The Strategic Role of Gainsharing* ROBERT M A N G E L Temple University, Philadelphia, PA 19122 M I C H A E L USEEM University of Pennsylvania, Philadelphia, PA 19104 Gainsharing is reported to greatly enhance organizational performance, but there is little empirical research that explains how gainsharing works and why companies adopt it. We employ a human resource control framework to examine the role gainsharing plays within the organization. Our analysis of 802 organizations suggests that gainsharing plans are used to enhance employee participation in organizations that employ market-based control methods. Organizations that use bureaucratic or clan controls are not likely to adopt gainsharing; they are more likely to use participation initiatives that do not involve a group bonus. I. Introduction Large group incentives such as gainsharing, profit sharing, and employee stock ownership plans (ESOPs), have been enjoying a resurgence as employers strive to increase productivity (Blinder, 1990; Hammer, 1988; O'Dell, 1987). Mounting evidence suggests that these plans can, under the fight conditions, positively affect productivity and organizational performance (for reviews, see Blasi, 1988; Blinder, 1990; Bullock and Lawler, 1984; Kruse, 1993). However, research also suggests that there is substantial variety in both plan design and impact (Mitchell et al., 1990; O'Dell, 1987; Towers Perrin, 1990). For these plans to reach their full potential, it is important to understand how they work and under what conditions they best work, yet there is little research that examines these questions (Hammer, 1988; Hanlon et al., 1992; Mitchell et al., 1990; Weitzman and Kruse, 1990). Gainsharing in particular has received a great deal of attention for its potential to increase worker productivity. Gainsharing utilizes operational or productivity measures of performance, which are usually more controllable by employees than the financial measures of profit sharing or the market measures of ESOPs (Hammer, 1988; Lawler, 1981; Weitzman and Kruse, 1990). As such, gainsharing is seen to be the largegroup incentive system most directly focused on enhancing employee motivation (Hammer, 1988; McKersie, 1986). The precise way that gainsharing increases productivity is not well understood, although research identifies employee participation in decision making as a key factor in its success (Blinder, 1990; Hammer, 1988; Mitchell et al., 1990). Some form of JOURNAL OF LABOR RESEARCH

Volume XX1, Number 2 Spring 2000

328

JOURNAL OF LABOR RESEARCH

employee involvement usually accompanies the adoption of gainsharing, and the increases to firm productivity are thought to come primarily through employee participation. The participation thesis focuses specifically on the improved decision making resulting from increased information flows. Employee suggestion systems and other forms of direct participation in decision making encourage employees to provide productivity-enhancing suggestions. The contingent group bonus, as one part of the compensation package, is designed to reinforce this employee participation in decision making, as opposed to the workers' "primary" job efforts (Hammer, 1988). Given that employee participation is a productivity-enhancing initiative itself, one might question whether the group bonus of gainsharing adds to productivity above and beyond the participation component. Hammer (1988) argues that gainsharing plans redefine the labor-effort exchange between employee and organization and that employee benefits (e.g., intrinsic satisfaction) gained through participation initiatives are not substitutes for pay in the redefined labor exchange. Rather, participation is viewed as work above and beyond the employee's normal job, and thus a bonus is required to reward this additional effort. Of course, organizations may use a means other than the group bonus to reward employee participation. A company could reward employees individually for their participative efforts or might attempt to structure the employment relationship such that the employees do not believe that they need specific rewards for their participative efforts. It is these alternative options for rewarding employee participation that set the framework for our arguments here. In order to examine the strategic role of gainsharing within the organization, we conceptually and empirically compare the efficiency of gainsharing bonuses with other potential options for rewarding participation. More specifically, we evaluate the organization's choice to adopt gainsharing by examining the relative fit of gainsharing plans and other participation-oriented initiatives with several HR practices that indicate the organization's HR control approach. Increasingly, researchers have been studying the way different HR practices interrelate and how bundles of these HR practices fit with their organizational and environmental contexts (Baird and Meshoulam, 1987; Cappelli and Singh, 1992; Dyer, 1985; Mahoney and Deckop, 1986). These bundles of HR practices work to control employee behavior by mediating the exchange between the individual and the organization. Although firms may choose different sets of practices to control behavior, research increasingly suggests that there must be a substantial degree of internal consistency within a given set of practices. The underlying assumption is that there is no one best way to manage human resources, but that HR practices must be mutually consistent and supportive in order to be effective and efficient. II. Control Theory and the Employment Relationship Many control frameworks have their roots in the transaction cost economics (TCE) theory of Williamson (1981), which models the efficiency of different forms of transactions. TCE focuses on the informational characteristics of contexts as a means for

