National Academies Press: OpenBook

Technology in Services: Policies for Growth, Trade, and Employment (1988)

Chapter: Measuring Productivity in Services Industries

« Previous: Technology and the Services Sector: America's Hidden Competitive Challenge
Suggested Citation:"Measuring Productivity in Services Industries." National Academy of Engineering. 1988. Technology in Services: Policies for Growth, Trade, and Employment. Washington, DC: The National Academies Press. doi: 10.17226/764.
×
Page 139
Suggested Citation:"Measuring Productivity in Services Industries." National Academy of Engineering. 1988. Technology in Services: Policies for Growth, Trade, and Employment. Washington, DC: The National Academies Press. doi: 10.17226/764.
×
Page 140
Suggested Citation:"Measuring Productivity in Services Industries." National Academy of Engineering. 1988. Technology in Services: Policies for Growth, Trade, and Employment. Washington, DC: The National Academies Press. doi: 10.17226/764.
×
Page 141
Suggested Citation:"Measuring Productivity in Services Industries." National Academy of Engineering. 1988. Technology in Services: Policies for Growth, Trade, and Employment. Washington, DC: The National Academies Press. doi: 10.17226/764.
×
Page 142
Suggested Citation:"Measuring Productivity in Services Industries." National Academy of Engineering. 1988. Technology in Services: Policies for Growth, Trade, and Employment. Washington, DC: The National Academies Press. doi: 10.17226/764.
×
Page 143
Suggested Citation:"Measuring Productivity in Services Industries." National Academy of Engineering. 1988. Technology in Services: Policies for Growth, Trade, and Employment. Washington, DC: The National Academies Press. doi: 10.17226/764.
×
Page 144
Suggested Citation:"Measuring Productivity in Services Industries." National Academy of Engineering. 1988. Technology in Services: Policies for Growth, Trade, and Employment. Washington, DC: The National Academies Press. doi: 10.17226/764.
×
Page 145
Suggested Citation:"Measuring Productivity in Services Industries." National Academy of Engineering. 1988. Technology in Services: Policies for Growth, Trade, and Employment. Washington, DC: The National Academies Press. doi: 10.17226/764.
×
Page 146
Suggested Citation:"Measuring Productivity in Services Industries." National Academy of Engineering. 1988. Technology in Services: Policies for Growth, Trade, and Employment. Washington, DC: The National Academies Press. doi: 10.17226/764.
×
Page 147
Suggested Citation:"Measuring Productivity in Services Industries." National Academy of Engineering. 1988. Technology in Services: Policies for Growth, Trade, and Employment. Washington, DC: The National Academies Press. doi: 10.17226/764.
×
Page 148
Suggested Citation:"Measuring Productivity in Services Industries." National Academy of Engineering. 1988. Technology in Services: Policies for Growth, Trade, and Employment. Washington, DC: The National Academies Press. doi: 10.17226/764.
×
Page 149
Suggested Citation:"Measuring Productivity in Services Industries." National Academy of Engineering. 1988. Technology in Services: Policies for Growth, Trade, and Employment. Washington, DC: The National Academies Press. doi: 10.17226/764.
×
Page 150
Suggested Citation:"Measuring Productivity in Services Industries." National Academy of Engineering. 1988. Technology in Services: Policies for Growth, Trade, and Employment. Washington, DC: The National Academies Press. doi: 10.17226/764.
×
Page 151
Suggested Citation:"Measuring Productivity in Services Industries." National Academy of Engineering. 1988. Technology in Services: Policies for Growth, Trade, and Employment. Washington, DC: The National Academies Press. doi: 10.17226/764.
×
Page 152
Suggested Citation:"Measuring Productivity in Services Industries." National Academy of Engineering. 1988. Technology in Services: Policies for Growth, Trade, and Employment. Washington, DC: The National Academies Press. doi: 10.17226/764.
×
Page 153
Suggested Citation:"Measuring Productivity in Services Industries." National Academy of Engineering. 1988. Technology in Services: Policies for Growth, Trade, and Employment. Washington, DC: The National Academies Press. doi: 10.17226/764.
×
Page 154
Suggested Citation:"Measuring Productivity in Services Industries." National Academy of Engineering. 1988. Technology in Services: Policies for Growth, Trade, and Employment. Washington, DC: The National Academies Press. doi: 10.17226/764.
×
Page 155
Suggested Citation:"Measuring Productivity in Services Industries." National Academy of Engineering. 1988. Technology in Services: Policies for Growth, Trade, and Employment. Washington, DC: The National Academies Press. doi: 10.17226/764.
×
Page 156
Suggested Citation:"Measuring Productivity in Services Industries." National Academy of Engineering. 1988. Technology in Services: Policies for Growth, Trade, and Employment. Washington, DC: The National Academies Press. doi: 10.17226/764.
×
Page 157
Suggested Citation:"Measuring Productivity in Services Industries." National Academy of Engineering. 1988. Technology in Services: Policies for Growth, Trade, and Employment. Washington, DC: The National Academies Press. doi: 10.17226/764.
×
Page 158
Suggested Citation:"Measuring Productivity in Services Industries." National Academy of Engineering. 1988. Technology in Services: Policies for Growth, Trade, and Employment. Washington, DC: The National Academies Press. doi: 10.17226/764.
×
Page 159

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.

