Click for next page ( 25


The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 24
1 Introduction and Impact of Information Technology at the Macroeconomic Level Information technology has changed the way that most companies do business. It is hard to imagine returning to an era in which payroll account- ing, check processing, airline reservations, or stock exchange trades were handled without computer technology. In a typical business day now, over 200 million shares are traded on the New York Stock Exchange alone; in 1989, U.S. commercial banks processed almost 58 billion payment transac- tions.~ Transaction volumes of this magnitude would have been impossible to handle with the technology of 50 years ago. Innovations in electronics made over the past 40 years have been incorporated into new computers, office automation equipment, communications equipment, and their associ- ated software systems. Information technology has revolutionized services both in service-sector industries and in the service activities that are an increasingly integral part of goods-producing industries. For purposes of this report, the service industries consist primarily of transportation, communications, wholesale and retail trade, financial ser- vices, insurance, real estate, utilities, and personal and professional ser- vices; unless otherwise specified, information technology (IT) includes computer and communications hardware, as well as the software and associated ser- vices required to exploit that hardware. Despite the obvious and dramatic changes in the economy brought about by use of IT, a strange paradox emerges from analysis of aggregate U.S. economic data. According to standard methods of calculation, productivity in the business sector has grown slowly since the widespread introduction 24

OCR for page 24
IMPACT OF IMFORMATION TECHNOLOGY AT THE MACROECONOMIC LEVEL 25 of IT. However, the slow growth of measured U.S. productivity has not been occumng uniformly throughout the economy. Measured productivity in the manufacturing sector, after slowing in the 1970s, has been growing at a rate of about 3 percent a year, as gauged by the average annual rate of increase in measured output per labor-hour over the past 10 years. The farming sector has also achieved solid growth in productivity. Much of the rest of the economy, however, has lagged behind especially construction and elements of the service sector. The result is that business as a whole has achieved only about a 1-percent-a-year improvement in output per hour in recent years. Particularly puzzling is the weak measured performance of the service sector, which has accounted for the bulk of the nation's expendi- tures in IT. The dilemma is this: although data show that over $750 billion has been invested in IT hardware alone in the 1980s (Figure 1.1), and although 120 110 100 ~ 90 o a, o In o ._ m 80 70 60 50 40 - 30 20 - 10 O- . 1981 1982 1983 1984 1985 - - 1986 1987 1988 1989 Year FIGURE 1.1 Growing investment in information technology (IT) by the service sector. The only systematic data available on IT expenditures account for hardware components but exclude software and services. Thus this figure shows expenditures only for office, computing, and accounting equipment, communications equipment, instruments, and photocopy and related equipment; it does not include expenditures for software, electronic information services, data processing and network services, computer professional services, custom programming, systems integration, consult- ing, or training services. SOURCE: Stephen Roach, Morgan Stanley & Co.

OCR for page 24
26 INFORMATION TECHNOLOGY IN THE SERVICE SOCIETY BOX 1.1 One Assessment of the Information Technology Paradox Stephen Roach's assessment of the information technology (iT) para- dox hinges on what he believes are the unique characteristics of the service sector's competitive environment. While U.S. manufacturing firms faced intense competition in the 1980s, service companies were pro- tected by regulation, he says, and by the lack of foreign competition in effect shielding them from the harsh realities of market pressures. In response, he maintains, service companies became complacent and al- lowed their costs to become increasingly bloated, leading to ~ most disappointing record of productivity growth that is consistent with the macroeconomic evidence presented in Chapter 1. Roach points out that the service sector's lagging efficiencies are particularly disturbing because it is the segment of the economy that has invested most heavily in IT. Indeed, over the past 10 years, the service sector spent $862 billion on IT hardware alone-accounting her about 85 percent of the nation's total IT expenditures This surge of spending has been enough to equip the average white-callar worker In the service sector with about $10,()00 worth of IT hardware in 199'-essentially double the endowment Mat existed in 1980~ Such extraordinary commit- ments to IT have not yielded measurable benefits far productivity, Roach maintains, because service companies have had little incentive to strive for efficiencies, given the unique protection they enjoyed from compet'- tive realities. As a result of this rapid spending on IT, Roach argues, the service sector has experienced an ominous transformation of its cost structure. Service companies used to be largely vartabJe-cost producers, win the salaries at white-collar workers constituting the bulk of overall expenses. When business conditions toughened, wh':e-collar job aeration was an effective means of cost control. Now' however, courtesy of the massive infrastructure of installed IT, service eornpan~es have a new layer of fixed costs in effect transforming them from variable- to increasingly fixed- cost producers. The commitment to IT has, therefore, crimped the abiT'ty of service managers to adapt to changing business conditions. Roach also stresses that this lack of flexibility has come at a most inopportune n~omer~t for services-~ point in time when co~npetit'Ye pressures have intensified dramatically. Regulation has been replaced by deregulation. Moreover, the pressures of foreign competition have been increasingly evident in the farm of a dramatic increase ire foreign direct investment in the U.~. service sector inflows of Capital that have deep- ened the packets of many players battling for market share in the world$s richest and deepest service market. in response to this intensified competitions a new wave of restruc- turing has been unIcashed in the vast service sector-highlighted in the

OCR for page 24
IMPACT OF IMFORMATION TECHNOLOGYAT THE MACROECONOMIC LEVEL 27 last ~ years by nearly a dozen mega-bank mergers, the Failures of a number of major airlines, and dramatic downsizing In retailing, telecom- munic~ions, accounting, advertising, and the legal profession In each of these industries, wh~te-coliar jobs are now being permar~er~tly eliminated. Roach maintains that this development would not be occurring if IT were delivering what it was intended to deliver increased output through meaningful improvements in white-collar productivity. Thus, he views the reality of serv~ce~sector restructuring as an independent confirmation of the IT paradox-a growing recognition by managers that there Is, in fact, considerable Validity to the disturbing results conveyed through the macroeconomic data on measured produ~:iY'ty. At the same time, he also views restructuring as a powerful catalyst that is now forcing manag- ers to rethink ap,olicat~ons of IT setting In motion a process that could well culminate in Me long-awaited macroeconomic payback from the use of IT SOURCE: Stephen Roach, Morgan Stanley ~ Co. experienced executives and observers of business note that IT has contrib- uted to important positive changes in the way that business is conducted, standard measures of productivity at the macroeconomic level still reveal few benefits of such changes over the past two decades.2 What are some possible explanations?3 . One possibility is that there has been wasteful or inefficient use of information technology in the service sector. Although it provides a power- ful tool, IT can be used well or badly. The view that mismanagement has kept the new technology from providing much impetus toward improving productivity has been expressed in the past by Stephen Roach, a member of the committee (Box 1.1~.4 An alternative view is that the use of IT has in fact raised produc- tivity significantly and that other problems have caused the slowdown in productivity growth. IT is only one of several factors that affect productiv- ity. Thus the slow growth of measured productivity in services may be the result of such problems as poor performance in U.S. schools, weaknesses in technology development in areas other than IT, changed energy prices, gen- erally inadequate investment in capital equipment or structures, including infrastructure, and/or a shift in employment to sectors that present limited opportunities for increasing productivity. Or perhaps a slowing of growth in measured productivity was inevitable as the economies of industrialized countries matured and the relatively easily obtained gains in productivity

