COMPENSATION MANAGERS' BELIEFS ABOUT STRATEGIES THAT AFFECT COMPENSATION PROGRAM GOALS Nancy E. Day University of Missouri-Kansas City Kansas City, Missouri Abstract A study was conducted to investigate whether individual compensation program goals of recruitment effectiveness, motivational effectiveness, admin- istrative manageability, cost effectiveness, retaining employees and encourag- ing poor performers to leave were differentially impacted by eleven compen- sation strategies. It was found that compensation strategies did have differen- tial effects on perceptions of the effectiveness of the compensation goals. While participation in compensation plan design and internal consistency were the most consistent predictors, there was significant variation in the compensation strategies that were predictive across all effectiveness perception variables. Based on these results, compensation managers should carefully consider the organization s strategy and desired employee behaviors before designing com- pensation programs. Introduction More than ever before, compensation programs are being challenged to provide a definite and measurable impact on the bottom line results of the company. Many organizations are designing pay programs that de-emphasize individual effort in favor of team, group, or organizational performance. Oth- ers are radically altering the traditional conception of base pay by creating broad bands rather than the typically narrow grades. Some organizations are putting less money into base pay programs and more into pay that requires employees to take significant amounts of risk. However, no matter the extent and type of compensation innovations, the ultimate goals of compensation sys- tems are to attract and retain good employees and motivate their performance. Although organizations utilize a variety of different methods to achieve these compensation objectives, there is currently little research that identifies com- pensation strategy components that are associated with these specific goals of effective compensation programs. Previous researchers (Balkin & Gomez-Mejia, 1990; Gomez-Mejia, 1992) found that compensation program effectiveness and organizational effectiveness were higher when particular pay strategies were appropriately matched with or- ganizational strategies. Specifically, they found that compensation strategies that 66 Journal of Business Strategies Vol. 13, No. 1 emphasize more flexible, adaptive and incentive-based pay plans were most effective in organizations that focus on a single product or service and are in a growth stage. Mature organizations that are more diversified, focusing on a dominant rather than a single product or service, and that use traditional, job- based, hierarchical and less flexible compensation strategies tend to be most successful. However, the effectiveness of any compensation system hinges on its success in meeting a number of critical objectives. In their 1990 study, Balkin and Gomez-Mejia measured compensation effectiveness through compensation managers' overall ratings of their compensation systems. They did not focus on the specific goals of compensation, that is, to attract, retain and motivate employee performance and behaviors. The study described in this paper investigates the pay strategies that compensation managers perceive as being most associated with achievement of each of the compensation system's primary goals. Compensation programs have multiple goals, most commonly expressed as attraction, retention and motivation. The decisions to accept a job offer, to remain with or leave the employer and to perform well are controlled to some degree by the compensation program, and there is little reason to believe that simply because a compensation system attracts and retains employees that it also is able to motivate them (Wallace & Fay, 1988). Different pay strategies would theoretically have differential effects on the achievement of each of these multiple goals. For example, evidence exists that well-designed individual incentives, such as bonuses that are contingent on an employee achieving certain goals, have a positive effect on productivity (Gerhart & Milkovich, 1992). However, pay pro- grams that require a significant percentage of pay to be "at risk," (employees' wages are contingent upon achieving a certain performance level) may be more problematic in recruiting workers, since the guaranteed wage is often lower than at competitor organizations (BNA, 1988; Milkovich & Newman, 1993). Thus, a pay-at-risk strategy may differentially affect two compensation system goals, attraction and motivation. Further, pay that is equal to or better than that found at competing orga- nizations may greatly affect attraction and retention, but has little effect on mo- tivation once the workers have been hired (Mahoney, 1979). Although research in efficiency wage theory has found that higher pay does seem to increase the number of candidates (Krueger, 1988; Holzer, 1990; Rynes & Barber, 1990), attract and retain more employees (Lawler, 1971) and attract better quality and longer tenured employees (Brown & Medoff, 1989; Holzer, 1990) there is little research to indicate that employees' motivation is affected by pay level during employment. Additionally, the decision processes involved in the acceptance of a job offer and maintaining organizational membership may rely on different types of information about payor other factors. Comparisons involving internal eq- uity (the internal value the organization places on jobs, usually established through some type of job evaluation program) are less likely to be made in Spring 1996 Day: Compensation Program Goals 67 the initial job-choice decisions, since the internal hierarchy is not obvious to job candidates. Factors that affect the attractiveness of the organization, such as management practices, quality of supervision or organizational climate may be less known to job candidates, but may make a larger contribution to turn- over decisions than characteristics