Information Technology in the Service Society: A Twenty-First Century Lever (1994)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

2 Impacts of Information Technology at the Industry Level As discussed in Chapter 1, it is difficult to generalize about the impacts of information technology (IT) on productivity or performance in the ser- vice sector, because the effects of using IT cannot be isolated from the effects of other factors and because the service sector is highly heteroge- neous. A better picture of IT's impact can therefore be obtained by focus- ing on individual industries. Although firms differ within a given industry, they must generally share a common environment whose factors are rel- evant to their performance for example, an industry's requirements for particular kinds of workers and skills, the quality of its labor-management relations, the presence or absence of regulation, how capital- or labor-inten- sive the industry is, and the kinds of technology applications associated with it all influence the performance of the firms within that industry. Likewise, an industry's environment affects the nature of the benefits that companies can realize from use of IT. Although the dimensions on which performance may be evaluated such as the nature of improvements in product quality, increases in product variety, and product innovations- will therefore vary across industries, industry context is important for un- derstanding both how and how much the use of IT has affected an industry's performance in providing services. In addition, the experience of one in- dustry may provide insights relevant to understanding or anticipating ef- fects in another. To explore how industry context affects the use of IT and performance, the committee focused on six industries: air transport, telecommunications, 52

IMPACTS OF INFORMATION TECHNOLOGY AT THE INDUSTRY LEVEL 53 retail and wholesale trade, health care, banking, and insurance. Together the industries chosen accounted for about 27 percent of the gross domestic product and 35 percent of U.S. employmentin 1989; they are also propor- tionately much higher users of IT than are some other industries (Tables 2.1 and 2.21.i The sectors to which some of these industries belong are rela- tively capital-intensive; for example, the ratio of capital investment to value added in 1991 for transportation and telecommunications exceeded the comparable TABLE 2.1 Six Selected Service Industries Compared to All U.S. Industry and All U.S. Service Industries, 1989 (figures in billions of 1982 dollars) Investment in IT in 1989 (as % GPO in 1989 of GPO) Value of IT Capital Stock in 1989 GNP GDP All industry (manufacturing, construction, mining, and non-goods-producing) Mining Construction Manufacturing All services (non-goods-producing industries) Transportation Communications Utilities Wholesale Retail Financial, insurance, real estate Business, personal, professional services Six selected service industries (air transport, telecommunications, retail and wholesale trade. health care, banking, insurance) Six industries as fraction of "all services" Six industries as fraction of "all industry" Six industries as fraction of GDP and GNP 4117.7 4087.6 3610.5  127.2 179.0 929.0 2375.3 156.3 109.4 136.6 304.7 412.0 604.0 652.3 1119.7 47% 31% 27% 128.4(3.6) 1.2(0.9) 0.7(0.3) 24.3(2.6) 102.3(4.3) 5.5(3~5) 17.6(16.1) 6.3(4.6) 11.6(3.8) 10.6(2.6) 33.7(5.6) 16.9(2.6) 62.5(5.6) 61% 49% 513.3 8.8 3.2 78.0 456.1 20.6 124.7 26.0 54.0 44.8 123.2 62.8 291.1  64% 57% SUMMARY: In 1989, six service industries (air transport, telecommunications, retail and wholesale trade, health care, banking, insurance) accounted for 27% of GNP, 31% of the 1989 GPO of "all industry," and 47% of the 1989 GPO of "all services." However, they accounted for 49% of 1989 IT expenditures by "all industry" and 61% of 1989 IT expenditures by "all services." Similarly, they accounted for 57% of the 1989 IT capital stock value of "all industry" and 64% of the 1989 IT capital stock value of "all services." In addition, these six industries averaged 5.6% in their 1989 IT investment as a percentage of GPO, while for all industries the comparable percentage was 3.6% and for all service industries it was 4.3%.

54 INFORMATION TECHNOLOGY IN THE SERVICE SOCIETY TABLE 2.2 Snapshots of Selected Service Industries, 1981 and 1989 Air Transport Telecommunications Retail Trade 1981 GPO (billions of 1982 dollars) 16.8 75.7 286.4 Revenues (billions of 1982 dollars) 45.2 91.8 453.6 Number of employees (millions of jobs) 0.5 1.1 16.9 Number of hours worked (billions) 0.9 2.4 27.4 Annual investment in IT (billions of 1982 dollars) 1.0 21.0 5.1 IT capital stock (billions of 1982 dollars) 1.8 125.3 13.2 GPO per labor-hour (1982 dollars) 17.70 31.97 10.44 Annual IT/GPO 6.2% 27.8% 1.8% Annual rate of change in GPO per labor-hour -2.7% 5.2% 1.6% 1989 (1989 Value as Percent of 1981 Value) GPO (billions of 1982 dollars) 31.7(189) 98.8(131) 412.0(144) Revenues (billions of 1982 dollars) 66.6(147) 100.6(110) 628.3(138) Number of employees (millions of jobs) 0.7(149) 0.9(82) 21.2(125) Number of hours worked (billions) 1.4(148) 2.0(83) 32.8(120) Annual investment in IT (billions of 1982 dollars) 3.0(288) 13.8(66) 10.6(207) IT capital stock (billions of 1982 dollars) 9.8(543) 113.7(91) 44.8(340) GPO per labor-hour (1982 dollars) 22.64(128) 49.97(156) 12.54(120) Annual IT/GPO 9.4%(153) 14.0%(50) 2.6%(144)  Annual rate of change in GPO per labor-hour -8.0% 4.0% 1.2% NOTE: See Box 2.1 for explanation of terms and sources of data given.

IMPACTS OF INFORMATION TECHNOLOGY AT THE INDUSTRY LEVEL 55 TABLE 2.2 Continued Wholesale Trade Health Care Banking Insurance 1981 GPO (billions of 1982 dollars)212.7138.1 64.2 35.4 Revenues (billions of 1982 dollars)317.6204.8 88.7 66.8 Number of employees (millions of jobs)5.75.9 2.0 1.2 Number of hours worked (billions)11.410.2 3.9 2.4 Annual investment in IT (billions of 1982 dollars)5.01.8 1.9 0.7 IT capital stock (billions of 1982 dollars)13.28.9 5.1 1.7 GPO per labor-hour (1982 dollars)18.6413.49 16.66 14.76 Annual IT/GPO2.4%1.3% 3.0% 2.0% Annual rate of change in GPO per labor-hour4.7%-1.2% 0.0% -9.0% 1989 (1989 Value as Percent of 1981 Value) GPO (billions of 1982 dollars)304.7(143)164.4(119) 71.4(111) 36.7(104) Revenues (billions of 1982 dollars)400.8(126)286.6(140) 99.8(112) 77.5(116) Number of employees (millions of jobs)6.6(116)7.9(134) 2.3(112) 1.4(116) Number of hours worked (billions)13.1(115)13.5(132) 4.3(110) 2.8(116) Annual investment in IT (billions of 1982 dollars)11.6(230)3.6(204) 13.8(715) 6.2(856) IT capital stock (billions of 1982 dollars)54.0(409)15.7(177) 35.1(690) 17.9(1080) GPO per labor-hour (1982 dollars)23.30(125)12.14(90) 16.80(101) 13.18(89) Annual IT/GPO3.8%(160)2.2%(171) 19.3%(643) 16.8%(826)  Annual rate of change in GPO per labor-hour2.3%-3.6% 0.2% -1.1%

56 INFORMATION TECHNOLOGY IN THE SERVICE SOCIETY BOX 2~1 Notes on Terms and Data Presented in Tables 2. ~ Through 2.9 Data Source GPO (billions of Pleasures clef TYPO tgross domestic product originating in the 1982 dollars) tJ.S. economy) output prepared by the Bureau of Economic Analysis (BEA3, U.S. Depa~mem of Commerce. Major reYi sions covering the period from ~ 977 to 1988 were released by flue BEA in the January 199} Survey occurrent Business. In April 1991, the BEA further revised APE) figures for 1986 to 1988 and released 1989 figures Revenues reflect measures ~ gross Output by industry pre pared by the Bureau clef Labor Statistics, U.S. Department of Labor. Revenues (billions of 1982 dollars) I Employment I "millions of jobs) ! Number of hours worked Millions) Annual investment in IT (trillions of 1982 dollars) IT capital stack (billions Off 1982 dollars) Employment includes both full-time and '~art-time employs  ees; unpublished data, Bureau of Labor Statistics, U.S. De- partment Of Labor, Total hours worked by ail employees (including proprietors), unpublished data, Bureau of Labor Statistics' U.S. Depart- ment of Labor. Annual investment In IT derived from data of Stepl~en Roach of kid organ Staniey, which present IT spending In current dollars and in 1987 dollars. To obtain t982 dollar equ~va~ }ents, the 1987 dollar figures have been multiplied by the ratio of the f982 expenditure on IT for the entire service sector in current dollars (t.e., in 1987 dotters) to the 1982 expenditure on IT for the entire service sector in 1987 dc~t- }ars. This approximation omits possible adjustments due a different technology mix in ~ 9B' compared to that of 1987 but is the best that can be done with the aYaTJable dam. Figures on investment In IT ~r 1969 (Tables 2.3 to 2.9 of this reports not ,oresented in Roach's data, are estimated an the basis of the 1970 figure, adjusted by the annual rate of change between 1970 and 1975. Same as annual investment In IT. GPO per GPO (Row 1) dreaded by number of hours worked (Row 43. labor-hour ~ ~ 982 dollars) Annual ITJGPO Annual investment in IT as ~ percentage of (iPO (Row 5 divided by Row 1 )+ Annual rate ~Annual compounded percentage change between any year change in GRO per and the preYTous year. Iabor-hc~ur

IMPACTS OF INFORMATION TECHNOLOGY AT THE INDUSTRY LEVEL 57 ratio for durable manufacturing, while the capital expenditure per full-time employee for transportation, communications, and financial services, insur- ance, and real estate (FIRE) was larger than or comparable to that for durable manufacturing.2 The discussion of these industries was informed by a variety of sources: the individual expertise and knowledge of committee members, interviews with senior executives from the six industries, data drawn from federal and other sources, and published articles and books. However, because the data available at the industry level are limited in scope and detail, this chapter's brief vignettes of each industry are intended in part to capture the direct and intuitive observations of experienced industry executives and industry ana- lysts. Given the multidimensional nature of performance in the service industries and the resulting difficulty of isolating the specific impact of using IT vis-a-vis other factors, these vignettes are largely anecdotal but suggest the important factors that have affected overall industry perfor- mance and the scale and impact of the use of IT in each industry. Box 2.1 describes the origin of the data on each industry presented below. AIR TRANSPORT In the last two decades, the U.S. air transport industry has undergone remarkable change, expanding its equipment inventories and facilities rap- idly to handle substantial increases in passenger volume. In 1989, it oper- ated a fleet of some 4000 large jet aircraft,3 produced 433 billion revenue passenger-miles (RPMs),4 generated revenues of $67 billion (1982 dollars), and employed about 691,000 peoples (Table 2.3~. However, poor profit- ability and even sometimes net losses have characterized the industry for much of the last decade.6 The industry has also undergone considerable consolidation and today is highly concentrated. As of September 1991, four large air carriers (America West, Continental, Midway, and Pan American) had filed for bankruptcy protection, and Eastern Airlines ceased operating in January 1991. Twenty- two large carriers (out of over 200 certificated U.S. air carriers) account for 95 percent of total airline revenues for scheduled passenger operations.7 A major event in the last decade was the deregulation of the airline industry. Beginning in the early 1980s, there was a gradual lifting of gov- ernment regulations whose effect had been to reduce the efficiency of air- craft use and to generate incentives for the acquisition of certain kinds of equipment and for the maintenance of routes that did not optimize the use of resources for the country as a whole. Deregulation has allowed airlines to employ a much more efficient "hub-and-spoke" routing system, rather than the point-to-point routing used in a more regulated environment.8 Other