OCR for page 24
28 INFORMATION TECHNOLOGY IN THE SERVICE SOCIETY were realized. In fact, most other industrialized countries have also slowed in measured growth at about the same time as the U.S. economy.s . A third perspective on this issue is that IT may have been used to transform business activities but that these changes are being missed by current methods of measuring productivity.6 In this view, the use of IT is bringing about real increases in productivity that are not captured in the data. Thus the slowdown in productivity growth may be partly or wholly the result of errors in measurement. A fourth view argues that IT-based systems have the potential to raise the productivity of the economy significantly, but that this result will take time. Economic historians have pointed out that major changes in technology, such as the introduction of steam power or electrification, have affected productivity only after many years.7 Thus, the payoff from invest- ments in IT hardware may have been delayed because software develop- ment and systems reengineering have lagged behind and because it takes many years to train people to use new technology productively. This opti- mistic view of the potential for future growth in productivity suggests that payoffs can be expected in the years ahead as U.S. service industries learn to use IT more effectively. . . A fifth view underscores the importance of examining the impact of IT at much lower levels of aggregation than is possible using macroeconomic data. A 1993 study by Erik Brynjolfsson and Lorin Mitts found that infor- mation technology "has made a substantial and significant contribution to the output of firms," with a return on investment for IT that is significantly greater than that of other capital investments. These various alternative approaches to explaining the IT paradox are not mutually exclusive. It could well be that some mistakes were made in early adoptions of IT, leading to wasteful investment, but that as software development, worker retraining, and management understanding progress, investments will pay off in measurable forms. For example, IT may be applied initially to increase information about markets, whereas productiv- ity may increase only later as the information is used to improve operations. In considering these alternatives it may be useful to keep in mind the difference between the average and the marginal contribution of using IT. The average contribution of using IT can be determined, in principle if not in practice, by asking how much lower productivity in the economy might be if there had been no electronics revolution to bring down computing costs: what would the level of productivity be if companies did not have access to IT at all? Determining the marginal contribution of using IT requires asking a different question: Given the existing stock of IT capital, what increment to productivity is gained by adding another dollar of invest- ment in additional IT? It is possible in principle that using IT has contrib

OCR for page 24
IMPACT OF IMFORMATION TECHNOLOGY AT THE MACROECONOMIC LEVEL 29 uted greatly to productivity on average, but that overinvestment at the mar- gin has led to a low return on the last dollar. Often without explicitly stating so, some discussions of the productivity paradox focus on the aver- age contribution of using IT, whereas others actually address the question of overinvestment at the margin. THIS STUDY-APPROACH, SCOPE, AND TERMS Motivated by questions concerning how the use of information technol- ogy contributes to productivity in U.S. service industries and whether the potential benefits of information technology are being full realized, the committee initially sought to assess the impact of IT on productivity in the service sector. However, the committee soon determined that this was the wrong issue. Although it attempted to investigate key factors in each of the five approaches to explaining the productivity paradox, the committee found no way to determine with precision the relative significance of the various alternatives as they pertain to use of IT. There was general agreement that existing methods for measuring productivity at the aggregate level do not entirely capture the changes taking place in the U.S. economy as a result of using IT and that focusing on the apparent productivity paradox gives an overly narrow picture of what is happening in services. Indeed some com- mittee members argued that the whole concept of productivity, as it is currently measured, may be outmoded and that other measures of perfor- mance provide a better indication of the contribution of IT. There was a consensus that the use of IT has changed the way that businesses operate, and that the locus of innovation and of the production of value has shifted away from traditional manufacturing activities and toward service indus- tries and service activities. Based on its assessment of current knowledge, the committee concluded that productivity is an important but not a sufficient measure of benefit. As measured, productivity does not capture other important elements of perfor- mance such as the quality, flexibility, convenience, variety, responsiveness, reliability, and new opportunities that the use of IT in services can permit. In addition, data on assessments of service-sector productivity often do not reflect the substantial benefits customers or suppliers have received from service-sector investments in IT. If captured at all, these benefits have shown up in measures of the customer- or supplier-industries' productivity. Finally, the use of IT in services has often transformed the structures of both manufacturing and service industries, created cross-competition among different industry segments, allowed development of global supplier and customer networks, created whole new industries, and changed the very way in which commerce and management are performed.

