Bulletin Board Pix

The Asset Price Meltdown and the Wealth of the Middle Class

The Asset Price Meltdown and the Wealth of the Middle Class

Edward N. Wolff

New York University

November 2013

Abstract: I find that median wealth plummeted over the years 2007 to 2010, and by 2010 was at its lowest level since 1969. The inequality of net worth, after almost two decades of little movement, was up sharply from 2007 to 2010. Relative indebtedness continued to expand from 2007 to 2010, particularly for the middle class, though the proximate causes were declining net worth and income rather than an increase in absolute indebtedness. In fact, the average debt of the middle class actually fell in real terms by 25 percent. The sharp fall in median wealth and the rise in inequality in the late 2000s are traceable to the high leverage of middle class families in 2007 and the high share of homes in their portfolio. The racial and ethnic disparity in wealth holdings, after remaining more or less stable from 1983 to 2007, widened considerably between 2007 and 2010. Hispanics, in particular, got hammered by the Great Recession in terms of net worth and net equity in their homes. Households under age 45 also got pummeled by the Great Recession, as their relative and absolute wealth declined sharply from 2007 to 2010.

1. Introduction

The last two decades have witnessed some remarkable events. Perhaps, most notable is the housing value cycle which first led to an explosion in home prices and then a collapse, affecting net worth and helping to precipitate the Great Recession. The housing bubble, in turn, was based on questionable mortgage practices and then speculative over-building.

The median house price remained virtually the same in 2001 as in 1989 in real terms.1 However, the home ownership rate shot up from 62.8 percent in 1989 to 67.7 percent in 2001 according to data from the Survey of Consumer Finances (SCF). Then, 2001 saw a recession (albeit a short one). Despite this, house prices suddenly took off. The median sales price of existing one-family homes rose by 17.9 percent in real terms nationwide. However, from 2004 to 2007 housing prices slowed, with the median sales price of existing one-family nationwide advancing only 1.7 percent over these years in real terms. Over the years 2001 to 2007 real housing prices gained 18.8 percent. The home ownership rate continued to expand, though at a somewhat slower rate, from 67.7 to 68.6 percent.

Then, the Great Recession and the associated financial crisis hit at the end of 2007 and asset prices plummeted. From 2007 to 2010, in particular, the median price of existing homes nose-dived by 21 percent in nominal terms and 24 percent in real terms.2 Moreover, for the first time in 30 years, the share of households owning their own home fell, from 68.6 to 67.2 percent.

The housing price bubble was fueled in large part by a generous expansion of credit available for home purchases and re-financing. This took a number of forms. First, many home owners refinanced their primary mortgage. However, because of the rise in housing prices, these home owners increased the outstanding mortgage principal and thereby extracted equity from their homes. Second, many home owners took out second mortgages and home equity loans or increased the outstanding balances on these instruments. Third, among new home owners, credit requirements were softened, and so-called “no-doc” loans were issued requiring none or little in the way of income documentation. Many of these loans, in turn, were so-called “sub-prime” mortgages, characterized by excessively high interest rates and “balloon payments” at the expiration of the loan (that is, a non-zero amount due when the term of the loan was up). All told, average mortgage debt per household expanded by 59 percent in real terms between 2001 and 2007 according to the SCF data, and outstanding mortgage loans as a share of house value rose from 0.334 to 0.349, despite the 19 percent gain in real housing prices (see Table 4 below).

In contrast to the housing market, the stock market boomed during the 1990s. On the basis of the Standard & Poor (S&P) 500 index, stock prices surged 171 percent between 1989 and 2001.3 Stock ownership spread and by 2001 over half of U.S. households owned stock either directly or indirectly. However, the stock market peaked in 2000 and dropped steeply from 2000 to 2003, recovered somewhat in 2004, and then rebounded from 2004 to 2007. Over the period from 2001 to 2007, the S&P 500 was up 6 percent in real terms. However, the share of households who owned stock either directly or indirectly fell somewhat to 49 percent from 52 percent in 2001. Then came the Great Recession. Stock prices, based on the S&P 500 index, crashed from 2007 to 2009 and then partially recovered in 2010 for a net decline of 26 percent in real terms. The stock ownership rate also once again declined, to 47 percent.

Real wages, after stagnating for many years, finally grew in the late 1990s. According to BLS figures, real mean hourly earnings gained 8.3 percent between 1995 and 2001. From 1989 to 2001, real wages rose by 4.9 percent (in total), and median household income in constant dollars inched up by 2.3 percent. 2001.4 Employment also surged over these years, growing by 16.7 percent.5 The (civilian) unemployment rate remained relatively low over these years, at 5.3 percent in 1989, 4.7 percent in 2001, with a low point of 4.0 percent in 2000, and averaging 5.5 percent over these years.6 Real wages then rose very slowly from 2001 to 2007, with the BLS real mean hourly earnings up by only 2.6 percent, while median household income gained only 1.6 percent. Employment also grew more slowly over these years, gaining 6.7 percent. The unemployment rate remained low again, at 4.7 percent in 2001 and 4.6 percent in 2007 and an average value of 5.2 percent.

Real wages, on the other hand, picked up from 2007 to 2010, with the BLS real mean hourly earnings increasing by 3.6 percent. In contrast, median household income in real terms declined sharply over this period, by 6.4 percent . Moreover, employment contracted over these years, by 4.8 percent, and the unemployment rate surged from 4.6 percent in 2007 to 10.5 percent in 2010, though it did come down a bit to 8.9 percent in 2011.

There was also an explosion of consumer debt leading up to the Great Recession. Between 1989 and 2001, total consumer credit outstanding in 2007 dollars surged by 70 percent and then from 2001 to 2007 it rose by another 17 percent.7 There were a number of factors responsible for this. First credit cards became more generally available for consumers. Second, credit standards were relaxed considerably, making more households eligible for credit cards. Third, credit limits were generously increased by banks hoping to make profits out of increased fees from late payments and from higher interest rates.

Another source of new household indebtedness was from a huge increase in student loans. According to the SCF data, the share of households reporting an educational loan rose from 13.4 percent in 2004 to 15.2 percent in 2007 and then surged to 19.1 percent in 2010.8 The mean value of educational loans in 2010 dollars among loan holders only increased by 17 percent from $19,410 in 2004 to $22,367 in 2007 and then by another 14 percent to $25,865 in 2010. The median value of such loans first went up by 19 percent from $10,620 in 2007 to $12,620 in 2007 and then by another 3 percent to $13,000 in 2010. These loans were heavily concentrated among younger households and, as we shall see below, was one of the factors (though not the principal one) which led to a precipitous decline in their net worth between 2007 to 2010.

Another important change over the last three decades affecting household wealth was a major overhaul of the private pension system in the United States. As documented in Wolff (2011b), in 1989, 46 percent of all households reported holding a defined benefit (DB) pension plan. DB plans are traditional pensions, such as provided by many large corporations, the federal government, and state and local governments, which guarantee a steady flow of income upon retirement. By 2007, that figure was down to 34 percent. The decline was more pronounced among younger households, under the age of 46, from 38 to 23 percent, as well as among middle-aged households, ages 47 to 64, from 57 to 39 percent.

Many of these plans were replaced by so-called defined contribution (DC) pension accounts, most notably 401(k) plans and Individual Retirement accounts (IRAs). These plans allow household to accumulate savings for retirement purposes directly. The share of all household with a DC plan skyrocketed from 24 percent in 1989 to 53 percent in 2007. Among younger households, the share rose from 31 to 50 percent, and among middle-aged households it went from 28 to 64 percent.

This transformation is even more notable in terms of actual dollar values. While the average value of DB pension wealth among all households crept up by 8 percent from $56,500 in 1989 to $61,200 in 2007, the average value of DC plans shot up more than 7-fold from $10,600 to $76,800 (all figures are in 2007 dollars).9 Among younger households, average DB wealth actually fell in absolute terms, while DC wealth rose by a factor of 3.3. Among middle-aged households, the value of DB pensions also fell in absolute terms while the value of DC plans mushroomed by a factor of 6.5.

These changes are important for understanding trends in household wealth because DB pension wealth is not included in the measure of marketable household wealth whereas DC wealth is included (see Section 3 below). Thus, the substitution of DC wealth for DB wealth is likely to lead to an overstatement in the “true” gains in household wealth, since the displacement in DB wealth is not captured (see Wolff, 2011b, for more discussion).

The other big story was household debt, particularly that of the middle class, which skyrocketed during these years, as we shall see below. Despite the recession, the relative indebtedness of American families continued to rise from 2007 to 2010.

What have all these major transformations wrought in terms of the distribution of household wealth, particularly over the Great Recession? How have these changes impacted different demographic groups, particularly as defined by race, ethnicity, and age? This is the subject of the remainder of the paper.

2. Plan of the Paper

The paper focuses mainly on how the middle class fared in terms of wealth over the years 2007 to 2010 during one of the sharpest declines in stock and real estate prices. As discussed below, the debt of the middle class exploded from 1983 to 2007, already creating a very fragile middle class in the United States. The main interest here is whether their position deteriorated even more over the Great Recession. The paper also investigates trends in wealth inequality, changes in the racial wealth gap and wealth differences by age, and trends in homeownership rates, stock ownership, and mortgage debt. The period covered spans the years from 1962 to 2010. The choice of years is dictated by the availability of survey data on household wealth. By 2010, we are able to see what the fall-out was from the financial crisis and associated recession and, in particular, which groups were impacted the most.

There are six specific issues addressed in the paper. (1) Did the inequality of household wealth rise over time, particularly during the Great Recession? (2) Did median household wealth continue to advance over time or did it fall? (3) Did the debt of the middle class increase over time, especially over the Great Recession? (4) What are the time trends in home ownership and home equity? (5) What happened to stock ownership? (6) How did time trends in average wealth, household debt, the home ownership rate, home equity, and stock ownership vary among different racial and ethnic groups and by age group?

The paper is organized as follows. The next section, Section 3, discusses the measurement of household wealth and describes the data sources used for this study. Section 4 presents results on time trends in median and average wealth holdings, Section 5 on changes in the concentration of household wealth, and Section 6 on the composition of household wealth. In Section 7, I provide an analysis of the effects of leverage on wealth movements over time, particularly over the Great Recession. Section 8 investigates changes in wealth holdings by race and ethnicity; and Section 9 reports on changes in the age-wealth profile. A summary and concluding remarks are provided in Section 10.

Previous work of mine (see Wolff, 1994, 1998, 2002a, and 2011a), using SCF data from 1983 to 2007, presented evidence of sharply increasing household wealth inequality between 1983 and 1989 followed by little change between 1989 and 2007. Both mean and median wealth holdings climbed briskly during the 1983-1989 period as well as from 1989 to 2007. However, most of the wealth gains from 1983 to 2007 were concentrated among the richest 20 percent of households. Moreover, despite the buoyant economy over the 1990s and 2000s, overall indebtedness continued to rise among American families.

The ratio of mean wealth between African-American and white families was very low in 1983, at 0.19 and about the same in 2007. In 1983, the richest households were those headed by persons between 45 and 74 years of age, and the relative wealth holdings of both younger and older families fell between 1983 and 2007, particularly those of the former.

In this study, I look at wealth trends from 1962 to 2010. The most telling finding is that median wealth plummeted over the years 2007 to 2010, and by 2010 was at its lowest level since 1969. The inequality of net worth, after almost two decades of little movement, was up sharply between 2007 and 2010. Relative indebtedness continued to expand during the late 2000s, particularly for the middle class, though the proximate causes were declining net worth and income rather than an increase in absolute indebtedness. In fact, the average debt of the middle class in real terms was down by 25 percent. The sharp fall in median net worth and the rise in its inequality in the late 2000s are traceable to the high leverage of middle class families in 2007 and the high share of homes in their portfolio. The racial and ethnic disparity in wealth holdings widened considerably in the years between 2007 and 2010. Hispanics, in particular, got hammered by the Great Recession in terms of net worth and net equity in their homes. Finally, young households (under age 45) also got pummeled by the Great Recession, as their relative and absolute wealth declined sharply from 2007 to 2010.

3. Data sources and methods

The primary data sources used for this study are the 1983, 1989, 1992, 1995, 1998, 2001, 2004, 2007, and 2010 SCF conducted by the Federal Reserve Board. Each survey consists of a core representative sample combined with a high-income supplement. The high income supplement was selected as a list sample derived from tax data from the IRS Statistics of Income. This second sample was designed to disproportionately select families that were likely to be relatively wealthy (see, for example, Kennickell, 2001, for a diiscussion of the design of the list sample in the 2001 SCF). The advantage of the high-income supplement is that it provides a much "richer" sample of high income and therefore potentially very wealthy families. Typically, about two thirds of the cases come from the representative sample and one third from the high-income supplement. In the 2007 SCF the standard multi-stage area-probability sample contributed 2,915 cases while the high-income supplement contributed another 1,507 cases.

The principal wealth concept used here is marketable wealth (or net worth), which is defined as the current value of all marketable or fungible assets less the current value of debts. Net worth is thus the difference in value between total assets and total liabilities or debt. Total assets are defined as the sum of: (1) owner-occupied housing; (2) other real estate; (3) demand deposits; (4) time and savings deposits, certificates of deposit, and money market accounts; (5) government, corporate, and foreign bonds and other financial securities; (6) the cash surrender value of life insurance plans; (7) the cash surrender value of pension plans, including IRAs, Keogh, and 401(k) plans; (8) corporate stock and mutual funds; (9) net equity in unincorporated businesses; and (10) equity in trust funds. Total liabilities are the sum of: (1) mortgage debt, (2) consumer debt, including auto loans, and (3) other debt such as educational loans.

This measure reflects wealth as a store of value and therefore a source of potential consumption. I believe that this is the concept that best reflects the level of well-being associated with a family's holdings. Thus, only assets that can be readily converted to cash (that is, "fungible" ones) are included. As a result, consumer durables such as automobiles, televisions, and furniture, are excluded here, since these items are not easily marketed, with the possible exception of vehicles, or their resale value typically far understates the value of their consumption services to the household. Another justification for their exclusion is that this treatment is consistent with the national accounts, where purchase of vehicles is counted as expenditures, not savings.10 As a result, my estimates of household wealth will differ from those provided by the Federal Reserve Board, which includes the value of vehicles in their standard definition of household wealth (see, for example, Kennickell and Woodburn, 1999).

Also excluded is the value of future Social Security benefits the family may receive upon retirement (usually referred to as "Social Security wealth"), as well as the value of retirement benefits from private pension plans ("pension wealth"). Even though these funds are a source of future income to families, they are not in their direct control and cannot be marketed.11

Two other data sources are used in the study. The first of these is the 1962 Survey of Financial Characteristics of Consumers (SFCC), also conducted by the Federal Reserve Board (see Projector and Weiss, 1966). This is also a stratified sample which over-samples high income households. Though the sample design and questionnaire are different from the SCF, the methodology is sufficiently similar to allow comparisons with the SCF data (see Wolff, 1987, for details on the adjustments). The second is a synthetic dataset, the 1969 MESP database. A statistical matching technique was employed to assign income tax returns for 1969 to households in the 1970 Census of Population. Property income flows (such as dividends) in the tax data were then capitalized into corresponding asset values (such as stocks) to obtain estimates of household wealth (see Wolff, 1980, for details).

4. Median wealth plummets over the late 2000s

Table 1 documents a robust growth in wealth from 1983 to 2007, even back to 1962. From 1962 to 1983, median wealth in real terms increased at an annual rate of 1.63 percent. Median wealth then grew slightly faster between 1989 and 2001, 1.32 percent per year, than between 1983 and 1989, at 1.13 percent per year. Over the 2001-2007 period it increased by 19 percent or an annual rate of 2.91 percent, even faster than during the 1970s, 1980s, and 1990s, though comparable to the 1960s. Then between 2007 and 2010, median wealth plunged by a staggering 47 percent! Indeed, median wealth was actually lower in 2010 than in 1969 (in real terms). The primary reasons, as we shall see below, were the collapse in the housing market and the high leverage of middle class families.12

Mean net worth also grew vigorously from 1962 to 1983, at an annual rate of 1.82 percent. Mean wealth grew quite a bit faster between 1989 and 2001, at 3.02 percent per year, than from 1983 to 1989, at 2.27 percent per year. There was then a slight increase in wealth growth from 2001 to 2007 to 3.10 percent per year. This modest acceleration was due largely to the rapid increase in housing prices counterbalanced by the reduced growth in stock prices between 2001 and 2007 in comparison to 1989 to 2001, and to the fact that housing comprised 28 percent and (total) stocks made up 25 percent of total assets in 2001. Another point of note is that mean wealth grew more about twice as fast as the median between 1983 and 2007, indicating widening inequality of wealth over these years.

The great Recession also saw an absolute decline in mean household wealth. However, whereas median wealth plunged by 47 percent, mean wealth fell by (only) 18 percent.13 In this case, both falling housing and stock prices were the main causes (see below). However, here, too, the relatively faster growth in mean wealth than median wealth (that is, the latter’s more moderate decline) was coincident with rising wealth inequality.

Median household income in real terms, based on the Current Population Survey (CPS), advanced at a fairly solid pace from 1962 to 1983, at 0.85 percent per year. Then, after gaining 11 percent between 1983 and 1989, it grew by only 2.3 percent from 1989 to 2001 and another 1.6 percent from 2001 to 2007. From 2007 to 2010, it fell off by 6.4 percent. This reduction was not nearly as great as that in median wealth. Mean income, after advancing at an annual rate of 1.2 percent from 1962 to 1983, surged by 2.4 percent per year from 1983 to 1989, advanced by 0.9 percent per year from 1989 to 2001, and then dipped by -0.1 percent per year from 2001 to 2007. Mean income also dropped in real terms from 2007 to 2010, by 5.0 percent, slightly less than that of median income.

In sum, while median household income virtually stagnated for the average American household over the 1990s and 2000s, median net worth grew strongly over this period. From 2001 to 2007, mean and median income changed very little while mean and median net worth grew strongly. The Great Recession, on the other hand, saw a massive reduction in median net worth but much more modest declines in mean wealth and both median and mean income.

5. Wealth inequality jumps in the late 2000s

The figures in Table 2 also show that wealth inequality in 1983 was quite close to its level in 1962. Then, after rising steeply between 1983 and 1989, it remained virtually unchanged from 1989 to 2007. The share of wealth held by the top 1 percent rose by 3.6 percentage points from 1983 to 1989 and the Gini coefficient increased from 0.80 to 0.83. What was behind the sharp rise in wealth inequality? There are two principal factors accounting for changes in wealth concentration (also see Section 8). The first is the change in income inequality and the second is the change in the ratio of stock prices to housing prices. As we shall see below, there was a huge increase in income inequality between 1983 and 1989, with the Gini coefficient rising by 0.041 points. Second, stock prices increased much faster than housing prices. The stock market boomed and the S&P 50 Index in real terms was up by 62 percent, whereas median home prices increased by a mere two percent in real terms. As a result, the ratio between the two climbed by 58 percent.

Between 1989 and 2007, the share of the top percentile actually declined sharply, from 37.4 to 34.6 percent, though this was more than compensated for by an increase in the share of the next four percentiles. As a result, the share of the top five percent increased from 58.9 percent in 1989 to 61.8 percent in 2007, and the share of the top quintile rose from 83.5 to 85.0 percent.14 The share of the fourth and middle quintiles each declined by about a percentage point from 1989 to 2007, while that of the bottom 40 percent increased by almost one percentage point. Overall, the Gini coefficient was virtually unchanged -- 0.832 in 1989 and 0.834 in 2007. 15

In contrast, the years of the Great Recession saw a very sharp elevation in wealth inequality, with the Gini coefficient rising from 0.83 to 0.87. Interestingly, the share of the top percentile showed less than a one percentage point gain.16 Most of the rise in wealth share took place in the remainder of the top quintile, and overall the share of wealth held by the top quintile climbed by almost four percentage points. The shares of the other quintiles, correspondingly, dropped, with the share of the bottom 20 percent falling from 0.2 percent to -0.9 percent.

The top 1 percent of families (as ranked by income on the basis of the SCF data) earned 17 percent of total household income in 2009 and the top 20 percent accounted for 59 percent -- large figures but lower than the corresponding wealth shares.17 The time trend for income inequality also contrasts with that for wealth inequality. Income inequality showed a sharp rise from 1961 to 1982, with the Gini coefficient expanding from 0.428 to 0.480 and the share of the top one percent from 8.4 to 12.8 percent.18 Income inequality increased sharply again between 1982 and 1988, with the Gini coefficient rising from 0.48 to 0.52 and the share of the top one percent from 12.8 to 16.6 percent. There was then very little change between 1988 and 1997. However, between 1997 and 2000, income inequality again surged, with the share of the top percentile rising by 3.4 percentage points, the shares of the other quintiles falling again, and the Gini index advancing from 0.53 to 0.56.19 This was followed by a modest uptick in income inequality, with the Gini coefficient advancing from 0.562 in 2000 to 0.574 in 2006. All in all, years 2001 to 2007 witnessed moderate rises in both wealth and income inequality.

Perhaps, somewhat surprisingly, the Great Recession witnessed a rather sharp contraction in income inequality. The Gini coefficient fell from 0.574 to 0.549 and the share of the top one percent dropped sharply from 21.3 to 17.2 percent. Property income and realized capital gains (which is included in the SCF definition of income), as well as corporate bonuses and the value of stock options, plummeted over these years, a process which explains the steep decline in the share of the top percentile. Real wages, as noted above, actually rose over these years, though the unemployment rate also increased. As a result, the income of the middle class was down but not nearly as much in percentage terms as that of the high income groups. In contrast, transfer income such as unemployment insurance rose, so that the bottom also did better in relative terms than the top. As a result, overall income inequality fell over the years 2006 to 2009.20 One of the puzzles we have to contend with is the fact wealth inequality rose sharply over the Great Recession while income inequality fell sharply. I will return to this question in Section 8 below.

5.1 The share of overall wealth gains, 1983 to 2010

Over the years 1983 to 2010, is period, the largest gains in wealth and income in relative terms were made by the wealthiest households (see Table 3). The top one percent saw their average wealth (in 2010 dollars) rise by 71 percent. The remaining part of the top quintile experienced increases from 52 to 101 percent and the fourth quintile by 21 percent, while the middle quintile lost 18 percent and the poorest 40 percent lost 270 percent!

Another way of viewing this phenomenon is afforded by calculating the proportion of the total increase in real household wealth between 1983 and 2010 accruing to different wealth groups. This is computed by dividing the increase in total wealth of each percentile group by the total increase in household wealth, while holding constant the number of households in that group. If a group's wealth share remains constant over time, then the percentage of the total wealth growth received by that group will equal its share of total wealth. If a group's share of total wealth increases (decreases) over time, then it will receive a percentage of the total wealth gain greater (less) than its share in either year. However, it should be noted that in these calculations, the households found in a given group may be different in the two years.

