Which is used to measure directly the average standard of living across countries?

12.2.2 From the Household's Disposable Income to the Standard of Living at the Individual Level: The Statistical Approach

The statistical notion of living standard (or “equivalent income” or “income per consumption unit”) is intended to make comparable the economic well-being of households of different sizes and composition. Its quality as a proxy for economic well-being is much debated (see also Decancq et al., 2014, Chapter 2, in this volume), but for the time being, it remains the most often used and the basis for the measurement of poverty thresholds (or, in the United States, poverty lines).

The standard of living is a construction based on the disposable income (or consumption) of a one-person household taken as a measure of his or her economic well-being. When the household counts more than one person, it is measured so as to take into account the fact that the needs of two (three, etc.) individuals living together are less than twice (three times, etc.) those of one person living alone because of the economies of scale that result from sharing a dwelling and durable and consumption goods and the benefits of the household's production. The additional income needed to keep the household at the same level of economic well-being when additional members are included in the household is difficult to measure because individuals’ consumption within households is not observed. It is most often estimated on the basis of the observed expenditure of households of different sizes and demographic composition. These estimations result in “equivalence scales,” which give a weight, assumed to reflect the additional income needed relative to a one-person household, to each additional household member. Whatever the equivalence scale actually implemented, the weight of any additional member is less than 1 (because, as mentioned above, adding a second person to a one-person household does not double the needs of the household). The dominant statistical approach to the standard of living (or equivalent income) currently uses the so-called “OECD-modified” equivalence scale,5 which gives a weight of 0.5 to an additional adult in the household and a weight of 0.3 to an additional child (a child being an individual younger than 14 years old).

While in this section we are more interested in the assumptions behind the measurement of the household's standard of living and its meaning at the individual level, it is worth briefly illustrating the difference between the household disposable income (INC), a “per head” approach (INCPH), and the standard of living, or “equivalent” disposable income (INCEQ). There is, of course, no difference for a one-person household: in this case INCPH = INCEQ = INC/1. If the household is composed of two adults, then INCPH = INC/2 and INCEQ = INC/(1 + 0.5); if it is a couple with one child, INCPH = INC/3 and INCEQ = INC/(1 + 0.5 + 0.3); INCEQ is always greater than INCPH; the difference accounts for economies of scale. Leaving aside a possible debate on the weightings, it is reasonable (and widely accepted) that INCEQ is a better basis than INC or INCPH for comparing the level of economic well-being between these three households. But current statistical practice goes a step further, since each individual in a given household is considered to achieve the level of economic well-being he or she would achieve if living alone with an income equal to the household's equivalent income; in other words, all the household's members have the same living standard—that of the household. As pointed out by Woolley and Marshall (1994): “The standard approach solves the problem of measuring resource distribution within households by ignoring it” (p. 429). In turn, this practice raises many questions about the actual meaning of indicators at the individual level (including indicators of poverty) based on the household's equivalent income: if one of the household members holds back some or all of his or her income from the common pool, or if the pooled income is not equally distributed among the household members (or if the equivalence scale does not allow for economies of scale to be “distributed” at individual level; see also Browning et al., 2006a), the approach is much less relevant. This highlights how essential the assumptions of income pooling and equal sharing within the household are both necessary conditions for using equivalence scales (Lise and Seitz, 2011) and deriving individual-level indicators from variables measured at the household level. The standard approach also results in a measure of individual well-being, which, by construction, ignores the possibility of inequality within the household and, by construction, makes intrahousehold inequality virtually impossible to assess.

11.4.3 PPP Exchange Rates

International comparisons of living standards require the use of PPP exchange rates to convert national currencies into a common numéraire.17 Two standard sets of PPPs are publicly available: those produced by the International Comparison Program (ICP) of the World Bank (World Bank, 2008) and those produced by the Penn World Tables (PWT), which also uses the underlying price survey data collected by the 2005 ICP.18 PPPs for years before and after the “benchmark” year 2005 are derived from each country's domestic price indices.

The price surveys undertaken for the 2005 ICP were both more detailed and more representative globally than in previous rounds of the ICP. China had never taken part in an ICP before the 2005 round, and India had not taken part since 1985, but both countries were surveyed in the 2005 ICP. Previous estimates of PPPs were therefore based on imputations. Partly for this reason, the results from the 2005 ICP have in some cases led to dramatic changes in estimated GDP. Both China and India were found to have real GDPs nearly 40% lower than previous estimates19 because prices were found to be higher than previously estimated. In the case of China, at least some of this downward revision appears to have been due to sampling problems: its price surveys took place in cities and their environs and did not cover rural areas. For this reason Chinese prices are likely to have been overestimated, and its real income underestimated. Following Chen and Ravallion (2010), and like Milanovic (2012), we make an adjustment to account for this (described later). Milanovic (2012) found that the revisions in the 2005 ICP make a substantial difference to estimated global inequality, raising the Gini by 4.4–6.1 percentage points over the period 1988–2002 and Theil T by 12.5–16.4 percentage points. Other studies that use the 2005 PPPs are Lakner and Milanovic (2013) and Bourguignon (2011), and we discuss their findings later.

