispersion (Shawet al., 2002).

Trevor, Reilly, and Gerhart (2012) emphasize that pay dispersion, even in interdependent work settings, has the potential for anet positive effect for at least two reasons. First, as noted, equity theory does not say that equal pay (low dispersion) is necessarilyseen as equitable. Indeed, equal pay will be perceived as inequitable if credible performance differences exist, and would thus beexpected to undermine motivation. Individual motivation in teams is a classic challenge, as indicated by longstanding program ofstudy on free rider (Olson, 1965) and social loafing (Shepperd, 1993) problems in groups (Gerhart & Rynes, 2003). So, it isprobably unwise to overlook or downplay the importance of maintaining individual motivation, ability, and performance in teamsettings. In fact, Shepperd (1993, p. 70) argues that monetary rewards tied to individual inputs “can serve as powerful incentivesfor behavior, countering the reduction in effort typically exhibited by participants who are combining their efforts.”

Second, Trevor et al. (2012) observe that even if low pay dispersion produces cooperation and collaboration in teams and even ifthe motivation of individuals who join and remain in the group does not suffer, the (negative) sorting effects of low pay dispersionmay still contribute to lower team performance by affectingwho joins and stays with the team. The reason is that PFIP is not possiblewith low pay dispersion. As discussed earlier, we know from previous research that high performers are more likely to join and staywith organizations that have stronger PFIP. Thus, even if low pay dispersion is the path to collaboration and cooperation in teams, itrisks an adverse selection/sorting effect such that high level of collaboration and cooperation takes place, but without the mostcapable team members, that would seem an unlikely path to high team performance. Indeed, Trevor et al. found in their sample ofNational Hockey League teams and players that the teams having the strongest PFIPweremore likely to attract/retain top talent. Also,team PFIP was tightly intertwined with team pay level. Thus, for a team to follow a low pay dispersion strategy, it might also have tofollow a low pay level strategy. As such, a teamwould have to overcome the closely related negative sorting consequences of not onlylow PFIP, but also low pay level (Gerhart & Rynes, 2003; Shaw, Delery, Jenkins, & Gupta, 1998). In hockey and in other settings wereindividual performance differences are substantial, measurable, and visible to the labormarket, it is difficult to see howpay equality isfeasible because under these conditions, high performers have the option of moving to an employer where, in return for maintainingtheir same high level of performance, they will get paid more.

4.4. Claim: PFIP does not fit many national cultures

Textbooks in human resource management and organizational behavior frequently include a section on national culturedifferences between countries and the implications of such differences for the design of management practice (e.g., PFIP) indifferent countries. Our textbooks seem to generally take the view that organizations must “localize” their practices to fit thenational culture of the country where operations are located. The traditional basis for this view is the work of Hofstede. Accordingto Hofstede (1983, p. 88), “Management is ‘culturally dependent’ and if we see what effective organizations in different cultureshave done, we recognize that their leaders did adapt foreign management ideas to local cultures” (p. 88). Although Hofstede(2001) recognizes that “there have been examples of multinationals successfully reforming local cultural traits,” he cautions that“this is a difficult task” and “for best results a multinational's management practices should fit the local culture” (pp. 441–442).

However, the basis for Hofstede's managerial practice recommendations, his classic work that assessed themagnitude of nationalculture differences (Hofstede, 1980, 2001), has been challenged (McSweeney, 2002; Gerhart & Fang, 2005). The key questionaddressed by Hofstede (1980) was the degree to which cultural values of individuals could be explained by knowing their country.Based on his analyses, he concluded that “nationality of respondents… affects [individual values] highly significantly” (1980, p. 71).Later summaries of his work sometimes go even further, in one case, for example, stating that “Hofstede found striking culturaldifferences within a single multinational corporation. In his study, national culture explained 50% of the differences in employees'attitudes and behaviors” (Adler, 2002: 67).