Measuring Productivity in Services Industries JEROME A. MARK The increased importance of services industries over the last two decades and concern over U.S. productivity growth have stimulated interest in pro- ductivity measures for this sector of the economy. As a result, the Bureau of Labor Statistics (BLS), which has responsibility for developing the gov- ernment's measures of productivity, has concentrated a great deal of its efforts on expanding the number of services industries for which it publishes pro- ductivity data. These data are in the form of indices of output per unit of input derived from dividing an index of output for an industry by the cor- responding index of input. This has been and is a difficult undertaking, but the number of services industries for which productivity measures are available has increased sub- stantially in recent years. Nevertheless, much remains to be done. This chapter describes these efforts, including some of the problems of measuring productivity in services industries and the way in which some of the measures have been developed. It also presents the results of BLS work, including the movements of productivity in the various services industries.) Some economists have argued that the productivity slowdown in the United States over the last decade and a half may have been more apparent than real because of the increased importance of the services sector and the weak- nesses of the productivity measures in services industries. Others have ques- tioned whether productivity measures can be derived at all for services industries because of the very nature of their activities. The traditional arguments have generally been that the outputs of goods industries are tangible and storable, and therefore measurable, while those of services industries are neither, and 139

140 JEROME A. MARK therefore not measurable. Related to this is the belief that units of services are less homogeneous than units of goods, which reflects the greater differ- ences in quality among units and thereby presents additional problems of measurement. There are indeed serious problems of measurement for some parts of the services sector, and it may not be possible even with intensive study to resolve some of the conceptual difficulties or to develop the data necessary for reliable measures of output and, in turn, productivity for these activities. The lack of homogeneity in many legal, medical, educational, and enter- tainment services, for example, clearly presents difficulties of measurement that may preclude the derivation of satisfactory productivity measures. On the other hand, there are many parts of the services sector for which the problems of measurement are no more severe than for parts of the goods- producing sector. Successful approaches for developing measures of goods activities can be applied to many services activities. To understand the diverse movements of productivity among industries, it is important to attempt to develop separate productivity measures for ser- vices industries and, to the extent possible, resolve the difficulties of mea- surement. That is what the BLS has been attempting to do. DEFINITIONS In this chapter the services sector is defined broadly to include all non- commodity-producing industries. This encompasses the major industrial groupings of transportation, communications, electricity, gas, and sanitary services (public utilities), trade, finance, insurance, real estate, and govern- ment, as well as business and personal services. The industries excluded from this group are those in agriculture, mining, construction, and manu- facturing. Other definitions have been used by private groups presently in- cluded in this definition. For example, Fuchs (1968) in his work on services industries' output and productivity measurement excludes transportation, communications, and public utilities. Marimont (1969) in describing output measurement in services industries in the national income and product ac- counts limits his coverage to finance, insurance, real estate, and business and personal services. Any definition is perhaps arbitrary. The BLS also uses the broader definition to ensure inclusion of as many industries as possible. Productivity measures are those that relate physical output to physical input. As such, they encompass a family of measures including single-factor input productivity measures, such as output per unit of labor or output per unit of capital input, as well as multifactor productivity measures, such as output per unit of labor and capital combined. The most extensively developed and widely used productivity measure is the one relating output to labor input. It is a measure relevant for analyzing

MEASURING PRODUCTIVITY IN SERVICES INDUSTRIES 141 labor costs, real income, and employment. It is also the type of productivity indicator for which data are more readily available to derive adequate mea- sures. As a result, almost all BLS services industries productivity measures developed are of this type and the ones on which this chapter concentrates. Work has been undertaken to develop multifactor productivity measures including capital, materials, and purchased services inputs, as well as labor for telephone communications and for electric, gas, and sanitary services, but these measures are not yet available. Changes in output per employee hour, as with all single-factor productivity measures, do not imply that labor is solely responsible for the changes in productivity. Movements in output per hour reflect technological innovations, changes in capital input, scale economies, education, management, and other factors, as well as skills and efforts of the work force. PROBLEMS OF MEASUREMENT In many ways, the problems of measuring output and, in turn, productivity for services industries are similar to those for goods industries. That is, the output indicator must be quantifiable and independent of the input measure. An industry output measure that is based on an input measure, as it is in some instances in the national income and product accounts (e.g., the prod- ucts of general government, households, and nonprofit institutions) obviously results in an incorrect measurement of change in productivity for that activity. Similarly, the coverage of the output measure should be the same as that of the labor input measure. If not, an imputation is implicitly being made that the movement of the portion of the input that is not included in the coverage of the output measure is the same as the covered portion. It is important also to distinguish between intermediate and final services. Productivity measurement attempts to ensure that the indicators used represent output flowing from the industry being measured and are not part of an intermediate step in the services flow. In this sense, productivity measurement differs from work measurement, which generally refers to the analysis of the labor requirements at operational stages of an activity. For work measurement the technology associated with the activity is fixed. Productivity measurement refers to the final services provided by the organization or industry and their relationship to input. Changes in these measures do reflect, among other things, changes in technology. OUTPUT In the case of an industry providing one type of service, output is merely a count of the units of this service, however defined. This assumes that there

142 JEROME A. MARK is homogeneity in the service being counted. In the more usual case of an industry providing a number of heterogeneous services, the various units must be expressed in some common basis for aggregation. In measures of an output per unit of labor input, this basis is in terms of the base year labor input requirements for the different types of services. In this way, the output measures for developing labor productivity measures differ from the more traditional production measures which are based on total price or value-added weighting of the components.2 When there are quality changes within a service, adjustments must be made in the output measure to account for the fact that the service is no longer the same homogeneous unit. However, the meaning of quality change for labor productivity measurement differs from the usual concept of quality change associated with consumer price measurement in that it reflects dif- ferences in producers' labor requirements or labor costs rather than differ- ences in consumer utility. This difference in the meaning of quality change for productivity mea- surement versus consumer price measurement is illustrated in the goods area with regard to the introduction of the catalytic converter in automobiles. The catalytic converter was added to automobiles in the 1970s as a device for meeting the antipollution requirements mandated by the government. The cost of purchasing an automobile increased when the device was added. The utility or satisfaction to the consumer who purchased the automobile did not increase, and in developing a measure of changes in consumer prices as an indicator of changes in the cost of living, this could be viewed as a price increase (similar to a tax). From the point of view of the producer, this change represented a quality increase. Labor and capital resources went into the production of the item, and it was an addition to the car in effect there was now more car per car. After careful consideration the BLS, which is responsible for providing both the government's consumer price indices and productivity indexes, decided to treat it as a quality increase. Ideally, then, output measures should incorporate data on the number of services provided differentiated by unit labor requirements and in sufficient detail to account for quality differentials. In practice, however, such data are generally not available. As a result, approximations based on alternative approaches utilizing various assumptions are used. In the absence of quantitative information on the units or amount of ser- vices, the principal alternative is to remove the change in price from the change in value (reflecting both price and quantity) of the volume of services. This approach is tantamount to weighting the quantities of services provided with price weights. Insofar as price relationships among the various com- ponent services of a services industry are similar to unit labor requirements or unit labor costs relationships, this measure approximates the desired mea-