ROBERT MANGEL and MICHAEL USEEM

329

determining which organizational forms and which types of employment relationships are most efficient at mediating the transactions among organization members. HR policies and practices are one important way that organizations can mediate the exchange between individual and organization. Strategic HR researchers have thus applied economic control frameworks to explore how firms choose different HR practices and bundles of HR practices (Beer et al., 1984; Snell, 1992). To illustrate the basic TCE approach, consider the information requirements of three basic control forms: output, behavior, and input control. Output-control systems by their nature require clear knowledge of performance standards to be efficient. In situations where it is easy, and hence inexpensive, to measure performance, output-oriented control systems can be efficiently used. Likewise, in situations where there is clear knowledge of the transformation process from inputs to outputs~ it is relatively inexpensive to monitor behaviors that are consistent with high performance. In situations where both performance standards and the transformation process are ambiguous, input controls are the only choice for the organization, assuming there is some knowledge of the skills needed to perform the work (Snell, 1992). Even if information characteristics would support more than one type of control system, there are several factors that lead us to expect some congruence among HR practices. For example, developing information systems to measure employee performance requires the investment of significant organizational resources. If firms invest in these information systems to support output control, cost efficiency concerns would generally dictate that other initiatives be designed to take advantage of this investment when possible. Moreover, as current trends illustrate, basic characteristics of the employment relationship and organizational structure put strong constraints on the nature of specific control systems that may be used in an organization. Recently, the employment relationship has been moving away from long-term, bureaucratic relationships, where employees are tightly tied to organizations, toward more flexible arrangements, where employees are less firmly attached (Pfeffer and Baron, 1988). At the same time, many organizational structures are evolving into flatter and leaner forms, with the hope of operating more efficiently and responding more quickly to rapidly changing environmental conditions (Kanter, 1989; Useem, 1993). As hierarchies shrink and job security decreases, it becomes more difficult for organizations to rely on promotions to motivate employees (Osterman, 1996). Moreover, the flatter hierarchies lead to wider spans of control, which diminish the capacity of organizations to closely supervise the behavior of employees (Eisenhardt, 1988). Organizations that have short job tenures and flat structures need to augment or replace career ladders with other reward mechanisms, and these same organizations also need mechanisms to substitute for the reduced ability of supervisors to directly monitor employee behavior. Relying more on output-control systems, such as performance pay, is one option. Providing yearly performance-based bonuses can partly substitute for the motivational value of potential promotions (Cappelli, 1997; Useem,

330

JOURNAL OF LABOR RESEARCH

1996). Whereas flatter firms may have limited promotional opportunities, they still have the flexibility to reward employees with bonuses. Yearly bonuses also provide a timely payback that fits well with decreased job tenure. Likewise, the output-control systems, such as productivity bonuses, provide an efficient way of controlling employees in organizational contexts with large spans of control, where it is difficult for supervisors to directly monitor the behavior of subordinates. Hence, strategic choices about organizational design and general employment practices (levels of hierarchy, perceptions of job security) have implications for control systems and may lead to congruence pressures beyond both simple cost considerations and the nature of employees' work. For example, even in situations where the nature of work might be suitable for behavior-based control (i.e., the nature of transformation is well understood) organizations with large spans of control may not have enough supervisors to effectively monitor behavior. While there is certainly opportunity for variance within individual organizations, basic issues like firm structure and propensity to downsize set constraints at the organizational level within which individual HR practices function. These constraints effectively define an organization-level control approach that provides the context within which gainsharing and other participative initiatives must operate. Understanding the basic nature of control at the organization-level is an essential part of understanding the strategic role of gainsharing and other participation initiatives. Three organization metaphors, markets, bureaucracies, and clans (Ouchi, 1980; Beer et al., 1984), capture the nature of output, behavioral, and input control, respectively, at the level of the organization. These metaphors present a framework for understanding how different HR initiatives fit together to mediate the exchange between individual and organization, including the specific pay for participation exchange that is at the heart of the model presented herein. Since we are focusing on rewarding participatory behavior such as suggestion making, and not the employee's primary job, the question of fit does not concern the meterability of the employee's primary tasks. Rather, the question is how do the different options for rewarding participation fit with the overall organizational control approach of markets, bureaucracies, or clans. In general, market-based control appears to be the approach that is best suited to gainsharing plans. Market control describes the emerging form of employment relationship that is typified by relatively short employment tenures, the flattening of hierarchies, and the need to manage many relationships across boundaries. Given the flat hierarchies and short employment tenures, the market approach is characterized by the principle of explicit and immediate exchanges between an organization and its members (Beer et al., 1984). With employee expectations of short organizational tenure and limited promotional opportunities, output-control approaches, such as performance pay, provide the timely payback consistent with market forms. Likewise, output-oriented performance pay does not put excessive demands on supervisors, who, in market forms of organization, often have a large number of subordinates to supervise. Middle management productivity incentives are one key indicator of market control, and one of particular relevance to gainsharing. Although companies have readily available firm-level performance measures, many do not have measures developed at

ROBERT MANGEL and MICHAEL USEEM

331

the plant or profit-center level. The use of middle management incentives indicates the existence of performance measures at the plant or profit-center level, and such measures are often used for operationally based gainsharing plans (Towers Perrin, 1990). The use of middle management productivity incentives also indicates the use of performance pay as an alternative to promotion up the corporate ladder. Firms that rely more on productivity bonuses are likely to emphasize timely payback to employees for their participatory efforts, and an output-oriented group bonus is likely to fit well with their overall control approach. Whereas employees' primary work responsibilities may be rewarded by individual incentives, employees are rewarded as a group for their participation based on a measure of their group-level performance or "output." If these market-control firms have flatter hierarchies, and the attendant wide spans of control, the group bonus presents a relatively efficient mechanism for specifically rewarding employee participation in decision making. The alternative of supervisors directly monitoring employee participatory efforts becomes less efficient as the spans of control increase, and the ability to use potential promotion as a motivational tool decreases as job ladders collapse. Hence: Hypothesis la: Firms that have adopted middle management productivity incentives are likely to adopt gainsharing plans. Firms that do not use market-based control may still choose to adopt some form of employee participation, but these firms are more likely to use mechanisms other than a group bonus to reward employee participation, as we discuss below. Conversely, firms with market-based control are not likely to adopt participatory initiatives unless they also adopt gainsharing, since the group bonus is likely to be the most efficient means of rewarding participation for these finns. The second half of Hypothesis 1 thus concerns the proposition that firms with market-control systems are not likely to adopt forms of participation that do not involve a group bonus. We use the term "pure participation" to refer to situations in which firms adopt some form of participation, e.g. employee involvement teams, supervisory training in participation, without also adopting a group-level bonus. Hypothesis lb: Firms that have adopted middle management productivity incentives are not likely to adopt "pure participation" systems. The importance of fit between the general control systems and participatory programs is also strongly illustrated by examining the mismatch between gainsharing and traditionally bureaucratic firms. The term "bureaucratic organization" is used to describe firms that utilize HR and organizational practices consistent with behavioral control. A bureaucratic organization operates according to a system of hierarchical surveillance, evaluation, and direction (Ouchi, 1980). A well-developed hierarchy provides a clear chain of command, and job ladders provide employees with the motivation to remain with the organization (Pfeffer and Cohen, 1984; Williamson, 1975). Such organizations also control employees by defining jobs narrowly, providing close supervision, and tying wages to jobs rather than to individuals (Beer et al., 1984).

