1
Introduction
Gross domestic product (GDP) is the most widely used measure of economic activity. There are international standards for its calculation, and much thought has gone into its statistical and conceptual bases. But GDP mainly measures market production, though it has often been treated as if it were a measure of economic well-being. Conflating the two can lead to misleading indications about how well-off people are and entail the wrong policy decisions.
One reason why money measures of economic performance and living standards have come to play such an important role in our societies is that the monetary valuation of goods and services makes it easy to add up quantities of a very different nature. When we know the prices of apple juice and DVD players, we can add up their values and make statements about production and consumption in a single figure. But market prices are more than an accounting device. Economic theory tells us that when markets are functioning properly, the ratio of one market price to another is reflective of the relative appreciation of the two products by those who purchase them. Moreover, GDP captures all final goods in the economy, whether they are consumed by households, firms or government. Valuing them with their prices would thus seem to be a good way of capturing, in a single number, how well-off society is at a particular moment. Furthermore, keeping prices unchanged while observing how quantities of goods and services that enter GDP move over time would seem like a reasonable way of making a statement about how society’s living standards are evolving in real terms.
As it turns out, things are more complicated. First, prices may not exist for some goods and services (if for instance government provides free health insurance or if households are engaged in child care), raising the question of how these services should be valued. Second, even where there are market prices, they may deviate from society’s underlying valuation. In particular, when the consumption or production of particular products affects society as a whole, the price that individuals pay for those products will differ from their value to society at large. Environmental damage caused by production or consumption activities that is not reflected in market prices is a well-known example.
There is yet another problem. While talking about the concepts of “prices” and “quantities” might be straightforward, defining and measuring how they change in practice is an altogether different matter. As it happens, many products change over time—they disappear entirely or new features are added to them. Quality change can be very rapid in areas like information and communication technologies. There are also products whose quality is complex, multidimensional and hard to measure, such as medical services, educational services, research activities and financial services. Difficulties also arise in collecting data in an era when an increasing fraction of sales take place over the internet and at sales as well as discount stores. As a consequence, capturing quality change correctly is a tremendous challenge for statisticians, yet this is vital to measuring real income and real consumption, some of the key determinants of people’s well-being. Underestimating quality improvements is equivalent to overestimating the rate of inflation, and therefore to underestimating real income. For instance, in the mid-1990s, a report reviewing the measurement of inflation in the United States (Boskin Commission Report) estimated that insufficient accounting for quality improvements in goods and services had led to an annual overestimation of inflation by 0.6%. This led to a series of changes to the U.S. consumer price index.
The debate in Europe has tended to go the opposite way: official price statistics have been criticized for underestimating inflation. This has been partly because people’s perception of inflation differs from the national averages presented in the consumer price index, and also because it is felt that statisticians overadjust for quality improvements in products, thereby painting too rosy a picture of citizens’ real income.
For market prices to be reflective of consumers’ appreciation of goods and services, it is also necessary that consumers are free to choose and that they dispose of the relevant information. It takes little imagination to argue that this is not always the case. Complex financial products are an example where consumer ignorance prevents market prices from playing their role as carriers of correct economic signals. The complex and ever-changing bundles of services offered by telecommunications companies are another case in point where it is difficult to ensure the transparency and comparability of price signals.
All the above considerations imply that price signals have to be interpreted with care in temporal and spatial comparisons. For a number of purposes, they do not provide a useful vehicle for the aggregation of quantities. This does not imply that the use of market prices in constructing measures of economic performance is generally flawed. But it does suggest prudence, in particular with regard to the often overemphasized measure, GDP.
This chapter suggests five ways of dealing with some of the deficiencies of GDP as an indicator of living standards. First, emphasize well-established indicators other than GDP in the national accounts. Second, improve the empirical measurement of key production activities, in particular the provision of health and education services. Third, bring out the household perspective, which is most pertinent for considerations of living standards. Fourth, add information about the distribution of income, consumption and wealth to data on the average evolution of these elements. Finally, widen the scope of what is being measured. In particular, a significant part of economic activity takes place outside markets and is often not reflected in established national accounts. However, when there are no markets, there are no market prices, and valuing such activities requires estimates (“imputations”). These are meaningful, but they come at a cost, and we shall discuss them before turning to the other proposals.
Imputations—Comprehensiveness Versus Comprehensibility
Imputations exist for two related reasons. The first is comprehensiveness. There are productive activities and associated income flows (typically non-monetary) that take place outside the market sphere, and some of them have been incorporated into GDP. The single most important imputation is a consumption value for the services that home-owners derive from living in their own dwellings. There is no market transaction and no payment takes place, but the national accounts treat this situation as if home-owners paid a rent to themselves. Most people would agree that if two persons receive the same money income but one of them lives in his/her own house while the other rents, they are not equally well-off—hence the imputation in order to better compare incomes over time or between countries. This brings us to the second reason for imputations, the invariance principle: the value of the main accounting aggregates should not depend on the institutional arrangements in a country. For example, if exactly the same medical services are provided in one case by the public sector and in another case by the private sector, overall measures of production should be unaffected by a switch between the two institutional settings. The main advantage of adhering to the invariance principle is better comparability over time and between countries. Therefore, for instance, measures of “adjusted disposable income” for households (see below) include an imputation for government services provided directly to citizens.
