HUMAN CAPITAL IN SUSTAINABLE WEALTH MEASUREMENT

(1)

HUMAN CAPITAL IN SUSTAINABLE WEALTH

MEASUREMENT

A Proposal for Improved Measurement

MSc thesis International Economics and Business

University of Groningen – Faculty of Economics and Business

Abstract

This thesis focuses on the definition of human capital in sustainable wealth measurement, and implements the lifetime income approach of human capital into adjusted net savings (ANS) for eighteen countries over country-specific five year periods. The modified model contributes to the idea that policymakers need a proper instrument to assess whether current policies cause any over-consumption or underinvestment that negatively affects the future generation. The results show that the modified ANS estimates are significantly different from the traditional ANS estimates that use (public) education expenditure as a proxy for human capital investments. The changed human capital definition is used to assess whether or not countries have non-negative investments in genuine wealth that prove compliance to the sustainability criterion. After adjusting for both population and total factor productivity change, Argentina, Denmark and Japan do not comply with the sustainability criterion, i.e. its current policies do not result in non-negative investments in wealth. These results should be interpreted with caution due to the use of a disputable GDP/wealth ratio in the methodology. Based methodological limitations of both the lifetime income itself and the implementation of HCI in ANS, it is concluded that the lifetime income approach is too premature to apply and that future research should focus on the use of a different global sustainable wealth measure.

Key words: human capital, lifetime income approach, sustainability, (genuine) wealth, adjusted net savings

Author: M.M. Vermeulen

Student ID number: 1934309

Date: 19/06/2014

(2)

2

LIST OF TABLES

TABLE 4.1 REAL INCOME GROWTH RATES AND DISCOUNT RATES ... 24

TABLE 4.2 AVERAGE EE AND HCI AS PERCENTAGE OF GDP ... 32

TABLE 5.1 GENUINE INVESTMENTS (INCL. HCI) AND COMPONENTS AS PERCENTAGE OF REAL GDP ... 34

TABLE 5.2 GROWTH RATES OF PER CAPITA GENUINE WEALTH ... 37

TABLE 5.3 SENSITIVITY ANALYSIS GDP/WEALTH RATIO ... 39

TABLE 5.4 SENSITIVITY ANALYSIS CO2 EMISSIONS ... 40

TABLE B.1: LIFETIME INCOME DATA AVAILABILITY ... 51

TABLE C.1 EDUCATIONAL ATTAINMENT SPAIN 2001-2006 ... 53

TABLE D.1 RANKING COUNTRIES WITH LIFETIME INCOME AS HUMAN CAPITAL COMPONENT (% OF GDP) ... 57

TABLE D.2 RANKING COUNTRIES WITH EDUCATION EXPENDITURE AS HUMAN CAPITAL COMPONENT (% OF GDP) ... 57

TABLE E.1 DECOMPOSITION NATURAL RESOURCE DEPLETION ... 58

TABLE F.1 ADJUSTED GROWTH RATES PER CAPITA GENUINE WEALTH (0.20 AND 0.30 GDP/WEALTH RATIO) ... 59

LIST OF FIGURES

FIGURE 2.1 WEALTH MEASUREMENTS – CLASSIFICATION OF APPROACHES ... 8

FIGURE 3.1 HUMAN CAPITAL MEASUREMENT - CLASSIFICATION OF APPROACHES ... 16

FIGURE 4.1 AVERAGE RATIO HUMAN CAPITAL STOCK TO REAL GDP ... 25

FIGURE 4.2 POPULATION AGES 0-14 (% OF TOTAL)... 26

FIGURE 4.3 EDUCATION EXPENDITURE AS % OF REAL GDP ... 27

FIGURE 4.4 HUMAN CAPITAL INVESTMENTS AND EDUCATION EXPENDITURE AS % OF REAL GDP ... 28

FIGURE C.1 INCREASE IN EDUCATIONAL ATTAINMENT AND ANNUAL EARNINGS ROMANIA 2002-2006 (%) ... 52

FIGURE C.2 CHANGE IN ANNUAL EARNINGS SPAIN 2001-2006 (%) ... 53

FIGURE C.3 INCREASE IN EDU AND ANNUAL EARNINGS NETHERLANDS 2002-2006 (%) ... 54

FIGURE C.4 INCREASE IN EDUCATIONAL ATTAINMENT AND ANNUAL EARNINGS THE UK 1997-1998 (%) ... 54

FIGURE C.5 INCREASE IN ANNUAL EARNINGS JAPAN 2002-2007 (%) ... 55

FIGURE C.6 CHANGE IN ANNUAL EARNINGS USA ‘00-’02 ... 55

FIGURE C.7 CHANGE IN ANNUAL EARNINGS NEW ZEALAND 2005-2006 ... 55

FIGURE C.8 CHANGE IN EDUCATIONAL ATTAINMENT NEW ZEALAND 2005-2006 (%) ... 55

FIGURE C.9 INCREASE IN EDUCATIONAL ATTAINMENT AND ANNUAL EARNINGS ITALY 2000-2002(%) ... 56

FIGURE C.10 INCREASE IN EDUCATIONAL ATTAINMENT AND ANNUAL EARNINGS ISRAEL 2003-2004(%) ... 56

(4)

4

LIST OF ABBREVIATIONS

AIN Annual income

ANS Adjusted net savings

CW Comprehensive wealth

CFC Consumption of fixed capital CO2D CO2 damages

DAS Depletion adjusted savings EDU Educational attainment EE Education expenditure EF Ecological footprint EMR Employment rate END Energy depletion GDP Gross domestic product GNI Gross national income GNS Gross national savings

HCI Human capital investment according to lifetime income approach HCP Human capital per capita

HCV Human capital stock HDI Human development index IC Intangible capital

LIN Lifetime income

MEW Measure of economic welfare MID Mineral depletion

MPV Maximize present value NFA Net financial assets NFD Net forest depletion NNS Net national savings NUM Number of individuals

OECD Organization of Economic Co-operation and Development PMD Particulate matter damages

POP Population

RD Resource depletion

SMEW Sustainable measure of economic welfare SNA Systems of national accounts

SUR Survival rate

(5)

5

1. | INTRODUCTION

Gross domestic product (GDP) is the most commonly used measure of economic activity and growth. It measures the monetary market value of all final goods and services produced in a country in one year of time. Real GDP per capita is used to compare a country’s position in economic development over time or relative to other countries, often referred to as ‘the standard of living’. Although it was originally intended to measure economic activity of a country, GDP has been incorrectly interpreted as a measure for social welfare of a country’s population. Therefore, ever since the development of GDP by Keynesian economists in the 1930s and its institutionalization in the internationally agreed standard Systems of National Accounts (SNA) in the 1950s, GDP has been widely criticized for not adequately capturing human welfare and progress (Van der Bergh, 2009). Most national governments, however, have continued to use GDP as the indicator for growth and welfare. Governmental policies promote economic growth; however, if our measure of growth is unsound it may imply that the current policy decisions are distorted (Stiglitz et al., 2010).