332

JOURNAL OF LABOR RESEARCH

Bureaucracies are less likely to use gainsharing as a means to foster participation for several reasons. Since immediate supervisors are traditionally the conduit for participation in bureaucracies (Beer et al., 1984), supervisors may reward employees for their contributions, on an individual basis, through promotions, raises, and other rewards. The relatively low spans of control in bureaucracies facilitate the direct observation of participation, and so there is less need for the output control that group bonuses provide. Furthermore, supervisors and managers in bureaucracies are likely to strongly resist the adoption of gainsharing. Managerial resistance to participation is frequently cited as a reason for the slow proliferation of participation plans, as managers are thought to view participation as a threat to their job security and as extra work (Klein, 1984; Ogden, 1992). Given the prominent role of managers and direct supervision in bureaucracies, such organizations are likely to be particularly resistant to the gainsharing forms of participation. Participation in gainsharing plans often involves suggestion systems and employee evaluation committees (Blinder, 1990; Hammer, 1988), which generally bypass the power of the supervisor. Furthermore, managers may resist tying employee pay to firm-level performance because they fear that firm-level performance pay will stimulate employee demands for increased influence (Ogden, 1992). Bureaucratic firms are thus likely to resist the adoption of gainsharing because both the type of participation and the nature of the group bonus are at odds with the traditional corporate hierarchy. Although several organizational characteristics may point to bureaucratic control, such as low spans of control or narrowly defined jobs, a formal career ladder is one initiative that clearly indicates the organization's reliance on potential promotion as a motivational tool (Osterman, 1996; Pfeffer and Cohen, 1984). In organizations where career ladders are a central part of the HR strategy, management is likely to use the hierarchy as the primary method for both monitoring and rewarding participation. The presence of career ladders would thus be inconsistent with the gainsharing system of rewarding employee suggestions with a group bonus. Organizations with strong career ladders are also likely to direct efforts at increasing participation toward the supervisor, since the supervisor is the commonly used conduit for participation, hence: Hypothesis 2a: Finns that have adopted career ladders are adopt gainsharing.

not

likely to

Hypothesis 2b: Firms that have adopted career ladders are likely to adopt "pure participation" plans. Hypothesis 2c: Firms that have adopted career ladders are likely to adopt "pure participation" initiatives that emphasize the supervisory role, such as supervisory training in greater worker participation. Substantial goal incongruence between employee and employer is a common assumption underlying market and bureaucratic control methods as well as the traditional transaction cost framework. The goals of the organization and the individual are

ROBERT MANGEL and MICHAEL USEEM

333

assumed to be sufficiently at odds so as to require formal systems for mediating the exchange between employee and organization. However, a growing body of literature contends that the interests of the employee are often closely aligned with those of the employer, or at least that it is possible for organizational control systems to bring employee interests more closely in line with organizational goals. For example, Van Dyne et al. (1994) define a covenantal employment relationship, characterized by shared values, organizational identification, and trust. Employees are motivated more by intrinsic issues and less by financial concerns. Similarly, Fox and Hamilton (1994), Davis et al. (1997), and others have conceptualized the role of managers as stewards, as opposed to agents. Stewardship theory defines situations in which managers are motivated by organizational as opposed to individual goals, because the managers' interests are aligned with the interests of the owners (Davis et al., 1997). A common theme in each of these approaches is that substantial goal divergence is not always an accurate assumption; there may be times when an individual's personal interests motivate behavior that is in fact in line with organizational goals. Ouchi (1980) was one of the first to argue that substantial goal incongruence is not universal, and that certain organizational actions may work to reduce the level of goal incongruence itself. He explains that many Japanese firms, for example, operate by hiring inexperienced workers and socializing them to accept company goals. These so-called clan forms of organization thus have little need for mechanisms to mediate the exchange between individual and organization, because the employee's natural (or socialized) inclination is to behave in ways that are consistent with the organization's objectives. Rather than controlling outcomes or behaviors, clan-control organizations focus on selecting employees who already possess the appropriate values and attitudes, and clan organizations work to further instill the appropriate values. There is then less need to control behaviors or outcomes after the fact. Clan control thus represents a third option for rewarding employee participation. In situations where employees feel that their interests are largely aligned with the organization, there may be less of a need to directly compensate employees for each effort they put forward. If these clan organizations are somewhat successful in controlling employees' primary job efforts through value alignment, they are also likely to foster an environment in which employees do not expect a specific and immediate monetary reward for their participatory behavior. Employees are likely to feel that by helping the organization through suggestion making and other participatory efforts, they are helping to further goals in which they believe. To a certain degree, they believe that by helping the organization they are helping themselves. In situations of substantial value alignment, a group bonus designed to reward participation is likely, at minimum, to be redundant. Since group bonuses average about seven percent of worker salaries (Towers Perrin, 1990), the group bonus of gainsharing represents a substantial and unnecessary expense if employees are already inclined to participate. Organizational selection approaches are one key indicator of clan-control efforts (Beer et al., 1984; Snell, 1992). Organizations attempt to build a work force with