The imputations can be smaller or larger, depending on the country and on the national accounts aggregate considered. In France and Finland, for example, the main imputations account for about one-third of adjusted household disposable income and for just over 20% in the United States. Thus, in the absence of imputations the living standards of French and Finnish households would be understated relative to the United States.
But imputations come at a price. One is data quality: imputed values tend to be less reliable than observed values. Another is the effect of imputations on the comprehensibility of national accounts. Not all imputations are perceived as income-equivalent by people, and the result may be a discrepancy between changes in perceived income and changes in measured income. This problem is exacerbated when we widen the scope of economic activity to include other services that are not mediated by the market. Our estimates below for household work amount to around 30% of conventionally-measured GDP. Another 80% or so are added when leisure is valued as well. It is undesirable to have assumption-driven data so massively influencing overall aggregates.
There is no easy way out of the tension between comprehensiveness and comprehensibility except to keep both elements of information available for users and to maintain a distinction between core and satellite accounts. A full set of household accounts, for example, may not be well placed in the core of national accounts aggregates. But a satellite account that comes up with a valuation of comprehensive forms of household production would represent a significant improvement.
What Can Be Done Within the Existing Measurement Framework?
Emphasize National Accounts Aggregates Other Than GDP
A first step towards mitigating some of the criticism of GDP as a measure of living standards is to emphasize national accounts aggregates other than GDP, for example, by accounting for depreciation so as to deal with net rather than gross measures of economic activity.
Gross measures take no account of the depreciation of capital goods. If a large amount of output produced has to be set aside to renew machines and other capital goods, society’s ability to consume is less than it would have been if only a small amount of set-aside were needed. The reason that economists have relied more heavily on GDP than on net domestic product (NDP) is, in part, that depreciation is hard to estimate. When the structure of production remains the same, GDP and NDP move closely together. But in recent years, the structure of production has changed. Information technology (IT) assets have gained importance as capital goods. Computers and software have a shorter life expectancy than do steel mills. On those grounds, the discrepancy between GDP and NDP may be increasing, and by implication, volume NDP may be increasing less rapidly than GDP. For example, real GDP in the United States rose by about 3% per year during the period 1985–2007. Depreciation rose by 4.4% over the same period. As a consequence, real net national product grew at a somewhat slower rate than GDP.
Of greater concern for some countries is that standard depreciation measures have not taken into account the degradation in quality of the natural environment. There have been various attempts to widen the scope of depreciation to reflect environmental degradation (or improvements, if such is the case), but without much success. The hurdle is the reliable measurement and monetary valuation of changes in environmental quality.
The case of natural resource depletion is slightly different—there is at least a market price, even if it does not reflect environmental damage attributable to the use of the natural resource. Depletion could be captured by excluding the value of the natural resources harvested from the production value of sectors like mining and timber. Their production would then consist only in a pure extraction or logging activity, with a corresponding decrease in GDP. A second possibility would be to take resource depletion into account in depreciation measures. In this case, GDP would be unchanged, but NDP would be lower.
In a world of globalization, there may be large differences between the
income of a country’s citizens and measures of domestic
production, but the former is clearly more relevant for measuring the well-being of citizens. We shall argue later that the household sector is particularly relevant for our considerations, and for households the income perspective is much more appropriate than measures of production. Some of the income generated by residents is sent abroad, and some residents receive income from abroad. These flows are captured by
net national disposable income, a standard variable in national accounts.
Figure 1.1 shows how Ireland’s income declines relative to its GDP—a reflection of an increasing share of profits that are repatriated by foreign investors. While the profits are included in GDP, they do not enhance the spending power of the country’s citizens. For a poor developing country to be told that its GDP has gone up may be of little relevance. It wants to know whether its citizens are better-off, and national income measures are more relevant to this question than GDP.
Moreover, the prices of imports evolve very differently from the prices of exports, and these changes in relative prices have to be taken into account in assessing living standards.
Figure 1.2 shows the divergence between real income and production in Norway, an oil-rich OECD country whose income has risen faster than GDP in times of rising oil prices. In many developing countries, whose export prices have tended to fall relative to import prices, the opposite will be true.