AIEASURING PROD UCTl VI 7Y IN SER VICES IND USTRIES 143 sure. Also, since it is generally easier to measure price change for services defined with detailed specifications, this approach is most generally used when adequate quantity information is not available. However, this approach requires price data in sufficient detail to represent adequately the price trends of services included in the change in value of the services. Otherwise, price movements of the covered areas are implicity imputed to the uncovered areas. However, because the relationship among the price movements of similar services is stronger than the relationship among quantity changes of various services, this alternative has greater vi- ability than imputing quantity changes of covered areas to those of uncovered areas. Nevertheless, the use of price deflators still requires ideally that adjust- ments for quality change be made. As mentioned earlier with regard to direct quantity measurement, this adjustment to the price measure should also be made on a cost basis.3 In actual practice, the BLS output measures for its services industries productivity measures are mixtures. Some, such as those for transportation and public utilities industries, are based on quantity data. Others are based on price deflation because of inadequate quantity information. Others utilize the deflation approach at lower levels of aggregation and labor input weight- ing at higher levels of aggregation. LABOR INPUT To derive a productivity measure that relates the output to the correspond- ing labor involved in the production or services-generating process, it is important to have data on the hours worked by all persons involved in the production process. In addition, the hours should refer to hours worked differentiated by types of employees in the particular industries. Unfortu- nately, the data available have serious gaps for meeting these requirements. The data required cover the hours of nonsupervisory, supervisory, and unpaid family workers, as well as the self-employed. The principal source of data on employment and hours is the BLS survey of establishments' payrolls, the Current Employment Survey (CES). This survey provides good measures of employment and hours of nonsupervisory workers by industry, but it does not provide data on the average hours of supervisory workers or the employment and hours of the self-employed and unpaid family workers. Information on the self-employed and unpaid is derived from a survey of households of the noninstitutional population, the Current Population Survey (CPS). These data, which are based on a survey of 60,000 households in the United States, are adequate for the measures of the business economy and major sectors but present limitations when used for measures of services industries.

144 JEROME A. MARK At the present time, the average weekly hours of supervisory workers are assumed to be equal to those of nonsupervisory workers in services industries. This assumption presents fewer limitations for developing measures of change than for developing measures of levels. As mentioned earlier, a desirable measure of productivity is one that reflects the change in labor input actually involved in the generation of the services provided. The hours data in the CES are based on hours paid and include paid vacations, holidays, sick leave, and other time off in addition to actual hours worked. To the extent that leave practices change, the resultant productivity measures over- or understate the actual change in output per hour. To develop a better series of hours at work, the BLS has been conducting an annual survey since 1982 of some 4,000 establishments to collect data on hours at work and hours paid for nonsupervisory workers in the private nonagricultural business sector. From this survey, ratios are developed to adjust the present hours-paid measures to an hours-at-work basis. The def- inition of hours at work was established, after careful study, as time on the job or at the place of work. It includes coffee breaks, short rest periods, paid cleanup time, and other paid time at the workplace, in addition to actual time worked. This definition was considered to be conceptually the most acceptable one for which data could be extracted from establishment records. Although the appropriate hours information is currently available from this survey for the aggregate measures, a substantial expansion of this survey will be required to develop reliable data for specific services industries. This will be very costly. The BLS measures of productivity based on the hours of all persons assume that workers are homogeneous with respect to skill. However, a highly skilled worker can be viewed as providing more labor services per hour than a less skilled one. When skill differences are ignored, increases in skill levels are measured as increases in productivity. As a result, shifts from less skilled to more skilled labor because of increased education or experience are not reflected as increases in the measures of labor input. This would not be a problem if the proportions of workers at different levels of productivity were constant over time. However, to the extent that there are changes in the composition of the work force with respect to education and experience, which result in skill differences, it may be desirable to adjust the labor input measure for these changes, which otherwise would be reflected in the productivity measure. To address this problem, previous studies have usually taken the position that relative wage or income level differentials associated with specific worker characteristics reflect marginal productivity of these attributes. Generally, the included characteristics are the number of years of schooling, age, sex, and possibly industry and occupation (Gollop and Jorgenson, 19801. Weight-

MEASURING PRODUCTIVITY IN SERVICES INDUSTRIES 145 ing the quantity of labor (measured in hours or employees), classified by these characteristics of the work force, by relative wage or income differ- entials results in an aggregate measure of labor input intended to reflect the composition of the work force. This procedure is not without problems. For example, workers with similar characteristics have different earnings in different occupations and industries. However, this correlation between industry or occupation and earnings may also be due to influences other than productivity, such as differences in the cost of living or degree of unionization. The BLS is currently developing new measures of labor input based solely on changes in the amount of work experience and schooling workers acquire (Waldorf et al., 19861. The methodology used follows directly from the economic theory of human capital developed by Mincer (1974) and Becker (19751. It assumes that increased schooling and on-thejob training increase one's stock of skills and productivity. It also assumes that economic returns to higher education and additional work experience reflect the marginal pro- ductivity of these characteristics. The BLS has developed a multidimensional data base which cross-classifies the annual hours of workers grouped by schooling and experience. The data base is developed from various models that make use of decennial census data, a matched sample from the CPS and from social security records. As mentioned above, it is recognized that hourly wages differ not only because of skill differences but also because of factors unrelated to produc- tivity. Accordingly, simple averages of hourly wage rates for each education and experience group are not necessarily suitable approximations of marginal productivity. Consequently, the BLS has developed an econometric model that provides measures of wages dependent on changes in education and experience but simultaneously controls for other types of variation. Skill-adjusted labor input measures have been developed for the business sector, and this work is currently being extended to determine the feasibility of developing corresponding measures for specific industries, especially ser- vices industries. MEASURES OF PRODUCTIVITY FOR SERVICES INDUSTRIES At present, the BLS publishes indices of output per hour (i.e., output per unit of labor input) for industries in each of the major services activities. This includes all groups of Standard Industrial Classification (SIC) 40 and higher: transportation, communications, public utilities, trade, finance, in- surance, real estate, and business and personal services. Productivity mea- sures for the federal government are also published with separate detail for the common functions provided by federal agencies, such as recordkeeping, library services, buildings and grounds maintenance, and loans and grants.