334

JOURNAL OF LABOR RESEARCH

the appropriate values through intensive screening efforts, and more specifically, by screening for values, hence: Hypothesis 3a: Firms that intensively screen applicants, and those that screen for values, are not likely to adopt gainsharing plans. Hypothesis 3b: Firms that intensively screen applicants, and especially those that screen for values, are likely to adopt "pure participation" systems. III. Research Design

Sample. Towers Perrin, a human resources consulting firm, and the Hudson Institute, a research firm, conducted a survey in 1990 that examined a wide variety of human resource practices, with an emphasis on employers' responses to shifting demographics. The survey was part of a follow-up study to the Hudson Institute's original Workforce 2000 study. The survey administrators sent questionnaires to senior executives responsible for human resources at almost 3,000 organizations, and 849 organizations responded, for a response rate of approximately 28 percent. After listwise deletion for missing variables, a total of 802 respondents remain in the study, representing approximately 27 percent of the initial sample. The strategy behind the sample is to identify as many large organizations across the U.S. as possible. The focus on large organizations is helpful as these firms are more likely to have formal HR mechanisms. We do not have access to the mailing list for the survey, so we cannot formally identify the response bias. The mean organizational size of approximately 8,275 employees indicates a skew toward larger organizations - - the sample average is close to the average number of employees for Fortune 1000 companies in the 1990s. The sample is also designed to cover a wide variety of organizations. The respondents represent a wide variety of industries and sizes, with both public and private organizations included. The heterogeneous mix of organizations should help to maximize the variation in the independent variables (Harrigan, 1983).

Dependent Variables. If the company has a full-fledged gainsharing plan, a pilot plan, or is planning to adopt a plan, GS is coded as 1.l There are several reasons for combining those companies which either have a full-fledged plan, a pilot plan, or are planning to adopt a plan. First, if a company has a pilot plan or is planning to adopt gainsharing, management is assumed to believe the plan is consistent with the rest of the HR system, even though not all of the firms will follow through in adopting a plan. Management's perception of fit is one important aspect of the fit perspective. Also, since much of the changes in the nature of the employment relationship have taken place in the 1980s and early 90s, the planning variables are more contemporaneous and capture a lot of relevant information that might otherwise be lost. Our use of a binary measure to indicate the existence of gainsharing is in line with similar such studies. Osterman (1994), for example, assessed the existence of innova-

ROBERT MANGEL and MICHAEL USEEM

335

tive pay systems (e.g., gainsharing, profit-sharing) by asking establishments whether or not they had each of the systems. His pretest results had suggested that it would be too difficult for the respondents to provide more detailed information, i.e., to provide information such as the percentage of employees covered by each program. We use two dependent variables to examine employee participation in decision making. The first measure, PARTICIPATION, BUT NO GS, is created to indicate whether any one of three forms of participation is used in the organization: supervisory training for greater participation, employee involvement teams, or autonomous work groups. The second participation variable, SUPV TRAINING, BUTNO GS, is created to indicate if the specific participation initiative - - supervisory training for greater participation - - is present. For both the composite participation variable and the supervisory training variable, a 1 is coded if the organization uses the relevant participation initiative(s) and the organization does not have gainsharing. The idea is to look at firms that have adopted participation but have not adopted gainsharing, a situation that corresponds to the pure participation scenario discussed above. Again, as with the gainsharing variable, a firm is considered to "have" the participation initiative if the respondent indicates that their company has a full-fledged plan, a pilot plan, or is planning to adopt a plan.

Independent Variables. MIDDLE MGMT INCENTIVES equals 1 if firms have fullfledged incentives, a pilot plan, or if they plan to adopt middle management productivity incentives. The primary variable for career ladders, FORMAL CAREER LADDER, is taken from a binary question that asks whether the company has a formal career advancement program for all employees. A yes is coded 1. As for employee screening, the screening variable, LN (NUMBER SCREENED), is taken from a question that asks firms the average number of applicants screened per hire. The question has seven categories: 1-5, 6-10, 11-20, 21-50, 51-100, 101-500, 500+. We code the midpoint for each category and cap the highest at 750. We also perform a natural log transformation to achieve a more normal distribution. VALUE SCREEN is coded 1 if the firms responded that inability to adapt to the work environment is one of two chief reasons that applicants are rejected. Control Variables. Given the administrative costs of installing a plan, small firms may not have sufficient resources to do so. Also, the larger the organization, the more subgroups it is likely to have and thus the more potential gainsharing sites. Firm size may thus impact the analysis through several different venues and hence the analysis will control for size. Size is measured by the natural log of the number of full-time workers LN (WORKERS). The use of gainsharing plans is likely to vary by industry, given that core technologies and human resource practices vary by industry. Likewise, previous studies by White (1979) find a preponderance of gainsharing in manufacturing settings, so we control for industry. We create dummy variables for industry at the 1-digit level, as follows: AGRICULTURAL, NATURAL RESOURCES, PRODUCTS, MANUFACTURING, TRANSPORT~UTILITIES, WHOLESALE~RETAIL, FINANCIAL, MISC SERVICES, and