Starting from the vector of prices in each country provided by the ICP, the World Bank and PWT use different methods to calculate PPPs. World Bank PPPs are based on the Eltetö–Köves–Szulc (EKS) method, whereas PWT uses the Geary–Khamis (GK) method (both with a variety of adjustments made in the process of estimation).20 EKS arose from a statistical approach to index numbers (Deaton and Heston, 2010) and is a multilateral generalization of the Fisher index for two countries (for further discussion, see Anand and Segal, 2008, p. 71). However, under certain assumptions EKS applied to incomes yields an index of real living standards, or utility, and for this reason Neary (2004) included it as an example of the “economic” approach to index numbers. Under the economic approach it is assumed that observed quantities arise from the optimizing behavior of some representative agent with a well-defined utility function. Real relative incomes measured using EKS PPPs represent relative utility levels when utility is quadratic (i.e., in these circumstances it is a “true” index).

GK, on the other hand, is an example of the “test” or “axiomatic” approach. The GK index has no interpretation in terms of optimizing behavior, but its putative advantage with respect to EKS is that it passes the test, or obeys the axiom, of matrix consistency. That is to say, GK provides a vector of “international prices” for individual goods that enable disaggregation of the economy into subsectors whose values at those prices sum to the total value of the economy. This is not true of EKS, which computes the relative size of aggregate incomes but does not provide a set of international prices with which economies can be consistently disaggregated. If one is interested in analyzing the structure of economies, then matrix consistency would seem to be a useful property. For instance, it is hard to interpret the relative size of manufacturing in two different countries when manufacturing plus nonmanufacturing within each country does not add upto 100% of its economy.

Matrix consistency would seem less relevant, however, when our concern is international comparisons of living standards. In this case, it is the overall value of consumption, not its composition, that concerns us. More important for our purposes is the drawback of the GK method, which is that it suffers from Gershenkron (or substitution) bias. Because consumers tend to substitute away from goods that are relatively expensive and toward goods that are relatively cheap, valuing the output of both country A and country B at country B's prices will lead to an overestimation of the income of country A relative to that of country B. The relative prices arising from the standard GK method more closely resemble those in rich countries than in poor countries, leading to an overvaluation of the incomes of poor countries relative to rich countries and therefore to an underestimation of inequality between countries. Ackland et al. (2004) found that the GK method overvalues the incomes of poorer countries compared to EKS. They regress log per capita GDP from GK on log per capita GDP from EKS and find the slope to be 0.94 and to be significantly less than 1.0. Deaton and Heston (2010) found that the Gini for concept 2 (between-country) global inequality, with per capita GDP as the income concept, is slightly higher using EKS than GK, at 0.533 as opposed to 0.527.

Almås (2012) also found that PWT PPPs underestimate global inequality when accounting for both substitution bias and differences in the quality of goods across countries. However, her estimates are based on the strong assumption that “there is a stable relationship between the budget share for food and household income; i.e., there is a unique Engel relationship for food in the world” (Almås, 2012, p. 1094). Deaton and Heston (2010, p. 5) pointed out that “there are many places in the world, such as North and South India, where there are large differences in consumption patterns of food in spite of only modest differences in relative prices.”

Neary (2004) presents a method that he denotes “Geary–Allen International Accounts” (GAIA) for constructing PPPs that is “economic” in the sense of being based on the assumption of optimizing behavior and therefore does not suffer from substitution bias, but that also satisfies a form of matrix consistency. However, the form of matrix consistency satisfied is not the form that GK satisfies; the sectoral quantities that sum to the value of the whole economy are not the actual observed sectoral quantities, but virtual quantities that a reference consumer, whose preferences are estimated from the data, would have chosen. So it is also the case in the GAIA method that observed manufacturing plus observed nonmanufacturing within an economy will not, in general, add upto 100% of the economy.

The theoretical advantage of GAIA over EKS is that it is a “true” index (i.e., produces estimates of relative real incomes that are consistent with optimizing behavior) for a wider range of utility functions. But because all such indices make the false assumption of identical tastes in all countries worldwide, this seems a rather limited benefit. EKS, on the other hand, has the advantage of being relatively transparent. Although GAIA requires the estimation of a demand system, the EKS exchange rate for a country is simply the geometric mean of that country's Fisher price indices relative to every other country and, as already mentioned, has a natural statistical interpretation that is attractive to national income accountants if not to consumer theorists (Deaton and Heston, 2010).

In our calculations that follow, we use the EKS-based World Bank consumption PPPs from the 2005 ICP. Following Chen and Ravallion (2008, 2010) we make the following adjustments. For both India and China, where the survey data are provided separately for rural and urban strata, we deflate urban incomes relative to rural incomes from price indices used for the construction of domestic urban and rural poverty lines. For India we assume that the World Bank estimated PPP is a weighted average of the urban and rural PPPs. For China we assume that the reported PPP is for urban areas and adjust rural prices downward. This is because the price surveys in China in 2005 were restricted to 11 metropolitan areas, which did not include any rural areas (Chen and Ravallion, 2010). The result is a lower overall price level for China, and thus higher average living standards, than those implied by the use of the 2005 ICP.

