Download The Persisting Influence of GDP: A Critique and Alternatives and more Essays (university) Economics in PDF only on Docsity!
The GDP paradox
Jeroen C.J.M. van den Bergh *
,
Department of Economics and Economic History and Institute for Environmental Science and Technology, Autonomous University of Barcelona, Edifici Cn – Campus UAB, Bellaterra, 08193 Cerdanyola, Spain ICREA, Barcelona, Spain
a r t i c l e i n f o
Article history: Received 8 July 2008 Received in revised form 10 December 2008 Accepted 12 December 2008 Available online 24 December 2008
JEL classification: D E O
Keywords: Bounded rationality Information failure Lock-in Policy indicators Economic impact of GDP information
a b s t r a c t
Despite all theoretically and empirically motivated criticism of GDP as a social welfare and progress indicator, its role in economics, public policy, politics and society continues to be influential. To resolve this paradox, one has to recognize that many economists accept the criticism of the GDP indicator but deny its relevance. This paper evaluates the reasons for denial. This entails five steps: (1) a brief review is offered of the extensive literature show- ing that GDP per capita (growth) is far from a robust indicator of social welfare (progress); (2) the influence of GDP information on economic decisions by firms, consumers, investors and governments is examined; (3) behavioural explanations for a widespread belief in the relevance of GDP are discussed; (4) the customary arguments in favour of the GDP indica- tor are analysed; and (5) proposed alternatives to GDP are evaluated. The paper ends with outlining the implications of giving less attention to GDP information in policy and politics. It is argued that removal of the information failure which GDP represents, in monitoring economic progress and guiding public policy, will lead to decisions and developments being more in line with improving human well-being. Moreover, ignoring GDP information is consistent with a perfectly neutral stance regarding economic (GDP) growth. Indeed, an unconditional anti- or pro-growth imperative acts as an unnecessary constraint on our search for human progress. Ó 2008 Elsevier B.V. All rights reserved.
- Introduction
Gross domestic product (GDP) is the monetary, market value of all final goods and services produced in a country over a period of a year. 2 The real GDP per capita (corrected for inflation) is generally used as the core indicator in judging the posi- tion of the economy of a country over time or relative to that of other countries. The GDP is thus implicitly, and often even explicitly, identified with social welfare – witness the common substituting phrase ‘standard of living’. This approach does
0167-4870/$ - see front matter Ó 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.joep.2008.12.
- Address: Institute for Environmental Science and Technology, Autonomous University of Barcelona, Edifici Cn – Campus UAB, 08193 Bellaterra (Cerdanyola), Spain. Tel.: +34 93581 3747; fax: +34 93581 3331. E-mail address: [email protected] (^1) Also affiliated with the Faculty of Economics and Business Administration, and the Institute for Environmental Studies, Free University, Amsterdam, The Netherlands. Fellow of NAKE and Tinbergen Institute. (^2) Gross domestic product (GDP) is the market output generated within a country’s boundaries by both its citizens and foreigners. Gross national income
(GNI) (formerly gross national product, or GNP) is the output produced by the citizens of a country, irrespective of where production takes place. For most countries the difference between GDP and GNI is not very large (a notable exception is Ireland). Moreover, the conceptual difference does not matter for our discussion here: all the shortcomings of GDP as a welfare indicator identified hereafter apply equally to GNI.
Journal of Economic Psychology 30 (2009) 117–
Contents lists available at ScienceDirect
Journal of Economic Psychology
j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / j o e p
not follow from any theory about GDP as a measure of social welfare, but has grown to become like this in the course of time.^3 For over half a century now, the GDP (per capita) has been severely criticized as not adequately capturing human welfare and progress. All the same, the GDP has maintained a firm position as a dominant economic indicator, which can be regarded as a paradox. Indeed, most economists in business and government, teachers of economics at various levels of education, and journalists, policy makers and politicians (regardless of their political preferences) continue to give much importance to GDP and calling for unconditional GDP growth. An innocent observer might conclude that they are ignorant of the widespread criticism of GDP. To resolve the paradox, or explain its persistence, one has to recognize the ambivalence with which many academic econ- omists approach the criticism of the GDP indicator: they accept it but deny its relevance. This denial comes in two forms: (1) a belief that the size of the impact of GDP information on economic reality is modest; and (2) a belief that despite all criticism the GDP still provides useful information. The first response in fact says that the shortcomings of the GDP indicator do not much affect the direction of economic development, while the second response suggests that the shortcomings are rather unimportant and hardly affect the social optimality of economic development. The arguments in support of either type of response will be examined here. Although a slowly increasing number of economists explicitly recognize the shortcomings of the GDP indicator,^4 a large majority of them seems to accept the role of GDP as a welfare or progress measure in empirical economics and politics as a fact of life. This is not just a paradox, but a very unsatisfactory state of affairs from the viewpoint of assuring a solid empirical information basis for public policy and human development. This article aims to arrive at a clearer, explicit position on what to do with the GDP indicator than earlier writings on this issue have accomplished. In particular, the question will be ad- dressed whether the present use of GDP information in politics and public policy should be continued, whether an adapted or entirely different indicator should replace it, or whether GDP should anyway be ignored as an indicator of welfare or pro- gress (though not necessarily as a model variable). For this purpose, we will briefly evaluate and compare available alterna- tive aggregate indicators. The discussion offered here has immediate relevance for economics and economists, and only indirectly for politics and politicians, or society at large. The reason is that as long as economics has not seriously dealt with the question ‘‘what are the full implications of the sustained criticism of GDP as a progress indicator?”, we cannot expect politicians to act as if the an- swer to this question is already known. The organization of the paper is as follows. Section 2 briefly recaptures the main criticism of GDP as an indicator of wel- fare or progress. This section claims no innovation, but provides a necessary basis for discussions in subsequent sections. Section 3 examines the influence of GDP information on the economy and economic development, by considering impacts on private and public decisions. Section 4 offers a number of behavioural explanations of the widespread belief in the rel- evance of GDP. Section 5 evaluates the customary arguments in favour of the GDP indicator. Section 6 looks into the argu- ments for and against the main empirical measures of social welfare that have been proposed as an alternative to the GDP indicator. Section 7 outlines the implications of giving less importance to GDP information for public policy. Section 8 concludes.
- Shortcomings of the GDP indicator
Since the 1960s, the use and implicit interpretation of GDP (per capita) as a proxy of social welfare – and regarding an increase in GDP as progress – has received much criticism. This criticism has come from some of the most respected econ- omists of the 20th century, including various Nobel laureates. Important contributions include Kuznets (1941), Hicks (1948), Galbraith (1958), Samuelson (1961), Mishan (1967), Nordhaus and Tobin (1972), Easterlin (1974), Hueting (1974), Hirsch (1976), Sen (1976), Scitovsky (1976), Daly (1977), Frank (1985, 2004), Hartwick (1990), Tinbergen and Hueting (1992), Ar- row et al. (1995), Weitzman and Löfgren (1997), Dasgupta and Mäler (2000), Dasgupta (2001), Ng (2003) and Kahneman, Krueger, Schkade, Schwarz, and Stone (2004a). The following provides a very brief summary of the most important criticisms (the Appendix offers more details, additional arguments and supporting references).
2.1. Principles of proper accounting
The use and calculation of the GDP indicator is inconsistent with two principles of good bookkeeping: (i) divide clearly between costs and benefits; and (ii) correct for changes in stocks and supplies. GDP really represents an estimate of the costs instead of the benefits of all market-related economic activities in a country. In addition, GDP does not capture all social costs as it omits external costs.
(^3) It is well-known that the GDP indicator was never developed for the purpose of welfare measurement. In 1665 Sir William Petty produced the first estimate of a national income, namely for England. His work aimed to determine which outlays on warfare could be supported by means of tax revenues. Work in the early twentieth century by Nobel Laureates Simon Kuznets for the USA and Richard Stone and James Meade for the UK allowed for the rapid diffusion of the GDP indicator in economic research and politics. Again, war acted as a stimulating factor: there was a need to determine the production capacity of the Allied Forces just before and during World War II. (^4) Even The Economist (11 February 2006, p. 70) has characterized GDP as ‘‘badly flawed as a guide to a nation’s economic well-being”.