Improving the Measurement of Services in General
In today’s economies, services account for up to two-thirds of total production and employment, yet measuring the prices and volumes of services is more difficult than for goods. Retail services are a case in point. In principle, numerous aspects should be taken into account in measuring the services provided: the volume of goods transacted but also the quality of service (accessibility of the shop, general service level of the staff, choice and presentation of products and so forth). It is difficult even to define these services, let alone to measure them. Statistical offices generally use data on the volume of sales as indicators for the volume of trade services. This method leaves aside most quality change in the trade services provided. What is true for retail holds for many other service industries, including those that are often publicly provided, such as health and education. A greater effort will be needed to come to grips with tracking the quantity and quality of services in modern economies.
Improving the Measurement of Government-Provided Services in Particular
Governments play an important part in today’s economies. Broadly speaking, they provide two types of services—those of a “collective” nature, such as security, and those of an “individual” nature, such as medical services and education. This does not imply that government is necessarily the only provider of these services, and indeed, the mix between private and public provision of individual services varies significantly across countries. And while one can argue about the contribution of collective services to citizens’ living standards, individual services, particularly education, medical services and public sports facilities, are almost certainly valued positively by citizens. These services tend to be large in scale but badly measured. Traditionally, for government-provided non-market services, measures have been based on the inputs used to produce these services rather than on the actual outputs produced. An immediate consequence of this procedure is that productivity change for government-provided services is ignored, because outputs are taken to move at the same rhythm as inputs. It follows that if there is positive productivity growth in the public sector, our measures underestimate growth.
Source: OECD Annual National Accounts

Work has started in many countries to develop output measures for these government-provided services that are independent of inputs, but the task is formidable. Take the following example: the United States spends more per capita on health care than many European countries, yet in terms of standard health indicators, outcomes are worse. Does this mean that Americans receive less health care? Or does it mean their health care is more expensive and/or delivered less efficiently? Or does it mean that health outcomes also depend on factors specific to American society other than health expenditure? We need to break down the change in health expenditure into a price and an output effect. But what exactly are the volumes of output that one is looking for? It is tempting to measure them by the population’s state of health. The problem is that the link between health care expenditure and health status is tenuous at best: expenditures relate to the resources that go into the institutions providing health services, whereas the health status of the population is driven by many factors—and the situation is much the same for education. For example, people’s lifestyles will affect health outcomes, and the time parents spend with their children will affect exam scores. Attributing changes in health or education status solely to hospitals or schools and the money spent on them neglects all these factors and can be misleading.
The quest is for more accurate measures of the volume growth of public services. A number of European countries as well as Australia and New Zealand have developed output-based measures for key government-provided services. One major challenge to these efforts is, once again, to capture quality change. Without a good measure of quality (or equivalently, a good estimate of increases in productivity), it is impossible to ascertain whether the conventional input measures underestimate or overestimate growth. If undifferentiated quantity measures are used, such as a simple number of students or of patients, changes in the composition and quality of the output may be missed. But one has to start somewhere and, because the numbers involved are important, the issue cannot be ignored, For example, with output-based measures, the UK economy grew at a rate of 2.75% per year between 1995 and 2003, whereas if the previous convention had continued to be used, the growth rate would have been 3% (Atkinson 2005). Similar effects could be observed in the case of France. A Danish study on the measurement of health output points the other way: output-based production of health services grew more rapidly than input-based production (
Figure 1.3).
An important criterion for the reliability of output-based measures is that they are based on observations that are detailed enough to avoid mixing up true volume changes with compositional effects. We can ask how many students are educated, and simply count their numbers. If spending per student increases, one might conclude that the unit cost of educational services has increased. This may be misleading, however, if costs have gone up because students are taught in smaller classes or if there is a larger share of students that take up engineering studies, which are more costly. The measurement mistake arises because the simple number of students is too undifferentiated an output measure to be meaningful, so a more detailed structure is needed. It would help, for instance, to treat one hour taught to a graduate engineering student as a different product from one hour taught to a first-year arts student, and thus to account for some quality and compositional change. A similar reasoning applies for health care: the treatments of different diseases have to be considered as different medical services. As it turns out, the health-care systems of some countries do provide the administrative data needed to obtain this detailed information. We conclude that despite this being a daunting task, the better measurement of government-provided individual services is central to the better assessment of living standards. Exploiting new administrative data sources is one way of making progress in this direction. Ideally, the information would also capture service quality, for instance, the way patients are accommodated in hospitals or the time devoted to them by the medical staff, though such data may be hard to collect. In this case, new primary data sources such as surveys may be necessary.
Figure 1.3. Volume output of health services in Denmark
Source: Devci, Heurlén and Sorensen (2008) “Non-Markt Health Care Service in Denmark—Empirical Studies of A, B and C Meythods”; paper presented at the meeting of the International Association for Research on Income and Wealth, Slovenia
Improving the volume measures of outputs does not dispense with the need to improve—and publish—the volume measures of inputs. Only if both the outputs and inputs of service production are well captured will it be possible to estimate productivity change and undertake productivity comparisons across countries.