The results indicate that the richest one percent received over 38 percent of the total gain in marketable wealth over the period from 1983 to 2010. This proportion was greater than the share of wealth held by the top one percent in any of the 9 years. The next 4 percent received 36 percent of the total gain and the next 15 percent 27 percent, so that the top quintile collectively accounted for a little over 100 percent of the total growth in wealth.

A similar calculation using the SCF income data reveals that the greatest gains in real income over the period from 1982 to 2009 were made by households in the top one percent of the income distribution, who saw their incomes grow by 59 percent. Mean incomes increased by almost half for the next 4 percent, over a quarter for the next highest 5 percent and by 13 percent for the next highest ten percent. The fourth quintile of the income distribution experienced only a 3 percent growth in income, while the middle quintile and the bottom 40 percent had absolute declines in mean income. Of the total growth in real income between 1982 and 2009, 39 percent accrued to the top one percent and over 100 percent to the top quintile. These figures are very close to those for wealth.

These results indicate rather dramatically that the growth in the economy during the period from 1983 to 2010 was concentrated in a surprisingly small part of the population -- the top 20 percent and particularly the top one percent.

6. Household debt continues to remain high

In 2010, owner-occupied housing accounted for 31 percent of total assets (see Table 4). However, net home equity -- the value of the house minus any outstanding mortgage -- amounted to only 18 percent of total assets. Real estate, other than owner-occupied housing, comprised 12 percent, and business equity another 18 percent. Liquid assets (demand and time deposits, money market funds, CDs, and the cash surrender value of life insurance) made up 6 percent and pension accounts 15 percent. Bonds and other financial securities amounted to 2 percent; corporate stock, including mutual funds, to 11 percent; and trust equity to 2 percent. Debt as a proportion of gross assets was 17 percent, and the debt-equity ratio (the ratio of household debt to net worth) was 0.21.

There were some significant changes in the composition of household wealth over the years 1983 to 2010. First, the share of gross housing wealth in total assets, after fluctuating between 28.2 and 30.4 percent from 1983 to 2001, increased to 32.8 percent in 2007 and then fell to 31.3 percent in 2010. There are two main factors behind this – the homeownership rate and housing prices. According to the SCF, the homeownership rate, after falling from 63.4 percent in 1983 to 62.8 percent in 1989, picked up to 67.7 percent in 2001 and 68.6 percent in 2007 but then fell to 67.2 percent in 2010. Median house prices for existing homes rose by 19 percent in real terms between 2001 and 2007 but then plunged by 26 percent from 2007 to 2010. A substantial share of the movement of the proportion of housing in gross assets can be traced to these two time trends.21

Second, net equity in owner-occupied housing as a share of total assets, after falling from 24 percent in 1983 to 19 percent in 2001, rose to 21 percent in 2007 but then fell sharply to 18 percent in 2010. The difference between gross and net housing as a share of total assets can be traced to the changing magnitude of mortgage debt on homeowner's property, which increased from 21 percent in 1983 to 33 percent in 2001, 35 percent in 2007 and then 41 percent in 2010. Moreover, mortgage debt on principal residence climbed from 9.4 to 11.4 percent of total assets between 2001 and 2007 and then to 12.9 percent in 2010. The sharp decline in net home equity as a proportion of from 2007 to 2010 is attributable to the sharp decline in housing prices.

Third, relative indebtedness increased, with the debt-equity (net worth) ratio climbing from 15 percent in 1983 to 18 percent in 2007 and then to 21 percent in 2010. Likewise, the ratio of debt to total income surged from 68 percent in 1983 to 119 percent in 2007 and then to 127 percent in 2010, its high for this period. If mortgage debt on principal residence is excluded, the ratio of other debt to total assets actually fell off from 6.8 percent in 1983 to 3.9 percent in 2007 but then rose slightly to 4.5 percent in 2010. The large rise in relative indebtedness between 2007 and 2010 could be due to a rise in the absolute level of debt and/or a fall off in net worth and income. As shown in Table 1, both mean net worth and mean income fell over the three years. There was also a slight contraction of debt in constant dollars, with mortgage debt declining by 5.0 percent, other debt by 2.6 percent, and total debt by 4.4 percent. Thus, the steep rise in the debt to equity and the debt to income ratio over the three years was entirely due to the reduction in wealth and income.

A fourth change is a dramatic increase in pension accounts, which rose from 1.5 percent of total assets in 1983 to 12 percent in 2007 and then to 15 percent in 2010. There was a huge increase in the share of households holding these accounts between 1983 and 2001, from 11 to 52 percent. The mean value of these plans in real terms climbed dramatically. It almost tripled among account holders and skyrocketed by a factor of 13.6 among all households. These time trends partially reflect the history of DC plans. IRAs were first established in 1974. This was followed by 401(k) plans in 1978 for profit-making companies (403(b) plans for non-profits are much older). However, 401(k) plans the like did not become widely available in the workplace until about 1989.

From 2001 to 2007 the share of households with a DC plan leveled off and then from 2007 to 2010 the share fell modestly, from 52.6 to 50.4 percent. The average value of DC plans in constant dollars continued to grow after 2001. Overall, it advanced by 21 percent from 2001 to 2007 and then by 11 percent from 2007 to 2010 among account holders and by 22 percent and 7 percent, respectively, among all households. Thus, despite the stock market collapse of 2007-2010 and the 18 percent decline of overall mean net worth, the average value of DC accounts continued to grow after 2007. The reason is that households shifted their portfolio out of other assets and into DC accounts.

Fifth, the share of corporate stock and mutual funds in total assets rose rather briskly from 9 percent in 1983 to 15 percent in 2001, and then plummeted to 12 percent in 2007 and even further to 11 percent in 2010. If we include the value of stocks indirectly owned through mutual funds, trusts, IRAs, 401(k) plans, and other retirement accounts, then the value of total stocks owned as a share of total assets more than doubled from 11 percent in 1983 to 25 percent in 2001, tumbled to 17 percent in 2007, and then rose slightly to 18 percent in 2010. The rise during the 1990s reflected the bull market in corporate equities as well as increased stock ownership, while the decline in the 2000s was a result of the sluggish stock market as well as a drop in stock ownership.

6.1 Portfolio composition by wealth class

The tabulation in Table 4 provides a picture of the average holdings of all families in the economy, but there are marked class differences in how middle-class families and the rich invest their wealth. As shown in Table 5, the richest one percent of households (as ranked by wealth) invested over three quarters of their savings in investment real estate, businesses, corporate stock, and financial securities in 2010. Corporate stocks, either directly or indirectly owned, comprised 21 percent. Housing accounted for only 9 percent of their wealth, liquid assets 5 percent, and pension accounts 8 percent. The debt-equity ratio was only 3 percent, the ratio of debt to income was 61 percent, and the ratio of mortgage debt to house value was 19 percent.

Among the next richest 19 percent of U.S. households, housing comprised 30 percent of their total assets, liquid assets 7 percent, and pension assets 21 percent. Investment assets – non-home real estate, business equity, stocks, and bonds – made up 41 percent and 20 percent was in the form of stocks directly or indirectly owned. Debt amounted to 14 percent of their net worth and 118 percent of their income, and the ratio of mortgage debt to house value was 30 percent.

In contrast, almost exactly two thirds of the wealth of the middle three quintiles of households was invested in their own home in 2010. However, home equity amounted to only 32 percent of total assets, a reflection of their large mortgage debt. Another 20 percent went into monetary savings of one form or another and pension accounts. Together housing, liquid assets, and pension assets accounted for 87 percent of total assets, with the remainder in investment assets. Stocks directly or indirectly owned amounted to only 8 percent of their total assets. The debt-equity ratio was 0.72, substantially higher than that for the richest 20 percent, and their ratio of debt to income was 135 percent, also much higher than that of the top quintile. Finally, their mortgage debt amounted to a little more than half the value of their principal residences.

Almost all households among the top 20 percent of wealth holders owned their own home, in comparison to 68 percent of households in the middle three quintiles. Three-quarters of very rich households (in the top percentile) owned some other form of real estate, compared to 49 percent of rich households (those in the next 19 percent of the distribution) and only 12 percent of households in the middle 60 percent. Eighty-nine percent of the very rich owned some form of pension asset, compared to 83 percent of the rich and 46 percent of the middle. A somewhat startling 74 percent of the very rich reported owning their own business. The comparable figures are 30 percent among the rich and only 8 percent of the middle class.

Among the very rich, 89 percent held corporate stock, mutual funds, financial securities or a trust fund, in comparison to 61 percent of the rich and only 15 percent of the middle. Ninety-five percent of the very rich reported owning stock either directly or indirectly, compared to 84 percent of the rich and 41 percent of the middle. If we exclude small holdings of stock, then the ownership rates drop off sharply among the middle three quintiles, from 41 percent to 29 percent for stocks worth $5,000 or more and to 24 percent for stocks worth $10,000 or more.

The rather staggering debt level of the middle class in 2010 raises the question of whether this is a recent phenomenon or whether it has been going on for some time. Table 6 shows the wealth composition for the middle three wealth quintiles from 1983 to 2010. Houses as a share of assets remained virtually unchanged from 1983 to 2001 but then increased from 2001 to 2010. It might seem surprising that despite the steep drop in home prices from 2007 to 2010, housing as a share of total assets actually increased slightly. The reason is that the other components of wealth fell even more than housing. While housing fell by 30 percent in real terms, other real estate was down by 39 percent, liquid assets by 48 percent, and stocks and mutual funds by 47 percent.

Pension accounts rose as a share of total assets by almost 13 percentage points from 1983 to 2010 while liquid assets declined as a share by 16 percentage points. This set of changes paralleled that of all households. The share of all stocks in total assets mushroomed from 2.4 percent in 1983 to 12.6 percent in 2001 and then fell off to 8.2 percent in 2010 as stock prices stagnated and then collapsed and middle class households divested themselves of stock holdings. The proportion of middle class households with a pension account surged by 41 percentage points between 1983 and 2007 but then fell off sharply by almost 8 percentage points in 2010.

Changes in debt, however, represent the most dramatic movements. There was a sharp rise in the debt-equity ratio of the middle class from 0.37 in 1983 to 0.61 in 2007, with all of the increase occurring between 2001 and 2004, a reflections mainly of a steep rise in mortgage debt. The debt to income ratio more than doubled from 1983 to 2007. Once, again, much of the increase happened between 2001 and 2004. The rise in the debt-equity ratio and the debt to income ratio was much steeper than for all households. In 1983, for example, the debt to income ratio was about the same for middle class as for all households but by 2007 the ratio was much larger for the middle class.

Then, the Great Recession hit. The debt-equity ratio continued to rise, reaching 0.72 in 2010 but there was actually a retrenchment in the debt to income ratio, falling to 1.35 in 2010. The reason is that from 2007 to 2010, the mean debt of the middle class in constant dollars actually contracted by 25 percent. There was, in fact, a 23 percent reduction in mortgage debt as families paid down their outstanding balances, and an even larger drop in other debt of 32 percent as families paid off credit card balances and other forms of consumer debt. The steep rise in the debt-equity ratio of the middle class between 2007 and 2010 was due to the sharp drop in net worth, while the decline in the debt to income ratio was almost exclusively due to the sharp contraction of overall debt.

As for all households, the ratio of net home equity to assets fell for the middle class from 1983 to 2010 and mortgage debt as a proportion of house value rose. The decline in the ratio of net home equity to total assets between 2007 and 2010 was relatively small despite the steep decrease in home prices, a reflection of the sharp reduction in mortgage debt. On the other hand, the rise in the ratio of mortgage debt to house values was relatively large over these years because of the fall off in home prices.

6.2 The “middle class squeeze”

Nowhere is the middle class squeeze more vividly demonstrated than in their rising debt. As noted above, the ratio of debt to net worth of the middle three wealth quintiles rose from 0.37 in 1983 to 0.46 in 2001 and then jumped to 0.61 in 2007. Correspondingly, their debt to income rose from 0.67 in 1983 to 1.00 in 2001 and then zoomed up to 1.57 in 2007 This new debt took two major forms. First, because housing prices went up over these years, families were able to borrow against the now enhanced value of their homes by refinancing their mortgages and by taking out home equity loans. In fact, mortgage debt on owner-occupied housing (principal residence only) as a proportion of total assets climbed from 29 percent in 1983 to 47 percent in 2007, and home equity as a share of total assets fell from 44 to 35 percent over these years. Second, because of their increased availability, families ran up huge debt on their credit cards.

Where did the borrowing go? Some have asserted that it went to invest in stocks. However, if this were the case, then stocks as a share of total assets would have increased over this period, which it did not (it fell from 13 to 7 percent between 2001 and 2007). Moreover, they did not go into other assets. In fact, the rise in housing prices almost fully explains the increase in the net worth of the middle class from 2001 to 2007. Of the $16,400 rise in median wealth, gains in housing prices alone accounted for $14,000 or 86 percent of the growth in wealth. Instead, it appears that middle class households, experiencing stagnating incomes, expanded their debt in order to finance normal consumption expenditures.

The large build-up of debt set the stage for the financial crisis of 2007 and the ensuing Great Recession. When the housing market collapsed in 2007, many households found themselves “underwater,” with larger mortgage debt than the value of their home. This factor, coupled with the loss of income emanating from the recession, led many home owners to stop paying off their mortgage debt. The resulting foreclosures led, in turn, to steep reductions in the value of mortgage-backed securities. Banks and other financial institutions holding such assets experienced a large decline in their equity, which touched off the financial crisis.

7. The role of leverage in explaining the steep fall in median wealth and the sharp rise in wealth inequality over the Great Recession

Two major puzzles emerge from the preceding analysis. The first is the steep plunge in median net worth between 2007 and 2010 of 47 percent. This happened despite a moderate drop in median income of 6.4 percent in real terms and steep but less steep declines in housing and stock prices of 24 and 26 percent in real terms, respectively.

The second is the steep increase of wealth inequality of 0.035 Gini points. It is surprising that wealth inequality rose so sharply, given that income inequality dropped by 0.025 Gini points (at least according to the SCF data) and the ratio of stock prices to housing prices was essentially unchanged. In fact, as shown in Wolff (2002), wealth inequality is positively related to the ratio of stock to house prices, since the former is heavily concentrated among the rich and the latter is the chief asset of the middle class. A regression run of the share of wealth held by the top one percent of households (WLTH) on the share of income received by the top five percent of families (INC), and the ratio of the Standard and Poor 500 index to housing prices (RATIO), with 21 data points between 1922 and 1998, yields:

(1) WLTH = 5.10 + 1.27 INC + 0.26 RATIO, R2 = 0.64, N = 21 (0.9) (4.2) (2.5)

with t-ratios shown in parentheses. Both variables are statistically significant (INC at the 1 percent level and RATIO at the 5 percent level) and with the expected (positive) sign. Also, the fit is quite good, even for this simple model.

Changes in median wealth and wealth inequality from 2007 to 2010 can be explained to a large extent by leverage (the ratio of debt to net worth). The steep fall in median wealth was due in large measure to the high leverage of middle class households. The spike in wealth inequality was largely due to differential leverage between the rich and the middle class.

7.1 Two arithmetic examples

A simple arithmetical example might illustrate the effects of leverage. Suppose average assets are 50 and average debt is zero. Also, suppose that asset prices rise by 20 percent. Then average net worth also rises by 20 percent. However, now suppose that average debt is 40 and asset prices once again rise by 20 percent. Then average net worth increases from a base of 10 (50 minus 40) to 20 (60 minus 40) or by 100 percent, Thus, leverage amplifies the effects of asset price changes. However, the converse is also true. Suppose that asset prices decline by 20 percent. In the first case, net worth falls from 50 to 40 or by 20 percent. In the second case, net worth falls from 10 to 0 (40 minus 40) or by 100 percent. Thus, leverage can also magnify the effects of an asset price bust.

Another arithmetical example can illustrate the effects of differential leverage. Suppose the total assets of the very rich in a given year is 100, consisting of 50 in stocks and 50 in other assets, and its debt is zero, for a net worth of 100. For the “middle class”, suppose total assets are 70, consisting of 60 in housing and 10 in other assets, and their debt is 30, for a net worth of 40. The ratio of net worth between the very rich and the middle is then 2.5 (100/40).

Suppose the value of both stocks and housing falls by 20 percent but the value of “other assets” remains unchanged. Then, the total assets of the rich fall to 90 (40 in stocks and 50 in other), for a net worth of 90. The total assets of the middle falls to 58 (48 in housing and 10 in other) but its debt remains unchanged at 30, for a net worth of 28. As a result, the ratio of net worth between the two groups rises to 3.21 (90/28). Here it is apparent that even though housing and stock prices fall at the same rate, wealth inequality goes up. The reason is differential leverage between the two groups. If asset prices decline at the same rate, net worth decreases at an even greater rate for the middle than the rich, since the debt-equity ratio is higher for the former than the latter. The converse is also true. A proportionate increase in house and stock prices will result in a decrease in wealth inequality.

7.2 Rates of return

Table 7 shows estimated average annual rates of return for both gross assets and net worth over the period from 1983 to 2010. Results are based on the average portfolio composition over the period. It is first of interest to look at the results for all households . The overall average annual rate of return on gross assets rose from 2.20 percent in the 1983-1989 period to 3.25 percent in the 1989-2001 period and then to 3.34 percent in the 2001-2007 period before plummeting to -6.95 percent over the Great Recession. The largest declines in asset prices over the years 2007 to 2010 occurred for residential real estate and the category businesses and non-home real estate. The value of financial assets, including stocks, bonds, and other securities, registered an annual rate of return of “only” -2.23 percent because interest rates on corporate and foreign bonds continued to remain strong over these years. The value of pension accounts had a -2.46 percent annual rate of return, reflecting the mixture of bonds and stocks held in pension accounts.

The average annual rate of return on net worth among all households also increased from 3.17 percent in the first period to 4.25 percent in the second and then to 4.31 percent in the third but then fell off sharply to -7.98 percent in the last period. It is first of note that the annual rates of return on net worth are uniformly higher – by about one percentage point – than those of gross assets over the first three periods, when asset prices were generally rising. However, in the 2007-2010 period, the opposite was the case, with the annual return on net worth 0.44 percent lower than that on gross assets. These results illustrate the effect of leverage, raising the return when asset prices rise and lowering the return when asset prices fall. Over the full 1983-2010 period, the annual return on net worth was 0.87 percentage points higher than that on gross assets.22

There are striking differences in returns by wealth class. The highest returns on gross assets were registered by the top one percent of wealth holders, followed by the next 19 percent and then by the middle three wealth quintiles. The one exception was the 2007-2010 period when the next 19 percent was first, followed by the top one percent and then the middle three quintiles. The differences are quite substantial. Over the full 1983-2010 period, the average annual rate of return on gross assets for the top one percent was 0.55 percentage points greater than that of the next 19 percent and 1.39 percentage points greater than that of the middle quintiles. The differences reflect the greater share of high yield investment assets like stocks in the portfolios of the rich and the greater share of housing in the portfolio of the middle class (see Table 5 for details on portfolio composition).

This pattern is almost exactly reversed for rates of return for net worth. In this case, in the first three periods when asset prices were generally rising, the highest return was recorded by the middle three wealth quintiles but in the 2007-2010 period, when asset prices were declining, the middle three quintiles registered the lowest (that is, most negative) return. The exception was the first period when the top one percent had the highest return. The reason was the substantial spread in returns on gross assets between the top one percent and the middle three quintiles – 1.79 percentage points. Interestingly, returns for the top one percent were greater than that of the next 19 percent and for the same reason.

Differences in returns between the top one percent and the middle three quintiles were quite substantial in some years. In the 2001-2007 period, the return on net worth was 5.95 percent per year for the latter and 4.03 percent per year for the former – a difference of 1.92 percentage points. Over the Great Recession the rate of return on net worth was -7.10 percent for the top one percent and, astonishingly, -11.39 percent for the middle three quintiles – a differential of 4.27 percentage points. The spread in rates of return between the top one percent and the middle three quintiles reflects the much higher leverage of the middle class. In 2010, for example, the debt-equity ratio of the middle three quintiles was 0.72 while that of the top one percent was 0.04. The debt-equity ratio of the next 19 percent was also relatively low, at 0.14. Indeed, except for years 2007 to 2010, the rate of return on net worth for the middle quintiles was more than double its return on gross assets.

The huge negative rate of return on net worth of the middle three wealth quintiles was largely responsible for the precipitous drop in median net worth between 2007 and 2010. This factor, in turn, was due to the steep drop in asset prices, particularly housing, and the very high leverage of the middle wealth quintiles. Likewise, the very high rate of return on net worth of the middle three quintiles over the 2001-2007 period (5.95 percent per year) played a big role in explaining the robust advance of median net worth, despite the sluggish growth in median income. This in turn, was a result of their high leverage coupled with the boom in housing prices.

The substantial differential in rates of return on net worth between the middle three wealth quintiles and the top quintile (over four percentage points) helps explain why wealth inequality rose sharply between 2007 and 2010 despite the decline in income inequality. Likewise this differential over the 2001-2007 period (a spread of about two percentage points in favor of the middle quintiles) helps account for the stasis in wealth inequality over these years despite the increase in income inequality.

8. The racial divide widens over the Great Recession

Striking differences are found in the wealth holdings of different racial and ethnic groups. In Tables 10 and 11, households are divided into three groups: (i) non-Hispanic whites (“whites” for short), (ii) non-Hispanic African-Americans (“blacks” for short), and (iii) Hispanics.23 In 2007, while the ratio of mean incomes between white and black households was an already low 0.48 and the ratio of median incomes was 0.60, the ratios of mean and median wealth holdings were even lower, at 0.19 and 0.06, respectively.24 The homeownership rate for black households was 49% in 2007, a little less than two thirds that among whites, and the percentage of black households with zero or negative net worth stood at 33, more than double that among whites.

Between 1982 and 2006, while the average real income of white households increased by 42 percent and the median by 10 percent, the former rose by only 28 percent for black households but the latter by 18 percent. As a result, the ratio of mean income slipped from 0.54 in 1982 to 0.48 in 2006, while the ratio of median income rose from 0.56 to 0.60.25 The contrast in time trends between the ratio of means and that of medians reflects the huge increase in income for a relatively small number of white households – a result of rising income inequality among whites.

Between 1983 and 2001, average net worth in constant dollars climbed by 73 percent for whites but rose by only 31 percent for black households, so that the net worth ratio fell from 0.19 to 0.14. Most of the slippage occurred between 1998 and 2001, when white net worth surged by a spectacular 34 percent and black net worth advanced by only a respectable 5 percent. Indeed, mean net worth growth among black households was slightly higher in the 1998-2001 years, at 1.55 percent per year, than in the preceding 15 years, at 1.47 percent per year. However, between 2001 and 2007, mean net worth among blacks gained an astounding 58 percent while white wealth advanced by 29 percent, so that by 2007 the net worth ratio was back to 0.19, the same level as in 1983.

