Social dimensions of macroeconomic measurement


caption Figure 1: Macroeconomics in Context. (Source: Global Development And Environment Institute, Tufts University)

The traditional macroeconomic model portrays a hypothetical economy in which only businesses engage in production, and in which the natural environment plays no role. Increasingly, however, people have raised questions about whether this gives an adequate picture of the macroeconomy.

People have come to realize that economic activity actually takes place within the context of human social institutions which in turn are inextricably embedded in the natural environment. This embeddedness is illustrated by the outer rings labeled “Social Context” and “Physical Context” in Figure 1. In addition, the contributions to production of households and community groups (within the core sphere), and of non-profit as well as government institutions (within the public purpose sphere) have recently received more attention, as illustrated in the center of Figure 1. Of course, the role of businesses, both foreign and domestic, is recognized in both the traditional and newer approaches.

Many researchers argue that national governments need to start gathering new kinds of data in order to face the challenges of 21st century concerns. Building on these new kinds of information, some researchers are concentrating on developing refined measures of national assets and production, keeping as close as possible to the framework of the National Income and Product Accounts (NIPA).

Other researchers make it their aim to design indicators that more directly measure social and economic well-being. Rather than seeking to measure the volume of production, these researchers seek to develop indicators of the quality of life.

Measuring Household Production

Another significant omission in the national accounts, as currently constituted, is the value of much household production. Only two aspects of household production are currently counted in gross domestic product (GDP): one, the services of the house itself (the rent paid explicitly or implicitly by residents) and, two, the services provided by paid household workers such as housekeepers and gardeners. The household production of goods and services such as childcare, housecleaning, laundry services, meal preparation, landscaping, and transportation using unpaid labor and household capital goods (such as automobiles and appliances) is not counted.

Even the most conservative estimates of the total value of household production come up with numbers equal to about 25-35% of standard GDP in the U.S., and less conservative estimates put the value as equal to or greater than the value of marketed production. How did this substantial area of productive activity come to be overlooked, and what is being done to remedy this omission?

The History of Exclusion

Various reasons have been advanced over the decades for why household activities should not count in GDP.

“Households are nonproductive.” One reason given was that households were not part of the economy because they did not produce economic goods. This depends on a definition of “economic” in which factories, farms and office buildings, and the people in them, were taken as defining “the economy.” Dating from Victorian times, households and families were often thought of, conversely, as “non-economic,” running along more primitive lines of social development, and as somehow “closer to nature” and naturally self-renewing. There was also a gender aspect to this split: for much of the 19th and 20th centuries, “the economy” was a man’s world, while “the home” was assigned to women.

This omission of much household production from the national accounts may contribute to a subtle bias in the perceptions of policy-makers who make economic decisions based on them. Since household work is not measured, it may be easy to think that its not important, or not even an “economic” matter at all. The U.S. Social Security retirement system, for example, gives stipends to people based only on their market wages and years in paid work. Some advocates suggest that people should also get credit for time spent in child raising, for example up to a year of Social Security credit for time taken off with each child, in recognition of the contribution that such unpaid work makes to social and economic life. Having home production counted in GDP might help keep policy-makers more aware of its productive contributions.

“It’s too hard to distinguish household production from consumption.” Sometimes an argument has been made on the basis of convenience: it is simply too hard, some commentators have argued, to distinguish production activities from consumption activities within the household. If diapering your baby is considered work, should cuddling be considered work, too? If you enjoy cooking as a hobby, isn’t that activity more like a form of entertainment than like the work of a paid cook? Should the time you spend playing tennis be considered “productive” and counted in GDP? It’s just easier, such commentators argue, to draw the line at the household door than to try to sort out the question of “What part of the activities are productive?”

Yet the matter of distinguishing production from consumption has been discussed intelligently for decades. A frequently-applied rule is what is called the third person criterion: an activity should only be classified as household production if it is in principle replaceable by market or third-person services. Cooking a meal is in this sense “production,” since you might alternatively pay someone to cook for you. Playing a game of tennis clearly is not, since paying someone to play on your behalf would be absurd.