OCR for page 24
30 INFORMATION TECHNOLOGY IN THE SERVICE SOCIETY In addition, IT is a relatively new technology, and hence long delays can be expected before its effective deployment. Many of the service in- dustries into which it has been deployed have not traditionally managed major technological changes. Even in industries inured to such changes, introduction of a new technology typically takes significant time and is accompanied by mistakes in implementation and evaluation and by lags in profitability increases. In the course of its deliberations, the committee recognized that under- standing and improving the impact of IT on the overall performance of services activities, not simply determining the productivity of IT in ser- vices, is the real challenge. Moreover, the committee came to believe that the primary impediments to the full exploitation of IT to improve service- sector business practices and performance were not as much issues of inad- equate technology as they were issues of management. For this reason, the report focuses primarily on the roles of management in the exploitation of IT, although the remarkable dynamism that characterizes the evolution of information technology itself has not gone unnoticed. To better understand these complex issues, the committee chose to in- vestigate the full range of impacts of IT use on performance in the service industries, including measured productivity as only one element in overall performance. To do so, the committee analyzed the impacts of IT use at four levels: the macroeconomic level, the industry level, the enterprise level, and the activity level within the enterprise. The purpose of these inquiries was not so much to critique past studies or results as to offer practical insights to executives and policymakers concerning the utility of available data and the application of practices that might improve IT's im- pact on the performance of service activities in the future. CURRENT DATA AND MEASURES OF PRODUCTIVITY Based on existing data, it is impossible to determine exactly what pay- off the U.S. economy has achieved from use of IT. To understand this point one must look at the ways in which productivity is currently measured, particularly in the service industries. There are both strengths and weak- nesses in the current statistical base. The recently updated productivity series developed by the Bureau of Labor Statistics for selected industries shows strong growth in many service industries. The aggregate data from the Bureau of Economic Analysis also offer useful insights about the pro- ductivity of the service sector. But both series may ignore or understate some key factors that contribute to productivity in several service indus- tries. Alternative measures of performance may be useful to supplement measures of productivity or to substitute altogether for them as indica- tors of economic performance.

OCR for page 24
IMPACT OF IMFORMATION TECHNOLOGYAT THE MACROECONOMIC LEVEL 31 National Income and Product Accounts Prepared by the Bureau of Economic Analysis The Bureau of Economic Analysis (BEA) of the U.S. Department of Commerce has responsibility for preparing the National Income and Prod- uct Accounts (NIPA). The data in these accounts are comprehensive, cov- ering economic activity throughout the U.S. economy. When the frame- work of the accounts was developed after World War II, the concern of public policy was to avoid a return to the disastrous period of the Great Depression. The data that are prepared are used when monetary and fiscal policy are adjusted in order to influence the level of demand in the overall economy. The need to be comprehensive and the focus on policies affect- ing demand influence in many ways the nature of the data collected. Thus, at times, the data are limited in their usefulness for studying aggregate supply productivity and long-term growth. Traditionally, the gross national product (GNP) has been the principal aggregate measure of all of the goods and services produced in the economy. How this figure is constructed is a factor in its limitations and general validity for assessing productivity. The basic data that go into GNP de- scribe the sales of goods and services for consumption, the sales of goods (including construction) for investment, and government purchases of goods and services (including the services government itself produces). Included in GNP are final sales, that is, sales to consumers or to government or the sales of durable goods to businesses. Sales of components or raw materials that are then processed further are considered sales of "intermediate" goods and are not included directly in GNP. Additions to inventory are counted as part of investment and are added to final sales in computing GNP (or sub- tracted in the case of any depletion of inventories). The GNP includes the income that is earned by U.S.-owned assets lo- cated overseas minus the earnings of foreign-owned assets located in the United States. When this net income is subtracted out, the resulting aggre- gate is called the gross domestic product (GDP). In its recent revisions of the NIPA, BEA has highlighted GDP rather than GNP. GDP gives the best aggregate indicator of economic activity within U.S. borders. The BEA collects information on final sales and inventories to estimate GDP in current dollars, that is, with everything priced at its actual sales price. It then uses price information collected by the Bureau of Labor Statistics and, for some components, by BEA itself to construct an aggre- gate price deflator for GDP. This deflator is used to transform GDP in current dollars into GDP in constant dollars, also known as real GDP. Real GDP reflects the total of goods and services produced in the economy priced in base-year prices, currently 1987 prices. Putting everything in common prices is designed to remove the effect of inflation on GDP and leave only

OCR for page 24
32 INFORMATION TECHNOLOGY IN THE SERVICE SOCIETY an index of the increase in the physical quantities of goods and services produced in the economy, often called the gross domestic product originat- ing in the U.S. economy (GPO). Once total GDP has been calculated, the activities of the government sector,9 owner-occupied housing, and the nonprofit sector are often sub- tracted out, and the resulting measure of output is then described as the GDP originating in the private business sector of the economy (i.e., the GPO of the business sector). Business GDP can be subdivided further into the GDP produced in each sector or industry in the economy (GPO by industry), based largely on the wage and capital income generated by the industry. Obtained in this way, the output of an industry reflects the value added produced within the in- dustry, that is, the real value of the sales of the industry (adjusted for any change in inventory) minus the value of materials and services purchased from other industries.~ Estimating the contribution to total GDP made by each constituent industry often involves some educated guesswork, because many companies have plants in several different industries, and so their income has to be divided up to account for their various activities. Another difficulty encountered in determining the real (constant-dollar) output of each industry comes in developing suitable industry price defla- tors. The dollar value of income generated in each industry is divided by an industry price deflator in order to compute real output of the industry in 1987 dollars. This approach is analogous to the procedure used to compute aggregate real GDP in 1987 dollars, but some of the industry price deflators are more difficult to construct than the aggregate price deflator. The prices associated with both the outputs and the inputs of some industries are in- trinsically hard to measure, or data are simply not collected that allow accurate measurement. ~ ~ Although it develops data on business GDP and GDP by industry, the BEA does not publish data on productivity. Instead, the responsibility for measuring productivity in the U.S. economy rests with the Bureau of Labor Statistics. Productivity from the NIPA Data: What do the Numbers Show? The Bureau of Labor Statistics (BLS) makes some adjustments to the NIPA data on output associated with business GDP and GDP by industry and combines those data with data on labor input by industry to give average labor productivity (Box 1.2), computed as real GDP per employee-hour. (Box 1.2 discusses different concepts of productivity, such as labor, capital, and multi- factor productivity.) BLS publishes data on productivity only for the business sector as a whole, for nonfarm business, and for manufacturing. However, BLS will release on request the productivity data for individual industries, and the resulting productivity information is often widely quoted.

