![Sample our Humanities journals, sign in here to start your FREE access for 14 days]({"type":"image" , "src" : "/sda/5939/TOC-Subject-Sample-2021-Humanities.jpg"})
Abstract
This paper analyses the impact of household wealth heterogeneity on inequality and macroeconomic stability in financialized economies. Based on the case of the USA since the 1980s it argues that transformation of financial sector operations has generated inequality by influencing gains from wealth ownership and leverage levels across the income distribution. Securitization and the subprime lending expansion have led to the emergence of a new class of leveraged homeowners, experiencing large increases in wealth prior to the Great Recession, followed by substantial losses after the crisis. Simultaneously, capitalists have diversified their asset portfolios while earning the highest and fastest-growing wages in the economy when employed as financial sector executives. In this light, the paper proposes a new conceptualization of households in macroeconomic models, defined by balance sheet composition rather than income sources alone. To inform this taxonomy, inequality and leverage indicators are simulated in a stock-flow consistent model calibrated to US data with three classes of households distinguished by their wealth composition, and a securitized financial sector. The proposed framework is found to produce more empirically accurate levels of income inequality and greater macroeconomic instability than the two-class division and establishes an equalizing effect of housing for wealth distribution.
Acknowledgments
I am grateful to two anonymous Reviewers, and Gary Dymski, Giuseppe Fontana and Peter Phelps for their guidance in writing this research paper. I would also like to thank Jan Toporowski, Paul Anand, and Alan Shipman for their comments on a draft of this manuscript, as well as Yannis Dafermos, Antoine Godin, Yun Kim, Marc Lavoie, Maria Nikolaidi, Özlem Onaran, Cem Oyvat, Marco Passarella Veronese, Mark Setterfield, Engelbert Stockhammer, and Gennaro Zezza for their feedback on earlier work on which this paper is based. Any errors remain my own.
Notes
1 The U.S. SCF measures household income before transfers and taxes for the calendar year prior to the survey wave, and accounts for wage income, business income, interest and dividend income, realized capital gains, social security and retirement income, and income from social transfers.
2 Note that estimates of income inequality using data from the U.S. SCF tend to be higher than other estimates as the dataset oversamples households at the top of the distribution. For instance, estimates based on data from Annual Social and Economic Supplement to the Current Population Survey yield the Gini index for pre-tax household at 0.431 in 1989, rising to 0.463 in 2007 and 0.476 in 2013 (U.S. Census Bureau Citation2019).
3 Note that the estimates of the Gini index of net wealth in Wolff (Citation2017) exclude the value of vehicles owned by a household on the grounds that the resale value of a vehicle does not accurately reflect its consumption value to the household (Wolff Citation2017, 6).
4 In the U.S. SCF assets are defined as follows: Primary residence is measured as the reported market value. Business equity is measured in net terms. Transaction accounts include call, checking, and saving accounts, money market deposit accounts, and money market mutual funds. Financial investment assets include certificates of deposits, savings bonds, bonds, stocks, other managed assets, pooled investment funds, i.e. non-money market mutual funds, and other financial assets. Retirement and insurance assets include the Individual Retirement Accounts, Keogh accounts, 401(k), and other retirement accounts, as well as the cash value of life insurance plans. Liabilities are measured as follows: Secured debt is measured as the amount outstanding on mortgages and home equity lines of credit secured by primary residence and other property. Unsecured debt is defined as credit card balances and instalment loans (which include vehicle, student, and consumer loans). Other debt is defined as other unsecured lines of credit and other miscellaneous forms of debt (e.g. debt to family members, borrowing against insurance policies or pension accounts, margin debt, etc.).
5 As the objective of Figure 1 is to examine the breakdown of household asset and debt holdings at different points of the income distribution, vehicles are included in this figure as they constitute a substantial portion of the value of net wealth for households in the bottom income quintile.
6 Neoclassical theory states that income inequality is a natural outcome of market processes as it reflects the marginal contributions to production, rewarding those with high or scarce skills (Stiglitz Citation2012, 37). Skill-biased technological change, differences in human capital, and trade openness are seen as the key determinants of inequality (Galbraith Citation2016, 74). This approach has been criticized for ignoring structural and institutional factors generating inequality in modern advanced economies, which are outlined in the previous section.
7 A distinction can be made between capitalists as entrepreneurs, who realize variable profit income dependent on the difference between expected and actual investment, and more passive rentiers, who receive fixed income in the form of unproductive rents based on their ownership of companies and financial institutions (Hayes Citation2006; Toporowski Citation2015). In the context of financial sector transformation and the existence of derivative trading, the capitalist class can be analyzed as including both entrepreneurs and rentiers, who pursue capital returns through investing in financial markets and ownership of financial assets (Toporowski Citation2001).
8 In this paper, we focus exclusively on changes in financial sector operations (primarily the development of structured finance and subprime lending) and their impact on the economy occurring since the 20th century, and avoid the term financialization to avoid its ambiguity (see definition by Epstein Citation2005). This is because the processes of financialization related to the development of credit, money, financial instruments, and interest rates have been argued to take place for as many as 5,000 years (Graeber Citation2011; Sawyer Citation2013). Consequently, financialization is not limited to any particular time or place, can take a variety of forms, and at times may also go in reverse (Sawyer Citation2017). For this reason, the preferred term used in this paper is financial sector transformation, which refers to the processes of financial liberalization and deepening in the USA since the second half of the 20th century. Financial deepening refers to increasing provision of financial services, diversity of financial instruments, and a greater number of financial institutions (Shaw Citation1973).