58 TABLE 2.3 Air Transport Industry INFORMATION TECHNOLOGY IN THE SERVICE SOCIETY 1969 GPO (billions of 1982 dollars) Revenues (billions of 1982 dollars; SIC 45a) Output measure (billions of RPMsC) Employment (millions of jobs) Number of hours worked (billions) Annual investment in IT (billions of 1982 dollars) IT capital stock (billions of 1982 dollars) GPO per labor-hour (1982 dollars) Annual IT/GPO RPM C/labor-hour Annual rate of change in GPO per labor-hourd 1980 1981 1982 1983 15.1 34.8 132b 0.4 0.8 0.0 0.0 20.0018.19 0.1%3.8% 174b267 o.g%e  17.3 16.8 19.0 22.8 52.1 45.2 41.4 44.2 254 248 260 281 0.5 0.5 0.4 0.5 1.0 o.s 0.9 0.9 0.7 1.0 0.9 1.2 1.0 1.8 2.4 3.2 17.70 6.2% 262 -2.7% 20.88 4.5% 285 17.9% 302 24.49 5.3% NOTE: See Box 2.1 for explanation of terms and sources of data given. aStandard Industrial Classification system code. b 1969 figure. important influences on the airline industry in the last decade have included significant increases in fuel costs, the 1981 strike of air traffic controllers, expensive purchases of new airplanes owing to need or to regulations call- ing for greater fuel efficiency or quieter aircraft, and a recessionary economy for much of the l980s that drove down demand for air travel. Extensive deployment of IT has been necessary to help the air transport industry deal with larger volumes of passengers and inquiries, higher operat- ing tempos, and more complex flight and ground operations. Indeed, the air transport industry is a major consumer of IT, having spent $3 billion in 1989 (1982 dollars) on IT. The importance of IT is suggested, though not proven, by the fact that in both the air-cargo and passenger-transport segments, air transport companies prospering in the last decade have tended to be those with strong IT systems (e.g., American Airlines, United Airlines). However, the wide range of external influences on the industry in the last decade makes it difficult to identify with confidence the impact of any single factor such as IT. Indeed, some analysts assert that the benefits of IT, substantial though they

IMPACTS OF INFORMATION TECHNOLOGY AT THE INDUSTRY LEVEL 59 TABLE 2.3 Continued 1984 1985 1986 1987 1988 1989 GPO (billions of 1982 dollars) Revenues (billions of 1982 dollars; SIC 45a) Output measure (billions of RPMsC) Employment (millions of jobs) Number of hours worked (billions) Annual investment in IT (billions of 1982 dollars) IT capital stock (billions of 1982 dollars) GPO per labor-hour (1982 dollars) Annual IT/GPO RPM C/labor-hour Annual rate of change in GPO per labor-hourd 23.4 23.2 28.4 33.0 32.4 31.7 47.9 51.1 305 336 0.5 57.7 62.3 65.1 66.6 367 404 423 433 0.5 0.6 0.6 1.0 1.1 1.2 1.2 1.1 3.8 1.5 3.1 1.7 4.6 6.9 7.6  21.30 24.27 6.4% 11.1% 23.15 4.5% 302 308 313 -5.5% -8.0% 13.9% 0.7 0.7 1.3 1.4 1.8 3.0 8.2 9.8 26.74 24.60 22.64 5.2% 5.5% 9.4% 328 321 309 10.2% -8.0% -8.0% CRPM, revenue passenger mile. dThe compounded annual rate of change in GPO per labor-hour from 1981 to 1989 was 3.1 percent. Labor productivity is sometimes measured in terms of GPO per labor-hour. eAverage annual rate of change in GPO per labor-hour from 1969 to 1980. may be to the leaders in the airline industry, are small compared to the general economic problems that the industry faces at present.9 In the air transport industry, IT has had an impact on both business and operational dimensions. In the business dimension, IT has been used to improve both the cost-efficiency and quality of performance as well as to provide services not otherwise available. Perhaps the most significant of these services is the use of computerized reservation systems (CRSs). CRSs are a classic example of IT use that has had a large-scale strategic impact on an industry. Generally owned by individual airlines, these systems con- tain fare and route information for both the CRS owner and other carriers, and they make it possible for agents to book passage on any carrier in the database. CRSs have enabled airlines to manage passenger loads and hub- and-spoke toutings that would otherwise have been extremely cumbersome and time consuming if possible at all-under less automated conditions. Two CRSs, SABRE (American Airlines) and APOLLO (United Air- lines), dominate the industry: in 1981, nearly 80 percent of travel agencies

60 INFORMATION TECHNOLOGY IN THE SERVICE SOCIETY with connections to CRSs were tied to one or both of these systems.~° Today, virtually all airline bookings are made through one of four airline- owned Cress. Carriers without their own CRSs pay CRS owners to have their flights and fares listed, thereby providing CRS owners with additional revenue. 12 Initially, CRSs simply replaced manual reservation systems and did not communicate either with the systems of other companies or with other de- partments within the owning company. However, by making relevant infor- mation about most scheduled flights available in a single database,~3 CRSs now enable passengers to use independent travel agents as well as agents from each participating airline as one-stop points of contact to gather sched- ule and fare information on all carriers providing the desired service and to make complete trip reservations, rather than having to call individual air- lines directly. In the future, it will be common for CRSs to provide gate- ways to other non-airline systems through which hotel reservations and automobile rentals can be obtained. CRSs have played an important role in facilitating the interlinkage of major and regional carriers through a process known as cross-coding. Re- gional carriers operate in specialized markets, whereas major carriers pro- vide what could be called "trunk" service on major routes. Flying between two small airports on opposite ends of the country usually requires coordi- nation of flights operated by a major carrier and by smaller regional carri- ers; from listing its flights on CRSs owned by major airlines, it is a rela- tively small step for a regional carrier to affiliate formally with a major carrier to provide a higher degree of operational integration. CRSs have enabled implementation of a fare structure that can be very finely tuned to maximize revenues through yield management,~4 and they have provided a mechanism for the rapid implementation of price changes and for monitoring the prices of competitors. Moreover, CRS-based IT figures prominently in marketing mechanisms that have flourished since deregulation (e.g., incentives and bonuses for travel agents) and indeed in the growth of the travel industry itself. For example, CRSs have been linked to "frequent-flyer" databases that have themselves been made opera- tionally feasible by IT. The potential drawback of CRSs is their ability to communicate the fares of one carrier very quickly to other carriers. As of this writing, the U.S. Department of Justice (DOJ) has alleged that carriers participating in various CRSs use postings of future fare changes to fix prices at artificially high levels. For example, the DOJ suit alleges that one airline announced on January 15 a price increase to take effect on February 1, and that other airlines noticed the posting and raised prices upward accordingly. In effect, CRSs are alleged to have replaced face-to-face meetings to set prices that are illegal under antitrust laws. In response, the airlines have argued that

IMPACTS OF INFORMATION TECHNOLOGYAT THE INDUSTRY LEVEL 61 they provide notices of future price changes for the benefit of consumers, who may wish to know when price increases are to take effect. The DOJ is seeking to enjoin the airlines from posting changes that will occur in the future. is The impact of IT on the air transport industry's operations has also been quite significant. For example, IT-based systems connected to CRSs keep track of flight arrivals and departures, and they manage the many changes in planned schedules that are inherent in complex planning opera- tions. Some airplane maintenance crews make use of knowledge-based "maintenance advisors" that guide mechanics through certain repairs. In- ventories of spare parts are maintained and supplies are replenished via IT- based systems, and airline food services are also being automated.~7 Computer-based flight simulators are used extensively to train pilots for reasons of cost and safety. An hour in a flight simulator costs far less than an hour of actual flight (hundreds of dollars vs. thousands of dollars), thereby reducing fixed costs associated with crew training. In addition, emergency situations can be simulated and pilot responses rehearsed with no real dan- ger to life or to an airplane, enabling the pilot to obtain experience and familiarity with situations that occur only rarely in actual flight. Today's simulators are so good that pilots for some commercial airplanes can be certified entirely on the basis of their experience in flight simulators. Flight safety is critically dependent on IT. Highly automated air traffic control systems manage thousands of en route flights. Weather an impor- tant influence on safety and flight schedules is monitored and predicted using IT-based systems. In-flight communications and navigation depend on state-of-the-art ITS Weight, and therefore fuel costs, are reduced by on-board IT that monitors every aspect of an airplane's flight performance. On-board computers complement pilots' abilities in flying airliners, thus enhancing safety.20 IT is becoming increasingly important to airplane design, since com- puter-aided design and engineering save time and design effort and reduce construction costs. In some cases, airframe designs stored in digital form are tested in simulated wind tunnels, and high-fidelity results can be ob- tained without a single model having to be built. The logistics planning requirements of major airlines managing hun- dreds of flights per day are formidable, especially with hub-and-spoke rout- ing. Today, centrally located dispatchers that previously operated out of three or four locations use IT-based systems to assist pilots with crew scheduling, flight plans, and fuel-load and weather-change calculations, an approach resulting in significant improvements in efficiency.22 (One estimate sug- gests that Delta Airlines was able to reduce the amount of unproductive crew time from 11 percent to 8 percent, thereby saving $14 million per year;23 United Airlines saved $6 million per year by installing, at a cost of

62 INFORMATION TECHNOLOGY IN THE SERVICE SOCIETY $500,000, a computer-based planning system for scheduling shift work.24) Working with expert systems that take on-line variables from local sites, such as the number of bags loaded onto a plane, crew schedulers and load planners are able to balance loads and to communicate directly with pilots. Airlines also use expert systems to generate least-cost routings of flights. The benefits of using IT to assist with operations are relatively easy to demonstrate quantitatively: for example, computer-assisted route planning provides for more optimal use of fuel and assignment of crews than would manual planning. However, specific financial bottom-line benefits of using IT are more difficult to quantify. A vice president of a major airline noted, We want to convert our unintelligent workstations to intelligent ones be- cause we've developed a number of programs that we believe are going to be significant in terms of their support and enhancement of customer ser- vice activities and other types of activities at the airports. Though the benefits are there, we cannot quantify them specifically. We're doing this to enhance the overall environment for the passenger. U.S. air transport companies have consistently led the worldwide indus- try in major innovations such as customer reservation systems, overnight package handling, special customer services (e.g., frequent-flyer programs and package tracking), and differential pricing systems to maintain load factors. Well-automated major U.S. airlines have maintained their relative competitiveness with other international carriers and have extended their penetration of U.S. and foreign markets. Indeed, none of the several senior executives from major airlines interviewed by the committee believed they could operate without extensive IT systems, and all pointed to important cost and quality improvements enabled by IT. Given the relatively poor financial performance of the industry as a whole,25 the ability to make necessary investments in IT may have been an important element in the survival of current industry leaders. In the future, IT may be used by the airline industry to create a variety of additional benefits. For example, airport operations and procedures could be greatly streamlined if the current manual system of check-in were auto- mated to link CRSs and airport systems.26 In-flight time could be used to entertain travelers or to help them be more productive; a number of pilot projects to provide relevant services have been launched or planned.27 IT could also be used in customized, proactive marketing for example, poten- tial customers could be alerted to the existence of inexpensive fares. TELECOMMUNICATIONS Over the last decade, the telecommunications industry has undergone dramatic structural change. It has also been highly successful in financial

IMPACTS OF INFORMATION TECHNOLOGY AT THE INDUSTRY LEVEL 63 terms, although whether this success is despite or because of these changes is open to debate. Revenues and contributions to the gross domestic prod- uct have grown substantially for telecommunications (Table 2.4), and the industry's rate of growth in productivity of 4.8 percent per year over the period from 1979 to 1989 exceeds that of manufacturing (3.3 percent per year), as indicated in Table 1.1 in Chapter 1. In 1990, the telecommunica- tions industry supported over 136 million access lines.28 Although data on total minutes of local calls are not available, the Federal Communications Commission reports that 308 billion minutes in interstate calls were con- nected in 1990 through these access lines.29 Two major factors have shaped contemporary telecommunications: (1) deregulation, in combination with the divestiture of AT&T, and (2) advances in technology, with concomitant reductions in cost and changes in product offerings. The divestiture of AT&T and the consequent restructuring of the regulatory environment completely changed the character of the domestic tele- communications industry. Prior to December 1983, the industry operated as a regulated monopoly. However, in August 1982 the courts approved a consent decree that required divestiture of AT&T intra-LATA service (i.e., service within local calling areas) to seven regional holding companies by January 1, 1984. This action created more direct competition in inter-LATA (long-dis- tance) services and equipment. The combination of the breakup of the Bell system, deregulation, and the proliferation of new, technology-based busi- nesses has fragmented the telecommunications industry; the result has been the emergence of about a half-dozen major long-distance/interexchange carri- ers such as AT&T, MCI, and U.S. Sprint, and several hundred smaller carriers operating via COMSAT and leased private lines. Such competition, combined with federally regulated charges for local access, has driven long-distance calling prices down and volume up, and much of the growth in the telephone industry in the last decade has been associated with the provision of long-distance services. Moreover, the separation of long-distance service from local carriers has forced local carriers to seek additional sources of revenue, and they have responded to these pressures by using IT to improve operations (e.g., collection of revenues from unbilled calls30) and to provide new services (e.g., Caller ID) to supplement tradi- tional local telephone service ("plain old" telephone service). Some have argued that today's regulatory environment, which mandates equal charges for local access to all long-distance providers and thus has reduced pricing differentials among providers, has placed an increased emphasis on quality of service and reliability. However, the growth in the number of services that are critically dependent on telecommunications and the experience of several major and highly publicized telecommunications outages have raised the question of whether current levels of reliability are as high as those that existed before the breakup of AT&T.3i