In contrast, Gerhart and Fang (2005), upon re-analyzing Hofstede's data, found that country was much less important thanindicated above. One reason is thatHofstede relied exclusively on tests of statistical significance to assess importance. But,with a largesample size like he used, statistical significance is not very informative in cross-country comparisons (Fang, 1993). Gerhart and Fanginstead estimated the effect size for country and found that country explained 2 to 4% of the variance in individual values. That(modest) level of effect size does not support a view of country differences in individual values as being so large and systematic as tobe likely to act as an important constraint on human resource practices such as PFIP.

Gerhart and Fang (2005) further emphasized, building on work of Bloom and Milkovich (1999), that sorting effects (also)needed to be considered in assessing the degree to which management practices will work in different countries. They note that

50 B. Gerhart, M. Fang / Human Resource Management Review 24 (2014) 41–52

even if some management practices are, on average, not a good fit to some countries, for management to be truly constrained, “itmust [also] be assumed that companies attract, select and retain employees in a random fashion (Bloom and Milkovich, 1999).Otherwise, if, for example, a company operates in a country with a mean level of individualism that is too low for its needs,” thecompany could “simply establish a hiring hurdle above that mean,” which would enable it to “find enough employees that havesufficiently high individualism to fit its management system” (Gerhart & Fang, 2005, p. 974). See Gerhart (2009) for further workon sorting effects in this context.

Given the information above, it is perhaps not surprising that recent empirical evidence indicates that country and/or nationalculture differences are less of a contingency variable than expected. Indeed, it may be that if there is a national norm for humanresource practices, deviating from that norm may in some cases be the path to higher performance (Rabl, Jayasinghe, Gerhart, &Kϋhlmann, 2010), consistent with the resource-based view of the firm (Barney, 1991). In the specific case of PFIP, there is no gooddirect evidence onwhether its effectiveness depends on the country. However, there is evidence available on usage of PFIP by country.Conventional wisdom would be that respondents from more collectivistic countries such as China and Japan would be much lesslikely to use what is called the equity rule (i.e., differentiating pay based on individual performance) than are respondents fromindividualistic countries like the United States (e.g., Early & Erez, 1997). Instead, those from collectivistic countries would be expectedto rely more on either the equality or need reward allocation rule. However, empirical findings, including a meta-analysis (Fischer &Smith, 2003) show that such differences, when they exist, are generally quite small in practical terms (Milkovich et al., 2013, Chapter16). For example, in a study by Zhou andMartocchio of Chinese and U.S. respondents, individual performance explained 64.2% of thevariance in bonus allocations, indicating the dominance of the equity rule across nationalities. In contrast, the interaction betweennationality and use of the equity (and need) rules explained only 1.1% of the variance (Gerhart, 2009). Data from Hewitt Associates(summarized inMilkovich et al., 2013.) shows that the ratio ofmerit increases given to high versus average performers in recent yearsis higher in China (and Japan) than in the United States.

5. Conclusion

We have provided a survey of selected issues concerning the use and effectiveness of PFIP in organizations. There are bothpotential advantages and disadvantages to using PFIP. We have focused more in this paper on the “case for” using PFIP. Inparticular, we have highlighted the importance of sorting effects in evaluating the effectiveness of PFIP and we believe we haveshown that explicitly recognizing the role of sorting effects paints a more positive (and accurate) picture of PFIP's likely impact oneffectiveness. Especially upon taking such sorting effects into consideration, our logic is that while PFIP is not a panacea and pay isnot the only motivator and things can (and do) sometimes go terribly wrong with PFIP, it is difficult to think of a better overallalternative. Winston Churchill said that “democracy is the worst form of government except all the others that have been tried.”Our evaluation of PFIP is similar. Based on what we know from research and practice, it is difficult to envision how individuals,companies, and economies would fare better, on average, by significantly diminishing the role of PFIP in organizations. As withmost things, the success of PFIP will depend both on the design and execution of the strategy. It also seems likely that the strongerthe incentive intensity of PFIP (most typically when results-based plans are used), the more likely it is that substantial gains inperformance can be obtained, but also the higher the risk that serious unintended negative consequences can occur. That meansthat organizations must not only think carefully about how to balance their use of results-based and behavior-based performancemeasures and individual-based and group/organization-based performance measures, but also their appetite for risk in choosingincentive/PFIP intensity.