Revisit the Concept of “Defensive” Expenditures
Expenditures required to maintain consumption levels or the functioning of society could be viewed as a sort of intermediate input—there is no direct benefit, and in this sense they do not give rise to a final good or service. Nordhaus and Tobin, in their seminal 1973 paper, for example, identify as “defensive” those activities that “are evidently not directly sources of utility themselves but are regrettably necessary inputs to activities that may yield utility.” In particular, they adjust income downwards for expenditures that arise as a consequence of urbanization and a complex modern life. Many such “defensive expenditures” are incurred by government, while others are incurred by the private sector. By way of example, expenditure on prisons could be considered a government-incurred defensive expenditure and the costs of commuting to work a privately-incurred defensive expenditure. A number of authors have suggested treating these expenditures as intermediate rather than final products. Consequently, they would not be part of GDP.
At the same time, difficulties abound when it comes to identifying which expenditures are “defensive” and which are not. For instance, if a new park is opened, does this constitute defensive expenditure against the disamenities of urban life or is it a non-defensive recreational service? What are the possible ways forward? Some options include:
First, focus on household consumption rather than total final consumption . For many purposes, the former is a more meaningful variable. And all of governments’ collective consumption expenditures (which would include things like prisons, military expenditure and the clean-up of oil spills) are automatically excluded from household final consumption.
Second, widen the asset boundary. In many cases, defensive expenditures include elements of investment and capital goods. In those cases, they should be treated much like maintenance expenditures in the case of conventional production. For example, health expenditures could be seen as investment in human capital instead of as final consumption. If there is an asset that captures environmental quality, expenditures made to improve or maintain it could also be considered an investment. Conversely, the consequences of economic activity that is detrimental to this asset could be captured in an extended measure of depreciation or depletion so that the net measure of income or production is reduced accordingly. And net measures, it was argued earlier, should be our benchmark for living standards rather than gross measures.
Third, widen the household production boundary. Some “defensive” expenditures cannot reasonably be treated as an investment. Take the case of commuting to work. Households produce transportation services—they use their time (labor input) and money (commuter ticket) for this purpose. With the exception of the consumer’s purchase of a ticket for a commuter train, which counts as final consumption, none of the above flows enter measures of production and income. This could be remedied by allowing for the household production of transportation services, which would be considered as an unpaid delivery of intermediate inputs to firms, “subsidized” by private households. Although this treatment would not change overall GDP, it would show a larger contribution to production by households and a smaller contribution by firms.
The biggest obstacle to these approaches lies in their implementation. How exactly should the scope of defensive expenditures be determined? How should new assets and in-kind flows be valued? And, of course, widening the scope of asset and production measures brings with it more imputations.
Income, Wealth and Consumption Have to Be Considered Together
Income flows are an important gauge for the standard of living, but in the end it is consumption and consumption possibilities over time that matter. The time dimension brings in wealth. A low-income household with above-average wealth is better off than a low-income household without wealth. The existence of wealth is also one reason why income and consumption are not necessarily equal: for a given income, consumption can be raised by running down assets or by increasing debt, and consumption can be reduced by saving and adding to assets. For this reason, wealth is an important indicator of the sustainability of actual consumption.
The same holds for the economy as a whole. To construct the balance sheet of an economy, we need to have comprehensive accounts of its assets (physical capital—and probably human, natural and social capital) and its liabilities (what is owed to other countries). To know what is happening to the economy, we need to ascertain changes in wealth. In some instances, it may be easier to account for changes in wealth than to estimate the total value of wealth. Changes in wealth entail gross investments (in physical, natural, human and social capital) minus depreciation and depletion in those same assets.
Although information about some central aspects of household wealth is in principle available from national accounts balance sheets, it is often incomplete. Furthermore, certain assets are not recognized as such in the standard accounting framework, not least of all human capital. Studies that have computed monetary estimates of human capital stocks found that they account for an overwhelming part of all wealth (80% or more). A systematic measurement of human capital stock is of interest from a number of perspectives. It constitutes an integral part of an extended measure of household production (see below), and it is an input for the construction of sustainability indicators.
Note a fundamental problem with valuing stocks. When there are markets for assets, the prices at which assets are bought and sold serve to value the stock as a whole. But there may be no markets for certain assets or no trading on the markets, as has recently been the case for certain financial assets. This raises the question of how to value them. And even when market prices do exist, transactions correspond only to a small fraction of the existing stock, and they may be so volatile as to put a question mark on the interpretability of balance sheets. That said, basic information on assets and liabilities is key to assessing the economic health of the various sectors and the financial risks to which they are exposed.
Bringing Out the Household Perspective
Income can be computed for private households as well as for the economy as a whole. Some of citizens’ income is taken away in the form of taxes, and so is not at their disposal. But the government takes this money away for a reason: to provide public goods and services, to invest, for example, in infrastructure and to transfer income to other (normally more needy) individuals. A commonly employed measure of household income adds and subtracts these transfer payments. The resulting measure is referred to as a measure of household disposable income. However, disposable income captures only monetary transfers between households and the government, thereby neglecting the in-kind services that government provides.