of the compensation system. In addition to attraction, retention and motivation, compensation systems must also meet secondary objectives which are related to the constraints in which all organizations operate, such as efficiency and equity (Milkovich & Newman, 1993). Compensation systems must have goals of cost-effectiveness and administrative manageability, and include considerations of the organization's ability to pay, compliance with legislation and labor contracts, and effective responses to internal and external labor markets (Hills, Bergmann & Scarpello, 1994). For example, pay secrecy is an administrative strategy that may increase the manageability of the system, in that less infonnation has to be communicated and fewer challenges to fairness may have to be negotiated. However, pay secrecy would also theoretically make it difficult to use the sys- tem as an incentive to perfonnance (Wallace & Fay, 1988), since employees may be unaware of the magnitude of potential salary increases available through advancement through their current salary grade or promotions. Simi- larly, pay levels that are high relative to the market may improve recruitment but have a negative effect on cost effectiveness. Thus, it is clear that different compensation strategies should theoretically affect compensation effectiveness outcomes differentially. Knowing if, when and how these differential effects may occur could be of great benefit to com- pensation managers responsible for the design of these programs. For example, if an organization's strategy necessitates attraction and retention of highly skilled workers, what compensation strategy mix would maximize these goals? On the other hand, if the organization needs to increase productivity, enhanc- ing motivation through a properly designed compensation system is essentiaL Knowing the strategies that will maximize the desired compensation effective- ness outcomes would be invaluable. In this study, we propose to begin an in- vestigation into these design issues by asking compensation managers which strategies they believe are most effective for each of the major compensation program goals. The proposition to be addressed in this study is: Proposition: Compensation managers' perception of the effectiveness of the individual goals of compensation (recruitment, retention, moti- vation, etc.) will be differentially predicted by the type of compensa- tion strategy utilized. Because there is little previous work that would guide and direct specific pre- dictions, this proposition does not delineate precise relationships between com- pensation strategies and measures of compensation program goal effectiveness. One of our goals is exploratory; we seek to investigate possible relationships 68 Journal of Business Strategies Vol. 13, No. 1 that can be built upon in future research. The proposition formally investigated is that no one strategy or strategies will predict effectiveness perception mea- sures, but we are also interested in the pattern of relationships that emerges from our findings. Method A questionnaire was sent to 1,121 midwestern organizations, taken from two samples. One sample was from a Dun & Bradstreet database (866 organi- zations of 200 or more on-site employees from five states: Arkansas; Iowa, Kansas, Missouri, Nebraska); the other was a membership directory of a local HR manager's professional organization (The Human Resource Management Association of Greater Kansas City; 255 organizations). The person in charge of the compensation system, who was identified through telephone calls (for the Dun & Bradstreet sample) or through the membership roster (for the HR professional organization), was sent a copy of the survey. The six page ques- tionnaire asked respondents to answer questions regarding their compensation policies as of January 1, 1994 only for non-union employee groups. From the Dun & Bradstreet (D&B) sample, 121 organizations responded, resulting in a response rate of 14%; for the HR managers' group, 52 responded, for a re- sponse rate of 20%. The total response rate was 15%. To ensure that the re- spondents were comparable to non-respondents, analyses of variances were conducted between these groups for the D&B database (comparable data were not available for the HRMA database). There were no significant differences between the D&B respondents and non-respondents in sales or whether they were private or publicly owned. Respondents did tend to be from slightly larger companies (the mean number of employees for respondents was about 225 more than for non-respondents) and to have been in business longer (the mean of number of years in business was about 10 years more for respon- dents than for non-respondents). These findings are to be expected, since larger, more established companies would not only have more advanced com- pensation practices but also would have more staff available to complete the survey. For the HRMA database, the only comparative data available on non- respondents was industry sector, and for both samples chi square analysis found no significant differences between respondents and non-respondents in this variable. To ensure that the respondents from the two samples were com- parable and could be analyzed together, analyses of variances were conducted that revealed no significant differences in the major variables. Measures Compensation Strategy Compensation strategy components were measured using Gomez-Mejia's (1992) scale (see Appendix). In Gomez-Mejia's original study, the individual Spring 1996 Day: Compensation Program Goals 69 items used were averaged to fonn 15 scales, which were then analyzed using principal component analysis. The end result was one factor, with negative loadings indicating an "algorithmic" strategy (using more routinized and mechanistic pay practices) and positive loadings indicating an "experiential" strategy (using more flexible and participative pay practices). We initially rep- licated this analysis, but one single factor did not emerge. Further