A limitation to all standard PPP estimates is that they assume all households within a country face the same price level for their expenditure basket. This may be problematic for at least two reasons. First, urban and rural areas typically have different price levels, and although we have taken this into account for China and India, where the urban and rural price surveys are distinct, it is not possible to do so for most countries. Second, different quantiles of a national income distribution will typically consume different baskets of goods and services,21 and hence face different costs of living. For instance, the poor may face higher unit costs for a good because they have to buy it in smaller quantities. Moreover, they purchase goods in different proportions from the nonpoor so the prices of goods will have different expenditure weights for them. At the other end of the distribution, the very rich (such as those captured by the top income data) may tend to buy more goods from outside their country of residence, to which market exchange rates would apply. But to the extent that the very rich spend their income on nontradable goods and services—for example, country estates, urban mansions, and domestic labor within their country of residence—PPPs with different expenditure weights may be more appropriate than market exchange rates.

4.5.3 Indicator of Standard of Living

The third HDI component, standard of living, was measured by real Gross Domestic Product (GDP)af per capita in Purchasing Power Parity Dollars (PPP$)ag between 1990 and 2010. It was replaced with Gross National Income (GNI)ah per capita in constant PPP$.ai Many limitations of the GDP were pointed out over the years—starting with the limitations inherent in its own construction by excluding those goods and services not traded in markets, ignoring household productive activities such as taking care of own children, housekeeping, and food production for own consumption. Such activities are omnipresent and universally important, especially in developing countries. GDP includes all effects of economic activities, but ignore the costs of activities such as those that cause pollution.

The GNI per capita replaced the GDP per capita in the HDI in 2010. In a globalized world, differences can be large between the income of a country’s residents (GNI) and its domestic production (GDP). Some of resident earnings are sent abroad, some residents receive international remittances and some countries receive sizeable aid flows. In other words, GNI expresses the income accrued to residents of a country, including some international flows, and excluding income generated in the country but repatriated abroad. Thus, GNI is a more accurate measure of a country’s residents’ ability to instrumentally use income to expand capabilities. For example, because of large remittances from abroad, GNI in the Philippines greatly exceeds GDP (for about 10% in 2019). In Ireland's case, GDP is actually larger than GNI because of repatriation of profits by companies resident in Ireland and repayments on the foreign elements of Ireland’s national debt.

The 2009 Report by the Commission on the Measurement of Economic Performance and Social Progressaj mentions that “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 Report then recommends using income or consumption rather than production; moreover, the Report suggests the use of wealth to take into account the temporal dimension. Income misses the various dimensions of wealth including financial and real wealth and natural wealth, which is central to evaluating sustainability. Foster (2013) entertained the idea of using one flow variable to indicate what is available for transformation into capabilities and to generate standard of living now, and one stock variable to indicate what is saved and transferred to the next generation.

The Commission also recommended that instead of using the averages of income, consumption, and wealth, it might be more insightful to use the medians as measures that pertain better to the “typical” individual or household than the mean (which may be distorted by extreme values at either end of the distribution). However, in practice, moving from means to medians may be difficult given that medians require microdata from household income surveys, whereas macroeconomic measures from the national accounts are based on a range of different macroeconomic sources and may not pertain to the same population. Many of the important properties of means, as well as the theories developed around the concepts such as welfare standards, may not translate directly to the median-based measures.

The standard of living dimension was settled for GNI per capita to account for income generated in the country plus some income received from abroad minus some sent abroad. Data on household income or consumption are neither readily available nor harmonized.

24.5.1 Extended Income, Consumption and Indirect Taxes

Although disposable income is the most-used indicator of living standard, it is widely recognized that economic well-being is a multidimensional concept (see Chapter 2). The economic value of the consumption of goods and services, including interhousehold transfers, in-kind benefits, and homeowners’ imputed rent related to the main accommodation, is often considered a better indicator than income when measuring individual well-being on both theoretical and pragmatic grounds (Meyer and Sullivan, 2011). The exclusion of consumption expenditure and noncash income from empirical studies of the redistributive effect of tax-benefit systems might also hamper cross-country comparison given the different degree of monetization of the economy across countries. Moreover, the distributional impact of policy changes may be rather different if noncash incomes and indirect taxes are included, with important implications for the design of policies aiming to fight poverty and social exclusion, because such an omission may lead to imperfect targeting and misallocation of resources. Notwithstanding their importance, most microsimulation models do not include either in-kind benefits or indirect taxes, mainly due to data limitations.