OCR for page 24
IMPACT OF IMFORMATION TECHNOLOGYAT THE MACROECONOMIC LEVEL 33 Table 1.1 shows the rates of growth of average labor productivity for the business sector as a whole and for the major industry sectors as pre- pared by BLS using the BEA data. The figures for the productivity of the business sector show the pattern that has become familiar. Productivity grew at a rate of a little less than 3 percent a year in the business sector as a whole from 1948 to 1973 and then fell to 0.63 percent a year from 1973 to 1979 before making a partial recovery from 1979 to 1989. The business sector is divided into goods- and service-producing indus- tries, and shows solid growth in both parts of the economy over the period ~. , TABLE 1.1 Average Annual Growth in Gross Domestic Product per Labor-Hour for Major Sectors of the U.S. Economy, 1948 to 1989 Average Annual Rate of Growth (% per year) Difference in Rate of Growth, 1948-73 and 1979-89 1973-891973-89 Sector 1948-73 Business 2.88 0.63 1.35 1.08 -1.80 Goods producing 3.21 0.71 2.31 1.71 -1.50 Farming 4.64 0.11 3.22 2.04 -2.60 Mining 4.02 -5.S6 2.13 -0.82 -4.84 Construction 0.58 -2.02 -0.71 - 1.20 - 1.78 Manufacturing 2.87 1.80 3.33 2.75 -0.12 Durable goods excluding nonelectrical machinery 2.56 1.55 2.35 2.05 -0.51 Nonelectrical machinerya 2.03 1.06 9.10 6.01 3.98 Nondurable goods 3.40 2.37 2.37 2.37 -1.03 Service producing 2.49 0.58 0.84 0.74 -1.75 Transportation 2.31 0.15 0.95 0.65 -1.66 Communication 5.22 4.27 4.84 4.63 -0.59 Utilities 5.87 2.66 2.35 2.46 -3.41 Trade 2.74 -0.35 2.10 1.18 -1.56 Wholesale 3.14 -1.21 2.65 1.18 -1.96 Retail 2.40 0.14 1.72 1.13 -1.27 FIREb 1.44 0.36 0.05 0.17 -1.27 Services 2.17 0.84 0.01 0.32 -1.85 Government enterprise -0.15 0.62 0.04 0.26 0.41 General government 0.21 0.03 0.53 0.34 0.13 aNonelectrical machinery hours from Bureau of Labor Statistics. bFinance, insurance, and real estate. SOURCE: U.S. Department of Commerce, Bureau of Economic Analysis. 1991. Survey of Current Business, April, Tables 6.2 and 6.11.

OCR for page 24
34 INFORMATION TECHNOLOGY IN THE SERVICE SOCIETY BOX I.2 Measuring Capital and Multifactor Productivity Pleasures of productivity relate physical output to physical input. As such, they encompass ~ family of measures ranging from measures of single-fac$or-input productivity, such as output per unit of labor input or output per unit of capital input, to measures of multi~ctor productivity, such as output per unit of labor and capital combined. }n the form of indexes these measures present the change in output associated with the corresponding change in input. in other words, the indexes show the change in resources (labor, capital, materials, or all combined) used to produce the output of the activity being measured. They do not measure how much of the change in output came from any of the individual input factors, but rather how the many Actors affecting the production process resulted in Flanges in resource use. The most extensively developed and widely used measure of produc- tivity is the one relating output to labor input-labor productivity. It is a ~tool relevant for analyzing labor costs, real income, and employment and, I as a practical matter, is most easily measured. However, an increase or decrease in output per emplayee-hour does not imply that labor is solely or necessarily even primarily responsible for improved or worsened pro duct~vity. Movements in output per hour reflect technological innova tions, changes in capital input, scale of production, education, manage ment, and many other factors as well as the skits and efforts of the work force. 1 Average labor productivity is the ratio of output to labor input. Labor input is chosen as the most important factor of produ~ion the cc~mpensat'an of labor is about two-thirds of fatal Flue added. But labor productivity can change for reasons that have nothing to do with the work force. Various factors can cause labor productivity to change oYer time. For example' there have been substantial increases in the amour,' Of capital used per worker. Techr`oiogy itself arid the organization of work may changes These considerations do riot invalidate Me concept of labor productivity; they simply mean that it is important to ~r`terpret it correctly Also, it nosy be necessary to choose among a fancily of meat sures rel~ir~g out'cut to labor, capital, and other inputs, some of which adjust {or quality and some of which do not. Many economists prefer a concept called multihctor productivity or total-factor productiY'ty, When the value-added measure Of output is used, grow - h in multihctor productiY'~g, is computed as Me growth rate of output minus a weighted average of the growth rates of labor and capital inputs The weights are then shares of labor and capital costs in total cost. Increases in multifactor productivity reflect increases In the efficiency with which labor and capital are used to produce Output.

OCR for page 24
IMPACT OF IMFORMATION TECHNOLOGY AT THE MACROECONOMIC LEVEL 41 are published for 39 service industries, covering about 42 percent of all workers in the service industries in the private business sector. The BLS also publishes measures of productivity for the federal gov- ernment with separate detail for the common functions provided by federal agencies, such as record keeping, library services, building and grounds maintenance, loans and grants, and so on. In addition, work is currently being conducted on the development of indexes of productivity for hospitals and additional wholesale trade activities. The measures of output used in these BLS series are independently developed from data on sales, prices, and physical units whenever possible, combined with labor input weights. In some instances the data on labor input weights are not available, and approximations such as gross margins have to be used. The data on labor input cover the hours of all persons employed in a particular industry, including the self-employed, and are based on the BLS surveys of establishment employment and hours. The "hours of employees" are based on hours-paid data from the surveys of establishment employment and hours. The hours of the self-employed are derived from the household survey of employment and hours conducted for the BLS by the Bureau of the Census. What Do the BLS Data on Selected Industries Show? Table 1.3 shows the compound annual percentage change in output per employee-hour for selected service industries over the period from 1973 to 1990, the longest period for which continuous measures from BLS are available. Contrary to the general belief that service industries are characterized by uniformly low growth in productivity, a wide range in the rates of growth is evident. Over the most recent decade, 1980 to 1990, the rates range from a decline of 4 percent per year for gas utilities to an increase of 9 percent per year for railroads. The increase for railroads actually exceeds that for almost all manufacturing industries, in which significant rates of growth in productivity have occurred. Even over the longer period from 1973 to 1990, when the range of rates of growth was narrower, there was still considerable variation across industries. Gaps in BLS Data and Measures of Productivity Although BLS publishes at least one measure of productivity for each major service activity group transportation, communications, and so on- there are major areas within each of these groups for which no measure of productivity has been developed and published. One reason for these gaps is that it is difficult to define and develop meaningful indicators of output and/or input in many service industries. In