The incorrect interpretation of GDP fueled the discussions on ‘beyond GDP’ by researchers and policy makers, resulting in the development of new measures of welfare. The motivation behind the creation of new measures is that measurements on the development of countries actually determine how countries will develop. This implies that governments are currently merely interested in increasing the production of final goods and services resulting in higher GDP values, which is perceived as positive economic development. The impact of this ‘development’ on the well-being or wealth of the current and future population, however, is not captured by GDP. Consequently, the implicit task of the government to ensure intergenerational well-being of its population is not fulfilled by development consistent with GDP growth only.

The developed new measures of welfare assess either current well-being or sustainability. Whereas current well-being is associated with both economic and non-economic aspects of the current generation, sustainability focuses on the maintenance of the current level of well-being for future generations. The concern for sustainable development was driven by initial research (Nordhaus and Tobin, 1972), the Brundtland Report (1987) and the United Nations’ Rio Summit (1992) which indicated that economic growth in many countries was not sustainable due to natural resource depletion or environmental damages. This means that the measurement for development in place (GDP) caused economic growth, but also environmental deterioration. Literature on sustainable development grew following these concerns and defined sustainability as

development that meets the needs of the present without compromising the needs of future generations

(Brundtland Commission, 1987; Pezzey, 1989). Measurements of sustainability can address whether or not a nation’s current policy is sustainable in the sense that future generations will have at least the same level of well-being; nations should not be over-consuming or underinvesting.

The importance of a sustainability measure should be stressed: by measuring sustainability, countries are driven to develop sustainably. However, not all indicators that tried to incorporate sustainability have done so adequately. According to Stiglitz et al. (2010), assessing sustainability is complementary to current well-being and should therefore be measured separately. Indicators of sustainability that appropriately comprehend all potential determinants of future well-being are monetary stock-based indicators. These monetary indicators rely on the notion that all determinants of intergenerational well-being are the multitude of capital assets (Hamilton and Clemens, 1999; Dasgupta and Maler, 2000). A country’s ‘wealth’ is therefore expressed as the weighted sum of produced, natural and human capital.

(6)

6

depletion must be reinvested in human or physical capital to maintain the capital stock at least at the same level.

From these three capital stocks, human capital is considered to be the most important form of capital that drives economic growth (Romer, 1986; World Bank 2006, 2011; Liu, 2011). While the World Bank (2011) provides human capital estimates of 55 percent of total wealth, Jorgenson and Fraumeni (1989) estimate human capital to be 80 percent of total wealth. Although the range between these estimates is large, both indicate that human capital is of considerable importance to a nation’s wealth account. Human capital has been addressed by using diverse methods, resulting in the difference in estimates. The stock-based welfare measures developed by the World Bank (2011) use two different approaches in measuring human capital; either measuring human capital indirectly, by subtracting estimates of natural and physical capital from a calculated total wealth, or directly, as the total annual public expenditure on education in a country. Both methods have been widely criticized for not addressing human capital investment and divestment appropriately. The indirect method in measuring human capital is sensitive to measurement errors; human capital estimates are reliant on total wealth as well as on produced and natural capital calculations. Taking annual public expenditure on education as a proxy for human capital is criticized for ignoring the efficiency and time dimension with which the inputs (education expenditures) are used to produce an output (human capital). Both the indirect method and the direct estimations of the inputs of human capital as a representative of the output are therefore considered inappropriate.

The lack of consensus on the measurement of human capital and the importance of human capital for total wealth were incentives for the OECD to launch the Human Capital Project in October 2009. This project aimed to develop national human capital stocks in a common metric. Based on the work by Jorgenson and Fraumeni (1989), the OECD, under the lead of Liu (2011), developed human capital accounts for sixteen countries using the lifetime income approach.

This recent development in the measurement of human capital makes it possible to address the shortcomings in the measurement of human capital in stock-based measurements of sustainable wealth (World Bank, 2011; Arrow et al, 2012). The aim of this thesis is to provide an improved holistic sustainability measure that combines insights from both the existing literature on stock-based welfare measures and recent research on human capital (Liu, 2011). In this way, a greater understanding of the contribution of human capital to the wealth of nations can be attained. This research contributes to the notion that policymakers need a proper instrument to assess whether current policies cause any overconsumption or underinvestment that negatively affects future generations. The urgency of the need of these instruments is stressed by the increased interest in sustainable development by governments’ and their population. This is supported by the previously introduced idea that “how we measure development will drive how we do development” (Worldbank, 2011, p. 18).

(7)

7

2. | MEASURING SUSTAINABLE WEALTH

The purpose of this section is twofold; it discusses literature on measurements on current well-being and sustainability, and explains the methodology behind the sustainable wealth measures of interest: Comprehensive Wealth (CW) and Adjusted Net Savings (ANS). The combination of all sub-sections provides a holistic assessment on the development of wealth measurement as a result of the discussion on ‘beyond GDP’. Sections 2.3 until 2.6 are most important to understand the intuition behind sustainable wealth and the corresponding methodologies. The definition of sustainability, the central concepts of ‘social rate of return’ and ‘residual approach’, and the difference between capital stock versus capital net investments are introduced and explained. The discussion on the methodologies of CW and ANS suggests that the component human capital can be improved to more appropriately measure sustainable wealth, as I attempt later in this thesis.

Research on measurements of current well-being and sustainability as a complement to or substitute for GDP is widespread and diverse. Whereas some authors focused on modest corrections to GDP, others have developed completely new indices based on entirely different data. In general, there are three types of measurements that were developed since the introduction of GDP: subjective measures of well-being, composite indices and monetary approaches. Subjective measures are based on surveys about the happiness or well-being self-reported by the respondents and can be misleading in their results due to differences in perception across nations (Fleurbaey and Blanchet, 2013). In composite indices, income or GDP per capita is usually one component among others such as health, poverty and employment. Monetary approaches come in different forms; either by correcting GDP, focusing on ‘equivalent income’ or calculating capital stocks and investments (stock-based approach). Although it is not the intention of this thesis to provide a complete survey1_{of wealth}

measurements, several contributions to the literature provide a background to the discussion on ‘beyond GDP’. Arrow et al. (2004) encountered the difficulty that arises in assessing both current well-being and sustainability in one wealth measure. Their efforts direct towards distinguishing between assessing current well-being and measuring sustainability; the goals as well as the implications for policy-makers of each of these approaches are different.