In European countries, in-kind benefits, such as services related to child and elderly care, education, health, and public housing, represent about half of welfare state support and contribute to reducing the inequality otherwise observed in the cash income distribution. The economic value of public in-kind benefits can be imputed at individual and household levels on the basis of per capita spending, considering the average cost of public services (such as providing care and education services), the gain from paying below-market rent or no rent at all for public housing, or the risk-related insurance value approach that considers public health care services equivalent to purchasing an insurance policy with the same cost for individuals who have the same sociodemographic characteristics. See Aaberge et al. (2010b) and Paulus et al. (2010) for empirical evidence across European countries and for methodological insights on the derivation of needs-adjusted equivalence scales that are more appropriate for extended income. However, survey data usually do not include enough information to simulate changes in the value of the benefit due to policy reforms, nor do they take into account the real utilization by the individual, the quality of the public service, or the discretion in the provision usually applied by local authorities (Aaberge et al., 2010a).

A more comprehensive measure of individual command over resources should include the income value of home ownership as well. This is because the consumption opportunities of homeowners (or individuals living in reduced or free rent housing) differ from those of other individuals due to the imputed rent that represents what they would pay if they lived in accommodation rented at market prices. The inclusion of imputed rent in microsimulation models is becoming more common due to the refinement of different methods for deriving a measure of imputed rent (Frick et al., 2010) and also a renewed interest in property taxation. From a cross-country perspective, Figari et al. (2012b) analyze the extent to which including imputed rent in taxable income affects the short-run distribution of income and work incentives, showing a small inequality-reducing effect together with a nontrivial increase in tax revenue. This offers the opportunity to shift the fiscal burden away from labor and to increase the incentive for low-income individuals to work.

Indirect taxes typically represent around 30% of government revenue. With only a few exceptions, household income surveys providing input data for microsimulation models do not include detailed information on expenditures either, preventing micro-level analysis of the combined effect of direct and indirect taxation. The solution usually adopted to overcome this data limitation is to impute information on expenditures into income surveys (Sutherland et al., 2002). Decoster et al. (2010, 2011) provide a thoughtful discussion of the methodological challenges and a detailed explanation of the procedure implemented in the context of EUROMOD for a number of European countries. Detailed information on expenditure at the household level is derived from national expenditure surveys, with goods usually aggregated according to the Classification of Individual Consumption by Purpose (COICOP), identifying, for example, aggregates such as food, private transport, and durables. The value of each aggregate of expenditure is imputed into income surveys by means of parametric Engel curves based on disposable income and a set of common socioeconomic characteristics present both in income and expenditure datasets. In order to prevent an unsatisfactory matching quality in the tails of the income-expenditure distributions, a two-step matching procedure can be implemented by first estimating the total expenditures and total durable expenditures upon disposable income and sociodemographic characteristics and then predicting the budget share of each COICOP category of goods. Moreover, the matching procedure takes into account the individual propensity for some activities, such as smoking, renting, using public transportation, and education services, which are not consumed by a large majority of individuals. Individual indirect tax liability is then simulated according to the legislation in place in each country, considering a weighted average tax rate for each COICOP category of goods imputed in the data.

Most microsimulation models that include the simulation of indirect taxes rely on the assumption of fixed producer prices, with indirect taxes fully passed to the final price paid by the consumer. To relax such an assumption one should go beyond a partial equilibrium framework and link the microsimulation models to macro models (see Section 24.3.4) in order to consider the producer and consumer responses to specific reforms or economy-wide shocks. There is some variety in the ways in which the models deal with the estimation of changes in spending patterns due to the simulated reforms (Capéau et al., 2014). Some models simulate only a nonbehavioral first round impact (i.e., quantities or expenditures are kept fixed at the initial level), and others estimate partial behavioral reactions taking into account the income effect on demand for goods and services by means of Engel curves (Decoster et al., 2010) or even full demand systems accounting for the real income effect and the relative price effects (Abramovsky et al., 2012).

The inclusion of indirect taxes also raises the question of how to measure their incidence. Table 24.1 shows the incidence of indirect tax payments for three European countries expressed as a percentage of disposable income and as a percentage of expenditure, by decile of equivalized disposable income. In the first case (see the left panel of Table 24.1), the regressivity of indirect tax payments is clear: poorer individuals pay a larger proportion of their income in indirect taxes compared to richer individuals, mainly due to a larger propensity to consume or even dissaving reflected by average expenditures exceeding incomes for the individuals at the bottom of the income distribution (Decoster et al., 2010). However, survey data might suffer from measurement error, in particular from income underrecording (Brewer and O'Dea, 2012), which could give a misleading snapshot of the income-consumption pattern at the bottom of the income distribution. In the second case (i.e., the right panel of Table 24.1), indirect tax payments are progressive, and poorer individuals pay a slightly smaller proportion of their total expenditure in VAT and excises compared to richer individuals. The main reason for this is that the goods that are exempt from VAT or subject to a lower rate (e.g., food, energy, domestic fuel, children's clothing) represent a much larger share of the total spending of poorer individuals than of richer individuals (Figari and Paulus, 2013). The distributional pattern of the indirect taxes being regressive with respect to disposable income and proportional or progressive with respect to expenditure reinforces, on empirical grounds, the importance of the choice of the measurement stick that should be used as a benchmark in the welfare analysis (Capéau et al., 2014; Decoster et al., 2010).

Table 24.1. Incidence of indirect tax payments

Notes: Decile groups are formed by ranking individuals according to equivalized household disposable income, using the modified OECD equivalence scale.