2.7. Environmental externalities and depletion of natural resources
An important subcategory of unpriced effects relates to use of natural resources and the environment. This involves neg- ative external effects as well as goods and services delivered by nature. The presence of externalities means that the current set of market prices insufficiently reflects the total of private and external costs, which makes these prices unreliable signals and thus an incorrect basis for calculation of a social welfare indicator. Moreover, if air, water, or natural areas are being pol- luted any resulting damage does not enter GDP, but when pollution is being cleaned up this will increase GDP. In addition, the capital depreciation associated with environmental change (fish stocks, forests and biodiversity) and the depletion of re- source supplies (fossil energy and metal ores) is missing from the GDP calculation. As a result, GDP suggests we are richer than we really are. The previous list of arguments and the more extensive discussion in the Appendix make clear that it is difficult to defend GDP as a social welfare or general progress indicator.
- How serious is the influence of GDP information on the economy?
Faced with so many fundamental criticisms of GDP as a measure of social welfare, a recurrent response in debates with colleague economists has been not to worry too much about the shortcomings of GDP as the latter does not have much influ- ence on the real economy. Such a response is of course entirely consistent with the lack of urgency shown in repairing, or in any other way overcoming, the shortcomings of the GDP. Many signs, however, indicate that GDP information has consid- erable influence on reality. In the first place, one wonders why governments invest structurally in calculating and predicting GDP. This investment is shared by all countries, and since the GDP is standardized through the United Nations System of National Accounts, it allows for an international comparison of countries in GDP terms. All of this would represent a forthright waste of money if GDP information indeed had no impact on the economy at large. A more obvious explanation is that many economic agents regard GDP information as an important input to their deci- sion-making. This is certainly true for politicians and public servants – witness the common appearance of GDP information in governmental documents on performance and goals of the economy, or the broad concern arising when GDP growth fig- ures turn out to be substandard. The importance of GDP information for firms, investors and citizens/consumers is illustrated by the media – television, radio, newspapers, financial and other magazines, and internet – informing us on a daily basis about the status of our national GDP, both over time and in comparison with other countries. To illustrate the widespread influence of GDP, note that a search on internet on July 7th 2008 for ‘‘GDP” delivered 44,700,000 hits (and ‘‘Gross domestic product” 4,970,000 more). This is, for example, more than five times the number of hits for ‘‘social welfare”, viz. 8,470,000. One can identify more concrete indications of the influence of GDP information on economically-relevant decisions. Financial markets are sensitive to realizations and predictions of GDP growth, which if positive create optimism and if neg- ative pessimism about the ‘‘economic climate”. Central banks generally formulate their interest policy on the basis of expec- tations about growth and inflation. Private companies regard GDP growth as an important element of the general investment climate. Even the confidence of consumers, which determines their purchasing behaviour, is influenced by expectations of GDP growth. Politicians want to avoid low GDP growth rates because they fear negative voter responses. To some extent this is moti- vated by the belief that insufficient growth will lead to economic instability characterized by much unemployment. If GDP does not grow according to hopes and expectations, politicians at both national and supranational levels become very ner- vous, and will not relax until GDP retains its old growth pace. With the formation of the EU, GDP growth has become an even more explicit and important goal, witness the unconditional 3% growth objective of the Lisbon strategy. In addition, GDP growth allows for rising tax revenues, as a result of which public expenditures can increase – a nice prospect for pol- iticians in power. All in all, there is no denying that public and private agents assign much importance to the national GDP indicator. The influence of GDP information is easily underestimated, as it runs through multiple channels – government, politics, public officers, private businesses, financial markets, investors, consumers and international agencies such as the IMF and the OECD. Moreover, these channels reinforce each other. For example, when growth rates go down, the media and financial markets will focus all attention on it, and politicians will respond with appropriate measures, objectives and promises, which again will make headlines in the media, which in turn will influence expectations and therefore decisions on financial mar- kets. Over longer periods of time, the influence of GDP information is further established and reinforced as education at all levels, economic and financial research, and economic advisory councils of governments give much attention to GDP infor- mation. The newspaper reader is influenced by the media, and the student of economics by her education, with the idea that GDP growth is relevant. The consequence is a large influence of GDP information on consumption, investment and policy decisions, with evident repercussions for economic structure and social conditions. A recent indirect but important effect of GDP information relates to climate policy. Among the most influential climate policy studies are economic analyses in which the policy cost is expressed in terms of a reduced rate of GDP growth (an early study is Nordhaus, 1991; for an overview see Tol, 2008; and for a critical evaluation Söderholm, 2007). This type of research has received much attention from policy makers, notably since it was widely diffused through reports of the IPCC (Intergov-
ernmental Panel on Climate Change). In fact, in deciding to not ratify the Kyoto Protocol, the Bush administration referred to Nordhaus’ work as providing an important motivation. Various other policy links can be observed. Especially in the United States, many commentators in public media and on the Internet try to justify tax cuts or a flat tax by arguing that these will boost GDP growth. Apart from the fact that this is not generally true (as it depends on many conditions), irony wants that GDP information is used to justify policies which will negatively affect income distribution, whose neglect (or implicit treatment) is an important reason for the GDP indicator being an invalid measure of well-being (Section 2 and Appendix A.5 of the Appendix). This is not the end of the story. Through pessimistic (optimistic) responses by individuals, firms, and governments to fore- casts of a low (high) rate of GDP growth, GDP information creates a pro-cyclic effect. This resembles the way in which behav- iour in financial markets is steered by perceptions, leading to herd behaviour which causes expectations to become true. Likewise, a large number of individuals acting on the basis of publicly available GDP information can give rise to a consid- erable macroeconomic response. Beliefs and expectations become then reality, i.e. a ‘self-fulfilling prophecy’ is at stake. At best, the effect of GDP information may be reinforcing the cycle. The recent financial (credit) crisis clearly illustrates this. Daily information in various news media about reduced GDP growth only serves to reinforce feelings of consumers, investors and others that things are going from bad to worse. The current financial crisis is illustrative of this phenomenon. We are overwhelmed by articles showing how little GDP growth the economy is expected to attain in the coming period. This cre- ates a bleak atmosphere influencing expectations and through it behaviour of economic agents. Unfortunately, rigorous empirical studies of the influence of GDP information on the economy are lacking. As economists study virtually everything, the reason must be that the topic does not seem to gather much interest. Anyway, readers who are not convinced by the arguments in this section and feel that GDP information does not have much impact on the econ- omy at large should really be sympathetic to reducing the role played by GDP information in the public sphere, as it serves no purpose while its provision is costly.^5 Taking the insights from this section and combining these with the conclusions of Section 2 gives the following overall conclusion: GDP not only is an inadequate proxy of social welfare but also has a considerable impact on public and private economic decisions. By implication, GDP represents a serious information failure.