analyses of the 15 scales revealed very low reliabilities for our sample, although Gomez- Mejia's work showed acceptable reliabilities (ranging from .74 to .89). Because of these dramatically different results, which indicate that this scale may not be generalizable across populations, we conducted further analyses, using prin- cipal component analysis with varimax rotation, to detennine how these vari- ables were meaningful in the current data set. Because our replication of the previous analysis was so significantly dif- ferent, assumptions based on the earlier work about the constructs the scale measures were not relevant. This led us to choose principal component analy- sis, since our goal was not to define an underlying causal model among the variables, as factor analysis would do. Principal component analysis linearly summarizes the data into simpler components (Kim & Mueller, 1978), and we believed it would more adequately satisfy our mission of showing that differ- ing compensation strategy choices affect compensation program goals. Also, we wanted to use factor scores, and this method produces exact scores rather than estimates (Hair, Anderson & Tatham, 1987). Further, factor analysis us- ing maximum likelihood extraction was conducted on these data, resulting in substantially the same factor structure. Varimax rotation was used because of its ability to produce more stable results over different sets of data and to sim- plify the columns, thus making interpretation easier (Hair, Anderson & Tatham, 1987). Again, other rotation methods (quartimax and equimax) did not sub- stantially change the factor structure. Using principal component analysis, 11 meaningful factors emerged (us- ing the criteria of eigenvalues greater than 1.0), accounting for 61 % of the variance. To ascertain the reliability of these factors, or the degree of associa- tion between each component and its factor scale (Kim & Mueller, 1978), we conducted separate principal component analyses on the factors that had heavy loadings (greater than .4) on each component, as recommended by Carmines and Zeller (1979). All 11 of these analyses showed that a single phenomenon was being measured by the major variables of each component (they loaded on one factor, which accounted for 40% or more of the variance; Carmines & Zeller, 1979). Using factor scores, we then created 11 variables that represented compensation strategy choices (Table 1). Using factor scores also had the ad- vantage of providing independent variables that were uncorrelated. 70 Journal of Business Strategies Vol. 13, No. 1 Table 1 Factor Loadings for Principal Component Analysis of Gomez-Mejia's Compensation Strategy Variables 1 2 3 4 5 6 7 8 9 10 11 Component eigenvalue 5.42 3.43 2.52 1.96 1.74 1.61 1.55 1.42 1.27 1.19 1.16 Percentage of variance accounted for 14.3 8.8 6.6 5.2 4.3 4.2 4.1 3.7 3.3 3.1 3.1 Use a job-based system -.00 .32 -.00 -.00 -.00 -.00 -.00 -.00 -.59 -.00 -.00 Do not use a skill-based system -.33 .36 -.36 Emphasis on job evaluation -.65 Rewards based on contributions .51 .43 Individual over team emphasis .72 Emphasis on long term goals .61 Seniority not important .79 Emphasis not on short term goals -.56 Pay based on group performance .69 Large part of pay is variable .79 Employees should be risk takers .65 Corporate performance is a pay criterion .71 Internal equity is important goal .74 Try hard to achieve comparable pay across org. .83 Higher priority to interna than external equity .45 -.31 Large pay differentials based on performance .30 .43 .39 Provides perks to top management .80 Spring 1996 Day: Compensation Program Goals 71 Table 1 Factor Loadings for Principal Component Analysis of Gomez-Mejia's Compensation Strategy Variables cont'd 1 2 3 4 5 6 7 8 9 10 11 Base pay is not an important part of pay package .72 Base salary low relative to other types of pay .74 Benefits are not an important part of compensation .56 -.39 Benefit package is not generous .86 Special pay packages to top management .39 .55 .30 Incentives are important part .82 Incentives provide significant amount of pay .68 Bonuses provided often .67 Psychological needs not considered important .38 .53 Focus on monetary v. intrinsic rewards .72 Pay not egalitarian; special rewards to elite groups .72 Policies are not uniform across units .59 .30 Line managers can make pay decisions Pay information is not kept secret .30 .65 Pay policies do not require secrecy .71 We openly disclose pay development procedures .66 • Employee preferences are considered .35 .57 72 Journal of Business Strategies Vol. 13, No. I Table 1 Factor Loadings for Principal Component Analysis of Gomez-Mejia's Compensation Strategy Variables cont'd 1 2 3 4 5 6 7 8 9 10 11 Employees have say in pay policies .84 Pay decisions are not autocratic .71 Preferred salary position is above market .30 .66 Preferred benefit position is above market .83 Components: 1. Pay based on corporate performance 2. Internal consistency and equity 3. Participation in designing compensation programs 4. Benefits 5. Disclosure of pay practices a key strategy 6. Hierarchical rather than egalitarian pay 7. Base salary not a major component of pay 8. Pay based on individual performance 9. Pay not based on the job 10. Monetary rewards versus psychological rewards emphasized 11. Long term competitiveness Compensation Program Effectiveness Respondents were asked to rate the effectiveness of the current pay sys- tem for both exempt and nonexempt employees using a four-point scale, from highly effective to highly ineffective. Nine dimensions were measured for both exempt and nonexempt employees, and six factors emerged (see Table 2). Based on these factors, six averaged variables were created: compensation program ef- fectiveness in recruitment (a = .89), motivating and retaining top employees (a = .89), administrative manageability (a = .89), cost effectiveness (a = .91), retaining average employees (a = .89) and encouraging poor performers to leave (a = .90). Appropriateness of Effectiveness