Rich and ever richer: Differential returns across socio-economic groups

Ederer, Stefan; Mayerhofer, Maximilian; Rehm, Miriam Working Paper Rich and ever richer: Differential returns across socioeconomic groups ifso working paper, No. 2 Provided in Cooperation with: University of Duisburg-Essen, Institute for Socioeconomics (ifso) Suggested Citation: Ederer, Stefan; Mayerhofer, Maximilian; Rehm, Miriam (2019) : Rich and ever richer: Differential returns across socio-economic groups, ifso working paper, No. 2, University of Duisburg-Essen, Institute for Socio-Economics (ifso), Duisburg This Version is available at: https://hdl.handle.net/10419/215634 Standard-Nutzungsbedingungen: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may be saved and copied for your personal and scholarly purposes. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. https://creativecommons.org/licenses/by/4.0/ uni-due.de/soziooekonomie/wp ifso working paper Jan Behringer Till van Treeck 2019 no.5 The Corporate Sector and the Current Account ifso working paper Stefan Ederer Maximilian Mayerhofer Miriam Rehm 2019 no.2 Rich and Ever Richer: Differential Returns Across Socio-Economic Groups uni-due.de/soziooekonomie/wp 1/28 Rich and Ever Richer: Differential Returns Across Socio-Economic Groups Stefan Ederera, Maximilian Mayerhoferb, and Miriam Rehmc* aAustrian Institute of Economic Research (WIFO) & Department of Socioeconomics, Vienna University of Economics and Business (WU); b Department of Socioeconomics, Vienna University of Economics and Business (WU); c Institute for Socio-Economics, University of Duisburg-Essen, Duisburg, Germany & Department of Socioeconomics, Vienna University of Economics and Business (WU), Vienna, Austria; Lotharstraße 65, 47057 Duisburg, [email protected] June 2019 2/28 Rich and Ever Richer: Differential Returns Across Socio-Economic Groups This paper estimates rates of return across the gross wealth distribution in eight European countries. Like differential saving rates, differential rates of return matter for Post Keynesian theory, because they impact the income and wealth distribution and add an explosive element to growth models. We show that differential rates of return matter empirically by merging data on household balance sheets with longrun returns for individual asset categories. We find that (1) the composition of wealth differentiates between three socioeconomic groups: 30% are asset-poor, 65% are middle-class home owners, and the top 5% are business-owning capitalists; (2) rates of return rise across all groups; and (3) rates of return broadly follow a log-shaped function across the distribution, where inequality in the lower half of the distribution is higher than in the upper half. If socioeconomic groups are collapsed into the bottom 95% workers and top 5% capitalists, then rates of return are 5.6% for the former and 7.2% for the latter. Keywords: rate of return; differential; wealth; distribution Subject classification codes: D31 (Personal income, wealth, and their distributions), D33 (Factor income distributions), E43 (Interest Rates: Determination, Term Structure, and Effects), E12 (General aggregative models: Post-Keynesian), E21 (Consumption, saving, wealth) 3/28 Introduction The distribution of wealth plays an important role in determining economic capabilities and in shaping the position of individuals in the socioeconomic structure. Wealth is not only distributed much more unequally than income, but it is also linked to political power. For all these reasons, Post Keynesian economics has a long tradition of investigating the theory of wealth inequality (Pasinetti 1962, Dutt 1990, Palley 2012). Recently, advances in data availability have garnered renewed interest in empirical wealth research (Piketty 2014; Saez and Zucman 2016; Alvaredo et al. 2017). In conceptualizing inequality, the Post Keynesian literature traditionally differentiates between classes. Workers differ systematically from capitalists in their insertion into the economy, i.e. their options and outcomes (Wright 1997), which is reflected in their respective income sources. Recent Post Keynesian thought has thus paid detailed attention to the functional distribution of income (Bhaduri and Marglin 1990; Barbosa and Taylor 2006; Stockhammer and Ederer 2008; Stockhammer, Onaran, Ederer 2009). This has shifted the focus somewhat from classes to the source of income. For understanding wealth inequality, however, classes or socioeconomic groups are important (Rehm and Schnetzer 2015; Rehm, Naqvi, Hofmann 2016; Ederer and Rehm 2019): The ownership of businesses conveys very different economic capabilities than, say, the ownership of one’s main residence. Since returns differ between wealth categories, such differences in ownership translate directly into differential returns on wealth for different socioeconomic groups. The Post Keynesian literature has long provided theoretical arguments for differential rates of return (Kahn 1959, Laing 1969, Harcourt 1972, Pasinetti 1974, 1983; Gupta 1977) analogous to differential saving rates. This debate revolved around the question whether interest rates (received by workers) are lower than profit rates (received 4/28 by capitalists) and already raised several key points