146 JEROME A. MARK In addition, work is currently being conducted on the development of indices of productivity for hospitals, wholesale trade activities, and automobile repair shops. The following sections describe the procedures utilized and some of the problems present in developing the particular measures. Because the methods and the sources used for deriving the labor input measures are the same for each of the separate industries, this discussion focuses on the output measures developed for the productivity indices. Transportation The BLS publishes productivity indices for five transportation industries: railroads, intercity trucking, intercity buses, air transportation, and petroleum pipelines. These measures cover 57 percent of transportation employment. Conceptually, the measures for transportation industries are easier to de- velop than those for other services industries because the transportation in- dustry output the movement of goods or passengers over distances is more easily quantified. Output units in transportation have two dimensions, amount and distance, that reflect how much has been transported how far. As such, ton-miles, passenger-miles, barrel-miles, and so forth are the pri- mary output indicators for these services. Historically, these data have been available mostly from the regulatory agencies of the transportation industries. In many cases, however, there are data gaps that place certain limitations on the measures. For example, it is sometimes impossible to adjust the output measures adequately for changes in the average length of haul. The unit labor requirements associated with the movement of freight and passengers are usually greater for short hauls than for long hauls. Therefore, a shift from a long haul to a series of short- haul trips or vice versa could be interpreted as a change in productivity, although only the mix of trips had changed. For the two major freight-carrying industries, railroads and trucking, un- differentiated ton-mile data are reported for total freight operations. In truck- ing, the ton-mile data are also reported separately for general, contract, and other carriers. However, the output measures should reflect the kinds of commodities handled and the average distances they are moved since these represent separate types of services. The preferred way to develop such measures would be to combine the tonnage and average haul of each com- modity by its respective labor input requirements and aggregate the results for all commodities transported. Unfortunately, this cannot be done with available data. However, separate commodity information on tonnage for railroads is available from the Interstate Commerce Commission for about 200 com- modity lines, ranging from agricultural and mining products to motor vehicles

MEASURING PRODUCTIVII'Y IN SERVICES INDUSTRIES 147 and scientific instruments. Several years ago similar information was avail- able for the trucking industry, but it was discontinued. The BLS uses these data to adjust the overall measure of freight ton-miles for changes in the composition of goods carried. Although this commodity adjustment represents an improvement over un- differentiated ton-miles, refinements cannot be developed to the extent de- sired. The commodity index adjustments are made in terms of unit revenue weights. The underlying assumption, therefore, is that differences in labor requirements among commodities are similar to differences in unit revenues. Since labor costs constitute more than half of each industry's total operating costs, this assumption may not be unreasonable. However, the proportions could conceivably differ by commodity. In deriving the total industry output index for each of the transportation industries, the freight ton-miles measure (adjusted or not) is combined with the revenue passenger-miles measure. The deregulation of many transportation industries has resulted in elimi- nation of some of the operating statistics that were previously published and used to develop output indices. As a result, some of the transportation industry measures have had to be extended on the basis of more limited inflation. The BLS has been cooperating with other government agencies to ensure that adequate statistics for transportation industries remain available. Communications The BLS productivity index for telephone communications covers about four-fifths of the employment of the communications sector. The output index is derived from revenues of all telephone companies reporting to the Federal Communications Commission. The revenues are stratified by major source (i.e., local, toll, and miscellaneous) and deflated by specially prepared price indices reflecting these different services. The deflated revenue measure includes revenues from private line services. It also accounts for television, radio, and computer transmission by telephone industry facilities, as well as directory services. Despite the details included in the output measure, improvements could be made if information were available on the intensity of use of telephone equipment by customers. The number of calls made can vary without revenue varying proportionately because of flat charges, such as WATS (Wide-Area Telecommunications Service) line or local call charges. To the extent that the number of such calls varies over time, the index over- or-understates output change.

148 JEROME A. MARK Electric, Gas, and Sanitary Services Services rendered by public utilities range from the provision of light, heat, and water to the disposal of solid and liquid wastes. In this area, the BLS currently publishes productivity indices for electric utilities, gas utilities, and a combination of the two. The measure of electric utility output covers all privately owned utilities, which account for roughly four-fifths of the total output generated in the United States. Output is defined in terms of kilowatt hours generated and distributed. The measure of output of gas utilities is defined in terms of therms (1 therm equals 100,000 British thermal units) delivered to customers by all privately owned companies (which account for 95 percent of total gas output). Since the labor requirements per kilowatt hour or per therm are higher for residential than for commercial and industrial customers, and higher for small establishments than for larger ones, the BLS differentiates both the kilowatt- hours and the therms by type of customer. The BLS is currently attempting to develop a productivity index for sanitary services, including sewage and refuse systems. For this measure, output is measured in terms of disposal of metered liquid sewage and tons of solid waste processed differentiated by type of waste insofar as different types of solid waste require more or less labor per unit processed. Retail Trade The BLS has been publishing measures for retail trade industries since 1975; however, the number of industries has increased markedly in recent years. At present, indices are available for 12 important industries: retail food stores, franchised new car dealers, gasoline service stations, apparel and accessory stores, furniture and appliance stores, eating and drinking places, drug and proprietary stores, and liquor stores. Apparel and accessory stores are further broken down into men's and boys' clothing stores, women's ready-to-wear stores, family clothing stores, and shoe stores. Also, furniture and appliance stores are disaggregated into furniture and furnishings stores and appliance stores. Work is currently under way on measures for hardware stores, auto and home supply stores, and department stores. For most retail trade industries, data on gross sales in current dollars, deflated by appropriate price indices, are used to estimate real output. This method can yield good measures of real output if adequate price indices reflecting the price movements of the various commodities sold by the es- tablishments can be developed. The recent improvements that have been made in the BLS price index program have enabled it to develop output and productivity indices for more components of retail trade.