It is not clear how much of the sharp drop in the racial wealth gap between 1998 and 2001 and the turnaround between 2001 and 2007 are due to actual wealth changes in the African-American community and how much is due to sampling variability (since the sample sizes of African Americans are relatively small in all years). However, one salient difference between the two groups is the much higher share of stocks in the white portfolio and the much higher share of homes in the portfolio of black households. In 2001, the gross value of principal residences formed 46 percent of the total assets of black households, compared to 27 percent among whites, while (total) stocks were 25 percent of the total assets of whites and only 15 percent that of black households. In the case of median wealth, the black-white ratio fluctuated over time but was almost exactly the same in 2007 as in 1983, 0.06 compared to 0.07.

The homeownership rate of black households grew from 44 to 47 percent between 1983 and 2001 but relative to white households, the homeownership rate slipped slightly from 0.65 in 1983 to 0.64 in 2001. From 2001 to 2007, the black homeownership rate rose slightly from 74.1 to 74.8 percent, and the ratio of homeownership rates advanced slightly, to 0.65. The percentage of black households with zero or negative net worth fell from 34 percent in 1983 to 31 percent in 2001 (and also declined relative to the corresponding rate for whites). However, in the ensuing six years the share rose back to 33 percent in 2007 (though relative to white households remained largely unchanged).

The picture is rather different for Hispanics (see Table 9). The ratio of mean income between Hispanics and non-Hispanic whites in 2007 was 0.50, almost the same as that between black and white households. However, the ratio of median income was 0.70, much higher than the ratio between black and white households. The ratio of mean net worth was 0.26 compared to a ratio of 0.19 between blacks and whites. However, the ratio of medians was 0.06, almost identical to that between blacks and whites. The Hispanic homeownership rate was 49 percent, almost identical to that of black households, and 34 percent of Hispanic households reported zero or negative wealth, almost the same as African- Americans.

Hispanic households made considerable progress over the years1983 to 2007. Mean income grew by 18 percent and median income by 16 percent, so that while the ratio of mean income slid from 60 to 50 percent, that of median income advanced from 66 to 70 percent. Between 1983 and 2001, mean wealth doubled for Hispanic households and the ratio of mean net worth increased slightly from 16 to 17 percent. Mean net worth among Hispanics then climbed by another 82 percent between 2001 and 2007, and the corresponding ratio advanced to 26 percent, quite a bit higher than that between black and white households. The surge in Hispanic wealth from 2001 to 2007 can be traced to a five percentage point jump in the Hispanic home ownership rate (see below).

From 1983 to 2007, median wealth among Hispanics remained largely unchanged, so that the ratio of median wealth between Hispanics and whites stayed virtually the same. In contrast, the homeownership rate among Hispanic households surged from 33 to 44 percent between 1983 and 2001, and the ratio of homeownership rates between the two groups grew from 0.48 in 1983 to 0.60 in 2001. Between 2001 and 2007, the Hispanic homeownership rose once again, to 49 percent, about the same as black households, and the homeownership ratio rose sharply to 0.66. The percentage of Hispanic households with zero or negative net worth fell rather steadily over time, from 40 percent in 1983 to 34 percent in 2007 (about the same as black households), and the share relative to white household tumbled from a ratio of 3.0 to 2.3.

Despite some progress from 2001 to 2007, the respective wealth gaps between minorities and whites were still much greater than the corresponding income gaps in 2007. While mean income ratios were of the order of 50 percent, mean wealth ratios were of the order of 20-25 percent and the share with zero or negative net worth was around a third, in contrast to 15 percent among non-Hispanic white households (a difference that appears to mirror the gap in poverty rates). While blacks and Hispanics were left out of the wealth surge of the years 1998 to 2001 because of relatively low stock ownership, they actually benefited from this (and the relatively high share of houses in their portfolio) in the 2001-2007 period. However, all three racial/ethnic groups saw an increase in their debt to asset ratio from 2001 to 2007.

The racial picture really changed radically by 2010. While the ratio of both mean and median income between black and white households changed very little between 2007 and 2010 (mean income, in particular, declined for both groups), the ratio of mean net worth dropped from 0.19 to 0.14. The proximate causes were the higher leverage of black households and their higher share of housing wealth in gross assets (see Table 10). In 2007, the debt-equity ratio among blacks was an astounding 0.55, compared to 0.15 among whites, while housing as a share of gross assets was 0.54 for the former as against 0.31 for the latter. The ratio of mortgage debt to home value was also much higher for blacks, 0.49, than for whites, 0.32. The sharp drop in home prices from 2007 to 2010 thus led to a relatively steeper loss in home equity for the former, 25 percent, than the latter, 21 percent, and this factor, in turn, led to a much steeper fall in mean net worth for black households than white households.26 Indeed, in terms of rates of return, while the overall annual rate of return on net worth among white households plummeted from 4.1% in 2001-2007 to -7.7% in 2007-2010, it collapsed from 6.4% to -10.9% among black households. It is of note that black households had a much higher return on net worth than white households in the 2001-2007 period but a much lower return in 2007- 2010.

The Great Recession actually hit Hispanic households much harder than blacks in terms of wealth. Mean income among Hispanic households rose a bit from 2007 to 2010 and the ratio with respect to white households increased from 0.50 to 0.57. On the other hand, the median income of Hispanics fell, as did the ratio of median income between Hispanics and whites. However, the mean net worth in 2010 dollars of Hispanics fell almost in half, and the ratio of this to the mean wealth of whites plummeted from 0.26 to 0.15. The same factors were responsible as in the case of black households. In 2007, the debt-equity ratio for Hispanics was 0.51, compared to 0.15 among whites, while housing as a share of gross assets was 0.53 for the former as against 0.31 for the latter (see Table 10). The ratio of mortgage debt to home value was also higher for Hispanics, 0.45, than for whites, 0.32. As a result, net home equity dropped by 48 percent among Hispanic home owners, compared to 21 percent among white home owners, and this factor, in turn, was largely responsible for the huge decline in Hispanic net worth both in absolute and relative terms. In terms of the annual return on net worth, it nosedived from 6.7% in 2001-2007 to -11.8% in 2007-2010. The drop was even steeper than that for black households. In fact, while Hispanic households had a higher return than white or black households in 2001-2007, it had the lowest return in 2007-2010.

There are two reasons that might explain the extreme drop in Hispanic net worth. First, a large proportion of Hispanic home owners bought their home in the interval from 2001 to 2007, when home prices were peaking. This is reflected in the sharp increase in their home ownership rate over this period. As a result, they suffered a disproportionately large percentage drop in their home equity. Second, it is likely that Hispanic home owners were more heavily concentrated than whites in parts of the country like Arizona, California, Arizona, and Nevada (the “sand states”) and Florida, where home prices plummeted the most.

There was also a steep drop in the home ownership rate among Hispanic households of 1.9 percentage points from 2007 to 2010. Indeed, after catching up on white households in this dimension from 1983 to 2007, Hispanic households fell back in 2010 to the same level as in 2004.

9. Wealth shifts from the young to the old

As shown in Table 11, the cross-sectional age-wealth profiles generally follow the predicted hump-shaped pattern of the life-cycle model. Mean wealth increases with age up through age 65 or so and then falls off. Home ownership rates have a similar profile, though the fall-off after the peak age is much more attenuated than for the wealth numbers (and in 2004 they actually show a steady rise with age). In 2010, the wealth of elderly households (age 65 and over) was 2.1 times as high as that of the non-elderly and their homeownership rate was 19 percentage points higher. Despite the apparent similarity in the profiles, there were notable shifts in the relative wealth holdings by age group from 1983 to 2007. The relative wealth of the youngest age group, under 35 years of age, declined from 21 percent of the overall mean in 1983 to 17 percent in 2007. In 2007, the mean wealth of the youngest age group was $95,900 (in 2010 dollars), which was only slightly more than the mean wealth of this age group in 1989 ($93,100). Though as noted in Section 1, educational loans expanded markedly over the 2000s and by 2007 one third of households in this age group reported a student loan outstanding, still 74 percent of the total debt of this age group was mortgage debt and only 9.5 percent took the form of student loans.

The mean net worth of the next youngest age group, 35-44, relative to the overall mean, collapsed from 0.71 in 1983 to 0.58 in 2007. The relative wealth of the next youngest age group, 45- 54, also declined, from 1.53 in 1983 to 1.19 in 2007. The relative wealth of age group 55-64 was about the same in 2007, 1.69, as in 1983,1.67. The relative net worth of age group 65-74 plummeted from 1.93 in 1983 to 1.61 in 1989 but recovered to 1.86 in 2007. The wealth of the oldest age group, age 75 and over, gained ground, from only 5 percent above the mean in 1983 to 16 percent in 2007.

Changes in homeownership rates tend to mirror these trends. While the overall ownership rate increased by 5.2 percentage points from 63.4 to 68.6 percent between 1983 and 2007, the share of households in the youngest age group owning their own home increased by only 2.1 percentage points. The homeownership rate of households between 35 and 44 of age actually fell by 2.3 percentage points, and that of age group 45 to 54 years of age declined by 0.9 percentage points. Big gains in homeownership were recorded by the older age groups: 3.9 percentage points for age group 55-64, 7.1 percentage points for age group 65-74, and 7.6 percentage points for the oldest age group.27 By 2007, homeownership rates rose monotonically with age up to age group 65-74 and then dropped for the oldest age group. The statistics point to a relative shifting of home ownership away from younger towards older households between 1983 and 2007.

Changes in relative wealth were even more dramatic from 2007 to 2010. The relative wealth of the under 35 age group plummeted from 0.17 to 0.10 and that of age group 35-44 from 0.58 to 0.41, while that of age group 45-54 fell somewhat from 1.19 to 1.14. In actual (2010) dollar terms, the average wealth of the youngest age group collapsed from $95,500 in 2007 to $48,400 in 2010, is second lowest point over the 27 year period (the lowest occurred in 1995),28 while the relative wealth of age group 35-44 shrank from $325,00 to $190,000 its lowest point over the whole 1983 to 2010 period. One possible reason for these steep declines in wealth is that younger households were more likely to have bought homes near the peak of the housing cycle.

In contrast, the relative net worth of age group 55-64 increased sharply from 1.69 to 1.81 (though it shrank in actual 2010 dollar terms from $950,400 to $841,000) and that of the oldest age group from 1.16 to 1.36 (though once again it was down in absolute terms from $653,700 to $629,100), though the relative wealth of age group 65 to 74 declined from 1.86 to 1.74 (and fell in absolute dollars as well, from $1,048,600 to $808,500). Home ownership rates fell for all age groups from 2007 to 2010 (except the very oldest) but the percentage point decline (3.3 percentage points) was greatest for the youngest age group.

Changes in the relative wealth position of different age groups depend in large measure on relative asset price movements and differences in asset composition. The latter are highlighted in Table 12 for the year 2007. Homes comprised over half the value of total assets for age group 35 and under, and the share fell off with age to about a quarter for age group 55-64 and then rose to 30 percent for the oldest age group. Liquid assets as a share of total assets remained relatively flat with age group at around 6 percent except for the oldest group for whom it was 11 percent, perhaps reflecting the relative financial conservativeness of older people. Pension accounts as a share of total assets rose from 4 percent for the youngest group to 16 percent for age group 55 to 64 and then fell off to 5 percent for the oldest age group. This pattern reflects the build-up of retirement assets until retirement age and then a decline as these retirement assets are liquidated.29 Corporate stock and financial securities showed a steady rise with age, from a 4 percent share for the youngest group to a 26 percent share for the oldest. A similar pattern was evident for total stocks as a percentage of all assets. Unincorporated business equity and non-home real estate were relatively flat as a share of total assets with age, about 30 percent. There was a pronounced fall off of the debt-equity ratio with age, declining from 0.93 for the youngest group to 0.02 for the oldest, as well as the debt to income ratio from 1.68 to 0.30 and mortgage debt as a share of house value from 0.65 to 0.05. As a result of the latter, net home equity as a proportion of total assets rose from 19 to 29 percent from the youngest to oldest age group.

Younger households were thus more heavily invested in homes and more heavily in debt whereas the portfolio of older households was more heavily skewed to financial assets, particularly corporate stock. As a result, younger households benefit relatively when housing prices rise and inflation is strong while older households benefit relatively from rising stock prices. Changes in the relative net worth position of age groups over the 1983 to 2007 period were thus largely due to these relative asset price movements. In particular, as with minority households, the higher leverage of younger age groups made them vulnerable when asset prices, particularly housing prices, declined. The steep decline in house prices from 2007 to 2010 thus led to a relatively steeper loss in home equity for the youngest age group, 59 percent, than overall, 26 percent, and this factor, in turn, led to a much steeper fall in net worth . Indeed, the annual return on the net worth of this age group nosedived from a considerable 8.1% in 2001-2007, the highest of any age group, to -15.0% (!) in 2007-2010, the lowest of any age group.

The story is very similar for age group 35 to 44. Their debt-equity ratio was 0.41 in 2007, their ratio of mortgage debt to house value was 0.51, and their share of housing in gross assets was 0.44, all much higher than overall. As with the youngest age group, the drop in home prices from 2007 to 2010 caused a large fall in home equity of 49 percent, which in turn caused a steep fall off in their relative net worth. The annual rate of return on net worth for this group tumbled from 5.9% in 2001-2007, the second highest of any age group, to -10.5% in 2007-2010, the second lowest of any age group.

10. Summary and concluding remarks

Median wealth showed robust growth during the 1980s and 1990s and an even faster advance from 2007 to 2010. Then the Great Recession hit. From 2007 to 2010, house prices fell by 24 percent in real terms, stock prices by 26 percent, and median wealth by a staggering 47 percent. Median income also dropped but by a relatively modest 6.4 percent. The percent of households with nonpositive net worth rose sharply from 18.6 to 22.5.

Wealth inequality after remaining relatively stable from 1989 to 2007 showed a steep increase over the Great Recession. The Gini coefficient climbed from 0.834 to 0.870 and the share of the top 20 percent from 85 to 89 percent. The share of the bottom 40 percent experienced a precipitous drop from 0.2 to -0.9 percent. In contrast, income inequality, after rising moderately from 2000 to 2007 (an increase of 0.12 Gini points), dropped substantially from 2006 to 2009 (a decrease of 0.25 Gini points).

The percentage increase in net worth (also income) from 1983 to 2010 was much greater for the top wealth (and income) groups than for those lower in the distribution. The greatest gains were enjoyed by the upper 20 percent, particularly the top one percent, of the respective distributions. Between 1983 and 2010, the top one percent received 38 percent of the total growth in net worth and 39 percent of the total increase in income. The figures for the top 20 percent are 101 percent and 104 percent, respectively – that is to say, the upper quintile got it all!.

The years 2001 to 2007 also saw a sharply rising debt to income ratio, reaching its highest level in almost 25 years, at 1.19 among all households in 2007. The debt-equity ratio was also way up, from 0.14 to 0.18. Most of the rising debt was from increased mortgages on homes. From 2007 to 2010 both ratios rose, the former moderately from 1.19 to 1.27 and the latter more steeply from 0.18 to 0.21. This was true despite a moderate retrenchment of overall average debt of 4.4 percent and reflected the drop in both mean wealth and income.

Home values as a share of total assets among all households remained relatively unchanged from 1983 to 2010 (around 30 percent). However, net home equity as a share of total assets fell from 0.24 in 1983 to 0.18 in 2010, reflecting rising mortgage debt on homeowner's property, which grew from 21 percent in 1983 to 35 percent in 2007 and then jumped to 41 percent in 2010. The large increase in the ratio from 2007 to 2010 was a result of falling home values (average mortgage debt actually declined by 5.0 percent in absolute terms).

Trends are more pronounced for the middle class. Among the middle three wealth quintiles, there was a huge increase in the debt-income ratio from 1.00 in 2001 to 1.57 in 2007 and an almost doubling of the debt-equity ratio from 0.32 to 0.61 percent. The debt-equity ratio was also much higher among the middle 60 percent of households in 2007, at 0.61, than among the top one percent (0.028) or the next 19 percent (0.121). However, from 2007 to 2010, while the debt-equity ratio continued to advance to 0.72, the debt to income ratio actually fell off to 1.35. The reason is the substantial retrenchment of average debt among the middle class over these years. Overall debt fell by 25 percent in real terms, mortgage debt by 23 percent, and other debt by 32 percent. The fact that the debt-equity ratio rose over these years was a reflection of the steep drop in median net worth.

Despite the 24 percent plunge in house prices (in real terms) from 2007 to 2010, the share of home owners who were underwater” was “only” 8.2 percent in 2010. However, average home equity among home owners did decline by 26 percent. This reduction would have been higher except for the contraction of mortgage debt noted above. Hispanics, younger households, and middle income households were hit particularly hard in terms of the loss of home equity.

One piece of mainly positive news is that among all households there was no deterioration in pension accumulations in DC-type pension plans over the Great Recession. The share of households with a DC account, after rising from 11 percent in 1983 to 53 percent in 2007, did fall off a bit to 50 percent in 2010. However, average DC pension wealth continued to grow from 2007 to 2010. The main reason was a shifting of household portfolios. Pension accounts as a share of total assets, after rising from 1.5 percent in 1983 to 12 percent in 2007, jumped to 15 percent in 2010. However, among middle class families, the share with a DC plan, after growing robustly from 12 percent in 1983 to 53 percent in 2007, fell off sharply to 46 percent in 2010, and the change in dollar terms from 2007 to 2010 was -24 percent. Thus, in terms of retirement preparedness from DC accounts, there was generally an improvement from 2007 to 2010 except for middle class households.

The key to understanding the plight of the middle class over the Great Recession was their high degree of leverage and the high concentration of assets in their home. The steep decline in median net worth between 2007 and 2010 was primarily due to the very high negative annual rate of return on net worth of the middle three wealth quintiles (-8.9 percent). This, in turn, was attributable to the precipitous fall in home prices and their very high degree of leverage. High leverage, moreover, helps explain why median wealth fell more than house (and stock) prices over these years and declined much more than median household income.

The large spread in rates of return on net worth between the middle three wealth quintiles and the top quintile (over a point and a half lower) also largely explained why wealth inequality increased steeply from 2007 to 2010 despite the decline in income inequality. Indeed, the middle class took a bigger relative hit on their net worth from the decline in home prices than the top 20 percent did from the stock market plunge. This factor is also reflected in the fact that median wealth dropped much more in percentage terms than mean wealth over the Great Recession. The evidence, moreover, suggests that middle class households went into debt partly in order to increase their leverage and to raise their rate of return, at least when asset prices were rising. Of course, the increased leverage also made them very vulnerable when asset prices collapsed.

The racial disparity in wealth holdings, after fluctuating over the years from 1983 to 2007, was almost exactly the same in 2007 as in 1983. However, the Great Recession hit black households much harder than whites and the ratio of mean wealth between the two groups plunged from 0.19 in 2007 to 0.14 in 2010, mainly due to a 34 percent decline (in real terms) in African-American wealth. The relative (and absolute) losses suffered by black households from 2007 to 2010 are ascribable to the fact that blacks had a higher share of homes in their portfolio than did whites and much higher debt-equity ratios (0.55 and 0.15, respectively).

Hispanic households made sizeable gains on (non-Hispanic) white households from 1983 to 2007. The ratio of mean net worth grew from 0.16 to 0.26, the homeownership rate among Hispanic households climbed from 33 to 49 percent, and the ratio of homeownership rates with white households advanced from 48 percent in 1983 to 66 percent in 2007. However, in a reversal of fortunes, Hispanic households got hammered by the Great Recession. Their mean net worth plunged in half, the ratio of mean net worth with white households fell from 0.26 to 0.15, their home ownership rate fell by 1.9 percentage points, and their net home equity plummeted by 48 percent. The relative (and absolute) losses suffered by Hispanic households over these three years are also mainly due to the much larger share of homes in their wealth portfolio and their much higher debt-equity ratio (0.51 versus 0.15). Another likely factor is that a high percentage of Hispanics bought their homes close to the housing cycle peak.

Young households also got pummeled by the Great Recession. The ratio of net worth between households under age 35 and all households fell from 0.21 in 1983 to 0.17 in 2007 and then plunged to 0.10 in 2010. In (real) dollar terms, their mean net worth declined by 49 percent from 2007 to 2010. Among age group 35-44, the ratio of their net worth to the overall figure fell from 0.71 in 1983 to 0.58 in 2007 and then declined precipitously to 0.41 in 2010. In dollar terms, their wealth fell by 42 percent over the latter three years. The same two factors explain the losses suffered by young households – the higher share of homes in their wealth portfolio and their much higher leverage ratios.

References

Kennickell, Arthur B. 2001. "Modeling Wealth with Multiple Observations of Income: Redesign of the Sample for the 2001 Survey of Consumer Finances," October, at: http://www.federalreserve.gov/pubs/oss/oss2/method.html.

Kennickell, Arthur B., and R. Louise Woodburn. 1999. "Consistent Weight Design for the 1989, 1992, and 1995 SCFs, and the Distribution of Wealth." Review of Income and Wealth 45: 193- 216.

Wolff, Edward N. 1980. “Estimates of the 1969 Size Distribution of Household Wealth in the U.S from a Synthetic Data Base.” In James Smith ed., Modeling the Distribution and Intergenerational Transmission of Wealth, Chicago: Univ. of Chicago Press, 1980, pp. 223-271.

Wolff, Edward N. 1979. "The Distributional Effects of the 1969-75 Inflation on Holdings of Household Wealth in the United States." Review of Income and Wealth series 25(2): 195-207.

-----. 1987. "Estimates of Household Wealth Inequality in the United States, 1962-83." Review of Income and Wealth series 33: 231-256.

-----. 1994. "Trends in Household Wealth in the United States, 1962-1983 and 1983-1989." Review of Income and Wealth 40: 143-174.

-----. 1998. "Recent Trends in the Size Distribution of Household Wealth." Journal of Economic Perspectives, 12: 131-150.

-------. 2002a. TOP HEAVY: A Study of Increasing Inequality of Wealth in America. Newly updated and expanded edition, New York: the New Press.

------. 2011a. “Recent Trends in Household Wealth in the U.S.: Rising Debt and the Middle Class Squeeze.” In Jason M. Gonzales, ed., Economics of Wealth in the 21st Century. Nova Science Publishers, Inc., pp. 1-41.

-----. 2011b. The Transformation of the American Pension System: Was It Beneficial for Workers? , Kalamazoo, Michigan : W.E. Upjohn Institute for Employment Research.