More comprehensive reviews of compensation theory, research, and practice (Gerhart & Rynes, 2003; Gerhart, Rynes, & Fulmer,2009; Gomez-Mejia et al., 2010) provide recommendations for future research. As such, we focus here on only a few selected areas forfuture research. One avenuewould be to better assess the implementation and survival of PFIP strategies. To assess effectiveness, onemust observe not only the strategies that survive, but also the ones that do not and, if possible, why they do not survive. Suchinformation would go a long way toward providing a better grasp of not only how well PFIP can work, but also how often it goeswrong and, when it does, how badly it goes wrong. In doing so, it would be important to move beyond anecdotes and case studies ifone wishes to be able to estimate success and failure rates. Any management strategy, including in the area of compensation, willsometimes succeed andwill sometimes fail. The important questions are: what is the ratio of success to failure and howmuch value iscreated on balance over time?

A second avenue of needed research is tomore comprehensively define andmeasure PFIP and study its consequences. Thatmeansnot limiting attention only to annual merit increases. As we have shown, the merit pay aspect of PFIP has declined in relativeimportance.Morework is (still) needed onmerit pay's effects, but PFIPmust includemore than traditionalmerit pay (see Exhibit 1) tomore accurately examine the effects of PFIP when more fully captured. Third, more work is needed to examine whether PFIP candemotivate employees under some circumstances. Our feeling at this point is that relevant theory, when applied to the workplace, isunlikely to find a general or practically meaningful effect of PFIP as a de-motivator (Fang & Gerhart, 2012). However, as we havediscussed, this idea of de-motivating effects of PFIP seems to have captured the imagination of scholars inmany disciplines (aswell asothers) and the bestway to address this idea is to address it directlywith data.We believe that the broader issue of how PFIP interactswith other rewards to influence intrinsic motivation, extrinsic motivation, overall motivation, and performance (including creativeperformance) is worth further examination. Finally, wewould like to seemorework on the role of hypothesized contextual variables,such as national culture and work interdependence, in the use and effectiveness of PFIP. Although we believe that such contextualfactors do not necessarily constrain PFIP in as much of a wholesale manner as sometimes suggested, we do think it is important tounderstand under what conditions these factors may serve as important moderator variables.

51B. Gerhart, M. Fang / Human Resource Management Review 24 (2014) 41–52

References

Abosch, K. (2012). Why can't (won't) we differentiate. Presentation at Wisconsin School of Business, University of Wisconsin—Madison, May.Adler, N. J. (2002). International dimensions of organizational behavior (4th ed.) Cincinnati: South-western.Aggrawal, R. K., & Samwick, A. A. (1999). The other side of the trade-off: The impact of risk on executive compensation. Journal of Political Economy, 107, 65–105.Arnold, H. J. (1985). Task performance, perceived competence, and attributed causes of performance as determinants of intrinsic motivation. Academy of

Management Journal, 28, 876–888.Barney, J. (1991). Firm resources and sustained competitive advantage. Journal of Management, 17, 99–120.Bartol, K.M., & Locke, E. A. (2000). Incentives andmotivation. In S. L. Rynes, & B. Gerhart (Eds.), Compensation in organizations (pp. 104–147). San Francisco, CA: Jossey-Bass.Bebchuck, L. A., Cremers, K. J. M., & Peyer, U. C. (2011). The CEO pay slice. Journal of Financial Economics, 102, 199–221.Berle, A. A., & Means, G. C. (1932). The modern corporation and private property. NewYork: Macmillan Publishing.Bloom, M. (1999). The performance effects of pay dispersion on individuals and organizations. Academy of Management Journal, 42(1), 25–40.Bloom, M., & Milkovich, G. T. (1999). A SHRM perspective on international compensation and rewards. In Patrick Wright et al. (Eds.), Research in Personnel and

Human Resources Management, Supplement 4.Byron, K., & Khazanchi, S. (2012). Rewards and creative performance: A meta-analytic test of theoretically derived hypotheses. Psychological Bulletin, 138,

