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.
Welfare Dimensions of Productivity Measurement MILTON MO S S University of Pennsylvania Efficient production of goods and services is a primary goal of economic effort, and statistical measurement of productivity is an important tool for monitoring and promoting its advance. The Panel to Review Productivity Statistics was asked to evaluate and recommend improve- ments in existing measurements of productive efficiency. In addition to meeting its prime responsibility, the Panel wished to have a paper that would discuss measurement issues in the wider con- text of the effects of economic effort upon the workplace, the home, and in society generally. One of the main purposes of such an extended view is to put into a useful perspective the frequent criticisms of the national income accounts for sometimes showing increases in economic well-being when, it is alleged, by some reckoning welfare is not increasing or might even be declining. Such a paper can also serve two other pur- poses. First, by contrasting the measurement of economic production with welfare considerations, it can help make clear that the national accounts do not, strictly speaking, provide a measure of welfare, how- ever important economic production may be in contributing or failing to contribute to welfare. Second, by illustrating attempts to measure welfare in particular areas of life it can point out potentially useful Comments on an earlier draft by Arthur Broida, Sharon De Sha, Solomon Fabricant, Gary Fromm, Beatrice Goley, William Kruskal, Jerome Mark, Margaret Martin, Dave O Neill, Albert Rees, and Markley Roberts are much appreciated. 276
Welfare Dimensions of Productivity Measurement 277 and appropriate ways of relating strictly economic measures of produc- tion to welfare measures. Although the issues underlying welfare measurement are exceedingly complex, the conclusions reached in this paper can be stated rather simply: First, goods and services as measured by the national accounts are a substantial component of welfare in their own right and are a basic input to welfare considered in broader terms. Many useful analyses of economic and social welfare can be undertaken by staying within the established definitions of economic output, input, and productivity. For example, it has become apparent that a socially acceptable standard of living for the elderly requires a rising level of national economic productivity as customarily defined. Second, other measures of output and input are needed for those analyses that try to evaluate the "effectiveness" with which certain social or human objectives are met, such as the effectiveness with which health facilities in fact promote health, or the effectiveness with which antipollution devices clean the air. Nevertheless, even with such measures of effectiveness, which go beyond the bounds set by the national accounts, the national accounts provide a basis for determining the costs of achiev- ing some level of effectiveness. Third, the construction of a single index of welfare for the nation as a whole would require so many arbitrary imputations and weights, that the results, however interesting, would have a journalistic rather than a scientific value. Moreover, even if such an index of welfare were avail- able, it probably would not be possible to determine the real cost or efficiency of resource use associated with changes in the index, that is, to measure "welfare productivity." However, useful partial analyses can be made in a number of areas discussed in this paper. These areas are consumption, including private and collective consumption, income distribution, quality of working life, hours of work and use of time, health facilities, and the environment. A general point emphasized is that the national income accounts measure output from the point of view of a producer, not a consumer. Economic production is an input to consumption, leading to a broader outcome, which may be called welfare. There would be little sense in combining such an input measure with the broad measure of outcome, or abandoning the input measure by changing it in order to approxi- mate the outcome measure, particularly if it is possible to have separate measures, and we shall see that it is. Keeping economic production measurements distinct and relating them to welfare objectives would seem to be the better approach, both for defining welfare objectives and
278 PAPERS for using the measure of economic output as a tool for social better- ment. The final section of the paper raises the question whether a social accounting system is feasible that relates the national accounts to demo- graphic and social data in an integrated fashion, keeping the economic boundary distinct. This is a different question from whether to seek a single index of welfare. The work on a social accounting system, including social indicators, is still in an early stage; hence our con- clusions about it are necessarily tentative. We believe, however, the work should be encouraged. GENERAL CONSIDERATIONS CONCERNING WELFARE DIMENSIONS Welfare has been viewed in many dimensions, both economic and social. We discuss these in general terms in this section and examine them in more detail later. Measuring well-being was long considered synonymous with mea- suring the standard of living or, as some purists would have it, the "level of living" to distinguish between norms and facts. Such measures referred to the physical quantity of goods and services available for consumption: the volume of food, clothing, housing, transportation services, medical services, recreational services, and SO forth, combined on the basis of the prices in some reference period and often expressed on a per-capita basis. In short, well-being meant economic well-being in the sense of the volume of material goods and services as they are presently measured in the national economic accounts of various coun- tries. Changes in such measures over time and differences in living stan- dards within and between countries indicated relative well-being, and frequently called dramatic attention to marked differences in the avail- ability of material goods and services. Problems of comparing different aggregates of goods and services (so-called index number problems) were acute, particularly in connection with comparisons over long periods of time or between countries of widely different tastes, economic struc- tures, and levels of development. Analysts have given some thought from time to time to dimensions of well-being other than those based on sheer quantities of goods and services, which for short we will henceforth call "GNP-type goods." For example, it has been suggested on occasion that variety as distinct from sheer quantity, is an important dimension of welfare; that a wider choice of what to buy with a given income would result in greater con- sumer satisfaction. But no one to my knowledge has been successful in
Welfare Dimensions of Productivity Measurement 279 developing a useful measure of variety. Not only is it hard to assign a meaningful value to the existence of certain additional choices within some category of a consumer good, but there is also the problem that in advancing economies, certain inexpensive varieties of a good disappear at the same time that more advanced, expensive, though more various types appear. For example, when automatic transmissions became available in a wide variety of automobile makes, models, and body styles, it became hard to obtain cars with standard gear shifts, which sortie people preferred. Many deplored the disappearance of the simple and inexpensive Model-T Ford when the newer more complicated models were introduced. Perhaps a more important dimension of welfare associated with GNP- type goods is quality as distinct from quantity. Improvements that yield more satisfaction to consumers arising from changes in the design, performance characteristics, and other features of goods and services over time are often not captured in the usual measures of physical volume. During World War II there was concern that existing measures of prices and quantities failed to take into account deterioration in quality of many goods (Mitchell 19451. The section on consumption in this paper takes up these issues in more detail. Welfare has been viewed not only in economic terms but in much more extended dimensions. The grounds for wanting to look well beyond the strict boundary of economic production are to be found in both positive and negative reactions to affluence. One of the earliest and most eloquent of the positive positions was expressed during the depths of the great depression by Keynes, who predicted that the force of compound interest on economic growth and productivity would in about 100 years solve the transitory economic problem of scarcity and pose the new and more enduring problem of what to do with our leisure time. He said (Keynes 1930, p. 367~: Thus for the first time since his creation man will be Aced with his real, his permanent problem-how to use his freedom from pressing economic cares, how to occupy the leisure, which science and compound interest will have won for him, to live wisely and agreeably and well. Since that time, negative voices about the "bads" of economic growth have been louder, although the recent recession and inflation have given some pause to an overemphasis on bads. Critics of economic growth have asserted that it wastes national resources, damages the environ- ment, increases the congestion in cities, and even worsens the dist~ibu- tion of income. In the extreme, these views have questioned the goal of
280 PAPERS economic efficiency itself and have advocated zero economic growth (ZEG). In response to these positions, and for other reasons associated with the search far quantitative measures of economic and social perfor- mance, efforts to measure welfare have gone forward. With some per- missible oversimplification, these efforts can be said to fall within two frames of reference. Economists treat welfare in terms of basic choices that are more or less deliberately made by productive agents seeking some definable product, while other social scientists are concerned with "ultimate outcomes," which may or may not be deliberately sought, wanted, or intended. This distinction between choices and outcomes, while much too sharp for some purposes, is useful as can be seen by the two sets of analyses we have in mind. Economists view choice as a productive effort whose objective is to maximize utility. For example, will a person increase his total utility by working longer hours and thereby having more GNP-type goods, or by working less and having more leisure? This question is discussed in the section of the paper on hours of work and use of time. In more general terms, a person or family is viewed, in more recent statements of economic theory, as seeking to choose the most efficient combination: of market purchases of goods and services; of use of time at work to earn income to buy these goods; of time that must be spent in consuming these goods; and of time spent and income foregone acquiring knowledge and skills or "human capital" for application both in the workplace and in the home all of these to be allocated so as to obtain a maximum level of utility. This is discussed mainly in the section on consumption. In the case of social choice (say, the choice between "private" goods and "public" goods), economists ask, for example, whether social welfare would be enhanced by a given increase (however measured) in cleaner air, obtained at the cost of giving up some amount of GNP-type goods. This type of question is examined in the section dealing with the environment. In comparing one person's welfare with another's, can total social welfare be enhanced by bettering the condition of some at a cost to others? This is discussed in the section on income distribution. Given the fact that perfect economic organization is not possible (that is to say, demand and supply are never in perfect equilibrium and monopoly and institutional restrictions result in a less than optimum allocation of resources), how can welfare be enhanced with "second-best" solutions? This question comes up by implication in a few of the areas discussed later, but it is not treated explicitly.