MEASURING PRODUCTIVITY IN SERVICES INDUSTRIES 149 Productivity indices based on deflated value of sales output measures reflect shifts among services with different values but the same trade labor requirements. Therefore, the overall industry productivity index can show movements without any change in component elements. In retail industries, a large portion of the value of sales is provided by the manufacturers and the wholesalers of the products sold. A net output measure would be desirable because it would correspond most closely to the functions provided by the retailer. Unfortunately, net output measures based on sep- arately deflated final sales and cost of materials data can result in large errors of measurement when the cost of materials is a large proportion of the final value. All the errors in the current value of sales, the current value of goods purchased, the product price indices, and the materials price measures affect this residual. A gross or total sales measure will yield the results of a net or value-added measure with perhaps less measurement error if the value added as a percent of sales (gross margin) does not change over time. Available data indicate that for most of the industries published, average gross margins have not changed substantially over time. In order to introduce labor input weighting at some stage, the indices for retail trade industries for the most part are developed in two stages. Deflated output measures are first developed for detailed merchandise lines. These are aggregated to higher levels, and the resultant indices are then combined with labor cost weights. For example, in retail food stores, sales for 13 key merchandise lines are deflated by using specially prepared price indices based on the BLS Consumer Price Index components. The merchandise lines indices are aggregated into five depart- ment lines: meat, produce, frozen food, dry groceries and dairy, and all others. These, in turn, are aggregated with labor cost weights to develop the overall output measures for grocery stores. Wholesale Trade At present, the BLS does not publish any indices for industries in the wholesale trade sector. It is currently examining data for four industries in considerable detail in order to derive reliable measures: metal service centers; scrap and waste materials dealers; petroleum bulk stations; and beer, wine, and distilled spirits distributors. These industries include about half a million workers or 10 percent of the employment in the sector. Physical quantity data are available to develop output measures for these industries. The quantity data for disaggregate commodities will be combined with fixed period labor input weights reflecting the services provided by the wholesalers to retailers and other users. Several measurement problems concerning these industries-need to be resolved. Some firms perform work on commodities that they distribute to retailers, but this varies substantially among wholesale distributors. Whether

150 JEROME A. MARK the labor input weights can adequately take this into account is questionable. In addition, in some instances a regional wholesaler distributes commodities to local wholesalers. This creates a problem of duplicate counting in the overall industry output measure. F. Inance In finance, the BLS publishes a measure for the commercial banking industry. The output measure for this industry is in terms of the three major services that commercial banks render their customers: deposits, loans, and trust services. While banks also perform non-fund-using services, such as safe deposits and customer payroll accounting, lack of adequate data pre- cludes deriving a measure for them. However, because the proportion of employees engaged in such services is small, the overall measure is little affected by the omission. There has been a great deal of discussion over the years about the appro- priate measure of banking output. Much of the discussion narrows down to whether the appropriate concept is one based on what has been called the liquidity approach or one based on a transactions approach (Gorman, 19691. In the former, banks are viewed as providers of money to hold, and their output is measured in terms of the interest received on the volume of deposits. Such interest received by banks is assumed to be equivalent to income foregone by depositors due~to their preference for holding their financial assets in banks as deposits rather than investing directly. The interest that the depositors forego represents the value of the banks' services in meeting their depositors' preference for liquidity. This approach can be ex- tended to savings accounts and other time deposit accounts, the assumption being that the foregone net interest is the value of the banks' services. The other approach is that banks are providing a series of services to their clients which are reflected in the transactions performed. The volume of the banks' output is proportional to the volume of transactions handled. In de- veloping the productivity measure, the BLS adopted the second approach for measuring output. Thus, the final output of banks represents an array of services provided to bank customers relating to depository, lending, and fiduciary functions of banks. Estimates of the number of transactions in each of these services functions are derived. In some instances, no direct count of transactions is available, so the number of transactions is estimated from data on the total value of transactions and on the results of surveys of average transactions amounts. While these estimates have some limitations, a count of the number of transactions is the measure used to reflect the quantities of services provided. Deposit activity is measured in terms of the number of checks transacted and the number of time and savings deposits and withdrawals. The data for

MEASURING PRODUCTIVITY IN SERVICES INDUSTRIES 151 demand deposit activities are derived ultimately from Federal Reserve counts and official benchmark surveys. For time and savings deposit activity, the output measure is derived from data published by the Federal Deposit In- surance Corporation and the Functional Cost Analysis conducted annually by the Federal Reserve Board. Lending services provided by banks are also measured in terms of units. As in the case of deposit activity and trust department activity, the BLS does not use banks' financial data to arrive at component output measures. Such use of financial data would be misleading even if appropriate price deflators could be developed; a rise in the constant dollar value of the loans might simply reflect a few large loans having been made and a fall could reflect the repayment of a few large loans even as the number of small loans may have increased. The BLS measures 10 types of loan output, using data generated by the Federal Reserve Board, the Department of Housing and Urban Development, the Federal Home Loan Bank Board, and others. Included in the loan output measure are residential and commercial mortgage loans, consumer loans, single-payment loans, credit card loans, and commercial loans. The number of loans is usually derived by dividing the average face value of a loan into the total value of all the loans in a given category. The output measure for the fiduciary or trust department services is derived from the number of accounts stratified by five major categories, including employee benefits trusts, personal trusts, and estates. The output indices for the three segments are then aggregated with em- ployment weights to derive the overall output index. Federal Government In addition to developing private sector measures, for the past 14 years, the BLS has been conducting a program of developing labor productivity indices for all federal government agencies with 200 or more employees. Currently, the measures include 390 organizational units representing 59 agencies which cover 69 percent of the civilian work force. The BLS calculated productivity, output, and employee years indices for the overall sample and for each of the 390 organizational units. Indices for individual organizations are not published but are returned to the responsible agency for its own use. Organizational data are grouped into 28 functional areas, such as audit, electric power, and personnel, for which the indices are published. Where possible, the relevant concept of output of a government agency is its final product output, that is, what the given organization produced for use outside the organization. In this work, the output data included in the