The fact that some household productive activities—like some parts of cooking and child care—may sometimes be sources of intrinsic pleasure to some of the people doing the work is, arguably , not particularly relevant. If we insisted that only work that is perceived as neutral or unpleasant should count as “work,” then substantial portions of time at paid jobs should obviously be excluded as well. There is an increasing realization that the definition of “work” should be centered around the idea that it is an activity that produces something valuable—a needed report, an emotionally healthy child—and not around whether the activity is unpleasant (or not) or paid (or not).

“GDP measures market production.” Sometimes it is granted that people do engage in productive activities in their roles as household members. For example, in the 1993 United Nations System of National Accounts, it is granted that many household activities “are productive in an economic sense”. But then you will often hear it said that that GDP aims to only measure production for the market. Since household outputs are not sold, this argument goes, it is consistent to exclude them from GDP.

The problem with this argument is that a substantial portion of GDP already reflects nonmarket production. Most government production of goods and services is never sold on a market (almost 12% of GDP in 2002). Nor do owner-occupiers really pay rent on their houses to themselves (about 7.5% of GDP in 2001). As more market substitutes have developed for formerly home-produced goods and services (day care for home child care, prepared foods for home-cooked, etc.) it has become less and less tenable to draw the line for defining “the economy” at the household door, imputing “market values” for the services of governments and houses, but not for the unpaid work done within homes.

caption Figure 2(a): Proportion of Full-Time Homemakers Among All Workers, 1870. In 1870, about 40 out of every 100 workers (both paid and unpaid, male and female) worked full time at home doing unpaid household production. (Source: Global Development And Environment Institute, Tufts University)

“Including household production would make too big of a change in the accounts.” The designers of the UN System of National Accounts granted that households do, in fact, produce many goods and services, and that estimated monetary values of these flows would be very large. However, the United Nations statisticians who design the System of National Accounts went on to argue that “The inclusion of large non-monetary flows of this kind in the accounts together with monetary flows can obscure what is happening on markets and reduce the analytic usefulness of the data” for addressing topics such as inflation and economic fluctuations.

caption Figure 2(b): Proportion of Full-Time Homemakers Among All Workers, 2000. By 2000, only 16 out of every 100 workers were full-time homemakers. (Source: Global Development And Environment Institute, Tufts University)

In reply, some economists argue that current GDP figures are less accurate for having neglected household production. Most obviously, GDP is understated—a substantial area of valuable productive activity has been overlooked. Perhaps more importantly, changes in an economy over time, such as the true growth rate of production, may be misstated. One of the major economic shifts that occurred during the 20th century was the movement of a large proportion of women from unpaid employment as full-time homemakers into paid employment in market work. In 1870, 40% of all U.S. workers (paid and unpaid, male and female) were full-time homemakers, as illustrated in Figure 2(a); by 2000, the proportion had dropped to 16%, as shown in Figure 2(b). This increase in market work, simultaneous with the increase in purchases of substitutes for home production such as paid child care and prepared foods, were counted as increases in GDP. The value of lost household production, however, was not subtracted off. This failure to account for reductions in some home-produced goods and services would tend to mean that GDP growth during the period was overstated.

The picture is a bit more complex, however, if you also take into account the changes in the productivity of household labor time during this period. Homemakers in 1870 had very little in the way of “capital stock” to work with—probably a coal stove that needed daily cleaning (instead of a range), a washtub and clothesline (instead of a washer/dryer), and an icebox (instead of a refrigerator). Productive investments in household technology have, the evidence suggests, led to people now enjoying cleaner clothes and more interesting and varied diets than previously. While the household sector was shrinking relative to other sectors in terms of labor hours devoted to it, the value of its real product, as least in some areas like cooking and cleaning, was not necessarily falling. This growth in true national product due to changes in household productivity has been entirely missed in standard accounting.

Accounting for household production might also change how we see the cyclical behavior of the economy, since there is reason to believe that the level of household production probably moves counter-cyclically. That is, when the economy is down—in recession—people’s “do-it-yourself” work probably goes up. While the financial and emotional consequences of lack of needed paid employment should not be trivialized, it is likely that many of the unemployed use some of their extra time doing additional childcare and household tasks. When people are in financial straits, they also tend to economize by replacing market purchases with home-produced goods and services.