64 TABLE 2.4 Telecommunications INFORMATION TECHNOLOGY IN THE SERVICE SOCIETY 1969 1980 1981 1982 1983 GPO (billions of 1982 dollars) Revenues (billions of 1982 dollars; (SIC 481,2,9a) Minutes of calls (billions) Employment (millions of jobs) Number of hours worked (billions) Annual investment in IT (billions of 1982 dollars) IT capital stock (billions of 1982 dollars) GPO per labor-hour (1982 dollars) Annual IT/GPO Annual rate of change in GPO per labor-hourb Minutes of calls per labor-hour 33.2 71.4 75.7 77.3 83.2 36.3 87.7 91.8 92.8 92.3 1,733 1,787 1,853 1,923 0.9 1.1 1.1 1.1 1.0 1.9 2.3 12.7 22.7 54.1 115.8 17.17 30.40 38.2% 31.8% 5.3% 2.4 2.3 2.1  21.0 17.3 125.3 31.97 27.8% 5.2% 126.5 33.02 22.3% 3.3% 737.8 754.6 791.5 14.3 124.6 39.49 17.2% 19.6% 912.7 NOTE: See Box 2.1 for explanation of terms and sources of data given. The growth of telecommunications parallels advances in information technology, because the technological substrate of telecommunications is wholly dependent on IT. The telecommunications industry's investments in IT have been very large, amounting to $12 billion to $15 billion per year in the late 1980s (1982 dollars), and those figures are for hardware only; they do not include the substantial investments in software associated with the expansion and upgrading of telecommunications networks and services. Meanwhile, during the 1980s the capabilities of underlying microelectronic technologies for telecommunications were improving at average rates of 40 to 50 percent per year. Over the years, major innovations have improved underlying service and hardware technology, including satellite, cellular, and fiber-optic com- munication systems. Historically, applications of IT have driven changes in long-distance telephone service, beginning with the introduction of direct dialing between 1950 and 1952. Since then, the telephone industry has witnessed the introduction of computer-based toll operator assistance sys

IMPACTS OF INFORMATION TECHNOLOGY AT THE INDUSTRY LEVEL 65 TABLE 2.4 Continued 1984 1985 1986 1987 1988 1989 GPO (billions of 1982 dollars) Revenues (billions of 1982 dollars; (SIC 481,2,9a) Minutes of calls (billions) Employment (millions of jobs) Number of hours worked (billions) Annual investment in IT (billions of 1982 dollars) IT capital stock (billions of 1982 dollars) GPO per labor-hour (1982 dollars) Annual IT/GPO Annual rate of change in GPO per labor-hourb Minutes of calls per labor-hour 80.8 81.7 85.0 93.2 97.8 98.8 89.2 91.5 91.1 92.8 98.4 100.6 2,045 2,145 2,207 2,267 2,385 2,460 1.0 1.0 2.1 2.1 13.8 15.9 122.0 121.2 -3.9% 0.9 2.0 1.0 ok  2.0 16.9 15.9 121.1 39.35 42.42 19.4% 19.9% 3.7% 2.0 2.0 13.8 116.7 45.98 48.06 49.97 17.1 % 14.2% 14.0% 7.8% 8.4% 4.5% 4.0% 960.1 1033 1101.3 1118.4 1172 1244.3 aStandard Industrial Classification system code. bThe compounded annual rate of change in GPO per labor-hour from 1981 to 1989 was 5.7 percent. Labor productivity is sometimes measured in terms of GPO per labor-hour. tems, automatic credit card validation and billing, automated directory as- sistance systems, automatic coin recognition in pay-telephone systems, and, recently, voice-recognition systems that build on the digitization of network and voice-recognition technology.32 Such innovations have reduced the amount of labor required to provide telecommunications services33 and have facilitated a gradual shift in employment toward executive and staff occupa- tions. There has also been ongoing progress relating to encoding, process- ing, and transmitting messages (e.g., voice and video compression), all of which provide for more efficient use of telecommunications facilities. A substantial amount of IT is used to support the people operating and managing the network. Operations support systems monitor traffic as well as the health and condition of facilities and switches, provide telemetry, detect faults, and provide alerts. Databases record what is in the network and how it is hooked up. Operations support systems help in keeping track of circuits and reconfiguring them. IT-based systems are used to collect vast amounts of data from multiple sources to support rate filing and net

66 INFORMATION TECHNOLOGY IN THE SERVICE SOCIETY work planning. Today's use of test equipment often incorporates expert- system technology in applications that automatically signal when mainte- nance is necessary and in techniques for optimizing network routing, espe- cially under emergency conditions. Managing the huge volumes of billing data that are generated in provid- ing telecommunications services would not be possible without sophisti- cated IT systems. Billing is complex because companies must continuously record each use of every service, and then rate it and price it based on multiple variables (distance, time, specific service, quality of service, pric- ing plans, and so on). For this reason, automated billing systems are built directly into central switches, and most billing systems are therefore rela- tively inflexible. Enhancements to billing systems have transformed them from being sim- ply support systems that emphasize rates, prices, invoicing, and billing, to being strategic systems that emphasize accuracy, speed, calling patterns, and other attributes of customer service. For example, MCI developed a new marketing program by using IT for sophisticated billing and data processing: Friends and Family involves accurate monthly tracking of customers and call- ing circles (designated groups of people who call each other) across their use of various telecommunications services. Other billing system enhancements have given rise to customer billing and usage management tools that have historically been focused on corporate users. For example, MCI has a rela- tively new product, Service Without Paper, that provides corporate telecom- munications managers with billing and usage data on optical disk. o Telephone credit cards (calling cards) were made possible by using IT,34 and they represent an area of current innovation. Calling cards serve as marketing tools, promoting easy access to products and services, and they are linked with proliferating services: electronic message delivery, stock quotations and weather forecasts, language interpretation, speed dial- ing, and teleconferencing. Long-distance carriers are now developing voice- recognition systems that can substitute for operators in handling routine interactive telephone transactions (e.g., collect calls). In the future, en- hanced calling cards may be indistinguishable from automated teller ma- chine cards.35 Deregulation and technological advances have both contributed to shifts in the mix of products produced by the telecommunications industry. Fall- ing prices for long-distance calls have contributed to a shift in the mix of local to long-distance calls. Telecommunications companies have evolved from defining the service they offer exclusively in terms of providing facili- ties to transmit the electrical impulses representing voices or data to rede- fining themselves as enterprises that help customers collect, process, and distribute data. More specifically, using IT, telecommunications companies provide more information to customers on the way infrastructure is used

IMPACTS OF INFORMATION TECHNOLOGY AT THE INDUSTRY LEVEL 67 (e.g., billing and network management services), new information-based modes of access to information (virtual private networks, calling cards), and gateways to information services, if not information services themselves. Thus there is a blurring of the traditional categories of telecommunications and information processing, as well as a convergence of public and private networking.36 In recent years, much of the capital investment in the telecommunica- tions industry has been driven by the enhancement of the communications infrastructures installed by both regional and national carriers. Today, the infrastructure of modern telecommunications makes increasing use of broad- band transmission and digital switching. Compared to analog transmissions, digitized communications improve signal fidelity and, through signal processing, use available bandwidth more efficiently. Digitized communications also facilitate data transmission (al- though voice signals are increasingly digitized), thereby changing the mix of voice and data traffic, and enable as well the use of companion overlay networks for network management functions (e.g., call setup) and for new services (e.g., automatic number identification Box 2.2~. The entire AT&T interchange network is digital today, as are about 40 to 50 percent of central office switching facilities in local exchanges.37 Broad-band transmission is associated primarily with optical fiber cable, which offers huge gains in capacity over that provided by copper cable. The bandwidth provided by fiber will greatly exceed what is needed to handle the expected traffic volume upon installation and will thus be able to accommodate increased traffic or new services. (At the same time, the prospect of "dark fiber" or unused capacity contributes to the controversy over how fast to increase network bandwidth.) Thanks in part to deregulation, new telephone services are now intro- duced much more rapidly than before. IT in particular, programmable digital switches has been integral to the rapid development of such inno- vative services as call waiting, caller identification, and call forwarding. By extending the traditional notion of plain old telephone service, such services increase the number of options for local customers and generate substantial revenue for local telephone companies today. IT has also been used in the telecommunications industry to enhance and improve corporate image and the quality of customer service, even when these considerations do not relate directly to calling volume. MCI's Richard Liebhaber described a $35 million investment to provide the capa- bility for customer service representatives to display customer invoices and correspondence on graphics terminals, saying, "It is an opportunity to im- prove the image, the feel, the touch of MCI. Intuitively, I would tell you it's going to provide market share, customer satisfaction, bottom line. I can't prove it. I'm not going to waste my time trying."

68 INFORMATION TECHNOLOGY IN THE SERVICE SOCIETY BOX 2.2 Applications for Automatic Number Identification Autornatie Number Identification (ANI) is a service usually (though not always) provided through an integrated services digital network (ISDN) that enables ~e telephone number of the calling party to be transmitted along with the signal that carries the call itself. Telemarketing organizat~c~ns initially used Alp! to identify callers; the calling number would trigger a database inquiry, enabling the marketer to identify the caller and his or her previous history. However many callers were disconcerted by the fact that their identities were known to the markewr without their providing any identification. New applications of ANI are being used to provide different levels of service. For example, if a caller Is known from ,oreY'~us interactions to speak Spanish, ANI can be used to route calls from that party to Spanish- speaking operators. Credit card holders can be routed to different ser- vice staffs depending on the type of card they hold Equipment service organizations can route calls SO the appropriate staffs based an databases that coordinate telephone numbers with equipment present at given sites. SOURCE. Sweeney, Terry. 1993. `'Carriers, Users Discover New Appli- cations for Automatic Number ID," Commanicatio'?s Week Manual 4, p. 27. Future advances in telecommunications are likely to include multime dia transmission (e.g., combined two-way video and audio) made possible with the wide availability of digital networks, the assignment of telephone numbers to individuals rather than locations (resulting in a single telephone number for an individual regardless of where he or she is), and automatic translation services (giving, e.g., an American and a Japanese businessperson the opportunity to converse through a telephone, but both in their native tongues). Broad-band networking technologies supporting transmission speeds in excess of 45 megabits per second will be deployed widely in the next decade; by contrast, the technology that supports most home telephones today is limited to transmission speeds measured in tens of kilobits per second. Broad-band networking technologies will require substantial de- ployment of fiber optics to end users (initially commercial ones). Wireless communications (cellular telephone, personal communications services) is a market in which considerable growth is expected (subject to availability of spectrum bandwidth). Advanced intelligent network technology is expected to continue facilitating the development of additional customized communi cations services.

IMPACTS OF INFORMATION TECHNOLOGY AT THE INDUSTRY LEVEL 69 Implementing these advances is likely to have a substantial impact both on users and providers of telecommunications technology. Using sophisti- cated telecommunications technology most effectively will most probably require substantial learning of new skills comparable to the new skills de- manded of those using increasingly powerful software packages on personal computers. As importantly, sophisticated telecommunications technology will depend increasingly on the capability to develop complex software, a task that routinely pushes the state of the art of current knowledge in com- puter science and software engineering. While both financial performance and growth in productivity in the telecommunications industry have been strong, the benefits of technological innovation and investments have sometimes appeared in other segments of the economy. Private companies, whether manufacturing or service firms, may realize improvements in productivity from the installation of their own switches and terminal equipment, but these improvements are not reflected in the data for public providers of communications services. Nonregulated companies outside the industry have been able to install their own wide- area networks and "cherrypick" profitable service opportunities by selling such systems and services to others. Returns from these activities show up in sellers' and users' books of account, rather than those of telecommunica- tion providers. There is a substantial cross-substitution between the "prod- uct" and "service" components of the telecommunications field and an in- creasing merging of the telecommunications business between equipment manufacturers and service providers as currently defined. RETAIL AND WHOLESALE TRADE i] The retail and wholesale trade industries are responsible for the order- ng, collection, distribution, and sale of goods. Both industries have been growing. Retail revenues increased from $447 billion in 1980 to $628 billion in 1989 (1982 dollars; Table 2.5~. Wholesale revenues grew from $310 billion in 1980 to $401 billion in 1989 (1982 dollars; Table 2.6~. For wholesale and retail trade taken together, the 1989 investment in IT in constant dollars was 2.6 times the amount invested in IT in 1980. The retail industry consists of a broad collection of segments with rather different businesses (department stores, restaurants, specialty stores, gro- cery stores, and so on) that vary greatly in terms of the categories of the items they sell, the range of prices of the items, the market size for the items, and profit margins. Retail trade is dominated by relatively few, large, national or regional chain operations specializing in merchandise trade, food, and pharmaceuticals.38 A much larger number of smaller "mom-and- pop" retail operations are locally based; such operations are relatively easy to start (compared to ventures in telecommunications or air transport) be