152 JEROME A. MARK overall sample are, in fact, final from the perspective of the organization and from the functional groupings in which these organizations are classified. However, since the outputs of one organization may be consumed wholly or partially by another federal organization in the production of its final outputs, all output indicators will not be final from the perspective of the entire federal government. Therefore, the overall statistics presented in the study do not represent federal government productivity, but rather the average of the productivity changes of the measured organizations included in the sample. The data base for the output indicators is quite extensive, having expanded fourfold since the program's inception. Currently, output information is in- cluded on about 3,000 indicators. By expanding the output detail of the survey participants, the BLS is better able to measure the services-oriented activities of the federal government. The indicators are diverse and illustrate the many functions and services provided by the government. For the auditing function, typically reported indicators are inspections completed and audit reports prepared. Providing information to the public is a major function and is measured by such things as statistical reports issued, maps produced, and weather forecasts made. Regulatory activities cross a wide spectrum and include food inspections conducted, drug arrests made, applications and licenses issued, drugs ap- proved, and patents approved. In transportation and utilities, the measures are similar to those employed for the private sector and include revenue ton- miles and kilowatt-hours. To the extent possible, data in sufficient detail are sought on these indicators to take account of shifts in the mix of services provided. State and Local Government More recently, the BLS has begun to explore the possibility of computing productivity indices for state and local government. There are a number of services areas in which final outputs can be specified, including electric power, drinking water, solid waste collection, and alcoholic beverage sales. Most have private sector counterparts. However, there are many services areas in which a lack of research or agreement exists on correct organizational outputs, including such important areas as education, fire, and police. There is no comprehensive system for collecting data on state and local government outputs. In a few select services areas, including electric power and mass transit, output statistics are collected as part of a federal agency's program and oversight responsibilities. However, in most services areas, output statistics are not available.

MEASURING PRODUCTIVITY IN SERVICES INDUSTRIES Business and Personal Services 153 In the area that includes not only business, personal, and repair services, but also education, health and social services, and political organizations, the BLS currently publishes measures of productivity for three activities: hotels and motels, laundry and dry cleaning services, and beauty and barber shops. Measures of productivity for automotive repair shops and for hospitals are also being developed. The employment in the industries for which mea- sures have been or are being developed covers about 23 percent of the total employment in this sector. Because physical output data are not available for the three industries that are presently published, the output indices for these industries are developed by using a deflated value approach. The techniques are similar to those described above in that the changes in revenues for specific services are divided by appropriate price indices and these output measures are then aggregated with employment weights to derive the industry output index. As mentioned above, the BLS is currently developing a measure of pro- ductivity of hospitals. The industry is defined as including all nonfederal short-stay hospitals (i.e., with patient stays of 30 days or less). Because the underlying concepts and measurement procedures are of particular interest and are currently being developed, they are discussed in some detail. Hospitals provide services designed to eliminate, retard, or neutralize pa- thologies. They also provide gynecological, neonatal, and other services. Treatments and related procedures can be regarded as the producer technol- ogies by which those services are rendered. Treatments, of course, are spec- ified to produce desired outcomes. However, since outcomes depend on factors other than, and in addition to, treatments for example, preadmission health of the patients there is some question whether the output measure related to hospital labor and other resources used for a productivity measure should be based on outcomes (e.g., see Scott, 19791. The BLS does not adjust for outcomes in deriving its output measure for its productivity index. Illnesses are classified as diagnoses once they have been identified clini- cally. Diagnoses tend to be standardized, and each diagnostic category or diagnosis-related group (DRG) implies a complex of treatments or proce- dures, which in turn indicate certain kinds and amounts of resource utilization. The concept of DRG was evolved by the Health Systems Management Group at the Yale University School of Organization and Management during the 1970s and has been modified gradually. It is based on the several thousand entries in the International Classification of Diseases (19781. To begin with, all diagnosis codes have been condensed into 23 mutually exclusive and exhaustive major diagnostic categories (MDC) (U.S. National Center for Health Statistics, 19851. The MDCs are generally based on diseases that tend to be diagnosed and treated similarly by specialists. Hospital discharges are coded first by their pertinent MDCs. Each MDC is, in turn, partitioned into

154 JEROME A. MARK numerous DRGs. The major partitions are between surgical and nonsurgical procedures. Surgical procedures are divided into procedural categories based on resource utilization. Nonsurgical procedures are similarly classified hi- erarchically with further partitioning by age, comorbidities (having more than one disease), and complications. Comorbidities and complications are defined in terAms of an increase in length of stay of one day or more, over and above the standard length of stay for a given DRG. The output measure is derived from data on number of inpatient discharges and number of outpatient visits weighted on the basis of revenue weights. The data on inpatient discharges are drawn from the National Discharge Survey of the National Center for Health Statistics of the U.S. Department of Health and Human seArvi~ ~n(1 are rl~cifi-A Hal ,1;~;~ MA A^~ MA—~ V,7 ~ACI~B11~J~ L1~ ~a~l~E;~1 Y . lne alscnarges in each agnostic category are aggregated with weights based on the average hospital operating costs per diagnostic category. The weights are derived from data from the Blue Cross Federal Employee Plan for the period 1972-1978 and, after that, from data on average cost relatives for DRGs determined by the Health Care Financing Administration (HCFA) for Medicare and various state commissions. The cost relatives reflect labor costs and exclude capital costs. Varying intensities of care are captured by the output measure because of the different DRG weights assigned to the various treatment modes. For example, intensity of care varies substantially according to whether or not surgery is perfo``'Aed. The separation of MDCs and, in turn, DRGs into surgical and nonsurgical classifications takes this into account. Rising in- tensities have been partially offset by declining lengths of stay, which are also reflected in the DRGs. When a new service is introduced for example, a new imaging device- intensity of care may not be affected. If it is affected, treatment protocols may not be. When a shift occurs from invasive to noninvasive procedures, such as from surgery to remove kidney stones to extracorporeal shock-wave lithotripsy, intensity declines. A new DRG will be introduced or the weight of an existing one recalibrated. If so the. new weight it intr^A''r~ ;~ + output measure. As with all BLS industry productivity measures when they are being developed, the hospital measure is currently being reviewed by industry and 'suer represen~arAves. Sudanese reviews provide insights into problems not an- ticipated and suggestions for improving the measures that are being devel- oped. ~ 7 ~ ~~ $~ AROMA lll~V LIAR WHAT DO THE DATA SHOW Table 1 shows the productivity growth rates of the various services in- dustries for which the BLS has published measures. As can be seen, the growth rates of the services industries vary substantially. Some, such as

