Why do the hundreds of billions of dollars spent annually in the United States on technology to improve the productivity of individuals and groups appear to have so little impact on the productivity of the organizations in this country? Why are increases in individual productivity not reflected in measures of organizational productivity? These and related questions frame the productivity paradox addressed by this report. The position of the Panel on Organizational Linkages, a study panel convened by the National Research Council, is that the answers are to be found in a better understanding of the linkages among individual, group, and organizational productivity.
THE PRODUCTIVITY PARADOX
The ability of nations, and organizations within nations, to enhance the standard of living of the world's growing population depends on continued increases in the productivity of the systems that provide goods and services. In an increasingly competitive global economy, productivity growth is also essential for maintaining or advancing economic opportunities for individuals and societies. Moreover, it is apparent as never before that the peoples and institutions of the world are highly interconnected and that, as a consequence, each nation has a vested interest in the productivity of other nations. A nation might be able to gain short-term advantage over a marginally productive competitor, but over the long term all nations lose from slow productivity growth
regardless of where it occurs. Thus, a productive nation is desirable for the contributions it makes not only to the quality of life of its people but, ultimately, to the quality of life of those in other nations as well. Some management theorists have identified productivity growth, particularly in knowledge and service workers, to be the greatest challenge now facing the developed countries of the world. They predict that it will determine the fabric of society and the quality of life in every industrialized nation and that without productivity growth the world will face increasing social tensions, polarization, and radicalization (e.g., Drucker, 1991).
Trends of Productivity Growth in the United States
The United States has experienced more than 25 years of declining productivity growth. Between 1965 and 1985, for example, the U.S. position in the international automobile, steel, shipbuilding, and textile industries deteriorated significantly. More recently, the U.S. position in electronics, computers, robotics, and biotechnology has slipped (Johnson and Packer, 1987; Wohlers and Weinert, 1988). The U.S. labor force does not appear to be making the contribution it once did to the productivity of the world economy.
In a more recent analysis of U.S. productivity growth, the Urban Institute (Sawhill and Condon, 1992) reported that between 1973 and 1990 the hourly output of an American worker grew only 0.7 percent a year. In contrast, the annual rate of growth between 1948 and 1973 was 2.5 percent. According to this analysis, if worker productivity—the basic determinant of wages—had continued to grow at the same rate after 1973 as it did before, the typical family's income in the United States in 1990 would have been $47,600 instead of $35,300.
What is the problem? Has the United States not been investing in productivity growth? There are indeed areas (e.g., infrastructure and education) in which inadequate investment may be inhibiting U.S. productivity. On the other hand, the United States has actually been investing heavily in advanced technologies to enhance productivity growth. The returns, however, do not appear commensurate with the investments. For example, one analysis showed that the data processing budgets for U.S. corporations increased by about 12 percent a year over the previous decade. Productivity increases from those investments, however, averaged less than 2 percent a year (Weiner and Brown, 1989).
In Chapter 2, Attewell reviews a number of investigations of the
impact of investments in information technology (IT) on organizational productivity. Such investments are of particular interest because, for many years, they have accounted for a very large proportion of the U.S. industrial investment. As discussed in Chapter 2, however, the huge annual private sector investments in computers and related technologies (an estimated $154 billion in 1990) have had no apparent effect on measures of organizational productivity. Although specific applications of IT have made positive contributions to productivity, the overall investment does not seem to have improved industrial productivity in the United States.1 This and similar evidence presents a paradox: Why have the enormous investments in IT not resulted in clear-cut increases in organizational productivity? It is clear from the analyses reviewed in Chapter 2 that enhancing productivity is a major national challenge and that this productivity paradox must be understood. As a nation, we need to understand better the factors that influence the productivity of our organizations and the methods for addressing the facilitators and inhibitors of organizational productivity.
Focus and Assumptions of Previous Research
The question that has been addressed by most research into the productivity of work units can be framed as follows: How can the productivity of X be increased? X might be an individual, group, or larger organization. To this end, research has examined a variety of interventions and their impact on the productivity of X. They include the design and implementation of new technology; the application of techniques for the selection, training, and motivation of personnel; the design and redesign of jobs and tasks; innovations in organizational development and management methods; and the introduction of new compensation and incentive systems. Traditionally, researchers have focused on a single level of analysis—improving performance at the individual, group, or organizational level. Interventions are made and, in some cases, measurements are made to determine the impact of the interventions, invariably at the same level at which the intervention was made. The linkage between an intervention at one level and the impact on productivity at another level has largely been ignored.