809–830.Cable, D. M., & Judge, T. A. (1994). Pay preferences and job search decisions: A person-organization fit perspective. Personnel Psychology, 47, 317–348.Cadsby, C. B., Song, F., & Tapon, F. (2007). Sorting and incentive effects of pay-for performance: An experimental investigation. The Academy of Management Journal, 50,

387–405.Combs, J., Liu, Y., Hall, A., & Ketchen, D. (2006). How much do high-performance work practices matter? A meta-analysis of their effects on organizational

performance. Personnel Psychology, 59, 501–528.Dalton, D. R., Daily, C.M., Certo, S. T., & Roengpitya, R. (2003). Meta-analyses of financial performance and equity: Fusion or confusion? Academy ofManagement Journal,

46(1), 13–26.Deci, E. L. (1975). Intrinsic motivation. New York: Plenum Press.Deci, E. L., Koestner, R., & Ryan, R. M. (1999). A meta-analytic review of experiments examining the effects of extrinsic rewards on intrinsic motivation.

Psychological Bulletin, 25, 627–668.Deci, E. L., & Ryan, R. M. (1985). Intrinsic motivation and self-determination in human behavior. New York: Plenum Press.Dreher, G. F., & Cox, T. H., Jr. (2000). Labor market mobility and cash compensation: the moderating effects of race and gender. Academy of Management Journal, 43,

890–900.Early, P. C., & Erez, M. (1997). The transplanted executive: Why you need to understand how workers in other countries see the world differently. New York: Oxford

University Press.Eisenberger, R. (1992). Learned industriousness. Psychological Review, 99, 248–267.Eisenberger, R., Rhoades, L., & Cameron, J. (1999). Does pay for performance increase or decrease perceived self-determination and intrinsic motivation? Journal of

Personality and Social Psychology, 77, 1026–1040.Fama, E. F., & Jensen, M. C. (1983). Separation of ownership and control. Journal of Law and Economics, 26, 301–325.Fang, M. (1993). Differences in work definitions: A study of the United States, Japan, and West Germany. MSc thesis, Cornell University.Fang, M., & Gerhart, B. (2012). Does pay for performance diminish intrinsic interest? A workplace test using cognitive evaluation theory and the attraction-selection-

attrition model. International Journal of Human Resource Management, 23, 1176–1196.Fischer, R., & Smith, P. (2003). Reward allocation and culture: A meta-analysis. Journal of Cross-Cultural Psychology, 34, 251–268.Fisher, C. D. (1978). The effects of personal control, competence, and extrinsic reward systems on intrinsic motivation. Organizational Behavior and Human

Performance, 21, 273–288.Frey, B. S., & Jegen, R. (2001). Motivation crowding theory. Journal of Economic Surveys, 15, 589–611.Frey, B. S., & Oberholzer-Gee, F. (1997). The cost of price incentives: An empirical analysis of motivation crowding-out. American Economic Review, 87, 746–755.Gagné, M., & Deci, E. L. (2005). Self-determination theory and work motivation. Journal of Organizational Behavior, 26, 331–362.Garen, J. E. (1994). Executive compensation and principal-agent theory. Journal of Political Economy, 102, 1175–1199.Gerhart, B. (2000). Compensation strategyandorganizational performance. In S. L. Rynes, & B. Gerhart (Eds.), Compensation in organizations. San Francisco, CA: Jossey-Bass.Gerhart, B. (2001). Balancing results andbehaviors inpay for performanceplans. In C. Fay (Ed.), The executive handbook of compensation (pp. 214–237).NewYork: Free Press.Gerhart, B. (2004). Comment on ‘promise and peril in implementing pay-for-performance’. Human Resource Management, 43, 29–31.Gerhart, B. (2009). Does national culture constrain organization culture and human resource strategy? The role of individual mechanisms and implications for

employee selection. Research in Personnel and Human Resources Management, 28, 1–48.Gerhart, B., & Fang,M. (2005). National culture and human resourcemanagement: Assumptions and evidence. International Journal of Human ResourceManagement, 16,

975–990.Gerhart, B., &Milkovich, G. T. (1989). Salaries, salary growth, and promotions ofmen andwomen in a large, private firm. InR.Michael, H. Hartmann, & B.O'Farrell (Eds.),