2.1 The Maximize Present Value criterion

Arrow et al. (2004) developed two criterions to assess whether a nation’s consumption is excessive or not: (1) the maximize present value (MPV) criterion and (2) the sustainability criterion2_{. The MPV criterion uses a}

measure of inter-temporal social welfare, which discounts the value of current and future utility of consumption from present to infinity, based on economic theory. Under MPV, current consumption is excessive when it is greater than the consumption prescribed by the optimal consumption path. This means that if lowering consumption today and increasing investment in capital assets could raise future utility to more than compensate for the loss in utility today, current consumption is excessive. The optimal consumption path takes into account both current consumption and future consumption, and can therefore be interpreted as an assessment of both current and future well-being. The authors, however, point out that “no one can seriously claim to pinpoint the optimal level of current consumption for an actual economy” (Arrow et al., 2004, p. 155). The optimal path depends on the discount rate; a higher value of the discount rate means that less weight is placed on future utility. The correct value of this discount rate is highly debated and no consensus exists on the trade-off between current and future utility.

Theoretically, current overconsumption can be assessed by evaluating the market rate of return on investment in comparison to the social rate of interest on consumption. According to Arrow et al. (2004), the optimal path requires the two to be equal. In practice, however, it is difficult to calculate the social rate of interest, since it is dependent on an assumed social rate of pure time preference, an elasticity of marginal utility and the growth rate of consumption3_{. According to the authors, excessive consumption is caused by the absence of complete}

1_{See Fleurbaey and Blanchet (2013) and Stiglitz, Sen and Fitoussi (2010) for extensive surveys on measurements of current well-being}

and sustainability.

2_{The sustainability criterion is discussed in section 2.4.}

3_{The Ramsey formula states that r=δ+ g, where r equals the social rate of interest on consumption, δ is the social rate of pure time}

(8)

8

FIGURE 2.1 WEALTH MEASUREMENTS – CLASSIFICATION OF APPROACHES

pooling of risks resulting in uncertain return on investments. Furthermore, excessive consumption is triggered by the taxation of capital income and the underpricing of natural resources relative to social costs.

The difficulty of combining both current and future consumption, based on a social rate of interest, resulted in Arrow et al. (2004) to focus on their sustainability criterion instead of the MPV criterion. Others have instead focused on measures of current well-being. Therefore, in the subsequent sections a clear distinction is made between measurements of current well-being and measurements of sustainability; this is visually represented in figure 2.1. Although all the displayed approaches are discusses, emphasis lies on monetary stock-based measures of sustainability and the corresponding human capital component; arguments for this emphasis are discussed below.

2.2 Current well-being

Current well-being is associated with both economic and non-economic aspects of the current generation, without taking into account the future generation. The Human Development Index (HDI), published yearly by the United Nations since 1990, is a composite index including both these economic and non-economic aspects. The components of HDI are Gross National Income (GNI), life expectancy at birth as a proxy for health, and educational achievement measured by enrollment rates and adult literacy rates4_{. The measure has been}

criticized for being too arbitrary (Kelley, 1991; Sagar and Najam, 1998; Ravallion, 2010) and numerous proposals for extension have been written. There is, however, no economic theory that can guide the composition of indicators into a holistic measure.

Whereas HDI measures current well-being using objective data, subjective measures of well-being rely on population survey data such as the World Values Survey or the World Gallup Poll. Deaton (2008) uses this latter survey to evaluate people’s life satisfaction compared to their income and health situation. Although economists generally presume that higher incomes increase satisfaction, Easterlin (1974, 1995) notably concluded that no long run relationship exists between a nation’s income and the average life satisfaction of its population. Deaton (2008), however, does find a strong global relationship between life satisfaction and GDP per capita. Nevertheless, his results, especially regarding health, show that subjective measures can be misleading in their outcomes as they may deviate significantly from objective measurements. People in countries such as Vietnam, Thailand and Cuba have greater confidence in their health care system than the United States (Deaton, 2008) despite quality of care being much lower in these low-income countries.

(9)

9

Subjective measurements are therefore not suited to compare well-being across nations. Consequently, policy decisions cannot be made based on differences found through subjective survey data.

Jones and Klenow (2010) do aim to compare welfare across nations and across time by developing a money-metric approach, based on objective measures, that has foundations in economic theory. They extend the economic approach focused on consumption, such as with the MPV criterion, by including the calculated consumption-equivalent of income inequality, leisure, and life expectancy in their model. Jones and Klenow (2010) make international comparisons by computing the change in per capita consumption that would be required in each country to bring it up to the welfare level of the United States, taking into account its own modeled non-GDP characteristics. In this attempt to create an objective money-metric measure of welfare, the authors recognize the limitation of this consumption-equivalent measure, as it neglects the natural environment and measures health quantitatively instead of qualitatively. The calculation of equivalent consumption compares individuals in a hypothetical world in which all else is held equal and differences are attributed to consumption. Contrasting the economic approach by Arrow et al. (2004), Jones and Klenow (2010) focus on data specified to one year only; future consumption is not taken into account.

The main caveat of measuring current well-being in general is related to this exclusion of the future generation; it does not provide policy-makers with a specific benchmark. With the growing importance of achieving sustainable growth, policy-makers are interested in whether the current development is leading to any overconsumption or underinvestment. Although the MPV criterion (Arrow et al., 2004) concentrates on overconsumption, the criterion does not require well-being to be equal or higher for a future generation. The welfare of future generations is largely neglected by measures of current well-being, although the task of governments and other policy-makers is to ensure inter-generational social welfare (Arrow et al. 2012). It is of low interest to promote current well-being if this is at the expense of future generations; we should therefore focus on “meeting the needs of the present without compromising the future to meet their own needs” (Brundtland Commission, 1987). Measuring sustainability focuses on this sustainable development of nations.