OCR for page 24
42 INFORMATION TECHNOLOGY IN THE SERVICE SOCIETY TABLE 1.3 Growth in Productivity in Selected U.S. Service Industries, 1973 to 1990 Period of Time and Output per Labor-Hour (average annual % change)a Service Industry SICb Codes 1973-90 1973-801980-90 Transportation Railroad 4011 5.8 1.59.0 Bus carriers, class 1 411, 13, 14 0.7c o.s_0.7d Trucking, except local 4213 3.2 2.33.9 Air transportation 4512, 13, 22 2.5 3.22.1 Petroleum pipelines 4612, 13 0.6 -0.71.6 Utilities Telephone communications 481 5.5 6.84.5 Electric utilities 491, 3 1.4 0.91.7 Gas utilities 492, 3 -2.5 -0.3-4.0 Trade Scrap and waste materials 5093 n/ae n/ae1.4d Hardware stores 5251 1.9 1.81.9 Department stores 5311 2.6 2.62.6 Variety stores 5331 -0.4 -1.60.5 Grocery stores 5411 -1.0 -0.1-1.7 Retail bakeries 546 -0.7 -1.1-0.4 New- and used-car dealers 5511 1.4 0.52.1 Auto and home supply stores 5531 2.2 2.42.0 Gasoline service stations 5541 2.9 2.92.9 Men's and boys' clothing stores 5611 1.7 0.52.6 Women's clothing stores 5621 3.2 3.33.2 Family clothing stores 5651 1.8 1.81.8 Shoe stores 5661 1.5 1.41.6 Home furniture stores 571 1.0 0.31.5 Household appliance stores 5722 2.5c 2.72.6d Radio, TV, and computer stores 573 4 8c 4 35.7d Eating and drinking places 581 -0.6 -0.5-0.6 Drug stores 5912 0.5 1.4-0.1 Liquor stores 5921 -0.4 0.2-0.8 Business and personal services Commercial banks 602 1.2 -0.52.7 Hotels and motels 7011 -0.4 0.4-1.0 Laundry and cleaning services 721 -0.8 -1.2-0.5 Beauty and barber shops 7231, 41 0.8 0.01.4 Automotive repair shops 753 -0.2 -1.20.4 aBased on compounding formula. bStandard Industrial Classification system code. C1973 to 1989. dl 980 to 1989. en/a, not available. SOURCE: Bureau of Labor Statistics.

OCR for page 24
IMPACT OF IMFORMATION TECHNOLOGY AT THE MACROECONOMIC LEVEL 43 most cases adequate measures have to be derived from data obtained in surveys conducted for purposes other than measuring productivity. For measuring productivity, the output indicator should be final to the activity being examined, quantifiable, and consistent in coverage with, but indepen- dent of, the input indicators; should reflect differences in the quality of the activity being evaluated; and should represent the major part of the activi- ties of the sector. For many service activities, the available data do not meet these criteria. The major gap within communications is radio and television broadcasting; in electricity, gas, and sanitary services, it is sani- tary services. In trade, the gap in coverage is in the area of wholesale trade; in finance, insurance, and real estate the gap is in the real estate component. In business and personal services, no measures are available for health, education, and legal services. Finally, in the area of government, the major gap is in state and local government. Conclusions on Productivity Measures Developed from BLS Data The BLS has made considerable progress in developing indexes of pro- ductivity for selected service industries. The measures of productivity for some of these industries are very accurate, clearly reflecting changes in output per unit of labor input. Because the concepts of output are relatively straight- forward and appropriate data are readily available, the measures for transpor- tation, communications, and public utilities fall into this category. The mea- sures for some of the trade industries present some difficulties primarily because of the heterogeneity of the categories of merchandise line sales and the limited availability of adequate price deflators. However, these indexes of productiv- ity are useful indicators of trends over time and are continually being im- proved as the number of detailed price indexes increases.~7 Many difficulties remain in clarifying some of the basic conceptual problems of defining the output of certain service industries, and many inadequacies are present in the data available. In particular, the BLS does not have the resources to follow the rapidly changing nature of the output in many services industries, changes that have been enabled by the use of IT in many cases. BLS's measurement of productivity is good if output over time can be captured adequately by a simple measure such as passenger- miles. It does not do so well when other characteristics are important, and in such situations it suffers from the same kinds of problems that the BEA data suffer from. For example, in the case of airlines, no account is taken of changes in the frequency of service, ease of booking or changing reserva- tions, flying time, number of stops, on-time record, and so on. For many service industries, these aspects of service quality are crucial.

OCR for page 24
44 INFORMATION TECHNOLOGY IN THE SERVICE SOCIETY ALTERNATIVE MEASURES OF PRODUCTIVITY AND PERFORMANCE The slowdown in the growth of measured productivity in services has been the subject of great concern among policymakers, but productivity is only one measure of economic performance. For the business manager, other measures of performance may provide a better guide to long-term profitability and success. Productivity and profitability are linked in that, other things being equal, a company that is able to raise its productivity relative to that of its competitors will be able to lower its costs and raise its margins or reduce prices to increase market share. But other things are not always equal. For example, a company that provides a unique new product or service may be able to raise its profitabil- ity by more than a company that simply concentrates on raising productivity in its existing lines of business. Or an industry may find that its market conditions change. The deregulation of many service industries -has forced them to adapt and, in many cases, to raise productivity. But the increase in competition within these industries has lowered prices, and margins and profits are hard to come by. Airlines, railroads, and trucking are all in this position. For workers, employment may be more important than productivity, especially in the short run. In fact some people see a trade-off between jobs and increases in productivity. The U.S. economy has been very successful over the years in increasing the number of jobs it has generated, and the largest number of these have been in the service sector. The fear in some quarters is that as some labor-intensive companies in the service sector seek to boost their productivity via restructuring, the creation of full-time jobs may slow.~9 However, downsizing is not a long-term recipe for sustained growth in productivity. Once companies have eliminated excess staff, they will need to look beyond downsizing and restructuring and embark on a path of judi- cious expansion that includes increases in the rate of hiring and in capacity. Recent evidence suggests that this may be starting to happen. Over the period from October 1992 to May 1993, job growth averaged 165,000 jobs per month, well in excess of the paltry gains of 18,000 per month that occurred in the first 17 months of the recovery. For example, during 1992 business spending on capital equipment has risen at an annual rate of 16 percent, with 60 percent of the increase concentrated in IT hardware. For some purposes international competitiveness is as important or more important than productivity. Again, the two are linked, but external circum- stances may change so that the two indicators of performance move differ- ently from each other. For example, the U.S. auto industry and the U.S. textile industry have both succeeded reasonably well in raising productivity