Pay equity: Empirical inquiries (pp. 23–43). Washington, DC: National Academy Press.Gerhart, B., & Milkovich, G. T. (1990). Organizational differences in managerial compensation and financial performance. The Academy of Management Journal, 33,

663–691.Gerhart, B., & Milkovich, G. T. (1992). Employee compensation: Research and practice. In M. D. Dunnette, & L. M. Hough (Eds.), Handbook of industrial & organizational

psychology (pp. 481–569) (2nd ed.). Palo Alto, CA: Consulting Psychologists Press, Inc.Gerhart, B., & Rynes, S. L. (2003). Compensation: Theory, evidence, and strategic implications. Thousand Oaks, CA: Sage.Gerhart, B., Rynes, S. L., & Fulmer, I. S. (2009). Pay and performance: Individuals, groups, and executives. The Academy of Management Annals, 3, 251–315.Gerhart, B., Trevor, C., & Graham, M. (1996). New directions in employee compensation research. Research in Personnel and Human Resources Management,

143–203.Gibbons, R. (1998). Incentives in organizations. The Journal of Economic Perspectives, 12, 115–132.Gomez-Mejia, L. R., & Balkin, D. B. (1992). Compensation, organizational strategy, and firm performance. Cincinnati, OH: Southwestern Publishing.Gomez-Mejia, L. R., Berrone, P., & Franco-Santos, M. (2010). Compensation and organizational performance. Armonk, N.Y.: M.E. Sharpe.Gupta, N., & Shaw, J. D. (1998). Let the evidence speak: Financial incentives are effective!! Compensation and Benefits Review, 26(2), 28–32.Guzzo, R. A., Jette, R. D., & Katzell, R. A. (1985). The effects of psychologically based intervention programs on worker productivity: A meta-analysis. Personnel Psychology,

38, 275–291.Hall, B. J., & Liebman, J. B. (1998). Are CEOs paid like bureaucrats? The Quarterly Journal of Economics, 113, 653–691.Hansen, F. (2000). The turnover myth. Workforce Management, 84(6), 34–38.Harrison, D. A., Virick, M., & William, S. (1996). Working without a net: Time, performance, and turnover under maximally contingent rewards. Journal of Applied

Psychology, 81, 331–345.Heneman, R. L., & Werner, J. (2000).Merit pay: Linking pay to performance in a changing world (2nd ed.)Charlotte, NC: Information Age Publishing.Herzberg, F. (1968). One more time: how do you motivate employees? Harvard Business Review, 46(1), 53–62.Hitt, M. A. (2005). Management theory and research: Potential contribution to public policy and public organizations. Academy of Management Journal, 48,

963–966.Holmstrom, B. (1979). Moral hazard and observability. Bell Journal of Economics, 10, 74–91.Hofstede, G. (1980). Culture's consequences: International differences in work-related values. Beverly Hills, CA: Sage Publications.Hofstede, G. (1983). The cultural relativity of organizational practices and theories. Journal of International Business Studies, 14, 75–89.Hofstede, G. (2001). Culture's consequences: Comparing values, behaviors, institutions, and organizations across nations (2nd ed.)Thousand Oaks, CA: Sage Publications.

52 B. Gerhart, M. Fang / Human Resource Management Review 24 (2014) 41–52

House, R. J., Hanges, P. J., Javidan, M., Dorfman, P. W., & Gupta, V. (2004). Culture, leadership, and organizations: The globe study of 62 societies. Thousand Oaks, CA:Sage Publications.

Jenkins, D. G., Jr., Mitra, A., Gupta, N., & Shaw, J. D. (1998). Are financial incentives related to performance? A meta-analytic review of empirical research. TheJournal of Applied Psychology, 83, 777–787.