2.3 Sustainability

Nordhaus and Tobin (1972) made a seminal contribution to the literature by developing the first “Sustainable Measure of Economic Welfare” (SMEW). The authors applied the approach of the “corrected GDP” type, which follows the idea to retain as much of the accounting consistency of GDP as possible (Fleurbaey and Blanchet, 2013). To develop the SMEW, the authors first derived a Measure of Economic Welfare (MEW), which started with total private consumption, subtracted components that negatively influenced welfare and added components that positively influenced welfare. MEW was converted to SMEW by taking into account changes in total wealth. A clear distinction is made in this conceptual framework between measuring current well-being (MEW) and sustainability (SMEW).

This indicator of the “corrected GDP” type inspired others to develop new measures in adjusting GDP, such as the Index of Sustainable Economic Welfare and the Net or Green Domestic Product5_{. The early attempt by}

Nordhaus and Tobin (1972), however, aimed to separate current well-being and sustainability, whereas merely correcting GDP focuses on one indicator that combines both categories. Clearly separating measurements of current well-being and its sustainability is what is advocated in this thesis, which in line with the literature (Stiglitz et al., 2010; Fleurbaey and Blanchet, 2013; Arrow et al., 2003, 2004, 2012): “measuring how well we are or how much we consume is one thing, measuring whether this corresponds to overconsumption is another” (Fleurbaey and Blanchet, 2013, p. 21).

There are only two types of indicators that do address the question of sustainability stricto sensu: the monetary stock-based indicators and footprint indicators. One of the footprint indicators is the Ecological Footprint (EF). The EF was developed by Wackernagel and Rees (1996) and addresses the sustainability issue along the

5_{“Greening” GDP corrects GDP for the depletion of natural resources or pollution; it does not deal with the more holistic sustainability}

(10)

10

environmental dimension. The approach focuses on land area that is required to support the flows of energy and material to and from an economy. In other words, the value of the EF measure for a country “represents the land area necessary to sustain current levels of resource consumption and waste discharge by that population” (Wackernagel and Rees, 1996, p. 5) and is displayed in global hectares. The EF deals with the specific ecological aspects of sustainability and is considered to be in line with the strong sustainability view. This view advocated by ecologists relies on the idea that there are critical levels of natural capital that must be preserved to maintain welfare; thus, there are no substitutes for certain natural assets.

Although the EF has potential in measuring sustainability along the environmental dimension, the method has been criticized on multiple grounds by Van den Bergh and Verbruggen (1999). Among others, the EF is biased against trade and does not distinguish between sustainable and unsustainable land area (Van den Bergh and Verbruggen, 1999). Moreover, the EF only focuses on one determinant of future well-being –natural capital– whereas monetary stock-based indicators encompass all potential determinant of future well-being. Therefore, it does not assess whether a country is over-consuming or underinvesting its current total wealth. Monetary stock-based measures of wealth do address sustainability in this broader form and is addressed next.

2.4 Monetary stock-based wealth measures

Monetary stock-based wealth measures are based on the sustainability criterion (Arrow et al., 2004). Contrasting the MPV criterion, the sustainability criterion prescribes that the inter-temporal social welfare must not decrease over time and therefore focuses on the change in welfare, not on the welfare level. Whereas the MPV criterion is based on the discounted value of current and future consumption, this criterion focuses on

capital. The former criterion assumes capital assets as the productive base for current and future consumption,

but does not require that this base needs to increase over time. Therefore, the MPV criterion does not automatically measure whether development is sustainable, whereas the sustainability criterion does.

The sustainability criterion can only be satisfied if the change in capital assets is non-negative, where capital assets consist of produced-, natural- and human capital. The intuition behind this definition of sustainability can be explained by the production function; capital assets are the factors of production to produce GDP as an output. If capital assets (input) deteriorate over time, GDP or wealth (output) is unable to sustain over time with an unchanged production function. This implies that a negative change in capital assets does not conform to the sustainability criterion. The production function, however, is influenced by technological change. Positive technological change causes the production of more outputs with the same amount of inputs (capital) due to efficiency improvements. This means that technological change can influence the interpretation of sustainable development. Population change affects sustainable development similarly; a country with non-negative change in capital assets, but with significant population change can have negative changes in capital assets per

capita. In sum, development is only sustainable if and only if the change in capital assets per capita is

non-negative after adjustments for technological change.

(11)

11

There are two approaches within this segment of wealth measurements; either the capital stocks are calculated or the net investments in the capital assets from one period to another are computed. The World Bank (2011) is a leader in both approaches by using “Comprehensive Wealth” (CW) and “Adjusted Net Savings” (ANS) to measure the sustainability of a nation’s current policies. Whereas CW measures total wealth and the stock of each capital category’s contribution, ANS focuses on changes in wealth by gross investments in the capital assets minus depreciation and depletion of those same assets. The difference in the applied methodology between the two approaches is especially consequential for the human capital component. The theoretical background and methodology of each of these approaches are important to understand the implications and limitations, which are the basis for possible improvement in sustainable wealth measurement, as I attempt later in this thesis.

2.5 Comprehensive Wealth

Implemented by the World Bank (2011) and measured with five year intervals from 1995 until 2005, CW is based on work by Hamilton and Hartwick (2005). The authors base their estimates on the assumption of a competitive economy with constant returns to scale. For the production function Y=F(K,L,R) with production factors K (produced capital), labor L, natural resource flow R and interest rate r, comprehensive wealth is calculated as:

∫ ( ) ( ) _(2.1)

where is the total value of wealth (or capital) at time t; is the asset value of produced capital at time t; is the asset value of human capital at time t; equals the asset value of the natural capital stock at time t; C(s) is consumption in year s; and r is the social rate of return on investment.

The equation suggests two ways of calculating comprehensive wealth; either by adding up all capital assets (K,H,S) at time t, or by measuring the calculated present value of future consumption along the optimal consumption path, which is in line with the methodology of the MPV criterion (Arrow et al., 2004). The intuition behind this definition is that consumption is interpreted as the (present) service value of the capital assets’ output during the lifetime of the assets.

Equation 2.1 implies that in the long run, a country can only consume within the limits of the sum of all its assets. The equation implicitly assumes that the level of savings is enough to account for the depletion of natural resources, and is thus sustainable. Some countries, however, have high levels of depletion resulting in a negative depletion-adjusted savings. Consumption of these natural resources impedes prospects of future consumption, especially for non-exhaustive natural resources such as oil and gas. To account for this issue, the World Bank (2011) implements a measure of sustainable consumption to adjust for sustainability in the optimal consumption path, so that

∫ ( ) ( ) (2.2)

where is sustainable consumption at time t.