AlEASURl\G PRODUCTIVITY IN SERVICES INDUSTRIES 155 TABLE 1 Productivity in Services Industries for which BLS Productivity Measures Are Available 1965- 1973- 1973- 1981- Industry 1973 9~ it': Nisi Transportation Railroads 4.2 4.6 2.3 10.8 Bus carriers —1.5 - 1.0a —1.4 - 0.3b Intercity trucking 2.7 0.4 0.5 0.6 Postal services 1.3c 1.3 1.4 1.0 Airlines 5.3 3.9 3.5 6.8 Petroleum pipelines 7.9 0.1 - 0.7 5.6 Communications Telephone communications 4.7 6.2 6.7 5 Public utilities Gas and electric utilities 4.9 - 0.5 0.3 - 0.6 Electric utilities 5.4 0.2 0.8 0.8 Gas utilities 3.9 - 2.1 - 0.4 - 4.4 Trade Retail food stores 2.2 - 1.0 - 1.1 - 0.5 Franchised new car dealers 2.6 1.2 0.6 3.3 Gasoline service stations 4.9 3.2 3.2 4.2 Apparel and accessory stores 2.8C 3.9 3.0 5.3 Men's and boys' clothing stores 3.3c 2.3 1.4 3.4 Women's ready-to-wear stores 4.6c 6.0 4.5 6.4 Family clothing stores 5.9C 3.6 1.7 4.1 Shoe stores _0.4C 2.1 2.5 2.9 Furniture and appliance stores 4.3c 2.7 2.5 3.4 Furniture and furnishing stores 4.3c 1.2 1.2 2.6 Appliance stores 4.4c 4.8 4.6 4.7 Eating and drinking places 1.1 - 0.7 - 0.5 - 1.1 Drug and proprietary stores 6.2 0.9 1.5 - 1.6 Liquor stores N.A. 0.3 0.1 - 1.0 Finance, insurance, and real estate Commercial banking 2.1c o.6a 0.2 5.4' Services Hotels, motels, and tourist courts 1.8 0.4 0.7 2 Laundry and cleaning services 1.7 - 1.1 - 1.1 0.1 Beauty and barber shops N.A. 0.7 1.0 - 2 Beauty shops N.A. 1.1 0.9 - 1 a1973_lg84. bl981 - 1984. C 1967- 1973. N.A.: not available. SOURCE: Bureau of Labor Statistics.

156 JEROME A. MARK telephone communications, gasoline service stations, and air transportation, had very high growth rates from 1973 to 1985, ranging from 3.2 to more than 6.2 percent per year in telephone communications when the business sector as a whole was experiencing a productivity growth rate of 0.9 percent per year and the goods sector 1.0 percent per year. Other services industnes, such as men's clothing stores and shoe stores, also had much higher growth rates than the business sector as a whole but were more in line with manu- facturing industries, which averaged 2.4 percent per year over the period. Some services industries, such as gas and electric utilities or laundry and cleaning services, showed very small growth rates and in some cases actual declines in productivity after 1973. In general, the range of growth rates in services industries was very broad and similar to that of pon`1.c-nrorl''~ino industries. ~ ~~ 4~ r~ ~J~1115 It has sometimes been mentioned that the falloff in productivity growth in the business economy of the United States is a reflection of the shift to services in the U.S. economy. Services industries are believed by some to have lower productivity growth than goods industnes; thus, their greater importance over the last decade and a half is believed to have contributed to the productivity deceleration in the business economy after 1973. The wide variation in growth rates does not support that conclusion for the initial There was some falloff in r~e procuc~v~y growth rates ot services industries after 1973, and this falloff contributed to the deceleration (Table 11. However, a sharp falloff in pro- ductivity growth occurred in the goods-producing sector, which until 1981 was greater than the falloff in services industries as a whole (Table 21. Since 1981, however, the picture has changed. Services industries have continued to show lower productivity growth on the average, whereas goods-producing part of the slowdown period from 1973 to 1981. +~ ~ ~~ a.. ~ . . _ ~ .. ~ ~ ~ . . . . . TABLE 2 Output per Hour, Business Sector and Goods and Nongoods Sectors, 1948-1985 (compound annual average rates) 1948-1973 Q/h Q Business Goodsa Nongoodsb Goods—nongoods (difference) u. / - 1 .u —1.4 (). 1 - 2. 1 - 2.3 2.2 - 0.9 - 3.2 aGoods sectors are the farm, mining, construction, and manufacturing sectors. bNongoods sectors include transportation; communications; electric, gas, and sanitary services; trade; finance, insurance, and real estate services; and government enterprises. Q. quantity; h, hour. SOURCE: Bureau of Labor Statistics. 3.0 3.2 2.5 h 3.7 0.7 3.0 0.0 4.0 1.4 0.7 -1.0 -1.4 Q/h 0.6 0.6 0.5 h 1.4 0.1 2.4 0.1 -2.1 -2.3 1981-1985 Q/h Q h 1.4 2.9 0.7 3.0 1.6 2.6 -0.3 3.5 2.9

MEASURING PRODUCTIVITY IN SERVICES INDUSTRIES 157 industries have experienced higher growth. The services industries' slower growth has been contributing to the continued sluggishness in the U.S. pro- ductivity growth in the overall business sector since 1981. CONCLUSION Measurement of productivity change for services industries is not a simple task. Problems of measurement are different for each of the services indus- tries, and the approaches followed to meet those problems vary. Each services industry has to be examined in some detail to determine whether or not reliable measures can be derived. For many services industries, adequate productivity measures can be and have been developed by the BLS. These industry productivity indices can also be useful to companies that have introduced major technological changes. To provide some insights on the impact of various innovations on the performance of the establishments, it is desirable to develop productivity measures for these organizations. More- over, it is helpful to have some basis for comparison. The industry produc- tivity measures presented here can serve as benchmarks for these comparisons. However, company measures must be developed which are reasonably con- sistent, at least in concept, with the overall industry measure. In this way, a company productivity measure showing a rate lower than that for the industry can provide a meaningful warning signal which reflects substantive and not just measurement differences. The expansion in the number of ser- vices industries for which BLS productivity indices are available greatly facilitates this process. Nevertheless, despite recent progress, many difficulties remain in clari- fying some of the basic conceptual problems of defining the output of certain services industries, and many inadequacies are present in the data available for the measurement of productivity in this area. With regard to data, more and improved data are needed on prices, value of output, and capital input. The BLS and other government statistical agencies have undertaken initiatives to expand the coverage in the services area, and the results of these efforts should lead to data suitable for deriving more services industries productivity measures. For example, more services industries price measures are being collected, expanded hours-worked data will provide more disaggregate i~- formation, and continued BLS work on public sector productivity measure- ment will result in more productivity measures for this important sector. However, some areas within the services sector, such as education, en- tertainment, legal and social services, and many segments of medical ser- vices, present severe conceptual and data problems for measuring productivity. Another rapidly growing area that provides great difficulties in measurement is the generation of software. Because of the complexity and heterogeneity of the products developed by this industry and the absence of data on quan-