Jensen, M. C., & Murphy, K. J. (1990). Performance pay and top management incentives. Journal of Political Economy, 98, 260–299.Judiesch, M. K. (1994). The effects of incentive compensation systems on productivity, individual differences in output variability and selection utility. University of Iowa

(Doctoral dissertation).Kanfer, R. (1990). Motivation theory and industrial and organizational psychology. In M. Dunnette, & L. Hough (Eds.), Handbook of industrial and organizational

psychology (2nd ed.), Vol. 1. (pp. 75–170). Palo Alto, CA: Consulting Psychologists Press, Inc.Kepes, S., Delery, J. E., & Gupta, N. (2008). Contingencies in the effects of pay range on organizational effectiveness. Personnel Psychology, 62, 497–531.Kohn, A. (1993). Why incentive plans cannot work. Harvard Business Review, 71(5), 54–63.Kreps, D. M. (1997). Intrinsic motivation and extrinsic incentives. American Economic Review, 87, 359–364.Kuntz, A. H., & Pfaff, D. (2002). Agency theory, performance evaluation, and the hypothetical construct of intrinsic motivation. Accounting, Organizations, and

Society, 27, 275–295.Larkin, I., Pierce, L., & Gino, F. (2012). The psychological costs of pay-for-performance: Implications for the strategic compensation of employees. Strategic

Management Journal, 33, 1194–1214.Lawler, E. E., III (1971). Pay and organizational effectiveness: A psychological view. New York: McGraw-Hill.Lazear, E. P. (1986). Salaries and piece rates. Journal of Business, 59, 405–431.Lazear, E. (2000). Performance pay and productivity. The American Economic Review, 90, 1346–1361.Locke, E. A., Feren, D. B., McCaleb, V.M., Shaw, K. N., & Denny, A. T. (1980). The relative effectiveness of fourmethods ofmotivating employee performance. In K. D. Duncan,

M. M. Gruenberg, & D. Wallis (Eds.), Changes in working life (pp. 363–388). New York: Wiley.Martin, G. P., Gomez-Mejia, L. R., & Wiseman, R. M. (2013). Executive stock options as mixed gambles: Revisiting the Behavioral Agency Model. Academy of

Management Journal, 56(2), 451–472.McSweeney, B. (2002). Hofstede's model of national cultural differences and their consequences: A triumph of faith - a failure of analysis. Human Relations, 55,

89–118.Milgrom, P., & Roberts, J. (1992). Economics, organization, and management. Englewood Cliffs, NJ: Prentice Hall.Milkovich, G. T., Newman, J. M., & Gerhart, B. (2013). Compensation (11th ed.)New York: McGraw-Hill/Irwin.Milkovich, G. T., & Wigdor, A. K. (1991). Pay for performance. Washington, DC: National Academy Press.Nyberg, A. J. (2010). Retaining your high performers: Moderators of the performance–job satisfaction–voluntary turnover relationships. The Journal of Applied Psychology,

95, 440–453.Nyberg, A., Fulmer, I., Gerhart, B., & Carpenter,M. A. (2010). Agency theory revisited: CEO returns and shareholder interest alignment. The Academy ofManagement Journal,

53, 1029–1049.Ordóñez, L. D., Schweitzer, M. E., Galinsky, A. D., & Bazerman, M. H. (2009). Goals gone wild: The systematic side effects of overprescribing goal setting. Academy of

Management Perspectives, 23, 6–16.Pfeffer, J. (1998). The human equation: Building profits by putting people first. Boston, MA: Harvard Business School Press.Pfeffer, J., & Langton, N. (1993). The effect of dispersion on satisfaction, productivity, and working collaboratively: Evidence from college and university faculty.

Administrative Science Quarterly, 38(3), 382–407.Prendergast, C. (1999). The provision of incentives in firms. Journal of Economic Literature, 37, 7–63.Prendergast, C. (2008). Intrinsic motivation and incentives. American Economic Review, 98, 201–205.Rabl, T., Jayasinghe, M., Gerhart, B., & Kϋhlmann, T. M. (2010). How much does country matter? A meta-analysis of the HPWP systems–business performance

relationship. Academy of Management Annual Meeting Proceedings.Roy, D. (1952). Quota restriction and gold bricking in a machine shop. American Journal of Sociology, 57, 427–442.Ryan, R. M., & Deci, E. L. (2000). Self-determination theory and the facilitation of intrinsic motivation, social development, and well-being. American Psychologist,