Measures Although we believe these compensation effectiveness measures are ap- propriate, they have some limitations which should be discussed. The first is that they are self-reported responses from compensation managers, who may have a stake in seeing positive relationships between the programs they have Spring 1996 Day: Compensation Program Goals 73 Table 2 Factor Loadings for Principal Component Analysis of Compensation Effectivess Variables 1 2 3 4 5 6 Component eigenvalue 6.17 2.48 2.03 1.55 1.31 1.09 Percentage of variance accounted for 34.3 13.8 11.3 8.6 7.3 6.0 Attracting large enough pool of exempt applicants .85 Attracting large enough pool of nonexempt applicants .86 Attracting highly qualified exempt applicants .76 Attracting highly qualified nonexempt applicants .82 Motivating top performance in exempt employees .82 Motivating top performance in nonexempt employees .74 Retaining top exempt employees .85 Retaining top nonexempt employees .77 Retaining average exempt employees .93 Retaining average nonexempt employees .93 Encouraging below average exempt employees to leave .95 Encouraging below average nonexempt employees to leave .95 Being a cost effective compensation system- exempt plan .86 Being a cost effective compensation system- nonexempt plan .90 Being administratively manageab1e- exempt plan .82 Being administratively manageab1e- nonexempt plan .78 Being easy to communicate to exempt employees .86 Being easy to communicate to nonexempt employees .80 Factor names: 1. Recruitment effectiveness 2. Administrative manageability 3. Motivation effectiveness 4. Encourage poor performers to leave 5. Retaining average employees 6. Cost effectiveness 74 Journal of Business Strategies Vol. 13, No. 1 designed and achievement of compensation objectives. However, the responses from the sample contain ample variance and means that do not reflect an ex- tremely strong positive bias (see Table 3 below, variables 5 through 10). If the respondents had greatly inflated their programs' benefits, there may have been more consistently positive results than were seen. Second, because many compensation managers ultimately have limited im- pact on the final compensation programs, they may not feel totally responsible for their outcomes and thus are likely to provide fairly objective responses re- garding compensation effectiveness. Top management can dramatically affect budgets, philosophies and programs, and the compensation manager does not have omnipotent power over the final package. Additionally, because of the confidentiality of the survey (questionnaires were returned directly to the researcher's university), there is little motivation for respondents to deliberately bias their responses. Therefore, it is unlikely that the respondents would have greatly overestimated the impact of their programs. Third, as in all self-report studies, the possibility of common method vari- ance should be addressed. Common method variance refers to the fact that "be- cause both measures come from the same source, any defect in the source con- taminates both measures" (Podsakoff & Organ, 1986, page 533). Common method variance is present when correlations between measures are not due to "true" relationships between the constructs but simply because the same re- spondents provide the measures for both constructs. One method of investi- gating whether this condition exists is Harmon's one factor test (Podsakoff & Organ, 1986), which consists of a factor analysis of all relevant variables. If a large degree of common method variance is present, one factor will emerge (Podsakoff & Organ, 1986; Miceli, lung, Near & Greenberger, 1991). Such an analysis was conducted on the compensation effectiveness perception and strategy variables in this sample. Fourteen factors emerged, with the first fac- tor (which, in cases of common method variance, would account for a major- ity of the variance) only accounting for 15% of the variance. Thus, common method variance is unlikely to be at issue here. Fourth, researchers (Balkin & Gomez-Mejia, 1987; Gomez-Mejia, Page & Tornow, 1982) have asserted that management opinions about effectiveness may actually be more valid indicators than "objective" data such as profitabil- ity, market share and shareholder value, since these measures are subject to a vast number of contaminating variables, including trends in the economy and other environmental factors. Indeed, researchers have argued that the "bottom line" of compensation effectiveness is its value as perceived by its users. Other researchers concur that self-report measures may, in some cases, represent more accurate descriptions than more objective measures (Howard, Maxwell, Weiner, Boynton & Rooney, 1980; Podsokoff & Organ, 1986). In the present study, since we are interested in the technical success of the compensation program, the only people with the breadth and depth of knowledge to adequately evalu- ate these dimensions are the compensation managers. Spring 1996 Day: Compensation Program Goals 75 Finally, since we were interested in_assessing the separate components of a successful pay system, we were limited in the number of objective measures that were available and within the scope of this study. Because of the previ- ously stated arguments. we concluded that the expert opinions of compensa- tion managers would be valid and appropriate for this study. As Podsakoff and Organ argue, "the practical utility of these types of measures makes them vir- tually indispensable in many research contexts" (page 540) and may be ac- ceptable if adequate controls, such as Harmon's one factor test, are reported for the data. While we believe that further research into these effectiveness constructs, using multiple measures from multiple constituents are essential, for this initial study in our research we believe that they are acceptable. Control Variables Four variables were analyzed as control variables, since past research has found relationships with these variables and compensation strategies and