In attempting to increase productivity, human factors specialists,
industrial engineers, industrial and organizational psychologists, and others, have largely followed the pattern of focusing on a single level of analysis. The principal level has been that of the individual, but groups and higher organizational units are being given increasing attention. Human factors specialists, for example, have attempted to enhance individual and group performance by matching improvements in technology with the capabilities and limitations of operators. The emphasis of this approach has been on increasing productivity by improving the design of the task and job. The goal has been to fit the task to the requirements of operators and, in doing so, to maximize the speed and accuracy of performance and to minimize other measures, such as learning time, work load, and accidents (Sanders and McCormick, 1987). Industrial engineers, on the other hand, have focused on the work system, often in alliance with human factors specialists and industrial and organizational psychologists. They have attempted to integrate people, technology, and methods to improve the performance of systems, with emphasis on quality and productivity. They have assessed the results of their efforts through such measures as efficiency, timeliness, and defect rates (Sink and Tuttle, 1989).
Industrial and organizational psychologists have attempted to enhance productivity principally by improving the capabilities of the individuals performing the work or through interventions in groups or organizations. Their approaches have included matching individuals to jobs and tasks, training individuals in job skills and knowledge, changing the structure of groups, and motivating individuals and groups toward job objectives. The effects of these efforts have been measured mainly in terms of the facility with which tasks are learned, self-reports of job satisfaction, ratings by others of job performance, indicators of group output, and other behavioral indices (e.g., absenteeism and turnover) presumed to be related to organizational effectiveness (Campbell, Campbell and Associates, 1988). A principal assumption underlying these efforts has been that increases in individual or group productivity will ultimately contribute to increases in the productivity of the enterprise. However, as discussed in Chapter 2, there appear to be factors or processes that inhibit the extent to which changes in individual productivity are reflected in changes in the productivity of aggregates of individuals—groups and organizations.
The Question Addressed in This Report
The question addressed in this report is, given an increase in the productivity of X, under what conditions will there also be an increase in the productivity of Y? Y might be a work unit of one or more people at the same or different level of analysis—the group or the organiza-
tion. Hence, this is a study of organizational linkages. In this report, a linkage is a change (or hypothesized change) in the performance of one work unit as the result of a change in the performance of another. Say, for example, that the introduction of new computer-based workstations results in an increase of productivity averaging 10 percent across the individuals in a work group. The question of interest here then becomes, under what conditions will the productivity gain of the work group as a whole be the same, less, or greater than 10 percent?
This question is addressed from the views of many different disciplines—psychology, engineering, information technology, and others. As a consequence, the report does not speak with a single voice in answering the question, nor does it examine from a single perspective the various influences that inhibit or facilitate productivity linkages in organizations. It is, rather, constructed as a series of essays. At the present state of understanding, this diversity seemed both necessary and useful. The final chapter summarizes the common themes and findings of the report and presents the conclusions that emerged from the panel's analyses and deliberations.
No claim is made that every possible problem involving organizational linkages has been addressed. There are certainly problems specific to work domains, organizational levels, and so on not examined by the panel. However, the panel believes that the interdisciplinary approach resulted in the identification and examination of the key issues in organizational linkages. Moreover, the panel believes that the findings and conclusions relative to these issues can be generalized to linkage problems that are not specifically addressed in this report.
Linkages and Influences
The unique perspective of this report on productivity is that it addresses the productivity linkages among different levels of analysis—individuals, groups, and organizations—and the factors (processes and mechanisms) that influence those linkages. The goal is to explicate the conditions under which linkages can be positively influenced. How can an organization promote the facilitators and diminish the inhibitors? Once that is understood, the organization can create the circumstances under which new technology can be introduced and the productivity of the organization increased as the expected gains in individual productivity are realized. The important influences on linkages are likely to be found in the nature of the relationships that exist between work units, differences in the states and structures of different work units, and differences in the processes that operate on the organizational linkages.