70 TABLE 2.5 Retail Trade INFORMATION TECHNOLOGY IN THE SERVICE SOCIETY 1969 1980 1981 1982 1983 GPO (billions of 1982 dollars) 212.7 281.7 286.4 287.5 307.8 Revenues (billions of 1982 dollars; SIC 52-59a) Employment (millions of jobs) Number of hours worked (billions) Annual investment in IT (billions of 1982 dollars) IT capital stock (billions of 1982 dollars) GPO per labor-hour (1982 dollars) Annual IT/GPO Annual rate of change in GPO per labor-hourb 326.5 446.5 453.6 12.3 16.8 16.9 22.9 27.4 27.4 0.1 4.6 5.1 0.4 9.31 0.1% 10.0 13.2 10.28 10.44 1.6% 1.8% o.9% 1.6% 459.9 16.9  27.2 5.8 6.6 485.4 17.4 27.8 7.0 20.7 10.56 11.06 2.0% 2.3 1.2% 4.7 NOTE: See Box 2.1 for explanation of terms and sources of data given. cause their requirements for capital and technology infrastructure are rela- tively small. For example, a means for displaying merchandise is often the minimum capital requirement, a task that can in some cases be accom- plished by publishing a catalog. Wholesalers provide all activities involved in selling goods or services to those who buy for resale or business use. They exist as an industry because they provide efficiencies that manufacturers often are unable to support themselves.39 Given the heterogeneity of the products sold by retailers and wholesal- ers, aggregate trade output is best measured in terms of financial value. (Output measures for trade are discussed in greater detail in Appendixes B and C.) However, measured productivity varies greatly among such differ- ent elements of the retail trade industry as supermarkets and other food stores, restaurants, apparel and accessory stores, new-car dealers, gas sta- tions, home furnishings and equipment stores, and drug stores. Moreover, financial measures do not necessarily capture other impor- tant elements of performance. For example, the 1980s saw improvement in measures of quality such as the variety of products, geographic dispersion, demographic specialization within stores, fashion responsiveness, and hours

IMPACTS OF INFORMATION TECHNOLOGY AT THE INDUSTRY LEVEL 71 TABLE 2.5 Continued 1984 1985 1986 1987 1988 1989 GPO (billions of 1982 dollars) Revenues (billions of 1982 dollars; SIC 52-59a) Employment (millions of jobs) Number of hours worked (billions) Annual investment in IT (billions of 1982 dollars) IT capital stock (billions of 1982 dollars) GPO per labor-hour (1982 dollars) Annual IT/GPO Annual rate of change in GPO per labor-hourb 334.0 354.4 377.5 521.7 549.5 574.9 18.3 18.9 29.3 8.9 29.9 8.5 25.9 30.2 11.40 11.86 2.7% 2.4% 3.1% 4.0% 19.5 30.6 9.0  34.2 12.35 2.4% 4.2% 371.6 399.2 595.6 623.5 20.1 20.6 31.5 32.2 8.0 36.8 10.5 41.1 11.7912.39 2.1%2.6% 628.3 21.2 32.8 10.6 44.8 12.54 2.6 5.1% 1.2 aStandard Industrial Classification system code. bThe compounded annual rate of change in GPO per labor hour from 1981 to 1989 was 2.3 percent. Labor productivity is sometimes measured in terms of GPO per labor-hour.  i] open. Significant increases in the diversity of their products, the growth of chain store branches, and the proliferation of small specialty chains may have masked other elements of improvements in productivity in large chain operations. In addition, because of "overstoring" and competitive pres- sures, margins at both retail and wholesale levels dropped during the 1980s, despite major improvements in operational performance enabled by new IT capabilities. Retail and wholesale enterprises seek faster receipt of goods, lower nventories and more inventory turns per year,40 a better fit between goods offered and goods bought, and reduced labor costs. In retailing, consider- able effort is made to attract customers to retail establishments, because the marginal cost of providing service to a new customer is relatively low. In retailing, investments in IT increased from $4.6 billion in 1980 to $10.6 billion in 1989 (1982 dollars; Table 2.5~. In wholesaling, invest- ments in IT approximately tripled in the 1980s, from $4 billion in 1980 to $11.6 billion in 1989 (1982 dollars; Table 2.6~. The vast bulk of these expenditures was made by large establishments, which now depend on IT for planning and control of inventories, merchandising, and logistics man- agement. By contrast, smaller retailers use relatively little IT, except in

72 TABLE 2.6 Wholesale Trade INFORMATION TECHNOLOGY IN THE SERVICE SOCIETY 1969 1980 1981 1982 1983 GPO (billions of 1982 dollars) Revenues (billions of 1982 dollars; SIC 50,1a) Employment (millions of jobs) Number of hours worked (billions) Annual investment in IT (billions of 1982 dollars) IT capital stock (billions of 1982 dollars) GPO per labor-hour (1982 dollars) Annual IT/GPO Annual rate of change in GPO per labor-hourb 149.0 210.0 4.2 8.8 0.3 0.8 200.1 212.7 310.3 4.0 10.4 16.97 17.81  0.2% 2.0% 0.4% 219.0 222.2 317.6 5.6 5.7 11.2 11.4 5.0 13.2 18.64 2.4 4.7 318.8 315.4 5.6 11.2 5.9 16.5 5.3% 5.6 11.3 10.4 23.2 19.70 4.7% 0.4% NOTE: See Box 2.1 for explanation of terms and sources of data given.  catalog sales operations. IT has contributed to the growth observed in these two industries and also to the growing competition between them: large retail chains now use IT to provide many of the services wholesalers used to provide, and wholesalers have linked "downstream" to tie retail customers together into voluntary chains. IT has become critically important to the retail and wholesale trade industries as a competitive tool; instead of merely providing information at greater speeds, IT can be used to help manage information that can then be coupled with strategic planning to optimize the efficiency of a business. Indeed, the National Retail Federation credits the use of IT with helping to cushion the impact of the recent recession: "Technology helped to cushion the blow. Many retailers had better information about their sales and inven- tories than in the last recession and were thus able to more selectively promote merchandise."4~ By enabling better and faster matching of inven- tories and customer needs, IT can decrease the need for retail markdowns, a major cost in retailing. In the 1980s, wholesalers often installed direct links to suppliers through electronic data interchange and support systems for retailers to deal with stock management, merchandising, and price marking. Wholesalers for

IMPACTS OF INFORMATION TECHNOLOGY AT THE INDUSTRY LEVEL 73 TABLE 2.6 Continued 1984 1985 1986 1987 1988 1989 GPO (billions of 1982 dollars)254.9267.1284.7284.3290.6 304.7 Revenues (billions of 1982 dollars; SIC SO,la)347.2362.2371.8373.4386.9 400.8 Employment (millions of jobs)5.96.16.16.26.4 6.6 Number of hours worked (billions)11.912.112.212.312.8 13.1 Annual investment in IT (billions of 1982 dollars)13.912.313.79.612.6 11.6 IT capital stock (billions of 1982 dollars)32.338.845.547.651.7 54.0 GPO per labor-hour (1982 dollars)21.4522.0023.4023.0522.78 23.30 Annual IT/GPO5.5%4.6%4.8%3.4%4.3% 3.8% Annual rate of change in GPO per labor-hourb8.9%2.6%6.4%-1.5%-1.2% 2.3% aStandard Industrial Classification system code. bThe compounded annual rate of change in GPO per labor hour from 1981 to 1989 was 2.8 percent. Labor productivity is sometimes measured in terms of GPO per labor-hour. foods, pharmaceuticals, and hospital supplies placed computerized ordering systems directly on customers' premises to great effect. These investments tended to shift shares of the market among wholesalers (or between whole- salers and retailers) rather than increase the total volume of products, so that benefits were frequently not captured in measures of output. Although volume grew during the 1980s without proportionate increases in employ- ment, competitive pressures kept wholesalers from realizing significant gains in profitability. The number of products and the complexity of services offered by wholesalers expanded enormously during this period. Conse- quently, the data on profits suggest that wholesalers were passing benefits from their investments in IT through to their customers and suppliers. A current trend is for large retailers to develop their own direct connec- tions to suppliers and manufacturers; both Kmart and Wal-Mart have direct links to over 2000 vendors. IT is an enabling element of such connections. In one case, point-of-sale data from the retailer have been used to adjust the mix of sizes of a manufactured product based on how frequently various sizes are sold.42 The fact that the trade industries are very labor-intensive has led to inter- est in technologies and processes that reduce labor costs. For example, a top

74 INFORMATION TECHNOLOGY IN THE SERVICE SOCIETY executive of a major supermarket chain noted that "scanning thas] had the highest impact of any technological advancement from IT in this business because of its improvement on productivity at the front end and through im- proved [e.g., faster] checkout and price removal, taking the prices off the products and not having to do price changes." Similarly, retailers have used IT to improve labor scheduling and to minimize costs. Scanning systems are used to read a product's universal pricing code label at the cash register, look up the product's price and the applicable tax, and record this information at the register as part of the transaction. Thus price changes can be recorded in a central database at the store rather than on each item or on the shelves. Moreover, the cashier is freed from having to enter prices on the register manually. Both attributes can improve accu- racy and reduce requirements for personnel; they also change the mix of skills required of cashiers. In addition, data made available at the point of sale facilitate accurate analysis of what gets sold, when, and, sometimes, to whom, while faster access to information facilitates better ordering, market- ing, and merchandising; improved management of inventory; and improved scheduling of personnel. Communications systems that link point-of-sale systems in stores to headquarters allow daily and even hourly analysis of sales and trends. A major benefit can be better insights into what sells well and poorly, thus minimizing overstocking and deep discounting. In some cases, point-of-sale systems (whether or not using scanners) have been tied to inventory control systems. Such systems are often used to link retail stores and wholesale vendors through electronic data interchange systems and allow nearly paperless and highly automatic ordering. In some instances, these systems allow orders to be shipped from wholesalers di- rectly to retail stores, bypassing warehouses. The large-scale deployment of point-of-sale systems and their integra- tion into centralized ordering and inventory control networks have been important elements in the expansion of the giants in the retail industry. For example, by using standardized ordering practices and inventory control procedures and drawing on the expertise of the entire corporation regarding sales and operations techniques, a new branch of a retailer like McDonalds can be opened more economically than a store of comparable size that lacks an existing support infrastructure. One result of this trend has been a division in the number of companies as a function of size- the retail indus- try now consists primarily of a handful of giants and a plethora of small institutions with one or a few stores under their control. IT has been used effectively in specialized trade applications. In furni- ture retailing, video catalogs are being used to show a large variety of fabrics and to aid in room planning, special orders, and follow-up, allowing the customer to more easily visualize and compare products. Restaurants today often use IT to perform location and layout analyses; manage recipes

IMPACTS OF INFORMATION TECHNOLOGY AT THE INDUSTRY LEVEL 75 and menus; track supplies, markets, and costs; and handle general ledger accounting (accounts receivable and payable, payroll accounting and labor scheduling, and financial reporting). Supermarkets are using IT to generate cents-off coupons at the point of sale (i.e., at the cash register) that are customized to a particular consumer's shopping patterns.43 In other cases, IT is applied across functional activities. An example is the use of a computer-aided design (CAD) system in the general merchan- dise segment. Without CAD, several individuals with three separate spe- cialties are needed to design floor plans, determine shelving requirements, and designate specific items to be placed in exact locations. Today, through the combined use of technology, historical data, and projections, individual functions can be reengineered into one process to best achieve the final goal of shelf-space management. -A - ' ' IT is also used to support buyers in making decisions about the mix of products for individual stores and to determine optimal item-to-store allocations. IT can be used to extend the activities of a business. A retail store may sell to other organizations the customer information captured in its market- ing database. Telemarketing, which first relied on telephones and now involves combinations of telephones and computer systems, is wholly de- pendent on IT for collecting information on existing customers and new prospects, for disseminating information through promotional efforts, and for conducting various other sales-related activities. In addition, computer systems are used throughout the telemarketing process, for sorting, merg- ing, and purging lists of prospective customers; distributing leads to agents; preparing for calls; dialing contacts; delivering sales presentations; refer- encing information; scheduling the next contact; recording contact informa- tion; carrying out the fulfillment process; and management reporting.44 A most promising future application of IT is its use to track the number of people entering a store, an approach pioneered by Kmart. At present, few stores can count how many people pass through its doors; measures of success (e.g., for sales and advertising) derive from measures of sales vol- ume (number of purchases). Yet the conversion of potential customers (those who visit the store) into actual customers (those who make a pur- chase) is a critical but poorly understood element of the retail business, and a small upward change in conversion ratios could have a dramatic effect on sales volume. Kmart believes that the installation of IT to count customers will enable managers to determine conversion ratios with unprecedented accuracy and will guide them in their improvement. HEALTH CARE In recent years, public concern over the health care industry has reached unprecedented levels. The industry provides a huge volume of service:

76 TABLE 2.7 Health Care INFORMATION TECHNOLOGY IN THE SERVICE SOCIETY 1969 1980 1981 1982 1983 Health services GPO (billions of 1982 dollars)a 73.4133.8138.1142.0143.7 National health expenditures (GPO, SIC 80,b estimated billions of 1982 dollars)a n/a309.6321.4326.1330.1 Revenues (billions of 1982 dollars; SIC 80b) 120.1194.5204.8213.6223.2 Employment (millions of jobs) 3.25.75.96.26.3 Number of hours worked (billions) s.6c9.810.210.710.8 Annual investment in IT (billions of 1982 dollars) 0.91.91.81.52.8 IT capital stock (billions of 1982 dollars) 3.38.68.98.910.0 GPO per labor-hour (1982 dollars) 13.0113.6413.4913.2713.25 Annual IT/GPO 1.2%1.4%1.3%1.1%2.0% Annual rate of change in GPO per labor-hourd 0.4%-1.2%-1.6%-0.2% NOTE: See Box 2.1 for explanation of terms and sources of data given. aGPO figures differ substantially from figures often reported as "national health expenditures" but are presented for the sake of consistency with other industries. National health expendi- tures are the most comprehensive measure, including both public and private spending on health care. The national health expenditure (NHE) figures presented in constant 1982 dollars are estimated on the basis of NHE figures in current dollars widely available in various publi Americans made 1.3 billion visits to physicians for ambulatory care and 31 million visits to hospitals in 1988.45 The health care industry absorbs a large and increasing percentage of the income of workers in all sectors: the private health care industry generated $120 billion in revenues in 1969 and $287 billion in 1989 (1982 dollars; Table 2.7~.46 It is also a large employer that in 1989 provided 7.9 million jobs. Table 2.7 describes the evolution of the health care industry over the last decade. A key element of the U.S. health care system is the massive presence of third-party payers, both private and government, that pay for most or all of  the health care received by individuals. A large number of third parties obtain the resources to pay for these services from some combination of the individual and the employer for whom the individual works, or through taxes levied on the population at large.

IMPACTS OF INFORMATION TECHNOLOGY AT THE INDUSTRY LEVEL 77 TABLE 2.7 Continued 1984 1985 1986 1987 1988 1989 Health services GPO (billions of 1982 dollars)a National health expenditures (GPO, SIC 80,b estimated billions of 1982 dollars)a Revenues (billions of 1982 dollars; SIC 80b) Employment 145.0 148.6 150.5 334.3 340.2 341.7 230.6 239.8 252.8 266.0 158.0 161.1 164.4 345.3 351.4 362.6 279.2 286.6 (rrullions of jobs) 6.5 6.6 6.9 7.2 7.5 7.9 Number of hours worked (billions) 11.1 11.4 11.7 12.2 12.8 13.5 Annual investment in IT (billions of 1982 dollars) 2.8 2.7 3.0 3.1 3.5 3.6 IT capital stock (billions of 1982 dollars) 11.1 11.8 12.7 13.6 14.7 15.7 GPO per labor-hour (1982 dollars) 13.08 13.07 12.85 12.97 12.59 12.14 Annual IT/GPO 2.0% 1.8% 2.0% 2.0% 2.2% 2.2% Annual rate of change in GPO per labor-hourd -1.3% -0.1% -1.7% 0.9% -2.9% -3.6% cations. Current-dollar NHE figures are deflated by a multiplicative factor that is the ratio of the GPO for health services for a given year in 1982 dollars divided by the same quantity in current dollars for that year. bStandard Industrial Classification system code. CHealth care hours worked in 1969 are extrapolated from 1979 ratio of jobs to hours worked. dThe compounded annual rate of change in GPO per labor-hour from 1981 to 1989 was -1.3 percent. Labor productivity is sometimes measured in terms of GPO per labor-hour. Rising health care costs have been attributed to many different causes, including the prescribing of newer, more complex, and more costly medical procedures with only marginally better clinical outcomes; a litigious cli- mate that encourages the practice of defensive medicine; the transformation of the underpaid and partly volunteer nursing force of the 1960s and early  1970s into the more professional hospital and nursing staffs of the 1980s; third-party payment of the medical expenses of large numbers of patients who are therefore not motivated to seek lower-cost care; and an aging popu- lation that can be expected to incur higher medical costs. These factors suggest that output measured by health care expenditures is problematic at best. The physical output of the health care system is also notoriously diffi- cult to measure. For example, using treatment outcomes as a measure in

78 INFORMATION TECHNOLOGY IN THE SERVICE SOCIETY individual cases is complicated by the fact that outcomes depend on factors other than the treatments provided, such as preadmission health of the pa- tient. Measures of public health, for example, incidences of disease, are complicated by external factors such as the increasing medical needs of an aging population. Issues such as unequal access to health care further compound the difficulties of measurement. Another key issue is measuring the quality of care that patients receive. Against this backdrop, technology, including IT, has had an undeniable impact on the success with which many medical problems are treated, e.g., cancer as treated with computer-controlled beams of radiation. IT is today an integral part of diagnosis and therapy. Microelectronics and embedded computing are essential parts of digital thermometers and sphygmomanom- eters; high-speed computing is at the heart of computerized tomography scanning and magnetic resonance imaging. Laboratory tests are increas- ingly automated, as computer-based instrumentation is used to take readings with consistency, reliability, and speed equal to or greater than that of human technicians. Monitoring of the vital signs of hospital patients and automatic signaling of the need for clinical action are performed more ef- fectively and reliably by IT than by human monitors. The use of IT in the business side of the health care industry is a different matter. As a whole, the health care field has been slow to use IT in its business operations and record-keeping activities. Nevertheless, pressure on the health care industry to explore the use of IT is now being exerted by a variety of factors, including the explosion of medical knowl- edge, the larger volume of clinical information potentially relevant to in- dividual patients, the increasingly cooperative nature of medicine, the need for more detailed clinical audit trails to defend against potential malprac- tice litigation and to ensure proper third-party payment, greater regulation, greater competition among health care institutions, and greater pressure for outcomes assessment and improvement of quality and patient satisfac- tion. By far the most common use of information systems in health care is for billing, accounting, and administrative tasks, which typically include the admission, transfer, and discharge of hospital patients; materials manage- ment; and scheduling and management of human and physical resources. Increasingly, hospitals recognize the need for integrating these tasks across departments. Integrated health care data systems can pass information elec- tronically in a common format among all departments in the system, thus reducing or eliminating the need to reenter data manually. With an inte- grated system, billing information from the laboratory and the pharmacy can be electronically integrated in a hospital's financial office. Billing for consumable items used by a patient can be integrated with the hospital's inventory control system. In some cases, inventory control systems them

IMPACTS OF INFORMATION TECHNOLOGY AT THE INDUSTRY LEVEL 79 selves are tied directly to hospital supply companies, and replacement or- ders are generated automatically when necessary. The actual clinical practice of medicine has not effectively used IT for cost reduction as widely as would be desirable. A recent study examined the potential of IT-based physician order entry in an Indianapolis hospital to steer physicians toward lower-cost treatments; the result was that patient bills were reduced by an average of 13 percent.47 The networked PC-based system used software that alerted physicians to the cost of various drugs and tests, warned about potential drug interactions, and specifically discour- aged the use of certain expensive drugs under particular circumstances. In the future, computer-based patient records (CPRs) hold considerable promise for improving health care and reducing costs. Today, patient records are contained on numerous slips of paper (e.g., laboratory tests, physician notes perhaps produced by transcribing the physician's dictation using a word processor) stuffed into a folder and perhaps arranged in chronological order. However, in some institutions CPRs are beginning to supplement paper records, over which they have many potential advantages. CPRs are or can be more legible, more current, more reliable and accurate, more comprehensive, more easily accessible, and more easily searchable. CPRs could also be the basis of longitudinal databases that would allow develop- ment of "best-clinical-practice" guidelines coupling treatment to expected or likely outcomes. Outcome-based treatment is thought by many to have considerable potential for reducing health care costs. A second growing application of IT in health care is its use in giving physicians remote access to hospital information systems through dial-up modems or leased lines, allowing them to check on the status of patients through CPRs. Physicians practicing in rural areas can thus use wide-band video to consult with specialists, who tend to reside in urban areas; one often-cited use is the transmission of x-ray images by network from rural family practitioners to urban radiologists. City-to-city teleconferencing can also be used so that the best specialists can consult together. Remote access provided by IT-based systems is also very important to emergency medical teams sent into the field. Ambulance teams or rescue squads do not typi- cally include physicians. However, radio communications can link on-site personnel with physicians elsewhere. Voice communications provide ad- vice and direction from the physician, and both voice and data downlinks from the field to the physician provide information about the condition of the patient. Automated systems to support clinical decision making have been the subjects of research since the mid-1960s. Given a set of data on a patient, these systems commonly act as a consultant that advises the physician on diagnosis or treatment. Such systems attempt to capture in machine-read- able form the expertise that clinical physicians have about their patients and

80 INFORMATION TECHNOLOGY IN THE SERVICE SOCIETY diseases. Two relatively old examples of medical "expert" systems are MYCIN, a system designed to help physicians manage different types of infection, and INTERNIST, a system designed to support diagnosis of dis- eases in internal medicine. The widespread use of IT in clinical practice will require IT that can be used to support the wide range of clinical functions for which the medical profession is responsible, taking into account the high degree of variability in clinical practice from practitioner to practitioner and the reluctance of physicians to alter their practices on the basis of outside influences. Never- theless, if physicians are willing to use data on outcomes and to adapt clinical practice to take the best advantage of IT's potential to reduce costs and improve efficiency, there are many opportunities to use information more effectively in providing health care. Future applications may include multimedia CPRs, explicit coupling of medical and related knowledge (e.g., the medical literature, government regulations regarding practice, or guide- lines for practice) to a patient's medical record and/or to a specific treat- ment program for the individual, and universal electronic interchange of data and information. BANKING As a major component of the U.S. financial system, the banking indus- try is vital to the U.S. and world economies and to individual financial well- being. Banks mediate a wide range of commercial and personal transac- tions involving the exchange and investment of money, providing such services as loans, deposits, and trust services to individuals, small businesses, and corporations. The banking industry is a major factor in the economy. It generated revenues of $209 billion (current dollars) in 1990. At the end of 1989, insured commercial banks alone held about $3.3 trillion in assets, $2.0 trillion in loans and leases, and $2.5 trillion in deposits. There were 12,713 banks in this category, down from a peak of about 14,481 in 1984;48 how- ever, the number of commercial banking offices (central and branch) grew from 57,010 in 1984 to 61,162 in 1989.49 The industry exhibits consider- able concentration. For example, in 1989, 4.8 percent of federally insured commercial banking institutions (614 commercial banks) held 74.9 percent of all assets ($2.47 trillion); 23.4 percent of these institutions held 88.9 percent of assets.50 Savings banks exhibit similar concentration, although not as intense; in 1989, 15.8 percent of federally insured savings banks (534) held 77.2 percent of all assets ($1.2 trillion).5~ Table 2.8 describes the evolution of depository institutions over the last decade, of which banks were by far the largest component, about 90 percent. Although the banking industry has grown at a moderate rate over the

IMPACTS OF INFORMATION TECHNOLOGY AT THE INDUSTRY LEVEL 81 last two decades, the most significant changes in this period concern its character rather than its size. For most of its history, banking has been subject to extensive federal and state regulation. However, partial bank deregulation in the late 1970s and early 1980s led to a sharp increase in the variety of services and products offered by commercial banks. Driven by both technology and competition from nonbank financial institutions, in- creasing product diversification continues today in commercial banking, although it is still constrained to some extent by current regulations. The banking industry has experienced a variety of economic and finan- cial problems over the past several years, including losses from bad real estate loans, a large fraction of portfolio assets tied up in borrowings for takeovers and for loans to less developed countries, the failure of a number of well-known financial institutions, the widespread implications of the sav- ings and loan crisis, and, finally, the economic downturn of 1990-1991.52 In large part, these problems are ultimately responsible for the considerable consolidation that the banking industry has undergone in the last decade. Bank failures, contractions, mergers, and acquisitions have reduced the number of all banks and the number of large banks, and as a result, industry em- ployment has been shrinking. Most reductions in personnel have been con- centrated in back-office jobs. Given the magnitude of the banking industry's investments in IT over the last two decades, large increases in productivity might have been ex- pected. One reason these have not appeared in measures of macroeconomic productivity is that such measures in the banking industry remain highly problematic. Figures on the gross domestic product originating in the U.S. economy (GPO) for the banking industry are derived by the Bureau of Economic Analysis essentially from employment data (Table B.3, Appendix B), whereas measures of gross output that are generated by the Bureau of Labor Statistics take into account commercial banks' three major services- deposits, loans, and trust services (Appendix C). Neither approach is able to account for improvements in the quality of service offered to customers or for the availability of a much wider array of banking services. For example, the speed with which the processing of a loan application is com- pleted is an indicator of service that is important to the applicant, as is the 24-hour availability through automated teller machines (ATMs) of many deposit and withdrawal services previously accessible only during bank hours. Neither of these services is captured as higher banking output at the macroeconomic level. A second reason for the lack of large increases in measured productiv- ity is that early applications of IT proved to be costly and cumbersome. Software and equipment had to be updated and replaced frequently. A dizzying array of new products constantly called for new software and com- munication capabilities. IT systems required large amounts of tailoring,