Table 1: Mean and Median Wealth and Income, 1962- 2010 (In thousands, 2010 dollars)

Table 2. The Size Distribution of Wealth and Income, 1962-2010

Table 3. Mean Wealth Holdings and Income by Wealth or Income Class, 1983-2010 (In thousands, 2010 dollars)

Table 4. Composition of Total Household Wealth, 1983 – 2010 (Percent of gross assets)

Table 5. Composition of Household Wealth by Wealth Class, 2010 (Percent of gross assets)

Table 6. Composition of Household Wealth of the Middle Three Wealth Quintiles, 1983-2010 (Percent of gross assets)

Table 7. Average Annual Rates of Return by Period and Wealth Class, 1983 - 2010 (percent)

Table 8. Household Income and Wealth by Race, 1983-2010 (In thousands, 2010 dollars)

Table 9. Family Income and Wealth for Non-Hispanic Whites and Hispanics, 1983-2010 (In thousands, 2010 dollars)

Table 10. Composition of Household Wealth by Race and Ethnicity, 2007 (Percent of gross assets)

Table 11. Age-Wealth Profiles and Homeownership Rates by Age, 1983-2010

Table 12. Composition of Household Wealth by Age Class, 2007 (Percent of gross assets)

_______________

Notes:

1 The source for housing price data, unless otherwise indicated, is Table 935 of the 2009 Statistical Abstract, US Bureau of the Census, available at [http://www.census.gov/compendia/statab/].

2 The source is National Association of Realtors, “Median Sales Price of Existing Single-Family Homes for Metropolitan Areas,” available at: http://www.realtor.org/sites/default/files/reports/2012/embargoes/2012-q1-metro-home-prices- 49bc10b1efdc1b8cc3eb66dbcdad55f7/metro-home-prices-q1-single-family-2012-05-09.pdf.

3 The source for stock price data is Table B-96 of the Economic Report of the President, 2012, available at http://www.gpoaccess.gov/eop/tables12.html.

4 The wage figures are based on the Bureau of Labor Statistics (BLS) hourly wage series. The source is Table B-47 of the Economic Report of the President, 2012, available at op. cit. The source for the income data is Table B-33 of the Economic Report of the President, 2012, available at op. cit.

5 The figure is for civilian employment. The source is Table B-36 of the Economic Report of the President, 2012, available at op. cit.

6 The source is Table B-42 of the Economic Report of the President, 2012, available at op. cit.

7 These figures are based on the Federal Reserve Board’s Flow of Funds data, Table B.100, available at: http://www.federalreserve.gov/releases/Z1/.

8 Unfortunately, no data on educational loans are available in the 2001 SCF.

9 The computation of DB pension wealth is based on the present value of expected pension benefits upon retirement. See Wolff (2011b) for details.

10 Another rationale is that if cars are included in the household portfolio, their “rate of return” would be substantially negative since cars depreciate very rapidly over time (see Section 8 for calculations of the overall rate of return on the household portfolio).

11 See Wolff (2011b) for estimates of Social Security and pension wealth.

12 The percentage decline in net worth from 2007 to 2010 is lower when vehicles are included in the measure of wealth – “only” 39 percent. The reason is that automobiles comprise a substantial portion of middle class wealth.

13 The decline in mean net worth is 15 percent when vehicles are included in net worth.

14 Actually, the big slippage in the share of the top one percent occurred between 1998 and 2001. The main reason appears to be a sizeable drop in the share of households in the top one percent owning their own business, from 72 to 66 percent. Whereas the mean net worth of the top one percent increased by 13.5 percent in real terms, the mean value of unincorporated business equity and other real estate grew by only 6.2 percent.

15 It might seem somewhat surprising that wealth inequality remained relatively unchanged during the latter part of the George Bush administration, the Clinton administration, and the George W. Bush administration. However, as we shall see in Section 8, stability in wealth inequality over these years was due largely to the sharp increase in the relative indebtedness of the middle class.

16 Once again, the main culprit explaining the rather meager increase in the share of the top one percent is unincorporated business equity, whose mean value fell by 26 percent in real terms from 2007 to 2010, compared to a 16 percent overall decline in their mean net worth.

17 It should be noted that the income in each survey year (say 2007) is for the preceding year (2006 in this case).

18 The 1969 MESP data suggest a huge expansion in income inequality from 1962 to 1969 but it is likely that the income data in the MESP file are flawed.

19 It should be noted that the SCF data show a much higher level of income inequality than the CPS data. In the year 2000, for example, the CPS data show a share of the top five percent of 22.1 percent and a Gini coefficient of 0.462. The difference is primarily due to three factors. First, the SCF oversamples the rich (as noted above), while the CPS is a representative sample. Second, the CPS data are top-coded (that is, there is an open-ended interval at the top, typically at $75,000 or $100,000), whereas the SCF data are not. Third, SCF income definition includes realized capital gains whereas the CPS definition does not. However, the CPS data also show a large increase of inequality between 1989 and 2000, with the share of the top five percent rising from 18.9 to 22.1 percent and the Gini coefficient from 0.431 to 0.462. .

20 The CPS data, in contrast, shows little change in household income inequality, with the Gini coefficient falling slightly from 0.470 in 2006 to 0.468 in 2009. The source for the CPS data is: http://www.census.gov/hhes/www/income/data/historical/household/2010/H04_2010.xls. However, the work of Emmanuel Saez and Thomas Piketty, based on IRS tax data, reveals a sizeable decline in income inequality from 2007 to 2010. In particular, incomes at the 99.99th, 99.9th, and 99th percentile drop sharply over these years (the source is: New York Times, October 24, 2012, page A14).

21 It may seem surprising that the share of housing in gross assets declined very little between 2007 and 2010, given the steep drop in housing prices, but the price of other assets also fell over this period, particularly those of stocks and business equity.

22 An earlier analysis was conducted by the author for the 1969-1975 period in the U.S. See Wolff (1979) for details.

23 The residual group, American Indians and Asians, is excluded here because of its small sample size.

24 It should be noted that the unit of observation is the household, which includes both families (two or more related individuals living together), as well as single adults. As is widely known, the share of female-headed households among African-Americans is much higher than that among whites. This difference partly accounts for the relatively lower income and wealth among African-American households.

25 The 1988 income figure for black households appears to be an outlier. The low income for blacks in that year probably reflects the small sample size for blacks (and Hispanics as well) and the survey-to-survey sample variability.

26 There was almost no change in the relative home ownership rates of the two groups – both experienced moderate losses – while the share of households with non-positive net worth actually increased more in relative terms for white households than black ones. Unfortunately, there are no data available to separate out actual declines in house prices for white, black, and Hispanic homeowners.

27 As with racial minorities, the sample size is relatively small for the oldest age group, so that the 9 percentage point increase in their homeownership rate from 2001 to 2004 may be due to sampling variation.

28 As in 2007, the principal source of debt was mortgage debt, which comprised 70 percent of the total debt for the youngest age group in 2010. However, educational loans now amounted to 15 percent of their total liabilities, up from 10 percent in 2007, and 40 percent of households in this age group had an outstanding student loan in 2010.

29 This pattern may also be partly a cohort effect since 401(k) plans and other defined contribution plans were not widely introduced into the workplace until after 1989.

Economics in Context

The Great Recession’s Impact on Wealth Inequality in the United States

Update April 2015

Note: This update specifically relates to Microeconomics in Context Chapter 10, Macroeconomics in Context Chapter 15, Principles of Economics in Context Chapters 11 and 30. For more information about the books, teaching materials, and research, see www.gdae.org

The Great Recession clearly reduced overall wealth in the United States. According to analysis by economist Edward Wolff, median net worth in the U.S. declined by 44% from 2007 to 2010. But how did the Great Recession, and subsequent recovery, impact the distribution of wealth? Wolff’s analysis provides the first detailed estimates of how the assets of different groups were affected by the financial crisis.

One of Wolff’s main findings is that during the Great Recession (2007-2010) wealth inequality increased significantly in the United States. During this period the share of all wealth held by the richest 20% rose from 85% to 89%, and the Gini coefficient based on wealth increased from 0.834 to 0.866. The assets of the middle class were particularly hard-hit by the Great Recession. The main reason is that middle- class households tend to have a disproportionately high share of their assets in the equity value of their homes. When median home prices declined by 24% during the Great Recession, many households saw the positive equity in their homes become a net liability as the market price of their homes fell below their remaining mortgage balance. As Wolff writes:

The key to understanding the plight of the middle class over the Great Recession was their high degree of leverage and the high concentration of assets in their home. The steep decline in median net worth between 2007 and 2010 was primarily due to the very high negative rate of return on net worth of the middle three wealth quintiles (-10.6 percent per year). This, in turn, was attributable to the precipitous fall in home prices and their very high degree of leverage. (p. 43)

Wolff also found that wealth inequality based on race and ethnicity increased during the Great Recession. In 2007 median wealth among black households stood at only 19% of median wealth among white households. By 2010 this percentage had fallen to 14%. The relative impact of the Great Recession on Hispanic households was even more severe, with their median net worth plummeting from 26% to 15% of white households. For both groups, the primary explanation again seems to be found in the fact that a disproportionate share of their net assets were reduced (or became negative) as a result of the decline in home values.

Surprisingly, median assets in the U.S. did not recover over the period 2010-2013, despite rising home values and stock market prices. Wolff concludes that this is a result of the high dissavings of the middle class. So while the good news is that the outstanding debt of middle-class households fell during this period, it appears many drew down their assets to pay debts and finance consumption. With wages stagnating for many households, they were able to maintain existing consumption levels and reduce debt only by selling off some of their assets.

Overall, there was little change in wealth inequality from 2010 to 2013, with the Gini coefficient increasing from 0.866 to 0.871. Wealth inequality on the basis of race also remained about the same over this time period. Wolff’s data indicate that black and Hispanic households had relatively high rates of return on their assets over 2010-2013, but that this was offset by particularly high dissavings rates.

As of 2013, median net worth among those in the top 1% stood at over $18 million, an increase of 82% compared to 1983, even after adjusting for inflation. For those in the middle wealth quintile, their median net assets fell from about $79,000 to $68,000 over 1983-2013. But poorer households fared even worse. Those in the bottom 40% saw their median worth fall from about positive $7,000 in 1983 to negative $11,000 in 2013.

Note:

The wealth data presented above are all median values. The Wolff paper also documents trends in mean wealth. While real median wealth in the U.S. fell from $78,000 to $64,000 between 1983 and 2013, real mean wealth increased from $304,000 to $509,000. Given that the median value decreased while the mean value increased, we can conclude that wealth inequality increased during this time period.

NBER WORKING PAPER SERIES HOUSEHOLD WEALTH TRENDS IN THE UNITED STATES, 1962-2013: WHAT HAPPENED OVER THE GREAT RECESSION?

Edward N. Wolff

Working Paper 20733 http://www.nber.org/papers/w20733

NATIONAL BUREAU OF ECONOMIC RESEARCH 1050 Massachusetts Avenue Cambridge, MA 02138 December 2014

I would like to express my appreciation for financial support provided by the Institute of New Economic thinking (INET). The views expressed herein are those of the author and do not necessarily reflect the views of the National Bureau of Economic Research.

NBER working papers are circulated for discussion and comment purposes. They have not been peerreviewed or been subject to the review by the NBER Board of Directors that accompanies official NBER publications.

© 2014 by Edward N. Wolff. All rights reserved. Short sections of text, not to exceed two paragraphs, may be quoted without explicit permission provided that full credit, including © notice, is given to the source.

Household Wealth Trends in the United States, 1962-2013: What Happened over the Great Recession? Edward N. Wolff NBER Working Paper No. 20733 December 2014 JEL No. D31,J15

ABSTRACT

Asset prices plunged between 2007 and 2010 but then rebounded from 2010 to 2013. The most telling finding is that median wealth plummeted by 44 percent over years 2007 to 2010, almost double the drop in housing prices. The inequality of net worth, after almost two decades of little movement, was also up sharply. Relative indebtedness expanded, particularly for the middle class, though the proximate causes were declining net worth and income rather than an increase in absolute indebtedness. The sharp fall in median net worth and the rise in overall wealth inequality over these years are traceable primarily to the high leverage of middle class families and the high share of homes in their portfolio. The racial and ethnic disparity in wealth also widened considerably. Households under age 45 saw their relative and absolute wealth declined sharply. Rather remarkably, there was virtually no change in median wealth from 2010 to 2013 despite the rebound in asset prices. The proximate cause was the high dissavings of the middle class, though their debt continued to fall. Wealth inequality and the racial and ethnic wealth gap also remained largely unchanged, though there was some recovery of net worth for young households.

Edward N. Wolff Department of Economics New York University 19 W. 4th Street, 6th Floor New York, NY 10012 and NBER edward.wolff@nyu.edu

1. Introduction

The paper focuses mainly on how the middle class fared in terms of wealth over years 2007 to 2010 during one of the sharpest declines in stock and real estate prices and over years 2010 to 2013 as asset prices recovered. The debt of the middle class exploded from 1983 to 2007, already creating a very fragile middle class in the United States. The main question is whether their position deteriorated even more over the “Great Recession.”

In particular, there are five specific issues addressed in the paper. (1) Did median household wealth continue to advance over time or did it fall, particularly from 2007 to 2013? (2) Did the inequality of household wealth rise over time, particularly during years 2007 to 2013? (3) Did the debt of the middle class increase over time, especially over the Great Recession? (4) What are the time trends in home ownership and home equity and what happened, in particular, from 2007 to 2013? (5) How did time trends in average wealth, household debt, the home ownership rate, and home equity vary by race, ethnic group, and age group?

The period covered is from 1962 to 2013. In particular, results will be provided for years 1962, 1969, 1983, 1989, 1992, 1995, 1998, 2001, 2004, 2007, 2009, 2010, and 2013. The choice of years is dictated by the availability of survey data on household wealth. By 2013, we will be able to see what the fall-out was from the financial crisis and associated recession and, in particular, which groups were impacted the most.

As discussed in the next section, trends in household wealth have a direct bearing on household well-being and should therefore be of general public interest. Moreover, since the election of Barack Obama in 2012, the fortunes of the middle class have generated a large amount of political interest and media interest as well.

The paper is organized as follows. The next section, Section 2, provides some historical background. Section 3 discusses the measurement of household wealth and describes the data sources used for this study. Section 4 presents results on time trends in median and average wealth holdings, Section 5 on changes in the concentration of household wealth, and Section 6 on the composition of household wealth. In Section 7, I provide an analysis of the effects of leverage on wealth movements over time, particularly in regard to how it impacted households during the Great Recession. Section 8 investigates changes in wealth holdings by race and ethnicity; and Section 9 reports on changes in the age-wealth profile. A summary of results and concluding remarks are provided in Section 10.

Previous work of mine (see Wolff, 1994, 1996, 1998, 2001, 2002, and 2011a), using the 1983, 1989, 1992, 1995, 1998, 2001, 2004, and 2007 Surveys of Consumer Finances, presented evidence of sharply increasing household wealth inequality between 1983 and 1989 followed by little change between 1989 and 2007. Both mean and median wealth holdings climbed briskly from 1983 to 2007. However, most of the wealth gains from 1983 to 2007 were concentrated among the richest 20 percent of households. Moreover, despite the buoyant economy over the 1990s and 2000s, overall indebtedness rose among American families, particularly those in the middle class. The wealth gap between African-American and white families was much the same in 2007 as in 1983. However, the relative wealth holdings of younger families (under age 45) fell somewhat between 1983 and 2007.

In this study, I look at wealth trends from 1962 to 2013. Asset prices plunged between 2007 and 2010 but then rebounded from 2010 to 2013. The most telling finding is that median wealth plummeted by 44 percent over years 2007 to 2010, almost double the drop in housing prices, and by 2010 was at its lowest level since 1969. The inequality of net worth, after almost two decades of little movement, was up sharply from 2007 to 2010. Relative indebtedness expanded from 2007 to 2010, particularly for the middle class, though the proximate causes were declining net worth and income rather than an increase in absolute indebtedness. In fact, the average debt of the middle class fell by 25 percent in real terms. The sharp fall in median net worth and the rise in overall wealth inequality from 2007 to 2010 are traceable primarily to the high leverage of middle class families and the high share of homes in their portfolio. The racial and ethnic disparity in wealth holdings widened considerably in the years between 2007 and 2010. Hispanics, in particular, got hammered by the Great Recession in terms of net worth and net equity in their homes. Young households (under age 45) also got pummeled by the Great Recession, as their relative and absolute wealth declined sharply from 2007 to 2010.

Rather remarkably, there was virtually no change in median (and mean) wealth from 2010 to 2013 despite the rebound in asset prices, presenting a new puzzle. The proximate cause was the high dissavings of the middle class. Relative indebtedness fell for the middle class as outstanding debt continued to drop. Wealth inequality and the racial and ethnic wealth gap also remained largely unchanged, though there was some recovery of the net worth of young households.

2. Historical Background

The last two decades have witnessed some remarkable events. Perhaps, most notable is the housing value cycle which first led to an explosion in home prices and then a collapse, affecting net worth and helping to precipitate the Great Recession, followed by a modest recovery. The median house price remained virtually the same in 2001 as in 1989 in real terms.1 However, the home ownership rate shot up from 62.8 percent in 1989 to 67.7 percent in 2001 according to data from the Survey of Consumer Finances (SCF). Then, 2001 saw a recession (albeit a short one). Despite this, house prices suddenly took off. The median sales price of existing one-family homes spurted by 17 percent nationwide. However, from 2004 to 2007 housing prices slowed, with the median price advancing only 1.7 percent. Over the years 2001 to 2007 housing prices gained 19 percent. The home ownership rate continued to expand, though at a somewhat slower rate, from 67.7 to 68.6 percent.

Then, the “Great Recession” and the associated financial crisis hit. The Great Recession “officially” began in December, 2007, and “officially” ended in June, 2009.2 Over this period, real GDP fell by 4.3 percent and then from the second quarter of 2009 to the second quarter of 2013 it gained 9.2 percent. The unemployment rate shot up from 4.4 percent in May of 2007 to a peak of 10.0 percent in October of 2009 but by February of 2014 it was down to 6.7 percent.3

One consequence was that asset prices plummeted. From 2007 to 2010, in particular, the median home price nose-dived by 24 percent, and the share of households owning their own home fell off, from 68.6 to 67.2 percent. This was followed by a partial recovery, with median house prices rising 7.8% through September 2013, though still way below their 2007 value. However, the homeownership rate continued to contract, falling to 65.1 percent.

The housing price bubble in the years leading up to 2007 was fueled in large part by a generous expansion of credit available for home purchases and re-financing. This took a number of forms. First, many home owners re-financed their primary mortgage. However, because of the rise in housing prices, these home owners increased the outstanding mortgage principal and thereby extracted equity from their homes. Second, many home owners took out second mortgages and home equity loans or increased the outstanding balances on these instruments. Third, among new home owners, credit requirements were softened, and so-called “no-doc” loans were issued requiring none or little in the way of income documentation. Many of these loans, in turn, were so-called “sub-prime” mortgages, characterized by excessively high interest rates and “balloon payments” at the expiration of the loan (that is, a non-zero amount due when the term of the loan was up). All told, average mortgage debt per household expanded by 59 percent in real terms between 2001 and 2007 according to the SCF data, and outstanding mortgage loans as a share of house value rose from 0.334 to 0.349, despite the 19 percent gain in real housing prices (see Table 4 below for more details).

In contrast to the housing market, the stock market boomed during the 1990s. On the basis of the Standard & Poor (S&P) 500 index, stock prices surged 159 percent between 1989 and 2001.4 Stock ownership spread and by 2001 over half of U.S. households owned stock either directly or indirectly. However, the stock market peaked in 2000 and dropped steeply from 2000 to 2003 but recovered somewhat in 2004, so that between 2001 and 2004 the S&P 500 was down by 11 percent.

From 2004 to 2007, the stock market rebounded, with the S&P 500 rising 19 percent. Over the period from 2001 to 2007, stock prices were up 6 percent. However, the share of households who owned stock either directly or indirectly fell to 49 percent. Then came the Great Recession. Stock prices, based on the S&P 500 index, crashed from 2007 to 2009 and then partially recovered in 2010 for a net decline of 26 percent. The stock ownership rate also once again declined, to 47 percent. The stock market continued to rise after 2010 and by 2013 the S&P 500 index was up 39 percent over 2010 and above its previous high in 2007. However, the stock ownership rate continued to drop, to 46 percent.

Real wages, after stagnating for many years, finally grew in the late 1990s. According to BLS figures, real mean hourly earnings gained 8.3 percent between 1995 and 2001.5 From 1989 to 2001, real wages rose by 4.9 percent (in total), and median household income in constant dollars grew by 6.0 percent (see Table 1). Employment also surged over these years, growing by 16.7 percent.6 The (civilian) unemployment rate remained relatively low over these years, at 5.3 percent in 1989, 4.7 percent in 2001, with a low point of 4.0 percent in 2000, and averaging 5.5 percent over these years.7

Real wages then rose very slowly from 2001 to 2004, with the BLS mean hourly earnings up by only 1.5 percent, and median household income dropped by 1.6 percent. From 2004 to 2007, real wages remained stagnant, with hourly earnings rising by only 1.0 percent. Median household income showed some growth over this period, rising by 3.2 percent. From 2001 to 2007 it gained 1.6 percent. Employment also grew more slowly over these years, gaining 6.7 percent. The unemployment rate remained low again, at 4.7 percent in 2001 and 4.6 percent in 2007 and an average value of 5.2 percent over the period.

Real wages, on the other hand, picked up from 2007 to 2010, increasing by 3.6 percent. In contrast, median household income in real terms declined sharply over this period, by 6.7 percent (see Table 1 below). Moreover, employment contracted over these years, by 4.8 percent, and the unemployment rate surged from 4.6 percent in 2007 to 10.5 percent in 2010. From 2010 to 2013 employment grew by 4.7 percent, and the unemployment rate did come down to an average rate of 7.4 percent in 2013.

There was also an explosion of consumer debt leading up to the Great Recession. Between 1989 and 2001, total consumer credit outstanding in constant dollars surged by 70 percent and then from 2001 to 2007 it rose by another 17 percent.8 There were a number of factors responsible for this. First credit cards became more generally available for consumers. Second, credit standards were relaxed considerably, making more households eligible for credit cards. Third, credit limits were generously increased by banks hoping to make profits out of increased fees from late payments and from higher interest rates.

Another source of new household indebtedness was from a huge increase in student loans. This issue has recently received wide attention in the press. According to the SCF data, the share of households reporting an educational loan rose from 13.4 percent in 2004 to 15.2 percent in 2007 and then surged to 19.1 percent in 2010 and 19.9 percent in 2013.9 The mean value of educational loans in 2013 dollars exclusively among loan holders increased by 17 percent between 2004 and 2007, another 14 percent between 2007 and 2010, and then by an additional 5 percent to $29,110 in 2013. The median value of such loans first went up by 19 percent from 2004 to 2007, by another 3 percent between 2007 and 2010, and then by an added 22 percent to $17,000 in 2013. These loans were heavily concentrated among younger households and, as we shall see below, was one of the factors (though not the principal one) which led to a precipitous decline in their net worth between 2007 to 2010.