- Bounded rationality as an explanation of the widespread belief in GDP relevance
At this point, many readers may desire a meaningful theory of why so many people all over the world place such impor- tance on changes in GDP (per capita). Aside from some of the arguments given in the previous section, insights from behav- ioural economics and innovation studies may provide possible explanations, linked to concepts like bounded rationality, historical accidents, increasing returns and lock-in. Bounded rationality of humans can explain why the problem of evaluating the performance of a complex modern econ- omy is simplified into a focus on GDP information. A simplified, aggregate indicator like GDP contributes to ease of trans- mission by education and media institutions, and allows for quick even though inaccurate comparisons. Conceptual simplification of choices is not an unusual strategy followed by humans, and has been well documented in the literature on behavioural economics. For example, human choice is often in accordance with the ‘isolation effect’, which represents a simplifying approach to compare the performance of complex alternatives (Kahneman & Tversky, 1979). It involves deci- sions being purely based on differentiating characteristics and not on shared characteristics of all prospects. It reflects bounded rationality as it gives rise to inconsistent choices (and preferences) when the same choice alternatives are pre- sented in different ways. In addition, conformism, docility, socialization and imitation can explain why GDP is without much criticism accepted by the majority of citizens and economic students alike. This is aided by repeated emphasis of the relevance of GDP in education and public media, including newspapers, television, radio and internet. Simon (1990) suggests that humans are generally docile in the sense of tending to depend on the information, recommendations and persuasion arriving through social chan- nels. Docility allows one to quickly learn by absorbing ideas and suggestions from others in a rather uncritical way. Indeed, being critical on information obtained through social interactions, notably from parents and teachers, would hamper rapid learning and accumulation of knowledge. Docility affects both cognitive information and moral knowledge. The GDP indica- tor may in fact be seen to relate to both types of knowledge. Colander and Klamer (1987) concluded on the basis of data obtained with a questionnaire among students in six top- ranking graduate programs in economics that students are subject to a real socialization process. Revisiting this theme al- most 20 years later, Colander (2005, p. 175) notes: ‘‘Individuals are not born as economists; they are molded through formal and informal training. This training shapes the way they approach problems, process information and carry out research, which in turn influences the policies they favour and the role they play in society. [.. .] In many ways, the replicator dynamics of graduate school play a larger role in determining economists’ methodology and approach than all the myriad papers writ- ten about methodology.” As part of economists’ methodology, the widespread belief in GDP is something that is reinforced
(^5) Most economists who read this section agreed with its arguments and conclusion. Several, however, regarded a large impact as trivial and hardly in need of argumentation. A few commentators expressed the opinion that GDP information has a small, negligible impact on the economy, but they were unable to provide any evidence.
Related to this point is the belief that GDP growth is generally needed or sufficient for (close to) full employment. How- ever, the empirical evidence for this view is weak (Saget, 2000). Instead, the long run equilibrium employment rate is likely to depend on other factors than the rate of GDP growth, notably effects of search time (jobs and employees), structural mis- matches between education and work, the gap between gross and net income, and the gap between income and unemploy- ment benefits (e.g., Pissarides, 2000). Further, GDP growth does not necessarily reduce unemployment for several other reasons: it may involve outsourcing associated with retaining much of GDP domestically while moving jobs to elsewhere; and growth often goes along with creative destruction, i.e. disruption of old economic activities, which in turn implies (tem- porary) unemployment in specific sectors or job types. Of course, employment in combination with wages or productivity will affect the GDP, suggesting the relevance of a reverse causality (and correlation at times). GDP per capita is often interpreted to convey information about productivity. However, a correct productivity measure- ment needs to be related to the number of hours worked, which shows much variation between countries, as well as over time. GDP per hour is a more useful indicator of productivity than GDP per capita. In addition, it is evident that an increase of labour productivity cannot serve as an ultimate goal for a society. Productivity is indeed a very partial indicator from a wel- fare perspective. Another often expressed view is that GDP provides a basis for estimating tax revenues. This might then allow one to fore- cast taxes, to evaluate creditworthiness in the case of providing loans to countries (as done by the IMF and the World Bank), or to determine fair financial contributions of member states to a federation of states (e.g., USA, EU). In this case, GDP does not function as a performance indicator but more modestly as a model variable.^8 Of course, estimates based on disaggregated information (e.g. value added per sector) will be required in order to arrive at a sufficiently accurate estimate of tax revenues. In other words, for tax calculation purposes, there is no real need to aggregate national accounts into a GDP. Within environmental economics the environmental Kuznets curve (EKC) hypothesis received much attention during the 1990s. It represents the idea that initially economic growth goes along with an increase in environmental pressure up to a turning point, after which GDP per capita and environmental pressure become delinked. One explanation for such a delink- ing is that the environment is a luxury good: once income levels are sufficiently high, voting and buying will be redirected to ‘green’ policies and products. Another explanation is the older idea that with a higher income more funds will be available for environmentally relevant innovations and investments. Empirical evidence for the EKC hypothesis shows, however, that it holds only for a subset of indicators. Moreover, these indicators are all partial in nature and often unrelated to long term environmental sustainability. In particular, problem shifting may occur to other countries or environmental dimensions not captured by the environmental indicator. In addition, findings usually depend on cross-section analysis and can not be used straightforwardly to predict intertemporal patterns for single countries (de Bruyn & Heintz, 1999; Stern, Common, & Barbier, 1996). What comes out of these studies is that nearby problems relating to human health, like local water and air pollution, are solved if income rises, but that many other environmental problems are not solved or at best shifted in space (e.g., export of solid waste, incineration) or time (e.g., landfills). In general, it is therefore not true that economic growth solves environ- mental problems. This means that GDP information is not as relevant to understand the dynamics of solutions to environ- mental problems as was initially believed. It is customary to point to the importance of GDP growth for developing countries. One would expect welfare growth here to show a higher correlation with GDP growth than in rich countries (especially because of the arguments related to lexico- graphic preferences in Section 2). Even so, I have been unable to find a thorough study that supports this widely held belief. Kenny (2005, p. 10) summarizes the empirical evidence as follows: ‘‘There has been convergence across a wide range of indi- cators of the quality of life. Given that there has not been convergence in the standard income indicator, this may suggest that income is only one among a number of factors in determining quality of life outcomes. In turn, this suggests some hope that improvements can be sustained even in the absence of sustained income growth.” Finally, three common defences of a different nature need to be mentioned here as well. The first one emphasizes the role of GDP information as an input to (macro)economic models or analysis. Such a use of GDP information is not the particular aim of the criticism here, even though it is doubtful whether it is an important use, given that most empirical macroeco- nomic models are actually bottom-up models rather than developed around an aggregate notion of GDP.^9 The central point to notice is that the foregoing discussion was not about using GDP as an intermediate model variable but as a goal variable with an implicit or explicit social welfare interpretation. An example of a constructive exercise using GDP as an intermediate variable is Bird (2001), who shows a positive statistical relation between measures of risk and the share of social spending in GDP and concludes that the Welfare State encourages risk-taking which in turn may stimulate innovative behaviour and activities. A second common type of defence is that GDP growth is just one of multiple goals of macroeconomic policy, including also stable prices, low unemployment, acceptable income distribution, etc. However, from a methodological angle it should be clear that adding indicators does not undo or compensate for the shortcomings of the GDP indicator, certainly not if the specific GDP-related goal is unconditional GDP growth.
(^8) A corrected GDP – to accommodate the critiques documented in Section 2 – would lack a pure financial, cash flow interpretation (‘taxable income’), which would be a disadvantage in view of goals like predicting tax revenues (unless taxable income would be redefined in line with the GDP corrections). (^9) Some models for private companies or governments – to forecast car sales, credit losses (banks) or state revenues – may be top-down and use GDP as a
starting point (which does not mean they work better this way).
Third, the existing international standard for national accounts and GDP supports the uniformity of data on GDP. Many observers regard this as a strong feature of the GDP indicator, as it contributes to a clear economic comparison of countries. It should be realized, however, that international comparability is a necessary but an insufficient condition for the usefulness of any indicator.^10 In conclusion, the foregoing arguments in defence of GDP are unable to alter the view that GDP represents a serious infor- mation failure, as concluded in Section 3. The defences of the GDP indicator reflect mistaken beliefs, irrelevant consider- ations, or a focus on the role of GDP information as an intermediate variable. Moreover, all defences lack a clear interpretation in terms of overall social welfare and happiness.
