III. HUMAN RESOURCES AND INTERNATIONAL SECTIONS
REFEREED PAPERS

The Effect of Employee Suggestions and Union Support on Plant Performance Under Gainsharing

JEFFREY B. ARTHUR
Virginia Polytechnic Institute and State University

DONG-ONE KIM
Korea University

Abstract

      We use longitudinal data from two unionized manufacturing plants with different levels of union support for gainsharing programs to test hypotheses concerning the effect of employee suggestions on plant performance. In addition, we use an organizational learning perspective to test whether the pattern and content of suggestions over time differed in the two plants. Results from ARIMA time-series regression analysis provide support for the hypothesis that the increases in the level of implemented suggestions are significantly related to lower unit labor costs in both of the plants. Contrary to expectations, 2nd-order learning suggestions were associated with improved performance only in the case of the plant with relatively high union support for gainsharing. In addition, the predicted decline in the relative number of 1st-order learning suggestions over time was found to exist only in the plant with relatively high union support and involvement in the gainsharing program.

 

      Despite the long history of gainsharing, our understanding of how the introduction of gainsharing plans lead to performance outcomes in firms remains weak. Early research on the Scanlon gainsharing plan emphasized the importance of employee and union involvement in driving the success of the plan by transforming labor-management relations from a traditional adversarial relationship to a more cooperative one by developing a system of employee involvement through suggestion making and joint labor-management committees, and promising to share any savings from performance improvement equally between employees and management (Frost, Wakely, and Ruh 1974).

      Although employee suggestions and the quality of labor-management relations are predicted to play a critical role in understanding how gainsharing works, these topics have not been evaluated systematically in the literature. In particular, we know very little about how and under what conditions these factors are likely to affect organizational performance.

      In this study, we adopt an organizational learning perspective to further our understanding of the process issues involved in gainsharing. Organizational learning is a fundamental concept in organizational theory that has experienced a resurgence of interest by researchers and practitioners in recent years (Garwin 1993). Although not utilized in the industrial relations literature, this perspective is useful in that it provides us with a framework for understanding how employee participation through suggestion making affects the dynamics of gainsharing performance over time.

      The data for this study come from two manufacturing organizations. In one plant, the union was actively involved in the introduction and design of the gainsharing program and fully participated in and supported its implementation. The second plant installed a virtually identical Scanlon-type gainsharing plan, but did so without the full support and participation of the employees and the union. We use over 4 years of monthly data on plant performance and employee suggestions in one plant and 7 years in the other to test three hypotheses concerning the impact of employee involvement and union support on plant performance under gainsharing.

Hypotheses

      Employee suggestions are perhaps the most basic form of employee participation, with a history of over 100 years in the United States. It has been argued that employee suggestions influence organizational performance by enhancing the flow and use of important work-related information (e.g., Locke and Schweiger 1980; Miller and Monge 1986). Under a Scanlon-type gainsharing plan, suggestion making is formalized with the use of joint employee-management review committees and reinforced by the use of group-based monetary rewards (Gomez-Mejia, Welbourne, and Wiseman 2001; Kim 2000). Proponents of gainsharing argue that employee suggestions enable more upward communication, facilitate better utilization of information based upon better understanding of job and tasks, and generate more creative and innovative ideas which, in turn, improve plant performance (e.g., Frost, Wakely, and Ruh 1974). It follows that, all else equal, more implemented employee suggestions should be associated with better performance.

      Hypothesis 1: Higher volume of employee suggestions under gainsharing is associated with improved plant performance.

      In addition to examining the overall impact of suggestions on performance over time, the organizational learning perspective provides a framework for disaggregating suggestions into two distinct types. First-order learning suggestions are those that seek to improve performance by saving costs and improving the efficiency of existing operations and procedures. Although important, this type of learning does not challenge the existing procedures, norms, or values (Argyris and Schon 1996). In contrast, second-order learning "is characterized by the search for and exploration of alternative rules, technologies, goals, and purposes" (Lant and Mezias 1993). Suggestions of this type include changes that would challenge existing routines, norms, and values in the organization by altering existing procedures or products. The organizational learning framework, then, leads us to predict that these different types of suggestions will have different types of impact on organizational performance. Whereas first-order type learning suggestions are expected to result in incremental short-term improvements in cost savings and efficiency, the impact of second-order type suggestions is expected to be more profound and longer term.

      Hypothesis 2: Second-order learning suggestions have a larger, longer-term impact on plant performance than first-order suggestions.