Source: Figari and Paulus (2013), based on EUROMOD.

The potential of microsimulation models that are capable of simulating direct and indirect taxes within the same framework is reinforced by the renewed interest in the tax shift from direct to indirect taxation in order to enhance the efficiency of the tax system (Decoster and Van Camp, 2001; Decoster et al., 2010). In particular, microsimulation models have been used to assess the distributional consequences of a “fiscal devaluation,” a revenue-neutral shift from payroll taxes toward value-added taxes that might induce a reduction in labor costs, an increase in net exports, and a compression of imports, with an overall improvement in the trade balance (de Mooij and Keen, 2013; European Commission, 2013).

Two general considerations arise from the use of microsimulation models for the analysis of the redistributive effects of indirect taxes. On the one hand, the actual degree of regressivity of indirect taxes might be less than that observed if surveys tend to underreport income more than consumption at the bottom of the income distribution (Brewer and O'Dea, 2012; Meyer and Sullivan, 2011). On the other hand, a more systematic use of simulated income values, as generated by a microsimulation model rather than as observed in the data, can help in solving the underreporting of income values, closing the gap between reported income and consumption and providing a more robust indicator of living standards for those with a low level of resources.

5.2.2 The Neolithic

Agriculture was independently invented at least seven times, starting with the Levant, around 12,000 years ago. During the so-called Neolithic Revolution, our ancestors abandoned nomadic hunting and gathering and became settled farmers. This new lifestyle spread rapidly, averaging about a kilometer per year from the Middle East into Europe (Pinhasi et al., 2005), and eventually became by far the dominant lifestyle of humans on all continents, wherever there was enough rainfall and sunshine to support a crop.

The most important change was of course agriculture itself. As long as our ancestors had remained hunter gatherers, their food supply had been ultimately limited by the amount available naturally. They could become more efficient in collecting it, or learn to consume new sources, but decreasing returns to extra effort set in rather early. With farming, humans had a way of transforming labor into food, most obviously by planting some of the previous year's crops in suitable conditions, but also by removing competing inedible weeds, or diverting small streams to irrigate dry land. This meant that food production could expand almost linearly with population size, at least as long as there was new land to put into cultivation.

Once land in a particular location became scarce, part of the population could take some of those seeds and set up a new farming village in a new location. If the chosen destination was inhabited, farmers could use their greater numbers to push aside the previous occupants. In some cases, perhaps the expansion could occur through intermarriage, with the offspring of farmers bringing the seeds and farming techniques, and the current inhabitants providing the land. These twin processes led to the rapid expansion of agriculture into virtually all lands that could support it. As this was going on, rapid coevolution of farmers and their crops meant that the cultivated species soon became domesticated, meaning that their genetics changed to make them easier to cultivate and harvest (Harlan, 1998).

The Neolithic Revolution is of interest to economists for a number of reasons. First of all, unlike most inventions of comparable importance, agriculture was invented multiple times (Purugganan and Fuller, 2009), which means that it is possible to gain some idea of what characteristics made this crucial discovery more likely. Second, the arrival of agriculture made densely populated civilizations, with highly hierarchical institutions possible. The period is therefore a logical starting point for studies of the long run persistence of institutions and culture. Third, in most cases, the first farmers were noticeably shorter than the last hunter-gatherers, and also showed a decrease in general health (Cohen and Armelagos, 1984), a finding that has prompted a number of explanations.

The earliest modern theory for the Neolithic (based on actual archaeological excavations rather than theoretical speculation) was arguably presented by Vere Gordon Childe (1936). He posited that due to a drier climate at the end of the last glacial period, humans, plants, and animals had to necessarily live in much closer proximity, and that this cohabitation made agriculture possible, and ultimately inevitable.

After WWII, work by Braidwood showed that the desiccation event, which Childe had found limited evidence for, had not in fact occurred in the areas with the earliest evidence for agriculture. Boserup (1965) theorized that as population densities increased, humans had to adapt by finding new sources of food, and some of these efforts resulted in agriculture. Note that in this formulation, population growth is essentially constant.

In contrast, later work by Binford (1968) and Flannery (1973) assumed that a population of hunter-gatherers would over time reach an equilibrium with its environment, leveling off at a fixed sustainable population size. Under these conditions, there was no pressure to develop agriculture. This pressure came instead in the wake of climate change, which forced the local populations to move into more marginal zones. These areas were locally overpopulated and had the right incentives for developing farming.

Diamond's Guns, Germs, and Steel (1997) proposed a different climate change story, ascribing the birth of agriculture to the abundance of plants that could be domesticated easily, which became more abundant at the end of the last Ice Age. As described in Section 5.4 below, he then extended this reasoning into a theory for the different development paths undertaken by civilizations on different continents.