Sustainable consumption depends on the level of depletion-adjusted savings (DAS). If a country has negative depletion-adjusted savings, the optimal consumption path needs to be adjusted by DAS, so that consumption is on a sustainable path and within the limits of its assets. Put differently, correcting for negative DAS implies that the sum of its capital stocks must not decrease over time, which is consistent with the economic meaning of sustainability. For a country that has positive depletion-adjusted savings, the current consumption path already is sustainable as it leaves the capital stock intact and does not need adjustment for DAS. Therefore, sustainable consumption is defined as follows.

(2.3)

(12)

12

where is Sustainable Consumption at time t; is Consumption at time t; and is Depletion Adjusted Savings at time t, which is defined by equation 2.4.

(2.4)

where GNS is the Gross National Savings; CFC is the Consumption of Fixed Capital; END is the Energy Depletion; MID is the Mineral Depletion; and NFD is the Net Forest Depletion.

Note that DAS includes produced capital investments in the form of GNS minus CFC, and natural capital depletion by inclusion of END, MID and NFD. Human capital investments, however, are not included in depletion adjusted savings. Although an exact explanation cannot be given for the exclusion of human capital investments by the World Bank (2011), the difference between the approaches of measuring produced and natural capital on one side, and human capital on the other suggest that the methodology does not allow accounting for human capital investments in DAS6_.

Produced capital and natural capital are measured directly7_{. The World Bank (2011), however, chooses to}

measure human capital indirectly due to the given difficulties in obtaining monetary estimates for intangible assets (Ferreira and Hamilton, 2010). Consequently, total wealth is first calculated as the present value of future consumption dependent on the social rate of return r. Based on the Ramsey formula, which calculates how much a consumer would need to be compensated for deferring a unit of consumption from the current period for the next, the social rate of return to investment r is calculated as:

(2.5)

where represents the pure time preference rate of consumption; is the elasticity of utility with respect to consumption, and can be interpreted as the preference for equality of consumption among generations (Arrow et al., 2004); and is the growth rate of consumption.

For equation 2.4 to be applicable, several judgment calls have to be made related to multiple parameters. Recall the difficulty that Arrow et al. (2004) had in calculating the social rate of interest, which was discussed in section 2.1. The World Bank (2011) assumes and a constant consumption growth rate. This consumption growth rate is assumed to be 2.5 percent, although historical values are usually less than 1.5 percent. Furthermore, the time preference is assumed to equal 1.5, based on empirical estimates by Pearce and UIph (1999), who have determined that time preference usually to range between 1 and 2. With these three assumptions, the social rate of return on investment r, used as a discount rate to calculate the net present value, is set at 4 percent. Finally, although theory assumes infinite lifetime for the analysis as shown by , the World Bank (2011) chooses a limited time horizon of 25 years - roughly corresponding to one generation - to be able to solve the equation. Therefore, total wealth approximation using the CW method is based on the present value

of the current level of (sustainable) consumption, taken over 25 years and discounted using a social rate of return of 4 percent.

As stated above, CW measures human capital indirectly, which is considered to be highly controversial (Liu, 2011). The World Bank (2011) considers human capital to be the main component of the intangible capital residual. This residual is calculated as follows:

(2.6)

where is Intangible Capital at time t; and is Net (foreign) Financial Assets at time t for which the country receives an income or pays interest (World Bank, 2006)8_.

6_{The World Bank (2011) could have adjusted for human capital investments in DAS by including education expenditure, since the ANS}

methodology (section 2.6) and the CW methodology were developed during a similar period of time. The World Bank (2011), however, sees these methods as two distinct methods, which could be an explanation for the exclusion of human capital in DAS.

7_{See appendix A for the specific methodology on the decomposition of CW}

8_{The inclusion of NFA accounts for foreign assets owned by the country and assets in the country owned by a foreign country for which}

(13)

13

The calculation of intangible capital as a residual particularly shows that comprehensive wealth is a top-down method of calculating the components of wealth. Although human capital is believed to dominate the intangible capital residual, other components, such as institutional capital and missing natural capital, are included in the residual as well. The critical role of human capital in driving economic growth has been emphasized by several authors (Romer, 1986; Lucas, 1988; World Bank 2006, 2011); measuring human capital indirectly does not do justice to the importance of this asset.

The “top-down” CW approach makes the human capital residual highly sensitive to measurement errors of both total wealth and the components of natural and produced capital. Although the chosen values of the parameters are in the range used in previous research (Rosegrant et al. 1995; Lampietti and Dixon, 1995; Klenow and Rodríguez-Clare, 1997; Kunte et al, 1998; Pearce and Ulph, 1999; Merlo and Croitoru, 2005; Weil, 2007; Kaufman, Kraay and Mastruzzi, 2009), the values do influence the implications. The social rate of return used to discount consumption, and also natural capital depletion, is calculated using three assumptions: ƞ=1, ρ=1.5 and g=2.5. These parameters are held constant across countries; however, it is highly likely that the pure time preference, which is based on a typical individual’s time preference, is different across nations. Individuals with an extremely low income concentrate on survival, indicating that these individuals have a short-term focus and therefore a high time preference ρ. An increased income causes individuals to be able to invest and focus on the longer term, resulting in a lower ρ. It is therefore expected that low-income countries have a different time preference than high-income countries. Similar critique applies to the assumed consumption growth rate. Historical growth rates are usually lower than 1.5 percent; however the World Bank (2011) uses an assumed growth rate of 2.5 percent for all countries. The reasoning for this high value of consumption growth compared to historical growth rates is unclear.

Consequently, it can be argued that this “top-down” approach makes the estimates of human capital too sensitive to measurement errors. Furthermore, measuring human capital indirectly cannot explain what drives the observed changes in the stock of human capital over time (Liu, 2011). The goal of a welfare measure is to implement the measure for policy-makers’ decisions. This method, however, makes it impossible to identify the causes for growth of human capital.

From a policy-maker’s perspective, a sustainable wealth measure should facilitate the exploration of the causes of an (un)sustainable development path. The residual approach is unsupportive in the determination of the exact causes for growth, since a major component – human capital – is measured indirectly. The insertion of a direct approach in estimating human capital stocks in CW can reduce the intangible capital residual, and provide policy-makers with guidance to draw useful policy implications on the direct estimates of the components. Section 3 discusses direct approaches in human capital measurement and proposes to implement the lifetime income approach in sustainable wealth measurement. Although this direct approach can be inserted in CW, ANS provides more potential due to data availability; section 3.5 elaborates on this.

