From Business and the Contemporary World , Vol. V, No. 4 Fall, 1993


Impact of Information Technology on
Societal Productivity and Employment*

by

Terence Krell
Associate Professor of Technology Management
Western Illinois University
3614 24th Street
Rock Island, Illinois, 61201 USA
309 793-1998

and

Jeffrey Gale
Professor of Management
Loyola Marymount University
Los Angeles, California, 90045 USA
310 338-7406





*An earlier version of this article entitled "Societal Impact of Information Technology Induced Change on Middle Managment in the 1990s" was published in the Proceedings of the IFSAM conference High Technology and Management, September 7-9, 1992, Tokyo, Japan

Impact of Information Technology on
Societal Productivity and Employment
ABSTRACT
Numerous observers have lamented the apparent lack of productivity gain derived from the substantial investment by corporations in information technology during the 1980s. This assessment is challenged through the proposal of a model relating information technology development and diffusion, changes in managerial activity and organizational process, and macroeconomic performance. As a result of the interplay of these forces the productivity gains are being realized at the organizational level and are reflected in the current wave of middle management layoffs in the US. This change in the middle management structure has long term impact on the society, and similar patterns may be developing in other countries.
INTRODUCTION
Investment in information technology (IT) has become a significant capital expenditure for U.S. firms, estimated to have risen from 27.2% of all capital spending in 1981 to 35.2% in 1991 [1]. Shipments of all computer equipment in the U.S. more than doubled from 1979 to 1989 to a peak of over $54 billion, prior to dropping slightly in 1990. These figures understate the actual computing power derived from the investment because of a decrease in the cost/performance ratio of 20-30% per year [2]. This heavy investment in IT began in the late 1970s as an attempt to remedy the productivity disparity between white collar office workers and production workers through increasing the capital investment for white collar to more closely match that of production workers [3].
Many claim the promised increase in productivity has not come to pass. In a series of research reports, Stephen Roach of Morgan Stanley has documented a continuing stagnation in IT productivity payoff [4]: "While there are many factors behind such disappointments, technology has quite simply not delivered its long awaited productivity payback." The popular press, including Business Week and Fortune , has questioned the payoff as well [5]. Other academic studies have made similar points and some managers are questioning the productivity returns from IT investment [6].
We challenge the view that investment in IT has not resulted in a productivity increase and propose a model that relates diffusion-based employment effects at the firm and societal levels to macroeconomic performance and structural change in the economy. As a result of the interplay of these forces, the productivity gains are being realized at the organizational level and are reflected in the current downsizing of U.S. Corporations especially in the middle management ranks [7].
THE GREAT IT PRODUCTIVITY DEBATE
In attempting to make sense of the observed lack of productivity increase traceable to the substantial investment in IT in the U.S. economy, the explanations fall into two distinct groupings. One explains the phenomenon in terms of inadequate definition and measures of productivity. The other explains the failure in terms of inappropriate use of the technology.
Poor Measures of Productivity
The first of the explanations is based on measurement issues. For example, in addressing the series of analyses by Roach, Panko [3] focuses on the technical means by which productivity is calculated in the studies. The labor economics literature follows the U.S. Federal government's Bureau of Labor Statistics (BLS) measures in its published data series. Both inputs and outputs are measured in terms of units or deflated costs of inputs and outputs. Consequently numerous measures appear, most of them measuring output per hour, which assumes labor input as the most significant one. Most of the series are aggregate and both governmental and non profit sectors are excluded. There are output problems with financial institutions as well.
Panko suggests that a multi factor productivity analysis is more appropriate, since output per hour measures substitution as well as productivity, making its meaning murky. This problem may not occur using multi factor approaches. BLS has published a limited number of such series since 1983 but none were used in the Roach studies. Under more strict criteria which limit analysis to industries with limited substitution effect communications, commercial banking, and office intensive U.S. Government functions substantial productivity gains are present.
Others argue that using productivity as generally calculated for an efficiency measure is particularly inappropriate in evaluating white collar use of IT. Rather, they suggest, a measure of effectiveness is more realistic. Davis [8] suggests an approach which would combine effectiveness, efficiency and innovation in judging the impact of IT. Strassman [9] argues for a "value added" productivity measure as more appropriate.
Inappropriate Use of the Technology
The other explanation is a criticism of the manner in which IT is used. Leavitt and Whisler [10], in 1958, hypothesized that full use of IT would include a substantial rethinking of management activity in firms. Recently, Applegate, Cash and Mills [11] suggest that a more complete use of IT would include new organization forms (clusters), centralized control with decentralized decision making, and more autonomous managers with less narrowly defined jobs.
Izzo [12] argues that a strategy for restoring productivity involves substantial rethinking of the nature of the systems in the organization with a move to a more decentralized architecture, active executive involvement and integration of the technology. Keen [13], conceding the limited impacts, makes the most extensive case for a complete redesign of the organization around the use of IT. Seven components of business design competitive positioning, geographic positioning, organization, human capital use, information capital economics, the IT platform, and business technology alignment are involved.
Both of these interpretations have merit, yet they miss the mark. IT has increased the potential productivity of the adopting firms. Even under traditional productivity measures, despite a less than optimal use of the technology, potential gains are there. Firms must, however, be willing to take other actions such as reductions in their workforce or reorganization of their internal processes, to realize these gains.
DIFFUSION AND ADOPTION OF IT
A useful approach to understanding IT productivity effects through employment impacts is the examination of the diffusion of the technology throughout the society and the pattern of adoption in the individual firms in the society through the use of the concepts developed in the literature. Taking the employment effects of individual firm adoption and aggregating them based on the societal adoption pattern, we propose an underlying pattern for the societal employment which results.
IT Technology Diffusion
Although there are no explicit analyses of the diffusion of IT in the U.S. economy, the phenomenon of technology diffusion in general has been extensively studied and the concepts developed are useful in understanding the spread of the technology. All users of a new technology do not adopt it at the same time. Rather, as developed by Rogers [14] in his pioneering work, they separate into adopter categories based upon a bell shaped curve over time.
Innovators (2.5% of adopters) are venturesome and risk takers adopting the technology first. Early adopters (13.5%) successful and discrete users of the new and are often opinion leaders. The Early Majority (34%) adopt just before the average member and are characterized by their deliberate decisions about adoption. The Late Majority (34%) approach innovations with great caution and skepticism. Finally, the Laggards (16%) are the last to adopt being firmly rooted in the past. Following this bell shaped curve, and building on a two step communication model, Rogers further characterizes the societal diffusion pattern as an S shaped (sigmoid) curve when cumulative adoption is mapped against time, which has proven to be the standard pattern for a number of innovations. Numerous explanations of this curve have been proposed [15], and other writers, notably Bass [16], have examined the pattern of innovation when other influence patterns, including both internal and external influence, are present.
For our purposes here, Rogers' adopter categories and the standard S shaped curve for societal diffusion of a technology will be utilized. Data on sales of IT equipment, reflective of diffusion patterns, are supportive of that position (See Figure 1.)
FIGURE 1: Shipments of Computer Equipment ($ billions)