9 At this stage it is not endogenously explained why households in each group chose to rent or own their house. Households within each group are assumed to behave in the same way, which is described for each group by the model equations below. However, motivations and behaviour of households across these three groups are assumed to differ based on the theory developed in this section. Future research will extend this framework to consider endogenous movements of households between groups and endogenous choice of housing tenure.
10 As shown in , it is assumed that working class households account for 40% of all households in the model, while the middle class and rentier households account for 50% and 10% of the modelled population respectively. This assumption implies that the homeownership rate in the model is 60% (corresponding to the sum of the population size of the middle class and rentiers), which is consistent with the empirically estimated value of 63.7% in 2016 based on the U.S. Survey of Consumer Finances.
11 This assumption only refers to the growth rate of the population size, which is uniform for the entire household sector in aggregate terms. It does not refer to either the share of income/wealth accruing to each group or the number of households above a certain income/wealth threshold. The size of each household group is assumed to expand at the same rate because it represents a fixed proportion of the overall population size. Endogenous movements between group will be the subject of future research.
12 Expansion of the household sector to account for owner-occupiers without a mortgage will be the subject of future research.
13 Unlike in Zezza (Citation2008), rent payments in this model are not directly dependent on the growth of income. This corresponds to the observation that income growth in the USA has lagged behind the growth of rent prices and total housing stock between 2005 and 2016 (Federal Reserve Economic Data St Louis Fed Citation2017).
14 This corresponds to the assumption of the “pecking order” in Setterfield and Kim (Citation2013) stating that households treat savings as a “luxury that is foregone first” in the presence of debt repayments.
15 Exclusion of rent payments from household consumption decisions is consistent with Zezza (Citation2008). If rent payments are excluded from the consumption function (so that working class consumption depends on the entire value of gross income), consumption and loans of this group becomes so large that net wealth is persistently negative.
16 While a weakness of this backward-looking approach to the formation of expectations is the possibility of systematically erroneous predictions if the economic variable is unstable, such adaptive expectations are preferred to the rational expectations hypothesis due to the presence of fundamental uncertainty in the economy.
17 βw is assumed to be high during a boom, as in the early 2000s when lending norms were lax due to the perceived minimization of credit risk through securitization. In times of recessions, βw can be thought of as low as lenders are more concerned about creditworthiness, leading to stricter lending norms.
18 Some studies extend the upper limit to as much as 300% of the median income because the 125% cut-off places a disproportionately large portion of the population in certain countries into the top category (Pressman Citation2007).
19 This resembles the Haig-Simon income specification, where capital gains enter into the disposable income equation (Godley and Lavoie Citation2007, 392).
20 Due to empirical problems, the LTC/PIH framework has seen numerous extensions aiming to improve its explanatory power. These incorporate factors which impede accurate formation of future income expectations, namely liquidity constraints in credit markets (Gross and Souleles Citation2002), precautionary saving (Carroll Citation1997), bequest motives (Cagetti and De Nardi Citation2008), and wealth effects of asset price increases (Mehra Citation2001; Duca, Muellbauer, and Murphy Citation2012). Despite these various extensions of the standard LTC/PIH framework, its basic premise of rational optimizing agents carefully planning their consumption patterns over the lifecycle remains. Another problematic feature of this literature is its assumption that financial innovation and subprime credit expansion should act as a relief to credit-constrained households, allowing for a more optimal distribution of economic resources (Barba and Pivetti Citation2008, 119; Elul Citation1995).
21 Simulation was conducted using R and the code is available upon request.
22 As noted by Dafermos and Papatheodorou (Citation2015) the negative impact of the unemployment rate on the wage share reflects the reserve army effect.
23 The simulated steady-state value of the interest rates on mortgages to the middle class is 6.8%, while the interest rate on loans to the working class is 8.8%.
24 Note that N is the total number of households: N = Nw+Nm+Nr.
25 A different version of that latter scenario was run, where securitization was excluded entirely from the reduced specification without the middle class; the results obtained are similar to the scenario presented here.
26 The simulated value of the income Gini index is more empirically accurate that the simulated value of 0.22 in Dafermos and Papatheodorou (Citation2015).
27 In the subprime period, the value of the income Gini index is estimated at 0.54 in 2003 and 0.57 in 2006 (Wolff Citation2017)
28 The CBO defines market income as income earned from non-governmental sources, including labor, business, capital, and retirement income (CBO Citation2019). Income before taxes and transfers additionally accounts for social insurance benefits.
29 Details of the sensitivity analysis are available upon request.
30 The simulated values of the Gini index for income and wealth inequality are higher in response to: a fall in w0 and w1, a decrease in the profit retention rate sf, a lower marginal propensity to consume out of income for rentiers c5, a higher βw. Income inequality also increases following a reduction in the rate of productivity growth gλ. In addition, the Gini index for wealth rises when: the marginal propensity to consume out of income for the working class and the middle class increases (c1 and c3), the marginal propensity to consume out of wealth for the middle class (c4) increases and falls for rentiers (c6), and when θ0 or π0 increase.
31 Leverage measures increase in response to: an increase in w0 and w1, a higher marginal propensity to consume out of income for rentiers c5, a higher β and βw, and a fall in θ0. In addition, the debt-to-income ratio for the economy increases following: a lower α, an increase in η, a higher rcb, and a fall in π0 or π2. The debt-service-to-income ratio for households additionally rises when α and π0 increase.
32 Detailed results are available on request.
33 This is consistent with Zezza (Citation2008) and Caverzasi and Godin (Citation2015) where an increase in consumption emulation in simulated to raise the aggregate debt-to-income ratio. While in this model we do not observe changes to the results following the shock to the emulation parameter, comparison of results of the full model with the reduced specification without emulation shows that the presence of relative consumption concerns leads to higher macroeconomic volatility measured by the debt-to-income ratio.