OCR for page 24
IMPACT OF IMFORMATION TECHNOLOGY AT THE MACROECONOMIC LEVEL 45 in recent years. But both industries have come under intense competition from foreign companies that have done as well or better in increasing pro- ductivity and product quality or have been able to make use of less costly labor. A recent study indicates that compared to similar industries of other countries, U.S. service industries have generally higher absolute levels of productivity (Box 1.3), although U.S. rates of growth in productivity have lagged those of many other nations.20 Furthermore, the United States main- tains a healthy $52 billion service trade surplus with foreign nations. OBSERVATIONS AND CONCLUSIONS Many Factors Influence Productivity: IT Affects Many Aspects of Performance As noted at the beginning of this chapter, many factors besides the application of information technology determine growth in the productivity of the U.S. service sector. This is not surprising, since output itself is a function of multiple inputs and technology, the latter affecting how inputs are combined in production. Productivity is also influenced, for example, by other kinds of capital investment, such as buildings for retailers or air- planes for airline companies. The fact that over the past 15 to 20 years slow growth in productivity has been accompanied by rapid investment in IT has suggested to some people that the use of IT has actually caused the slowdown in productivity growth. But the committee determined that it is a mistake to conclude that IT is the culprit. It surely is in some cases, but in others it has probably made a large positive contribution. The key point is that the currently available macroeconomic data cannot precisely measure how investments in IT alone are influencing productivity in the services. As documented and discussed above in this chapter, the committee's find- ings have led it to make the following observations about current macroeconomic data on service-sector productivity and how investments in and the use of IT have influenced these data: . . The BEA macroeconomic data do not deal adequately with changes in the quality of the services provided. And IT is often used to improve the quality of services. The development and use of better measures of output might show a very different picture of growth in productivity. The BEA macroeconomic data do not adequately allow for the ef- fects of new services, and IT is often used to provide new services. Again, the development and use of better measures of output that account for new services could show a radically different picture of growth in productivity. Existing data on the capital input to production do not take account of the knowledge capital (see Box 1.2) that is considered by many to be the

OCR for page 24
46 BOX 1.3 A Recent Study of Service-Sector Productivity in the United States and Other Countries l 1 INFORMATION TECHNOLOGY IN THE SERVICE SOCIETY The Global Institute of McKinsey & Co., assisted by economists Mar- tin Baily, Francis Bator, and Robert Solow, completed in 1992 a com,oari- son of the productivity of the airline, banking, general merchandise retail- ing, telecommunications, and restaurant industries in the United States relative to the comparable industries in European countries and Japan. For three of the industries airlines, banking, and telecommunica- tions the study used quantitative output measures of the type used by the Bureau of Labor Statistics in its analysis of specific industries. For example, in airlines, different output measures were combined, such as passenger-miles, number of aircraft maintained, and number of passengers served Adjustments were made to allow for such factors as different route lengths and fleet structures. For the other two industries value added was the measure of output, with the international comparisons made using what are called purchasing power parity exchange rates. Except in the restaurant industry, where productivity was compa- rable in the countries studied, the findings indicated that the U.S. ~E,dus- tries had a lead in productivity, usually in the range of 20 to 40 percent. Several explanations for this productivity lead were examined. For ex- ample, it was found that U.S. airlines, larger average size gave them some advantage over European airlines. Differences in the mix of output and in capital intensity were also significant in some cases. The most important difference overatI, however, was found to be differences in how the work force is Organized and managed. There is greater competitive pressure and less restrictive regulation In the U. S. industries than in those of over countries studied. The ~cK'nsey study Is based an an analysis in a single year (1989) of productivity levels and therefore does not directly contradict the Bureau of Economic Analysis macroeconomic data that show a slow rate of growth of productivity in the service sector in the United States. But the finding that tJ.S. services have a substantial lead over other countries could reinforce the concerns about the BEA data that are presented in this chapter. In general the McKinsey study concluded that differences in technol- ogy were not the most important factor in the productivity d't~rences that they found, since the technology is available to enterprises in all of the countries studied if they wish to use it, For telecommunications and banking, Me study did conclude that a failure to adopt the most advanced information technology had hurt productivity in other countries. The teleeammuntc~ions industry in Germany, for example, had not intro- duced electronic switching and could not offer business customers high- speed data trar~sm~ss'an. And in the banking industry, both Germany and

OCR for page 24
IMPACT OF IMFORMATION TECHNOLOGY AT THE MACROECONOMIC LEVEL 47 the united Kingdom suffered a productivity disadvantage because fewer automatic teller machines had been deployed. U.S. banks had invested more heavily in terminals, allowing many activities to be centralized that were still carried colt less productively In branches overseas, particularly so In the United Kingdom. most important part of the capital base of the service industries. The use of any improved measure of capital input could also drastically change the estimates of growth in productivity. In some service industries, the BEA data exclude, by definition, any changes in measured productivity because real output is measured by the number of employees working in the industry. Banking and financial services are the most important examples. For the business decision maker, productivity may be an important con- sideration but will often not be the most important one. Productivity is important for business because a company that has weak productivity in its core business activities, relative to other companies, is tying one hand be- hind its back in its competitive struggle. But productivity is rarely suffi- cient for business success and should not be the sole or even the primary basis for most companies' decisions about investment in IT. Providing value to the consumer is often more important than increasing productivity, and such circumstances will require that measures of performance be used to track growth in the economy and to evaluate the part played by IT. Making better business and policy decisions about investing in IT requires having more information about what IT is being used for than can be pro- vided by the standard data on productivity. In Chapters 2 through 4, the committee discusses how IT has been used to improve performance at the industry, enterprise, and activity levels of the U.S. service sector. ORGANIZATION AND SCOPE OF THIS REPORT In refining its approach in response to its initial findings, the committee chose to investigate the full range of impacts of the use of IT on perfor- mance in the service sector. Thus the committee examined the nature and measurement of performance in services at progressively less aggregated levels of analysis: the industry level (Chapter 2), the enterprise level (Chapter 3), and the activity level (Chapter 4~. These chapters are supplemented by appendixes that provide more methodological detail and supporting data: Appendix A describes selected research on economic and strategic impacts