Another major change in the 1990s and the decade of the 2000s affecting household wealth was a major overhaul of the private pension system in the United States. As documented in Wolff (2011b), in 1989, 46 percent of all households reported holding a defined benefit (DB) pension plan. DB plans are traditional pensions, such as provided by many large corporations, the federal government, and state and local governments, which guarantee a steady flow of income upon retirement. By 2007, that figure was down to 34 percent. The decline was more pronounced among younger households, under the age of 46, from 38 to 23 percent, as well as among middleaged households, ages 47 to 64, from 57 to 39 percent.

Many of these plans were replaced by so-called defined contribution (DC) pension accounts, most notably 401(k) plans and Individual Retirement accounts (IRAs). These plans allow household to accumulate savings for retirement purposes directly. The share of all households with a DC plan skyrocketed from 24 percent in 1989 to 53 percent in 2007. Among younger households, the share rose from 31 to 50 percent, and among middle-aged households it went from 28 to 64 percent.

This transformation is even more notable in terms of actual dollar values. While the average value of DB pension wealth among all households crept up by 8 percent from $56,500 in 1989 to $61,200 in 2007, the average value of DC plans shot up more than 7-fold from $10,600 to $76,800 (all figures are in 2007 dollars).10 Among younger households, average DB wealth actually fell in absolute terms, while DC wealth rose by a factor of 3.3. Among middle-aged households, the mean value of DB pensions also declined, while the average value of DC plans mushroomed by a factor of 6.5.

These changes are important for understanding trends in household wealth because DB pension wealth is not included in the standard measure of marketable household wealth whereas DC wealth is included (see Section 4 below). Thus, the substitution of DC wealth for DB wealth is likely to lead to an overstatement in the “true” gains in household wealth, since the displacement in DB wealth is not captured (see Wolff, 2011b, for more discussion).

The other big story was household debt, particularly that of the middle class, which skyrocketed during these years, as we shall see below. Despite the recession, the relative indebtedness of American families continued to rise from 2007 to 2010, though it did fall off from 2010 to 2013.

What have all these major transformations wrought in terms of the distribution of household wealth, particularly over the Great Recession? How have these changes impacted different demographic groups, particularly as defined by race, ethnicity, and age? This is the subject of the remainder of the paper.

3. Data sources and methods

The primary data sources used for this study are the 1983, 1989, 1992, 1995, 1998, 2001, 2004, 2007, 2010, and 2013 Survey of Consumer Finances (SCF) conducted by the Federal Reserve Board. Each survey consists of a core representative sample combined with a high-income supplement. In 1983, for example, the supplement was drawn from the Internal Revenue Service's Statistics of Income data file. For the 1983 SCF, an income cut-off of $100,000 of adjusted gross income was used as the criterion for inclusion in the supplemental sample. Individuals were randomly selected for the sample within pre-designated income strata. In later years, the high income supplement was selected as a list sample from statistical records (the Individual Tax File) derived from tax data by the Statistics of Income Division of the Internal Revenue Service (SOI). This second sample was designed to disproportionately select families that were likely to be relatively wealthy (see, for example, Kennickell, 2001, for a more extended discussion of the design of the list sample in the 2001 SCF). Typically, about two thirds of the cases come from the representative sample and one third from the high-income supplement.

The principal wealth concept used here is marketable wealth (or net worth), which is defined as the current value of all marketable or fungible assets less the current value of debts. Total assets are the sum of: (1) housing; (2) other real estate owned by the household; (3) bank deposits, certificates of deposit, and money market accounts; (4) financial securities; (5) the cash surrender value of life insurance plans; (6) the value of pension plans, including IRAs, Keogh, and 401(k) plans; (7) corporate stock and mutual funds; (8) net equity in unincorporated businesses; and (9) equity in trust funds. Total liabilities are the sum of: (1) mortgage debt, (2) consumer debt, including auto loans, and (3) other debt such as educational loans.

This measure reflects wealth as a store of value and therefore a source of potential consumption. I believe that this is the concept that best reflects the level of well-being associated with a family's holdings. Thus, only assets that can be readily converted to cash (that is, "fungible" ones) are included. As a result, consumer durables such as automobiles, televisions, furniture, household appliances, and the like, are excluded here, since these items are not easily marketed, with the possible exception of vehicles, or their resale value typically far understates the value of their consumption services to the household. Another justification for their exclusion is that this treatment is consistent with the national accounts, where purchase of vehicles is counted as expenditures, not savings. As a result, my estimates of household wealth will differ from those provided by the Federal Reserve Board, which includes the value of vehicles in their standard definition of household wealth (see, for example, Kennickell and Woodburn, 1999).

Also excluded is the value of future Social Security benefits the family may receive upon retirement (usually referred to as "Social Security wealth"), as well as the value of retirement benefits from private pension plans ("pension wealth"). Even though these funds are a source of future income to families, they are not in their direct control and cannot be marketed.11

Two other data sources are used in the study. The first of these is the 1962 Survey of Financial Characteristics of Consumers (SFCC). This survey was also conducted by the Federal Reserve Board of Washington and is a precursor to the SCF. This is also a stratified sample which over-samples high income households. Though the sample design and questionnaire are different from the SCF, the methodology is sufficiently similar to allow comparisons with the SCF data (see Wolff, 1987, for details on the adjustments). The second is the so-called 1969 MESP database, a synthetic dataset constructed from income tax returns and information provided in the 1970 Census of Population. A statistical matching technique was employed to assign income tax returns for 1969 to households in the 1970 Census of Population. Property income flows (such as dividends) in the tax data were then capitalized into corresponding asset values (such as stocks) to obtain estimates of household wealth (see Wolff, 1980, for details).

4. Median wealth plummets over the Great Recession

Table 1 documents a robust growth in wealth from 1983 to 2007, even back to 1962 (also see Figure 1). Median wealth increased at an annual rate of 1.6 percent from 1962 to 1983, then slower at 1.1 percent from 1983 to 1989, about the same at 1.2 percent from 1989 to 2001, and then much faster at 2.9 percent from 2001 to 2007.12 Then between 2007 and 2010, median wealth plunged by a staggering 44 percent! Indeed, median wealth was actually lower in 2010 than in 1969 (in real terms). The primary reasons, as we shall see below, were the collapse in the housing market and the high leverage of middle class families. There was virtually no change from 2010 to 2013.13

As shown in the third row of Panel A, the percentage of households with zero or negative net worth, after falling from 18.2 percent in 1962 to 15.5 percent in 1983, increased to 17.9 percent in 1989 and 18.6 percent in 2007. However, this was followed by a sharp rise to 21.8 percent in 2010, at which level it remained in 2013.

Mean net worth also grew vigorously from 1962 to 1983, at an annual rate of 1.82 percent, a little higher than that of median wealth. Its growth accelerated to 2.27 percent per year over years 1983 to 1989, about double the growth rate of median wealth. Over the years 1989 to 2001, the growth rate of mean wealth was 3.02 percent per year, even higher than in the preceding periods. Its annual growth rate accelerated even more, reaching 3.10 percent between years 2001 and 2007. This acceleration was due largely to the rapid (19 percent) increase in housing prices over the six years counterbalanced by a reduced growth in stock prices in comparison to years 1989 to 2001, and to the fact that housing comprised 28 percent and (total) stocks made up 25 percent of total assets in 2001. Overall, its 2007 value was almost double its value in 1983 and about three quarters larger than in 1989. Another point of note is that mean wealth grew more about twice as fast as the median between 1983 and 2007, indicating widening inequality of wealth over these years.

The great Recession also saw an absolute decline in mean household wealth. However, whereas median wealth plunged by 44 percent between 2007 and 2010, mean wealth fell by (only) 16 percent.14 In this case, the main cause was both falling housing and stock prices (see below). However, here, too, the relatively faster growth in mean wealth than median wealth (that is, the latter’s more moderate decline) was coincident with rising wealth inequality. There was again virtually no change in mean wealth from 2010 to 2013.

Median household income (based on Current Population Survey data) advanced at a fairly solid pace from 1962 to 1983, at 0.61 percent per year (also see Figure 2). Then, after gaining 2.03 percent per annum between 1983 and 1989, its annual growth dipped to only 0.48 percent from 1989 to 2001 and then to 0.26 percent from 2001 to 2007, for a net change of 22 percent (overall) from 1983 to 2007. However, from 2007 to 2010, it fell off in absolute terms by 6.7 percent. Though this is not an insignificant amount, the reduction was not nearly as great as that in median wealth. From 2010 to 2013, median income slipped by another 1.3 percent (overall).

In contrast, mean income, after advancing at an annual rate of 0.93 percent from 1962 to 1983, gained 2.66 percent per year from 1983 to 1989, 1.21 percent per year from 1989 to 2001, and then -0.14 percent per year from 2001 to 2007, for a total change of 35 percent from 1983 to 2007. Between 1983 and 2007, mean income grew less than mean net worth, and median income grew at a much slower pace than median wealth. However, mean income also dropped in real terms from 2007 to 2010, by 5.2 percent, slightly less than that of median income, but gained 0.9 percent from 2010 to 2013.

In sum, while household income virtually stagnated for the average American household from 1989 to 2007, median net worth grew strongly. The Great Recession, on the other hand, saw a massive reduction in median net worth but much more modest declines in mean wealth and both median and mean income.

5. Wealth inequality jumps in the late 2000s

Net worth is highly concentrated, with the richest 1 percent (as ranked by wealth) owning 37 percent of total household wealth in 2013 and the top 20 percent owning 89 percent (see Table 2). The figures in Table 2 also show that wealth inequality in 1983 was quite close to its level in 1962 (also see Figure 3).15 Then, after rising steeply between 1983 and 1989, it remained virtually unchanged from 1989 to 2007. The share of wealth held by the top 1 percent rose by 3.6 percentage points from 1983 to 1989 and the Gini coefficient increased from 0.80 to 0.83.16

Between 1989 and 2007, the share of the top percentile actually declined sharply, from 37.4 to 34.6 percent, though this was more than compensated for by an increase in the share of the next four percentiles.17 As a result, the share of the top five percent increased from 58.9 percent in 1989 to 61.8 percent in 2007, and the share of the top quintile rose from 83.5 to 85.0 percent. The share of the fourth and middle quintiles each declined by about a percentage point from 1989 to 2007, while that of the bottom 40 percent increased by almost one percentage point. Overall, the Gini coefficient was virtually unchanged -- 0.832 in 1989 and 0.834 in 2007.

In contrast, the years 2007 to 2010 saw a very sharp elevation in wealth inequality, with the Gini coefficient rising from 0.834 to 0.866. Interestingly, the share of the top percentile showed a smaller relative gain -- less than a one percentage point gain. Most of the rise in wealth share took place in the remainder of the top quintile, and overall the share of wealth held by the top quintile climbed by almost four percentage points. The shares of the other quintiles correspondingly dropped, with that of the bottom 40 percent falling from 0.2 to -0.9 percent.

From 2010 to 2013 there was a very small rise in the Gini coefficient, from 0.866 to 0.871. The share of the top one percent did increase by 1.6 percentage points but there was virtually no change in the share of the top quintile. In constant dollar terms, the net worth of the top one percent grew by 5.9 percent over those years but that of the next 19 percent was down by 1.8 percent. The wealth of the fourth quintile also lost 1.7 percent, that of the middle quintile fell 0.7 percent, and that of the bottom forty percent declined 5.7 percent.

The top 1 percent of families (as ranked by income on the basis of the SCF data) earned 20 percent of total household income in 2012 and the top 20 percent accounted for 62 percent -- large figures but lower than the corresponding wealth shares.18 The time trend for income inequality also contrasts with that for net worth (also see Figure 3). Income inequality showed a sharp rise from 1961 to 1982, with the Gini coefficient expanding from 0.428 to 0.480 and the share of the top one percent from 8.4 to 12.8 percent. Income inequality increased sharply again between 1982 and 1988, with the Gini coefficient rising from 0.48 to 0.52 and the share of the top one percent from 12.8 to 16.6 percent.

Inequality again surged from 1988 to 2000, with the share of the top percentile rising by 3.4 percentage points, the share of the top quintile up by 3.0 percentage points, the shares of the other quintiles falling again, and the Gini index advancing from 0.52 to 0.56. As a result, the years from 1989 to 2001 saw almost the same degree of increase in income inequality as the 1983-1989 period.19 Inequality once again rose from 2001 to 2007, though the pace slackened. The Gini coefficient increased from 0.562 to 0.574, the share of the top one percent was up by 1.3 percentage points, the share of the top quintile was also up by 1.7 percentage points, and the shares of the other quintiles fell. All in all, the period from 2001 to 2007 witnessed a moderate increase in income inequality and a small rise in wealth inequality.

Perhaps, somewhat surprisingly, the years 2007 to 2010 witnessed a rather sharp contraction in income inequality. The Gini coefficient fell from 0.574 to 0.549 and the share of the top one percent dropped sharply from 21.3 to 17.2 percent. Property income and realized capital gains (which is included in the SCF definition of income), as well as corporate bonuses and the value of stock options, plummeted over these years, a process which explains the steep decline in the share of the top percentile. Real wages, as noted above, actually rose over these years, though the unemployment rate also increased. As a result, the income of the middle class was down but not nearly as much in percentage terms as that of the high income groups. In contrast, transfer income such as unemployment insurance rose, so that the bottom also did better in relative terms than the top. As a result, overall income inequality fell over the years 2006 to 2009.20

The second half of the Great Recession saw a reversal in this trend, with income inequality once again increasing sharply. The Gini coefficient increased by 0.025 points to 0.574, the same level as in 2007. The share of the top percentile rose to 19.8 percent, slightly below its level in 2007, while the share of the top quintile was up to 61.8 percent, slightly above its level in 2007. The same set of factors, though in reverse, help explain this turnaround in income inequality. Property income, realized capital gains, and associated income rose sharply over these years as the stock market recovered, accounting for the sharp rise in the share of the top percentile. The unemployment rate fell over these years but real wages were down, according to the BLS figures. As a result, the income of the middle class rose but not nearly as much in percentage terms as that of the high income groups. Transfer income such as unemployment insurance fell, as the extensions of benefits enacted in the early days of the recession ended.

All in all, income inequality increased much more than either net worth or non-home wealth inequality over years 1983 to 2013. On the basis of the Gini coefficient, net worth inequality was up by nine percent, while income inequality rose by 20 percent.

As a result, one of the puzzles we have to contend with is the fact that net worth inequality rose sharply over the first part of the Great Recession while income inequality fell sharply, at least according to the SCF data. A second is the reverse, namely that wealth inequality remained virtually unchanged over the second half of the recession while income inequality increased sharply over these years. A third is that income inequality increased much more than net worth inequality over years 1983 to 2013. I will return to these questions in Section 8 below.

5.1 The share of overall wealth gains, 1983 to 2013

Table 3 shows the absolute changes in wealth and income between 1983 and 2013. The results are even more striking. Over this period, the largest gains in relative terms were made by the wealthiest households. The top one percent saw their average wealth (in 2013 dollars) rise by over eight million dollars or by 82 percent. The remaining part of the top quintile experienced increases from 52 to 110 percent and the fourth quintile by 24 percent, while the middle quintile lost 14 percent and the average wealth of the poorest 40 percent fell by $17,500. By 2013, the average wealth of the bottom 40 percent had fallen to -$10,800.

Another way of viewing this phenomenon is afforded by calculating the proportion of the total increase in real household wealth between 1983 and 2013 accruing to different wealth groups. This is computed by dividing the increase in total wealth of each percentile group by the total increase in household wealth, while holding constant the number of households in that group. If a group's wealth share remains constant over time, then the percentage of the total wealth growth received by that group will equal its share of total wealth. If a group's share of total wealth increases (decreases) over time, then it will receive a percentage of the total wealth gain greater (less) than its share in either year. However, it should be noted that in these calculations, the households found in each group (say the top quintile) may be different in the two years.

The results indicate that the richest one percent received 41 percent of the total gain in marketable wealth over the period from 1983 to 2013. This proportion was greater than the share of wealth held by the top one percent in any of the intervening 10 years. The next 4 percent received 36 percent of the total gain and the next 15 percent 22 percent, so that the top quintile collectively accounted for almost 100 percent of the total growth in wealth, while the bottom 80 percent accounted for virtually none.

A similar calculation using the SCF income data reveals that the greatest gains in real income over the period from 1982 to 2012 were made by households in the top one percent of the income distribution, who saw their incomes grow by 90 percent. Mean incomes increased by over half for the next 4 percent, over a quarter for the next highest 5 percent and by 14 percent for the next highest ten percent. The fourth quintile of the income distribution experienced only a 3 percent growth in income, while the middle quintile and the bottom 40 percent had absolute declines in mean income. Of the total growth in real income between 1982 and 2012, almost half accrued to the top one percent and over 100 percent to the top quintile. These figures are very close to those for net worth.

These results indicate rather dramatically that the despite the Great Recession and its aftermath, growth in the economy during the period from 1983 to 2013 was concentrated in a surprisingly small part of the population -- the top 20 percent and particularly the top one percent.

6. Household debt finally recedes

In 2013, owner-occupied housing was the most important household asset in the average portfolio breakdown for all households shown in Table 4, accounting for 29 percent of total assets (also see Figure 4). However, net home equity -- the value of the house minus any outstanding mortgage -- amounted to only 17 percent of total assets. Real estate, other than owner-occupied housing, comprised 10 percent, and business equity another 18 percent.

Demand deposits, time deposits, money market funds, CDs, and the cash surrender value of life insurance (collectively, “liquid assets”) made up 8 percent and pension accounts 17 percent. Bonds and other financial securities amounted to 2 percent; corporate stock, including mutual funds, to 13 percent; and trust fund equity to 3 percent. Debt as a proportion of gross assets was 15 percent, and the debt-equity ratio (the ratio of total household debt to net worth) was 0.18.

There were some significant changes in the composition of household wealth over time. First, the share of housing wealth in total assets, after fluctuating between 28 and 30 percent from 1983 to 2001, jumped to 34 percent in 2004 and then declined to 29 percent in 2013. Two factors explain this movement. The first is the trend in the homeownership rate. According to the SCF data, the homeownership rate rose from 63 percent in 1983 to 69 percent in 2004 and then fell off to 65 percent in 2013. The second is the trend in housing prices. The median house price for existing one-family homes rose by 18 percent between 2001 and 2004 but plunged by 17 percent from 2004 to 2013.21

A second and related trend is that net equity in owner-occupied housing (the difference between the market value and outstanding mortgages on the property), after falling almost continuously from 24 percent in 1983 to 18 percent in 1998, picked up to 22 percent in 2004 but then fell again to 17 percent in 2013. The difference between the two series (gross versus net housing values as a share of total assets) is attributable to the changing magnitude of mortgage debt on homeowner's property, which increased from 21 percent in 1983 to 37 percent in 1998, fell back to 35 percent in 2004, and then rose again to 39 percent in 2013. Moreover, mortgage debt on principal residence climbed from 9.4 of total assets in 2001 to 12.7 percent in 2010 before receding to 11.2 percent in 2013. The increase in net home equity as a proportion of assets between 2001 and 2004 reflected the strong gains in real estate values over these years, while its sharp decline from 2007 to 2013 reflected the steep fall in housing prices over those years.

Third, relative indebtedness first increased, with the debt-equity ratio climbing from 15 percent in 1983 to 21 percent in 2010, and then fell off to 18 percent in 2013. Likewise, the debtincome ratio surged from 68 percent in 1983 to 127 percent in 2010 but then dropped to 107 percent in 2013. If mortgage debt on principal residence is excluded, then the ratio of other debt to total assets actually fell off over time from 6.8 percent in 1983 to 4.0 percent in 2013.

The large rise in relative indebtedness among all households between 2007 and 2010 could be due to a rise in the absolute level of debt and/or a fall off in net worth and income. As shown in Table 1, both mean net worth and mean income fell over the three years. There was also a slight contraction of debt in constant dollars, with mortgage debt declining by 5.0 percent, other debt by 2.6 percent, and total debt by 4.4 percent. Thus, the steep rise in the debt to equity and the debt to income ratio over the three years was entirely due to the reduction in wealth and income. In contrast, from 2010 to 2013, relative indebtedness declined. In this case, both net worth and incomes were relatively unchanged, so that the proximate cause was a sizeable reduction in household debt. In fact, average mortgage debt (in constant dollars) dropped by 13 percent, the average value of other debt by 11 percent, and average household debt by 13 percent.

A fourth change is that pension accounts rose from 1.5 to 16.5 percent of total assets from 1983 to 2013. This increase largely offset the decline in the share of liquid assets in total assets, from 17.4 to 7.6 percent, so that it is reasonable to infer that households to a large extent substituted tax-deferred pension accounts for taxable savings deposits.

Fifth, other (non-home) real estate fell from 15 percent of total assets in 1983 to 10 percent in 2013, financial securities declined from 4.2 to 1.5 percent of total assets, and unincorporated business equity held more or less steady over time at around 18 percent. Stocks and mutual funds rose from 9 to 13 percent of gross assets over these years. Its year to year trend mainly reflects fluctuations in the stock market. If we include the value of stocks indirectly owned through mutual funds, trusts, IRAs, 401(k) plans, and other retirement accounts, then the value of total stocks owned as a share of total assets more than doubled from 11.3 percent in 1983 to 24.5 percent in 2001, and then tumbled to 17.5 percent in 2010, before rising to 20.7 percent in 2013. The rise during the 1990s reflected the bull market in corporate equities as well as increased stock ownership, while the decline in the 2000s was a result of the sluggish stock market as well as a drop in stock ownership. The increase from 2010 to 2013 reflected the recovery of the stock market.

6.1. Portfolio composition by wealth class

The tabulation in Table 4 provides a picture of the average holdings of all families in the economy, but there are marked class differences in how middle-class families and the rich invest their wealth. As shown in Table 5, the richest one percent of households (as ranked by wealth) invested almost three quarters of their savings in investment real estate, businesses, corporate stock, and financial securities in 2013 (also see Figure 5). Corporate stocks, either directly owned by the households or indirectly owned through mutual funds, trust accounts, or various pension accounts, comprised 25 percent by themselves. Housing accounted for only 9 percent of their wealth (and net equity in housing 7 percent), liquid assets 6 percent, and pension accounts another 9 percent. Their ratio of debt to net worth was only 3 percent, their ratio of debt to income was 38 percent, and the ratio of mortgage debt to house value was 17 percent.

Among the next richest 19 percent, housing comprised 28 percent of their total assets (and net home equity 20 percent), liquid assets 8 percent, and pension assets another 22 percent. Investment assets -- real estate, business equity, stocks, and bonds – made up 41 percent and 23 percent was in the form of stocks directly or indirectly owned. Debt amounted to 12 percent of their net worth and 97 percent of their income, and the ratio of mortgage debt to house value was 30 percent.