- Evaluating alternatives to the GDP indicator
It seems wise to remove an indicator that is seriously misleading, irrespective of whether an acceptable alternative is available. Hence, the removal of GDP information would be an enormous improvement because a structural information fail- ure would be eliminated. In other words, ignoring or abolishing GDP should be unconditional upon the availability of sub- stitute indicators. Nevertheless, since many commentators seem to be of the opinion that one should not ignore or abolish GDP until a good alternative is available, the main alternatives are briefly reviewed hereafter. There are four types of alternative indicators available in the literature. A first type is based on rather pragmatic account- ing adjustments to GDP. The most influential example is the Index of Sustainable Economic Welfare (ISEW: Daly & Cobb, 1989). Derived indicators are the Genuine Progress Indicator (GPI), and the Sustainable Net Benefit Index (SNBI) (Lawn & Sanders, 1999). All these represent a correction of the regular GDP by repairing important deficiencies through adding or subtracting certain partially-calculated money amounts to/from GDP.^11 The various applications of the ISEW show that, whereas GDP follows a rising trend, the ISEW shows a constant or even decreasing pattern after a certain time. The temporal breakpoint varies with the country, but lies somewhere in-between the late 1960s and the 1980s. The various ISEW (and GPI) studies suggest that the costs of economic growth now outweigh the benefits, leading to ‘‘growth that is uneconomic” (paraphrasing Herman Daly). A second type of indicator also starts from GDP but focuses entirely on environmental externalities and natural resource depletion. Corrections here give rise to ‘sustainable’ or ‘green(ed)’ GDP type of indicators. ‘Sustainable income’ denotes a level of income that is based on a reproducible economic and environmental base. Recalculation of a GDP with externalities ‘inter- nalized’ is not simple, as it implies a completely different set of prices in the economy. It is not surprising, then, that there have been few empirical exercises aimed at calculating a green or sustainable income. The best known is the Sustainable National Income (SNI), which has been calculated for the Netherlands (Gerlagh, Dellink, Hofkes, & Verbruggen, 2002). It is based on the conceptual work by Hueting (1974). This can be seen to reflect the notion of ‘sustainable income’ as expressed by Hicks (1948), with the assumption that individuals are better off if vital environmental functions remain available ad infinitum (Hueting, 1974). The SNI approach uses a general equilibrium model that calculates the impact on national income of imposing sustain- ability constraints for the nine most important environmental themes (for the Netherlands). The policy interpretation of this approach is then that an economy is submitted to a strong sustainability policy with a tremendous impact on national income. Hofkes, Gerlagh, and Linderhof (2004) analyse SNI over the period 1990–2000, for 1990–1995 and 1995–2000. They find that although it increased substantially in this period and the relative gap between SNI and GDP decreased, the absolute gap be- tween them increased. This suggests that production in the Netherlands became less sustainable over the studied period. Over the whole period 1990–2000, the enhanced greenhouse effect appears to be the binding environmental constraint. A third type of indicator is genuine savings (or genuine investment). This focuses on maintaining or increasing wealth, defined as the sum of economic, human and natural capital (Dasgupta & Mäler, 2000; Hamilton & Clemens, 1999). Recently, genuine savings (GS) has been adopted as a central indicator by the World Bank, under the name of ‘adjusted net savings’. GS can be defined as traditional net savings subject to a number of corrections (Bolt, Matete, & Clemens, 2002): (i) the value of depletion of natural resources is deducted; (ii) the costs associated with pollution damage, including economic and health effects, are deducted; (iii) expenditures on education are treated not as consumption but as savings/investments in human capital; (iv) net foreign borrowing is deducted, while net official transfers are added; (v) capital depreciation (capital con- sumption) is deducted. Categories (i) and (ii) are the most difficult to estimate. Nevertheless, the World Bank has produced estimates for most countries in the world. Genuine savings turn out to be negative for the Middle East and North Africa, and Sub-Saharan Africa regions, positive for OECD countries, and the highest for the East Asia/Pacific region (World Bank, 2006). One main disadvantage of the GS indicator is characteristic of weak sustainability, namely that losses of natural capital are not regarded as worrisome as long as they are compensated by economic and human capital (known as ‘weak sustainabil- ity’). Another main shortcoming of the GS approach is that it only overcomes few imperfections of the GDP indicator, namely neglect of capital depreciation and (partially) valuation of informal activities. A fourth and final type of indicator of social welfare is a composite index that combines indicators that are considered to capture relevant aspects of human well-being. Unlike the previous types of indicators, this does not generate a monetary
(^10) Unfortunately, the international standard also acts as a barrier to the implementation of changes in the GDP calculation which undo some of the
shortcomings discussed in Section 2. (^11) This approach goes back to Nordhaus and Tobin (1972).
due to the threat of stagnating GDP growth are no longer possible. One will less dogmatically deal with stimulating devel- oping countries to enter a transition to a formal economy (for that matter, the World Bank and UNDP has already moved a long way in this direction but do not seem to dare taking the step to discard GDP entirely). See further Frank (1985, 2004), Ireland (2001) and Layard (2005) for a number of examples of alterations in economic policy that are in line with replacing the GDP indicator by information obtained from the subjective-empirical welfare literature: e.g. regulatory taxation of work- ing overtime and status goods, limiting commercial advertising, and restricting flexible labour contracts. Although from a GDP growth perspective these look like unnecessarily limiting measures, they are more positively evaluated from a real welfare or happiness perspective. Ng (2003) shows analytically that systematically striving for GDP growth leads to an over-spending on private consumption and an under-provision of public goods, which may cause economic growth to be welfare-reducing. Without GDP, there will also be less resistance against policies which improve social welfare at the cost of GDP growth. Whereas from a GDP perspective such policies might look as a sacrifice, from a social welfare angle they would seem very logical. A very topical example is the case of climate policy. Most economic studies of climate policy have analysed the devel- opment of GDP under alternative climate (policy) scenarios, and therefore have focused attention on the trade-off between GDP growth and risks related to climate change (Kelly & Kolstad, 1999). But this trade-off represents a misplaced concep- tualization of the problem as GDP is not a good measure of social welfare. Moreover, this trade-off concerns a period of 50–100 years in the future, during which the GDP of the rich countries will have grown far beyond any welfare-maximizing level.^16 From a real welfare perspective, one would arrive at conclusions about climate policy which likely support a more stringent reduction of greenhouse gases. This is illustrative of the difference that (absence of) GDP information can make in the context of concrete policy.
7.3. Different implications for developing, middle income, and rich countries
Does GDP per capita serve a more useful function in evaluating processes and policies in developing countries – defined by the World Bank as having an average GDP per capita of less than US$6000 – than in developed countries? It is often believed that especially poor countries require GDP growth in order to improve the well-being of their cit- izens. In Section 2 (point 6; see also Appendix A.6 in the Appendix) it was suggested, however, that in many cases income growth here just represents a shift of activities from the informal to the formal sector. Particularly in very poor countries this may be accompanied by a loss of local community and subsistence agriculture, as well as migration of farmers to urban slums, with predictable negative consequences for food availability, health and quality of life. More- over, income inequality may increase during the process (the initial part of the well-known Kuznets curve). Such very poor countries dominate among those that the World Bank classifies as ‘low income countries’ and the UN as ‘low human development countries’ (according to the HDI). Middle income countries may often show a positive correlation between income and welfare growth. However, negative impacts on welfare and health may result here from severe environmental pollution and resource degradation. This is illustrated by the current development of China. Finally, moving on to high income countries, the results reported in Section 2 (point 4; see also Appendix A.4 in the Appendix) support the idea that GDP growth there does not contribute much to welfare growth (the ‘threshold hypothesis’). In- deed, welfare seems to have stagnated for the richest countries in the world, despite continuing growth of GDP per capita. This is not surprising given that in these countries all basic needs are more than satisfied, so that the consum- ers in these countries are mainly involved in adaptation to economic changes, and a zero-sum rivalry game of income and status consumption.