      In addition to providing a conceptual framework for disaggregating employee suggestions, the organizational learning perspective implies that the content of employee suggestions will change over time. According to this perspective, learning results from a search process that is motivated by a gap between what exists and what is expected (or aspired to). This perspective leads us to predict that in the period immediately following the introduction of gainsharing, employee search processes will likely result in first-order learning suggestions. This is because employees will seek familiar solutions to problems and improve the "easy" things first by improving on existing conditions.

      There is, however, a finite amount of cost savings that can be gained from improvements in an existing operation or system. As the potential for cost savings declines, we expect employees to begin to focus relatively more attention toward searching for ways in which the current operations can be transformed or altered. Based on the organizational learning framework, we predict a decline in the relative proportion of first-order suggestions (and a corresponding rise in second-order suggestions) over time (e.g., Arthur and Aiman-Smith 2000).

      In this study, we add to the existing literature by testing the hypothesis that this organizational learning dynamic is contingent on the nature of the labor-management relations in the plant. Drawing on the IR literature, we note that employee suggestion making is rarely politically neutral. By their very nature, second-order learning suggestions change the implicit wage-effort bargain that exists in every employment relationship and is institutionalized in unionized settings through the collective bargaining process. Because employee suggestion making exists outside of the formal controls established in collective bargaining, we posit that employees' willingness to shift toward relatively more second-order learning suggestions will be contingent on the degree to which they believe that management is acting in their best interests. In other words, they require some assurance that management will not take advantage of employee suggestions to make unilateral changes in their favor in the wage-effort bargain. In a unionized setting, we hypothesized that the union's direct involvement and support for the gainsharing plan provides the basis for that assurance.

      Hypothesis 3: Union involvement and support for gainsharing will affect the type of suggestions submitted over time. Specifically, a decreasing proportion of first-order suggestions will be found only in the context of labor-management cooperation and union support for gainsharing.

Methods

      The data for this study come from two manufacturing organizations. The two plants are similar in many ways. Both are located in the Midwest. Employees in both plants are represented by a labor union. Both plants decided in the mid- to late 1980s to implement a Scanlon-type gainsharing program in order to improve performance. Consistent with the modified Scanlon plan format, both plants instituted an employee suggestion system and a bonus formula based on reductions in the amount of labor and other production costs compared to historical averages. In one plant (labeled "Plant A"), the union was actively involved in the introduction and design of the gainsharing program and fully participated in and supported its implementation. The second plant (labeled "Plant B") installed a virtually identical Scanlon-type gainsharing plan, but did so without the full support and participation of the employees and the union. Monthly data on employee suggestion and plant performance were obtained from both plants. By comparing the results from both plants, we are able to test the hypotheses described previously concerning the impact of suggestions on performance as well as the importance of labor-management relations in understanding how gainsharing works.

Measures

Employee Suggestions

      Data on the number of employee suggestions submitted each month were obtained from plant records. In both plants, we were able to obtain the total number of suggestions that were implemented in the plant each month (Plant A) or each 4-week period (Plant B) to create the variable Total Implemented Suggestions. For Plant A, data were obtained from January 1989 through December 1992 (48 months). For Plant B, we use the 89-month period from July 1985 through May 1992.

      The variables First-Order Suggestions and Second-Order Suggestions were created by content analyzing each suggestion based on distinctions found in the organizational learning literature. We used multiple raters and tested to insure that our categorizations were psychometrically reliable (i.e., had a statistically significant coefficient Kappa; see Arthur and Aiman-Smith 2001).

Plant Performance

      We used monthly data on unit labor costs (including direct and indirect labor costs) as the measure of plant performance. Labor costs represent a significant portion of total costs in both plants. The unit labor costs were calculated by dividing labor costs by the value of production. The focus on labor costs and productivity is consistent with previous studies on the impact of employee participation on performance (Locke and Schweiger 1979). Labor costs are also an appropriate measure of plant performance because they are an element of operating performance that employees are most likely to be able to influence directly through their efforts and ideas.

Controls

      We used the number of employee grievances and the production volume each month as controls in the regression analyses because of the potential confounding effects of these variables on performance outcomes. The number of grievances filed each month were obtained from plant records. In Plant A production volume was measured as the number of units produced that month. In Plant B monthly production volume was measured as the sales value of production for the 4-week period.