2.6 Adjusted Net Savings

In contrast to CW, Adjusted Net Savings (ANS) measures the change in total social capital assets, including human, natural and produced capital. The ANS capital account can be interpreted as the investments or disinvestments in a nation’s capital assets depending on positive or negative values. The main difference between CW and ANS is that ANS does not require estimating capital stocks before determining whether or not the change in capital stocks is non-negative. In other words, ANS measures the change in capital assets directly and is therefore less data intensive. This advantage of ANS over CW also shows in the years of data available published by the World Bank (2011): CW is measured with five year intervals from 1995 until 2005, whereas ANS is estimated yearly between 1970 and 2012.

ANS is based, amongst other, on work of Hamilton (1994) and Pearce et al. (1996), and is also termed “net savings”, “genuine savings” or “genuine investments” (Arrow et al., 2004). These true savings are calculated by taking into account investments in human capital, natural resource depletion and pollutions, so that

(14)

14

(2.7)

where equals the adjusted net savings in time t; is net national savings; is public education expenditures; is resource depletion, is damages; and equals the Particulate Matter damages. All components are denoted to time t.

Researchers at the World Bank were leaders in computing the genuine savings for a large number of nations. Whereas Pearce and Atkinson (1993) tracked the genuine investments for only 18 countries, Hamilton and Clemens (1999) computed average net savings for a large sample of countries for the period 1970-1993. This was among the first efforts in large scale comparative sustainability measurement. The authors show that “genuine savings” were positive for all high income industrialized countries, while a large number of developing countries had negative values of net savings in the same period9_{. The calculated negative values of}

ANS were mainly for resource-rich developing countries, which displays that these nations did not invest sufficiently in human and produced capital compared to the depletion of their natural resources. The World Bank (2011) has continued the efforts of measuring the genuine investments of nations, and published yearly ANS numbers for almost all nations based on the methodology of ANS. The detailed methodology on the estimates of the components of ANS is described in appendix A.

Note the difference between the calculation of depletion adjusted savings in equation 2.2 and the definition of ANS in equation 2.7. In contrast to DAS, ANS does include human capital investments in the equation by the term EE. Although DAS is used in CW to account for the sustainability of the optimal consumption path, human capital is not included in DAS nor is the calculation of human capital in CW appropriate10_{. Contrary to CW, ANS}

measures human capital, as well as the other components, directly. Measurement errors are therefore limited in this “bottom-up” approach (Hamilton and Liu, 2013) compared to CW’s “top-down” approach.

Human capital is calculated as the current public expenditures on education (World Bank, 2011); data on private expenditure on education are not widely available and therefore not included. The argument for choosing this particular human capital measure is that expenditures on education can be seen as investments in human capital. Public education expenditures, however, do not measure the outcome of the education and therefore not human capital per se. Furthermore, divestments of human capital are not accounted for (such as retirement) and changes in public education expenditures cannot explain the specific drivers behind the change, such as higher returns to education or higher education levels. These specific drivers are important for policy-makers to be able to make decisions accordingly.

Measuring human capital directly does justice to the importance of human capital in wealth accounting. However, the limitations of taking public education expenditures as a proxy for human capital investments show that the requirements of a proper human capital measure are not satisfied. EE does not measure human capital outcome nor does it provide policy-makers with specific causes for growth or (un)sustainability.

This section has made clear that the available monetary sustainability measurements can be improved in their human capital component. However, no clear consensus exists on “measuring human capital”. Section 3 therefore evaluates several approaches and proposes an appropriate methodology in measuring human capital, based on research by the OECD (Liu, 2011).

9_{Table 3 in Hamilton and Clemens (1999) provides an overview of the “genuine savings” of all countries in the sample. Developing}

countries with negative net savings include Trinidad and Tobago, Venezuela, Saudi Arabia, the Republic of Congo, Nigeria and other resource rich countries.

10_{Recall the two arguments: (1) human capital measurements in CW are subject to measurement errors, and (2) measuring human}

(15)

15

3. | MEASURING HUMAN CAPITAL

The human capital component in monetary sustainable wealth measurement has been criticized on multiple grounds in section 2. The criticism is an incentive to explore other methods of human capital estimation, which is the purpose of this section. The importance of human capital for economic growth is stressed ,and monetary direct approaches are discussed. Section 3.1 and 3.2 introduce the cost-based approach and “returns on years of schooling” respectively, and discuss literature that implemented these approaches. Section 3.3 emphasizes the advantages of the income-based approach over the other two approaches and proposes the implementation of the lifetime income approach in sustainable wealth measurement. The implementation procedure in either CW or ANS is explained in section 3.4. Although the proposed human capital approach can be implemented in both these sustainability measures, data availability causes this thesis to propose the implementation of the lifetime income method of human capital estimation in ANS. This proposal is the foundation for the subsequent sections of this thesis that apply the improved ANS measure empirically.

The concept of human capital was developed by economists (Schultz, 1961; Mincer, 1974) to explain the ‘puzzle’ of economic growth. The puzzle arises when examining the notion that income is the economic return on wealth. A “normal” rate of return on assets is approximately 5 percent (World Bank, 2011), so that wealth should be 20 times national income. When examining the accounts of the System of National Accounts (SNA), calculations imply unrealistically high returns11_{due to the exclusion of intangible capital, such as human}

capital, in these accounts. Schultz (1961) argues that a large part of this residual is explained by investments in human capital through education, training and investments in health. This has led to many empirical studies trying to prove the positive relationship between human capital and economic growth. Early findings on this relationship were mixed, mainly caused by measurement errors (Boarini et al., 2012). More recent work used improved data on educational attainment resulting in more robust estimates of the impact of human capital on economic growth. For example, Arnold et al. (2007) provide evidence of a strong effect of human capital accumulation on economic growth in a sample of 21 OECD countries over the period between 1971 and 2004. This suggests that better measures of human capital could improve our understanding of the drivers of economic growth, which has significant impact on future policy-making.

The importance of human capital has been stressed by many; however, no consensus exists on how to measure human capital. Similarly to wealth measurements, a broad distinction can be made between several approaches; this distinction is made in figure 3.1. Indicator-based approaches rely on various types of educational characteristics of a country’s population, such as literacy rates, years of schooling and school enrollment rates. Dashboards of indicators rely on a multiple of these educational characteristics, but cannot be aggregated into one overall measure due to the lack of a common metric. The indicator-based approach is therefore neither suitable for comparisons of human capital stocks across nations nor is it possible to implement this human capital measure into economic accounting. Monetary measures of human capital are appropriate for integration into conventional economic accounting, which is key given the role of the SNA. Monetary approaches can be classified into four methodologies. The residual approach was addressed and criticized in section 2.5. The subsequent sub-sections focus on the literature and methodology of the other three direct monetary approaches: the cost-based approach, return on years of schooling and the income-based

approach.