Welfare Dimensions of Productivity Measurement 281 A particularly difficult but important dimension of welfare concerns the choice between present and future welfare. Many people or societies are willing to undergo some discomfort early in life in order to have greater satisfaction later on. This question is discussed mainly in the section on consumption and income distribution in connection with measures of lifetime income, but it is brought up in other sections of the paper as well. Social psychologists, political scientists, and sociologists have been concerned, in terms of welfare, with developing measures of social outcomes that relate to the health, status, satisfaction, and social stability of the population. Thus, they have worked in such areas as (1) objective social indicators, or time series that measure changes in health, in rates of crime and victimization, in economic security, and in changes in family structure; (2) subjective measures of satisfaction of people in the workplace, the home, the neighborhood, and in society generally as, for example, satisfaction with or trust in government; (3) specific measures of effectiveness, in the case of the workplace in providing for self-esteem or absence of stress, or of hospitals in promoting health; and finally (4) in measures of achievement of different groups, particularly dis- advantaged minorities, as measured by social mobility. We discuss these "outcome-type" questions in several sections of the paper but mainly in connection with quality of work, health facilities, hours of work and use of time, and in a final section on the feasibility of a system of social and economic accounts. In recent years, economists and other social scientists have crossed the disciplinary boundaries so often that no hard and fast line should be drawn between what the different practitioners are doing. Moreover, these questions are highly interrelated as will be noted in our discussions in each of the specified areas. We start with consumption. CONSUMPTION PRIVATE AND PUBLIC An otten-stated social goal of economic production, apart from what may actually motivate individual firms, is to provide goods and services to meet the needs and wants of the population. People and households or families purchase goods and services to meet their private wants (private or personal consumer expenditures) and also obtain goods in some collective setting (public goods). The national accounts measure the provision or production of these goods and services for consumers; they do not measure actual consumption. The illustrations given below help show that the objectives sought by a national accounts definition
282 PAPERS and by an actual consumption definition can best be served if the two are kept distinct, though interrelated, rather than merged into a com- bined measure. THE PRODUCTION BOUNDARY IN THE NATIONAL ACCOUNTS Many fine points about what constitutes the boundary of economic output are not discussed here: for example, the geographical boundary (whether the limits are U.S. territory or the output of all U.S. nationals wherever they reside), or the time boundary (whether the reference is to a year or to a longer or shorter period), or the different phasing or timing of different types of output and input. Our immediate concern is with "value" boundaries. The national accounting value system, in so far as it is relevant to welfare purposes, measures the value of goods and services in "real" or constant-price terms and measures the value of production rather than consumption. Under the national accounts definition of real output, a one-pound loaf of bread of given price delivered to a consumer during a chosen reference period counts, say, for one unit of output of bread. Another pound loaf selling for twice that price in the same period and market place counts for two units. Why the second loaf sells for twice the price of the first may be determined on the basis of the different characteristics of the two loaves. But whatever the reasons (the consumer perception of taste, nutrition, texture from the demand side, or resource costs from the supply side), the relative market price determines the real value of output of the given product at the given time and place. In measuring the change in real output from the chosen reference period to any other period, we determine how many more or less of these loaves were produced, remembering to count each type of loaf in proportion to its relative price in the base period. If a new loaf of bread, not previously marketed, is sold in some later period, its price relative to the old type must be determined. This process is often somewhat arbitrary, particularly if at the time of introduction of the new loaf, the old one is no longer produced. In any case, some judgment is made about what the relative prices would have been if the two types had been sold in the same market at the same time, say, in the reference base period. With much oversimplification this describes the meaning of real output and its change. In the aggregate, for bread and all other con- sumer goods and services, real output is so measured in the national income accounts.
Welfare Dimensions of Productivity Measurement 283 PRIVATE OR PERSONAL CONSUMPTION IN THE NATIONAL ACCOUNTS The national accounts provide detailed measures of the value of "personal consumption expenditures," but they are not directly concerned with consumption. ' Personal consumption expenditures are, strictly speaking, the market value of goods and services delivered by enterprises to eon- sumers. While the activities of consumers in buying and using the goods purchased (e.g., driving to the shopping center, cooking at home' are reflected in the accounts, say, through consumer purchase of fuels, the accounts are not concerned with the use by consumers of the goods and services they buy. Although the market value of the goods purchased may reflect con- sumer perceptions of how these goods may meet their basic needs and wants, the national accounts accept as data the quantities and prices of purchases. They are not required within their definitional limits to inquire into the nature of these basic needs and wants, nor into how well they are being met by these products. The application of productivity measurement to personal consumption expenditures indicates essentially the efficiency with which business enterprises produce and deliver goods to consumers. It is not concerned with how well these goods serve consumers' needs and wants or how efficiently consumers make use of these goods in satisfying their needs and wants. This would take us outside the production boundary. As an aside, it is interesting to note that any increase in the produc- tion capabilities of a producer's good that is not reflected in its relative price will count as an increase in future productivity after its installation and use. For a consumer good, on the other hand, future performance is not measured in the accounts except to the extent that relative price predicts future relative performance, because this is beyond the presently defined production boundary. Having briefly described the national accounting value system applied to consumption, we now discuss in more detail welfare issues and the measurement problems they pose. First, we emphasize that, apart from how well GNP-type goods actually meet the needs and wants of the population, a useful analysis of welfare issues can be undertaken within the bounds set by the economic accounts. iThe only measures of consumption in the national accounts are ~ goods consumed in the production process, including capital, materials, and fuels consumed in produc- tion.