ef- fectiveness. These control variables are sales, size of the organization in num- ber of employees (Day, 1995; Gerhart & Milkovich, 1990; Gomez-Mejia, 1992), extent and process of diversification and life cycle stage (Balkin & Gomez-Mejia, 1990; Gomez-Mejia, 1992; Gomez-Mejia & Balkin, 1992). The magnitude of sales was evenly distributed across respondents: 33% reported sales of over $500 million; 32% reported sales of under $50 million. The mean number of employees was 1,253. The measures of diversification and life cycle stage (see Appendix) were taken from the work of Gomez-Mejia and Balkin (Balkin & Gomez-Mejia, 1990; Gomez-Mejia, 1992; Gomez-Mejia & Balkin, 1992) who had used Rumelt's (1974) earlier work on diversification. The four diversification ex- tent categories include: single product/service firm (95% or more of revenues are derived from a single product/service line); dominant product/service firm (may be several products/services but a dominant product/service accounts for 70% to 94% of revenues); related product/service firm (70% or less of rev- enues are derived from any single product/service); unrelated product/service firm. Most respondents (46%) judged their organizations to be dominant prod- uct; similar numbers judged their organizations to be either single product (26%) or related products (29%). Only one respondent judged the organiza- tion to be unrelated. Process of diversification included two categories: organizations that are vertically integrated with a commitment to an existing product/service and that expand primarily in that product/service area; organizations that prefer to ex- pand by acquiring new businesses, even if unrelated to existing product/ser- vice lines. Most respondents (77%) believed their organizations were vertically integrated. Twenty-three percent believed that their organizations expanded outside current product/service lines. Similarly, the life cycle stages of organizations were analyzed. These stages included: start-up (a small company five or fewer years old, run by an 76 Journal of Business Strategies Vol. 13, No. 1 entrepreneur); growth (sales are growing at 20% or more annually, with tech- nology and company structure changing); mature (growth is stable and slow and products/services are familiar to most prospective users); decline (growth is declining). Most respondents (64%) judged their organizations to be mature. Twenty-seven percent believed their organizations were growing. Only 2% saw their organizations as start-up and only 7% saw them as declining. Data were analyzed using hierarchical regression analysis for equations for each of the six dependent variables. The five control variables were en- tered in the first step of the equations. In order to determine the unique vari- ance accounted for over and above the control variables, the eleven compen- sation strategy variables were entered in the second step of the equation. Results Means, standard deviations and the correlation matrix for the major vari- ables can be found in Table 3. In the first step, for all six dependent vari- ables, the control variables entered as a block did not account for signifi- cant variance (Table 4). In the second step, the proposition was confirmed for five of the six dependent variables (Table 4). As predicted, there was variation in the specific strategy variables that affected each compensation effectiveness goal, and the compensation strategy variables accounted for at least 20% of the variance for the five measures of effectiveness perceptions in recruitment, motivation, administration, cost and encouraging poor per- formers to leave. No compensation strategy variables were significantly re- lated to the perceived effectiveness of the compensation system in retaining average performers. Two compensation strategies showed significant relationships with four of the six dependent effectiveness perception measures. Participation in plan design was predictive of effectiveness in recruitment, motivation, cost effec- tiveness and encouraging poor performers to leave the organization. An emphasis on internal consistency was predictive of effectiveness perception in recruitment, motivation, administration and cost effectiveness. An emphasis on paying for individual performance was predictive of three effectiveness perception measures: motivational effectiveness, cost effectiveness and en- couraging poor performers to leave. Contemporary thinking about compen- sation practices would predict that participation and incentive pay, since they are innovative and nontraditional approaches, would be associated with ef- fectiveness. However, it is surprising that an emphasis on internal consis- tency, which has been maligned because it has been seen as unsupportive of organizational goals (Lawler, 1990; Schuster & Zingheim, 1992), would so consistently predict effectiveness perceptions. This finding may indicate that internal consistency is, as many compensation analysts have claimed for decades, a critical component of compensation effectiveness, regardless of organizational strategy. Table 3 CI':l Means, Standard Deviations and Correlation Matrix 't:l.,..... ::s Mean I7C/ SD 1 2 3 4 5 6 7 8 9 10 -1,0 Control Variables 1,0 0'. I. Sales 2.80 1.49 1.00 2. Number full-time employees 1248.28 2460.71 .46 1.00 3. Diversification extent 2.04 .75 .13 .10 1.00 4. Life cycle stage 2.76 .60 .12 .01 -.17 1.00 t:l Compensation effectiveness variables ~ 5. Recruitment effectiveness 3.14 .62 .26 .18 .04 -.01 1.00 ~6. Motivational effectiveness 2.84 .60 .16 .12 .08 -.10 048 1.00 7. Administrative manageability 3.11 .68 .02 .01 .15 .02 .32 .32 1.00 ~ ('; _. ~ f1Q ...... '0 '0 0'1 Step Ii Control Variables Sales Number full-time employees Diversification extent Diversification process Life cycle stage Step 2: Compensation Strate&ies Participation in design of plan Pay for corporate perfonnance Pay disclosure Long tenn competitiveness Monetary rewards Low emphasis on base pay Pay for individual performance Benefits major part of compo Skills