In contrast, over three-fifths of the assets of the middle three quintiles of households was invested in their own home in 2013. However, home equity amounted to only 31 percent of total assets, a reflection of their large mortgage debt. Another quarter went into monetary savings of one form or another and pension accounts. Together housing, liquid assets, and pension assets accounted for 87 percent of the total assets of the middle class. The remainder was about evenly split among non-home real estate, business equity, and various financial securities and corporate stock. Stocks directly or indirectly owned amounted to only 10 percent of their total assets. The ratio of debt to net worth was 64 percent, substantially higher than for the richest 20 percent, and their ratio of debt to income was 125 percent, also much higher than that of the top quintile. Finally, their mortgage debt amounted to about half the value of their principal residences.

Almost all households among the top 20 percent of wealth holders owned their own home, in comparison to 67 percent of households in the middle three quintiles. Three-quarters of very rich households (in the top percentile) owned some other form of real estate, compared to 44 percent of rich households (those in the next 19 percent of the distribution) and only 12 percent of households in the middle 60 percent. Eighty-nine percent of the very rich owned some form of pension asset, compared to 84 percent of the rich and 44 percent of the middle. A somewhat startling 77 percent of the very rich reported owning their own business. The comparable figures are 26 percent among the rich and only 7 percent of the middle class.

Among the very rich, 84 percent held corporate stock, mutual funds, financial securities or a trust fund, in comparison to 60 percent of the rich and only 14 percent of the middle class. Ninety-four percent of the very rich reported owning stock either directly or indirectly, compared to 85 percent of the rich and 41 percent of the middle. If we exclude small holdings of stock, then the ownership rates drop off sharply among the middle three quintiles, from 41 percent to 30 percent for stocks worth $5,000 or more and to 25 percent for stocks worth $10,000 or more.

Table 6 looks at trends in the wealth composition of the middle three wealth quintiles as well as asset ownership rates. Perhaps, the most striking development is with regard to the homeownership rate. After rising almost continuously over time from 72 percent in 1983 to 78 percent in 2004, it plunged by 11 percentage points to 67 percent in 2013. This trend was more pronounced than that among all households, among whom the homeownership rate dropped from 69 percent in 2004 to 65 percent in 2013. A similar trend is evident for the share of home values in total assets. It remained virtually unchanged from 1983 to 2001 but then rose sharply in 2004. This increase was largely a result of rising house prices and secondarily a consequence of the continued gain in the homeownership rate. The share then declined from 2004 through 2013 as housing prices fell and the homeownership rate plummeted.

It might seem surprising that despite the steep drop in home prices from 2007 to 2010, housing as a share of total assets actually fell only slightly. The reason is that the other components of wealth fell even more than housing. While the mean value of housing among households in the middle three quintiles fell by 31 percent in real terms, the mean value of other real estate was down by 39 percent and that of stocks and mutual funds fell by 47 percent.

Likewise, despite the modest recovery in housing prices from 2010 to 2013, the share of housing in total assets dropped by 2.3 percentage points. The mean value of housing fell by 7.3 percent. Of this, the decline in the homeownership rate accounted for only 19 percent of the overall decline, while the main culprit was the decline in the mean values of houses, which explained 81 percent. This result seems contrary to the finding that the median value of existing homes rose by 8 percent in real terms according to data from the National Association of Realtors (see footnote 1 for the reference). The most likely reason for the difference in results is that the 8 percent figure is based on data for existing homes only whereas the SCF data includes the value of homes that were owned by the household prior to the current year as well as newly bought homes. Another difference is that the former include all families whereas my figure is based on households in the middle three wealth quintiles. In fact, according to the SCF data, the median value of homes among middle class households was down by 14 percent in real terms from 2010 to 2013. This result, in turn, may be due to the fact that the new homes bought by families in the SCF sample were cheaper than existing homes.

The share of pension accounts in total assets rose by 15 percentage points from 1983 to 2013, while that of liquid assets declined by 13 percentage points. This trend was more or less continuous over time. This set of changes paralleled that of all households. In contrast, the share of middle class households holding a pension account, after surging by 41 percentage points from 12 percent in 1983 to 53 percent in 2007, collapsed by 7.6 percentage points to 46 percent in 2010 and then declined further to 44 percent in 2013. From 2007 to 2010 the mean value of pension accounts fell quite sharply, by 25 percent, though this was less than that of average overall assets, so that the share of pension accounts in total assets rose. From 2010 to 2013, in contrast, mean pension accounts were up by 12 percent, despite the slight decline in the ownership rate, so that the share of pension accounts in total assets strengthened considerably (by 2.2 percentage points).

The share of all stocks in total assets mushroomed from 2.4 percent in 1983 to 12.6 percent in 2001 and then fell off to 8.1 percent in 2010 as stock prices stagnated and then collapsed and middle class households divested themselves of stock holdings. The proportion then rebounded to 9.5 percent in 2013 as the stock market recovered. The stock ownership rate among the middle class also shot up quickly from 17 percent in 1983 to 51 percent in 2001, when it peaked. It then declined steeply to 41 percent in 2010, where it remained in 2013. In similar fashion, the share of middle class households owning either corporate stock, financial securities, mutual funds or a personal trust rose from 22 percent in 1983 to 28 percent in 2001 and then collapsed almost by half to 14 percent in 2013. Much of the decline took place between 2007 and 2010, as middle class households got scared off by the stock market collapse of those years.

6.2 Middle Class Debt

The rather staggering debt level of the middle class in 2013 raises the question of whether this is a recent phenomenon or whether it has been going on for some time. The debt-income ratio peaked in 2010 and then receded in 2013, while the debt-equity ratio peaked in 2007 and then contracted substantially in 2010 and a bit more in 2013.

There was a sharp rise in the debt-equity ratio of the middle class from 37 percent in 1983 to 61 percent in 2007. There was a particularly steep rise between 2001 and 2004, a reflection mainly of a steep rise in mortgage debt. The debt to income ratio skyrocketed from 1983 to 2007, more than doubling. Once, again, much of the increase happened between 2001 and 2004. In constant dollar terms, the mean debt of the middle class shot up by a factor of 2.6 between 1983 and 2007, the mean mortgage debt by a factor of 3.2, and the average value of other debt by a factor of 1.5. The rise in the debt-equity ratio and the debt-income ratio was much steeper than those for all households. In 1983, for example, the debt to income ratio was about the same for middle class as for all households but by 2007 the ratio was much larger for the middle class.

Then, the Great Recession hit. The debt-equity ratio continued to rise, reaching 72 percent in 2010 but there was actually a retrenchment in the debt to income ratio, falling to 134 percent in 2010. The reason is that from 2007 to 2010, the mean debt of the middle class actually contracted by 25 percent in constant dollars. Average mortgage debt declined by 23 percent, as families paid down their outstanding balances, while the mean value of other debt plummeted by 32 percent, as families paid off credit card balances and other forms of consumer debt. The significant rise in the debt-equity ratio of the middle class between 2007 and 2010 was due to the steeper drop off in net worth than in debt, while the decline in the debt-income ratio was exclusively due to the sharp contraction of overall debt.

Both the debt-equity and the debt-income ratios fell from 2010 to 2013. The proximate cause was a decline in overall mean debt, which fell by 8.2 percent in real terms over these years. This, in turn, was due to a decline in average mortgage debt, which dropped by 10.4 percent. The average balance on other debt actually increased slightly, by 1.6 percent.

As for all households, net home equity as a percentage of total assets fell for the middle class from 1983 to 2013 and mortgage debt as a proportion of house value rose. The decline in the former between 2007 and 2010 was relatively small despite the steep decrease in home prices, a reflection of the sharp reduction in mortgage debt. There was virtually no change from 2010 to 2013. On the other hand, the rise in the ratio of mortgage debt to house values was relatively large over years 2007 to 2010 because of the fall off in home prices. This ratio actually contracted somewhat from 2010 to 2013 as outstanding mortgage debt fell.

6.3 Concentration of assets by asset type

Another way to portray differences between middle class households and the rich is to compute the share of total assets of different types held by each group (see Table 7). In 2013 the richest one percent of households held about half of all outstanding stock, financial securities, trust equity, and business equity, and a third of non-home real estate. The top 10 percent of families as a group accounted for about 85 to 90 percent of stock shares, bonds, trusts, and business equity, and over three quarters of non-home real estate. Moreover, despite the fact that 46 percent of households owned stock shares either directly or indirectly through mutual funds, trusts, or various pension accounts, the richest 10 percent of households accounted for 81 percent of the total value of these stocks, though less than its 91 percent share of directly owned stocks and mutual funds.

In contrast, owner-occupied housing, deposits, life insurance, and pension accounts were more evenly distributed among households. The bottom 90 percent of households accounted for 59 percent of the value of owner-occupied housing, 33 percent of deposits, 35 percent of life insurance cash value, and 35 percent of the value of pension accounts. Debt was the most evenly distributed component of household wealth, with the bottom 90 percent of households responsible for 74 percent of total indebtedness.

The concentration of asset ownership by asset type remained remarkably stable over the three decades despite the dramatic changes in the economy over this time period discussed in Section 2. However, there were three exceptions. First, the share of total stocks and mutual funds held by the richest 10 percent of households declined from 90 to 85 percent from 1983 to 2004 but then rose back to 91 percent in 2013, while their share of stocks directly or indirectly owned fell from 90 percent in 1983 to 77 percent in 2001 but then rose to 81 percent in 2013. Second, the proportion of total pension accounts held by the top 10 percent fell from 68 percent in 1983 to 51 percent in 1989, reflecting the growing use of IRAs by middle income families, and then rebounded to 65 percent in 2013 from the expansion of 401(k) plans and their adoption by high income earners. Third, the share of total debt held by the top 10 percent declined from 32 to 27 percent between 1983 and 2013.

6.4. The “middle class squeeze”

Nowhere is the middle class squeeze more vividly demonstrated than in their rising debt. As noted above, the ratio of debt to net worth of the middle three wealth quintiles rose from 37 percent in 1983 to 46 percent in 2001 and then jumped to 61 percent in 2007. Correspondingly, their debt to income rose from 67 percent in 1983 to 100 percent in 2001 and then zoomed up to 157 percent in 2007! This new debt took two major forms. First, because housing prices went up over these years, families were able to borrow against the now enhanced value of their homes by refinancing their mortgages and by taking out home equity loans (lines of credit secured by their home). In fact, mortgage debt on owner-occupied housing (principal residence only) as a proportion of total assets climbed from 29 percent in 1983 to 47 percent in 2007, and home equity as a share of total assets fell from 44 to 35 percent over these years. Second, because of their increased availability, families ran up huge debt on their credit cards.

Where did the borrowing go? Some have asserted that it went to invest in stocks. However, if this were the case, then stocks as a share of total assets would have increased over this period, which it did not (it fell from 13 to 7 percent between 2001 and 2007). Moreover, they did not go into other assets. In fact, the rise in housing prices almost fully explains the increase in the net worth of the middle class from 2001 to 2007. Of the $16,400 rise in median wealth, gains in housing prices alone accounted for $14,000 or 86 percent of the growth in wealth. Instead, it appears that middle class households, experiencing stagnating incomes, expanded their debt in order to finance normal consumption expenditures.

The large build-up of debt set the stage for the financial crisis of 2007 and the ensuing Great Recession. When the housing market collapsed in 2007, many households found themselves “underwater,” with larger mortgage debt than the value of their home. This factor, coupled with the loss of income emanating from the recession, led many home owners to stop paying off their mortgage debt. The resulting foreclosures led, in turn, to steep reductions in the value of mortgage-backed securities. Banks and other financial institutions holding such assets experienced a large decline in their equity, which touched off the financial crisis.

6.5. The housing market

It is perhaps no surprise that the housing sector took an especially large hit in the financial crisis — the prime culprits in this crisis were the mortgage industry and the creation of faulty financial instruments by the financial sector that were tied to the fate of the housing market. The housing bubble in the early part of the last decade, which artificially inflated home prices to unprecedented levels, certainly set the stage for a major market ‘correction’. Indeed, as noted in Section 2 above, from 2007 to 2010, the median price of existing homes plummeted by 24 percent in real terms. Because housing makes up over 30 percent of total assets for all households and over 60 percent of the assets for the middle class, any economic downturn that affects the housing market will naturally hurt the wealth of the middle class.

As discussed above, the overall home ownership rate declined from 68.6 percent in 2007 to 67.2 percent in 2010 according to the SCF data, for a drop of 1.4 percentage points (see Table 8). This seems pretty modest, given all the media hype about home foreclosures over these years. Percentage point reductions were sharper for African-American and Hispanic households (1.9 percentage points) than for white households (almost no change); for single males (2.6 percentage points) than for married couples or single females (actually a net increase); for high school graduates (4.3 percentage points) than other educational groups; younger age groups in comparison to age group 75 and over (a large net increase); and for households with annual incomes below $25,000 and, surprisingly, above $75,000 than for middle income households.

The collapse in home values led to a surprisingly modest uptick in the number of families “underwater,” or with negative home equity. In 2007, only 1.8 percent of homeowners reported that their net home equity was negative on the basis of the 2007 SCF. By 2010, according to the SCF data, 8.2 percent of homeowners were “underwater.” As discussed above, though housing prices dropped by 24 percent in real terms from 2007 to 2010, there was also a substantial retrenchment of mortgage debt, which accounts for the relatively small share of home owners underwater in 2010.

Normally, we might think that the poorest households had the greatest incidence of being underwater but this was not always the case. Minorities did have a somewhat higher incidence of negative home equity than (non-Hispanic) whites but the differences were quite small. Somewhat surprisingly, single females, the poorest of the three family types, and single males had a somewhat lower incidence of negative home equity among homeowners than married couples. The reason for this is likely the lower mortgage debt of single females and single males (that is, they had less expensive houses to begin with). Also, again somewhat surprisingly, the lowest educational group, those with less than 12 years of schooling, had the smallest incidence of negative home equity among their homeowners, only 5 percent.22 In contrast, the incidence ranged from 8 percent to 11 percent among high school graduates, those with some college, and college graduates.

The age pattern is more consistent with expectations. Homeowners in the youngest age group, under age 35, had by far the highest incidence of negative home equity, 16 percent. The incidence of negative home equity declined almost directly with age, reaching only 3 percent for the oldest age group, 75 years and older. This reflects the fact that mortgages are generally paid off as people age. Moreover, the overall ratio of debt to net worth also declined directly with age (see Table 15).

However, the pattern by income class is again unexpected. The overall pattern is U-shaped, with the lowest incidence of negative home equity being for the lowest income class (under $15,000 of annual income) and the highest income class ($250,000 or more). The incidence of negative home equity among homeowners peaked at the $50,000 to $75,000 income class. Thus, the middle class was hit hardest by the collapse in housing prices. The reason is that they took out much higher mortgage debt, through re-financing, secondary mortgages, and home equity lines of credit, relative to their home values than the poor or the rich (see Table 5 above).

I also show the percentage decline in the average value of home equity among home owners from 2007 to 2010. For all home owners, the average decline was 29 percent (in constant dollars). This, again, is a surprisingly low figure given the 24 percent decline in real housing prices. The reason is that if average mortgage debt had remained constant over the three years, average home equity would have dropped by 37 percent.23 It was only the contraction of average mortgage debt over these years that kept the percentage decline in home equity at 26 percent instead of 37 percent.

The pattern by demographic group in the change in net home equity tends to be different than that of the fifth column, the share of households who were underwater in 2010. Hispanic home owners suffered by far the largest percentage decline in home equity – 47 percent – of the three racial/ethnic groups. Black households experienced a somewhat larger percentage decline than white home owners. Single female households experienced a somewhat larger decline than single males or married couples. The less schooled households suffered a larger decline than college graduates (only 26 percent for the latter). There is tremendous age variation, with older households more immune to the housing price collapse. The youngest age group experienced a 53 percent fall in home equity while the oldest age group had “only” a 19 percent decline.

There is a U-shaped pattern with regard to household income, with the lowest income class experiencing only a 0.6 percent depreciation in home equity, income class $75,000-$99,999 suffering the greatest percentage decline – 32 percent – and the highest income class undergoing a 18 percent loss. It is likely that this pattern is due to the fact that Hispanic, black, and younger households came later into the home buying market and therefore were more likely to buy when prices were peaking. Indeed, during the early 2000s mortgage companies and banks were using all kinds of devices to permit households with low income and low credit ratings to get into very risky mortgages. This particularly affected minorities and low income whites.

Generally speaking (though not always) the groups with the highest ownership rates -- non-Hispanic whites, married people, people with higher education, older people (over age 64), and people with higher income -- also had the lowest share of homeowners with negative home equity and the lowest percentage loss in home equity. Here, too, the difference likely reflects when the families in these groups bought their home. Of those who were home owners, minorities, married individuals, those with some college education, younger people and people with incomes between $50,000 and $100,000 had the highest percentages with negative home equity. Likewise, among home owners, Hispanics, single females, those with less than a high school degree and those with some college, younger households, and those with incomes between $75,000 and $990,000 suffered the largest percentage declines in home equity. Young homeowners under the age of 35 (16.2 percent with negative home equity and a 59 percent decline in net home equity) were the hardest hit by the recession.

An update to 2013 is also provided in Table 8. Did the housing situation change by 2013? The overall homeownership rate, as noted above, fell by 2.1 percentage points between 2010 and 2013. Blacks and Hispanics suffered larger declines than whites but declines were about equal among family types. Those with some college, middle aged families, particularly age group 45- 54, and middle income families, especially income class $25,000 to $49,999, experienced the largest drops in homeownership. On the other hand, the homeownership rate picked up among age group 65-74, the lowest income class, and income class $75,000-$99,999.

There was a modest reduction in the overall share of homeowners underwater between 2010 and 2013, from 8.2 to 6.9 percent. The share fell among white households but continued to rise among black and Hispanic households, by 5.0 and 2.9 percentage points, respectively. By 2013, 14 percent of black homeowners had negative home equity, the largest of the three groups. The “underwater rate” fell among the three family types – most strongly among single females – It declined among the educational groups except the lowest one, where it showed a very modest uptick. There was a sizeable decline in the underwater rate among the youngest age group, bringing it down to 9.4 percent from 16.2 percent, and more modest decreases for age groups 35- 44 and 75 and over. For the oldest group, the underwater rate was down to 0.4 percent. There were relatively small changes among the other age groups. Changes were also relatively minor by income class, except for middle income ones which experienced very substantial declines in their underwater rate (decreases of 4.7 percentage points for income class $50,000-$74,999 and 2.9 percentage points for income class $75,000-$99,999).

Overall, mean home equity declined by 3.8 percent in real terms from 2010 to 2013. Among African-Americans, it fell by 20 percent, compared to 3.4 percent for whites, and among Hispanics a slight increase was recorded, offsetting the steep falloff in the previous three years. Home equity among single males rose by 4.8 percent but dropped by 1.6 percent among married couples and almost 10 percent among single females, compounding the previous precipitous decrease. Net home equity recovered somewhat among those with less than a high school education but fell sharply among high school graduates. Some recovery in net home equity was found for the two youngest age groups but it continued to fall among middle age (45-54 and 55- 64) households and among the oldest age group. The record is mixed by income classes, though the middle income group ($25,000-$49,999) showed the steepest falloff in home equity.

7. The role of leverage in explaining the trends in median wealth and wealth inequality

7.1 The Six Puzzles

Six puzzles emerge from the preceding analysis. Before proceeding to a discussion of these, it is helpful to review previous work on these issues. As discussed in Wolff (2002), wealth inequality is very sensitive and positively related to the ratio of stock prices to housing prices, since the former is heavily concentrated among the rich and the latter is the chief asset of the middle class. A regression was run of a wealth inequality index, measured by the share of marketable wealth held by the top one percent of households (WLTH) on income inequality, measured by the share of income received by the top five percent of families (INC), and the ratio of stock prices (the Standard and Poor index) to housing prices (RATIO), with 21 data points between 1922 and 1998. It yields:

(1) WLTH = 5.10 + 1.27 INC + 0.26 RATIO, R2 = 0.64, N = 21 (0.9) (4.2) (2.5)

with t-ratios shown in parentheses. Both variables are statistically significant (INC at the 1 percent level and RATIO at the 5 percent level) and with the expected (positive) sign. Also, the fit is quite good, even for this simple model.

The first puzzle is why median wealth surged from 2001 to 2007 while median income was sluggish. The second is why wealth inequality was flat over these years when income inequality grew. The third is why there was such a steep plunge in median wealth between 2007 and 2010 of 47 percent. This happened despite a moderate drop in median income of 6.4 percent in real terms and steep but less steep declines in housing and stock prices of 24 and 26 percent in real terms, respectively.

The fourth is why there was such a steep increase of wealth inequality, of 0.035 Gini points, over years 2010 to 2013. It is surprising that wealth inequality rose so sharply, given that income inequality dropped by 0.025 Gini points (at least according to the SCF data) and stock and housing prices declined at about the same rate.

The fifth and, perhaps, most perplexing one is why median (and mean) wealth failed to recover in years 2010 to 2013 when asset prices surged. The last is why wealth inequality increased so moderately from 2010 to 2013 when income inequality shot up and the stock to house price ratio climbed from 66 to 85.

Changes in median wealth can be explained to a large extent by the high leverage (that is, debt to net worth ratio) of middle class households. This is particularly the case for the strong gains in median net worth over the 2001 to 2007 period and its steep fall over years 2007 to 2010. Trends in wealth inequality are largely accountable by differential leverage between the rich and the middle class. This factor can help explain the constancy of wealth inequality over the 2001-2007 and 2010-2013 periods and its spike over years 2007 to 2010. With regard to the fact that median net worth showed no improvement over years 2010 to 2013, a different explanation is called for. It appears that substantial dissavings over this period accounts for the failure of wealth to grow over these years.

7.2 Two arithmetic examples

A simple arithmetical example might illustrate the effects of leverage. Suppose average assets are 50 and average debt is zero (see Table 9a). Also, suppose that asset prices rise by 20 percent. Then average net worth also rises by 20 percent. However, now suppose that average debt is 40 and asset prices once again rise by 20 percent. Then average net worth increases from a base of 10 (50 minus 40) to 20 (60 minus 40) or by 100 percent, Thus, leverage amplifies the effects of asset price changes.

However, the converse is also true. Suppose that asset prices decline by 20 percent. In the first case, net worth falls from 50 to 40 or by 20 percent. In the second case, net worth falls from 10 to 0 (40 minus 40) or by 100 percent. Thus, leverage can also magnify the effects of an asset price bust.

Another arithmetical example might illustrate the effects of differential leverage. Suppose the total assets of the very rich in a given year is 100, consisting of 50 in stocks and 50 in other assets, and its debt is zero, for a net worth of 100 (see Table 9b). In contrast, among the “middle class”, suppose their total assets are 70, consisting of 60 in housing and 10 in other assets, and their mortgage debt is 30, for a net worth of 40. The ratio of net worth between the very rich and the middle is then 2.5 (100/40).