A parallel research program has investigated the wellbeing of our ancestors as agriculture was invented and adopted. Until the mid 1960s, researchers thought hunting and gathering was a universally precarious subsistence strategy, and agriculture therefore a clear improvement. In 1966, Marshall Sahlins introduced his Original Affluent Society hypothesis, which noted that most hunter gatherers that had survived to be studied by anthropologists appeared anything but precarious in their day to day life. Indeed, they appeared to be able to satisfy their primary needs with a three-day workweek, which left them time to socialize and play. Further, he argued that this was also likely to be true for primitive hunter gatherers. While he conceded that some hunters in marginal locations were indeed at times at risk of starvation, humanity is still all too acquainted with starvation, so that hunger can hardly be said to be a trait distinctive of hunter gatherers.

The comparatively leisurely conditions were largely confirmed by Cohen and Armelagos (1984), which collected and analyzed skeletal evidence from a number of distinct excavation campaigns in both the New and Old World covering the transition from hunting and gathering to agriculture in each location. The nearly unanimous finding was that as agriculture was adopted, our ancestors became shorter and in worse general health, suggesting that they ate less, worked more, and were more exposed to infectious diseases. Diamond (1997) argued that the initial increase in food availability was more than matched by runaway population growth, which resulted in worse health conditions.

Economists have tested these theories for the reduction in average standard of living and proposed their own. Since this chapter focuses on archaeological data, we are going to focus mainly on the contributions with an explicitly empirical component (this however covers the vast majority of the economic literature of the Neolithic of the last two decades).1

The literature on the Neolithic transition can be divided into two parts: the first one focuses on the factors that made the invention and the adoption of agriculture more likely and is discussed below while the second part investigates the long-term impact of agriculture on human political and economic disparities and will be discussed in the next section.

The Diamond (1997) hypothesis, as stated in the book, relied on a fairly general overview of the data on domesticable species by comparing Eurasia with Africa, the Americas, and Australasia. Olsson and Hibbs (2005) sought to systematically test this hypothesis at the country level, using a sample of 112 countries. The data on plant domestication (the same used by Diamond) comes from Blumler (1992). It should be noted that this data is specified for only nine macro regions worldwide and is therefore only appropriate for very large-scale analysis such as this one.

Ashraf and Michalopoulos (2015) were able to use much finer grained data to test their novel theory for the Neolithic Revolution, which built on features of the Broad-Spectrum Revolution and climate change hypotheses. As in the theories of Binford and Flannery discussed above, they argue that where there was too little climate variability at the scale of decades or centuries, human populations would have little incentive to innovate. But as in e.g. Richerson et al. (2001), too much climate variability made it too risky to farm, since harvests of individual species would have been too variable. They therefore reason that farming was more likely to be invented in areas that were experiencing intermediate climatic volatility, and they test this hypothesis at the global scale using gridded climate data for the period 1901-2000 CE, as provided by Mitchell et al. (2004),2 and using the country data on Neolithic adoption from Putterman and Trainor (2006). They find that, as predicted, areas with intermediate levels of volatility were the first to adopt agriculture. The analysis is then repeated, with similar results, for the 765 archaeological sites in Western Eurasia with C14 dates reported in the Pinhasi et al. (2005) dataset. Their datasets are available online from the journal website and are a logical entry point for researchers interested in probing the causes for the differential timing of the Neolithic.

Matranga (2020) also argues that climate volatility was the driving factor for the invention and spread of agriculture, but focuses on predictable seasonal changes within a year, rather than random differences across years. The difference is important because seasonal changes can be rather easily smoothed by storing food, as long as a suitable storage technology is present. The author shows that well understood cyclical changes in the parameters of Earth's orbit caused a marked increase in climate seasonality around the time when humans first became sedentary, and then started to farm. He argues that this was caused by climate seasonality making it essential to store wild foods for winter, even before agriculture was developed. Since food stores are too heavy to carry on nomadic migrations, this meant they had to become sedentary, which in turn facilitated the development of agriculture. The empirical part uses the same outcome data as Ashraf and Michalopoulos (2015),3 but uses reconstructed panel climate data for the last 22,000 years from He (2011) for the explanatory variables, instead of cross-sectional climate data from the present. The author shows that agriculture first appeared at times and in places where climate was very seasonal, and that more seasonality also favored the spread of farming techniques. Further, he shows that areas that had a lot of different microclimates in a very short radius (accessible by a sedentary population) started farming earlier.

Why Consider a Career in Toxicology?

4.4 Beyond GDP

It has long been recognized that GDP is an imperfect measure of living standards. Pollution, leisure, life expectancy, inequality, crime, and even freedom are some of the factors that are incorporated imperfectly, if at all, in GDP. Various attempts have been made over the years to repair some of the omissions, or at least to judge how important they might be. Examples include Nordhaus and Tobin (1972), Becker et al. (2005), Fleurbaey and Gaulier (2009), and Stiglitz et al. (2009).