336

JOURNAL OF LABOR RESEARCH

HEALTH SERVICES. Conglomerates are coded as OTHER and omitted from regressions to avoid multicollinearity. Management values have been cited by several researchers as an important determinant of HR practices and policies (Kochan et al., 1986; Osterman, 1994). Indeed, Osterman (1994) finds managerial values to be a strongly significant predictor of flexible work organizations. Our underlying framework is based on an efficiency argument, with the assumption that managers are choosing systems of practices because they form a logical fit. Hence, it is important to explore and to control for the degree to which managerial values influence the adoption of gainsharing and participation plans, and, hence, the degree to which the adoption of these plans is driven by motives that may not be directly related to organizational efficiency. We expect that managers who support employees would be likely to adopt empowerment-oriented initiatives such as gainsharing and other employee participation plans; we therefore control for whether the organization is said to have a supportive culture. The measure for supportive culture, SUPPORTIVE CULTURE, is a binary variable indicating whether the company has developed a more supportive culture to reduce turnover and enhance productivity. Our measure for supportive culture is in line with a similar study by Osterman (1994). In his survey of workplace transformation, a management values variable was constructed from one question with a five-point scale. The top two categories were combined and assigned a 1 for the binary variable. In addition to potential productivity increases, certain incentive plans are thought to offer advantages to firms by facilitating employment stability and by generally replacing fixed compensation costs with variable compensation costs. If some portion of regular compensation is replaced with pay tied to firm performance, firms may have greater ability to maintain employment levels in the face of business downturns. However, as Kruse (1993) points out, the group bonus must be a substitute rather than an add-on to compensation if group incentives are to have this theorized impact on employment stability. Although increasing compensation flexibility is itself a potential role for gainsharing, our main focus herein is on the role of facilitating participation. The type of control form may affect the decision of organizations to try to replace fixed wages. Bureaucratic organizations, for example, may rely more on fixed wages. We thus control for the use of gainsharing as a substitute for fixed wages. If a significant portion of gainsharing plans are substitutes for fixed wages, the existence of gainsharing will be negatively associated with the firm's tendency to raise salaries, and we will thus control for the firm's efforts to raise salaries. RAISE SALARIES is a binary variable that is coded 1 if the firm raises salaries to attract or retain workers. IV. Results

Descriptive Results. In total, 22 percent of the responding organizations have adopted, or are planning to adopt, a gainsharing plan. Breaking the number down, about 10 percent of the organizations have fully adopted gainsharing, 5 percent have adopted a pilot plan, and another 6 percent are planning to adopt one. These numbers are roughly corn-

ROBERT MANGEL and MICHAEL USEEM

337

parable to those of Osterman (1994), who finds that about 14 percent of organizations have a gainsharing plan. 2 As for other forms of participation, about 56 percent of the organizations have employee involvement teams, autonomous work groups, or supervisory training for greater participation. 3 One third of these organizations, or 19 percent of the total sample, also have gainsharing. For these respondents, we cannot be sure whether the gainsharing is intended to reward the separate participation initiatives or whether the plans operate independently in the organization. 4 In many cases, there is likely to be a connection between the gainsharing and the other forms of participation, as our results show that it is unlikely for a firm to adopt a gainsharing plan without also adopting one of these forms of participation. Indeed, 97 percent of the firms that have gainsharing also have one of the other participation forms. To examine alternative means for rewarding participation, we thus focus on the 37 percent of the total sample that have at least one of the participation forms but do not have gainsharing. It is these organizations, designated as PARTICIPATION, BUT NO GS, that we expect to be associated with career ladders or intensive employee screening. Similarly, 22 percent of the total sample have supervisory training for greater participation but do not have gainsharing. We predict that these organizations will be associated with the use of career ladders, since career ladders are a control mechanism that focus on the role of the supervisor and the formal hierarchy. Multivariate Results. Table 1 presents the results for the logistic regression equations. Model 1 presents the results for the dependent variable GS. Model 2 presents the results for the dependent variable PARTICIPATION, BUT NO GS. Model 3 presents the results for the dependent variable SUPV TRAINING, BUT NO GS. Examining the coefficients for middle management incentives across Models 1 and 2, we see partial support for Hypothesis 1. Consistent with Hypothesis la, the middle management incentives estimate in Model 1 indicates a strong positive relationship between middle management incentives and the adoption of gainsharing. Model 2 shows no significant relationship between middle management incentives and the adoption of PARTICIPATION, BUT NO GS. Thus, there is evidence that market-control organizations are more likely to adopt gainsharing as a means to foster participation, although there is no significant evidence here to support the hypothesis that these organizations are less likely to adopt pure participation plans. Each of the hypotheses concerning bureaucratic control are supported by the results. Model 1 shows that the existence of formal career ladders is negatively related to the adoption of gainsharing, as hypothesized in 2a, although the significance level is only marginal. Likewise, the coefficient on the formal career ladders variable in Model 2 shows a significant positive association between career ladders and the decision to adopt "pure participation" plans, as predicted in Hypothesis 2b. Career ladders may in some firms provide an efficient alternative to group bonuses for rewarding participation efforts. These results are consistent with arguments by Beer et al. (1984) that bureaucratic firms generally channel participation through supervisors.

338

JOURNAL OF LABOR RESEARCH

Table 1

Logistic Regressions Model 2

Model 3

GS

Participation -No GS

Supervisor Training -No GS

LN (Workers)

.17"* (.06)

.08* (.04)

.14"* (.O5)

Manufacturing

.50*** (.14)

-. 11 (.12)

-.07 (.15)

Health Services

-.01+ (.18)

.11 (.13)

.05 (.15)

Products

-. 10 (.21)

.!6 (.17)

.02 (.2)

-2.36 (3.61)

,42 (.35)

-2.40 (3.6)

Natural Resources

-. 16 (.38)

.18 (.32)

.06 (.34)

Wholesale~Retail

-.07 (.24)

.37* (.21 )

-.23 (.24)

Transport./Utils.