Comparisons between countries are also made more difficult by the lack of accounting for household production in GDP. In countries of the global South, where such activities make up a much higher proportion of total production, GDP is even more inadequate as an indicator of national production.

Time Use Surveys

Table 1: Average Hours per Day Spent in Primary Activities
Activity Total Men Women
Personal care 9.34 9.13 9.54
Eating and drinking 1.21 1.24 1.18
Household 1.83 1.33 2.30
-- Housework .62 .23 .98
-- Food preparation and cleanup .53 .25 .79
-- Lawn and garden care .20 .26 .14
-- Household management .13 .11 .15
Purchasing goods and services .81 .68 .94
Caring for and helping people .84 .60 1.06
Working and work-related activities 3.69 4.57 2.87
Educational activities .47 .45 .50
Organizational, civic, and religious activities .32 .29 .35
Leisure and sports 5.11 5.41 4.83
Other activities .38 .31 .44
TOTAL 24.00 24.00 24.00
Source: “Time-Use Survey: First Results Announced by BLS”
BLS NEWS, Sept. 14, 2004, Table 1. Accessed Oct. 4, 2004.

A first step in determining a value for household production is to find out how much time people spend in unpaid productive activities. In the past, estimates of time use for the U.S. came from small and sporadic surveys. However, in 2003, following the lead of many other industrialized countries, the U.S. Bureau of Labor Statistics began collecting data for the first national ongoing survey of time use. The American Time Use Study (ATUS), conducted by the U.S. Bureau of Labor Statistics, asks people age 16 or over in a nationally representative sample to report in detail how they used their time on one particular day.

The first results of the ATUS were released in 2004. The results of the survey indicate that on average on any given day, 84% of women and 63% of men spend some time doing household activities including housework, food preparation and cleanup, lawn and garden care, or household management (such as paying bills). When averaged over all responses (including those who had not spent any time on household activities), women spent an average of 2.3 hours per day on these activities, while men spent 1.3 hours.

Some other highlights are summarized in Table 1. These figures are averages for people in all employment categories, and include both weekdays and weekend days.

The largest blocks of time reported were for personal care (including sleeping), leisure and sports (including an average 2.6 hours per day of watching television), and working (for pay) and work-related activities (including commuting time).

Unpaid household production is spread over several categories in Table 1. Time spent caring for children or the elderly or ill is included in the category of “caring for and helping people,” though these data require closer examination. These tasks may be done as a primary activity, as reported in Table 1. Or they may be done as a secondary activity—that is, done while the person is primarily doing something else, such as shopping or watching television. Other tables released by the BLS reveal that in households with children under 13, women spend an average of 6.4 hours caring for children as a secondary activity, while men spend an average of 4.1 hours.

The most conservative approach to measuring household production would be to count only primary “household” and “caring and helping” activities as productive, yielding an average figure for household production of 2.67 hours per day, compared to 3.69 for paid work and related activities. Even taking this most conservative approach, household production would then account for 42% of total productive time. Less conservative approaches would also include as “productive” at least some of the time spent purchasing (analogous to how working as a “purchaser” in a business is considered productive), in education activities (that is, investing in human capital), and in caring done as a secondary activity. With such approaches, household production could easily be found to account for well over half of total productive time.

Methods of Valuing Household Production

Unlike the attempt to incorporate environmental assets and services into the national accounts, which requires considerable development of new techniques of measurement and valuation, the imputation of a value for household production can generally follow a similar procedure to that currently used to impute a value for government production.

Both government production and household production result in goods and services that are generally not sold on markets. Both government production and household production use manufactured capital goods and labor. Hence, as with government production, a quasi-market value for household production could be imputed by summing the values of the labor and capital services devoted to the productive activities.

The major difference is that in the case of household work the labor is unpaid. Once time use has been measured in terms of hours spent on various activities, by surveys such as the one just discussed, standard national accounting procedures demand that these hours be assigned a monetary value using some market or quasi-market prices. Economists have developed two main methods of assigning a monetary value to household time use: The replacement cost method and the opportunity cost method.