Jones and Klenow (2015) extend this literature by using a standard utility function and a “behind the veil of ignorance” approach to construct a consumption-equivalent welfare measure that values consumption, leisure, mortality, and inequality for a range of countries. Table 5 shows their baseline findings for a high-quality sample of countries for which household survey data can be used to compute welfare. The key finding comes in two parts. First, Western European countries like the United Kingdom and France have much higher living standards than their GDPs indicate. For example, compared to the United States, France has higher life expectancy, more leisure per person, and lower inequality of both consumption and leisure, and these differences make a substantial difference: whereas GDP per person in France is only about 2/3 of that in the United States, consumption-equivalent welfare is around 92% of the US level. Second, while many rich countries are richer than we might have thought, the opposite is true for poor countries. Life expectancy and leisure tend to be lower and inequality tends to be higher, all of which reduce welfare relative to GDP. As just one example, South Africa's GDP per person is about 16% of the US level, but consumption-equivalent welfare is only 7.4% of that in the United States.

Table 5. Beyond GDP: Welfare across countries

Notes: The consumption-equivalent welfare numbers in the first column use a conventional utility function to “add up” the contributions from consumption, leisure, mortality, and inequality and express them in a consumption-equivalent manner. The income column reports GDP per person. The “decomposition” columns report an additive decomposition of the log difference between welfare and income.

Source: These numbers are taken from table 2 of Jones, C.I., Klenow, P.J. 2015. Beyond GDP: Welfare across countries and time. Stanford University, unpublished manuscript, and are based on data from household surveys in each country, from the World Bank (for mortality), and from the Penn World Tables 8.0 for a year close to 2005.

In terms of growth rates, declining mortality has the largest impact: in most countries of the world—the notable exception being in sub-Saharan Africa—declining mortality has raised consumption-equivalent welfare growth substantially. In the United States and Western Europe, for example, growth rates since 1980 are arguably understated by around a full percentage point because of this factor.

A.4 The effect of increased resource scarcity

Let us study the implications for equilibrium household size and the standard of living consequent upon small exogenous shifts in the functions α(M), β(M) and γ(M). We take it that prior to the shifts inequality (A.7) holds. The perturbations will be taken to be sufficiently small so that (A.7) continues to hold in the new equilibrium.

Consider first the case where the perturbation consists of small upward shifts in α(M) and γ(M) and a small downward shift in β(M). Notice that if (A.8) holds, both n* and y* would be marginally smaller in consequence of the perturbation. This is the case we would expect intuitively: a small increase in resource scarcity results in poorer, but smaller, households.

Now consider the case where (A.9) holds. Suppose the perturbation consists of small upward shifts in each of the functions α(M), β(M) and γ(M). We can so set the relative magnitudes of the shifts that the small increase in resource scarcity results in poorer, but larger, households, that is, y* declines marginally but n* increases marginally. This is the timeless counterpart of the positive feedback mechanism between population size, poverty and degradation of the natural-resource base that was discussed in Section 8.4. Such a feedback, while by no means an inevitable fact of rural life, is a possibility. In this chapter I have argued that evidence of the experiences of Sub-Saharan Africa and northern Indian sub-continent in recent decades are not inconsistent with it.

21.2 Overview of Japanese consumer cooperatives

Japanese consumer co-ops can trace their history to the late 19th century. The first cooperative shops, which modeled the Rochdale Equitable Pioneer’s Society, were set up in Tokyo, Osaka, and Kobe in 1879, just 12 years after the Meiji Restoration. After years of trial and error in the course of industrialization, three types of co-ops emerged, namely co-ops attached to companies/factories for their employees, worker-oriented co-ops associated with a radical labor movement, and citizen co-ops organized by middle-class people. At the end of the Second World War, they were mostly destroyed; the left-wing co-ops were liquidated by the militaristic government, while the neutral ones were deprived of trading licenses and finally destroyed by air raids. The consumer movement had to start from scratch, although the legacy of the movement was inherited and kept alive through cooperative leaders.

The history of consumer co-ops in the postwar era can be divided into four epochs; the mushrooming of buying clubs seeking scarce food in the 1940s; the emergence of worker-oriented co-ops sponsored by trade unions in the 1950s, the flourishing of consumerism-based citizen co-ops during the 1960s through to the early 1990s; and then stagnating growth since the mid-1990s onward (Kurimoto, 2017).

Just after the Japanese surrender in 1945, the entire economy fell into chaos due to the massive destruction of production and distribution facilities. Faced with this, the US occupation introduced drastic reforms to democratize the economy such as the agrarian reform, legitimization of unions, and the antimonopoly legislation. The rationing system for staple food as a part of a wartime supply mechanism could not function effectively to support the daily life of consumers. Most of the urban population faced a serious shortage of food and daily necessities as well as being subject to rampant inflation. They had to rely on black markets, visit farmers to barter their valuables for food or starve to death. Under such circumstances, numerous buying clubs were formed by residents in the districts or by workers in the factories/offices. Their mission was to procure food for members from farms/factories and so were often called “buying associations.” They had mushroomed at an incredible speed; more than 6500 co-ops were operating in 1947. However, most of them lacked effective management and support systems. They largely collapsed soon after the rationing system began functioning. The number of co-ops had swiftly shrunk to roughly 1000 by 1948, when the Consumer Co-operative Act was enacted. As such the first boom of consumer co-ops came to a quick end. The Japanese Consumer Co-operative Union (JCCU) was set up as a national body in 1951 to help the revival of consumer co-ops under the new legislation.