versus jobs Internal consistency Hierarchical rewards *** p < .001 ** P < .01 * P < .05 + P < .10 Beta -,035 .080 .004- ,051 .089 -.059 .205+ .031 -.055 .086 -.100 .080 .195+ .118 .180+ -.077 .1.R2 .01 .14 Beta AR2 .03 -.035 .053 tJ .172+ ~ -.068 ~ .015 ~ ~ .20* ;::: .275** ~.... -.090 c' ;::: -.095 "tl .107 <3 OQ -.123 i:1 .065 ~ .283*** (J <::l .013 l::l '"'- "" -.063 ,lIS -.113 -J '0 80 Journal of Business Strategies Discussion Vol. 13, No. 1 Our proposition, that compensation strategies will differentially affect the objectives of an effective compensation system, was supported. However, some of the relationships suggested in the introductory section of this paper were not supported. For example, we suggested that internal consistency may be positively related to retention but not related to recruitment, since those out- side the organization would not understand the internal pay structure. Our find- ings were opposite this prediction. While we did not find relationships between internal consistency and retention, we found that it was predictive of recruit- ment effectiveness perceptions. We also suggested that pay secrecy would posi- tively affect administrative effectiveness because it requires less salary com- munication, but secrecy would negatively affect motivation, since employees would be unaware of potential pay increases. We did not find these relation- ships. We also believed that pay at risk would be positively associated with motivation effectiveness, which these data supported; basing pay on both cor- porate and individual performance was positively related to motivation effec- tiveness. However, we had suggested that paying for performance would be negatively associated with recruitment effectiveness, which was not found in these data. Predictors of Recruitment Effectiveness Compensation systems are in part designed to attract acceptable candidates to the organization. Most frequently this is conceptualized as offering a pack- age of pay and benefits, along with appropriate working conditions and other psychological accouterments that will be appealing to the appropriate type and quality of candidate. We found that four compensation strategies were signifi- cantly related to perceptions of recruitment effectiveness. First, as would be an- ticipated, was an emphasis on benefits. Since the value of benefits as a recruit- ment device has increased dramatically over the last few years (ACA, 1993), this finding is not surprising. Of some interest is that an emphasis on benefits did not predict any other effectiveness perception variables. Second, participa- tion in pay program design was predictive of recruitment effectiveness, perhaps indicating that active employee input produces pay programs that are more at- tractive, not only to current incumbents, but also to job candidates. Third, the relationship between pay disclosure and recruitment may indicate that the more pay information job candidates are able to gather, the more they are apt to per- ceive that the organization offers adequate financial incentives and is open with key job information. Finally, an emphasis on internal consistency was predic- tive of perceived recruitment effectiveness. While it is doubtful that external candidates are aware of the level of internal equity present in the organization, it may be that an emphasis on internal equity is positively correlated with pay level. Recent authors have claimed that traditional point-factor job evaluation programs, which are found in pay programs emphasizing internal equity, are Spring 1996 Day: Compensation Program Goals 81 more likely to pay over the market (Lawler, 1990; Schuster & Zingheim, 1992) and thus may be more attractive to recruits. Alternatively, if internal consis- tency is coupled with pay disclosure, this may communicate to candidates that the organization has an organized and equitable pay structure that rewards ap- propriately for skills and abilities that the organization values. Motivational Effectiveness Four compensation strategies were predictive of perceived motivational ef- fectiveness. Participation in design of pay programs, an emphasis on long-term competitiveness and paying for individual performance were associated with pay programs that were seen as motivational. These findings reflect current professional thinking regarding compensation programs and their ability to elicit appropriate employee behaviors (Lawler, 1990). Individual incentives, when designed properly, have been found to be motivational (Gerhart & Milkovich, 1992), and participation in plan development encourages acceptance and understanding of the plan (Crepanzano & Folger, 1989; Milkovich & Newman, 1993), which may result in increased motivation. It is interesting that an emphasis on internal consistency was also predic- tive of motivational effectiveness. This may refute the current philosophy that complicated point-factor job evaluation programs, typically used to establish internal equity, support the status and thus detract from the accomplishment of key organizational goals (Emerson, 1991; Lawler, 1990; Schuster & Zingheim, 1992). Indeed, these results indicate that internal consistency may assist the compensation system in motivating employees toward goal accom- plishment. Administrative Manageability Emphases on long term competitiveness, internal consistency and egali- tarian rather than hierarchical pay were found to be predictive of administra- tive manageability. These predictors indicate that emphasizing both competi- tiveness and internal equity are associated with a more easily managed com- pensation program, facts with which most compensation managers would prob- ably concur. Less emphasis on hierarchical relationships, especially in perqui- sites and status-related pay and benefit components, was associated with a pay system that is easier to administer, perhaps through fewer grade levels or oth- erwise simpler programs. Effectiveness in Retaining Average Performers No compensation strategy variables were predictive of this measure. How- ever, the lack of associations with the main effects of compensation strategies may be because the variable itself is not seen as a 9ritical goal of compensa- tion programs. While retaining employees who "meet expectations" is theoreti- cally an important compensation goal, it may not be one that is very promi- nent in the minds of compensation managers or is actively promoted. 