Replacement cost method. In the replacement cost method, hours spent on household labor are valued at what it would cost to pay someone else to do the same job. In the most popular approach—and the one used to generate the most conservative estimates—economists use the wages paid in a general category like “domestic worker” or “housekeeper” to impute a wage. A variant of this method, which usually results in higher estimates, is to value each type of task separately: child care time is valued according to the wage of a professional child care worker, housecleaning by the wages of professional housecleaners, plumbing repair by the wages of a plumber, etc.

Opportunity cost method. The opportunity cost method starts from a different view, based on microeconomic “marginal” thinking. Presumably, if someone reduces his or her hours at paid work in order to engage in household production, he or she must (if acting rationally) value the time spent in household production (at the margin) at least the wage rate that he or she could have been earned by doing paid work for another hour. That is, if you choose to give up $30 you could have earned working overtime in order to spend an hour with your child, you must presumably think that the value of spending that hour with your child is at least $30. This leads to using the wage rate the household producer would have earned in the market to value household work time. In this case, estimates of the value of nonmarket production can go quite a bit higher, since some hours would be valued at the wage rates earned by doctors, lawyers, managers, etc.

Neither approach to imputing a wage rate is perfect. However, it would be hard to argue that perfection has been achieved in any of the other measurements and imputations involved in creating the national accounts, and many argue that imputing some value for household labor time, even using minimal replacement costs, is more accurate than imputing a value of zero.

In addition to valuing the time used in household production, a value must be assigned to the capital services provided by appliances, vehicles, and the like. “Consumer durables” spending by households would need to be renamed “household investment” spending, and included as a subcategory of investment rather than consumption. Adding the flow of services that are yielded by these capital goods to the measurement of GDP would require that new calculations be made. But the techniques for estimating such service flows have already been designed for the case where cars, washing machines, etc. are owned by businesses. These same techniques could be extended to household capital goods.

While this section focuses on unpaid household work, similar arguments have also been made concerning unpaid volunteer work in communities and nonprofit organizations—the time people spend coaching children’s sports teams, visiting nursing homes, serving on church and school committees, etc. In the ATUS, 13% of the people surveyed reported participating in organizational, civic and religious activities on their surveyed day, a figure which includes organized volunteer activities. Were both these forms of work counted, the proportion of production attributed to the “households and institutions” sector would rise considerably.

Making Changes

Perhaps surprisingly, initial estimates of the value of household services for the United States predate the design of the National Income and Product Accounts (NIPA). In 1921 a group of economists at the National Bureau of Economic Research calculated that the value of household services would be about 25 to 30% of market national income. Decades later, economist Robert Eisner in 1988 reviewed six major proposed redesigns of the NIPA, all of which included substantial estimated values for household production. In spite of numerous demonstrations of its practicality dating back more than 80 years, however, actual inclusion of household production in the U.S. NIPA remains a project for the 21st century.

Internationally, there is interest in at least gathering data on household production. Many counties, including Australia, Canada, India, Japan, Mexico, Thailand and the United Kingdom, have conducted or are conducting national time use surveys to aid their understanding of unpaid productive activities. The United Nations Statistical Commission and Eurostat (the statistical office of the European Union) are encouraging countries to develop satellite accounts that, similar to the satellite accounts for the environment, provide the necessary information to adjust measures of GDP to take household production into account while not changing the definition of official GDP.

Measuring Economic Well-Being

Instead of trying to estimate a dollar value for domestic production, some economists insist that instead of – or in addition to – standardized national accounts, more direct indicators of economic well-being are needed. Since the goal of macroeconomics is human well-being, we need to be sure the indicators we pay most attention to are ones that relate to the goal we want to achieve!

Does Output Measure Well-Being?