284 PAPERS As an elementary proposition, it must be clear that productivity, viewed within the traditional production boundary, is a basic source of well-being. It is the primary source of increase in real compensation per labor-hour. The increase in all the physical, social, and psychological amenities of life may not necessarily be proportionate to the advance in real in- come, certainly not for everyone or at all stages of the life cycle, as will be discussed later. But choice available to everyone from this source can be jeopardized if productivity fails to advance. One of the most impor- tant economic concerns is the concern about inflation. Rising produc- tivity could be a significant weapon against secular inflation. If ways were found to advance the rate at which the real cost of scarce resources per unit of market output is reduced, long-run inflationary pressures could pari passe be lessened, unless these gains were offset by a corre- sponding increase in the rate of change in factor prices. A one percentage point reduction in the inflation rate at the present level of personal consumption expenditures amounts to a potential saving of roughly $13 billion, which is a large sum compared to the cost of any single present social welfare program. The consequences of diminished productivity, in the form of social tensions and economic distortions, cannot be elaborated here, though their importance for social stability must be considerable. The principle with which we are concerned, however, should be clear. Productivity advance helps consumers to make their welfare choices, and a decline in productivity hurts this process. The measurement issue comes to this: Whatever the process of consumer choice and the outcome of this choice, the input to this process of real income and of the goods and services chosen should be kept distinct and measured as best as we can within the economic production boundary. The area beyond the production boundary, which includes actual consumption, has many important dimensions, not as easily measured as production, which we now discuss. CONSUMPTION THEORY New formulations of consumer behavior are useful, at least in concept, in helping distinguish between output of GNP-type goods and consump- tion. According to the human capital version (~Michael and Becker 1973), people and families are viewed as production agents who produce satis- faction by using as inputs market goods and services purchased with their income, their own time at work and in the home or away from work, and their human capital, which not only provides skill and earnings
Welfare Dimensions of Productivity Measurement 285 on the job but knowledge useful in purchasing and using market com- modities. According to this theory, people and households make choices that can reveal their basic wants. Once these wants are revealed it is con- jectured that the ultimate structure of basic wants does not vary much among people in different circumstances. People do differ, of course, in their money incomes, in the prices and quantities of goods they pur- chase, in wage rates, in knowledge and skills, and in the family and social resources available to them. Because of these differences, people differ in their capability to produce basic satisfactions. And for any one person, large variations occur over the life cycle. Nevertheless, these forces constrain the choice of what goods to buy, how many hours to work for wages, how many hours to devote to consumption and house- hold production (including eating, cooking, cleaning, child rearing, recreation, and so forth) and how much to invest in future needs. Thus far, this theory has not sought or been able to develop a list of basic wants and needs. It has worked only selectively with a few areas of major choice such as marriage, labor force participation, health main- tenance, the number and rearing of children, travel expenditures, invest- ment in education, and migration. In any event, within this framework, income earned from production and the goods and services purchased in the market are separate elements in a welfare or utility function and are defined essentially as in the national economic accounts. QUALITY OF CONSUMER PRODUCTS Goods and services have many characteristics pertinent to well-being, but only some of them can be used to explain relative price in a market setting. Horsepower or the presence of automatic transmissions in automobiles, for example, accounts for an important part of the dif- ference in prices of models of a given make at a given time. Some makes have higher prices than others because they are perceived, correctly or not, to be more durable. But relative repair experience and other mea- sures of later performance of different makes are often impossible to predict, and the relative price at the time of purchase is rarely a guide to such differences in subsequent performance. As an aside, it may be noted that the welfare costs of "recalls," notoriously frequent at this writing, present a challenge to analysis, particularly when it is recog- nized that luckless purchasers of "lemons" are better off with than without recalls. In general, there are classes of outcomes in the process of consump
286 PAPERS tion that require different quantitative evaluation outside of, or different from, the national accounting measuring rod of quantity and price. These are cases in which some objective function for various products is specified and where the quantities and values enter as elements in the objective function. Such outcomes would include nutrition (say, in pro- tein calories of beef), decreased mortality in the case of certain types of medical care, and durability of consumer goods, not only of "durables," but also of"nondurables" such as men's shirts. Apart from performance that occurs after delivery, there are many cases in which new design and performance characteristics are specified and evident, but relative price does not appear to reflect these in any proportionate degree. Indeed, often the relative price may decrease even though performance by some measure increases. A dramatic case in point is cited in Chapter 5 of the Panel's report of electronic calculators whose capacity and speed have increased manyfold while their relative price has declined. It is useful and important for measures of consumer well-being to reflect changes in the quality of consumption that are not reflected in relative prices. A price index measure of constant utility, though diff~- cult to compile and interpret, would be useful to develop, at least on an experimental basis. Such measures would be helpful in analyzing con- sumer behavior particularly under inflation (e.g., extent to which sub- stitute commodities are sought) and producer behavior (e.g., ways in which commodity variation occurs under rising costs). Such measures would be useful to relate to or to supplement measures of market prices of goods produced. CONSUMER KNOWLEDGE AND EFFICIENCY A consumer with more information about sources for the goods he seeks, relative quality and prices of competing brands, and so forth, is potentially better off than one with less information but with the same income and time to choose. The more knowledgeable person or family may obtain the same bundle of satisfactions with a smaller outlay. A knowledgeable consumer is more efficient in both searching for and using products with benefits to society as a whole as well as to himself. An important current example is consumer use of energy. Society and consumers would benefit from more knowledge about cost-effective energy use in the home (Hirst and Carney 1978) and on the highway. While knowledge is treated here as contributing to individual welfare via private consumption, it is also a public good.
Welfare Dimensions of Productivity Measurement CONSUMER AND SOCLAL COSTS 287 In assessing costs beyond the national accounting boundary, we should recognize costs generated by consumers themselves. A case in point is the costs to consumers and to society of the use of alcohol, other drugs, and tobacco. Estimates of the total direct costs of illness, of motor vehicle accidents, and of earnings foregone in consequence of excessive use of these substances are not available, but the figure must be a large one. The national income, oddly enough, increases as a consequence of these expenses, including the additional expenses arising from preven- tive or curative measures such as drug treatment centers, alcohol abuse centers, etc. (this point is further discussed later in the section on health facilities). As far as welfare measurement is concerned, it is important to determine these costs to help society decide the amount of resources to be allocated toward their reduction. Also, analyses of the effectiveness of the preventive or curative measures should be part of the wider study of productivity beyond the national accounting boundary. PUBLIC GOODS OR COLLECTIVE CONSUMPTION Several issues pertinent to this section are discussed later in the section entitled Environment. At this point, we mention briefly two significant types of public goods that may require measurement beyond the estab- lished production boundary: public safety and defense outlays. Public Safety The national income accounts, under general government, measure the activity of the criminal justice system by counting the compensation paid to police, judges, wardens, parole officers, etc. Such a cost measure could be improved if we could count the activities actually performed and weight them by their costs. But even so, these accounts do not and should not measure the output of public safety. The national accounts appropriately provide a basis for measuring the resources allocated to promoting public safety in the community and in industry. But the public safety itself must be measured separately outside the national income boundary. A measure of public safety might include the changing probability over time or differences in likelihood among cities of losing property, limb, or life from criminal activity. As a corollary measure, the proba
288 PAPERS bility might be calculated that people in various circumstances will commit crimes of a given type. The choice of weights for combining offenses of different types or damages to victims of different degree would imply the meaning given to the measure of public safety. For example, the weights chosen might be proportionate to the cost to vic- tims, including an estimate of the value of a human life. Any such measure of public safety would not, of course, measure the effectiveness of the criminal justice system per se, since crime and vic- timization are affected by forces considerably beyond those of the arrest, punishment, rehabilitation, and other efforts now included in that system. Nevertheless, good measures are needed of the effectiveness of the criminal justice system in making a net contribution to crime reduc- tion. For some people, particularly those with a high probability of victimization, the prevention of crime is perceived as the most important welfare problem of all. We cannot go into the details of measuring the effectiveness of different modes of dealing with offenders, for example, whether a community treatment program is more cost effective than a jail sentence in preventing recidivism of a first offender. Our point is that the national accounting measure of criminal justice activity should not be adjusted to include some imputed contribution to safety. This does not preclude continuous evaluation, however difficult, of the effec- tiveness of the system, and such evaluation efforts should be encouraged. We mention in passing what appears to be an anomaly in the national accounts. Expenditures by private industry for protection against theft, etc., are intermediate costs, increases in which result in declines in mea- sured output per unit of input. Increased expenditures by government for public safety equipment and similar expenditures and by consumers for home burglar alarms, etc., are increases in final output. This pro- cedure is appropriate for national accounting purposes in distinguishing between marketed output purchased by final consumers and the costs of producing such output, which include safety outlays by business. But this procedure would need modification in a wider reckoning of costs of producing public safety regardless of which sector incurred the costs. Defense Outlays The question whether to count defense outlays as intermediate costs or final products is an old one in national accounting literature, but was given a new twist by Nordhaus and Tobin in their effort to construct a measure of economic welfare (MEW) (Nordhaus and Tobin 1972~. In this study, the authors sought a measure of consumption rather than of production. For the period 1929-1965, they determined that because the