82 Journal of Business Strategies Vol. 13, No. 1 Encouraging Poor Performers to Leave Participation in plan design and individually based incentive plans were predictive of encouraging poor performers to leave the organization. The lat- ter finding supports commonly held beliefs that incentive pay will not be at- tractive to poor performers, since they are unable to reap the same rewards as others while continuing to shirk or perform poorly (Gerhart & Milkovich, 1992; Lawler, 1990), and will thus be more likely to leave. Participation in design of the compensation program may create a program that maximizes perceived fairness and equity based on performance inputs, which sends a message to those who do not perform as well as others that their behavior will not be well rewarded. Recommendations for Compensation Managers Although further research is needed, some suggestions should be con- sidered regarding the development of compensation programs in specific or- ganizations based on this study. First, different compensation objectives were seen to be supported by different compensation strategies, and a careful analysis of the intent and purpose of compensation in the context of the entire organization's strategic plan should be done to ensure that pay strat- egies match the objectives for which the program is being designed. For example, does the organizational strategy demand that highly skilled employ- ees be recruited from a competitive labor market? If so, those compensa- tion strategies that affect recruitment should be examined to ensure that they are designed adequately. For example, is the benefits package competitive for the segment of the labor force the organization wishes to attract? Would employee participation in pay plan design produce a compensation package that would be attractive to job candidates? Alternatively, if the organization is geared toward improving productivity and cost-cutting, compensation strat- egists may carefully evaluate whether the program includes those strategy choices that seem to positively predict motivation and cost effectiveness. Can employees participate in plan design? Is the compensation program internally equitable? Does the organization pay for performance, both at the corporate and individual level? Ensuring that these compensation strategies are utilized may assist in accomplishing the organization's goals of improving produc- tivity and reducing costs. Two strategies, participation in plan design and internal consistency, were associated with four key compensation goals, and thus may prove to be criti- cal compensation strategies across a number of programs. While there has been much recent negative opinion about establishing internal equity, the respondents surveyed in this study believe that it is associated with a significant number of desirable outcomes. Indeed, since the basic psychological makeup of hu- man beings will probably not change dramatically, we may assume that basic equity judgments will continue to' be made. Thus, compensation strategies should take internal equity into consideration. Spring 1996 Day: Compensation Program Goals 83 Our results also indicate that participation in pay program design may be critical to many facets of compensation effectiveness. Thus, compensation man- agers may want to carefully consider to what degree and extent employees can become involved in this process. While probably not contributing to adminis- trative effectiveness (as indeed our results indicate), establishing compensation programs based on employee input may be worth the time and trouble since they may enhance other important goals of the compensation program. Compensation managers should consider those compensation strategies that showed limited relationships with effectiveness: an emphasis on benefits and pay disclosure. The extremely high cost of benefits in today's organiza- tions is well known and should be carefully considered in light of these re- sults. If benefits serve primarily a recruiting function, then the benefits pack- age needs to be designed so that it obtains maximum recruitment impact. While a competitive benefits package may have a dramatic effect on other non- compensation issues, such as a sense of security or organizational commitment, it may not produce broad positive compensation effects. Similarly, pay disclosure was related only to recruitment effectiveness per- ceptions. Since pay disclosure involves increased communication, potential challenges to fairness and associated complaints and grievances, compensation managers should carefully consider whether a policy emphasizing secrecy or disclosure is appropriate. Once information about a compensation program is disclosed, it will always be disclosed. Especially if disclosure does not result in added motivation or retention, it may be more prudent to lean towards a policy of non-disclosure. Finally, compensation managers should also take note of those compen- sation strategies that were not useful in predicting compensation effectiveness. For example, an emphasis on monetary over psychological rewards, low em- phasis on base pay and an emphasis on skills versus jobs were not associated with any of the effectiveness measures in this study. These compensation man- agers place greater value on other strategies as the real drivers of compensa- tion effectiveness. Recommendations for Further Research In interpreting these results, it must be kept in mind that our sample was not representative of all US firms. First, our analyses showed that these firms tended to be somewhat larger and older than the nonrespondents. They also reported