Measures of the value of national output (or income) per capita—even improved measures that incorporate environmental concerns and household production—can often be poor indicators of sustainable human well-being. Neglect of the questions of what, how and for whom can mean that growth in production per capita may not lead to increased welfare. Now we can go into more detail about the problems that arise from focusing on production alone. These include:

Well-being-reducing products. Some outputs decrease rather than increase human well-being. The production of unhealthy foods and drugs, dangerous equipment, and community-destroying urban developments, for example, may lower, not raise, well-being. Even if it is the case that people are apparently willing to pay for such goods and services, either individually (perhaps influenced by media advertising) or through their governments (perhaps influenced by interest group lobbyists), it may be that such decisions reflect poor information or bad judgment when looked at from a well-being point of view.

Defensive expenditures. As discussed earlier, some outputs merely compensate for, or defend against, harmful events. Environmental defensive expenditures are only one example. Others include armaments necessitated by an increase in international tensions, increased spending on police forces to combat increased crime, increased gun ownership due to fear or societal breakdown, or increased medical spending due to a rise in automotive accidents. These increased expenditures do not reflect an increase in welfare, but only an attempt to maintain a status quo situation.

Loss of leisure. A rise in output might come about because people expend more time and effort at paid—and, in expanded accounts, also in unpaid—work. Only looking at the increase in measured output doesn’t take into account the fact that overwork makes people more tired and stressed, and takes away from time they could use for enjoying other activities.

Loss of human and social capital formation. Measured output is lower to the extent that people eliminate or reduce immediately productive activities in order to invest in schooling or training, or participate in community-building activities. Yet if we consider these as investments in the creation of valuable human and social capital such activities should be seen as increasing sustainable well-being.

Well-being-reducing production methods. If people are miserable at their jobs, suffering alienation and unpleasant working conditions, their well-being is compromised. This is even more obvious in cases where their health or survival is threatened by unhealthy or dangerous working conditions, even if their work results in a high volume of marketed goods and services.

Unequal distribution. National income may be very unevenly distributed, making some people very rich but leaving others in extreme poverty.

For such reasons, and others, some economists argue that other indicators, including direct measures of well-being outcomes, are necessary either instead of—or more often, in addition to—improved official national production accounts.

Some of these base their work on an NIPA-like framework, but they include more qualitative judgments about whether specific kinds of “production” actually add to sustainable well-being or perhaps detract from it instead. Some of the redesigns of national accounting suggested by academic economists have included adjustments for well-being-enhancing, or well-being-damaging, production. In 1989, economist Herman Daly and theologian John Cobb, Jr. suggested an alternative measure to GDP which they called the Index of Sustainable Economic Welfare (ISEW), in a popular book entitled For the Common Good: Redirecting the Economy Toward Community, the Environment, and a Sustainable Future. That initiative helped to spawn the design of more such measures nationally and internationally, generally implemented by nonprofit groups. We will examine one such measure, the Genuine Progress Indicator.

Other academic and nonprofit researchers have moved more directly to the issue of whether the economy is getting better or worse for people by developing indicators of social and economic outcomes enjoyed by the population. We will discuss one such measure, the Human Development Index, developed by a United Nations agency.

Example: The Genuine Progress Indicator

Table 2: Calculating The 2000 Genuine
Progress Indicator (in billions of 1996 dollars)
The GPI ’s Starting Point
Personal Consumption in 2000 6,258
Costs That Are Subtracted From The GPI
Adjustment for unequal income distribution -959
Net foreign borrowing -324
Cost of consumer durables -896
Social costs
Cost of crime -30
Costs of automobile accidents -158
Cost of commuting -455
Cost of family breakdown -63
Loss of leisure time -336
Cost of underemployment -115
Environmental costs
Cost of household pollution abatement -14
Cost of water pollution -53
Cost of air pollution -39
Cost of noise pollution -16
Loss of wetlands -412
Loss of farmlands -171
Depletion of nonrenewable resources -1,497
Cost of long-term environmental damage -1,179
Cost of ozone depletion -313
Loss of old-growth forests -90
Benefits That Are Added To The GPI
Value of housework and parenting 2,079
Value of volunteer work 97
Services of consumer durables 744
Services of highways and streets 96
Net capital investment 476
The Total: The Genuine Progress Indicator 2,630
The Genuine Progress Indicator: 2000 Update” Clifford Cobb,
Mark Glickman, and Craig Cheslog, p. 3.