Suppose the value of both stocks and housing falls by 20 percent from year one to year two. Then, the total assets of the rich fall to 90 (40 in stocks and 50 in other), for a net worth of 90.24 The total assets of the middle falls to 58 (48 in housing and 10 in other) but its debt remains exactly the same at 30, for a net worth of 28. As a result, the ratio of net worth between the rich and the middle rises to 3.21 (90/28). Here it is apparent that even though housing and stock prices fall at the same rate, the inequality of wealth goes up. The key is the differential leverage between the rich and the middle. If asset prices fall, then the rate of return to net worth will be lower than that to assets alone if households are leveraged. In other words, if asset prices decline at the same rate, net worth decreases at an even greater rate for the middle class than the rich, since the ratio of debt to net worth is much higher for the middle class than the rich. The converse is also true. If the debt-equity ratio is higher for the middle class than the rich, then an equal percentage increase in house and stock prices will result in a decrease in wealth inequality.

7.3 Rates of return

Table 10 shows estimates of average annual rates of return for both gross assets and net worth over the period from 1983 to 2013. Results are based on the average portfolio composition over the period and assume that all households receive the same rate of return by asset type.25 It is first of interest to look at the results for all households (see Appendix Table 1 for the source data). The overall average annual rate of return on gross assets rose from 2.33 percent in the 1983-1989 period to 3.33 percent in the 1989-2001 period and then fell slightly to 3.10 percent in the 2001-2007 period before plummeting to -6.38 percent over the Great Recession. This was followed by a substantial recovery to 4.83 percent over years 2010 to 2013.

As shown in Appendix Table 1, the largest declines in asset prices over the years 2007 to 2010 occurred for residential real estate and the category businesses and non-home real estate. The value of financial assets, including stocks, bonds, and other financial securities, registered an annual rate of return of “only” -3.72 percent because interest rates on corporate and foreign bonds continued to remain strong over these years. The value of pension accounts had a -0.34 percent annual rate of return, reflecting the mixture of bonds and stocks held in pension accounts. From 2010 to 2013, all asset classes with the exception of liquid assets made a robust recovery. This was led by financial assets which recorded a 12.5 percent annual return and businesses and non-home real estate, which experienced a 7.4 percent annual return.

The average annual rate of return on net worth among all households also increased from 3.32 percent in the first period to 4.35 percent in the second , declined somewhat to 4.04 percent in the third and then fell off sharply to -7.28 percent in the 2007-2010 period. Once again, there was a strong recovery to 6.20 percent in the 2010-2013 period. It is first of note that the annual rates of return on net worth are uniformly higher – by about one percentage point – than those of gross assets over the first three periods and the last period, when asset prices were generally rising. However, in the 2007-2010 period, the opposite was the case, with the annual rate of return on net worth about one percentage point lower than that on gross assets. These results illustrate the effect of leverage, raising the return when asset prices rise and lowering the return when asset prices fall. Over the full 1983-2013 period, the annual return on net worth was 0.83 percentage points higher than that on gross assets.26

When we next consider rates of return by wealth class, we see some striking differences. The highest rates of return on gross assets were registered by the top one percent of wealth holders, followed by the next 19 percent and then by the middle three wealth quintiles. The one exception was the 2007-2010 period when the next 19 percent was first (the least negative), followed by the top one percent and then the middle three quintiles. The differences are quite substantial. Over the full 1983-2013 period, the average annual rate of return on gross assets for the top one percent was 0.59 percentage points greater than that of the next 19 percent and 1.52 percentage points greater than that of the middle quintiles. The differences reflect the greater share of high yield investment assets like stocks in the portfolios of the rich and the greater share of housing in the portfolio of the middle class (see Tables 5 and 6). Indeed, in the 2010-2013 period, there was a huge cleavage in rates of return between the top one percent and the middle group of 2.63 percentage points, reflecting the much higher gains on stocks and investment assets than on housing in those years.

This pattern is almost exactly reversed when we look at rates of return for net worth. In this case, in the first three periods and the last when asset prices were generally rising, the highest return was recorded by the middle three wealth quintiles but in the 2007-2010 period, when asset prices were declining, the middle three quintiles registered the lowest (that is, most negative) rate of return. The exception was the first period when the top one percent had a slightly higher return than the middle class. The reason was the substantial spread in returns on gross assets between the top one percent and the middle group – 1.72 percentage points.

Differences in returns between the top one percent and the middle three quintiles were quite substantial in some years. In the 2001-2007 period, the average annual rate of return on net worth was 5.58 percent for the latter and 3.92 percent for the former – a difference of 1.67 percentage points. The spread was less over years 2010 to 2013, only 0.79 percentage points. The smaller difference was due to the much higher returns on gross assets held by the top percentile than by the middle group. On the other hand, over years 2007 to 2010, when asset prices declined, the rate of return on net worth was -6.52 percent for the top one percent and -10.55 percent for the middle three quintiles – a differential of 4.04 percentage points in favor of the top one percent.

The spread in rates of return between the top one percent and the middle three quintiles reflects the much higher leverage of the middle class. In 2013, for example, the debt-equity ratio of the middle three quintiles was 0.64 while that of the top one percent was 0.026. The debt-equity ratio of the next 19 percent was also relatively low, at 0.118.

The huge negative rate of return on net worth of the middle three wealth quintiles was largely responsible for the precipitous drop in median net worth between 2007 and 2010. This factor, in turn, was due to the steep drop in asset prices, particularly housing, and the very high leverage of the middle wealth quintiles. Likewise, the very high rate of return on net worth of the middle three quintiles over the 2001-2007 period played a big role in explaining the robust advance of median net worth, despite the sluggish growth in median income. This in turn, was a result of their high leverage coupled with the boom in housing prices. However, somewhat puzzling is the fact that the rate of return on net worth of the middle group was very high over years 2010 to 2013 – in fact, the highest of any period – and yet median wealth stagnated over these years. We shall return to this issue later.

The substantial differential in rates of return on net worth between the middle three wealth quintiles and the top quintile (our percentage points lower) helps explain why wealth inequality rose sharply between 2007 and 2010 despite the decline in income inequality. Likewise this differential over the 2001-2007 period (a spread of 1.67 percentage points in favor of the middle quintiles) helps account for the stasis in wealth inequality over these years despite the increase in income inequality. The higher rate of return of the middle than the top group over years 2010 to 2013 also helps account for the relative constancy in wealth inequality despite the rise in income inequality.

8. The racial divide widens over the Great Recession

Striking differences are found in the wealth holdings of different racial and ethnic groups. In Tables 11 and 12, households are divided into three groups: (i) non-Hispanic whites, (ii) non- Hispanic African-Americans, and (iii) Hispanics.27 In 2006, while the ratio of mean incomes between non-Hispanic white (“white”) and non-Hispanic black (“black”) households was an already low 0.48 and the ratio of median incomes was 0.60, the ratios of mean and median wealth holdings in 2007 were even lower, at 0.19 and 0.06, respectively (also see Figure 16).28 The homeownership rate for black households was 49 percent in 2007, a little less than two thirds the rate among whites, and the percentage of black households with zero or negative net worth stood at 33.4, more than double the corresponding percentage among whites.

Between 1982 and 2006, while the average real income of white households increased by 42 percent and the median by 10 percent, the former rose by only 28 percent for blacks and the latter by 18 percent. As a result, the ratio of mean income slipped from 0.54 in 1982 to 0.48 in 2006, while the ratio of median income rose from 0.56 to 0.60. The contrast in the time trends for the ratio of means and that of medians reflects the fact that a relatively small number of white households increased their incomes by a huge amount over these years – a result of rising income inequality among white households.

Between 1983 and 2001, average net worth (in constant dollars) climbed by 73 percent for whites but rose by only 31 percent for black households, so that the net worth ratio fell from 0.19 to 0.14. However, between 2001 and 2007, mean net worth among black households gained an astounding 58 percent while white wealth advanced only 29 percent, so that by 2007 the net worth ratio was back to 0.19, the same level as in 1983.

In the case of median wealth, the black-white ratio increased from 7 percent in 1983 to 10 percent in 2001 but then dipped to 6 percent in 2007, a little less than the ratio in 1983. In this case, median wealth among white households grew by 37 percent between 1983 and 2001 but more than doubled among black households. However, between 2001 and 2007, median net worth among black households actually crashed by 26 percent, reflecting in part the rising share of black households with zero or negative net worth.

The homeownership rate of black households grew from 44 to 47 percent between 1983 and 2001 but relative to white households, the homeownership ratio slipped slightly to 0.64 in 2001. Homeownership rates continued to rise for both groups between 2001 and 2007 and the homeownership ratio remained about the same.

In contrast, the percentage of black households reporting zero or negative net worth fell from 34 percent in 1983 to 31 percent in 2001 (and also fell relative to the corresponding rate for white households). However, by 2007, the share was up to 33 percent (though a bit lower relative to whites). The share of households with zero or negative wealth very likely reflects the boom/bust cycle in the housing market. For example, if a family bought a home in 2001, its home value increased substantially as home prices surged but then tanked as home prices collapsed, leading to a sharp decline in net worth.

The picture is somewhat different for Hispanics (see Table 12). The ratio of mean income between Hispanics and (non-Hispanic) whites in 2007 was 0.50, almost the same as that between blacks and whites. However, the ratio of median income was 0.70, much higher than the black-white ratio. The ratio of mean net worth was 0.26 compared to a ratio of 0.19 between blacks and whites. However, the ratios of medians were 0.06 and 0.01, respectively, almost identical to those between blacks and whites. The Hispanic homeownership rate was 49 percent, almost identical to that of black households, and 34 percent of Hispanic households reported zero or negative wealth, almost the same as African-Americans.

Progress among Hispanic households over the period from 1983 to 2007 was generally a positive story. Mean household income for Hispanics grew by 18 percent and median household income by 16 percent, so that the ratio of mean income slid from 60 to 50 percent while that of median income advanced from 66 to 70 percent.

The ratio of mean net worth between Hispanic and white households was about the same in 2001 as in 1983. Mean net worth among Hispanics then surged by an astonishing 82 percent from 2001 to 2007, and the corresponding ratios climbed to 26 , quite a bit higher than that between black and white households. The steep rise in Hispanic wealth from 2001 to 2007 can be traced to a five percentage point jump in the Hispanic home ownership rate (see below).

From 1983 to 2007, median wealth among Hispanics remained largely unchanged, so that the ratio of median wealth between Hispanics and whites stayed virtually the same. In contrast, the homeownership rate among Hispanic households surged from 33 to 44 percent between 1983 and 2001, and the ratio of homeownership rates between the two groups advanced from 0.48 to 0.60. Then, between 2001 and 2007, the Hispanic homeownership rose once again, to 49 percent, about the same as black households, and the homeownership ratio grew to 0.66.

The percentage of Hispanic households with zero or negative net worth fell rather steadily over time, from 40 percent in 1983 to 34 percent in 2007, and the share relative to white household tumbled from a ratio of 3.55 to 2.30.

Despite some progress from 2001 to 2007, the respective wealth gaps between African- Americans and Hispanics on the one hand and non-Hispanic whites on the other were still much greater than the corresponding income gaps in 2007. While mean income ratios were of the order of 50 percent, mean wealth ratios were of the order of 20-25 percent. The percent with zero or negative net worth was around a third, in contrast to 15 percent among white households (a difference that appears to mirror the gap in poverty rates). While blacks and Hispanics were left out of the wealth surge of the years 1998 to 2001 because of relatively low stock ownership, they actually benefited from this (and the relatively high share of houses in their portfolio) in the 2001-2007 period. However, all three racial/ethnic groups saw an increase in their debt to asset ratio from 2001 to 2007.29

The racial/ethnic picture changed radically by 2010. While the ratio of both mean and median income between black and white households changed very little between 2007 and 2010 (mean income, in particular, declined for both groups), the ratio of mean net worth dropped from 0.19 to 0.14. The proximate causes were the higher leverage of black households and their higher share of housing wealth in gross assets (see Table 13). In 2007, the ratio of debt to net worth among African-American households was an astounding 0.55, compared to 0.15 among whites, while housing as a share of gross assets was 54 percent for the former as against 31 percent for the latter. The ratio of mortgage debt to home value was also much higher for blacks, 0.49, than for whites, 0.32. The sharp drop in home prices from 2007 to 2010 thus led to a relatively steeper loss in home equity for black homeowners, 26 percent, than for white homeowners, 24 percent (see Table 8), and this factor, in turn, led to a much steeper fall in mean net worth for black households than white households. In fact, the annual rate of return on the net worth of black families over years 2007 to 2010 was a staggering -9.9 percent, compared to -7.1 percent for white households.30

The early part of the Great Recession actually hit Hispanic households much harder than black households in terms of household wealth. Mean income among Hispanic households rose a bit from 2007 to 2010 and the ratio with respect to white households increased from 0.50 to 0.57. On the other hand, the median income of Hispanics fell, as did the ratio of median income between Hispanic and white households. However, the mean net worth in constant dollars of Hispanics fell almost in half, and the wealth ratio with respect to white households plummeted from 0.26 to 0.15. The same factors were responsible as in the case of black households. In 2007, the debt-equity ratio for Hispanics was 0.51, compared to 0.15 among whites, while housing as a share of gross assets was 53 percent for the former as against 31 percent for the latter (see Table 13). The ratio of mortgage debt to home value was also higher for Hispanics, 0.452, than for whites, 0.324. As a result, net home equity dropped by 47 percent among Hispanic homeowners, compared to 24 percent among white homeowners (see Table 8), and this factor, in turn, was largely responsible for the huge decline in Hispanic net worth both in absolute and relative terms. Indeed, the annual rate of return on the net worth of Hispanic families over these years was an astonishing -10.8 percent, compared to -7.1 percent for white households.

There are two reasons that might explain the extreme drop in Hispanic net worth. First, a large proportion of Hispanic home owners bought their home in the interval from 2001 to 2007, when home prices were peaking. This is reflected in the sharp increase in their home ownership rate over this period. As a result, they suffered a disproportionately large percentage drop in their home equity. Second, it is likely that Hispanic home owners were more heavily concentrated than whites in parts of the country like Arizona, California, Florida, Arizona, and Nevada where home prices plummeted the most. There was also a steep drop in the homeownership rate among Hispanic households of 1.9 percentage points from 2007 to 2010.

Was there any relative improvement over the second half of the Great Recession, 2010- 2013? Black households continued to suffer moderate losses in both mean and median household income in absolute terms, and declines relative to white households. The mean net worth of black households also continued to fall, in this case by 9 percent, and the ratio of mean net worth between black and white households dipped further to 0.13 from 0.14. Their median net worth actually fell from $6,700 to $1,700, and the ratio relative to white households plunged from 0.06 to 0.01.

One of the most notable developments was a sharp fall in the black homeownership rate from 48 to 44 percent, which followed a more modest 0.9 percentage point decrease from 2007 to 2010, and a decline in the homeownership rate relative to white households from 0.64 in 2010 to 0.60 in 2013. Equally striking is the steep uptick in the share of black households with no net worth, from 33 to 40 percent. Thus, by almost all indicators, the absolute and relative position of black household deteriorated even further from 2010 to 2013. This development actually seems surprising in light of the fact that the annual yield on the portfolio of black households was 7.14 percent, compared to 6.12 percent for white households. The key is the sharp decline in their homeownership rate. Indeed, this led to a considerable loss in home equity in the black portfolio, which fell by 26 percent overall and 20 percent among black homeowners (see Table 8).

Income trends were very similar for Hispanics but wealth trends were different. Mean incomes of Hispanics were down 15 percent from 2010 to 2013, and the ratio relative to white households plunged from 0.57 to 0.45. The story was similar for median income. On the other hand, the mean net worth of Hispanic households remained stable from 2010 to 2013, as did their position relative to white households, while their median wealth fell from $2,900 to $2,000. However, like black families, their homeownership rate continued to fall, in this case from 47 to 44 percent (back to where it was in 1992), and their homeownership ratio relative to whites also slipped from 0.63 to 0.60. The percent of Hispanics with non-positive wealth actually fell slightly from 2010 to 2013. Overall, Hispanic households had an average annual rate of return on their portfolio of 7.48 percent, compared to 7.14 percent for black households. The main difference between them and black households was a much smaller decline in home equity – only 5 percent overall – and an actual 1.6 percent increase among Hispanic homeowners alone.

9. Wealth shifts from the young to the old

As shown in Table 14, the cross-sectional age-wealth profiles of the various years between 1983 and 2013 generally follow the predicted hump-shaped pattern of the life-cycle model (see, for example, Modigliani and Brumberg, 1954). Mean wealth increases with age up through age 65 or so and then falls off. Homeownership rates also have a similar profile, though the fall-off after the peak age is much more attenuated than for the wealth numbers (and in 2004 they actually show a steady rise with age). In 2013, the wealth of elderly households (age 65 and over) was 2.0 times as high as that of the non-elderly and their homeownership rate was 24 percentage points higher.

Despite the apparent similarity in the profiles, there were notable shifts in the relative wealth holdings of age groups between 1983 and 2007 (also see Figure 7). The relative wealth of the youngest age group, under 35 years of age, slipped from 21 percent of the overall mean in 1983 to 17 percent in 2007. In 2007, the mean wealth of the youngest age group was $102,400 (in 2013 dollars), which was only slightly more than the mean wealth of this age group in 1989 ($99,500). The mean net worth of the next youngest age group, 35-44, relative to the overall mean tumbled from 0.71 in 1983 to 0.58 in 2007.

Changes in homeownership rates tend to mirror net worth trends. While the overall ownership rate increased by 5.2 percentage points between 1983 and 2007 (from 63.4 to 68.6 percent), the share of households in the youngest age group owning their own home increased by only 2.1 percentage points. The homeownership rate of households between 35 and 44 of age actually fell by 2.3 percentage points. By 2007, homeownership rates rose monotonically with age up to age group 65-74 and then dropped for the oldest age group. The statistics point to a relative shifting of homeownership away from younger towards older households between 1983 and 2007.

Changes in relative wealth were even more dramatic from 2007 to 2010. The relative wealth of the under 35 age group plummeted from 0.17 to 0.11 and that of age group 35-44 from 0.58 to 0.42. In 2013 dollars, the average wealth of the youngest age group collapsed almost in half, from $102,400 in 2007 to $55,400 in 2010, its second lowest point over the 30 year period (the lowest occurred in 1995), while the relative wealth of age group 35-44 shrank from $346,900 to $211,200 its lowest point over the whole 1983 to 2010 period. One possible reason for these steep declines in wealth is that younger households were more likely to have purchased their homes near the peak of the housing cycle. Home ownership rates fell for all age group from 2007 to 2010 (except the very oldest) but the percentage point decline (3.3 percentage points) was greatest for the youngest age group.

Changes in the relative wealth position of different age groups depend in large measure on relative asset price movements and differences in asset composition. The latter are highlighted in Table 15 for the year 2007. Homes comprised over half the value of total assets for age group 35 and under, and its share of total assets fell off with age to about a quarter for age group 55-64 and then rose to 30 percent for age group 75 and over. Liquid assets as a share of total assets remained relatively flat with age group at around 6 percent except for the oldest group for whom it was 11 percent, perhaps reflecting the relative financial conservativeness of older people. Pension accounts as a share of total assets rose from 4 percent for the youngest group to 16 percent for age group 55 to 64 and then fell off to 5 percent for the oldest age group. This pattern likely reflects the build-up of retirement assets until retirement age and then a decline as they are liquidated.31 Corporate stock and financial securities showed a steady rise with age, from a 4 percent share for the youngest group to a 26 percent share for the oldest. A similar pattern is evident for total stocks as a percentage of all assets. Unincorporated business equity and non-home real estate are relatively flat as a share of total assets with age, about 30 percent.

There was a pronounced fall off of debt with age. The debt-equity ratio declined from 93 percent for the youngest group to 2 percent for the oldest, the debt-income ratio from 168 percent to 30 percent, and principal residence debt as a share of house value from 65 to 5 percent. As a result of the latter, net home equity as a proportion of total assets rose from 19 to 29 percent from the youngest to oldest age group.

Younger households were thus more heavily invested in homes and more heavily in debt whereas the portfolio of older households was more heavily skewed to financial assets, particularly corporate stock. As a result, younger households benefit relatively when housing prices rise and inflation is strong while older households benefit relatively from rising stock prices. Changes in the relative net worth position of age groups over the 1983 to 2007 period were to a large extent due to differences in portfolio composition and relative asset price movements. Conversely, as with black and Hispanic households, the higher leverage of younger age groups made them vulnerable when asset prices, particularly housing prices, declined.

The steep decline in house prices from 2007 to 2010 then led to a relatively steeper loss in home equity for the youngest homeowners, 53 percent, than all homeowners, 29 percent (see Table 8), and this factor, in turn, led to a much steeper fall in net worth . Indeed, in terms of the annual rate of return on their wealth portfolio, this group, which had the highest over the 2001- 2007 period, 7.9 percent, had the lowest over the 2007-2010 period, -13.5 percent!

The story is very similar for age group 35 to 44. Their debt-equity ratio was 0.41 in 2007, their ratio of mortgage debt to house value was 0.51, and their share of housing in gross assets was 44 percent, all much higher than overall. As with the youngest age group, the drop in home prices from 2007 to 2010 caused a large fall in home equity of 48 percent among homeowners, which in turn caused a steep fall off in their relative net worth. In terms of the annual rate of return on their wealth portfolio, this group went from being the second highest in years 2001- 2007, 5.6 percent, to the second lowest in years 2007 to 2010, -7.4 percent.

Years 2010 to 2013 saw an 11 percent (real) increase in the net worth of the youngest age group and a slight rise in relative terms as well. On the surface, one might have expected an even larger rise since the rate of return on the portfolio of this age group was a robust 10.7 percent per year – the highest of any age group. However, further investigation indicates that the main reason why its net worth did not increase more was the continued decline in its homeownership rate, which fell by almost two percentage points.

Age group 35-44 made a big comeback in terms of net worth, which rose an astonishing 54 percent (in real terms) from 2010 to 2013. The average net home equity among homeowners in this age group jumped by 36 percent (see Table 8), and though the homeownership rate did fall by two percentage points, average home equity among all households in this age group expanded by 32 percent. This age group also had a 7.5 annual average return on its portfolio over these years, and, partly as a result, the mean value of other real estate was up by 39 percent, that of business equity by 137 percent, mean pension accounts by 42 percent, and mean corporate stock and mutual funds by 40 percent.

10. Summary and concluding remarks

After a period of robust growth, median wealth continued to climb by 19 percent from 2001 to 2007, even faster than during the 1990s (and 1980s). Median income, on the other hand, rose only 1.6 percent. Then the Great Recession hit. From 2007 to 2010, house prices fell by 24 percent in real terms, stock prices by 26 percent, and median wealth by a staggering 44 percent. Median income also dropped but by a more modest 6.7 percent and median non-home wealth plummeted by 49 percent. The share of households with zero or negative net worth rose sharply from 18.6 to 21.8 percent.