158 JEROME A. MARK titles, values, or price movements, it will be virtually impossible to derive reliable productivity measures for the industry for some time. Appropnate solutions to these problems may be a long time in coming. Even additional resources may have limited effect until the conceptual problems can be dealt with appropriately. In the meantime, the BLS will continue to expand the coverage for those industries in which the problems seem tractable. NOTES 1. This chapter expands on and updates an article (Mark, 1982) that described some of the earlier work of the BLS in developing measures for services industries. 2. In the case of multifactor productivity measures, the appropriate weight would be the sum of the factor costs and would be closer to the traditional value-added weight. 3. The literature on the problems of quality measurement for output and price indices is extensive. The approaches to developing adjustments for quality change to both types of indices are similar, but the extent to which they can be met differs. Perhaps more efforts have been devoted to obtaining adjustments for quality change to the measures for goods industries largely because of the longer history of developing output and price measures for goods industries. 4. The importance of preadmission conditions is perhaps best illustrated by a statement of Joseph A. Califano, Jr., former Secretary of the U.S. Department of Health, Education and Welfare, before the Committee on Labor and Human Resources, U.S. Senate, January 12, 1987: "Heart disease is America's number-one killer. People have the impression that coronary bypass surgery, modern cardiopulmonary techniques, miracle pills and human heart transplants are the way to battle heart diseases. Right? Couldn't be more wrong. Since 1970, our nation has experienced a dramatic 25 percent decline in deaths from coronary heart disease. The major reasons? Improved eating habits the reduction of cholesterol- and the decline in cigarette smoking were responsible for more than half of the decline in deaths from heart disease." REFERENCES Becker, G. 1975. Human Capital. Chicago and London: University of Chicago Press. Committee on National Statistics, National Research Council. 1986. Statistics About Service Industries. Washington, D.C.: National Academy Press. Fuchs, R. 1968. The Service Economy. New York: Columbia University Press. Gollop, F. M., and D. W. Jorgenson. 1980. U.S. productivity growth by industry, 1947-73. Pp. 17-124 in New Developments in Productivity Measurement and Analysis, J. W. Ken- drick and B. N. Vaccara, eds. Chicago: University of Chicago Press. Gorman, J. 1969. Alternative measures of the real output and productivity of commercial banks. Pp. 155-188 in Production and Productivity in the Service Industries, V. R. Fuchs, ed. New York: Columbia University Press. Marimont, M. 1969. Measuring real output for industries providing services: OBE concepts and methods. Pp. 15-40 in Production and Productivity in the Service Industries, V. R. Fuchs, ed. New York: Columbia University Press. Mark, J. A. 1982. Measuring productivity in service industries. Monthly Labor Review (June). Mincer, J. 1974. Schooling, Experience and Earnings. New York: Columbia IJniversity Press. Scott, W. R. 1979. Measuring outputs in hospitals. Pp. 255-275 in National Research Council, Measurement and Interpretation of Productivity. Committee on National Statistics. Wash- ington, D.C.: National Academy Press.

MEASURING PRODUCTIVITY IN SERVICES INDUSTRIES 159 U.S. National Center for Health Statistics. 1985. Diagnosis-related Groups Using Data from the National Hospital Discharge Survey, United States, 1982. R. Pokras and K. K. Kub- lishke. Advance Data from Vital and Health Statistics, No. 105. Department of Health and Human Services Pub. No. (PHS) 85-1250, January 18, 1985. Hyattsville, Md.: Public Health Service. Waldorf, W. H., K. Kunze, L. S. Rosenblum, and M. B. Tannen. 1986. New measures of the contribution of education and experience to U.S. productivity growth. Paper presented at the annual meetings of the American Economic Association, December 28-30, New Orleans, La.

Next: Air Cargo Transportation in the Next Economy »
Technology in Services: Policies for Growth, Trade, and Employment Get This Book
×
Buy Paperback | $60.00 Buy Hardback | $60.00
MyNAP members save 10% online.
Login or Register to save!
Download Free PDF

Beginning by dispelling some of the myths about services, this provocative volume examines the growth in services, the way technology has shaped this growth, and the consequences for the American economy. Chapters discuss such topics as the effects of technology on employment patterns and wages, international trade in services, and the relationship between services and the traditional manufacturing industries.

  1. ×

    Welcome to OpenBook!

    You're looking at OpenBook, NAP.edu's online reading room since 1999. Based on feedback from you, our users, we've made some improvements that make it easier than ever to read thousands of publications on our website.

    Do you want to take a quick tour of the OpenBook's features?

    No Thanks Take a Tour »
  2. ×

    Show this book's table of contents, where you can jump to any chapter by name.

    « Back Next »
  3. ×

    ...or use these buttons to go back to the previous chapter or skip to the next one.

    « Back Next »
  4. ×

    Jump up to the previous page or down to the next one. Also, you can type in a page number and press Enter to go directly to that page in the book.

    « Back Next »
  5. ×

    To search the entire text of this book, type in your search term here and press Enter.

    « Back Next »
  6. ×

    Share a link to this book page on your preferred social network or via email.

    « Back Next »
  7. ×

    View our suggested citation for this chapter.

    « Back Next »
  8. ×

    Ready to take your reading offline? Click here to buy this book in print or download it as a free PDF, if available.

    « Back Next »
Stay Connected!