Source: U.S. Annual Survey of Manufacturers, 1991
IT Adoption Patterns in the Firm
Several significant efforts have sought to examine the pattern of IT adoption within organizations. Working from Rogers efforts, they build upon the concept of stages of development in the adoption of the technology within the organization.
The earliest of these attempts to develop such a framework was Gibson and Nolan's [17] 1974 theory of EDP growth. This model attempts to explain the adoption within the EDP department in terms of four distinct stages. "Initiation" is the beginning use of the technology with early success leading to more interest. "Expansion" finds a proliferation of applications in functional areas. "Formalization" leads to control over the proliferation and measured evaluation. The final "Maturity" stage represents full adoption. The model has remained useful despite subsequent criticism [18][19].
Taking a broader organizational approach, McFarlan and McKinney [20] break the adoption/diffusion of new technologies in the organization into four phases.
Phase 1. "Technology Identification and Initial Investment" emphasizes the learning and application of the new technology.
Phase 2. "Experimentation and Learning" focuses on raising user awareness of the technology, assessing staffing and skill requirements, equipment acquisition and technology adaptations.
Phase 3. "Rationalization and Management Control" is characterized by the implementation of controls on the technology for short term efficiencies, upgrading skills, and cost effective implementation.
Phase 4. "Widespread Technology Transfer" disseminates the technology widely in the organization.
This model emphasizes the learning process within the organization and seeks to guide managers in the appropriate activities to further the implementation.
Employment Impacts at the Firm and Societal Level
The latter phase model, which addresses organization wide implementation of IT, is a useful basis from which to examine the impacts on firm employment needs. We assume that these needs are rooted in the firm's processes at a given level of production. The production level may be adjusted upward to absorb any now unused labor, or overall employment may be adjusted by reassignment or by trimming the size of the workforce.
Figure 2 shows the pattern of organizational employment resulting from this firm implementation pattern. At the early phases, some technical help may be added to deal with the new technology, especially if current staffing is geared to system maintenance activities. As the firm seeks to move the technology broadly into the organization, additional technical staff are needed for design, implementation and training. At the later stages of the implementation, this additional technical staffing may no longer be required.
FIGURE 2: Individual Firm Employment During IT Adoption