around differential returns: First, differential rates of return impact the income distribution if there are mixed income sources, because they imply workers’ income from profits on their invested capital is lower than capitalists’. Second, differential rates of return may lead to positive (i.e. explosive) feedback effects in the distribution of wealth. Third, this then has implications for the stability of a growth regime. The issue of differential rates of return is thus highly relevant for the Post Keynesian growth literature. Crucially for this paper, new and improved data sources now make it possible to investigate the question of differential rates of return empirically. Some recent empirical work (Bach, Calvet, and Sodini 2018; Fagereng et al. 2018) has done so for individual Nordic countries, often with a limited theoretical backdrop. The aim of this paper is thus to situate differential returns in a modern Post Keynesian debate, and to provide an empirical estimate for differential rates of return for socioeconomic groups in several European countries. This paper estimates rates of return across the distribution of gross wealth in European countries by combining two (relatively) novel data sets. We apply rates of return compiled by Jordà et al. (2018) to the Household Finance and Consumption Survey (ECB 2014) and show that (1) the composition of wealth differentiates between three socioeconomic groups: the asset-poor, middle-class home owners, and capitalists; (2) rates of return rise over those groups in all countries; and (3) inequality is higher between the middle and bottom than between the top and middle of the wealth distribution. The structure is as follows: Section 2 reviews the literature and section 3 describes the data sources. Section 4 contains our results, and section 5 concludes. 5/28 Literature Review The Post Keynesian literature has long recognized that socioeconomic groups or classes (i.e., workers and capitalists) differ in systematic ways. One key aspect here are differential saving rates, i.e. that capitalists on the whole save a larger fraction of their income than workers. This is a well-established concept in Post Keynesian theory that was introduced by Keynes (1936) and formalized by Kalecki (1937), who posits a bipolar distribution of functional income – where workers exclusively earn wages and rentiers receive only profits – and in addition assumes that workers do not save. With these restrictive assumptions, the functional and the personal income distribution coincide. Kaldor (1955) relaxes the assumption that workers do not save in his well-known savings function, and thus moves towards a differentiation by income source. However, wageearners still have a lower saving rate than profit-earners, because firms retain some profits to reinvest. Pasinetti (1962), in contrast, deals with classes (workers and capitalists) and assumes different saving rates for each class regardless of their income sources. Furthermore, he points out that workers who save and thus accumulate wealth must receive a part of profits as interest, even though he assumes a single rate of return for both workers and capitalists (Pasinetti 1962). 1 In modern Post Keynesian models, a positive saving rate for workers is by now a standard assumption (e.g. Bhaduri and Marglin 1990). In the recent literature, 1 . He states that "in a long-run equilibrium model, the obvious hypothesis is that of a rate of interest equal to the rate of profits" (Pasinetti, 1962, pp. 271-272). Samuelson and Modigliani (1966) point out that while it is necessary to make this assumption explicit in Pasinetti's framework, it is not relevant in Kaldor's framework, given that the latter distinguishes between income groups. 6/28 a number of works have built on Pasinetti in incorporating an endogenous wealth distribution in Post Keynesian frameworks (Palley 2012, 2017; Taylor et al. 2015; Ederer and Rehm 2019), in which the wealth share of workers and capitalists depend inter alia on their saving rates. If there are differential rates of return, then the question whether to focus on the functional or personal income distribution is important, because profit incomes (which are determined by the functional income distribution) differ by classes (i.e., the personal income distribution). That is, differential returns directly impact the income distribution. A rich Post Keynesian debate has examined the implications for Post Keynesian theory if the interest rate received by workers is lower than the rate of profit received by capitalists. First suggested by Kahn (1959) and picked up by Laing (1969), early contributions to the literature on differential rates of return were among others Balestra and Baranzini (1971), Harcourt (1972), Maneschi (1974), Moore (1974), Gupta (1977), Fazi and Salvadori (1981), and especially Pasinetti’s seminal work (1974, 1983). 2 This literature revolved around the recognition that an interest rate lower than (capitalists’ or the average) profit rate lends further support to a Pasinetti-like stable wealth distribution. 3 Differential returns however add an explosive element to the dynamic of wealth inequality, and thus have implications for the stability of growth. They raise the likelihood of landing in an 'anti-dual equilibrium' (Darity 1981), in which capitalists own all the wealth. 