OCR for page 24
48 INFORMATION TECHNOLOGY IN THE SERVICE SOCIETY of IT; Appendix B describes methods used by the Bureau of Economic Analysis to derive measures of output; Appendix C describes methods used by the Bureau of Labor Statistics for deriving measures of productivity; Appendix D describes how the committee conducted its study, including the interview process used by the committee to inform its deliberations and a summary of its results. Chapters 2 through 4 culminate in Chapter 5, which discusses implica- tions for managers in organizations wishing to improve their management of information technology, and Chapter 6, which presents issues and recom- mendations for public policy. To put this report in perspective, some observations on the committee's operation and scope of concern are appropriate. First, this report was shaped by interactions within a multidisciplinary committee that included business executives, economists, behavioral scientists, management theorists, and tech- nologists. Second, the committee considered the context of international competitiveness in conducting its analysis, but detailed investigation of in- ternational conditions was beyond its scope. Third, although smaller com- panies were represented in the data that support industry- and national-level analyses, the committee's resources did not permit a systematic examina- tion of the distinguishing characteristics of smaller enterprises. In addition to using the standard macroeconomic data collected and developed by government agencies, the committee drew on observations from managers and executives in industry (a group including some of the committee's own members, as well as numerous others). These observa- tions were obtained through interviews that were used to develop and check insights, not to generate quantitative data. Appendixes provide method- ological and supporting details about the committee's sources of informa- tion and its approach. Appendix E lists the interviewed executives, whose observations helped the committee to understand the processes by which IT projects are planned, implemented, and evaluated. At the macroeconomic level, the committee was concerned about the constraining effects of looking at services from the traditional perspectives of goods-production (Chapter 1~. At present, most of the terminology, methodology, and data for analyzing productivity (and performance) derive from earlier studies in the goods-producing industries, but to the committee, most of these seemed inadequate for understanding trends in the service sector. For example, whereas productivity in goods-producing activities is measured in terms that refer to relatively concrete units of output, dollar sales, or profits, performance in services may relate to more subjective quantities such as timing, quality, comfort, or convenience. Measurement difficulties are a theme that runs through the report.

OCR for page 24
IMPACT OF IMFORMATION TECHNOLOGY AT THE MACROECONOMIC LEVEL 49 NOTES AND REFERENCES 1This figure is for checks cleared, deposits processed, and cash distributed by human tellers. Cash distributed by automated teller machines is not included. Data reported in 1992 by The Global Institute of McKinsey and Company (see Box 1.3). 2See, for example, Baily, M., and R. Gordon, 1986, "The Productivity Slowdown: Mea- surement Issues and the Explosion of Computer Power," pp. 347-431 in Brookings Papers on Economic Activity, W. Brainard and G. Perry (eds.), Brookings Institution, Washington, D.C.; Roach, Stephen S., 1988, "Technology and the Services Sector: America's Hidden Competitive Challenge," in Technology in Services, Bruce R. Guile and James Brian Quinn (eds.), National Academy Press, Washington, D.C.; Roach, Stephen S., 1991, "Services Under Siege: The Restructuring Imperative," Harvard Business Review, September-October, pp. 82-91; and Strassmann, P., 1990, The Business Value of Computers, Information Economics Press, New Canaan, Conn. 3A paper by Paul Attewell that was brought to the attention of the committee while this report was in press addresses some of these issues. See Attewell, Paul. 1994. "Information Technology and the Productivity Paradox," in Understanding the Productivity Paradox: Or- ganizational Linkages, Douglas H. Harris (ed.), National Academy Press, Washington, D.C., in preparation. 40ne variant of this idea suggests that IT has transformed the way business is conducted and is essential to competitive success, but the changes it has facilitated do not raise the overall output of the economy. For example, individual companies can use IT to target their marketing campaigns more effectively, but overall industry sales are not affected and so nei- ther is productivity measured at the industry level. 5For a discussion of the slowdown in the United States and other countries, see Baumol, William J., Sue Anne Batey Blackman, and Edward N. Wolff, 1989, Productivity and Ameri- can Leadership: The Long View, MIT Press, Cambridge, Mass.; and Baily, Martin N., and Alok K. Chakrabarti, 1988, Innovation and the Productivity Crisis, Brookings Institution, Washington, D.C. However, note that U.S. productivity growth rates that lag behind those of its international competitors are a distinct concern for U.S. policymakers, even though U.S. base productivity levels may still be higher than those of its competitors. 6See, for example, Kendrick, J.W., 1987, "Service Sector Productivity," Business Eco- nomics, April, pp. 18-24. 7David, Paul A. 1990. "The Dynamo and the Computer: An Historical Perspective on the Modern Productivity Paradox," American Economic Review, May, pp. 355-361. 8Brynjolfsson, Erik, and Lorin Hitt. 1993. "Is Information Systems Spending Produc- tive? New Evidence and New Results," MIT Sloan School of Management, Working Paper 3571-93, September 24. (To appear in Proceedings of the 14th International Conference on Information Systems.) 9The part of the government sector that is subtracted is "general government." This includes the salaries of government workers in the various federal, state, and local agencies. The output of "government enterprises," such as the Postal Service and the Tennessee Valley Authority, are included in the business sector of the economy. 1OThe capital goods purchased by an industry are not subtracted from the gross output of the industry. Value added in an industry reflects the payments to both labor and capital inputs to the industry. 1lOne difficulty is that there is not a direct match between products and industries. Plants in several industries cast their own steel products, for example. An industry deflator should be based on the prices of the products that are produced by the plants that are classified within that industry. 12Also included in Table 1.1 is productivity for the government sector, although any productivity that shows up in the data is purely the result of shifts in the composition of government activities. By assumption there is no productivity growth within government activities.