At the same time, the impacts of the technology itself are felt in the organization, decreasing the necessary personnel for a given level of output. Capital substitution leads to employment reduction in clerical positions and hence to reduced supervisory and administrative needs as well. Broad IT induced redefinition and improved efficiency of the organization as a whole may also lead to reduced demand for middle managers.
While it is true that the implementation of IT may be geared to increasing the quality of decisions, as argued in much of the productivity literature, the net effect, at a given level of production is a reduction of the labor requirement. Osterman [21] examined the employment level impacts of the implementation of mainframe computers on clerks and managers. In the forty industries studied, from 1972 1978, a 10 percent increase in computer power was associated with a 1.8 percent decrease in clerical employment and a 1.2 percent decrease in managerial employment at a constant output level. The entire issue is clouded by the substitution effect and the growth of output which might result from the implementation.
FIGURE 3: Societal Employment and Technology Adoption

The societal diffusion pattern of IT described above, when combined with the individual firm adoption patterns and subsequent employment effects, yields, at the same level of output, an aggregate societal employment effect presented in Figure 3. There is a substantial time lag in this pattern from the employment additions due to early adopters until the decreases due to full implementation by the majority of the firms in the economy are in place. The start of the substantial IT investments in the US, some 10-15 years ago, would lead to the present overall decreases.
IT AND THE MACROECONOMY
An evaluation of the impacts of IT diffusion must also include consideration of the macroeconomic context in which it occurs. Changes in the structure of the U.S. economy and the increased integration into the world economy with global competitive pressures must be examined as well as the influence of the business cycle.
The underlying structure of the U.S. economy has continued to change during the diffusion of IT. The most important aspect of that change has been the shift to a more services-based structure. According to the BLS, from 1980-1989, services employment increased from 24% to 30% of the private sector total and increased in absolute terms by over 50 percent. At the same time, manufacturing employment decreased by 4 percent [22]. One of the early observers of this trend, Daniel Bell, characterized the result as a "post industrial society" and the shift as largely to information work [23]. The greater intensity of IT usage in the services sector has been well documented, for example, a McKinsey & Co. study showed that between 1981 and 1985, banks spent $30 billion on IT [24].
Unfortunately, the productivity in the service sector in the 1980s was extremely poor, despite IT investment. While manufacturing experienced a 4.1 percent average gain, spurred on by increased competition, services showed only a 1 percent average annual gain. The service gain was even lower than in the 1970s, and, in the period 1982-1986, nine of ten new jobs created were in the services sector [25].
In addition, the greater openness of the U.S. economy to global competition has increased pressures on business, primarily manufacturing, resulting in both increased competition at home and demands to expand in world markets to make up for losses at home [26]. The response has been a 50 percent increase in real capital expenditures from 1980 to 1989, which includes IT investment. The increase in manufacturing productivity described above is one result [27].
Finally, the pattern of business decision-making in the business cycle has an influence as well. Employers are relatively slow to trim their labor force during the high parts of the cycle. However, when the business cycle enters a downturn resulting in recession, the pressure and willingness to make necessary workforce adjustments increases. The U.S. recession which started in 1988 has provided just such an opportunity for adjustment and the increase in unemployment, especially among the managerial ranks, is a result.
A PROPOSED MODEL
It is the interplay of these forces, societal diffusion of IT, individual firm adoption of IT, and macroeconomic structure and performance which determines the societal employment impacts of IT. The linkages between these various factors are displayed in the model presented in Figure 4.