7.4. Abolishing GDP should not be confused with ‘anti-growth’, ‘anti-innovation’ or ‘anti-accounting’
Many people responding to an early draft of this paper concluded that the author must be against GDP growth. But they seem to confuse ‘no GDP indicator’ with ‘no GDP growth’. It is important to realize that without a GDP indicator GDP growth is no longer measured. Ignoring GDP information reflects the irrelevance of GDP growth and not an anti-growth position. Given that GDP is not a good welfare indicator, one should not be in favour of unconditional (always) GDP growth (or ‘‘growth-fetishism”). To strive under all circumstances for GDP growth puts an unnecessary constraint on the space within which we search for welfare growth.^17 In fact, it means that GDP^ growth cannot be traded-off against something else. History shows that, with the availability of a GDP indicator, structurally striving for (GDP) growth is not just a risk but inevitable. Evidently, the temptation is too large. Ignoring GDP information in policy must not be confused with a plea to restrict individual income growth. Individual in- come growth can mean satisfaction of more (basic) needs or more happiness due to an improved relative income position (Section 2, points 3 and 5; see also Appendices A.3 and A.5 in the Appendix). From an individual perspective it can therefore
(^16) Azar and Schneider (2002) show that the costs of reaching what the IPCC considers ‘‘safe” concentrations of CO 2 in the atmosphere, for the world as a
whole, fall in the range US$1 tot US$20 trillion. Although these are impressive figures, they imply less than 3 years delay of reaching a certain income level 100 years from now (given 2% GDP growth). Interpreted this way, the costs of a stringent climate policy are marginal in economic terms in the long run. (^17) Using the same logic, there is no reason to be generally against GDP growth. In fact, growth critics or anti-growth proponents should seriously consider
shifting to oppose GDP information instead of GDP growth. The logical position on growth that follows is a neutral one.
make sense to strive towards income growth. Moreover, even if individuals do not get happier from individual income growth (e.g. due to adaptation), then still status-seeking will make them strive for more. However, as argued in Section 2, from a social perspective continued income growth ultimately results in what is at best a zero sum game and possibly, given e.g. environmental externalities, a negative sum game. For this reason, a society and its government should not foster eco- nomic growth as a goal in itself. In addition, ignoring GDP does not imply a plea against innovation. It will, though, result in attaching more value to inno- vations that improve social welfare than to innovations that will promote GDP growth but have less clear or even negative social welfare benefits. Innovations with a positive welfare effect and negative growth effect will become more attractive. Interestingly, many of these can be found in the area of sustainable technologies, such as renewable energy and energy-effi- cient production and housing. Removal of the GDP information failure is likely to change, through its effect on financial mar- kets, investors and public policies, the incentives for innovations in the direction of welfare improvement as well as sustainable production and consumption. Furthermore, it should be stressed that criticism of GDP as a welfare indicator and its role in public debates and policy preparation should not be seen as a critique of the system of national accounts. Whereas GDP is the most aggregated descrip- tion of the economy, the national accounting system provides a detailed, disaggregated picture of the flows of goods and services along with complementary monetary transfers. As a result, the national accounts can support economic modeling and forecasting, analysis of productivity growth of sectors, and financial planning by the government. Interestingly, national accounts are being transformed and extended to eliminate various shortcomings related to informal markets, natural re- sources and environmental damage. But despite such improvements, the method of GDP calculation has remained largely the same. Finally, is it realistic to think that the information failure associated with GDP can be removed? Recalling the conceptu- alization of the GDP paradox as a problem of lock-in (Sections 3 and 4), one will understand the difficulties involved. Econ- omists and their training play a central role in maintaining the idea that GDP information matters. To begin with, therefore, economic education should pay due attention to the shortcomings of GDP information, and the entire curriculum (notably macroeconomics) should be screened for any uncritical use of GDP information. This will inevitably imply a thorough revi- sion of many textbooks. Next, economists should consistently and explicitly advice politicians and journalists to reduce attention and concern for GDP in the public arena. Research and related publications by economists should be filtered for use of GDP as an ultimate goal.
- Conclusions
This article has examined the paradox that despite all theoretically and empirically motivated criticism of GDP as a social welfare and progress indicator, its role in economics, public policy, politics and society remains influen- tial. To support the starting point of the analysis, Section 2 summarized the large literature which has identified the imperfections of GDP. This delivers the unambiguous conclusion that GDP suffers from many serious shortcomings. To understand the ‘GDP paradox’, one has to recognize that many academic economists accept the criticism of the GDP indicator but deny its relevance. This denial was suggested to come in two forms. The first is a belief that the size of the im- pact of GDP information on economic reality is modest. Section 3 examined this view and arrives at an opposite conclusion, namely that GDP information has a large impact on both private and public decisions, and therefore on the economy at large and the direction of long term economic development. Section 4 then explained on the basis of various aspects of bounded rationality – including the tendency to conceptually simplify complex problems, conformism and docility – in combination with uncritical or even biased information provision by education and public media, why so many people belief in the rel- evance of GDP information. A second type of denial of the relevance of GDP criticism is that despite its shortcomings, GDP still provides useful infor- mation. Section 5 considered this view, and concludes that the various arguments commonly mentioned in favour of GDP do not have a clear structural association with, or interpretation in terms of, social welfare. Moreover, the arguments reflect mistaken beliefs, irrelevant considerations, or a focus on the role of GDP information as an intermediate variable. Combining the insights of Sections 2–5 gives the conclusion that GDP represents a serious information failure: it suffers from many shortcomings and has a large influence on socioeconomic reality. A widespread view among economists is that we should not get rid of GDP until a good alternative aggregate indicator is available. Two counterarguments were given here. First, the many efforts to develop an improved indicator of social wel- fare and progress show that it is unlikely that a perfect indicator will arise in the near future. All current alternatives suffer from various shortcomings, even though most of them represent a clear improvement over GDP. Second, regardless of the availability of perfect alternatives, it seems wise anyway to remove the GDP indicator as it is seriously misleading infor- mation. One should especially take into consideration that this information failure is structural and affects long term eco- nomic structure through its influence on investment and innovation decisions by private and public agents. Removal of such an information failure really deserves an unreservedly enthusiastic response from economists. It is possibly the larg- est information failure we are facing, given that GDP is a central indicator in all countries and supranational governance structures.
increasing costs of economic change (whether progress or decline). This explains why GDP and welfare growth do not nec- essarily coincide. At a certain moment, GDP growth creates more costs than benefits, so that an optimal scale of economic activity will be surpassed (Daly, 1992). Economists are happy to argue in favour of cost-benefit analysis as a general method for policy evaluation and support. When it comes to the direction of the economy as a whole, many of them suddenly are satisfied with only information about costs, that is, GDP information. Finally, the correction of GDP for inflation is required to make estimates comparable over time. This leads to the particular problem that the correction is based on an average consumption basket, which is regarded as representative for the entire population. However, the more skewed is the income distribution or the more heterogeneous in terms of consumption (pur- chase) behaviour is the population, the more inaccurate and thus less representative this procedure will be.
A.2. Intertemporal considerations
Macroeconomics, and within it especially economic growth theory, is concerned with the dynamic aspects of the econ- omy as a whole. Macroeconomics does not offer any support for GDP as a measure of social welfare. Quite the contrary, opti- mal (normative) growth theory proposes models that explicitly use some theoretical notion of social welfare resulting in an objective function (based on continuous or overlapping generations) that does not generally come down to a GDP type of criterion. Weitzman (1976) has shown that under rather severe restrictions the national product can serve as a proxy (sta- tionary equivalent) of a utilitarian intertemporal welfare function formulated as a net present value of future consumption flows. Vellinga and Withagen (1996) have generalized this result. Three objections can be raised against the relevance of these theoretical findings. First, the instantaneous utility function is approximated by consumption (Weitzman) or by a Tay- lor series in the arguments of utility – consumption and capital (Vellinga and Withagen). Second, one needs to assume that there is no pure time dependence in the form of, for instance, exogenous technical change or exogenously changing world prices; time can only enter in the form of a time preference rate. Both these choices imply a deviation from actual social wel- fare. Third, a specific intertemporal function is posed, without any empirical support, as a suitable representation of social welfare. However, one might instead adopt an egalitarian or Rawlsian welfare approach, or any other (hypothetical) social welfare function. Next, the fact or belief that GDP growth in certain periods or regions has correlated positively with progress (however measured) should not be confused with the idea that GDP (growth) is a good measure of social welfare (progress) in general. In other words, the correlation may be low or even negative in certain periods and regions. If, by way of thought experiment, one extrapolates a constant tempo of real GDP growth towards the distant future, one will end up with an incredibly high GDP. But it is very unrealistic to suppose that social welfare will reach a comparably high level. To illustrate this, using 2% as a conservative estimate of the average yearly GDP growth rate over the past decades, extrapolation of this rate 1000 years into the future gives a GDP that is (1.02)^1000 400 million times as high as the current GDP. Surely, no one can believe – if only on the basis of introspection – that individual and social welfare can increase to such an extent. This shows that, in the long run, GDP can not serve as a good indicator or even rough approximation of social welfare. Definitely, at some point a de- linking of GDP and welfare must occur. In fact, it is very well possible that such de-linking has already occurred for many rich countries in the world. This is in fact suggested by a number of empirical studies (Section 2.4).