In the 1950s, trade unionism entered an expanding phase and started supporting “worker’s welfare businesses” to supplement its main function of collective bargaining. In this process, worker-oriented co-ops were created to undertake economic activities to meet workers’ various needs under the sponsorship of trade unions. The local trade union councils assisted to set up “worker-led community consumer co-ops” in the 1950s, enabling these co-ops to operate relatively large stores in comparison and competition to small retailers. These trade union-based co-ops provided a variety of food and consumer goods in local cities prior to the advent of supermarkets and achieved quick success by automatically enrolling unionists as co-op members and attracting a wider range of consumers. This triggered the reaction of retailers, leading to intense anti-cooperative campaigns. However, their success was short-lived due to the lack of management skills or member education; they failed to compete with supermarkets introduced by progressive retailers in the late 1950s. Some of them disappeared, while others transformed into citizen co-ops in the 1960s. On the other hand, trade unions and consumer co-ops worked together to set up worker-oriented co-ops such as labor banks, insurance co-ops, and housing co-ops. Today labor banks and insurance co-ops established themselves as “workers’ welfare enterprises” in which trade unions retain a strong influence.

The rapid economic expansion in Japan since the late 1950s drastically enhanced the standard of living, while it brought a massive migration of people to large cities. This process was synchronized with revolutionary changes in consumption and distribution patterns. Manufacturers established the mass production and wider distribution of packaged groceries using chemicals as food additives, which often caused serious health problems. Fresh food was now processed in more industrialized ways, making massive use of pesticides and antibiotics. Consumers were concerned with these chemicals as well as the high inflation, misleading labeling, and air/water pollution. Such circumstances gave momentum to the consumer movement seeking safer foods and a healthier environment. One of the responses that emerged was “citizen co-ops” backed by housewives in the 1960s and 1970s. Some co-ops invented the “joint buying” or home delivery to Han groups1 that made collective orders and received products on a weekly basis. Citizen co-ops were set up in all seats of prefectural government until 1980, attracting a wide range of consumers; the membership of the consumer cooperative movement expanded from 2 million to 14 million, while the turnover has grown 10 times between 1970 and 1990. Thus the Japanese model of consumer cooperatives was created with housewives as a driving force.

The consolidation of the buying functions, however, was slow in comparison with their European counterparts as primary co-op societies continued to buy from local suppliers and develop their own CO-OP brand products. In 1958 however, major co-ops established a wholesale federation to pool their buying power at the national level. This federation then integrated local buying functions and finally merged with the JCCU in 1962, aiming at strengthening the central buying and national coordination. During the 1980s, “core” primary co-ops were established through merging smaller societies in many prefectures under the JCCU’s guidance. In the 1990s, co-ops formed regional consortiums, beyond prefectural borders, which now cater for nearly 90% of the overall turnover of co-ops.

Since the mid-1990s, however, the turnover of consumer co-ops has stagnated because of the lingering recession and stiffer competition seeking to get a larger market share. The number of independent retailers continued to decline, but even some of the larger department stores and supermarket chains also failed. As a result, the concentration was intensified; the Aeon Group and Seven & I Holding became the biggest retailers through a number of mergers and acquisitions. To meet the challenges of more women working outside the home and with a more individualized lifestyle, co-ops introduced individual home delivery, partly supplementing and partly replacing joint buying. They also made efforts to strengthen their buying power through regional consortiums. They succeeded in maintaining their overall turnover in the past decade, while the declining store sales has been offset by the growing home delivery sales.

Consumer co-ops are regulated by the Consumer Co-operative Act (1948), which classifies co-ops into categories according to the type of business (retailing, healthcare, insurance, housing, and so on) and areas of operation (communities or workplaces). Retail co-ops provide members with food, nonfood products, and various services. The typical retail co-ops operating in communities are called citizen co-ops, which account for 70% of the total co-op membership. Workplace co-ops operate in companies and government offices to serve employees working in these institutions. Extended workplace co-ops are hybrids of these types and have incorporated local consumers living in the communities adjacent to the institutions. University co-ops and schoolteachers’ co-ops cater to the specific needs of the constituencies within these institutions including students, faculty members, and schoolteachers. Medical or health co-ops provide health and social care services at hospitals and clinics. Insurance co-ops or kyosai provide consumers with life and general insurance policies. Housing co-ops sell or rent mainly collective houses and provide maintenance and/or repair services. Other than these, there are also co-ops that specialize in elderly/child care provision, environmental conservation, and so on.

Japanese consumer co-ops have a three-tiered structure stretching from primary co-ops to prefectural/national unions/federations. According to the Consumer Co-operative Act, 300 or more consumers may establish a primary co-op, while federations can be set up at the secondary and tertiary levels. The prefectural unions and inter-prefectural consortiums are formed by primary co-ops. National federations composed of insurance, housing, and university co-ops also exist. The JCCU with several categories of affiliated members has the dual functions of being a national center for all types of co-ops as well as a national consortium for retail co-ops.