11_{The national balance sheet accounts include both produced assets and the value of commercial natural resources (natural capital)}

(16)

16

3.1 The cost-based approach

The cost-based method is a ‘backward-looking’ or retrospective approach as it focuses on the historical costs of the production of human capital (Le et al., 2003). These costs are past investments undertaken by households, individuals, employers and governments, which include teachers’ salaries, the consumption of fixed capital and household expenditures for school fees and educational material (Boarini et al., 2012). This input-based approach is commonly used, since data on public expenditures is readily available for most countries.

Kendrick (1976) and Eisner (1989) belong to the seminal work in this strand of human capital measures. Kendrick (1976) applied a cost-based approach to the human capital calculations of the United States. Human capital investments were divided into tangible and intangible investments. Child rearing costs were regarded as tangible investments as these are needed to develop the person, and were calculated as all household expenditures for children up to the age of 14. Intangible costs include expenditures on training and education, health and safety, as well as opportunity costs of students attending school. In contrast, Eisner (1989) classified all human capital as intangible and did not include child rearing costs.

By calculating the costs of human capital, a distinction must be made between a household’s consumption and investments, which involves judgment. Kendrick (1976), for example, includes half of household expenditure on health and safety into the human capital measure, while the other half is regarded as consumption. Most household expenditures involve both consumption and investment, so that the cost-based method is sensitive to the arbitrary allocation of consumption and investment of the particular research.

When applying the cost-based approach, the researcher also has to make a decision regarding the depreciation rate, which has significant impact on the estimates of human capital. Depreciation of human capital accounts for the fact that people forget some of what they learn and that some knowledge becomes obsolete (Ederer et al. 2007). Whereas Kendrick (1976) applied the double declining balance method, Eisner (1989) used the straight line rule to depreciate human capital. This researcher’s choice, as well as the often ignored appreciation of human capital, can bias the results.

The arbitrary allocation of consumption and investment, and the choice of the depreciation rate are limitations of the cost-based approach. Thirdly, by focusing on the inputs of production, the efficiency with which the inputs are used to produce an output is ignored. Rather than aiming at these production costs, the value of human capital should be determined by demand and supply of human capital. Le et al. (2003) provide an example that illustrates the problem of the cost-based approach: “an innately less able and less healthy child is more costly to raise, so the cost-based approach will overestimate its human capital while underestimating well-endowed children who, all else equal, should incur fewer rearing and educational expenses”(Le et al., 2003, p. 274). Therefore, estimating the inputs of human capital as a representative of the output is inappropriate. Finally, the time dimension of educational investment is not captured. According to Jorgenson

FIGURE 3.1 HUMAN CAPITAL MEASUREMENT - CLASSIFICATION OF APPROACHES

(17)

17

and Fraumeni (1992), a lengthy gestation period exists between the current expenditures on education and the emergence of the graduates’ human capital.

The cost-based approach is applied by the World Bank (2011) in the ANS in one of its simplest forms. The population’s human capital is valued at the public education expenditures of a country, which is data that is usually readily available. Although this method made it possible for the World Bank (2011) to calculate the ANS for almost all countries for the period 1970-2008, the limitations discussed imply that the values do not reflect the actual human capital stocks present in a country.

3.2 Returns on years of schooling

Mincer (1974) introduced the idea that human capital is based on the returns on years of schooling. Klenow and Rodríguez-Clare (2005) build on the early work of Mincer and state that human capital is dependent on the notion that investments in education earn a market rate of interest for the period of education. This is expressed in the following formula:

( ) (3.1)

where H is human capital; h is human capita per capita; L is the amount of workers; equals the rate of interest per year of education; and s is the average number of years of education of the population.

The approach is not monetary in its original form as suggested by Mincer (1974) and Klenow and Rodríguez-Clare (2005). Arrow et al. (2012), however, adjust the approach in such a way that it becomes monetary and implement this adjusted form to measure human capital in the wealth accounts for the United States, Brazil, China, India and Venezuela for the period 1995-2000. They calculate the shadow price of human capital as the total wage bill of a country divided by the stock of human capital defined by equation 3.1 with an assumed rate of return on human capital of 8.5 percent12_{. The shadow price for one unit of human capital differs for the}

nations in the sample; for the US it is estimated as greater than $100,000, and less than $10,000 for China and India. Due to the fixed return on years of education, the stock of human capital of a nation can increase either when the average level of education increases or if the number of adults increases. Although the authors include the income of the workers by the total wage bill and derived shadow prices, the approach has numerous limitations compared to the income-based approach described below. The calculations are based on the average level of education and do not categorize data for individual groups, which provides a more accurate picture of the human capital stock. Furthermore, the return of education is set at 8.5 percent for each nation; however, differences across nations and from one education level to another may exist.

3.3 The income-based approach

Contrasting the cost-based approach, the income-based approach is ‘forward-looking’ or prospective (Le et al., 2003) in its methodology since it values human capital as the discounted values of all future income streams of all individuals of a country’s population. This valuation of human capital is based on economic theory that argues that labor compensation reflects the service value that human capital provides during a specified period. This method values the earning power of individuals at market prices, which reflects the interaction of human capital supply and demand. Also, no depreciation rate choice and decision on the allocation of consumption and investment have to be made with the income-based approach. The lengthy gestation period that exists between undertaking education and the emergence of human capital is accounted for. The limitations of the indirect and cost-based approach do not apply to this method. Furthermore, the application of the income-based approach is more accurate than the returns on years of schooling as suggested by Arrow et al. (2012). The latter approach is too dependent on the constant rate of return on education and the average years of schooling of the whole population instead of the categorization proposed by the former. The income-based approach is therefore perceived as most reliable (Jorgenson and Fraumeni, 1989, 1992; Le et al., 2003; Liu, 2011; Hamilton and Liu, 2013; Liu, 2013).