82 TABLE 2.8 Banking INFORMATION TECHNOLOGY IN THE SERVICE SOCIETY 1969 1980 1981 1982 1983 GPO (billions of 1982 dollars) Revenues (billions of 1982 dollars; SIC 60a) Employment (millions ofjobs) Number of hours worked (billions) Annual investment in IT (billions of 1982 dollars) IT capital stock (billions of 1982 dollars) 0.2 4.4 5.1 GPO per labor-hour 39.6 61.9 56.4 90.5 n/a 2.0 n/a 3.7 0.1 1.6 64.2 88.7 2.0 3.9 65.2 66.1 96.4 100.9 2.1 3.9 1.9 1.8 5.6 2.1 3.9 2.9 6.9  (1982 dollars) 16.67 16.66 16.77 16.77 Annual ITtGPO 2.6% 3.0% 2.8% 4.4% Annual rate of change in GPO per labor-hourb 0.0% 0.6% 0.0% NOTE: Figures include all depository institutions. Commercial banks, Federal Reserve banks, and savings and loan banks account for about 90 percent of all figures. See Box 2.1 for explanation of terms and sources of data given. training, upgrading, and updating. Cost control, management skills, and productivity tracking systems lagged behind the new technologies in a rap- idly changing competitive marketplace. Margins long protected by regula- tions were squeezed by inverted yield structures (i.e., financial instruments whose yield varied inversely with their time to maturity), changing markets, and cross-competition with other savings and investment instruments. Risks and opportunities for profit both increased. The result was that tangible paybacks from IT investment were delayed. Consolidations in the industry have had the desirable effect of reducing excess IT capacity, particularly that associated with back-office processing functions, and also excessive branch network structures.53 IT has been an enabler of consolidation. However, studies of smaller mergers in the mid- 1980s suggest that the noninterest expenses of merged banks do not differ significantly from the expenses of nonmerged banks.S4 Other studies sug- gest that economies of scale characterize some but not all activities under- taken by banks.55 Today, IT is being used to support growing efficiency and greater dis- persion among banks. Back-office IT systems for internal accounting and

IMPACTS OF INFORMATION TECHNOLOGY AT THE INDUSTRY LEVEL 83 TABLE 2.8 Continued 1984 1985 1986 1987 1988 1989 GPO (billions of 1982 dollars) 67.2 Revenues (billions of 1982 dollars; SIC 60a) Employment (millions ofjobs) Number of hours worked (billions) Annual investment in IT (billions of 1982 dollars) IT capital stock (billions of 1982 dollars) GPO per labor-hour (1982 dollars) Annual IT/GPO Annual rate of change in GPO per labor-hourb 68.5 70.3 99.9 100.3 2.2 4.04.1 5.9 9.112.5 16.60 6.0% - 1.0%-0.2% 2.2 4.2 4.3 7.9 16.9 16.58 16.62 8.7% 11.2% 0.2% 71.1 70.8 71.4  100.3 99.5 99.8 2.3 2.3 9.8 22.2 4.2 4.3 12.6 13.8 28.8 35.1 16.70 16.77 16.80 13.8% 17.8% 19.3% 0.5% 0.4% 0.2% aStandard Industrial Classification system code. bThe compounded annual rate of change in GPO per labor-hour from 1981 to 1989 was 0.1 percent. Labor productivity is sometimes measured in terms of GPO per labor-hour. customer accounting were the focus of IT investment throughout the 1970s and early 1980s. Virtually all banks were interconnected to regional and national networks. A large percentage of paper-based transactions were transmitted and processed electronically. ATM services and direct elec- tronic deposits and withdrawals by large automated users replaced many paper processes. As new products and services expanded, and as margins became less predictable, commercial banks began investing in front-office automation to provide contact people with better information and to en- hance the delivery of products and services. 1 , Initially, investment in IT was concentrated in computers, since bank- ina involved input-output intensive processing and huge databases. Most bank work involves establishing accounts (or customer relationships) and handling transactions within those accounts; banks process an increasing number of transactions that now total many tens of billions per year, a number that would be unthinkable without automation.56 For example, Visa's peak capacity for processing credit card transactions grew from 30,000 per day in 1978 to over 1.4 million per day in 1991, while its response time for authorizations dropped from 5 minutes in 1973 to 1.1 seconds in 1991.57

84 INFORMATION TECHNOLOGY IN THE SERVICE SOCIETY Banks developed the technology for magnetic ink character recognition, which enables a high degree of automation in the check-processing system. Stand-alone systems for check processing, credit applications, and savings accounts are used by a large number of banks today. More recently, new IT-based systems have helped to consolidate mainframe-based operations. For example, the Bank of Boston revamped its securities-processing busi- ness in 1989, consolidating these operations from 11 separate locations into 2. This IT-based consolidation reduced staff requirements by 17 percent while volume increased by 80 percent. In the late 1980s, the banking industry began to focus on automation of data communications. The installation of on-line terminals in the early 1970s enabled automation of the customer interface and front-office appli- cations in such areas as corporate treasury. ATMs, first introduced in the late 1970s, have become an agent of a strategic change in banking. More than 75,000 ATM s in the United States handle about 6 billion transactions per year for cash dispensing, funds transfer, credit card payments, and ob- taining credit lines.59 The deployment of an ATM network in 1977 was a major contributor to Citibank's growth of market share from 4 percent to 13.4 percent; between 1977 and 1988 Citibank's branch staff grew by 18 percent while they served three times as many customers.60) Data commu- nications also lies at the heart of bank-to-bank transactions all over the world: the Clearing House for Interbank Payments System handles daily transactions worth nearly $2 trillion.6i IT enables banks to provide better service to customers. For example, Mellon Bank has installed an IT-based system to help resolve credit card billing disputes; this system enables the clerk to view in one place all documents relevant to a particular disputed transaction. Using this system, which also involved considerable reengineering of its process for handling customer complaints, Mellon was able to reduce its backlog of customer complaints from 5200 to 2200, resolving them in 25 days on average (ver- sus 45 days previously).62 IT has been used to support a plethora of new products, including ATM access to banking services, the ability to shift funds between multiple ac- counts, and overdraft privileges.63 Many of these new products are made possible by "securitization" of mortgages and credit card portfolios. Bank products are arguably information combined in new ways; for example, variable-rate consumer loans depend on computer-generated indexes and interest-calculating algorithms,64 while unified statement processing enables the display of information from all accounts at one bank on one statement and "sweep" services enable customers to consolidate dividends and inter- est from different investment funds in one central interest-paying money market account.

IMPACTS OF INFORMATION TECHNOLOGY AT THE INDUSTRY LEVEL 85 Changed by sophisticated IT-based systems, loan processing shifted from personal loan officers accessing branch cash reserves to automated access to credit databases and external funds acquisition with asset-backed security arrangements.65 Trade processors have eliminated the gap between front and back offices through automated interfaces.66 Automated credit scoring simulates the organizational structure of centralized approval; instead of branch personnel sending applications to a central site for approval, work- stations connect lenders to centralized computer systems that are used to help make decisions about granting credit.67 Dan Schutzer, a vice president of Citibank, asserts that IT has been used to create "whole new industries in investment banking that never ex- isted before with the growth of derivative products, mortgage-backed secu- rities, credit card receivables, commercial paper, corporate bonds to a large extent, and so forth. Many of these things didn't even exist as sources of revenue the way they are now." The banking industry has invested heavily in building a sophisticated information infrastructure, which now includes high-powered desktop sys- tems, artificial intelligence applications, automated credit scoring, electron- ics funds transfer, and links among a wide range of services and providers. IT will be used to enable an increasingly diverse set of banking interactions. More than likely, a higher proportion of banking will be carried out via ATMs, or more broadly over wires via telephone or screen-based transac- tions as well as by mail. Thus the interaction will be more indirect. In the future, there may be more card-based interactions, as the uses of credit cards grow and as cards themselves become "smarter" (e.g., smart card- telephone services). An increasing amount can be done at little extra cost, with relatively few people, a circumstance that promotes diversification into generation of assets (e.g., mutual funds), insurance products, and so on- building blocks for the industry. There will also be a high premium on integration of products and services that is, standardization up to some level followed by specialization to the customer. Lastly, the banking industry has come under considerable competition from "nonbank banks." While the relaxation of the regulatory environment in the last decade has been important, IT has enabled a variety of nonbanking institutions to take advantage of this new environment by offering services that previously only banks could offer. For example, supermarkets with ATMs are now surrogate bank branch offices for transactions such as cash withdrawal. Large corporate entities such as AT&T and General Motors now offer widely accepted credit cards similar to Mastercard and Visa. With the cooperation of local banks, investment firms offer mortgages at favorable rates across the nation. Thus the traditional boundaries of the banking industry are blurring, and the ultimate impact of new competition is uncertain.

86 INFORMATION TECHNOLOGY IN THE SERVICE SOCIETY INSURANCE The insurance industry is divided into two large segments: (1) life and health insurance, including pensions and annuities, and (2) property and casu- alty insurance. Life and health insurance is divided among individual prod- ucts (sold to individual consumers) and the group market (life, health, and pension insurance products sold to employee or other groups).68 Property and casualty insurance products are divided between personal lines (individual products, mainly automobile and homeowners' insurance) and commercial lines (e.g., coverage for liability, errors and omissions, business continuation, and employee fidelity bonding, as well as fire, theft, and so on). Total premium receipts for life and health insurance and annuities ap- proach $300 billion annually. Life and health insurance is provided by about 2350 companies;69 total life insurance in force (i.e., aggregate policy face value) exceeds $10 trillion, with the average U.S. household now hav- ing more than $100,000 in coverage. Net written property and casualty premiums total about $225 billion annually,70 with about 3800 companies providing such coverage.7i About 55 percent of this figure covers commer- cial liabilities, with the remainder covering personal liability. The scale of these dollar flows makes insurance companies a major element of the U.S. financial system and major institutional investors. The several thousand companies that provide insurance coverage range from very large multiline companies that sell a broad range of products in all states, to small companies that sell particular products to fill specialized customer or product niches. Policies are sold to the public through direct writers that sell through their own agents (often known as captive agents) or directly through telemarketing, the mail, and so on, and through agency companies that sell through independent agents who offer products from several insurance companies. In 1989 about 1.4 million people worked for insurance carriers (Table 2.9~. Expenditures on IT in the insurance industry have grown significantly more rapidly than those of most of the other ser- vice industries examined. Table 2.9 describes the evolution of the industry in more detail. In recent years, major changes have taken place in the insurance indus- try. Life insurance carriers have relied less on traditional whole-life poli- cies, having been pushed by market forces to offer products that compete with a wide range of other investment vehicles. In the very competitive market for pensions and annuities, life insurance companies offer a wide range of investment vehicles to individuals and corporate pension plans. In the health insurance market, traditional indemnity-based group health plans now are in a minority, as insurers more often offer and serve as administra- tors of managed health care plans, such as health maintenance organizations and preferred provider organizations. Insurance companies are facing in

IMPACTS OF INFORMATION TECHNOLOGY AT THE INDUSTRY LEVEL 87 creasing competition from a range of entities, including banks and other financial institutions, health insurance plan administrators, and corporations' increasing willingness to self-insure. Understanding the impact of these changes on productivity in the insur- ance industry is highly problematic, and the BLS has been unable to de- velop a meaningful measure of output for the insurance industry. Still, certain structural features of the industry can be identified. Much of the insurance industry, especially the property and casualty and health segments, is data-intensive, being characterized by huge volumes of accounting, analysis and statistical data, and transaction processing. In- surance companies are concerned about the quality of business generated by their agents, as measured, for example, by the level of loss payouts result- ing from the mix of products offered by a company. These companies need to be able to assess the risk to which a given portfolio exposes them, and thus they are large users of actuarial data that enable them to calculate statistically the likelihood of these risks with adequate levels of predictabil- ity. In addition, a large volume of data persists throughout an insurance product's life cycle, which includes underwriting (in which a substantial number of risk factors are evaluated in determining the acceptability of a risk and the premiums to be charged), maintaining policyholder and policy and coverage information, and keeping records of claims and payments. Thus, it is not surprising that insurers were among the first in the com- mercial and corporate community to invest heavily in computers of all sizes, to develop their own extensive software and to use packaged software, and to install major telecommunications networks. For example, the insurance industry was quick to automate claims processing, which involves verifying that a claim has been submitted by a current policy holder, validating the consistency between the particular claim made and the policy's coverage (perhaps checking other policies held to coordinate benefits), and generat- ing a check for the appropriate amount along with an explanation for how that amount was derived. All of these functions are performed today in a highly automated fashion. Automation has enabled the volume of claims processed to grow enormously with a minimal increase in personnel re- quirements, clearly an increase in productivity. As importantly, such sys- tems enable claims processing to be carried out more quickly on a decen- tralized basis by employees (or agents) in the field who can access the necessary databases remotely. Beyond reducing labor requirements and increasing processing speed in back-room operations such as accounting and records, insurance companies are using IT to do qualitatively different things. For example, portable computers are being used increasingly in field operations to communicate with the home office for a variety of functions. An agent tied by modem to the home office can serve a customer much more effectively now, by pro