55(1), 68–78.Ryan, R. M., Mims, V., & Koestner, R. (1983). Relation of reward contingency and interpersonal context to intrinsic motivation: A review and test using cognitive

evaluation theory. Journal of Personality and Social Psychology, 45, 736–750.Rynes, S. L. (1987). Compensation strategies for recruiting. Topics in Total Compensation, 2, 185–196.Rynes, S. L., Gerhart, B., &Minette, K. A. (2004). The importance of pay in employeemotivation: Discrepancies betweenwhat people say andwhat they do.HumanResource

Management, 43, 381–394.Rynes, S. L., Gerhart, B., & Parks, L. (2005). Personnel psychology: Performance evaluation and pay for performance. Annual Review of Psychology, 56, 571–600.Salamin, A., & Hom, P. W. (2005). In search of the elusive U-shaped performance turnover relationship: Are high performing Swiss bankers more liable to quit?

The Journal of Applied Psychology, 90, 1204–1216.Sanders, W. G., & Hambrick, D. C. (2007). Swinging for the fences: The effects of CEO stock options on company risk taking and performance. Academy of

Management Journal, 50, 1055–1078.Schneider, B. (1987). The people make the place. Personnel Psychology, 40, 437–453.Shaw, J. D., Delery, J. E., Jenkins, J., & Gupta, N. (1998). An Organization-Level Analysis Of Voluntary And Involuntary Turnover. Academy of Management Journal, 41,

511–525.Shaw, J. D., Dineen, B. R., Fang, R., & Vellella, R. F. (2009). Employee–organization exchange relationships, HRM practices, and quit rates of good and poor

performers. The Academy of Management Journal, 52, 1016–1033.Shaw, J. D., & Gupta, N. (2007). Pay system characteristics and quit patterns of good, average, and poor performers. Personnel Psychology, 60, 903–928.Shaw, J. D., Gupta, N., & Delery, J. E. (2002). Pay dispersion and work force performance: Moderating effects of incentives and interdependence. Strategic

Management Journal, 23, 491–512.Shepperd, J. A. (1993). Productivity loss in performance groups: A motivation analysis. Psychological Bulletin, 113, 67–81.Stajkovic, A. D., & Luthans, F. (1997). A meta-analysis of the effects of organizational behavior modification on task performance, 1975–1995. Academy of

Management Journal, 40, 1122–1149.Tosi, H. L., Werner, S., Katz, J. P., & Gomez-Mejia, L. R. (2000). How much does performance matter? A meta-analysis of CEO pay studies. Journal of Management, 26,

301–339.Trank, C. Q., Rynes, S. L., & Bretz, R. D., Jr. (2002). Attracting applicants in the war for talent: Differences in work preferences among high achievers. Journal of

Business and Psychology, 16, 331–345.Trevor, C. O., Gerhart, B., & Boudreau, J. W. (1997). Voluntary turnover and job performance: Curvilinearity and the moderating influences of salary growth and

promotions. The Journal of Applied Psychology, 82, 44–61.Trevor, C. O., Reilly, G., & Gerhart, B. (2012). Reconsidering pay dispersion's effect on the performance of interdependent work: Reconciling sorting and pay

inequality. The Academy of Management Journal, 55, 585–610.Viswesvaran, C., Ones, D. S., & Schmidt, F. L. (1996). Comparative analysis of the reliability of job performance ratings. The Journal of Applied Psychology, 81, 557–574.Watson, Towers (2012). Moderate pay raises on tap for U.S. workers, Towers Watson survey finds. http://www.towerswatson.com/press/7856 (August 24).Whyte, W. F. (1955). Money and motivation: An analysis of incentives in industry. NewYork: Harper Brothers.WorldatWork (2010). Promotional guidelines. (Scottsdale, AZ).WorldatWork (2012). Compensation programs and practices 2012. (Scottsdale, AZ).Wright, P. M., George, J. M., Farnsworth, R., & McMahan, G. C. (1993). Productivity and extra-role behavior: The effects of goals and incentives on spontaneous

helping. The Journal of Applied Psychology, 78, 374–381.