In Chapter 3, Goodman, Lerch, and Mukhopadhyay examine linkages, processes, and influences. Their goal is to develop a set of conceptual tools that will aid the analysis of linkages and promote the understanding of processes that influence linkages. To that end, they introduce and examine five processes that facilitate changes in individual and organizational productivity: (1) coordination, (2) problem solving, (3) focus of attention, (4) organizational evolution, and (5) motivation. In addition, they present a set of important concepts and definitions, an analytic strategy that addresses linkage inhibitors and facilitators, and a set of hypotheses about organizational linkages.
Linkages and influences are addressed to some extent in each chapter, which provides different perspectives on the issues identified. Some chapters focus on understanding the inhibitors and facilitators of linkages; others are more concerned with addressing the measurement issues raised by linkage concepts. Across all chapters there is a mix of theoretical considerations, interpretation of research findings, and exploration of linkage issues and processes in specific work domains. In addition, because the panel believes that understanding organizational linkages will require additional research, each of the chapters makes recommendations for research that derive from the discussion in that chapter.
Multiple Levels of Reciprocal Linkages
The concept of organizational linkages provides a useful framework within which to examine the productivity paradox. It has led the panel to conclude that one contributor to the productivity paradox is the common attempt to initiate change through the introduction of a single intervention (technology) at a single level in the organization (the individual). In Chapter 4, Schneider and Klein state this conclusion explicitly:
Changing a single aspect of an organization almost never results in a substantial change in organizational performance. Organizations are too complex, their performance too multidetermined, and their inertia is great for a single innovation at the individual level to have a substantial impact on organizational performance.
Schneider and Klein address organizational linkages in the domain of office automation—the application of information and communication technology to tracking, monitoring, recording, directing, and supporting information in the workplace. The launching point for their analysis is the report of the study ordered by the U.S. Congress to de-
termine why office automation has not yielded the improvements predicted (Office of Technology Assessment, 1985).
There are several reasons why innovations such as office automation can fail to yield improvements in organizational productivity. The introduction of new systems may not contribute to productivity at any level because the systems are not successfully implemented. Even if a system is successfully implemented and used as intended, it may do little to enhance, and in fact may impair, individual productivity. Finally, even if the system does in fact augment individual productivity there may be no resulting improvements in organizational productivity. This leads to the requirement for an organizational systems framework to clarify the multiple reciprocal linkages that determine organizational productivity.
An understanding of important processes that affect productivity, within the complexity of organizational linkages, can also be gained from decomposing the productivity paradox—identifying and examining the factors that produce the paradox. This is the approach taken by Pritchard in Chapter 7. He identifies three main types of factors that might account for the paradox: structural characteristics of the organization itself, intervention side effects (unintended consequences of the intervention), and problems associated with the measurement of organizational performance.
A number of structural factors associated with multiple levels of reciprocal linkages are identified and addressed in Chapter 7 and in other chapters of this report. An example is time lag. Because of the way the task is structured, improvements at one level can sometimes take considerable time to show up as improvements in the combined outputs at a higher level. Other examples of structural factors include slack in the process, the degree of centrality of the task to the process, and the degree of interdependence of work units.
An example of an intervention side effect is changing the focus of the effort, inappropriately, from one unit of analysis to another—the introduction of computers might result in low task interdependence when high task interdependence is required. Individual productivity might increase as a consequence, but the output of the group as a whole might decrease. Other side effects might be descriptions in communication patterns and the socialization process, or the generation of resistance to change as a consequence of the way in which the intervention is introduced.
Measurement issues have been a principal concern of the study panel. The productivity paradox could, of course, be explained by the inad-
equacy of the measures and processes used to assess organizational productivity. Moreover, the study of productivity linkages among multiple levels of analysis would surely require measurement methods of considerable complexity and sophistication to provide the required validity and sensitivity. Thus, this volume addresses measurement from two important perspectives—as an explanation of the productivity paradox and as a critical tool in the understanding of organizational linkages.
The panel deliberated at length about the appropriate concept and definition of productivity within which to address measurement issues, but without satisfactorily resolving the issue. Perhaps the panel is not alone in being unable to arrive at a consensus. In a review of the literature on productivity, Pritchard (1991) found that the term productivity was used to encompass constructs as diverse as efficiency, output, motivation, individual performance, organizational effectiveness, production profitability, cost-effectiveness, competitiveness, and work quality. Further, productivity measurement was used interchangeably with performance appraisal, production capability assessment, quality control measurement, and the engineering throughput of a system.