themselves as being mainly dominant product/service and vertically integrated. Additionally, they are Midwestern firms, often accused of being more conservative in outlook and practices than the rest of the country. Whether these results would generalize to a broader population is not known. The major limitation of this study is its use of compensation manager rat- ings of compensation effectiveness. Further research should attempt to mea- sure the pay effectiveness dimensions from multiple constituencies and, if pos- sible, objective sources. For example, recruiters should be surveyed regarding 84 Journal of Business Strategies Vol. 13, No. 1 the effectiveness of the compensation program. The relationship of compensa- tion strategies to cost effectiveness could be examined through analyzing ac- tual payroll and costs of activities associated with compensation. Administra- tive effectiveness could be ascertained through infonnation obtained from line managers who must administer the program on a day-to-day basis. Further fonnulation and definition of these constructs are needed and future research should focus on these. There also may be other compensation effectiveness di- mensions that would be of interest, such as ability of the system to withstand legal challenges or support labor-management relationships. Further, dimen- sions could be expanded to include sub-dimensions. For example, recruitment effectiveness could focus on ease of recruiting specific job types, such as R&D positions, clerical positions, etc. Finally, the compensation strategy measures used in this study were those used previously by other researchers, but we found dramatically different sta- tistical results in our sample. Future research should compare and contrast this scale with others to ascertain its validity and reliability, and/or to develop an- other more useful and generalizable measure of compensation strategies. References American Compensation Association. Certification Course 7: Fundamentals of Em- ployee Benefit Pr0i:rams. Scottsdale, AZ: American Compensation Association, 1993. Balkin, D.B., and L.R. Gomez-Mejia. "A Contingency Theory of Compensation Strat- egy." Strategic Management Journal 8 (1987): 169-182. Balkin, D.B., and L.R. Gomez-Mejia. "Matching Compensation and Organizational Strategies." Strategic Management Journal II (1990): 153-169. Brown, C. and J. Medoff. "The Employer Size-Wage Effect." Journal of Political Economy 96 (1989): 1027-1053. Bureau of National Affairs. Changing Pay Practices: New Developments in Employee Compensation. Washington, D.C.: Bureau of National Affairs, 1988. Carmines, E.G. and R.A. Zeller. Reliability and Validity Assessment. Beverly Hills, CA: Sage Publications, 1979. Crepanzano, R. and R. Folger. "Referent Cognitions and Task Decision Autonomy: Beyond Equity Theory." Journal of Applied Psychology 74 (1989): 293-299. Day, N.E. "An Investigation into Pay Level Policy: Relationships with Pay Structure Characteristics and Effectiveness." Paper presented at the annual meeting of the Mid- west Academy of Management, St. Louis, MO, March, 1995. Spring 1996 Day: Compensation Program Goals 85 Emerson, S. "Job Evaluation: A Barrier to Excellence?" Compensation and Benefits Review 23 (1991): 39-51. Gerhart, 8., and G.T. Milkovich. "Organizational Differences in Managerial Compen- sation and Financial Performance." Academy of Management Journal 33 (1990): 663- 691. Gerhart, B., and G.T. Milkovich. "Employee Compensation: Research and Practice." In M. Dunnette & L.M. Hugh (Eds.), Handbook of Indystrial/Organizational Psychol- Qgy. 2nd Edition, 481-569, Palo Alto, CA: Consulting Psychologists Press, 1992. Gomez-Mejia, L.R "Structure and Process of Diversification, Compensation Strategy and Firm Performance." Strategic Management Journal 13 (1992): 381-397. Gomez-Mejia, L.R. and D.B. Balkin. Compensation, Qrganizational StrateflY and Firm PerfOrmance. Cincinnati: South-Western Publishing Co., 1992. Gomez-Mejia, L.R., R.C. Page, and W.W. Tornow. "A Comparison of the Practical Utility of Traditional, Statistical and Hybrid Job Evaluation Approaches." Academy of Management JQurnal 25 (1982): 790-809. Hair, J.F., RE. Anderson, and RL. Tatham. Multivariate Data Analysis, 2nd Edition. New York: Macmillan Publishing Company, 1987. Hills, F.S., T.J. Bergmann, and V.G. Scarpello. Compensation Decision Making. Fort Worth: The Dryden Press, Harcourt Brace College Publishers, 1994. Holzer, H.J. "Wages, Employer Costs, and Employee Performance in the Firm." In:. dustrial and Labor Relations Review 43 (1990): 147S-164S. Howard, G.S., S.E. Maxwell, RL. Wiener, K.S. Boynton and W.M. Rooney. "Is a Behavioral Measure the Best Estimate of Behavioral Parameters? Perhaps Not." Ap- plied Psychological Measurement 4 (1980): 291-311. Kim, J.O. and C.W. Mueller. Factor Analysis: Statistical Methods and practical Issues. Beverly Hills, CA: Sage Publications, 1978. Krueger, A.B. "The Determinants of Queues for Federal Jobs." Industrial and Labor RelatiQns Review 41 (1988): 567-581. Lawler, E.E. m. Pay and OrganizatiQnai Effectiveness: A PSychoIQgical View. New York: McGraw-Hill, 1971. Lawler, E.E. m. Strategic Pay: Aligning Organizational Strategies and Pay Systems. San Francisco: Jossey-Bass Publishers, 1990. Mahoney, T.A. Compensation and Reward Perspectives. Homewood, IL: Richard D. Irwin, 1979. 86 Journal of Business Strategies Vol. 13, No. 1 Miceli, M.P., I. Jung, J.P. Near, and D.B. Greenberger. "Predictors and Outcomes of Reactions to Pay-for-Performance Plans." Jgurnal gf Applied Psychology 76 (1991): 508-521. Milkovich, G.T., and J.M. Newman. Compensation, 4th Edition. Homewood, IL: Irwin, 1993. Podsakoff. P.M., and D.W. Organ. "Self Reports in Organizational Research: Problems and Prospects." Journal of Management 12 (1986): 531-544. Rumelt, R.P. Strate~y, Structure and Economic PerfQrmance. Boston: Division of Re- search, Harvard Business School, 1974. Rynes, Sara L., and Alison E. Barber. "Applicant Attraction Strategies: An Organiza- tional Perspective." Aca