A.3. Lexicographic preferences
People have various basic needs, such as air, water, food, sex, shelter, company, respect and freedom. These cannot be traded off against luxury services and material goods – in fact, the latter often serve as a sublimation of the basic needs them- selves (e.g., status goods to gain respect from peers). In other words, substitution between broad categories of consumption is limited, even though within categories is may be present (e.g., one type of food for another). This is the core of the notion of lexicographic preferences (Day, 1971; Encarnación, 1964; Georgescu-Roegen, 1954; Tversky, 1972), which is closely con- nected with the older idea of the Maslow pyramid in psychology. Lexicographic preferences can be defined as having two characteristics: (i) individuals have limited needs in certain goods or services, as feelings of satisfaction occur after consump- tion reaches a certain level; (ii) ‘lower’ needs (e.g. the removal of thirst and hunger) need to be fulfilled before ‘higher’ needs (e.g. recreation) can appear. Within this framework, income growth and the associated growth of material consumption, notably in urban and polluted environments, is an imperfect compensation for a lack of satisfaction of basic needs, such as relaxation, space, serenity, clean air and water, and direct access to nature. As a result, one cannot be sure that (individual and social) welfare will increase, despite GDP and individual income growth. The previous point does not mean to suggest that GDP growth always implies more material consumption. It is quite pos- sible that it comprises an increase in services. What this in turn yields in terms of welfare is difficult to say in general. Sen’s (1999) concept of individual ‘capabilities’ may be useful here. This tries to bring goods and services into a single denominator by emphasizing freedom and opportunities to choose, as well as context-dependent functionality of goods and services, based on taking into account the peculiarities and environments of individuals. Examples of the latter are being disabled ver- sus being perfectly healthy, and living in a dense, busy city versus living in the countryside: different goods and services may be needed to realize the same level of welfare in these alternative circumstances. Income indicators do not correlate well with capabilities and opportunities in these various welfare-relevant dimensions.
A.4. Empirical analysis of individual happiness and social welfare
A growing field of subjective well-being analysis on the basis of empirical data has produced many insights about the determinants of welfare or happiness. 19 Studies of this type are being undertaken by economists, psychologists and sociol- ogists. This research has first of all delivered the insight that, somewhere in-between 1950 and 1970, the increase in welfare stagnated or even reversed into a negative trend in most western (OECD) countries, despite a steady pace of GDP growth. Blanchflower and Oswald (2004) offer such an analysis for the UK and the USA.^20 This insight is confirmed by the ‘Euroba- rometer surveys’, the half-yearly opinion polls of the inhabitants of the EU Member States, as well as by corrections of GDP that seem to point more in the direction of social welfare (e.g. the ISEW indicator of Daly & Cobb, 1989; see Section 6 on this). The income level at which de-linking occurs between GDP and (subjective) social welfare has been estimated to approximate $15,000 (Helliwell, 2003). This has been referred to as a ‘threshold hypothesis’, reflecting that the costs of growth exceed the benefits (Max-Neef, 1995, p. 117): ‘‘... for every society there seems to be a period in which economic growth (as convention- ally measured) brings about an improvement in the quality of life, but only up to a point – the threshold point – beyond which, if there is more economic growth, quality of life may begin to deteriorate.” Subjective well-being studies also show that, at the individual level, income does not perfectly correlate with welfare – indeed much less than is often taken for granted – so that absolute individual income is not a good, general proxy of individual welfare (Easterlin, 2001; Ferrer-i-Carbonell, 2005; Frey & Stutzer, 2002; van Praag & Ferrer-i-Carbonell, 2004). Relative income is relevant (see Section 2.5). In addition, other – income-independent – factors influence individual welfare or happiness. Important ones are: being employed, having a stable family (and having a partner), being healthy, personal freedom (political system), having friends, and belonging to a tight social community. This type of research further shows the relevance of unobservable or not easily observable factors, notably a pessimistic or optimistic attitude towards life in general. A recent empirical study by Ferrer-i-Carbonell and Frijters (2004) concludes, on the basis of an analysis of the effect of this attitude, that the belief that ‘‘being rich makes people happy” can better be replaced by ‘‘happy people are more likely to be rich”. For it appears that optimistic individuals are on average relatively happy and successful in life (ceteris paribus), and on the basis of the latter enjoy a relatively high average income (see also Lyubomirsky, King, & Diener, 2005). In addition, this type of research indicates that the well-being of men on average responds differently to income changes than that of women. Responses also differ among income brackets. Now if income does not render a reliable and robust measure of happiness at the micro-level, then it is very unlikely that the aggregation of individual incomes in a GDP provides a robust indicator of social welfare at the national level. Another important insight of this literature is that individuals adapt or get used to changed circumstances, such that their subjectively felt well-being does not increase (Frederick & Lowenstein, 1999). This relates to the fact that our senses can only handle a limited amount of stimuli, so that beyond a certain threshold a feeling of satisfaction or boredom arises. A change in circumstances can of course create a one-off or ephemeral welfare effect, but this typically will quickly fade away. Since peo- ple do not realize the phenomenon of adaptation they keep striving for ‘more’. Terms like ‘addiction’, ‘hedonic adaptation’ (Helson 1964), ‘hedonic treadmill’ (Brickman & Campbell, 1971) and ‘preference drift’ (van Praag, 1971) are used in this re- spect. There is further evidence that individuals adapt more when confronted with changes in pecuniary versus non-pecu- niary categories (Graham, forthcoming). Utilizing subjective well-being indicators for a large number of countries – on the basis of World Values Surveys data (Inglehart, Basanez, Deiz-Medrano, Halman, & Luijkx, 2004) – Layard (2005) concludes that, whereas countries with high in- comes show little variation in average reported happiness, this is quite different for countries with low incomes. The first group is dominated by countries of Protestant origin, which may point to a religious factor at stake. In addition, one cannot exclude a serious influence of climate conditions and political systems (notably communism versus enlightened capitalism) on happiness. This suggests that, at best, the relationship between GDP and happiness or welfare is conditional on various other aggregate features. In addition, the country comparison by Layard clarifies that happiness is characterized by dimin- ishing returns of increases in GDP per capita. Happiness evidently depends on leisure. But leisure is not captured by GDP. Quite the contrary, it has an opportunity cost of not being productive in terms of contributing to GDP. A study by the OECD (2006) makes adjustments of GDP by valuing leisure at average income (GDP) per worked hour, and finds that the result (in per capita terms) leads to a different ranking than according to GDP per capita. In this ranking, The Netherlands scores best of all OECD countries. Two factors contribute to this: the inactive part of the working force is relatively large, and part-time working is very common (de Groot, Nahuis, & Tang, 2006).^21
(^19) No sharp distinction is made here between notions like utility, welfare, well-being and happiness. One reason is that disciplines such as economics,
sociology, and psychology have developed their particular jargon, even though this is not always well defined or consistently used. More importantly, empirical research does not make a sharp distinction between these various notions. In fact, differences between them are so subtle that they defy measurement. This does not deny the relevance of different interpretations or types of happiness. (^20) The original influential study is Easterlin (1974). (^21) The introductory chapter in Bruni and Porta (2005) offers a discussion of different interpretations of happiness (short-term feelings of euphoria or long- term feelings of satisfaction). It notes a fundamental distinction between subjective hedonism (‘‘seeking pleasure and avoiding pain”) and objective eudaimonia (‘‘striving for perfection that represents the realization of one’s true potential”).
one guarantees that creating new preferences contributes to people’s happiness. Reference drift can then ultimately result from a combination of advertising and comparing with, as well as imitating, others. The phenomenon of relative income and context-dependent welfare does not just explain why humans strive for income rises. It also clarifies that an increase in relative income can improve the welfare of the respective individual, whereas social welfare is not being served by it. The reason is that status is a very scarce good, causing rises in relative income and welfare to resemble a zero sum game: one individual loses what another one gains (Frank, 1985, 2004; Layard, 2005).^27 In other words, you cannot make everyone increase in relative income. The relatively rich are generally happier than the relatively poor; this has always been so, and GDP growth will not change it. The rise of the relative income of an individual can be regarded as a negative external effect (external cost) on the welfare of the one whose relative income drops as a result of it. As is well known, externalities are harmful to social welfare, and need to be corrected.^28 The important conclusion of the foregoing for our purpose here is that, by being unable to consider relative income and rivalry in consumption, GDP will overestimate welfare/progress.