2 . For a comprehensive overview of this theoretical literature, see Baranzini and Mirante (2013). 3 . As opposed to the 'dual equilibrium' by Samuelson and Modigliani (1966), in which capitalists cease to exist and are replaced by workers, who own all wealth. 13/28 Figure 1 shows the relative distribution 7 of households’ asset categories across unconditional gross wealth vingtiles. The composition changes from ‘other’ wealth being the dominant asset category at the bottom of the wealth distribution, to housing in the (upper) middle, and to business wealth increasing in importance at the top of the wealth distribution. On (unweighted) average over all countries, housing supersedes the no-yield asset category ‘Other’ in the 6th vingtile in making up more than 50% of households’ balance sheets. At the other end of the unconditional wealth distribution, business wealth gains importance above 25% of households’ assets only in the top vingtile, i.e. the top 5% of households by wealth. These indicators can thus be used to divide the population into three groups with corresponding ownership patterns of wealth categories: the first, asset-poor group comprises on average the bottom 30% of the population, the second, home-owners, the middle 65%, and the third, business owners, the top 5%. We thus find substantial differences in the composition of assets between socio-economic groups. Combining these asset compositions with the rates of return of JKKST, we find that the rates of return rise across the wealth distribution, as Figure 2 shows. They are near zero for the bottom 5% in all countries except Finland (where they start above 1%) and remain well below 2% for the bottom 20% in all countries except Portugal. Rates of return then rise steeply and plateau for the upper half of the wealth distribution until the top 5% in most cases. The resulting pattern is roughly one of a log function for the bottom 95% of the wealth distribution with an inflection point between the 2nd and 8th gross wealth vingtile. Only Germany shows a more linear gradient in the bottom half of its 7 . For the distribution of absolute values of household asset categories across unconditional gross wealth vingtiles, see Figure 1A in the Appendix. 14/28 wealth distribution. Rates of return then rise steeply for the top 5% in most countries; the exceptions are the Netherlands and Portugal. Figure 1. Asset composition by gross household wealth vingtiles Source: ECB 2014 (data); own elaboration 15/28 Figure 2. Rates of return by country across gross wealth vingtiles Note: Countries are sorted by their rate of return level in the 20th vingtile. Source: ECB 2014, Jordà et al. 2018 (data); own elaboration Whereas the pattern thus shows some striking similarities across countries, levels differ. The “plateau” lies between roughly 3.5% annual real rate of return in Germany and about 8.5% in Finland at the other extreme. Real rates of return stabilize between 4 and 4.5% in Italy and Spain, and between 5.5% and 6.5% in Belgium, France, the Netherlands, and Portugal. 16/28 Table 4. Average annual real rates of return for three household groups Country Asset-Poor Homeowning Middle Class Workers (bottom 95 %) Businessowning Top 5% (1) (2) (1+2) (3) Belgium 2.6 6.6 6.5 8.3 Finland 3.4 8.7 8.6 10.8 France 2.6 5.7 5.6 7.6 Germany 2.1 3.7 3.6 5.9 Italy 2.4 4.5 4.4 5.7 Netherlands 3.2 5.7 5.5 6.1 Portugal 2.0 6.1 6.1 6.9 Spain 0.9 4.5 4.5 6.5 Average 2.4 5.7 5.6 7.2 Note: ‘Asset-Poor’ in column (1) captures all gross wealth household vingtiles at the bottom of the gross wealth distribution owning on average less than 50% of their gross wealth in their main residence (30% of all households). ‘Home-owning Middle Class’ in column (2) covers households with more than 50% of their gross wealth invested in their main residence, and less than 25% of their gross wealth in business wealth, on average (65% of all households). ‘Workers’ are column (1) and (2) combined and make up the bottom 95% of all households. ‘Business-owning Top 5%’ in column (3) are households in the top vingtile, owning on average more than 25% of their gross wealth in business wealth (except in the Netherlands). The average in the last row is an unweighted mean across countries. Source: ECB 2014, Jordà et al. 2018 (data); own elaboration 17/28 Table 4 combines our findings from the two data sources into our main result. It shows unweighted average real annual rates of return for the asset-poor, the home-owning middle class, and the business-owning top 5%. These amount to 2.4%, 5.7%, and 7.2%, respectively. The top 5% thus earn returns that are more than 2.5 times higher than those at the bottom. Over 50 years on an initial investment of 10,000 Euro, this amounts to a difference of about 290.000 Euro. Since the Post Keynesian literature typically refers to just two groups, workers and capitalists, collapsing the former two into a (very broad) workers’ group yields an average rate of return of 5.6% for workers and 7.2% for capitalists. Table 5. Differences in rates of return between selected vingtiles of the gross wealth distribution Country v20–v10 v10–v1 v20-v1 v19–v2 Belgium 1.7 6.4 8.1 7.1 Finland 2.3 