.17 (.18)

.20+ (.15)

.19 (.16)

Misc. Service

-.34 (.29)

.03 (.21)

-. 14 (.26)

Financial Services

-.61 *** (.29)

.16 (. 13)

-.05 (. 15)

Supportive Culture

.36*** (. 10)

.25*** (.08)

.32*** (.09)

Raise Salaries

-. 17* (. 10)

.04 (.08)

.01 (.09)

Model 1

Agricultural

Middle Mgmt. Incentives

.66"** (.10)

-.01 (.08)

-.01 (.09)

Formal Career Ladder

-. 15+ (.11)

.16" (.09)

.22* (.10)

LN (# Screened)

-.06 (.O9)

.10+ (.O6)

.04 (.08)

Value Screen

-. 17* (. 10)

.08 (.08)

.06 (.09)

-5.90 (3.77)

-.66 (.94)

-4.80 (3.80)

- 2 Log Likelihood

693

1018

814

Model Chi 2 Significance

.000

.001

.000

Constant

Notes: (Standard errors are in parentheses). N= 802 for all equations. All hypotheses are directional - - l-tailed significance levels are used, +(*, **, ***) Significant at the. 1 (.05, .01, .001) levels.

ROBERT MANGEL and MICHAEL USEEM

339

In these firms, the supervisor can monitor the participatory behavior of individual employees and compensate employees for their participation, on an individual basis, through promotions and salary increases. There is no need to reward employees collectively for their participation, and there is little use for employee suggestion systems that circumvent the supervisor. In order to promote greater participation, these companies often provide supervisors with training in facilitating participation. Supporting this specific argument, our results in Model 3 show that career ladders are significantly associated with supervisory training in greater participation, as predicted in Hypothesis 2c. The negative relationship between bureaucratic control and gainsharing may be driven in part by supervisory resistance to the adoption of gainsharing, as discussed above. Given the more prominent role of supervisors in bureaucratic-control systems, gainsharing plans may be viewed as a threat to the supervisors' role as the primary conduit for employee participation, and the supervisors may have power to block the adoption of gainsharing. The hypotheses concerning clan control receive partial confirmation. Firms that use clan control are less likely to adopt gainsharing, although not all of the tests display significant support. Hypothesis 3a is partially supported. Model 1 indicates that screening for value fit is significantly and negatively related to the adoption of gainsharing, as predicted. However, the estimate in Model 1 for the general level of screening, LN (NUMBER SCREENED), does not indicate a significant negative relationship to GS. Likewise, there is partial support for Hypothesis 3b that firms using clan control are likely to adopt participation plans without also adopting gainsharing. Model 2 indicates that screening overall, LN (NUMBER SCREENED), has the predicted positive association with the adoption of pure participation plans, although the significance is marginal. The coefficient for value screening, however, does not indicate a significant positive association between screening for values and adopting pure participation plans. Firms using clan-control systems may not necessarily compensate employees directly for the extra work of participation. However, because these firms screen employees carefully, their employees are more likely to perceive that their interests are aligned with those of the firm. There is then less need for extra compensation for employee participation in decision making. Overall, the results at least partially support each of the three hypotheses. Some of the tests showed results that were not significant, however, none of the tests showed results that were counter to the hypotheses. Control Variables. The estimates for two of the control variables, RAISE SALARIES and SUPPORTIVE CULTURE, are both significant and consistent with results from previous research. The coefficient for RAISE SALARIES in Model 1 indicates a significant negative association between raising salaries to recruit employees and adopting gainsharing. Thus, there is evidence that gainsharing, on average, at least partially substitutes for other forms of compensation. The important role played by managerial values in the adoption of participation programs is also evident. The coefficient for SUPPORTIVE CULTURE is positive and

340

JOURNAL OF LABOR RESEARCH

significant in each equation, indicating a positive association between supportive cultures and the adoption of both gainsharing and forms of participation that do not include a group bonus. Similarly, Osterman (1994), in his study of 694 manufacturing establishments, found that managerial values play an important role in the adoption of flexible workplace practices. Our results thus add to the growing evidence that supportive cultures are associated with innovative workplace practices. V. Discussion

Our results suggest that gainsharing plans are a mechanism for fostering employee participation that are more commonly found in firms that use market control. Firms that use either clan-control systems or bureaucratic-control systems are less likely to use gainsharing, and they are more likely to adopt forms of participation that do not involve a group bonus. This "fit" perspective provides possible explanations for why companies have not been adopting innovative HR initiatives as quickly as might be expected. Specifically, HR researchers have begun to question why participation and other empowerment initiatives are not more widespread, given the high potential impact of these initiatives on firm performance. General managerial resistance to employee participation in decision making may be one factor impeding the spread of these practices (Ogden, 1992; Pfeffer et al., 1995). Our results support this argument and also suggest that this resistance may be especially high in bureaucratic firms. More generally, our results suggest that the need for a fit among HR initiatives may play a role in hindering the widespread adoption of these plans. Successful adoption of different forms of participation may depend in part on the types of HR control systems and related initiatives that are currently in place in a given organization. If gainsharing plans fit best with market control, then many organizations may not have the appropriate control context for introducing gainsharing. Successful adoption of gainsharing may therefore first require large-scale organizational transformations. If all of a company's policies and practices are geared toward bureaucratic control (tall hierarchy, career ladders), adopting a gainsharing plan may be too difficult and expensive. Likewise, if a company attempts to adopt a participation system that is not consistent with its current HR control systems, the plan is less likely to succeed. Evidence suggests that some companies are in deed transforming themselves to become more focused on increasing shareholder value and are adopting performancepay initiatives at various levels (Useem and Gottlieb, 1990). We may thus expect more gainsharing as part of this transformation. The notion that the success of HR initiatives may depend on a firm's HR approach also has implications for the general literature that examines employee participation in decision making. Reviews examining the effectiveness of participation programs show little consensus as to whether these plans on the whole are substantially effective (Cotton et al., 1988; Wagner, 1994). In light of the equivocal results on the effects of participation, researchers are looking more at the reasons why participation may or may