Welfare Dimensions of Productivity Measurement 289 national security (in their view) had not increased, the "consumption" contribution of defense outlays was zero. This is not a fruitful approach to welfare analysis in my view. I do not consider it appropriate to measure welfare by imputing a consumption value to a measure of production. Viewed as a welfare function, at least in concept, national security depends on several factors, including the defense outlays of the United States and those of other countries, U.S. foreign policy and that of other countries, and the state of the U.S. economy, including its productivity. It is not clear to me what the response of other countries, say the U.S.S.R. would be were we to reduce our defense expenditures radically. Even if Nordhaus and Tobin were correct in their view that national security was no greater in 1965 than in 1929, it would surely not be appropriate to assume therefore that the net contribution of defense outlays to security had been zero. Whatever the degree of national security at any given time, it makes little sense for national accountants or their welfare-minded cousins simply to alter GNP in order to measure it. INCOME DISTRIBUTION Among the most important welfare outcomes of the economic system is the distribution of rewards and damages, including income from eco- nomic production, income from transfers, other benefits that are inci- dental or unintended, and damages not necessarily intended by eco- nomic agents. THE WELFARE SYSTEM Advance in economic productivity is a means of improving living con- ditions for everyone. In its absence, increased income for some can only come at the expense of reduced income for others. The way the term welfare is most commonly used today, in connection with our welfare or transfer system, makes it evident that increases in economic productivity are a crucial element in social welfare policy. The welfare of the dependent population, notably the children, the infirm, the retired, and the poor, is vitally affected by the productivity of those who produce the nation's food, clothing, housing, medical care, and other GNP-type goods. The dependency ratio is often calculated as the ratio of the number of dependents (various classes of nonworkers) to the number of maintainers, or workers. However, each maintainer should not be counted without change over time as one person. This is because changes occur in the relative importance of high- and low
290 PAPERS productive employees, hours of work per person and in output per labor- hour. The last mentioned factor has been the most important over a long period of years. Output per hour of workers doubled in the 25-year period 1947-1972. While the number of dependents increased, as did their benefits, the "ability" of a worker to support these welfare benefits also increased. Only through further increases in productivity, trans- lated into increases in real compensation per hour, can these transfers to the dependent population be increased in real terms, without reducing the real incomes of the working population. The slowdown in recent years in the trend rate of productivity growth gives rise to added social concern in the context of the problems facing the U.S. welfare system. The importance of productivity in the real financing (as distinct from fiscal or monetary financing) of welfare policy is clear. However, pro- ductivity is a two-way street. While increased productivity supports increases in welfare benefits, increases in the burden of income or pay- roll taxes to support welfare benefits affect the productivity of workers, either by affecting working time (e.g., early retirement) or the quality of effort, or through the alleged fiscal effects of social security financing in retarding the growth of capital investment (Feldstein 1975~. A full economic analysis of the impact of the transfer and tax system on eco- nomic productivity is still not at hand. (For an interesting philosophic discussion, see Okun 1975; see also Fabricant 1969.) Viewed in the context of such welfare considerations, the definition and measurement of production for income distribution purposes should be within the customary economic boundary. In going beyond this example, we must recognize that the subject of income distribution is a vast one. The relation of income distribution to productivity has many interesting facets, particularly that of the extent to which the marginal productivity of the factors of production deter- mines how increases in the national income are divided between labor and capital and among workers. Obviously, we cannot explore all these questions. We touch briefly on one aspect, that of wage-rate differentials. DO WAGE DIFFERENCES MEASURE DIFFERENCES IN PRODUCTIVITY AMONG WORKERS? This question has several dimensions including changes in productivity with age and differences in productivity between men and women and whites and nonwhites. Cutting across all these is the matter of differences among workers in their effort per working hour. The measurement issue here involves an apparent conflict between welfare considerations and national accounting criteria.
Welfare Dimensions of Productivity Measurement 291 In many productivity studies, different "qualities of labor" are dif- ferently weighted. Usually the wage rate in the base period measures differences among workers in their levels of productivity per hour. This assumes that output per hour is the same among workers with the same wage rate; that is, that there are no differences in the amount or quality of effort per hour. This procedure is not unreasonable when comparing large differences, say, in comparing a skilled machinist with an unskilled custodial worker. From a welfare point of view, however, it is important to recognize and assess the importance of differences between wage-rate measures and measures of actual (or potential) pro- ductivity that exist both within and among occupations, whether be- cause of age, sex, race, or other factors. This topic extends well beyond the economic boundary set by the national accounts in that the forces that result in discrimination between workers in their opportunities for productive work, while focused in the workplace, emanate from wider social conditions. The literature on forces that determine differences in personal earnings is vast (Juster 1976, Mincer 1970, 1974, Sahota 1978, Smith 1975~. Many of these forces may be viewed in a measurement context that extends the dimension of income from a current basis to a lifetime basis. Issues of income distribution are being increasingly viewed in a lifetime context partly because a lifetime measure refers to the long-run, total, or permanent economic welfare of a person rather than his transitory welfare. This is not to downgrade the importance for welfare policy of transitory conditions such as the temporary loss of a job or temporary disability, but rather to recognize that many of the concerns about relieving poverty are of a long-run nature, involving the need for a higher level and more stable growth rate of lifetime income. To what extent this growth in lifetime income requires improvement in basic skills, increases in the number of "good jobs" with promising career opportunities, or other social changes affecting the economic opportunity structure, is a question that continues to challenge social welfare policy (Moss 19784. There are other questions of income distribution involving extensions in the income concept beyond that in the national income accounts, which we do not explore, but list below: 1. Whether money income should be extended to include realized capital gains. 2. The measurement of the distribution of income from wealth or the distribution of wealth itself. 3. The identification of gainers and losers from inflation.
292 PAPERS 4. The distribution of environmental damages. In the section of this paper entitled "Environment" we discuss differences among parts of the country and groups of people in the impact of environmental damage. The above four topics are discussed in Juster (1976), Sahota (1978), and Smith (1975~. QUALITY OF WORKING LIFE Briefly, the difference between a strictly production view of working conditions and a welfare view is that the first regards workers and their working conditions as economic resources or factor inputs to be allocated as efficiently as possible; the second regards working conditions as contributing to the quality of workers' lives by enhancing economic security and a sense of well-being at given jobs. As Seashore has aptly stated, the first view is "what (workers) can give to their jobs"; the second is "what jobs can give to the workers" (Seashore 1977~. The two viewpoints are not necessarily inconsistent but conflicts do arise. We discuss each view separately to suggest the relevant welfare dimensions. PRODUCTION CRITERIA AND WORKING C ONDITION S In the production view, workers and their tasks are to be allocated so that labor costs per unit of output can be minimized and/or so that output per worker-hour can be maximized. As a production resource, worker skills must be maintained or improved, and the marginal costs of recruiting, training, and retraining of workers should not exceed marginal productivity over some meaningful accounting period. Added costs or diminished output arising from wasted effort, dishonesty, excess absenteeism, or turnover and the like must also be kept as low as eco- nomically feasible. Significant and enduring improvements in working conditions have occurred and proved viable because of certain economic advantages. Programs to improve conditions of the working environment may lead to a reduction in the firm's unit labor costs either because of increased productivity or decreased compensation per hour. Productivity may increase if wasted effort is reduced by improved design of jobs and better management and working conditions favored by workers. (As will be noted below, research to determine the relationship between produc- tivity and worker satisfaction is still inconclusive.) A decrease in labor costs can occur where employers offering better working conditions can
Welfare Dimensions of Productivity Measurement 293 attract better workers, recruit more economically, and have less turnover than competing employers. The historical forces that have determined the distribution of in- creases in productivity among higher wages, better fringes, lower hours, 2 more vacations, or in improving health, safety, and work conditions generally have reflected in good part employer and employee perceptions of the relative advantages of these gains. Government regulation and union pressures have also played a role. Nevertheless, regardless of the complex interplay of employer and employee initiative, and pressure from unions and society, significant and enduring improvements in the quality of working life have come about and will continue to come about largely, if not solely, through continued increases in economic produc- tivity as defined in the strict sense. For a discussion of such historical forces, see Rees (1973) and Ginzberg (1975~. We discuss further the economic issue of the relationship between improved working conditions and productivity. Improvements may be costly. Do they pay in increased productivity? The literature on this subject is vast and growing (e.g., Katzell, Yankelovitch, et al. 1975~. Experiments on new modes of work are being tried in the United States and abroad, including Yugoslavia, Israel, West Germany, The Nether- lands, and Sweden. These efforts are being undertaken with mixed motives, but a significant one is to seek ways of increasing economic efficiency by improving the economic, or physical, organizational, psychological, physiological dimensions of worker welfare. Three com- ments are in order here on this complicated subject. First, these efforts appear to have contributed to improving our under- standing of the diversity of dimensions of worker satisfaction and its dependence on various factors in the workplace. This is important and should be encouraged. Second, on the specific issue of the contribution of improvements in working conditions to increasing productivity, the results are incon- clusive (Katzell, Yankelovitch, et al. 1975, p. 23 et seq.~. This is not surprising. Measuring or seeking to enhance worker satisfaction is most difficult in view of its many highly interrelated dimensions: economic, including adequate pay, long-term security, upward mobility; psy- chological, including among others, the measurement of satisfaction itself, and other variables such as self-esteem, work motivation, mo- notony, stress, anxiety, depression, sense of participation; physiological, including detailed concerns about and measurement of cardiovascular, 2We do not discuss in this section the important question of hours of work but discuss the matter separately in the section on hours of work and use of time.