Adjusting Household Income Measures for Government Services in Kind
Earlier in this text we mentioned the invariance principle, according to which a movement of an activity from the public to the private sector, or vice versa, should not change our measure of performance, except to the extent that there is an effect on quality or access. This is where a purely market-based measure of income meets its limits and where a measure that corrects for differences in institutional arrangements may be warranted for comparisons over time or across countries. Adjusted disposable income is a national accounts measure that goes some way towards accommodating the invariance principle, at least where “social transfers in kind” by government are concerned.
The meaning of adjusted disposable income is best explained by way of an example (
Table 1.1). Assume that a society’s labor income equals 100 and that individuals who are active in the labor market buy private health insurance. They make an annual payment for the insurance equal to 10, which can be decomposed into 8 units of insurance premiums (the actuarial value of a payment of 8) and 2 units of consumption of insurance services. At the same time, persons who are sick receive 8 units as reimbursement of their health expenditures. In this case—let us call it Case A—no taxes are paid and insurance claims and premiums offset each other, so that household disposable income equals 100. Now, assume that the government decides to provide the same amount of health insurance coverage to everyone, funded through a tax of 10 units. Nothing has changed, other than that the government is now collecting the insurance payment and distributing the benefits (Case B). But according to standard national accounts statistics, household disposable income has fallen, to 90 currency units. Thus, disposable income here yields a distorted comparison. If one adds in the social transfers in kind that households receive from the government under Case B (8 units corresponding to the reimbursement of health expenditures and 2 units corresponding to the running costs of the insurance), the adjusted measure of household disposable income indicates equality between the two cases.
Table 1.1. Private and Public Insurance Schemes
| Private insurance scheme (Case A) | Public insurance scheme (Case B) |
---|
Labor income | 100 | 100 |
Tax | 0 | –10 |
Insurance premiums (excluding insurance services) | –8 | 0 |
Insurance claims | +8 | 0 |
Household disposable income | 100 | 90 |
Social transfers in kind: | | +10 |
| —reimbursements | 0 | +8 |
| —running costs of the insurance | | + 2 |
Adjusted household disposable income | 100 | 100 |
The above example leaves aside, however, any consideration about which insurance regime operates more cost effectively and about the profits that might be made by private insurance companies—it was simply assumed that the private and public insurance services are equivalent to 2 currency units. In practice, this is almost certainly not the case, although it is difficult to make a general observation about the relative efficiency of such schemes. If the insurance services industry is not perfectly competitive (a reasonable assumption in most countries), the transfer of responsibility from the private to the public sector will be reflected in decreased profits and decreased insurance prices. Even if profits are distributed to households in the form of dividends, the change in the form of provision (from private to public) can increase the accessibility of the insurance service. Having an opportunity to insure against certain types of risks has a positive impact on the well-being of people who are risk adverse.
While the failure to estimate the value of the insurance services provided causes one set of biases, there are other biases that arise from the fact that the value of some social transfers in kind (those corresponding to the running costs of the insurance in the example above) is measured by the cost of producing these services. In some countries, in particular in the developing world, the cost of these services may greatly exceed their value to households, who may receive little or nothing. In this situation, the result of using adjusted household income would be a large-scale overestimation of the level of household income and consumption. Some of this can be tackled by using output-based volume measures for the health and education services produced by government. It is also likely that different parts of the population benefit unequally from social transfers in kind provided by government. There is thus an important distributional aspect.
Major items included in social transfers in kind are health and education services, subsidized housing, sport and recreation facilities and the like that are provided to citizens at a low price or for free. In France, general government provides nearly all of these services, which in 2007 cost about € 290 billion. Education and health services each account for about one-third of total transfers in kind, and housing and recreational and cultural activities (museums, public parks, etc.) account for about 10% (
Figure 1.4).
Figure 1.4. Social transfers in kind from general governement,France 2007
Source: INSEE.
Medians and Means—Distribution of Income, Consumption and Wealth
Average measures of per-capita income and wealth give no indication of how the available resources are distributed across persons or households. Similarly, average consumption gives no indication of how people effectively benefit from these resources. For example, average income per capita can remain unchanged while the distribution of income becomes less equal. It is therefore necessary to look at disposable income, consumption and wealth information for different groups. A conceptually simple way of capturing distribution aspects is to measure median income (the income such that half of all individuals are above that income, and half below), median consumption and median wealth. The median individual is, in some sense, the “typical” individual. If inequality increases, the differences between medians and averages may well increase, so a focus on averages does not give an accurate picture of the economic well-being of the “typical” member of society. For example, if all the increases in societal income accrue, say, to those in the top 10%, median income may remain unchanged, while average income increases. Over the past two decades, the dominant pattern in OECD countries is one of a fairly widespread increase in income inequality, with strong rises in Finland, Norway, Sweden (from a low base) and Germany, Italy, New Zealand and the United States (from a high base). In these cases, medians and means would give different pictures of what is happening to societal well-being. Alternatively, changes in the disposable income of different income groups can be tracked. Such an approach would, for instance, look at the numbers of people below a critical income level, or the average income of those in the bottom or top decile. Similar calculations would be useful for consumption and wealth. Empirical research has repeatedly shown that the distribution of consumption can be quite different from the distribution of income. Indeed, the most pertinent measures of the distribution of material living standards are probably based on jointly considering the income, consumption and wealth position of households or individuals.