12_{The application of is based on calculations of Klenow and Rodríguez-Clare (2005). Arrow et al. (2012) use the data from}

(18)

18

The origin of scientific application of the income-based approach dates back to 1853 when Farr (1853), cited in Le et al. (2003, p. 272), introduced the methodology by calculating the net present value of future earnings of an individual in the United Kingdom, adjusted for the probability of death. The net human capital of a laborer in agriculture was valued at £15013_{using a discount rate of five percent. This approach has been the inspiration}

for many later researchers, who adjusted or extended the model. Graham and Webb (1979) extended the model by including economic growth and education levels, and estimated the value of human capital of the US male population in 1969. Interestingly enough, they compare their results to the cost-based approach applied in Kendrick (1976). The estimation of human capital by the resource-costs is much lower than the calculated present value of future earnings proposed by Graham and Webb (1979). Actually, a discount rate of over 20 percent is required to equate the two approaches. This illustrates the large differences in estimating human capital stock by different approaches; it is argued that estimations based on the cost-based approach applied by Kendrick (1976) are highly undervalued.

Although the income-based approach was widely applied since the 1960s, it was not until the influential work of Jorgenson and Fraumeni (1989, 1992) that significant interest in measuring human capital arose. They strengthened the methodology of Graham and Webb (1979) and simplified the procedure of discounting future income streams to the present. The model is better known as the Jorgenson-Fraumeni (J-F) model or the lifetime income approach, and was first applied to the US population for the period 1948-1984. The model defines human capital as follows:

Definition: An individual’s human capital is valued as the current labor income plus the present value of the

lifetime income in the subsequent period weighted by employment rates and survival probabilities.

Since this seminal research by Jorgenson and Fraumeni (1989, 1992), other researchers applied this definition and provided estimates for other countries than the United States. Ahlroth et al. (1997) applied the J-F model on Swedish data for the years 1967, 1973, 1980 and 1990. Ervik et al. (2003) modified the J-F model by excluding non-market labor activities and by limiting the estimates to higher education sector. They applied this modified method to Norway for the year 1995. Estimates for Australia were provided by Wei (2004, 2008). Whereas Jorgenson and Fraumeni applied the methodology to the whole population, Wei (2008) applied the method to the adult working age population, which they regard everyone between 18 years and 65 years. Using full Australian Census data for the years 1981, 1986, 1991, 1996 and 2001, Wei (2008) concludes that education has become an increasingly important driver for human capital and that the impact of population aging slowed down the growth of human capital stock. Le et al. (2006) also use population Census data from 1981 until 2001 for measuring the human capital for New Zealand. The estimates are restricted to the working age population (18-64), categorized in four educational levels. Furthermore, Gu and Wong (2008), Coremborg (2010), Li et al. (2013) and Jones and Chiripanhura (2010) have provided estimates for Canada, Argentina, China and the United Kingdom respectively. Although these efforts show the interest of many in developing human capital account by applying the lifetime income approach, estimates are currently still restricted to a small number of nations.

The drawback of the lifetime income approach is that data is not as widely available as with the cost-based or indirect approach in measuring human capital. However, its potential was recognized when the OECD set up the Human Capital Project to identify a common methodology and data requirements for building human capital accounts. This project was launched in October 2009 and identified the lifetime income approach, with its roots in the J-F model (1989, 1992), as having significant potential.

The original J-F model was further developed and human capital accounts were developed for sixteen14_OECD

countries under the lead of Liu (2011, 2013). Based on the results of this research as well as on the argument that the income-based approach is consistent with economic theory and the manner in which the tangible

13_{The difference between the estimated average salary of £349 and the average maintenance cost of £199.}

14_{Human capital accounts were created based on data availability for Australia (1997, 1999, 2001), Canada (1997-1999, 2003-2006),}

(19)

19

assets are measured in the SNA (Boarini et al., 2012), the income-based approach is regarded as most

appropriate for constructing human capital accounts. So far, it has not been the concept itself, but the

methodology and data availability that were the main challenge in applying this approach. The results of Liu (2011), however, show the feasibility of implementing this method. The specifics on the methodology are discussed below, based on explanations by Hamilton and Liu (2011) and Liu (2011,2013).

Although the definition of human capital according to the income-based approach is given, this definition was operationalized by previous research (Wei, 2004, 2008; Liu, 2011, 2013) by focusing on the current working population:

Operational definition: The human capital stock owned by categorized individuals is measured as the total

discounted present value of the expected future incomes potentially generated by the current working population.

The definition above highlights two important issues. First, it focuses on categorized individuals. The lifetime income approach takes a broad range of factors concerning the working age population into account (demographics, education, mortality rates and labor market characteristics) and is therefore more realistic than measures based on single indicators. Ideally, the lifetime income of the population is categorized across age, educational level and gender, so that comparisons can be made and trends can be identified across demographic groups. The inclusion of the broad range of factors allows for valuable information for policy interventions. A fundamental assumption related to the demographic characteristics is that an individual of a given age, gender and educational level will have in year t+1 the same labor income15_{as a person who, in year t,}

is one year older but has otherwise the same characteristics (gender and educational level).

Secondly, the approach calculates the present value of future income of the working age population only and excludes the categories (0-14 years) and (65 years and over). The latter category is assumed to be in the retirement stage of life and that their lifetime income is zero, since it is expected that no more income will be retrieved after withdrawal from the labor market. The youth category (0-14) is not accounted for because data on income is not available for this group; this category is not yet active in the labor market. It is recognized, however, that the lifetime income per capita of this group is higher than of the working age population, since the youth category has a longer timespan left in the working age. Not being able to calculate human capital for the whole population, but instead for the working age population, is a limitation resulting in downward biased estimates of human capital stocks.

In addition to these operational assumptions, there are a number of limitations of the lifetime income approach related to (1) the limitation of human capital to capital “in use” and (2) market activities, and (3) the attribution of income to individuals. These practical limitations were identified by Liu (2011, p.13). First, the human capital stocks calculated represent the human capital that is “in use” rather than the human capital that is “available” within a country; withdrawal from the labor market (affecting the employment rate) and earning gaps between males and females affect the human capital stocks of a country. Second, the methodology is limited to market activities. Education could increase household productivity, but this is excluded from the model. Finally, earnings of individuals are attributed to their highest achieved level of education, which ignores the effects of other factors on an individual’s income, such as trade-unions, health or performance. Despite these shortcomings, the lifetime income approach is identified as the most practical and appropriate approach in estimating human capital stocks by the OECD.

In the construction of human capital stocks, two stages for the working age population are identified by Liu (2011): ‘study-and-work’ (15-40 years) and ‘work only’ (41-64 years). The calculations for these two stages differ. The lifetime labor income for people in the ‘work only’ stage is based on age, survival rate, the annual growth rate of income and a discount rate, and calculated as follows:

{( ) ( )} (3.2)

HUMAN CAPITAL IN SUSTAINABLE WEALTH MEASUREMENT