FIGURE 4: A Proposed Model of Societal Impact of IT

The societal diffusion of IT occurs in a pattern represented by the S-Curve. The speed and extent of the diffusion is influenced by the competitive pressures from the macroeconomy and by individual firm experiences with the technology. Individual firms initiate the adoption of the technology at different times depending upon their adopter category. The pattern within the firm follows the four stages ending in technology transfer throughout the firm. The process of adoption within the firm ultimately has an effect on the firm's potential productivity, and on the level of employment at the firm. The individual firm employment, at a constant output, increases as it moves through the phases but declines with full implementation. Societal productivity potential is also affected. When the individual firms are aggregated over the society as a whole, with diffusion according to the S-curve, an increase in employment initially occurs but overall employment, both among technical IT specialists and among the managers and clerical personnel, decreases on a lagged basis with full societal diffusion. We suggest that this is the case in the transition to the 1990's.
Societal diffusion of technology also has an indirect effect on societal employment and societal productivity without direct adoption by firms, for example, students use computers to produce term papers, having more time to spend on other projects. The diffusion of technology also changes the cultural climate as the people come to expect more use of technology.
At the same time, the competitive conditions in the macroeconomy exert an influence on the IT-related societal employment situation as well. A more competitive economy, with more cost pressures resulting, leads to an accelerated adoption by individual firms and to earlier attempts to fully realize IT productivity increases by labor force savings. In addition, a downturn in the business cycle, with the resulting pressures on profits and slower growth, increases pressures to fully realize the gains and provides management of firms with a basis for making labor force adjustments.
In the U.S. during much of the 1980s, the economy was in the growth stage of the business cycle and sales of IT grew rapidly, stabilizing at roughly one third of new capital investment by the end of the decade. Managers were added in organizations that anticipated continued growth and based on pre-IT work patterns and organization processes. As more firms adopted the technology, the demand for technological experts increased, driving up the prices of those skills needed, and putting pressure on others to gain the skills necessary to operate IT.
A broad economic imperative developed in the U.S. in the late 1980s because of the cyclical downturn in the U.S. economy after expansion most of the decade; the structural shift to services, with their low productivity; and exposure to increased global competition in the mid- to late 1980s. In addition, a narrower economic imperative developed for those companies that had taken on debt and restructured in the mid 1980s. The imperatives thus created a situation in which the realization of the potential productivity from 15 years of IT diffusion could no longer be avoided. The gains which had been enabled by the large scale investment were now realized through the mechanism of labor force adjustment. There has been a substantial and continuing pattern of middle management cutbacks since 1986. White-collar productivity, declining much of the decade, finally started increasing in 1989 as a result [1].
The model suggests that these large scale management cutbacks, enabled by the imperative, are significantly IT induced, are based upon realization of real productivity gains, and are long term in nature. It also implies that the continuing presence of the economic imperative will cause companies to seek further productivity gains, and consequently further deplete the ranks of middle management.
NATIONAL DIFFERENCES IN SOCIETAL IMPACT
This basic process is not culture bound but is a part of global competition. Impacts, however, will vary based on societal norms, position of the country in the world economy, and national capital availability allowing organizations to make further capital investment in IT. There is a further dynamic in the willingness of the society and the political institutions to allow the process to have its full effect.
The most significant societal initiative to limit the full impact of the dynamic is government employment policy which varies with jurisdiction. For instance, German and Swedish policy attempts to limit the effects of labor market disruption through active job retraining and public initiatives to provide placement. France, Belgium and the Netherlands act to restrict hours worked ensuring broader employment. The Japanese system, in contrast, is centered in the private sector employment relationship and societal norms regarding employment practices. It is in the U.S., Canada, and the U.K. where the individual must bear the most cost for the societal effects of the dynamic. [28]