294 PAPERS endocrine, visual, and other bodily functions; and organizational, including plans for worker ownership, representation in management, rearrangement of tasks or of work modules; and other physical changes to promote health, safety, recreational, and other benefits or amenities of the workplace. Testing these variables separately, in combination, or as affecting worker satisfaction (which has been relatively successful), and then determining their effect on productivity (which has not) is a most exacting scientific task i whose difficulty is compounded further by differences among workers in their needs and wants in the work- place. The task is hard even to begin, given the problems of cooperation of management, unions, and workers themselves. It is a tribute to social scientists as well as management, unions, and workers that such efforts do go forward and produce beneficial results (Kahn 1960 and individual papers by Walton, Seashore, Sheppard, Lawler, Dyer and Hoffenberg, and Goodale in Davis and Cherns 1975~. A third comment concerns the specification of the productivity vari- able itself. It is possible that some of the inconclusiveness of attempts to determine the relationship between improvement in working conditions and productivity results from a poor specification of the productivity variable. Productivity is often defined in terms of fiscal success such as profitability, which depends on market prices and demand changes as well as on real output per unit of input. Improvements have been occur- ring in real cost accounting at the level of the firm, that is, in deter- mining labor costs per unit of real output as well as of unit capital and material costs. These methods should be used more precisely in studies relating productivity to job redesign. BEYOND ECONOMIC PRODUCTION CRITERIA Strictly speaking, we cannot go beyond economic criteria because any improvements in working conditions that threaten the viability of a firm are not likely to be instituted or retained. Nevertheless, the use only of economic criteria can have adverse welfare consequences, such as when they result in the closing of a plant and the terminations of many jobs. Studies by the University of Michigan's Institute for Social Research have reported serious damage to the mental and physical well-being of workers experiencing job loss, as compared with control groups of workers in comparable settings in plants that were in no danger of shutting down (Cobb and Kasl 19771. The detailed observations made in these studies showed that the plant closings took a great toll not only in economic terms, but also in terms of psychological and physiological health.
Welfare Dimensions of Productivity Measurement 295 Episodes of this nature give rise to serious criticisms of the national income accounts when they are mistakenly taken as literal measures of welfare. In these circumstances, apart from the fall in output oc- casioned by the plant closing, the national income could rise to the extent that expenditures for medical care or funeral services increased. The tragic consequences of plant closings could, if the national accounts stood alone as a measure of welfare, show up as a rise in the national well-being. Moreover, when employers take measures to cushion the effects of possible shut-downs, or in general make expenditures to safe- guard the health and safety of their workers, the cost of output as mea- sured in the accounts rises and measured productivity falls. Thus, the accounts appear to be perverse; they count benefits to workers as costs, and costs to workers as benefits. Such views may have journalistic value, but they have little scientific merit. The position taken in this paper is that the national accounts, in providing a systematic basis for itemizing such production transactions as employer costs for health and safety or worker expenditures for medical care, are highly useful in determining how scarce resources are allocated to meet human needs. They do not specify or judge these needs nor do they evaluate how well they are met. ILLUSTRATIVE STUDIES ON WORKER WELFARE To avoid misunderstanding, it should be said that it is not among the objectives of this paper to discourage studies in which values are im- puted to entries in the national accounts for the effects of production on the health, safety, and general sense of well-being of workers. The point here is to make clear that a measure so obtained is different from and not a substitute for a measure of production. But the separate study of desirable and unwanted outcomes in the workplace and of experi- ments in the job changes to foster desirable outcomes needs to be en- couraged for its own sake, since working life constitutes such a major part of life in general. Such studies of outcomes beyond the production boundary as presently defined are growing rapidly. I list several here: 1. Among economic outcomes, the added dimension of economic security on a lifetime basis to point up more clearly changes in the long- term opportunity structure of working conditions (Moss 19781. 2. Measures of absenteeism, job turnover, and differential effects of effort per hour as indications of the degree of attachment of workers to jobs or of motivation in the workplace (Wool 19731.
296 PAPERS 3. Studies that seek to determine under what circumstances improve- ments in working conditions can lead to significant and enduring in- creases in productivity. 4. Studies of psychological outcomes of work conditions to determine effects of work on self-esteem, interest in the job, and the risk of mental distress, and the effect of these upon work performance (see Kahn 19724. 5. Analysis of the extent, nature, and changes in jobs (and of workers in such jobs) that characteristically have unwholesome working con- ditions. These jobs typically pay low wages, generally carry the least social prestige, require little skill, and provide little or no expectation of stable employment or opportunity for a promising career. Physical conditions are often harsh, meeting only minimum government require- ments for health and salty. Such jobs are often found in marginal industries where employment fluctuates with seasonal and other changes in demand. Their lifetime income profile shows little or no rise with age, or declines earlier in life than those of more stable occupations. Although the relative number of poor jobs seem to be declining, it is apparently very difficult to eliminate them entirely. One important reason, according to M. Piore, is that industries or occupations in which productivity is high, and working conditions are relatively secure and salubrious, achieve such good performance in part because they are able to transfer excess demand to marginal firms in times of peak activ- ity. As long as there are firms that absorb the large instabilities and unpredictabilities of economic activity, and as a result provide relatively few amenities in the workplace, the other firms, free of such costs, can more easily afford the amenities they do provide to their more fortunate work force (Piore 19781. HOURS OF WORK AND USE OF TIME An obvious welfare dimension of productivity advance, defined in its strict economic sense, is that the increase in the national product over the years has been achieved with people working fewer and fewer hours per unit of output. This has meant that workers could choose more freely between work and leisure. From a production point of view, hours of work, or the length of the workweek or workyear, constitute a dimension of input. The production question is concerned with the effect on output of a change in hours, with the effect on total output of changes in total labor-hours, as deter- mined by changes in the number employed, and with the effect on
Welfare Dimensions of Productivity Measurement 297 output per labor-hour of changes in the number of hours worked per person. We discuss mainly the latter effect. The welfare implications of this matter are considerable. Increases in output per labor-hour that can be achieved by an optimal choice of working hours, or thwarted by less than optimal choices, can have significant effects on economic and social well-being. The historic trend toward shorter hours or shorter workweeks, and increases in time off for holidays and vacations with pay, has been accompanied by increases in output per labor-hour. But it is not clear how much is really known about this relationship either about the effects on output or on real income of shorter hours or about the value placed by workers on added leisure. Determining the production effects of shorter hours is a difficult and challenging task. It must take into account not only the effect of les- sening of fatigue, but many other motivational, managerial, and group behavioral variables. It also involves differences among industries and occupations. Differing worktime arrangements are important. Experi- ments with new time arrangements have included staggered hours, 4-day workweeks, flextime, sabbaticals, and early retirement. This last factor, early retirement, opens up an entirely new dimension to the study of worktime. Just what the impact on productivity would be of changing, reducing, or increasing the length of total working life is a question of increasing economic and social importance. In sum, while the technical and other difficulties of determining the economic productivity implications of shorter hours are enormous, the economic welfare implications of shorter hours could be considerable. Continued and perhaps more systematic and comprehensive research would be in the national interest. Determining the welfare dimension or value to put on time spent away from work is in its infancy, particularly on the empirical side. The theoretical work has given rise to a large and growing body of literature (Becker 1975; see also the discussion of the Nordhaus and Tobin valua- tion of leisure time in Moss 1973~. We touch on some of the main ele- ments of this theory and then mention what we hope will prove to be an important empirical development. Consider a person dividing a fixed endowment of time among paid work, consumption, including sleep, non-market productive activities, and investment. At the current standard workweek, a man spends less than a fifth of his lifetime on the average in paid working hours. A much smaller fraction of time is spent on paid work by women, al- though that fraction has been rising rapidly, particularly for married