In practice, moving from averages to medians is more difficult than meets the eye. Measures of averages are obtained by dividing aggregates by a population figure. To consider distributional elements, micro-economic information is needed that provides information for individual households or groups of households. Micro-economic measures refer to people living in private households and are typically derived from household income surveys, whereas macro-economic measures from the national accounts are based on a range of different sources, and include people living in collective households (such as prisons and institutions for long-term care).
An important choice also concerns the unit of measurement. Macro-estimates give totals for a whole country or sector, while micro-data retain the household (or the family) as the unit within which resources are pooled and shared, and adjust income for differences in “needs.” There are, for instance, fixed costs to running a household, allowing larger families with the same per-capita income to have a higher standard of living. Another step towards bringing demography and some distributional aspects into income measures is to calculate disposable income per consumption unit rather than per person. Consumption units are households whose size has been adjusted to take account of economies of scale in housing and other costs. This adjustment is of increasing importance as household size shrinks.
Against this background, we can consider the evolution of average and median household income in several countries (
Figure 1.5). Average income per capita and average income per consumption unit diverge, reflecting a trend towards a smaller household size. Survey income measures permit comparing average and median income. In the case of France, these two items move in parallel. At least from this perspective, there is no indication of a widening income distribution. The picture is different for the United States, where average income per capita and per consumption unit grow at the same rate but where there is a widening gap between median and average income, pointing to a more unequal income distribution.
Figure 1.5. Trends in different measures of household disposable income
Source: Computations based on OECD SNA and income distribution data.
There are many measurement issues that can influence the above statements. One source of discrepancy between micro- and macro-estimates is property income, whether imputed or not. If this aggregate is not well measured in micro-estimates, this could explain why average and median incomes in these estimates move in parallel in France, where wage inequalities are less important than property income inequalities. In addition, there is a possibility that top incomes are underrepresented in household income surveys. Finally, the international comparability between household surveys is far from perfect.
From the perspective of living standards, what matters is that the distribution of income, consumption and wealth determines who enjoys access to the goods and services produced within a society. Complementing measures of average income by measures with a distributional element is thus a crucial task for official statistics. Ideally, such distributional measures should be compatible in scope with average measures from the national accounts.
Similarly, the distribution of the volume of consumption is also important. The same dollar may buy different bundles of products, depending on the income group of the purchaser. Going from nominal to real income and from the value to the volume of consumption means applying a price index, raising the question of whose price index are we measuring. Conceptual discussions about price indices are often conducted as if there exists a single representative consumer. Statistical agencies calculate the increase in prices by looking at what it costs to purchase an average bundle of goods. The problem is that different people buy different bundles of goods, e.g., poor people spend more on food, and rich people on entertainment. People also buy goods and services in different types of stores, which sell “similar” products at very different prices. When all prices move together, having different indices for different people may not make much difference. But recently, with soaring oil and food prices, these differences have become marked. Those at the bottom may have seen real incomes be more affected than those at the top.
A price index for (actual) private consumption for major groups in society (age, income, rural/urban) is necessary if we are to appraise their economic situation. One of the recommendations of the Commission sur la mesure du pouvoir d’achat des ménages (2008) (Commission on the measurement of household purchasing power) in France was to develop consumer price indices for owners of dwellings, for households who rent dwellings and for households who are about to purchase dwellings. A full development of price indices differentiated by socio-economic group requires, however, that different prices be collected for different segments of the population, so that socio-economic aspects are taken into account in the data collection design. This is likely to prove difficult and costly, and should constitute a medium-term research objective—a recommendation that echoes a similar conclusion by the 2002 Panel on Conceptual, Measurement, and Other Statistical Issues in Developing Cost-of-Living Indices in the United States. Such work would not only enhance the quality of deflation procedures, it would also make it easier for citizens to compare their personal situation with some of the income and price data released by statistical offices.
Broader Measures of Household Economic Activity
There have been major changes in how households and society function. For example, many of the services people received from other family members in the past are now purchased on the market. This shift translates into a rise in income as measured in the national accounts and gives a false impression of a change in living standards, while it merely reflects a shift from non-market to market provision of services. Just as we argued that a shift from private to public provision of a particular good or service, or vice-versa, should not affect measured output, so too, a shift of production from household to market production, or vice-versa, should not affect measured output. We noted earlier that, in practice, current conventions do, however, lead to changes in measured income in both instances.