ROBERT MANGEL and MICHAEL USEEM

341

not be effective. For example, some research suggests that the specific form of participation determines the degree of its impact on performance (Cotton et al., 1988, 1990), although this line of research is as yet controversial (Wagner, 1994). Researchers are also suggesting that more attention be paid to contextual factors that may influence participation's success (Cotton et al., 1990; Schwochau et al., 1997). Levine and Tyson (1990) specifically argue that it is important for participation to be rewarded through group incentives, as quality circles and other forms of participation tend to be short-lived if they do not offer employees a stake in the returns. Our model also recognizes the importance of rewarding employee participation, but we suggest that the efficiency of the type of reward may depend on the strategic HR context. A group bonus may not be the most efficient means of rewarding participation in organizations that are not well suited for market control. In such firms, participation may be better rewarded through bureaucratic means such as career ladders, or firms may reduce the need for rewards by using clan-control systems to align the values of the work force with those of the organization. Our results suggest that characteristics of HR control mechanisms, and especially the alternative methods for rewarding participation, are important contextual factors to consider in future research on the efficacy of employee participation. Finally, research that looks at the performance effects of fit could also be quite useful. In our approach, we are looking for evidence of internal consistency. Research that examines whether better fits between gainsharing and other HR initiatives lead to better organizational performance would provide strong evidence for the fit perspective. Such research could help us to understand which fits are of greatest importance, and such research could help to further address the questions concerning how gainsharing plans work and why they are adopted.

NOTES *The authors thankPeter Cappelli, David Crawford, John Deckop, Peter Sherer, and Harbir Singh for comments on earlierdrafts, Towers Perrin for access to these data, and TempleUniversityfor fundingsupport. IFor the incentiveand participationquestions,the survey also presents an option for the respondentto indicate that the plan was discontinued.Althoughit is not possibleto know why the plan was discontinued,two possible explanationsincludethat the plan failed to fit in with the context or that the organizationalcontext changed to create a misfit.Thus, we countedthese responsesas indicatingthat the organizationdid not have the plan. Treatmentof the discontinuedcategory should have little impact as relativelyfew respondents (approximately 1%) indicatedthat a plan has been discontinued. 2Of course, caution should be used in comparingthe descriptive results here to those of Osterman (1994), given samplingand other procedural differences. 3Toclarifyexposition,we use the term "have" in the descriptiveresults sectionto includecompaniesthat have already adopted a plan, have a pilot plan, or are planningto adopt a plan. 4Gainsharingoften includesa form of participationexplicitlyas part of the plan.Thus, the respondentsmay consider the participationcomponentof the plan to be includedin the term gainsharing,or they may indicate that the participationcomponentis one of the other participationforms listed in the survey.

342

JOURNAL OF LABOR RESEARCH

REFERENCES Baird, Lloyd R. and Ilan Meshoulam. "Managing Two Fits of Strategic Human Resource Management." Academy of Management Review 13 (January 1987): 116-28. Beer, Michael, Bert Spector, Paul R. Lawrence, D. Quinn Mills, and Richard E. Walton. Managing Human Assets. New York: Free Press, Macmillan, 1984. Blasi, Joseph R. Employee Ownership: Revolution or Ripoff? Cambridge, Mass.: Ballinger Publishing, 1988. Blinder, Alan S., ed. Paying for Productivity. Washington, D.C.: Brookings Institution, 1990. Bullock, R. J. and Edward E. Lawler. "Gainsharing: A Few Questions, and Fewer Answers." Human Resource Management 23 (Spring 1984): 23-40. Cappelli, Peter. "Rethinking the Nature of Work: A Look at the Research Evidence." Compensation and Benefits Review 29 (July- August 1997): 50-9. and Harbir Singh. "Integrating Strategic Human Resources and Strategic Management." In Research Frontiers in Industrial Relations and Human Resources. David Lewin, Olivia S. Mitchell, and Peter Sherer, eds. Madison, Wisc.: Industrial Relations Research Association, 1992, pp. 165-92. Cotton, John L., David A. Vollrath, Kirk L. Froggatt, Mark L. Lengnick-Hall, and Kenneth R. Jennings. "Employee Participation: Diverse Forms and Different Outcomes." Academy of Management Review 13 (January 1988): 8-22. Cotton, John L., David A. Vollrath, Mark L. Lengnick-Hall, and Kirk L. Froggatt. "Fact: The Form of Participation Does Matter-A Rebuttal to Leana, Lock, and Schweiger?' Academy of Management Review 15 (January 1990): 147-53. Davis, James H., F. David Schoorman, and Lex Donaldson. "Toward a Stewardship Theory of Management." Academy of Management Review 22 (January 1997): 20-48. Dyer, Lee. "Strategic Human Resources Management and Planning." In Research in Personnel and Human Resources Management. Kendrith M. Rowland and Gerald R. Fen'is, eds. Greenwich, Conn.: JAI Press, 1985, pp. 1-30. Eisenhardt, Kathleen M. "Agency- and Institutional-Theory Explanations: The Case of Retail Sales CompensationY Academy of Management Journal 31 (September 1988): 488-511. Fox, Mark A. and Robert T. Hamilton. "Ownership and Diversification: Agency Theory or Stewardship Theory." Journal of Management Studies 31 (January 1994): 69-81. Hammer, Tove H. "New Developments in Profit Sharing, Gainsharing, and Employee Ownership." In Productivity in Organizations. John P. Campbell and Richard J. Campbell, eds. San Francisco: Jossey-Bass, 1988, pp. 328-66. Hanlon, Susan C., David G. Meyer, and Robert R. Taylor, "Reversal of a Successful Gainsharing Program: Experimental Evidence of Residual Effects." In Jerry L. Wall and Lawrence R. Jauch, eds. Academy of Management Proceedings, 1992, pp. 264-8. Harrigan, Katherine R. "Research Methodologies for Contingency Approaches to Business Strategy." Academy of Management Review 8 (April 1983): 398-405. Kanter, Rosabeth Moss. "The New Managerial Work." Harvard Business Review 67 (November-December 1989): 85-92. Klein, J.A. " Why Supervisors Resist Employee Involvement." Harvard Business Review 62 (SeptemberOctober 1984): 87-95. Kochan, Thomas A., Harry C. Katz and Robert B. McKersie. The Transformation of American Industrial Relations. New York: Basic Books, 1986. Kruse, Douglas L. Profit Sharing: Does It Make a Difference ? Kalamazoo, Mich.: W.E. Upjohn Institute for Employment Research, 1993. Lawler, Edward, III. Pay and Organization Development. Reading, Mass.: Addison-Wesley, 1981.