One measure of well-being expressed in monetary terms is the Genuine Progress Indicator (GPI), calculated for the United States by the nonprofit group Redefining Progress.

The GPI takes as its starting point the Personal Consumption Expenditures (PCE) component of GDP for each year, as calculated by the Bureau of Labor Statistics. The reasons for not wholly including government spending, (business) investment, and net exports—the other components of standard GDP—require some explaining:

  • Much of government spending is on a clearly defensive type of expenditure—national defense. Similarly, government regulatory activities defend against predatory businesses, police spending defends against crime, etc. The GPI excludes such defensive expenditures. Only certain government services (such as provision and maintenance of highways and streets) are added in, at a later stage.
  • Net capital investment in structures and equipment is added in at a later stage (though only to the extent that it increases such capital per worker). (Consumer durables—that is, household investment expenditures—are subtracted off, but their services are later added back in.)
  • No adjustment is made for net exports since the goal is a measure of domestic welfare, starting with a measure of people’s consumption, rather than of production. Imported consumer goods are already included in Personal Consumption Expenditures, and exports, of course, never make it to domestic consumers. However, the creators of the GPI also consider international capital flows to be important in judging the sustainability of welfare. If the U.S. continually imports more than it exports, the debt it incurs in order to do so is considered to be a drain on “genuine progress” and a separate adjustment is made.

The adjustments made to Personal Consumption Expenditures in order to arrive at the GPI for a recent year (2000) are described in Table 2. The GPI subtracts off a factor reflecting the assumption that income inequality reduces well-being, and also a factor for foreign indebtedness. The value of consumer durables is subtracted not because such spending is “bad,” but rather because the GPI recognizes that the benefit of such goods comes from enjoying a flow of services—added in the last set of adjustments—rather than from spending on the goods themselves.

The GPI then subtracts off other factors reflecting social costs, such as the cost of crime to households and the loss of leisure time. Then estimates of environmental costs, such as the costs of air pollution and loss of wetlands, are subtracted off. Lastly, some benefits are added in. An estimate of the value of time spent on household and volunteer work is added. The services of consumer durables and the services that come from government spending on highways and streets is added.

Real GPI for the year 2000 was calculated at $2.63 trillion, as can be seen in Table 2. This is substantially less than the Personal Consumption Expenditures figure—$6.259 trillion—from which the calculations started. Hence, on net, the subtracted costs—especially environmental costs—are more sizeable than the added benefits. Dividing by the population size, per capita GPI was calculated as $9,550.

Trends in real GDP per capita and real GPI per capita are tracked from 1950 to 2000 in Figure 4. Not only is per capita GPI lower than per capita GDP, it has also grown more slowly. That is, environmental and social costs have been increasing faster than the value of the benefits omitted from standard accounting. Looking at the GPI over time, you get a sense of a steady or even falling level of sustainable well-being, which contradicts the trend you see when looking only at measured GDP.

Example: The Human Development Index

Human Development Index 2000
Life Expectancy at
Birth (years)
GDP per
1 Norway 78.5 .98 29,918 .942
6 United States 77.0 .98 34,142 .939
27 Korea, Rep. of 74.9 .95 17,380 .882
46 United Arab
75.0 .74 17,935 .812
54 Mexico 72.6 .84 9,023 .796
89 Sri Lanka 72.1 .84 3,530 .741
122 Namibia 44.7 .81 6,431 .610
130 Cambodia 56.4 .66 1,446 .543
161 Angola 45.2 .36 2,187 .403
173 Sierra Leone 38.9 .33 490 .275
Source: UNDP Human Development Report 2002, pp. 149-152.
* Expressed in purchasing power parity U.S. dollars

Beginning in 1990, the United Nations Development Programme (UNDP), under the direction of Mahbub ul Haq and in consultation with Nobel laureate economist Amartya Sen, began publication of a Human Development Index (HDI) designed to compete with GDP as a crude index of human welfare.