Imagine a two-parent household with two children and an income of 50,000 currency units a year, in which only one parent works full-time for pay and the other specializes in home production. The parent who stays at home does all the shopping, cooks all the meals, does all the cleaning, and performs all the child care. As a result, this household does not need to devote any of its market income to purchasing these services. Now imagine a two-parent household with two children in which both parents work full-time for the same global pay (50,000 a year), and neither parent has any time left over for household production or child care. They must pay for all the shopping, cooking, cleaning and child care out of pocket. Their available income is therefore reduced. Conventional measures treat these two households as if they have identical living standards, but obviously they don’t. Focusing on market production provides a biased picture of living standards—some of the measured increase in market production may simply reflect a shift of the locus of production from households to the market.
To get a sense of how important home production is economically, one has to start by examining how people use their time.
Figure 1.6. provides a first comparison of time spent per household and per day on various activities. Household production comprises time spent on housework, purchasing goods and services, caring for and helping household and non-household members, volunteer activities, telephone calls, mail and email and travel time related to all these activities. “Personal care” consists mainly of sleeping, eating and drinking, whereas leisure was defined to include sports, religious and spiritual activities and other leisure activities.
Based on these definitions, more time is spent on household production in European countries than in the United States, and more time is spent on leisure in Finland, France, Italy, Germany and the United Kingdom than in the United States (
Figure 1.6). Note that some of the classifications are ambiguous, so the results should be read with care. For example, eating and drinking are included in the definition of personal care, whereas, arguably, some eating and drinking is time spent on leisure. The time-use picture would also change if eating time were allocated differently. We conclude that the allocation of specific activities to time-use categories as well as their international comparison leaves room for improvement and harmonization.
If we gloss over these issues, it is possible to come up with an illustrative calculation of the value of household production for France, Finland and the United States. The approach chosen here is simple: the value of the production of household services is measured by its cost. The value of labor is estimated by applying the wage rate of a generalist household worker to the number of hours that people spend on housework. Methodology matters in this context and results can differ markedly, depending in particular on the hypotheses chosen for the valuation of labor and capital. We also lack estimates for productivity changes in household production.
Minutes per day and person, latest year available*
Using normalized series for personal care; United States: 2005, Finland 1998, France 1999, Germany 2002, Italy 2003, United Kingdom 2001.
Source: OECD (2009), Growing Unequal? Income Distribution and Poverty in OECD Countries; Paris.
However, our estimates do provide orders of magnitude. It is apparent, and no surprise in light of previous studies, that imputations for own-account production of household services are sizeable in all countries. Household production amounts to about 35% of conventionally-measured GDP in France (average 1995–2006), about 40% in Finland and 30% in the United States over the same period.
Once one starts thinking about non-market income, one also has to think about leisure. With time spent on generating income (market or non-market), we buy or produce goods and services to meet our needs or for simple enjoyment. Time available for leisure obviously affects well-being. Changes in the amount of leisure over time and differences between countries represent one of the more important aspects of the situation of well-being in these respects. Focusing only on goods and services can therefore bias comparative measures of living standards. This is of particular concern as the world begins to come to terms with environmental constraints. It may not be possible to increase the production, especially of goods, beyond limit, because of the environmental damage that this would entail. Taxes and regulations may be imposed that will discourage production. However, it would be a mistake if, as a result of these measures, we were to conclude that living standards have fallen when leisure time (and environmental quality) has increased. As society progresses, it is not unreasonable to expect people to enjoy some of the fruits of that progress in the form of leisure. Different societies may respond differently to higher living standards, and we do not want to bias our judgments (e.g., of success) against societies that choose to enjoy more leisure.
Measurement of the value of leisure starts, once again, from time-use data. We multiply the average leisure time per day by the working-age population and then by the average wage rate in the economy. Again, this procedure raises many measurement issues, but the purpose here is to show that estimates are feasible and can produce meaningful cross-country comparisons. For the three countries at hand, the value of leisure roughly doubles net household disposable income in nominal terms. More interesting than nominal income levels is the question of how considering leisure affects the measured growth rates of
real income and their comparisons across countries. This is captured in
Table 1.2. It shows the evolution of household income, now adjusted for household work (upper panel) and for household work and leisure (lower panel). For all countries, the new real income measures grow more slowly than the traditional measures of income. When expressed as income per consumption unit (i.e., per household, adjusted for household size), the income growth rates of the three countries turn out to be very similar.
Percentage change at annual rate, 1995–2006
The imprecision associated with the above estimates should be reiterated here. These are orders of magnitude at best and should not be overinterpreted. However, it is clear that the recognition of broader measures of economic activity and of leisure does make a difference to comparisons over time and between countries. More work needs to be done to test methodologies, to single out the most critical parameters and to test the robustness of such measures. Only if there is sufficient confidence in extended measures of income will there be a broader take-up by statistical offices.