ROBERT MANGEL and MICHAEL USEEM

343

Levine, David I. and Laura A. Tyson. "Participation, Productivity, and the Firm's Environment." In Alan S. Blinder, ed. Paying for Productivity. Washington, D.C.: Brookings Institution, 1990, pp. 183-285. Mahoney, Thomas A. and John R. Deckop. "Evolution of Concept and Practice in Personnel Administration/Human Resource Management." Journal of Management 12 (June 1986): 223-41. McKersie, Robert B. "The Promise of Gainsharing." ILR Report 24 (Fall 1986): 7-11. Mitchell, Daniel J.B., David Lewin, and Edward E. Lawler, III. "Alternative Pay Systems, Firm Performance and Productivity." In Alan S. Blinder, ed. Paying for Productivity. Washington, D.C.: Brookings Institution, 1990, pp. 15-94. O'Dell, Carla. People, Performance and Pay. Houston, Tex.: American Productivity Center, 1987. Ogden, Stuart. "The Limits to Employee Involvement: Profit Sharing and Disclosure of Information." Journal of Management Studies 29 (March 1992): 229-48. Osterman, Paul. Broken Ladders. New York: Oxford University Press, 1996. . "How Common Is Workplace Transformation and Who Adopts It?" Industrial and Labor Relations Review 47 (January 1994): 173-88. Ouchi, William G. "Markets, Bureaucracies, and Clans." Administrative Science Quarterly 25 (March 1980): 129-41. Pfeffer, Jeffrey and James Baron. "Taking the Workers Back Out: Recent Trends in the Structuring of Employment." In Barry M. Staw and Larry L. Cummings, eds. Research in Organizational Behavior. Greenwich, Conn.: JAI Press, 1988, pp. 257-303. Pfeffer, Jeffrey, Robert B. Cialdini, Benjamin Hanna, and Kathleen Knopoff. "Why Managers Won't Empower Employees: Managerial Involvement and the Biased Evaluation of Worker Output." Annual Meeting of the Academy of Management, 1995. Pfeffer, Jeffrey and Yinon Cohen. "Determinants of Internal Labor Markets in Organizations." Administrative Science Quarterly 29 (December 1984): 550-72. Schwochau, Susan, John Delaney, Paul Jarley, and Jack Fiorito. "Employee Participation and Assessments of Support for Organizational Policy Changes." Journal of Labor Research 18 (Summer 1997): 379-401. Snell, Scott A. "Control Theory in Strategic Human Resource Managment: The Mediating Effect of Administrative Information." Academy of Management Journal 35 (June 1992): 292-328. Towers Perrin. Achieving Results Through Sharing: Group Incentive Program Survey Research. Valhalla, N.Y.: Towers Perrin, 1990. Useem, Michael. "Corporate Restructuring and the Restructured World of Senior Management." In Paul Osterman, ed. Broken Ladders: Managerial Careers in the New Economy. New York: Oxford University Press, 1996, pp. 23-54.

. Executive Defense: Shareholder Power and Corporate Reorganization. Cambridge, Mass.: Harvard University Press, 1993. _

_

and Martin M. Gottlieb. "Corporate Restructuring, Ownership-Disciplined Alignment, and the Reorganization Of Management." Human Resource Management 29 (Fall 1990): 285-306.

Van Dyne, Linn, Jill W. Graham, and Richard M. Dienesch. "Organizational Citizenship Behavior: Constrnct Redefinition, Measurement, and Validation."Academy of Management Journal 37 (August 1994): 765-803. Wagner, John A. III. "Participation's Effect on Performance and Satisfaction: A Reconsideration of Research Evidence." Academy of Management Review 19 (April 1994): 312-31. Weitzman, Martin L. and Douglas L. Kruse. "Profit Sharing and Productivity." In Alan S. Blinder, ed. Payingfor Productivity. Washington, D.C.: Brookings Institution, 1990, pp. 95-141. Williamson, Oliver. E. "The Economics of Organization: The Transaction Cost Approach." American Journal of Sociology 87 (November 198 l): 548-77.

. Markets and Hierarchies. New York: Free Press, 1975.