Most panel members held one or the other of two positions regarding the concept of productivity. Some wanted to define productivity as the ratio of outputs to inputs, in line with the original definition of the term by labor economists. They believe that this is the only definition that is unique to the concept. Others argued that this definition is too restrictive. They believe that productivity must encompass concepts such as quality and effectiveness to be meaningful. The panel's solution was to adopt the systems model of organizational performance (described by Sink and Smith in Chapter 6). In this model, productivity is but one of seven interrelated and interdependent criteria of organizational performance. The seven criteria, each of which is operationally defined in Chapter 6, are productivity, effectiveness, efficiency, profitability, quality, quality of work life, and innovation.
Within this model, productivity provides just one part of the total performance picture. The total picture requires the examination of all seven criteria, each of which might necessitate several different measures. The approach advocated in Chapter 6 is to consider the seven criteria as variables that explain variation in performance. Variables can be included and excluded from the analysis to determine which ones explain variation in performance for a particular work unit relative to a specified objective. In Chapter 6, Sink and Smith address
measurement issues and the design of measurement systems within the framework of this model. They conclude that the paradox of unrealized productivity improvements results from incomplete systems thinking and from failure to understand the nature of linkages among the individual, group, and organizational levels.
Measuring Individual Productivity
The characteristics of measures of individual productivity determine the extent to which the measures can be aggregated or related to higher levels of analysis. These characteristics—the definition and scope of individual productivity, the measurement systems employed, and the specific measurement metrics—are the focus of Chapter 5. In that chapter, Ruch introduces a variety of concepts, such as goal alignment, that put these measurement issues in perspective. Goal alignment is the ideal in any system that is assumed to be driven by goal-based measures. In such a system, individual productivity depends on the extent to which individual measures are in line with organizational goals and the extent to which those goals form a logical hierarchy across organizational levels. He also presents two models that provide alternative views of factors affecting individual productivity. The models encompass such variables as individual characteristics, psychological factors, sociological factors, technology, and the characteristics of systems and organizations. He then extends his analysis to examine four key measurement issues to be considered when individuals become groups—complexity, input factors, aggregation, and goal alignment.
Measurement and Its Implications for Research
In Chapter 8, Campbell addresses models of measurement and their implications for research on the linkages between individual and organizational performance. His central argument is that effective measurement depends on the substantive specification of productivity in the specific domain of interest, such as a specific aspect of IT. In support of this approach, he provides a hierarchical measurement model for research application that is consistent with the models provided in Chapters 5 and 6.
Domain-Specific Examinations of Linkages
The chapter authors provide many examples in an attempt to clarify the concepts and ideas they introduce. To reduce further the level of abstraction in this report, some chapters examine linkage issues within
a specific domain. The objective is to determine how linkages, and the processes that influence them, actually work within operational organizations. The panel considers this exercise to be a reality check on the formulations that emerged from its individual study and group deliberations.
Coordination in Software Engineering
Software engineering is the domain of Kiesler, Wholey, and Carley's examination in Chapter 9 of the role of coordination in the linkage between individual and group productivity. The technical project team approach employed in software engineering is actually a paradigm for how various types of technical work are now accomplished. Organizations create a project team when the required technical tasks transcend the assigned functions or capabilities of individuals. A project team can range in size from two to several hundred members, and the membership of larger teams is relatively diverse. The interdependence of tasks and jobs make coordination—those activities required to support group work—a critical factor in team productivity.
In Chapter 9, Kiesler, Wholey, and Carley discuss what is known about coordination in groups and apply that knowledge to the problem of coordination in software development teams. They show that the traditional model of coordination, with its emphasis on sharing ideas through direct communication, is not applicable because of the complexity, uncertainty, and interdependence that characterize software engineering. They emphasize team design and team communications as positive approaches to enhancing the linkages between individual and group productivity, and they provide a set of hypotheses relative to each approach. In addition, they present a set of stimulating research problems and directions, the pursuit of which will enhance understanding of linkages.