More instructive than estimating the rate of change in real income is assessing how household production and leisure bear on the comparison of income
levels across countries. Income levels should be compared in real terms, so we construct currency converters, so-called
Purchasing Power Parities (PPP) that permit comparisons of “full” income (including housework and leisure) across countries.
Figure 1.7 compares three income aggregates for France and the United States. The first comparison uses the established disposable income measure. Here, France’s per-capita income is about 66% of the comparable United States figure. Adding in government-provided services such as health and education narrows the gap to 79%. If, in addition, housework and leisure are accounted for, one ends up with a relative income level of 87%.
Distribution of Full Income
It was argued earlier that measures of average income should be accompanied by measures that also provide distributional information. The rationale for examining income distribution holds for market income, but also for broader measures, such as full income. The recognition of the own-account production of household services and leisure affects aggregate measures of income and production, but may also change the established picture of income distribution.
Figure 1.7. Real income per capita in France compared to the United States, 2005 United States=100
Developing distributional measures of full income is, however, a formidable task. The most difficult challenge is to allocate to various groups those income flows that have been imputed at the macro level when comprehensive measures of income were derived, for example, imputed rents from own-occupied housing. Other imputations for own-account services produced by households also fall under this category, as do the distributional effects of government services that are provided in kind.
Again, measurement difficulties should not prevent us from developing a more comprehensive picture of the distribution of income and wealth. The distribution of full income should be firmly anchored in the research agenda.
Main Messages and Recommendations
Recommendation 1: Look at income and consumption rather than production.
GDP is the most widely-used measure of economic activity. There are international standards for its calculation, and much thought has gone into its statistical and conceptual bases. But GDP mainly measures market production, though it has often been treated as if it were a measure of economic well-being. Conflating the two can lead to misleading indications about how well-off people are and entail the wrong policy decisions. Material living standards are more closely associated with measures of real income and consumption—production can expand while income decreases or vice versa when account is taken of depreciation, income flows into and out of a country, and differences between the prices of output and the prices of consumer products.
Recommendation 2: Consider income and consumption jointly with wealth.
Income and consumption are crucial for assessing living standards, but in the end they can only be gauged in conjunction with information on wealth. A vital indicator of the financial status of a firm is its balance sheet, and the same holds for the economy as a whole. To construct the balance sheet of an economy, we need comprehensive accounts of its assets (physical capital—and probably even human, natural and social capital) and its liabilities (what is owed to other countries). Balance sheets for countries are not novel in concept, but their availability is still limited and their construction should be promoted. There is also a need to “stress test” balance sheets with alternative valuations when market prices for assets are not available or are subject to bubbles and bursts. Measures of wealth are also central to measuring sustainability. What is carried over into the future necessarily has to be expressed as stocks—of physical, natural, human or social capital. Here too the right valuation of these stocks plays a crucial role.
Recommendation 3: Emphasize the household perspective.
While it is informative to track the performance of economies as a whole, trends in citizens’ material living standards are better followed through measures of household income and consumption. Indeed, the available national accounts data shows that in a number of OECD countries real household income has grown quite differently from real GDP, and typically at a lower rate. The household perspective entails taking account of payments between sectors, such as taxes going to government, social benefits coming from government and interest payments on household loans going to financial corporations. Properly defined, household income and consumption should also reflect the value of in-kind services provided by government, such as subsidized health care and educational services.
Recommendation 4: Give more prominence to the distribution of income, consumption and wealth.
Average income, consumption and wealth are meaningful statistics, but they do not tell the whole story about living standards. For example, a rise in average income could be unequally shared across groups, leaving some households relatively worse-off than others. Thus, average measures of income, consumption and wealth should be accompanied by indicators that reflect their distribution. Ideally, such information should not come in isolation but be linked, i.e., one would like information about how well-off households are with regard to all three dimensions of material living standards: income, consumption and wealth. After all, a low-income household with above-average wealth is not necessarily worse-off than a medium-income household with no wealth. The desirability of disposing of information on the “joint distribution” of dimensions will be encountered once again in Recommendation 3 of the chapter on the quality of life.
Recommendation 5: Broaden income measures to non-market activities.
There have been major changes in how households and society function. For example, many of the services people received from other family members in the past are now purchased on the market. This shift translates into a rise in income as measured in the national accounts and may give a false impression of a change in living standards, while it merely reflects a shift from non-market to market provision of services. Many services that households produce for themselves are not recognized in official income and production measures, yet they constitute an important aspect of economic activity. While their exclusion from official measures reflects uncertainty about data more than it does conceptual dissent, more and more systematic work in this area should be undertaken. This should start with information on how people spend their time that is comparable both over the years and across countries. Comprehensive and periodic accounts of household activity as satellites to the core national accounts should complement the picture.