298 PAPERS women. For men, the fraction has been falling as a result of earlier retirement, though that trend may halt or be reversed in future years as the rate of labor force growth diminishes. The contribution of paid work or income to the total value of time spent by a person or family is necessarily a highly significant one. lIow- ever, the determination of the value to a person of some small increase in nonpaid work time at the expense of a corresponding decrease in work time is at the heart of the problem we are discussing. Recent formulations of the theory of consumer behavior have sought to develop a production function for the household in which the outputs are ultimate satisfactions and the inputs are time, human capital, market income, and quantities and prices of market goods. This theory attempts to formulate the problem of choice between how many hours to work for wages and how many hours to devote to consumption and household production (including eating, cooking, cleaning, child rearing, recreation) and to investment in future needs. In principle, this formula- tion could provide a basis for determining the relative price in welfare terms of an hour of work and an hour of nonwork. Thus far, however, the empirical results have not moved much beyond the assumption that some variant of the prevailing wage rate is the basis for the exchange. The important empirical work growing out of this theoretical framework has dealt with several areas of major choice on a selective basis, including when and whether to marry, the labor force participation rate, health expenditures, having and rearing children, travel expenditures, invest- ment in education and migration. These studies tend to underscore the importance of current and expected future income (the importance of human capital, including education and skills) in the allocation of time. They do not yet provide a basis for decision making in determining the welfare gains or losses from changing the standard workweek. An empirical investigation that we hope will be of major significance is the gathering and analysis of data on actual time use, now being pursued at the University of Michigan under the direction of F. T. Juster. This work, based partly on diaries and on interviews, seeks to determine how people with different characteristics and in different circumstances actually divide a fixed endowment of time. This time- budget inquiry not only divides time spent among various activities but also determines for each activity how much is intermediate (onerous), consumption (satisfying, per se), and investment (activity for future consumption). Thus far, the total time to be divided is in terms of a year, not a lifetime. Work on the lifetime dimension, in which this writer hopes to participate, is planned for the future. The analysis of
Welfare Dimensions of Productivity Measurement 299 time use of parents and children in a family setting has seminal pos- sibilities for studying family behavior and outcomes within and beyond the family. HEALTH FACILITIES One of the most controversial areas in which a distinction can be drawn between production and welfare or consumption objectives is in connec- tion with health facilities, including pharmaceutical companies, hospitals, and medical services generally. In the production view, these facilities, like any other economic facilities, use labor, equipment and plant, and materials and fuels to deliver goods and services to consumers. In real terms, these goods and services are measured as quantities of product, each valued by their market price. The definition of these quantities often becomes troublesome, and questions are frequently raised about the imperfections of the market in determining their relative prices. In the case of drugs and medicines, quantities are fairly easy to dis- tinguish. Their relative prices, however, often given rise to serious ques- tioning of the workings of the market. For example, public debate often becomes intense over what appears to be unjustified differences in the prices of identical drugs, such as between a brand name and a generic version of the same drug, or between different brands of aspirin. It the price of brand ~ of a given medicine is twice that of brand B. then in the national accounts each unit of brand A counts as twice as much output as a unit of brand B. even though they may be identical in their use. But the national accounts do not measure use, they measure only what is produced and relative prices. Of course, production responds to increases in scientific and consumer information, notwithstanding the slowness of this response as seen by advocates of change. The banning of thalidomide is a dramatic instance in which the eventual reaction to the tragic consequences of the use of a drug has stopped its production. In the case of health services, including visits to a doctor s office or a stay in a hospital, quantities of service are often difficult to measure. Prom a national accounting point of view, these quantities should be defined as closely as possible in terms of what the production agents are doing, that is to say, what doctors, nurses, orderlies, and administrators are engaged in: performing health examinations, surgeries, and other services. Often the problem in such areas as health services is that he do not in fact know what the agents are doing. What they are paid to do may not be done or may be done with more or less effort per hour. If,
300 PAPERS for example, patients are more frequently left unattended in one hos- pital than another, or relative to some past period, we may never know whether the hospital was in fact less productive. This is so even in the strict national accounting sense, i.e., even apart from the welfare ques- tion of whether what such personnel are being paid to do is most con- ducive to the health of patients. Thus, while the detailed quantities of such services very often cannot be specified with available data, the concept of production in the national accounting sense should be clear. Production efficiency criteria require that goods and services in health facilities be produced at least cost per unit and that these unit costs decrease over time. To the extent these conditions hold, economic wel- fare is enhanced. Social welfare criteria, however, raise a different question: How effec- tive are these goods and services in promoting health? Measured over many years, there is little question that the improvement and increased availability of medical services and of medicines have had a major impact on decreasing mortality rates, particularly mortality at birth, both of infants and mothers. For risks at older ages, particularly the ages above 60, death rates at each specific age have also declined while the increase in medical services of all types, as a result of medicate for the elderly, has been enormous. Increased health insurance and medicaid has greatly increased the volume of medical services used by all age groups, and prices and costs of medical care have risen faster than those of nearly all other goods and services. We may be on the threshold of a new dimension in health care, should a national health insurance plan be adopted. An important policy question that helps contrast the social welfare and the produc- tion points of view is: What is the most effective form of health organi- zation? A proposed solution to rising medical costs has been to encour- age the growth of health maintenance organizations (HMOS), which agree to provide comprehensive medical services for a fixed annual payment for each person served, in contrast to the more prevalent practice charging a fee for each service performed. Several arguments are presented in favor of the HMOS. The HMO has an economic incentive to prevent illness because its returns are greater when its members are healthy than when they are sick; it also has an incentive to avoid prescribing too many expensive services, since it cannot charge individually for them. Various other economies associated with group practice are cited by advocates, such as more efficient record keeping and improved budgeting for services that can more easily be predicted under this form of medical organization. Note that many of