CONCLUSION
The adoption of IT in the U.S. through the 1980s resulted in an increase in capability which would ultimately lead to a decreased demand for middle managers. Rather than an instant change, this adoption of IT was a dynamic process. Application of diffusion models from the literature suggests both a time lag in reaping benefits and a temporary increase in personnel in order to adopt and implement the technology: Managers were added in organizations that anticipated continued growth and based on pre-IT work patterns. The subsequent implementation of IT resulting in greater productivity, when combined with an economic turndown culminated in lessened need for these middle managers.
We expect that an economic upturn will not restore middle management positions that have been lost as organizations alter their structures to accommodate the current state of IT. Will the overall increase in capability result in new industries and hence employment being created? That is the key question.
REFERENCES
[1]Coy, P., "The New Realism in Office Systems", Business Week , 6/15/92, 128
[2] Morton, S.. The Corporation of the 1990s , (New York: Oxford University Press, 1991) 9
[3] Panko, R. R., "Is Office Productivity Stagnant?," MIS Quarterly , June 1991, 191 203.
[4]. Roach, S. S., "America's Technology Dilemma: A Profile of the Information Economy," Economic Perspectives , (New York: Morgan Stanley, April 22, 1987.) and Roach, S. S., "Technology and the Services Sector: America's Hidden Competitive Challenge," in Guile & Quinn, eds., Technology in Services: Policies for Growth, Trade and Employment , (Washington, D.C.: National Academy Press, 1988) 118-138.
[5]. Bowen, W., "The Puny Payoff from Office Computers," Fortune , May 26, 1986, 20. and "Office Automation: Making It Pay Off," Business Week , October 12, 1987, 134-138.
[6]. Curley , C.F., "Are There Any Real Benefits from Office Automation?," Business Horizons , July-August, 1984, 37-42. and Klein, R.L., "Does Automation Necessarily Mean An Increase in Productivity?," Journal of Systems Management , May 1985, 32-34.
[7] Fischer, A. B. "Morale Crisis", Fortune, 11/18/91, 70
[8] Davis, T.R.V., "Information Technology and White Collar Productivity," Academy of Management Executive , Vol. 5, No. 1, 1991, 55 67.
[9] Strassman, P.l A., Information Payoff , (New York: Free Press, 1985.)
[10] Leavitt, H. J. and T.s L. Whisler, "Management in the 1980's," Harvard Business Review , Nov. Dec. 1958, 41.
[11] Applegate, L. M., J. I. Cash, Jr. and D. Q. Mills, "Information Technology and Tommorow's Manager," Harvard Business Review , Nov. Dec. 1988, 128 136.
[12] Izzo, J.A., The Embattled Fortress , (San Francisco: Jossey Bass, 1987)
[13] Keen, P. W., Shaping the Future , (Cambridge: Harvard Business School Press, 1991).
[14] Rogers, E. M., The Diffusion of Innovations , 3rd. ed. (New York: Free Press, 1984).
[15] Mahajan, V. and R. A. Peterson, Models for Innovation Diffusion , (Beverly Hills: Sage, 1985).
[16] Bass, F. M., "A New Product Growth Model for Consumer Durables," Management Science , Vol. 15 1969, 215 227.
[17] Gibson, C. F. and R. L. Nolan, "Managing the Four Stages of EDP Growth," Harvard Business Review , Jan. Feb. 1974, 76 88.
[18] Benbasat, I, A.S. Dexter, D.H. Drury and R.C. Goldstein, "A Critique of the Stage Hypothesis: Theory and Empirical Evidence," Communications of the ACM , Vol. 27, May 1984, 476 485.
[19] King, J. L. and K. L. Kraemer, "Evolution and Organizational Information Systems; An Assessment of Nolan's Stage Model," Communications of the ACM , Vol. 27, May 1984, 466 475.
[20] McFarlan, F. W. and J. L. McKinney, Corporate Information Systems Management , (Homewood, IL: Irwin, 1983)
[21]Osterman, P., "The Impact of Computers on the Employment of Clerks and Managers," Industrial and Labor Relations Review , January 1986.
[22] U.S. Bureau of Labor Statistics, BLS Bulletin 2370 , September 1990.
[23] Bell, D., The Post Industrial Society , (New York: Basic Books, 1962).
[24] Porat, M.U.., The Information Economy , (Washington, D.C.: Office of Telecommunications Policy, 1977).
[25] Samuelson, R., "The Squeeze on Services," Business Week , January 25, 1988, 50.
[26] Bureau of Labor Statistics, Outlook 2000 , (Washington, D.C.: U.S. Department of Labor, 1990), 21.
[27] U.S. Bureau of the Census, Plant and Equipment Expenditures and Plans , December 1990.
[28] Blyton, P., "New Technology and Future Work Patterns," in Piercy, N., The Management Implications of Information Technology, (New York: Nichols, 1984), 122-139. and Osterman, P., Employment Futures , (New York: Oxford University Press 1988), 108-134.