88 TABLE 2.9 Insurance INFORMATION TECHNOLOGY IN THE SERVICE SOCIETY 1969 1980 1981 1982 1983 GPO (billions of 1982 dollars) Revenues (billions of 1982 dollars; SIC 63a) Employment (millions of jobs) Number of hours worked (billions) Annual investment in IT (billions of 1982 dollars) IT capital stock (billions of 1982 dollars) GPO per labor-hour (1982 dollars) Annual IT/GPO Annual rate of change in GPO per labor-hourb 20.2 38.7 44.1 1.0 1.9 0.0 0.1 35.4 67.8 66.8 29.8 67.8 33.2 67.3  1.2 1.2 1.2 1.2 2.4 0.6 1.3 10.54 16.21 0.2% 1.6% 4.0% 2.4 0.7 1.7 14.76 2.0% -9.0% NOTE: See Box 2.1 for explanation of terms and sources of data given. 2.4 0.8 2.0 2.4 1.8 3.1 12.4513.97 2.5%5.3% -15.6% 12.2% viding customized quotes or claims payments instantaneously, than in the past. Insurance companies have connected their own agents and indepen- dents through networks they have designed and built themselves, or through  systems such as the Insurance Value Added Network System (IVANS) de- veloped by IBM. Development of insurance policies for different corporate clients has also benefited from the use of IT. For example, the Aetna Life and Casu- alty Company used IT to revamp its "case installation" process for develop- ing health plans for new clients.72 The networked PC-based system now used by Aetna's Health Plan business unit for case installation was based on reengineering Aetna's installation process to homogenize case installations across regional offices and to reduce the volume of paperwork associated with each new installation and its different requirements. Knowledge-based expert systems are being used increasingly by under- writers to help decide acceptance of applicants for coverage. Such systems facilitate the rapid comparison of a particular individual's risk profile with those in the community at large and make it easier to customize a policy based on the risks faced by that individual. They also provide the capabil- ity, based on quite sophisticated and elaborate data analysis developed by

IMPACTS OF INFORMATION TECHNOLOGYAT THE INDUSTRY LEVEL 89 TABLE 2.9 Continued 1984 1985 1986 1987 1988 1989 GPO (billions of 1982 dollars) 35.9 39.5 36.0 34.4 37.0 36.7 Revenues (billions of 1982 dollars; SIC 63a) 70.2 74.2 73.7 73.5 76.9 77.5 Employment (millions of jobs) 1.2 1.3 1.4 1.4 1.4 1.4 Number of hours worked (billions) 2.4 2.5 2.7 2.7 2.8 2.8 Annual investment in IT (billions of 1982 dollars) 3.3 3.6 4.6 5.6 6.4 6.2 IT capital stock (billions of 1982 dollars) 5.4 7.5 10.0 12.9 15.8 17.9 GPO per labor-hour (1982 dollars) 14.93 15.72 13.52 12.53 13.32 13.18 Annual IT/GPO 9.2% 9.1% 12.8% 16.4% 17.3% 16.8% Annual rate of change in GPO per labor-hourb 6.9% 5.3% -14.0% -7.4% 6.3% -1.1% aStandard Industrial Classification system code. bThe compounded annual rate of change in GPO per labor-hour from 1981 to 1989 was -1.4 percent. Labor productivity is sometimes measured in terms of GPO per labor-hour. the insurance industry, to implement widespread and varied cost-contain- ment techniques in paying group health claims. Quality in the insurance industry can be reflected in the extent to which insurance products can be tailored to meet the needs of individuals and in the manner in which customers are served. As is the case with many other service-providing activities, measuring the quality of the service delivered in the insurance industry is problematic. Although in some cases quality does have quantitative indicators (e.g., the average number of days needed to issue a policy or settle a claim), often the value of quality is subjective; i.e., it is in the minds of customers. Today, many companies are using customer surveys and focus group sessions to find out what is important to their customers. Many insurers are focusing their initial reengineering efforts in areas that will affect customer service, seeking ways to further differentiate themselves in what is being viewed increasingly by many customers as a commodity market. The ulti- mate goal is to sell insurance to a market of one single clients to whose individual circumstances a policy's payments and benefits are specifically tailored. The way each company uses IT will help determine whether and how it meets this goal.

9o INFORMATION TECHNOLOGY IN THE SERVICE SOCIETY Selling customized insurance to markets of one may change the tradi- tional notion of insurance as the sharing of risk among large groups of individuals. Insurance companies have always differentiated among some risk groups (e.g., male drivers under the age of 25 usually pay more than others for auto insurance). But prior to the widespread availability of IT, data to differentiate more finely among individuals were simply unavail- able. Today, the use of IT enables insurance companies to collect data on individuals from multiple sources and thereby define risk groups consist- ing of relatively small numbers of individuals, individuals with AIDS or cancer. Insurance may not be available to such groups at all, or only at a very high price a possibility that poses significant questions for social policy. IT has also been used effectively by insurance companies attempting to compete in new markets. For example, James Stewart, executive vice presi- dent and chief financial officer, noted that CIGNA invested in IT "to sig- nificantly improve and enhance the ability to deal with an individual participant' s statements and a whole bunch of other services that we felt were respon- sive, both to the competition and to the customer's needs. That decision was made, not on a cost-benefit basis, but rather [to answer such questions asi, What markets do we want to compete in? What do we need to do to position ourselves against both the insurance company and non-insurance company competition?" Insurance agencies that provide coverage to individuals have followed a similar path with respect to using IT. A recent survey of independent agents found that 85 percent of independent agencies have some automation and that a growing number are using integrated, comprehensive full-func- tion agency systems for back-office work (e.g., commission analyses, ac- counts receivable and payable). Regardless of whether they have agency systems, 90 percent have word processing, 85 percent have automated in- surance rating, 76 percent have automated accounting, and 72 percent have automated billing. Somewhat fewer have automated underwriting (33 per- cent) and automated interfaces to multiple companies (44 percent, up from 23 percent in 1987~. Seventy-two percent of these systems were installed before 1989.73 Agency systems, like corporate systems, have been used largely to au- tomate what were previously manual processes (accounting, billing, market- ing, and customer service). Although such systems enable the same number of people to handle a larger volume of work (and thus hold down labor costs), independent agents' desires and expectations for agency automation are, according to some observers, largely unfilled. Early agency users of IT experienced more difficulties than they had expected, partly as a result of their failure to understand the new systems, plan appropriately for their use, and train employees accordingly.

IMPACTS OF INFORMATION TECHNOLOGY AT THE INDUSTRY LEVEL 9 J. Raymond Caron of CIGNA Systems argues that There is a big evolution of IS (information systems) in an insurance envi- ronment. We went through an era with massive tab card sorting. We left that era and went into mainframe computers, data centers, and batch pro- cessing. We left that era and went into on-line applications, networks, and all that stuff. I think we're on the verge of a significant change in what IS organizations are all about, and I think that change is going to be radical. Reengineering, how we build applications, and how we work on solutions will be the new drivers. CIGNA's systems organization is going through the process of reengi- neering along with every other division within the company. What we are striving for in our reengineered systems organization is to be an integrated part of the business strategy. To accomplish this, we have had to help our people change their mind-set from a narrow technology perspective to a more global business perspective. Today, CIGNA systems professionals use their skills to solve business problems, not technology problems. And they understand that their performance will now be measured by customer satisfaction and business results. One of the things I'm planning is changing the way IS people think. If you get them to think about it right they can provide solutions that have nothing to do with technology. We need to get them to think that way because those might be the most powerful solutions rather than thinking the only thing IS can bring is technology. So we're restructuring IS, and I think the difference will be dramatic. Put differently, IT may be used to drive restructurings that might not happen in the absence of technology. Developing better ways to use IT is already an integral part of many insurance companies' strategic planning and is a key factor in the industry's growth and in the successes of indi- vidual companies. OBSERVATIONS AND CONCLUSIONS In both scale and nature, the use of IT is highly varied within and among the industries the committee examined. That is, IT has been used by companies for many different purposes, both operational (e.g., accounting, logistics scheduling) and strategic (e.g., changing competitive positioning or providing totally new services). Although IT has been used primarily to improve the efficiency and effectiveness of specific activities, the impact of IT in these individual instances has sometimes had a large influence on the structure of the industry as a whole. Because of the effects of other non-IT forces, the significance of IT's use for an industry may be observable in aggregate industry data only when its impact is truly overwhelming or revo- lutionary. A good example is the telecommunications industry, which liter- ally has been built on IT and which is now pervaded by IT. In other

92 INFORMATION TECHNOLOGY IN THE SERVICE SOCIETY instances that are more likely the norm, IT's impact may be measurable only at the level of individual enterprises (e.g., in shifts of share or profit) or in activities within various companies. At the same time, it is clear that the use of IT can bring about changes in the boundaries and definition of an industry. For example, deregulation and new technologies have allowed new players to provide services that were once the exclusive domain of traditional telephone carriers; today, private carriers and value-added carriers handle voice and data long-dis- tance traffic for paying customers, and even some corporate networks have spare capacity that can be sold to other outside users. IT has been used to alter existing relationships among wholesalers, retailers, manufacturers, and end users (e.g., to eliminate intermediaries and enable direct sales from manufacturers to end users). Insurance companies are moving ever deeper into the business of health care management, and the lines between the two are already blurring. Through the use of IT, different industries are moving into banking, thus threatening the traditional understanding of what consti- tutes the banking system. IT (in the form of CRSs) has been part of what one airline executive suggested has been the transformation of the airline industry into one of the largest retailers of perishable goods (i.e., empty seats that disappear when the flight takes off). Further examples and ampli- fications of this point are contained in Chapter 3. Lastly, IT may well create entirely new sectors for business that will encompass a variety of new and different industries. For example, a power- ful telecommunications and information infrastructure has the potential to open the private home to a host of new information-related services. A1- though most network-based services have been aimed at businesses and other organizations, many now envision the creation of large home-based markets for new products combining computing and communications de- vices and services (e.g., multimedia technology). Although past attempts to develop services for the home such as videotext information delivery or computer-based teleshopping were largely unsuccessful, greater potential is seen in the more sophisticated offerings promised by recent and announced efforts to blend telephone, cable television, software, video, audio, text, and image database services. Better services for the home imply changes in the existing communications, information services, and entertainment indus- tries; new markets for services; and new possibilities for home-based work. Consistent with the benefits-pass-through phenomenon discussed in Chapter 5, these new markets will likely generate public or private financial gains that far exceed those realized by the first companies entering these new markets. Although an industry perspective offers particular insights about the actual performance of IT-based systems versus IT's measurable effects on productivity, further analysis at the enterprise and activity levels is neces