OCR for page 24
so INFORMATION TECHNOLOGY IN THE SERVICE SOCIETY 13The narrowly defined services sector consists of hotels and other lodging places, per sonal services, business services, auto repair and similar services, motion pictures, amusement and recreation services, health care, legal services, education, and social services. By contrast, the term service sector is used throughout this report to denote the entire range of service industries, as noted on the opening page of this chapter. 14Though important when productivity is used as an indicator of living standards, adjust- ments for changes in product quality can be misleading if the issue at hand is product cost analysis. For example, if the number of teaching hours per student does not decline, college costs must rise as quickly as faculty salaries, whether or not the quality of teaching changes. 15Not everyone agrees that claims processing has speeded up. There have been com- plaints that some insurance companies have delayed payments. 16For example, the GDP originating in the commercial banking industry per full-time employee in 1980 ($37,600 in 1982 dollars) was virtually identical to that in 1988 ($37,900 in 1982 dollars). This small change was due entirely to a different structure in the commercial banking industry. 17The use of gross margin weighting as an approximation of labor input weighting is appropriate as long as labor costs represent a significant portion of gross margins and the elements of margins do not change over time. 18The difficulties that remain involve clarifying some basic conceptual problems of de- fining output for some service industries and filling gaps in data available. With regard to data, more and improved data are needed on prices, value of output, capital, and material input. The BLS and other government statistical agencies have undertaken initiatives to expand the coverage in the service sector, and the results of these efforts should lead to better and more comprehensive industry measures. Collection of more measures of service industry prices, expanded data on hours worked that provide more disaggregation, and continued work on public-sector productivity by BLS are but a few examples. However, some areas e.g., educa- tion, entertainment, legal and social services, and many segments of medical services present such severe conceptual and data problems that appropriate solutions may be a long time in coming. Additional resources may have only a limited effect until the conceptual problems can be dealt with. In expanding the number of industries for which measures are prepared, the BLS has had to take these problems, among others, into consideration, and this has limited the number of service industries for which measures have been developed. Nevertheless, adequate mea- sures of productivity have been developed for many service industries. The problems of measuring productivity are different for each of the various service industries, and the ap proaches followed vary. 19The recovery of 1993 is widely believed to be relatively weak by historical standards. For example, unpublished data from the Bureau of Labor Statistics, U.S. Department of Labor, indicate that jobs in the ninth quarter of this recovery rose by 0.3 percent as compared to their nadir in the first quarter of 1991. By comparison, the corresponding figure for the eight previous recessions is 6.8 percent. In this recovery, manufacturing firms seem to be relying more on overtime than on new hiring to meet expanding demand for goods, while service firms seem to be relying more on hiring part-time workers with few benefits (for one example, see Swoboda, Frank, 1992, "In Omaha, the Underside of a Jobs Promise," Washington Post, Octo- ber 25, p. H-1). The result in manufacturing has been fewer jobs with longer work weeks, and in services more jobs with shorter work weeks (see Uchitelle, Louis, 1993, "Fewer Jobs Filled As Factories Rely on Overtime Pay," New York Times, May 16, p. 1). Even strong companies are cutting labor costs to maintain or increase profit margins (see Uchitelle, Louis, 1993, "Strong Companies Are Joining Trend to Eliminate Jobs," New York Times, July 26). 20Recent data suggest that the slowdown in growth of U.S. productivity has been accom- panied by a slowdown in that of the other leading industrial countries, and that the percentage

OCR for page 24
IMPACT OF IMFORMATION TECHNOLOGYAT THE MACROECONOMIC LEVEL 51 decline in Japan has been almost exactly as great as that in the United States. Further, all industrial countries have moved closer to the United States in levels of productivity and GDP per capita, although they have not caught up with this country in real terms (i.e., in terms of purchasing power parity exchange rates). With the race narrowing, it is a tautology that the laggards must be running faster than the leader, thus explaining how the United States can be ahead of everyone else in productivity level, although its productivity growth rate lags behind those of many of the other countries. ~ . _ _ U.S. Department of Commerce, Bureau of Economic Analysis. 1992. Survey of Cur rent Business, June, Table 3, "Selected Services Transactions," p. 96. BIBLIOGRAPHY FOR CHAPTER 1 Brand, Horst, and John Duke. 1982. "Productivity in Commercial Banking: Computers Spur the Advance," Monthly Labor Review, Vol. 105 (December), pp. 19-27. Brand, Horst, and Zaul Ahmed. 1986. "Productivity in Beauty and Barber Shops," Monthly Labor Review, Vol. 109. Carnes, Richard B. 1978. "Laundry and Cleaning Services Pressed to Post Productivity Gains." Monthly Labor Review, Vol. 101 (February), pp. 38-42. Carnes, Richard B., and Horst Brand. 1977. "Productivity and New Technology in Eating and Drinking Places," Monthly Labor Review, Vol. 100 (September), pp. 9-15. Dean, Edwin, and Kent Kunze. 1990. "Productivity Measurement in Service Industries," Output Measurement in the Services Sector, National Bureau of Economic Research, Cambridge, Mass. Duke, John. 1977. "New-Car Dealers Experience Long-Term Gains in Productivity," Monthly Labor Review, Vol. 100 (March), pp. 29-33. Friedman, Brian. 1984. "Productivity in the Apparel and Accessories Stores Industries," Monthly Labor Review, Vol. 107 (October), pp. 37-42. Friedman, Brian, and John L. Carey. 1975. "Productivity in Gasoline Stations, 1958-1973," Monthly Labor Review, Vol. 98 (February), pp. 32-36. Kutscher, Ronald, and Jerome A. Mark. 1983. "The Service-Producing Sector: Some Com- mon Perceptions Reviewed," Monthly Labor Review, Vol. 106 (April), pp. 21-24. Mark, Jerome A. 1982. "Measuring Productivity in the Services," Monthly Labor Review, Vol. 105 (June), pp. 3-8. Mark, Jerome A. 1988. "Measuring Productivity in Service Industries," in Technology in Services: Policies for Growth, Trade, and Employment, Bruce R. Guile and James Brian Quinn (eds.), National Academy Press, Washington, D.C. Mincer, Jacob. 1974. Schooling, Experience and Earnings, Columbia University Press, New York. Roach, Stephen S. 1991. "Services Under Siege: The Restructuring Imperative," Harvard Business Review, September-October, pp. 82-91. Smith, Anthony D. 1972. The Measurement and Interpretation of Service Output Changes, National Economic Development Office, London. Waldorf, William H., Kent Kunze, Lawrence S. Rosenblum, and Michael B. Tannen. 1986. "New Measures of the Contribution of Education and Experience to U.S. Productivity Growth," paper presented at annual meetings of American Economic Association, Dec. 28-30, New Orleans, La.