Welfare Dimensions of Productivity Measurement 301 the arguments are essentially economic in nature. But the overriding test of effectiveness is whether the members of HMOS are in fact maintained in better health per dollar of cost than people served by other health facilities. There is still debate on this question, and we cannot hope to resolve it here, except to highlight the distinction between effectiveness in promoting welfare objectives and economic efficiency in the produc- tion of the services of health facilities. The task of measuring the effect of health services on people's health is a difficult one, largely because of the importance of the health condi- tions and health habits prior to and during treatment. This issue is dis- cussed in Scott (1977 and in this volume), in which longevity is used as a measure of health output. This brief treatment of a large subject mentions one other issue, a dimension of input. A number of studies (e.g., Fuchs 1975) have pointed to the critical importance to health of health habits. To such observers, perhaps the least costly way to improve health is for people of nearly all ages to adopt programs of balanced and moderate diets, to take appro- priate and regular exercise, to avoid excess use of alcohol and harmful drugs, and in some way to succeed in reducing mental stress. How health outcomes in the welfare sense should eventually be mea- sured depends on the objectives sought. It is possible that a more sophis- ticated measure may be developed in this field, using some medically agreed upon vitality index based on a weighted average of items in health examinations. The choice of weights would signify how much each component of the index contributed to some basic concept of total health, say, to the chances of illness-free days of life. Whatever the metric chosen, however, it should be distinct from, and not a substitute for, the more restricted measure of production of health services. In the provision of health care, wise use of the production and welfare measures in relation to each other should contribute to economic and social welfare. THE ENVIRONMENT The healthful properties of air, water, and land are affected not only by economic production (industrial wastes) but by consumption (auto and household emissions, household pets, recreational activities, e.g., power boating), and by nature itself. Pollution damages not only health but also property. It is clear that a welfare function in this domain is highly complex, quite apart from the question of how to set standards of
302 PAPERS quality, say, in terms of the oxygen content of water or the percentage of the most harmful emissions in the air. Within this complex domain, measurement of the production boundary as set by the national accounts has proved useful to environmental welfare concerns, as illustrated in the two following studies. Leontief has introduced applications of the input-output accounts for calculating the generation of pollutants and of the abatement costs (Leontief 1970, 1973, Herfindahl and Kneese 19734. Calculations can be made from this approach of changes in pollutants resulting directly and indirectly from change in the volume or composition of final prod- uct (GNP), or the net increase in abatement production needed to achieve a specified reduction in pollutants while holding GNP constant. Another important use of the measured production boundary pertinent to welfare analysis is the evaluation of the effect on measured output of new requirements to protect the physical environment against pollution (Denison 1978~. In addition to measuring costs associated with improving the physical environment, this study also calculated the costs incurred by business to protect the safety and health of workers, and those de- signed to reduce losses from crimes committed against business enter- prises. Denison determined that measured national product would have been 1 percent higher in 1975 had the new requirements for pollution abatement control introduced from 1967 to 1975 not been adopted. This type of calculation is a necessary ingredient in the analysis of economic and social welfare because it determines what the nation has to give up in the way of GNP goods and services to achieve a particular social welfare objective. WELFARE EFFECTS The Denison study says essentially that if an electric power company installs scrubbers in its smoke stacks to reduce toxic emissions, the opportunity cost per kilowatt hour of generated electricity is increased by the cost of those scrubbers. Going beyond the measured boundary would entail two conceivable steps, one a relatively short step and the other a great leap, perhaps into the unknown. The short step would determine the effectiveness of these scrubbers in cutting down on the toxic emissions by measuring the difference in the volume of emissions between stacks with and without scrubbers. The leap would be to try to determine the net effect of the new scrubbers on health of people in the area previously contaminated by the smoke stacks. Taking the shorter step, in fact, is part of a program planned by the Bureau of Economic
Welfare Dimensions of Productivity Measurement 303 Analysis employing data in selected cases from Census Bureau surveys of pollution abatement control expenditures by manufacturing establish- ments. An example of an attempt to take the greater step is an analysis of the relation between mortality and air pollution (Lave and Seskin 1973, Wyzga 1978~. See also Stone (1972) for an analysis of the problem of determining the point at which the gain from devoting a little more resources to pollution control is balanced by the loss of goods and ser- vices foregone in the effort. DISTRIBUTION EFFECTS Much environmental pollution is concentrated in particular areas of the country. Another welfare dimension, not captured in the studies cited above, is where these impacts occur and upon whom. One reason for examining distributional impacts from a welfare productivity point of view is to identify some of the worst cases and thereby improve the target efficiency of abatement measures. We cannot survey all pertinent studies in this field, but we mention one dealing with air pollution to illustrate the distributional dimension. This study (Gianessi et al. 1977) estimated the dollar value per capita of air pollution damages to health and property from major pollutants from some 149 geographic areas, called county groups, which could specify important SMSAS and cities as well as certain rural areas. A total national value of pollution damage of approximately $20 billion for 1968 in 1970 dollars was distributed in this study among geographic areas. The study brought out the great unevenness of impact by area as well as differences among areas caused by different types and sources of pollutants. In many rural areas, for example, nature is the dominant polluter. The study concluded that air pollution damage does not impact more severely on the poor than the rich; but this finding has been questioned, largely because the measured income differences across the areas mask significant unmeasured dif- ferences within them. This section of the paper logically should treat the general problem of externalities, both "goods" and "bade.'. One candidate for inclusion is knowledge or technological know-how, which was discussed earlier as a private good under consumption. But it is also a public good that is an outcome of more than the "knowledge industry" (schools, research organizations). The subject of knowledge as a public good or as "social capital," which unlike physical capital does not depreciate with use but rather appreciates with productivity-increasing dissemination, cannot be elaborated on here.
304 THE FEASIBILITY OF A SYSTEM OF SOCIAL AND ECONOMIC ACCOUNTS PAPERS Efforts have been going forward in a number of directions to measure quality of life, compiling social indicators, and to develop an integrated system of social and economic data. None of these efforts presently provides for a measure of productivity. They are nevertheless relevant to this paper because of their concern with a data system for measuring outcomes in society that go beyond, but still relate to, the national eco- nomic accounts. It is not feasible to review all of these efforts in this paper. Our main concern is with the data systems. By way of summary, the following main Matures or objectives of these efforts are highlighted. 1. They seek to be an accounting system in somewhat the same spirit as the economic accounts, in the sense of accounting for all activities without duplication. Thus, they seek to account for the total population in a number of statuses such as age, sex, household status, socio- economic status (occupation, income, education, and other dimen- sions), position in certain life sequences (such as in school, at work, in retirement), and various other measures of behavior and circumstance. The system for accomplishing this (in a more systematic fashion than a mere listing of the usual tables in a statistical abstract) is to seek an integration of data largely at a microlevel so that information gathered from different sources on, say, health, education, and financial security can be brought together in a consistent and integrated fashion (Stone 1973, United Nations 1978, 1975~. 2. These efforts seek to achieve an outcome measure. Just as the national accounts seek to measure final real output of the economy, the work on social accounting has sought to develop outcome measures for society as a whole. These outcome measures are presently in the form of social indicators covering areas of social concern such as health, public safety, educational attainment, social mobility (or its converse), social disadvantage, and other measures of well- or ill-being. The work on social indicators per se has helped in seeking definitions of social concerns, but it falls short of any in-depth analyses of well- being and of any integrated view of social performance (Organization for Economic Cooperation and Development 1976~. 3. These efforts also seek a numeraire. The national accounts have been held up as a model of social reporting in part because of the simple numeraire of money by which all economic transactions may be mea- sured, combined, and related. The social accounts do not presently have