IMPACTS OF INFORMATION TECHNOLOGY AT THE INDUSTRY LEVEL 93 sary to understand where and how deployment of IT can be most effective and how IT's use is changing the basic structure of the economy, individual organizations, and most jobs. NOTES AND REFERENCES sin 1989, the United States employed 116.6 million workers in full-time and part-time jobs. See U.S. Department of Commerce, Bureau of Economic Analysis. 1990. Survey of Current Business, July, Table 6.6B, Government Printing Office, Washington, D.C. 2U.S. Department of Commerce, Bureau of Economic Analysis. 1992. Survey of Current Business, July, Government Printing Office, Washington, D.C. Table 6.22C (p. 91) provides corporate capital consumption allowances by industry, Table 6.1C (p. 82) provides national income by industry, and Table 6.5C (p. 84) gives full-time-equivalent employees by industry. 3U.S. Department of Commerce. 1992. U.S. Industrial Outlook, Government Printing Office, Washington, D.C., p. 40-1. 4The revenue passenger-mile (RPM) is the principal measure of output for the air trans- port industry: a single paying passenger carried for a 1000-mile airplane flight counts as 1000 revenue passenger-miles. The RPM is thus a direct measure of the service being delivered: the movement of passengers between different points. 5U.S. Department of Commerce, Bureau of the Census. 1991. Table 1072, p. 627, Government Printing Office, Washington, D.C. U.S. Statistical Abstract, 6Scheduled passenger carriers suffered losses from 1981 to 1983, in large part due to the recession and high fuel prices. From 1984 to 1989, the industry earned positive net income in four of the six years and reported a minimal loss in another. However, the industry's financial losses were so severe in 1990 (an operating loss of $3.95 billion in 1990-1991) that any gains in the 1980s were wiped out; scheduled passenger carriers lost about $2 billion from 1981 to 1990. (U.S. Department of Commerce, 1992, U.S. Industrial Outlook, p. 40-2.) 7U.S. Department of Commerce, 1992, U.S. Industrial Outlook, p. 40-1. 8Recent trends suggest that hub-and-spoke routing systems are most profitable under conditions of relatively high demand. Lower demand reduces the revenues that airlines can generate, while the costs of maintaining a hub are considerably less elastic. As recessionary pressures have driven down the demand for air travel, many airlines are starting to scale back the number of hubs they maintain. See Hirsch, James. 1993. "Big Airlines Scale Back Hub- Airport System to Curb Rising Costs," Wall Street Journal, January 12, p. Al. 9See Margolis, Nell. 1992. "AMR Cuts Costs, Jobs," Computerworld, December 21, p. 2. 10Transportation Research Board, National Research Council. 1991. Winds of Change: Domestic Air Transport Since Deregulation, Special Report 230, National Academy Press, Washington, D.C., p. 54. U.S. Department of Commerce, 1992, U.S. Industrial Outlook, p. 40-4. 12For example, this revenue stream is worth $100 million per year to American Airlines (Quinn, James Brian. 1988. "Technology in Services: Past Myths and Future Challenges," p. 37 in Technology in Services: Policies for Growth, Trade, and Employment, Bruce R. Guile and James Brian Quinn (eds.), National Academy Press, Washington, D.C.). 13The CRS is conceptually a single database, although in fact the data it contains may be physically distributed over many different computer systems. Technologically, the CRS pro- vides a single standard interface to the human user (e.g., the travel agent) for access to mul-  tiple heterogeneous databases. 14The term yield management refers to the practice of continuously managing the number and fare class of seats on a given flight. By combining knowledge of how quickly seats of a given fare are selling with knowledge of competitors' pricing, yield management minimizes the number of empty or low-paying seats on any given flight, thus increasing revenues.

94 INFORMATION TECHNOLOGY IN THE SERVICE SOCIETY 15See Tolchin, Martin. 1992. "U.S. Sues 8 Airlines Over Fares," New York Times, December 22, p. D1. Air Transport World. 1991. "Ryder Automates Parts Location," September, p. 106. Air Transport World. 1991. "Recipes by VideoSpecs," August, pp. 78-81. 18The A320 Airbus is a case in point. See Stix, Gary. 1991. "Along for the Ride," Scientific American, July, p. 97. 19Air Transport World. 1991. "ICAO Looks to Space," December, pp. 96-98. 20Science. 1989. "Flying the Electric Skies," Vol. 244(June 30):1532-1534. 21Aviation Week and Space Technology. 1991. "Airbus Employs Advanced Procedures in A330/A340 Manufacturing, Assembly," November 4, p. 48. 22Air Transport World. 1991. "Command Center on the Prairie," May, pp. 38-39. 23Kindel, Sharen. 1991. "The Route Cause," Financial World, August 20, pp. 70-72. 24Larson, Richard. 1988. "Operations Research and the Services Industries," in Managing Innovation: Cases from the Services Industries, Bruce R. Guile and James Brian Quinn (eds.), National Academy Press, Washington, D.C. 25From 1990 to 1992, the U.S. airline industry lost $9 billion to $10 billion dollars, wiping out all that it had earned, and more, in the 1980s. See Lenorovitz, Jeffrey. 1993. "Panel Issues Guidelines for Airline Recovery," Aviation Week and Space Technology, August 9, p. 31. 26See O'Leary, Meghan, 1991, "Airport '91," CIO, June, pp. 40-52; Air Transport World, 1991, "Ionworks," May, pp. 62-66; and Air Transport World, 1991, "Automation Picks Up Speed," September, pp. 51-56. 27Air Transport World. 1992. "New Cabins in the Sky," April, pp. 75-80. 28U.S. Telephone Association. 1992. Statistics of the Local Exchange Carriers: 1992 for the Year 1991, U.S. Telephone Association, Washington, D.C., p. 2. 29Federal Communications Commission. 1991. "FCC Releases Report on Long-Distance Market," news release, October 2, FCC, Washington, D.C. 30Schatz, Willie. 1992. "Who's Calling, Please?" Computerworld, November 23, p. 69. 31U.S. Department of Commerce, 1992, U.S. Industrial Outlook, p. 28-3. 32In some countries, long-distance calls are still logged by operators with hand-written tickets. 33According to Peter Chinloy, labor costs constituted approximately 75 percent of the operating costs for the Bell system in the early 1980s. (See Chinloy, Peter. 1981. "Implica- tions for the Communications Workers of America of Technological Change at the Bell Sys- tem," reproduced in Productivity in the American Economy, 1982, Hearings before the Sub- committee on Employment and Productivity of the Committee on Labor and Human Resources, U.S. Senate, March 19 and 26 and April 2 and 16, Government Printing Office, Washington, D.C.) This study of the Bell system from 1947 to 1978 also found that labor input grew more slowly than output, resulting in a steady increase in labor productivity. However, note that "labor" costs refer only to the costs incurred by telecommunications service providers; cus- tomer-provided labor (e.g., the keying by customers of telephone and credit card numbers) is not included in such costs. In other words, "reductions" in the labor required are at times more accurately characterized as shifting labor costs from provider to customer. Customers can also benefit from such shifts, as they can complete their calls more quickly, more securely, and less expensively by keying numbers themselves. 34Calling cards were invented in the 1950s but became especially popular after 1982, when AT&T automated the calling and billing processes, eliminating the need for operator  intervention. See Ramirez, Anthony. 1992. "Plastic Keys to Phone Wizardry," New York Times, May 30, p. 48. 35See Wall Street Journal. 1992. "AT&T Unveils Card for ATMs That Has Wider Range of Uses," December 8, p. A6. 36Today, with a few exceptions (notably cable TV companies), telecommunications com- panies provide their customers with the ability to move and manage data and information that

IMPACTS OF INFORMATION TECHNOLOGY AT THE INDUSTRY LEVEL 95 the customer generates and/or owns. Local telephone companies, especially the regional Bell holding companies, are actively interested in providing information content as well as delivery. Legislative action may be taken in the future to remove the constraints placed by the 1982 consent decree on providing such services. 37Personal communication, Joseph Timko, vice president, AT&T Architecture, June 1992. 38See Business Week. 1992. "Clout! More and More, Retail Giants Rule the Market- place," December 21, pp. 66-73. 39For example, small manufacturers often have limited financial resources that prevent them from creating a direct-selling organization. To the retailer the wholesaler offers multiple lines of merchandise and a larger assortment than is available from an individual manufacturer; wholesalers also obviate the need to work with multiple manufacturers. The wholesaler has the ability to buy in huge bulk portions that can be broken down into smaller units for the retailer. Some manufacturers are engaged in the wholesale business through the use of their sales branches. Retailers may also choose to buy directly from the manufacturer through chain-store warehouses; however, this activity does not fall within the wholesale classification. 40Inventory can represent as much as 70 percent of a retailer's assets. See, for example, Standard & Poor's Corp., 1992, "Retailing: Current Analysis," Industry Surveys 160(2), sec. 1, January 9; and Standard & Poor's Corp., 1991, "Retailing: Basic Analysis," Industry Surveys 159(17), sec. 1, May 2. 41National Retail Federation, Financial Executives Division. 1991. Merchandising & Operating Results of Retail Stores in 1990, National Retail Federation, New York. 42Business Week. 1992. "Clout! More and More, Retail Giants Rule the Marketplace," December 21, pp. 66-73. 43See Elliot, Stuart. 1993. "A Last Hurdle for Shoppers: The Checkout-Counter Pitch," New York Times, January 11, p. D-7. 44Bencin, Richard L., and Donald J. Jonovic. Prentice Hall, Englewood Cliffs, New Jersey, p. 276. An 1989. Encyclopedia of Telemarketing, 4)U.S. Department of Commerce, U.S. Statistical Abstract, 1991, Table 162, p. 104, for ambulatory care and Table 173, p. 109, for hospital usage. 46But see also explanatory note in Table 2.7. 47Tierney, W.M., M.E. Miller, J.M. Overhage, and C.J. McDonald. 1993. "Physician Inpatient Order Writing on Microcomputer Workstations: Effects on Resource Utilization," JAMA 269(January 20):379-383. 48U.S. Department of Commerce, 1991, U.S. Statistical Abstract, Table 807, p. 501. 49U.S. Department of Commerce, 1991, U.S. Statistical Abstract, Table 804, p. 500. 50U.S. Department of Commerce, 1991, U.S. Statistical Abstract, Table 802, p. 499. 51U.S. Department of Commerce, 1991, U.S. Statistical Abstract, Table 802, p. 499. 52U.S. Department of Commerce, 1992, U.S. Industrial Outlook, p. 46-1. 53Humphrey, David B. 1991. "Productivity in Banking and Effects from Deregulation," Federal Reserve Bank of Richmond, Economic Review 77(March/April):16-28. 54Srinivasan, Aruna. 1992. "Are There Cost Savings from Bank Mergers?," Federal Reserve Bank of Atlanta, Economic Review 77(March/April):17-28. 55Andersen Consulting. 1991. Strategies for High Per~formance, Arthur Andersen, Chicago. 56Steiner, Thomas D., and Diogo B. Teixeira. 1990. Technology in Banking: Creating  Value and Destroying Profits, Dow Jones-Irwin, Homewood, Illinois, p. 29. 57Ganley, Oswald, and Gladys Ganley. 1992. "To Inform or to Control?," The New Communications Networks, Second Edition, Ablex Publishing, Norwood, New Jersey, p. 223. Also personal communication with Visa, 1992, cited in "Rewards and Risks: Communications and Information in the Global Financial Services Industries," PIRP, Harvard University, Cam- bridge, Mass., June 1992. 58Stewart, Thomas. 1992. "U.S. Productivity: First But Fading," Fortune, October 19, p. 56.

96 INFORMATION TECHNOLOGY IN THE SERVICE SOCIETY 59Bove, Richard X. 1991. "Bank Technology Reshapes Industry," The Bankers Maga- zine, May/June, pp. 17-20. 60Glaser, Paul. 1988. "Using Technology for Competitive Advantage: The ATM Expe- rience at Citicorp," in Managing Innovation: Cases from the Services Industries, Bruce R. Guile and James Brian Quinn (eds.), National Academy Press, Washington, D.C. 61Passell, Peter. 1992. "Fast Money," New York Times Sunday Magazine, October 18, p. 42. 62Fortune. 1992. "How to Steal the Best Ideas Around," October 19, p. 106. 63Steiner and Teixeira, 1990, Technology in Banking: Creating Value and Destroying Profits, p. 23. 64Steiner and Teixeira, 1990, Technology in Banking: Creating Value and Destroying Profits, p. 49. 65Bove, 1991, "Bank Technology Reshapes Industry." 66Andersen Consulting, 1991, Strategies for High Performance, p. ix. 67Andersen Consulting, 1991, Strategies for High Performance, p. 5. 68There are two large subgroups within the group arena: Medicare (health insurance for the elderly) and Medicaid (health insurance for the poor) are government-sponsored programs that are administered by companies under contract to states. Blue Cross and Blue Shield organizations are also state organized but privately administered. Over three-quarters of life insurers' health premiums come from the group health market (usually through employer plans). See U.S. Department of Commerce. 1992. U.S. Industrial Outlook, p. 50-2. 69U.S. Department of Commerce, 1991, U.S. Statistical Abstract, Table 851, p. 518. 70The Department of Commerce estimates life and health insurance premium receipts in 1991 at $283.5 billion and net written premium receipts for property and casualty insurance at $224.9 billion (U.S. Department of Commerce, 1992, U.S. Industrial Outlook pp. 50-1 and 50-6). 71U.S. Department of Commerce, 1992, U.S. Industrial Outlook, p. 50-7. 72Cafasso, Rosemary. 1993. "Manual No More: Aetna Unit Gets PCs, Results," Computerworld, January 18, p. 37. 73Acord Corporation. 1992. Moving Into the 90s: The 1992 Acord Agency Automation  and Interface Survey, Acord [Agency Company Organization for Research and Development], White Plains, New York.