The HDI aggregates three indicators of human well-being:

  • Life expectancy at birth
  • An index reflecting a combination of the adult literacy rate and statistics on enrollments in education
  • GDP per capita

That is, not only does the HDI take into account output per capita, but it also takes into account whether national production has been effective in raising well-being outcomes in terms of human longevity and knowledge.

The resulting index is a number between zero and one, and is computed by the UNDP for many countries every year. Examples taken from the 2002 Human Development Report are shown in Table 3.

Taking the 173 countries as a whole, there is a rough correspondence between rankings by HDI and rankings you would get by looking at GDP alone. At the top of the rankings by HDI 2002 is Norway, with a life expectancy of 78.5 years, an education index of .98, and GDP per capita of $29,918. At the bottom of the rankings is Sierra Leone, with both extremely low GDP per capita and abysmally low indicators for health (life expectancy of only 38.9 years) and education.

Comparisons between HDI and GDP can be revealing about the shortfalls of GDP as an indicator of welfare. Life expectancies are over a year longer in all five countries (Norway, Sweden, Canada, Belgium, and Australia) that rank above the United States in HDI, even though the U.S. has higher GDP per capita. Looking at Table 3, you can see that GDP per capita is very similar between the Republic of Korea ($17,380) and the United Arab Emirates ($17,935), but Korea has been more successful in educating its populace (an education index of .95 compared to .74). Sri Lanka makes a relatively strong showing on the HDI index, in spite of its low levels of GDP per capita ($3,530), compared to, for example, Namibia, which has a substantially higher measured GDP per capita ($6,431). Among very poor countries, Cambodia makes a similarly relatively strong showing compared to Angola, which has a higher income per capita but lower health and education achievements.

HDI and GDP rankings differ for a number of reasons. Countries that score lower on HDI than countries with similar levels of GDP per capita may be plagued by war or by diseases such as AIDS. They often have very unequal income distributions and/or have governments that put a low priority on spending for health and education. Table 3 makes it clear that what is important in determining human welfare is not only how much is produced, but also what is produced (such as armaments vs. clinics) and how it is distributed. Countries with good social infrastructure and a lack of extreme gaps between rich and poor tend to score relatively well on HDI, compared to their achievements measured by GDP alone.

The UNDP also gathers a wealth of other statistics that can be used to measure human development in greater detail. These include measures of disparities between males and females in health, literacy, and political and economic participation; measures of poverty and deprivation (for example, the percent of the population without access to safe water), and statistics on crime and environmental degradation. All of these are important to getting a full picture of human development. They also gather data on topics such as military expenditures, amounts of external debt, and public expenditures on health and education that can help to explain differences in economic and social outcomes.

The Future of Macroeconomic Indicators

Due to funding cutbacks, progress on improving macroeconomic measurement within the official accounts is currently largely at a standstill within the United States. Many private groups within the United States, however, as well as official statistical agencies in a number of other countries, are making progress in developing better measures to deal with the social and environmental issues of the 21st century.

No one—and especially not their creators—would argue that alternative macroeconomic indicators have been perfected. It is quite possible to argue about how “costs of family breakdown” can be valued in computing GPI, for example, or whether more direct measures of poverty should be included in HDI.

Of course, no one intimately involved in the creation of the official NIPA would argue that they are perfect, either! The creation of official statistics has often been compared to sausage making: the closer you are to observing the process, the less attractive the final product appears. Only from a distance—when you can remain naively unaware of all the conventions that their creators have had to impose on messy real-world data and all the imputations they have had to make—do “official” statistics look clean and elegant, just as only with a distance from its processing does sausage look appetizing.

All indicators have their flaws, whether due to their definitions and classifications, the data they rely on, or the statistical techniques applied. The important contribution of alternative indicators is the way they bring to our attention significant aspects of the economy, such as environmental and social sustainability, that were not taken into account in the design of traditional 20th century macroeconomic indicators.

Further Reading



Harris, J., Goodwin, N., Nelson, J., & Institute, G. (2006). Social dimensions of macroeconomic measurement. Retrieved from


To add a comment, please Log In.