Productivity Linkages in Computer-Aided Design
Computer-aided design (CAD) has been introduced into engineering organizations with the expectation of increasing the productivity of the organization by increasing the productivity of individual designers. The principal output of a design effort is a set of design specifications that meet agreed design objectives, guidelines, and constraints. Thus, the core definition of productivity within this domain is the ratio of design-specification output to the input of resources, mainly labor. Productivity gains are anticipated from the capabilities of CAD to automate routine functions, enhance the accuracy and efficiency of design tasks, promote the exchange of information, facilitate the performance
of sophisticated design tasks, and integrate better the design and production processes.
In Chapter 10, Harris examines productivity linkages and influences within the CAD domain. Many of the issues discussed in earlier chapters are relevant to this type of information work. They are examined in Chapter 10 as they relate to linkages among designers, design teams, and engineering design organizations. The central question Harris addresses is, when CAD technology increases the productivity of individual designers, under what conditions will those increases lead to increases in the productivity of the design team and, in turn, the design organization? Among the influences he discusses are the degree of physical isolation of designers, the extent of task specialization in the design team, the mode of team supervision, the nature of controls on the flow and access of information, the burden of design support and coordination, the manner in which technology is implemented, resource management, and system quality and reliability. Each of these influences is examined relative to its impact on organizational linkages.
The Case of Downsizing
Organizational downsizing has been one of the major initiatives undertaken by firms in the United States during the past decade to increase productivity. Downsizing encompasses shrinking, retrenching, or consolidating the organization, principally by reducing the number of employees and hierachical levels. However, according to the evidence in Chapter 11, the anticipated effects have not been realized. The mounting evidence that downsizing initiatives do not yield commensurate gains in productivity is another form of the productivity paradox. It is a particularly troublesome version of the paradox because removing sizable amounts of overhead slack from an organization would be expected to lead directly to increased organizational productivity. Using the perspective of organizational linkages, Whetten and Cameron examine this productivity paradox within the domain of manufacturing organizations. They report the results of their analyses in the form of a set of myths regarding the best way to design and implement a downsizing program. The prevalence of these myths was verified by a survey of 909 businesses in the United States. The extent to which downsizing programs have been based on these myths helps explain the productivity paradox associated with downsizing. Whetten and Cameron further examine organizational linkages by comparing two approaches to downsizing and their impact on organizational productivity.
In Chapter 12, the panel summarizes the principal themes and common issues that run through the various chapters. These themes and
issues encompass the productivity paradox and organizational linkages. In this chapter the panel also presents the broader conclusions it reached. As noted, the more specific conclusions and recommendations for research are provided at the end of each chapter.
Campbell, J.P., R.J. Campbell and Associates. 1988. Productivity in Organizations. San Francisco: Jossey-Bass.
Committee to Study Computer Technology and Service Sector Productivity. 1994. Information Technology in the Service Society: A Twenty-First Century Lever. Computer Science and Technology Board, Commission on Physical Sciences, Mathematics, and Applications. National Research Council. Washington, D.C.: National Academy Press.
Drucker, P.E., 1991. The new productivity challenge. The Harvard Business Review. November–December:69–79.
Johnson, W.B., and A. Packer. 1987. Workforce 2000: Work and Workers for the Twenty-First Century. Indianapolis, Ind.: Hudson Institute.
Office of Technology Assessment. 1985. Automation of America's Offices . Washington, D.C.: U.S. Government Printing Office.
Pritchard, R.D., 1991. Organizational productivity. In M.D. Dunnette and L.M. Hough, eds., The Handbook of Industrial and Organizational Psychology, 2nd ed. Palo Alto, Calif.: Consulting Psychologists Press.
Sanders, M.S., and E.J. McCormick. 1987. Human Factors in Engineering Design, 6th ed. New York: McGraw-Hill.
Sawhill, I., and M. Condon. 1992. Bidding war—or growth? The Washington Post February 27.
Sink, D.S., and T.C. Tuttle. 1989. Planning and Measurement in Your Organization of the Future. Norcross, Ga.: Industrial Engineering and Management Press.
Weiner, E., and A. Brown. 1989. Human factors: The gap between humans and machines. The Futurist May–June:9–11.
Wohlers, E., and G. Weiner. 1988. Employment Trends in the United States, Japan, and the European Community: A Comparative Economic Study. New Brunswick, N.J.: Transaction Books.