      Multiple regression analysis is used to test the various hypotheses. Because these are time series data, we used an ARIMA procedure to model the error term and provide the appropriate corrections for nonstationarity, autocorrelations, seasonality, and moving average processes (Ostrom 1980; SAS Institute 1993).

Results

      Results of the ARIMA multiple regression analysis are presented in Tables 1 and 2. To test for the possibility of delayed effects of suggestions on performance, in Table 2 we lagged the variable measuring accepted and implemented suggestions by 1 to 6 months and performed separate regression analyses for each of these lagged periods on performance. This allows us to measure effect of the number of implemented suggestions in each of the previous 6 months on plant performance.

      For both plants the negative relationships between the number of suggestions and performance at Lag 0 support the prediction in hypothesis 1 that employee suggestions are associated with improved performance (reduced labor costs).

      The ARIMA regression analysis in Table 2 tests whether second-order suggestions have a larger, long-term implant performance than first-order learning suggestions (hypothesis 2). The results provide some support for the hypothesis for Plant A but not for Plant B. In Plant A, second-order suggestions have a negative relationship with unit labor costs in all models except Lag 2 (statistically significant in the Lag 4 model). In contrast, all seven lag models for Plant B show a positive relationship between second-order suggestions and unit labor costs (statistically significant positive relationships in the Lag 1, 2, and 6 models).

      Finally, the analysis in Table 3 tests the hypothesis that differences in labor-management relations between Plants A and B will result in a different pattern of suggestion making in the two plants. This hypothesis is supported by the results of the regression analysis in Table 3 in which the trend variable Time is negatively related to the number of first-order suggestions submitted each month in the Plant A case, but not in the Plant B case.

Discussion

      The analyses presented in this paper provide some preliminary support for the hypothesis that employee suggestions are associated with improved performance. We found significant negative relationships between suggestions and unit labor costs in both plants. We also found mixed support for the hypothesis that different types of learning suggestions have a different impact on performance. Specifically, the results for Plant A indicated that second-order learning suggestions appear to have a stronger delayed effect on performance than first-order learning suggestions. The results for Plant B, however, did not follow this hypothesized pattern.

      Finally, we found strong support for the hypothesis that the pattern of first- and second-order learning suggestions is different in the two plants. Although there was a significant decline in the number of first-order learning suggestions over time for Plant A, no such significant decline was found in Plant B. These results appear to support the observation that labor-management relations play an important role in understanding how employee participation in gainsharing works. In particular, the willingness of employees to submit second-order learning suggestions, which could alter the wage-effort bargain, is associated with a higher level of union support and involvement in the design and implementation of the gainsharing plan.

References

Argyis, C., and D.A. Schön. 1996. Organizational Learning II: Theory Method and Practice. Reading, MA: Addison-Wesley.

Arthur, J.B., and Aiman-Smith, L. 2001. "Gainsharing and Organizational Learning: An Analysis of Employee Suggestions over Time." Academy of Management Journal. Forthcoming.

Frost, C.F., J.H. Wakeley, and R.H. Ruh. 1974. The Scanlon Plan for Organization Development: Identity, Participation, and Equity. East Lansing: Michigan State University Press.

Garwin, D.A. 1993. "Building a Learning Organization." Harvard Business Review, July-August, 73-91.

Kim, Dong-One. 2000. "The Impact of Employee Suggestions on Organizational Performance: A Longitudinal Study." Paper presented in the 52nd Industrial Relations Research Association Meetings, Boston, January 7-9.

Miller, K.I., and P.R. Monge. 1986. "Participation, Satisfaction, and Productivity: A Meta-analytic Review." Academy of Management Journal, Vol. 29, no. 4, pp. 7'-53.

Lant, T.K., and S.J. Mezias. 1993. "Organizational Learning Model of Convergence." Organizational Science, Vol. 3, pp. 47-71.

Locke, E.A., and D.M. Schweiger. 1979. "Participation in Decision-Making: One More Look." In M. Staw, eds., Research in Organizational Behavior, Vol. 1. Greenwich, CT: JAI Press, pp. 265-339.

Ostrom, C.W. 1980. Time Series Analysis: Regression Techniques. Newbury Park, CA: Sage.

SAS Institute, Inc. 1993. SAS/ETS User's Guide, 2 vol., 2nd ed. Cary, NC: SAS Institute, Inc.

Gomez-Mejia, L.R., T.M. Welbourne, and R.M. Wiseman. 2000. "The Role of Risk Sharing and Risk Taking Under Gainsharing." Academy of Management Review, Vol. 25, pp. 492-507.