7.2 9.5 8.6 France 2.3 5.3 7.6 5.9 Germany 3.1 2.8 5.9 3.7 Italy 1.5 4.1 5.6 4.8 Netherlands 0.1 5.7 5.9 5.7 Portugal 0.8 6.1 6.9 6.1 Spain 2.2 4.2 6.4 3.9 Average 1.8 5.2 7.0 5.7 Note: This table shows absolute differences between average rates of return of gross wealth vingtiles. The average in the last row is an unweighted mean across countries. Source: ECB 2014, Jordà et al. 2018 (data); own elaboration Finally, Table 5 presents selected distributional indicators in order to take another look at the question where in the distribution inequality in returns stems from. Taking the difference between the rate of return of the top vingtile and the median (v20-v10) yields 18/28 0.1 in the Netherlands and 0.8 in Portugal with their broadly stable rates of return in the top half of the wealth distribution. For the other countries, the difference lies between 1.5 (in Italy) and 2.3 (in Finland and France). We therefore find somewhat limited differences in returns between the middle and the top of the distribution. In contrast, inequality in rates of return is much starker in the bottom half of the distribution. The difference in rates of return received by the median and the bottom vingtile is between 2.8 (in Germany with its almost linearly rising rates of return) and 7.2 (in Finland). That is, on (unweighted) average, European households in the top 5% of the wealth distribution receive an annual real rate of return that is almost 7 percentage points higher than for households in the bottom 30%, whereas the difference between the median and the top is roughly 1.75 percentage points. Table 4 also shows that because of the upward tick of rates of return in the top 5%, using deciles underestimates the differential nature of rates of return in all cases except the Netherlands. 8 To sum up, we find that asset composition differs substantially across three socioeconomic groups, the asset-poor, the home-owning middle class, and the businessowning top 5%. These groups are faced with differentials in rates of return that are economically significant, rising from 2.4% to 5.7% and 7.2%, respectively. Roughly categorized in Post Keynesian terms as ‘workers’ and ‘capitalists’, the differentials in rates of return amount to 5.6% and 7.2%. Finally, across the gross wealth distribution rates of return can be approximated by a broadly log-shaped function, with the inflection point lying between the 2nd and 8th vingtile. That is, inequality between the bottom and 8 . The uptick in the 10th decile in the Netherlands is due to one household with very high business assets. However, we chose not to deleted single observations from our data. 19/28 the middle in rates of return is much larger than that between the middle and the top of the gross wealth distribution. Our results are in line with the (limited) existing literature, not just regarding patterns but also with regard to magnitudes. As discussed, Bach, Calvet and Sodini (2018) estimate a differential of 1.99 percentage points between the median and the top 5% of gross wealth, which is almost exactly equal to our result of roughly 2 percentage points. Fagereng et al. (2018, Figure 4A) find a roughly log-shaped function of differential returns from the 20th percentile; their data is noisy below that level since they use net wealth. For this reason, their average rates of return are markedly negative at the bottom of the distribution, but the differential between their rates of return between the median and the top 5% amounts to roughly 1.4 percentage points, which is not very far from our 1.75 percentage points. It should, however, be noted that our findings, while compatible with the existing literature, need to be interpreted cautiously. First, it is highly likely that we underestimate the differentials in rates of return due to the notorious under-coverage of high-wealth households (which can only be partially redressed by the oversampling of the HFCS), combined with the highly right-skewed distribution of wealth documented in Figure 1A in the Appendix. Second, our data is only able to capture one of several potential channels for differential rates of return, namely household balance sheet composition across the distribution. Taking other channels into account 9 is likely to lead to results finding higher differentials in rates of return. Third, our data does not permit us to capture capital gains, which likely play an important and differential role in wealth accumulation across the 9 . Or, ideally, using observed rates of return, which our data quality unfortunately does not permit. 