A.6. Formal versus informal economy
In general, GDP just covers activities and transactions that have a market price and thus completely neglects informal transactions between people that occur outside formal markets. The formal market dimension of human activities can com- prise a large or small part of total human activity, depending on whether one observes OECD countries (a large part), econ- omies in transition (medium) or less developed countries (small). The fact that the informal economy is left out of consideration explains why GDP per capita for many countries in the latter group can be so extremely low. At the same time, it can easily give a wrong picture of how (un)happy people really are. 29 This is amplified by the problem that the size of the informal economy relative to that of the formal economy may change considerably over time, both in developing and devel- oped countries (Bos, 2006). Actual GDP growth often comes down to a transfer of existing informal activities (unpaid labour) to the formal market. This means that the benefits were already enjoyed but the market costs were not yet part of GDP. This clearly illustrates the earlier point by Mishan and Daly (Section 2.1) that GDP reflects the costs of reaching a certain, unknown welfare level and not that welfare itself (i.e. the benefits). With transfer of existing activities from the informal to the formal circuit, economic growth means that the costs increase more rapidly than the benefits, and in the worst case only the costs rise. This holds, for instance, when informal activities like subsistence agriculture in developing countries, voluntary work, household work, and child care disappear. Such activities originally took place within the informal family circle and the local community. Transfer from the informal sphere to the formal market also occurs when people are born, die or are nursed in a hospital instead of at home. The GDP therefore does not recognize the value of all kinds of informal activities and services. As a result, public policy is often aimed at cutting back and discouraging informal activities. This can be interpreted as a strongly normative goal that is not entirely without risks for social welfare. The transition from an informal to a formal economy in itself offers no guarantee for a rise in happiness or welfare. For example, local social contacts – that form the basis for stable and happy lives – are much stronger and occur more frequently within informal than formal economies. In other words, ‘society’ in the strict meaning of the term has a value that is not captured by GDP. Obviously, it is not my intention to defend the extreme position that a transition from an informal to a formal economy automatically works out badly for social welfare. It is evident that labour division and specialization can be carried through more extensively in a formal than in an informal economy. It is possible, though not certain, that as a result of this produc- tivity increases, labour conditions improve and the choice spectrum (diversity of products) for consumers is widened.^30 These advantages do not neutralize the earlier mentioned negative welfare consequences of a transition to a formal market economy. Indeed, many other negative aspects should be taken into account. For example, if the labour market grows in scale, it stimulates commuting (distances), as well as people changing their house for a job. This erodes local community structures, with negative effects on individual happiness. In developing countries, the trajectory towards the formal econ- omy often goes together with a large-scale migration of ‘subsistence’ farmers with big families to the slums of large cities. The main message here is that GDP cannot serve as a measure to judge the welfare impact of fundamental changes that involve transitions from an informal to a formal economy. The expansion of markets to include informal activities is not al-
(^27) Empirical studies even suggest asymmetry in the sense that the ‘poor’ lose more happiness than the ‘rich’ gain (Ferrer-i-Carbonell, 2005). This would imply a negative-sum game. This asymmetry was already suggested a long time ago (Duesenberry, 1949). (^28) In addition, the attention for ‘market consumption’, including search, gathering information about (new) products, spending much time in shops and malls, and buying things that are hardly ever used, can divert adults as well as children from activities that contribute much more to happiness, such as more leisure (less working, less spending), playing with your kids, taking time for the extended family (grandparents), friends and neighbours, etc. (Schorr, 1998). (^29) For a large number of African and a small number of Asian countries, GDP has been regularly subject to corrections and adaptations based on estimates of the value added of the informal agricultural sector. Such corrections are less common for Latin-American countries (Charmes, 2000). (^30) According to recent field and experimental studies by psychologists a larger choice set may have a direct negative effect on individual motivation and
welfare because of ‘‘choice overload”, caused by incomplete information, search costs, and too much dependence on information offered by experts (Iyengar & Lepper, 2000). Herbert Simon’s notion of bounded rationality, i.e. limited brain capacity to process information and ‘satisficing’ behaviour , offers an additional, complementary explanation of this phenomenon (Simon’s famous statement ‘‘.. .a wealth of information creates a poverty of attention.. .” is also relevant here).
ways good for social welfare, even if GDP is raised. One might suspect that a certain combination of informal and formal (market) relationships between people would render the best of both worlds. With the GDP indicator, however, one cannot judge this in any way, since GDP omits the informal dimension of the economy. In the light of the continuous, idealistic pub- lic debate on the (un)desirability of expansion of the market domain it is therefore of utmost importance to not rely on GDP but to use adequate, real welfare indicators.
A.7. Environmental externalities and depletion of natural resources
The previous point is a specific case – through conceptually a very important one – of the more general criticism that GDP omits the value of non-market goods and services. Another example of unpriced effects relates to the natural environment and resources. This involves negative external effects as well as goods and services delivered by nature. The presence of externalities means that the current set of market prices insufficiently reflects the total (private + external) costs, which makes these prices unreliable signals in whatever calculation aimed at producing a social welfare indicator. Moreover, if air, water, or a natural area is being polluted any damage does not enter GDP, but when pollution is being cleaned this in- creases GDP. In addition, the (capital) depreciation associated with environmental changes (fish stocks, forests, biodiversity) and depletion of resource supplies (fossil energy, metal ores) is missing from the GDP calculation.^31 As a result, we are con- sidering ourselves ‘richer’ than we really are (Atkinson et al., 1997). It makes sense to define real, meaningful income as sus- tainable income: namely, as the maximum amount of income (or consumption) in one period without depleting capital or without harming the capacity to generate the same or higher level of income in future periods (Hicks, 1948). Maintaining this capacity requires sustainable capital utilization, regardless of the type of capital: human capital, machinery, and natural capital (natural resources and ecosystem services). Here we focus on sustaining the latter. A fundamental consequence of neglecting sustainability of natural capital has already been mentioned: namely, that the use of GDP as an indicator of welfare and progress means regarding substitution of basic conditions – like space, serenity, and direct access to nature and water – by market goods – like large houses, roads, cars, sewage systems and water purification, and expensive holidays in exotic locations – as progress. This in turn will unnecessarily stimulate the replacement of ‘nature’ by the ‘market economy’. A correct economic welfare approach would only characterize changes as real progress (welfare improvement) if they are accompanied by a sustainable use of environment and nature. Hueting (1974) already recognized this early on, and his elaboration of a measure of a green or sustainable income is based exactly on this insight (Gerlagh et al., 2002). The previous list of arguments is non-exhaustive. 32
References
Alpizar, F., Carlsson, F., & Johansson-Stenman, O. (2005). How much do we care about absolute versus relative income and consumption. Journal of Economic Behaviour and Organization, 56, 405–421. Arrow, K. J., Bolin, B., Costanza, R., Dasgupta, P., Folke, C., Holling, C. S., et al (1995). Economic growth, carrying capacity, and the environment. Science, 268 (April 28), 520–521. Arrow, K., Dasgupta, P., Goulder, L., Daly, G., Ehrlich, P., Heal, G., et al (2004). Are we consuming too much? Journal of Economic Perspectives, 18, 147–172. Arthur, W. B. (1989). Competing technologies, increasing returns, and lock-in by historical events. Economic Journal, 99, 116–131. Atkinson, G., Dubourg, R., Hamilton, K., Munasinghe, M., Pearce, D., & Young, C. (1997). Measuring sustainable development: Macroeconomics and the environment. Cheltenham: Edward Elgar. Azar, C., & Schneider, S. H. (2002). Are the economic costs of stabilising the atmosphere prohibitive? Ecological Economics, 42, 73–80. Beckerman, W. (1976). Defence of economic growth. London: Johathan Cape. Beckerman, W. (1999). A pro-growth perspective. In J. C. J. M. van den Bergh (Ed.), Handbook of environmental and resource economics (pp. 622–634). Cheltenham: Edward Elgar. Bird, E. J. (2001). Does the welfare state induce risk-taking? Journal of Public Economics, 80(3), 357–383. Blanchflower, D. G., & Oswald, A. J. (2004). Well-being over time in Britain and the USA. Journal of Public Economics, 88(7–8), 1359–1386. Bleys, B. (2006). The index of sustainable economic welfare: Case study for Belgium – First attempt and preliminary results. Belgium: Vrije Universiteit Brussel. Bolt, K., Matete, M., & Clemens, M. (2002). Manual for calculating adjusted net savings. Environment Department, World Bank. Bos, F. (2006). The development of the Dutch national accounts as a tool for analysis and policy. Statistica Neerlandica, 60(2), 225–258. Brickman, P., & Campbell, D. T. (1971). Hedonic relativism and planning the good society. In M. H. Apley (Ed.), Adaptation-level theory: A symposium (pp. 287–302). New York: Academic Press. Buss, D. M. (1989). Sex differences in human mate preferences: Evolutionary hypotheses tested in 37 cultures. Behavioral and Brain Sciences, 12, 1–49. Bruni, L., & Porta, P. L. (Eds.). (2005). Economics and happiness: Framing the analysis. Oxford: Oxford University Press. Charmes, J. (2000). The contribution of the informal sector to GDP in developing countries: Assessment, estimates, methods, orientations for the Future. In Fourth meeting of the Delhi Group on informal sector statistics, Geneva 28–30 August 2000. Saint Quentin and Yvelines: C3ED, University of Versailles.