Welfare Dimensions of Productivity Measurement 305 nearly that simple a numeraire, but work proceeds on two possibilities: one concerns the number of the population and the other the allocation of time. Both represent defined boundaries and measurable quantities. 4. Finally, these efforts seek a generalized system that may have the makings of a social framework if not a theory of social change. The national accounting system is grounded in the economic theory of balancing income payments, product flows, inputs and outputs, as well as saving, spending, and investment. No such theory is presently available in the field of social accounting. Nevertheless, some effort is going forward that has the possibility of defining elements of interest to the measurement of productivity in a broad sense. This effort envisages the achievement of well-being as con- strained by economic resources (measured by GNP and the economic accounts generally), by the time available to people and households, and by human capital (broadly defined to include personal skills as well as social support networks). In a broad sense, these constraints are inputs. Outcomes are measured by social indicators in this view. Relating inputs to outcomes is not presently achievable and may never be. Given the very early stage of conceptualization and even empirical work, no firm conclusion on the feasibility or usefulness of this effort is offered. Considering the interest in many of the issues raised, it would seem desirable to have such work continue. CONCLUDING NOTE We have seen that while the national accounts measure production; the extensions beyond the accounting boundary into the consumption behavior of people and families, of their health and safety, their working condi- tions, and of the air they breathe involve a very large number of dimen- sions, including a lifetime dimension, of the human condition. In each of the domains that can be separately delineated, as we have sought to illustrate, the economic boundary defined by the national accounts is best viewed as a tool for improving conditions. This tool responds even- tually to a combination of market forces and other forces exercised by producers, consumer groups, unions, trade associations, Sierra Clubs, our governments, other governments, etc. For statistical measurement to help in this process, continued improve- ments should be made both in the national accounting measures and in the more difficult welfare or social accounting measures. For this ongoing endeavor, the economic boundary that measures the production of goods
306 PAPERS and services bought by final consumers, valued at market prices or factor costs, should be kept distinct. Measures of impacts on the environ- ment, for example, should not merely be used to add to or subtract from the measures of market production. Knowledge of these impacts obtained from welfare-type estimates used in conjunction with national accounting cost estimates of controlling these impacts should help us make better choices of changes in output and consumption of marketed goods and services. We conclude with a caveat: The scientific formulation of economic and social concerns could have grave and unforeseen consequences, however, if the conclusions became the basis for dictatorial decisions. Insofar as these conclusions provide useful knowledge to producers, consumers, and regulators to be used in open public debate, the con- sumer and social welfare could be furthered. REFERENCES Becker, G. (1975) Human Capital. 2nd edition. National Bureau of Economic Research. New York: Columbia University Press. Cobb, S., and Kasl, S. V. (1977) Termination: The Consequences of Job Loss. National Institute For Occupational Safety and Health. U.S. Department of Health, Education and Welfare, Public Health Service, Center For Disease Control. Davis, L. E., and Cherns, A. B., eds. (1975) The Quality of Working Life: Cases and Commentary. Vol. 2. New York: The Free Press. Denison, E. F. (1978) Effects of selected changes in the institutional and human environ- ment upon output per unit of input. Survey of Current Business 58(1). Fabricant, S. (1969) A Primer Ott Productivity. New York: Random House. Feldstein M. (1975) Social security: time for reform. The Public I,'`eres~ No. 40:75-95. Fuchs, V. R. (1975) Who Shall Live? Health. Economics and Social Choice. New York: Basic Books. Gianessi, L. P., Peskin, H., and Wolf, E., (1977) The distributional implications of national air pollution damage estimates. In F. T. Juster ea., The Distribution of Economic Well-Being. NBER Studies in Income and Wealth, Vol. 41. Cambridge, Mass.: Ballinger Publishing Co. Ginzberg, E. (1975) Work structures and manpower realities. In L. E. Davis and A. B. Cherns, eds., The Quality of Working Life. Vol. 1. New York: The Free Press. Herfindahl, O. C., and Kneese, A. V. (1973) Measuring social and economic change: benefits and costs of environmental pollution. In M. Moss, ea., The Measurement of Economic and Social Performance. NBER Studies in Income and Wealth, Vol. 38. New York: Columiba University Press. Hirst, E., and Carney, J. ( 1978) Effects of federal residential energy conservation programs. Science (February):845-851. Juster, F. T., ed. (1976) The Distribution of Economic Well-Being. NBE:R Studies in Income and Wealth, Vol. 41. Cambridge, Mass.: Ballinger Publishing Co.
Welfare Dimensions of Productivity Measurement 307 Kahn, R. L. (1960) Productivity and job satisfaction. Personnel Psychology 13(3). Kahn, R. L. (1971) The meaning of work: interpretation and proposals for measure- ment. In A. Campbell and P. E. Converse, eds., The Human Meaning of Social Change. New York: Russell Sage Foundation. Katzell, R. A., Yankelevitch, D., et al. (1975) Work Productivity and Job Satisfaction. The Psychological Corporation. New York: Harcourt Brace Jovanovich, Inc. Keynes, J. M. (1930) Economic possibilities for our grandchildren. Reprinted in Essays in Persuasion. first published in 1931. Reissued 1952. London: Rupert-Hart Davis. Lave, L. B., and Seskin, E. P. (1973) Analysis of association between U.S. mortality and air pollution. Journal of the American Statistical Association 68(342):284-290. Leontief, W. (1970) Environmental repercussions of the economic structure: an input- output approach. Review of Economics and Statistics 52:262-271. Leontief, W. (1973) National income, economic structure, and environmental externali- ties. In M. Moss, ea., The Measurement of Economic and Social Performance. NBER Studies in Income and Wealth, Vol. 38. New York: Columbia University Press. Michael, T., and Becker, G. S. (1973) On the new theory of consumer behavior. The Swedish Journal of Economics. 75:379-396. Mincer, J. (1970) The distribution of labor incomes: a survey with special reference to the human capital approach. Journal of Economic Literature 8(March): 1-26. Mincer, J. (1974) Schooling, Experience, and Earnings. National Bureau of Economic Research. New York: Columbia University Press. Mitchell, W. (1945) Prices and the Cost of Living in Wartime-An appraisal of the Bureau of Labor Statistics Index of the Cost of Living in 1941-44. Report of the .ech- nical committee appointed by the Chairman of the President's Committee on the Cost of Living, by Wesley Clair Mitchell, Chairman, Simon Kuznets, and Margaret Reid (June 1944). Pp. 243-369 Report of the Presidents Committee on the Cost of Living, Office of Economic Stabilization. Washington, D.C.: U.S. Government Printing Office. Moss, M., ed. (1973). The Measurement of Economic and Social Performance. NBER Studies in Income and Wealth, Vol. 38, New York: Columbia University Press. Moss, M. (1978) Income distribution issues viewed in a lifetime income perspective. The Review of Income and Wealth 24(2):119-136. Nordhaus, W. D., and Tobin, J. (1972) Is growth obsolete?, in Economic Growth. Fiftieth Anniversary Colloquium V, National Bureau of Economic Research. Also published with extended discussion in M. Moss, ed. ( 1973) The Measurement of Economic and Social Performance. New York: Columbia University Press. Okun, A. M. (1975) Equality and Efficiency. Washington, D.C.: The Brookings Insti tution. Organization For Economic Cooperation and Development ( 1976) Measuring Social Well-Being. Paris: Organization for Economic Cooperation and Development. Piore, M. (1978) Unemployment and inflation: an alternative view. Challenge (May- June), particularly p. 28. Rees, A. (1973) The Economics of Work anal Pay. Chapter 14. New York: Harper and Row Publishers. Sahota, G. S. (1978) Theories of personal income distribution: a survey. Journal of Economic Literature 16(1 ): 1-55. Scott. W. R. (1977) Effectiveness of organizational ettectiveness studies. Pp. 63-95 in P. Goodman, M. Pennings and Associates, New Perspectives on Orgc~ni7atio'~c'1 Effective- ness. San Francisco, Cal.: Jossey-Bass, Inc.
308 PAPERS Seashore, S. E. (1977) Monitoring the quality of working life, in S. H. Hymans and H. T. Shapiro, eds. The Economic Outlook for 1977. Department of Economics, University of Michigan, Ann Arbor. Smith, J. D., ed. (1975) The Personal Distribution of Income and Wealth. Studies in Income and Wealth, Vol. 39. New York: National Bureau of Economic Research. Stone, R. (1972) The evaluation of pollution: balancing gains and losses. Minerva 10(3):412-425. Stone, R. (1973) A system of social matrices. Review of Income and Wealth 19(2). United Nations (1975) Towards a System of Social and Demographic Statistics. Studies in Methods, Series F. no. 18. New York: United Nations. United Nations (1978) Social Indicators: Preliminary Guidelines and Illustrative Series. Statistical Papers, Series M, no. 63. New York: United Nations. Wool, H. (1973) What's wrong with work in America? a review essay. Monthly Labor Review 96(3):38-44. Wyzga, R. E. (1978) The effect of air pollution upon mortality: a consideration of distribution lag models. Journal of the American Statistical Association 73(363): 463-471.