20/28 wealth distribution (Taylor et al. 2015). Wealthy households are likely to benefit more from higher capital gains, which again renders our estimates conservative. Conclusion This paper estimates rates of return across the distribution of gross wealth in European countries. We find that that three socioeconomic groups are marked by differential rates of return analogous to differential saving rates. This underscores the theoretical considerations of the Post Keynesian literature since the 1950’s on differential rates of return. Concretely, we merge annual real rates of return compiled by Jordà et al. (2018) with the detailed asset categories at the household level provided in the Household Finance and Consumption Survey (ECB 2014) for eight European countries. That is, we apply the rates of return for low-yield assets (deposit and saving accounts, vehicles etc.), housing (i.e., the primary residence), bonds (including voluntary pension plans), and equity (business wealth both in stocks and in direct stakes) to the wealth held by households at the micro level. Our findings are three-fold: First, the composition of wealth differs substantially between three socioeconomic groups. These are the asset-poor, who own mostly lowyield assets and comprise roughly 30% of the population on average; middle-class home owners, who make up most of the upper half of the population and whose household balance sheet is dominated by the main residence; and capitalists, who own economically significant amounts of business wealth and make up roughly the top 5% of the population. Second, we find differential rates of return across the wealth distribution. Rates of return rise largely monotonously over the entire population, and there are clear differences in average rates of return for the three socioeconomic groups. The asset-poor 21/28 receive a rate of return of about 2.4% on their wealth, the home owning middle and upper class receives roughly 5.7%, and capitalists about 7.2%. If these are collapsed into the standard groups of workers and capitalists, then this averages out to about 5.6% for the former and 7.2% for the latter. The magnitude of our findings is in line with the (limited) existing literature on differential rates of return. Third, rates of return across the wealth distribution roughly approximate log functions for the bottom 95% in most countries. That is, they rise steeply to a broad “plateau”, and then only tick upwards again at the very top of the wealth distribution. This implies that inequality is substantially higher between the middle and bottom than between the top and middle of the wealth distribution. Naturally, many important research questions remain unanswered. First and foremost, higher-quality data from administrative sources might yield additional insights into the distribution of rates of return, especially at the very top of the distribution. Second, capital gains play an important role in wealth accumulation and merit closer attention since it is likely that they are unevenly distributed, but our data does not permit us to study their dynamics. Third, it would be very interesting to investigate other potential channels for differential returns, such as differences in innate ability, in professionalism in investment management, in access to insider information, or in political clout. 22/28 References Bach, L., L. E. Calvet, and P. Sodini. 2018. “Rich Pickings? Risk, Return, and Skill in Household Wealth.” HEC Paris Research Paper FIN-2016-1126. http://dx.doi.org/10.2139/ssrn.2706207 Balestra, P., and M. Baranzini. 1971. “Some Optimal Aspects in a Two Class Growth Model with a Differentiated Interest Rate.” Kyklos 24 (2): 240–256. Baranzini M., and A. Mirante. 2013. “The Cambridge Post-Keynesian School of Income and Wealth Distribution.” In: Harcourt, G. C. and P. Kriesler, The Oxford Handbook of Post-Keynesian Economics: Volume 1 Theory and Origins. New York: Oxford University Press. Brainard, W. C., and J. Tobin. 1968. “Pitfalls in Financial Model Building.” The American Economic Review 58 (2): 99-122. Darity, W. A. 1981. “The Simple Analytics of Neo-Ricardian Growth and Distribution.” The American Economic Review 71 (5): 978–993. Dutt, A. K. 1984. “Stagnation, Income Distribution, and Monopoly Power.” Cambridge Journal of Economics 8 (1): 25–40. ECB. 2014. Household Finance and Consumption Survey, second wave. European Central Bank Ederer, S., and M. Rehm. 2019. “Will Wealth Become More Concentrated in Europe? Evidence from a Calibrated Neo-Kaleckian Model.” Cambridge Journal of Economics (forthcoming). Ederer, S., and M. Rehm. 2018. “Making Sense of Piketty’s "Fundamental Laws" in a Postkeynesian Framework.” INEQ Working Paper Series 10. Vienna University of Economics and Business. Erikson, R., and J. Goldthorpe. 1992. The Constant Flux: A Study of Class Mobility in Industrial Societies. New York: Clarendon Press. Fagereng, A., L. Guiso, D. Malacrino, and L. Pistaferri. 2018. “Heterogeneity and uni-due.de/soziooekonomie/wp Institute for Socio-Economics University of Duisburg-Essen Lotharstr. 65 47057 Duisburg Germany uni-due.de/soziooekonomie [email protected] ifso working paper ifso working papers are preliminary scholarly papers emerging from research at and around the Institute for Socio-Economics at the University of Duisburg-Essen. All ifso working papers at uni-due.de/soziooekonomie/wp ISSN 2699-7207 This work is licensed under a Creative Commons Attribution 4.0 International License Institute for Socio-Economics University of Duisburg-Essen Lotharstr. 65 47057 Duisburg Germany uni-due.de/soziooekonomie [email protected] ifso working paper ifso working papers are preliminary scholarly papers emerging from research at and around the Institute for Socio-Economics at the University of Duisburg-Essen. All ifso working papers at uni-due.de/soziooekonomie/wp ISSN 2699-7207 This work is licensed under a Creative Commons Attribution 4.0 International License