(^31) Non-renewable resource depletion presents a special problem. It is widely agreed upon that the method proposed by El Serafy (1989) should be followed here. This transforms the finite income stream from a non-renewable resource stock into a (lower) infinite stream of income from other types of capital (manufactured and human capital). The method computes the difference between these two income streams and deducts the resulting cost. The latter is referred to as the ‘user cost’ of non-renewable resource extraction. This approach thus assumes that (and works only if) one can sustain an infinite stream of income by substituting the non-renewable resource by other types of capital. Hartwick (1977) provides the theoretical basis for this line of thought. (^32) For example, Dasgupta (2001) suggests that expenditures on defence, a significant part of GDP, in many countries contributes to GDP overestimating social welfare. He notes: ‘‘But in poor countries the machinery for warfare is all too frequently used by governments against their own people ... This means we can ignore defence expenditure ...” (p. 53). Laband and Sophocleus (1988) and Dougan (1991) further draw attention to the enormous rent-seeking costs of modern economies, which boost GDP (e.g. through the activities of lawyers). Clark, Kahn, and Ofek (1988) use hedonic price techniques (based on income compensation) to assess the impact of urbanization on social welfare. Their findings suggest a GNP-as-welfare deflator in the order of 6–7%, which steadily increases at a rate of half a percent per decade.
Layard, R. (2005). Happiness: Lessons from a new science. London: Penguin. Lomborg, B. (2001). The skeptical environmentalist: Measuring the real state of the world (12th ed.). Cambridge: Cambridge University Press. Lyubomirsky, S., King, L., & Diener, E. (2005). The benefits of frequent positive affect: Does happiness lead to success? Psychological Bulletin, 131, 803–855. Max-Neef, M. (1995). Economic growth and quality of life: A threshold hypothesis. Ecological Economics, 15, 115–118. Mishan, E. J. (1967). The cost of’ economic growth. London: Staples Press. Ng, Y.-K. (2003). From preference to happiness: Towards a more complete welfare economics. Social Choice and Welfare, 20, 307–350. Ng, Y.-K. (2008). Environmentally responsible happy nation index: Towards an internationally acceptable national success indicator. Social Indicators Research, 85, 425–446. Noorbakhsh, F. (1998). A modified human development index. World Development, 26, 517–528. Nordhaus, W. D., & Tobin, J. (1972). Is growth obsolete? Economic Growth. 50th anniversary colloquium V. Columbia University Press for the National Bureau of Economic Research. New York. Reprinted In Moss Milton (Ed.), The measurement of economic and social performance, Studies in income and wealth (Vol. 38). National Bureau of Economic Research, 1973. Nordhaus, W. D. (1991). To slow or not to slow? The economics of the greenhouse effect. Economic Journal, 101, 920–937. OECD (2006). Going for growth. Paris: OECD. Pissarides, C. A. (2000). Equilibrium unemployment theory (2nd ed.). Cambridge, MA: MIT Press. Saget, C. (2000). Can the level of employment be explained by GDP growth in transition countries? Theory versus the quality of data. Labour, 14(4), 623–643. Samuelson, P. A. (1961). The evaluation of social income: Capital formation and wealth. In F. Lutz & D. Hague (Eds.), The theory of capital. New York: St. Martin’s Press. Schorr, J. (1998). The overspent American. New York: Basic Books. Scitovsky, T. (1976). The joyless economy. New York: Oxford University Press. Simon, H. A. (1990). A mechanism for social selection and successful altruism. Science, 250(4988), 1665–1668. Simon, J. (1981). The ultimate resource (2nd ed., 1996). Princeton: Princeton University Press. Sen, A. (1976). Real national income. Review of Economic Studies, 43(1), 19–39. Sen, A. (1979). The welfare basis of real income comparisons. Journal of Economic Literature, 17(1), 1–45. Sen, A. (1999). Commodities and capabilities. Oxford: Oxford University Press. Sen, A. (2000). Development as freedom. Oxford: Oxford University Press. Söderholm, P. (2007). Modelling the economic costs of climate policy, Research Report 2007, 14, Luleå University of Technology.. Solnick, S., & Hemenway, D. (1998). Is more always better? A survey on positional concerns. Journal of Economic Behaviour and Organization, 37, 373–383. Stern, D. I., Common, M. S., & Barbier, E. B. (1996). Economic growth and environmental degradation: A critique of the Environmental Kuznets Curve. World Development, 24, 1151–1160. Stiglitz, J. E. (2005). The ethical economist – A review of ‘‘The moral consequences of economic growth”. In Friedman, B. M. (Ed.) (Knopf, 2005), Foreign affairs (November/December). Tinbergen, J., & Hueting, R. (1992). GNP and market prices: Wrong signals for sustainable economic success that mask environmental destruction. In R. Goodland, H. Daly, & S. El Serafy (Eds.), Population, technology and lifestyle: The transition to sustainability. Washington, DC: Island Press. Tol, R. S. J. (2008). The social cost of carbon: Trends, outliers, and catastrophes. Economics, the Open-Access, Open-Assessment E-Journal, 2(25), 1–24. Tversky, A. (1972). Elimination by aspects: A theory of choice. Psychological Review, 79, 281–299. Tversky, A., & Simonson, I. (2000). Context-dependent preferences. In D. Kahneman & A. Tversky (Eds.), Choices, values and frames (pp. 518–527). Cambridge: Cambridge University Press. van Praag, B. M. S. (1971). The welfare function of income in Belgium: An empirical investigation. European Economic Review, 2, 337–369. van Praag, B., & Ferrer-i-Carbonell, A. (2004). Happiness quantified: A satisfaction calculus approach. Oxford: Oxford University Press. Veblen, T. (1899). The theory of the leisure class: An economic study in the evolution of institutions. New York: Macmillan. Vellinga, N., & Withagen, C. (1996). On the concept of green national income. Oxford Economic Papers, 49(4), 499–514. Weil, D. N. (2005). Economic growth. Boston: Pearson Education, Addison-Wesley. Weitzman, M. L. (1976). On the welfare significance of national product in a dynamic economy. Quarterly Journal of Economics, 90, 156–162. Weitzman, M. L., & Löfgren, K.-G. (1997). On the welfare significance of green accounting as taught by parable. Journal of Environmental Economics and Management, 32, 139–153. World Bank (2006). Where is the wealth of nations. Washington, DC: The World Bank.