However, from 2010 to 2013, asset prices recovered with stock prices up by 39 percent and house prices by 8 percent. Despite this, both median and mean wealth stagnated, while median income was down by 1.3 percent but mean income rose by 0.9 percent. The percent of households with zero or negative net worth remained unchanged.

Wealth inequality after remaining relatively stable from 1989 to 2007 showed a steep advance over years 2007 to 2010. The Gini coefficient climbed from 0.834 to 0.866 and the share of the top 20 percent from 85 to 89 percent. In contrast, income inequality, after rising moderately from 2000 to 2007 (an increase of 0.12 Gini points), dropped substantially from 2006 to 2009 (a decrease of 0.25 Gini points). Net worth inequality, on the other hand, remained relatively unchanged between 2010 and 2013, though the share of the top one percent was up by 1.6 percentage points. But income inequality showed a substantial rise from 2010 to 2013, with the Gini coefficient returning to its 2007 level.

Between 1983 and 2013, the top one percent received 41 percent of the total growth in net worth, 43 percent of the total growth in non-home wealth, and 49 percent of the total increase in income. The figures for the top 20 percent are 99 percent, 98 percent, and 103 percent, respectively – that is to say, the upper quintile got it all!

Another notable development was the sharply rising debt to income ratio during the early and mid 2000s, reaching its highest level in almost 25 years, at 119% among all households in 2007. The debt-equity ratio was also way up, from 14.3 percent in 2001 to 18.1 percent in 2007. Most of the rising debt was from increased mortgages on homes. From 2007 to 2010 both ratios continued to rise, the former moderately from 119 to 127 percent and the latter more steeply from 18.1 to 20.6 percent. This was true despite a moderate retrenchment of overall average debt of 4.4 percent and reflected the drop in both mean wealth and income. Both ratios fell off sharply by 2013, to 107 percent and 17.9 percent, respectively, as outstanding debt continued to shrink, by 13 percent in this case.

Home values as a share of total assets among all households remained relatively unchanged from 1983 to 2010 (around 30 percent). However, net equity in owner-occupied housing as a share of total assets fell from 24 percent in 1983 to 17 percent in 2010, reflecting rising mortgage debt on homeowner's property, which grew from 21 to 39 percent in 2013. The large increase in the ratio from 2007 to 2010 was a result of falling home values (average mortgage debt actually declined by 5.0 percent in constant dollars). The decline from 2010 to 2013 reflected a substantial reduction in average outstanding mortgage debt (13 percent).

Among the middle class (defined here by the middle three wealth quintiles) there was a huge increase in the debt-income ratio from 100 to 157 percent from 2001 to 2007 and of the debt-equity ratio from 46 to 61 percent. The debt-equity ratio was also much higher among the middle 60 percent of households in 2007, at 0.61, than among the top one percent (0.028) or the next 19 percent (0.121). However, from 2007 to 2010, while the debt-equity ratio continued to advance to 69 percent, the debt to income ratio actually fell off to 134 percent. The reason is the substantial retrenchment of average debt among the middle class over these years. Overall debt fell by 25 percent in real terms, mortgage debt by 23 percent, and other debt by 32 percent. The fact that the debt-equity ratio rose over these years was a reflection of the steep drop in median net worth of 44 percent. Both ratios dropped from 2010 to 2013 as outstanding debt levels fell -- overall debt by 8 percent and mortgage debt by 10 percent, though other debt rose by 1.6 percent.

The overall stock ownership rate (either directly or indirectly through mutual funds, trust funds, or pension plans), after rising briskly from 32 percent in 1989 to 52 percent in 2001, fell off moderately to 49 percent in 2007 and then to 47 percent in 2010 and 46 percent in 2013. Similar time trends are evident for the share of households with $5,000 or more of stocks (in 1995 dollars) and with $10,000 or more of stocks. The fall off from 2007 to 2010 was surprisingly modest in light of the very steep decline in stock prices over those years.

However, the concentration of investment type assets generally remained as high in 2013 as during the previous two and a half decades. About 90 percent of the total value of stock shares, bonds, trusts, and business equity, and about 80 percent of non-home real estate were held by the top 10 percent of households. Stock ownership is also highly skewed by wealth and income class. The top one percent of households classified by wealth owned 38 percent of all stocks in 2013, the top 10 percent 81 percent, and the top quintile 92 percent.

Despite the 24 percent plunge in house prices (in real terms) from 2007 to 2010, the share of home owners who were “underwater” was “only” 8.2 percent in 2010. However, average home equity among home owners did decline by 29 percent. This reduction would have been higher except for the contraction of mortgage debt noted above. Hispanics, younger households, and middle income households were hit particularly hard in terms of the loss of home equity. From 2010 to 2013, the share underwater fell to 6.9 percent as mortgage debt continued to decline and house prices recovered somewhat. Mean home equity dropped by only 3.8 percent. The decline was particularly great among black, single female, middle aged, and middle class households, while younger households recorded particularly large gains.

The one piece of mainly positive news is that among all households there was no deterioration in pension accumulations in DC-type pension plans over the Great Recession. The share of households with a DC account, after rising from 11 percent in 1983 to 53 percent in 2007, did fall off to 49 percent in 2013. However, average DC pension wealth among all households continued to grow from 2007 to 2010 and from 2010 to 2013. The main reason was a shifting of household portfolios. Pension accounts as a share of total assets, after rising from 1.5 percent in 1983 to 12.1 percent in 2007, jumped to 15.1 percent in 2010 and then to 16.5 percent in 2013. Moreover, the percent of middle class households with a defined contribution pension plan, after growing robustly from 12 percent in 1983 to 53 percent in 2007, fell off sharply to 46 percent in 2010 and then to 44 percent in 2013, and the change in dollar terms from 2007 to 2013 was -16 percent.

The key to understanding the plight of the middle class over the Great Recession was their high degree of leverage and the high concentration of assets in their home. The steep decline in median net worth between 2007 and 2010 was primarily due to the very high negative rate of return on net worth of the middle three wealth quintiles (-10.6 percent per year). This, in turn, was attributable to the precipitous fall in home prices and their very high degree of leverage. High leverage, moreover, helps explain why median wealth fell more than house (and stock) prices over these years and declined much more than median income.

However, this is not the whole story with regard to the collapse in median net worth from 2007 to 2010. On the basis of the rates of return computed for the middle three wealth quintiles (and assuming that all wealth classes receive the average rate of return by asset class), median wealth should have fallen by only 27 percent, instead of the actual 44 percent. If we ignore net flows of inheritances and gifts over the period,32 this discrepancy must be due to dissavings. Indeed, the results imply a substantial dissaving rate over this period, of 5.6 percent per year relative to initial wealth and 11.4 percent per year relative to initial (median) income.33

With regard to the fact that median net worth showed no improvement over years 2010 to 2013, a different explanation is called for. For the period from 2010 to 2013, the whole story is dissavings. As noted above, asset prices more than recovered from 2010 to 2013, except for housing, which was still up by 8 percent (in real terms). On the basis of rates of return computed for the three middle wealth quintiles, median net worth should have increased by 36 percent. Despite this, median wealth was down slightly over these years. It appears (once again ignoring net flows of inheritances and gifts) that substantial dissavings over this period accounts for the failure of wealth to grow over these years. In particular, the middle class must have had an annual dissavings rate of 8.1 percent relative to initial wealth and 9.9 percent relative to initial (median) income.34

The stagnation of median wealth from 2010 to 2013 can be traced to the depletion of assets. In particular, the middle class was using up its assets to pay down its debt, which decreased by 8.2 percent over these years. This shows up, in particular, in reduced asset ownership rates. The homeownership rate fell from 68.0 to 66.7 percent, that of pension accounts from 45.8 to 44.4 percent, that of unincorporated businesses from 8.2 to 6.6 percent, and that of stocks and financial securities from 15.3 to 14.2 percent. However, the reduction in assets was greater than the reduction of debt.

The likely reason for the high rate of dissavings of the middle class over both the 2007- 2010 and the 2010-2013 periods is income stagnation (actually, a reduction in median income over these years). It appears that the middle class was depleting its assets to maintain its previous level of consumption. The evidence, moreover, suggests that middle class households, experiencing stagnating incomes, expanded their debt (at least until 2007) mainly in order to finance normal consumption expenditures rather than to increase their investment portfolio.35

The large spread in rates of return on net worth between the middle three wealth quintiles and the top percentile (over four percentage points) also largely explains why wealth inequality advanced steeply from 2007 to 2010 despite the decline in income inequality and constancy in the ratio of stock to housing prices (both declined at about the same rate over these years). It was thus the case that the middle class took a bigger relative hit on their net worth from the decline in home prices than the top 20 percent did from the stock market plunge. This factor is also reflected in the fact that median wealth dropped much more in percentage terms than mean wealth over the Great Recession.

In contrast, there was relatively little change in wealth inequality from 2010 to 2013. This is true despite the large increase in income inequality over these years as well as a sharp rise of 29 percent in the ratio of stock to housing prices. The offsetting factor in this case was the higher rate of return on net worth of the middle class than the top one percent (a 0.79 percentage point difference).

The racial disparity in wealth holdings, after fluctuating over the years from 1983 to 2007, was almost exactly the same in 2007 as in 1983. However, the Great Recession hit African-American households much harder than whites and the ratio of mean wealth between the two groups plunged from 0.19 in 2007 to 0.14 in 2010, mainly due to a 33 percent decline (in real terms) in black wealth. The relative (and absolute) losses suffered by black households from 2007 to 2010 are ascribable to the fact that blacks had a higher share of homes in their portfolio than did whites and much higher leverage than whites (debt-equity ratios of 0.55 and 0.15, respectively). These factors led to a wide discrepancy in rates of return on their respective portfolios (-9.9 versus -7.1 percent per year). From 2010 to 2013, the wealth ratio slipped from 0.14 to 0.13, despite the fact that the rate of return on the portfolio of black families was greater than that of white families (7.14 versus 6.12 percent per year). The results imply that black families had a substantially higher dissavings rate than white families.

Hispanic households made sizeable gains on (non-Hispanic) white households from 1983 to 2007. The ratio of mean net worth grew from 0.16 to 0.26, the homeownership rate among Hispanic households climbed from 33 to 49 percent, and the ratio of homeownership rates with white households advanced from 48 to 66 percent. However, in a reversal of fortunes, Hispanic households got hammered by the first half of the Great Recession. Their mean net worth plunged in half from 2007 to 2010, the ratio of mean net worth with white households fell from 0.26 to 0.15, their home ownership rate fell by 1.9 percentage points, and their net home equity plummeted by 47 percent. The relative (and absolute) losses suffered by Hispanic households over these three years are also mainly due to the much larger share of homes in their wealth portfolio and their much higher leverage rate (a debt-equity ratio of 0.51 versus 0.15). These factors led to a wide disparity in returns on their respective portfolios (-10.8 versus -7.1 percent per year). Another likely factor is that a high percentage of Hispanics bought their homes close to the housing cycle peak. From 2010 to 2013, their net worth ratio remained unchanged despite the fact that they had a higher return on their portfolio than did whites (7.5 versus 6.1 percent per year). The implication here too is that Hispanics had higher dissavings rates than did whites.

Young households also got pummeled by the Great Recession. The ratio of net worth between households under age 35 and all households, after falling from 0.21 in 1983 to 0.17 in 2007, plunged to 0.11 in 2010. In (real) dollar terms, their mean net worth declined by 46 percent from 2007 to 2010. Among age group 35-44, the ratio of their net worth to the overall figure fell from 0.71 in 1983 to 0.58 in 2007 and then declined precipitously to 0.42 in 2010. In dollar terms, their wealth fell by 39 percent over the latter three years. The same two factors explain the losses suffered by young households as for minorities – the higher share of homes in their wealth portfolio and their much higher leverage ratios. In terms of rates of return, the youngest age group had an annual return of -13.5 percent and age group 35-44 had a return of -9.6 percent compared to -7.3 percent for all households. The relative net worth of the under 35 age group did recover slightly to 0.12 in 2013 while that of age group 35-44 rebounded to 0.64. These trends mainly reflected the high annual rate of return on their wealth portfolio – 10.7 percent for the under 35 age group and 7.5 percent for age group 35-44 compared to 6.2 percent overall.

References

Blau, Francine D., and John W. Graham. 1990. "Black-White Differences in Wealth and Asset Composition." Quarterly Journal of Economics, Vol. 105, No. 1 (May): 321-339.

Board of Governors of the Federal Reserve System. 1998. "Flow of Funds Accounts of the United States: Flows and Outstanding Second Quarter 1998." (September 11).

Feldstein, Martin, and Shlomo Yitzhaki. 1982. “Are High Income Individuals Better Stock Market Investors?” NBER Working Paper No. 948, June.

Gittleman, Maury, and Edward N. Wolff. 2004. "Racial Differences in Patterns of Wealth Accumulation,” Journal of Human Resources, Vol. 39, No. 1, Winter, pp. 193-227.

Kennickell, Arthur B. 2001. "Modeling Wealth with Multiple Observations of Income: Redesign of the Sample for the 2001 Survey of Consumer Finances," October, http://www.federalreserve.gov/pubs/oss/oss2/method.html.

Kennickell, Arthur B., and R. Louise Woodburn. 1999. "Consistent Weight Design for the 1989, 1992, and 1995 SCFs, and the Distribution of Wealth." Review of Income and Wealth Series 45, No. 2, June, pp. 193-216.

Modigliani, Franco, and Richard Brumberg. 1954. "Utility Analysis and the Consumption Function: An Interpretation of Cross-Section Data." In K. Kurihara ed., Post-Keynesian Economics. New Brunswick, N.J.: Rutgers University Press.

Oliver, Melvin L., and Thomas M. Shapiro. 1997. Black Wealth, White Wealth. New York: Routledge.

Saez, Emmanuel, and Gabriel Zucman. 2014. “Wealth Inequality in the United States since 1913: Evidence from Capitalized Income Tax Data,”NBER Working Paper No. 20625, October.

U.S. Council of Economic Advisers. 2004. Economic Report of the President, 1994. Washington, DC: United States Government Printing Office.

U.S. Council of Economic Advisers. 2012. Economic Report of the President, 2012. Washington, DC: United States Government Printing Office.

Wolff, Edward N. 1980. “Estimates of the 1969 Size Distribution of Household Wealth in the U.S from a Synthetic Data Base,” in James D. Smith ed., Modeling the Distribution and Intergenerational Transmission of Wealth, Chicago: Chicago University Press, pp. 223-271.

-----. 1979. "The Distributional Effects of the 1969-75 Inflation on Holdings of Household Wealth in the United States." Review of Income and Wealth, Series 25, No. 2, June, pp. 195-207.

-----. 1987. "Estimates of Household Wealth Inequality in the United States, 1962-83." Review of Income and Wealth, Series 33, No. 3, September, pp. 231-256.

-----. 1994. "Trends in Household Wealth in the United States, 1962-1983 and 1983- 1989." Review of Income and Wealth, Series 40, No. 2, June, pp. 143-174.

-----. 1996. TOP HEAVY: A Study of Increasing Inequality of Wealth in America. New York: New Press.

-----. 1998. "Recent Trends in the Size Distribution of Household Wealth." Journal of Economic Perspectives, Vol. 12, No. 3, Summer, pp. 131-150.

-----. 2001. "Recent Trends in Wealth Ownership, from 1983 to 1998", in Thomas M. Shapiro and Edward N. Wolff eds., Assets for the Poor: The Benefits of Spreading Asset Ownership, New York: Russell Sage Press, pp. 34-73.

-------. 2002. TOP HEAVY: A Study of Increasing Inequality of Wealth in America. Newly updated and expanded edition, New York: the New Press.

------. 2011a. “Recent Trends in Household Wealth in the U.S.: Rising Debt and the Middle Class Squeeze,” in Jason M. Gonzales editor, Economics of Wealth in the 21st Century. Nova Science Publishers, Inc., pp. 1-41.

-----. 2011b. The Transformation of the American Pension System: Was It Beneficial for Workers? Kalamazoo, Michigan : W.E. Upjohn Institute for Employment Research.

------. Forthcoming. Inheriting Wealth in America: Future Boom or Bust? New York: Oxford University Press.

Table 1: Mean and Median Wealth and Income, 1962-2013

Table 2. The Size Distribution of Wealth and Income, 1962-2013

Table 3. Mean Wealth Holdings and Income by Wealth or Income Class, 1983-2013

Table 4. Composition of Total Household Wealth, 1983 - 2013

Table 5. Composition of Household Wealth by Wealth Class, 2013

Table 6. Composition of Household Wealth of the Middle Three Wealth Quintiles, 1983-

Table 7. The Percent of Total Assets Held by Wealth Class, 2013

Table 8. Share of Homeowners with Negative Home Equity

Table 9a. The Effects of Leverage on the Rate of Return:

Table 9b. The Effects of Differential Leverage

Table 10. Average Annual Real Rates of Return by Period and Wealth Class, 1983 -

Table 11. Household Income and Wealth by Race, 1983-2013

Table 12. Household Income and Wealth for Non-Hispanic Whites and Hispanics, 1983-

Table 13. Composition of Household Wealth by Race and Ethnicity, 2007

Table 14. Age-Wealth Profiles and Homeownership Rates by Age Group, 1983-2013

Table 15. Composition of Household Wealth by Age Class, 2007

Appendix Table 1. Average Annual Nominal Rates of Return

Figure 1: Mean and Median Net Worth, 1962-2013

Figure 2: Mean and Median Household Income, 1962-2013

Figure 3: Wealth and Income Inequality, 1962-2013 (Gini coefficients)

Figure 4: Composition of Household Wealth, 1983-2013 (percent of gross assets)

Figure 5: Composition of Household Wealth by Wealth Class, 2013 (percent of gross assets)

Figure 6: Ratio of Mean Net Worth Between Racial and Ethnic Groups, 1983-2013

Figure 7:  Ratio of Mean Net Worth of Young Age groups to Overall Mean Net Worth, 1983-2013

_______________

Notes:

1 The source for years 1989 to 2007 is Table 935 of the 2009 Statistical Abstract, US Bureau of the Census, available at http://www.census.gov/compendia/statab/. For years after 2007, the source is: National Association of Realtors, “Median Sales Price of Existing Single-Family Homes for Metropolitan Areas,” available at: http://www.realtor.org/sites/default/files/reports/2012/embargoes/2012-q1-metro-home-prices- 49bc10b1efdc1b8cc3eb66dbcdad55f7/metro-home-prices-q1-single-family-2012-05-09.pdf [both accessed October 17, 2014].The figures are based on median prices of existing houses for metropolitan areas only. All figures are in constant dollars unless otherwise indicated.

2 The source is: http://www.nber.org/cycles/cyclesmain.html [accessed April 20, 2014].

3 The source is the U.S. Bureau of Labor Statistics at: http://data.bls.gov/timeseries/LNS14000000 [accessed April 10, 2014].

4 The source for stock prices is Table B-96 of the Economic Report of the President, 2013, available at http://www.gpoaccess.gov/eop/tables13.html, with updates to 2013 from: http://us.spindices.com/indices/equity/spcomposite- 1500 [both accessed October 17, 2014].

5 These figures are based on the Bureau of Labor Statistics (BLS) hourly wage series. The source is Table B-15 of the Economic Report of the President, 2014,available at http://www.gpo.gov/fdsys/pkg/ERP-2014/pdf/ERP-2014- table15.pdf [accessed October 17, 2014]. The BLS wage figures are converted to constant dollars on the basis of the Consumer Price Index (CPI-U).

6 The figure is for civilian employment. The source is Table B-14 of the Economic Report of the President, 2014, available at http://www.gpo.gov/fdsys/pkg/ERP-2014/pdf/ERP-2014-table14.pdf [accessed October 17, 2014].

7 The source is Table B-12 of the Economic Report of the President, 2014, available at http://www.gpo.gov/fdsys/pkg/ERP-2014/pdf/ERP-2014-table12.pdf [accessed October 17, 2014].

8 These figures are based on the Federal Reserve Board’s Flow of Funds data, Table B.100, available at: http://www.federalreserve.gov/releases/Z1/ [accessed October 17, 2014].

9 Unfortunately, no data on educational loans are available before the 2004 SCF.

10 The computation of DB pension wealth is based on the present value of expected pension benefits upon retirement. See Wolff (2011b) for details.

11 See Wolff (2011b) for estimates of Social Security and pension wealth.

12 Unless otherwise indicated, all dollar figures are in 2013 dollars.

13 The percentage decline in median net worth from 2007 to 2010 is lower when vehicles are included in the measure of wealth – “only” 39 percent. The reason is that automobiles comprise a substantial share of the assets of the middle class. However, median net worth with vehicles remained virtually unchanged from 2010 to 2013.

14 The decline in mean net worth is 16 percent when vehicles are included in net worth.

15 This is not to say that there was no change in wealth inequality over these years. Indeed, on the basis of estate tax data, Wolff (2002) documents a sharp reduction in wealth inequality from about 1969 to 1976 and then an equally sharp rise from 1976 to 1983.

16 What was behind the sharp rise in wealth inequality? As I shall discuss in Section 9, there are two principal factors accounting for changes in wealth concentration. The first is the change in income inequality and the second is the change in the ratio of stock prices to housing prices. As we shall see below, there was a huge increase in income inequality between 1983 and 1989, with the Gini coefficient rising by 0.041 points. Second, stock prices increased much faster than housing prices. The stock market boomed and the S&P 50 Index was up by 62 percent, whereas median home prices increased by a mere two percent. As a result, the ratio between the two climbed by 58 percent.

17 This difference, by the way, shows the danger of relying on the share of the top one percent as a measure of inequality, as has been done in many studies, since it tracks differently over time than the Gini coefficient.

18 It should be noted that the income in each survey year (say 2013) is for the preceding year (2012 in this case).

19 It should be noted that the SCF data show a much higher level of income inequality than the CPS data. In the year 2000, for example, the CPS data show a share of the top five percent of 22.1 percent and a Gini coefficient of 0.462. The difference is primarily due to three factors. First, the SCF oversamples the rich, while the CPS is a representative sample. Second, the CPS data are top-coded, whereas the SCF data are not. Third, the SCF income definition includes realized capital gains whereas the CPS definition does not. However, the CPS data also show a large increase of inequality between 1989 and 2000, with the share of the top five percent rising from 18.9 to 22.1 percent and the Gini coefficient from 0.431 to 0.462.

20 The CPS data, in contrast, shows little change in household income inequality, with the Gini coefficient falling slightly from 0.470 in 2006 to 0.468 in 2009.The source is: http://www.census.gov/hhes/www/income/data/historical/household/2010/H04_2010.xls. However, the work of Emmanuel Saez and Thomas Piketty, based on IRS tax data, reveals a sizeable decline in income inequality from 2007 to 2010. In particular, incomes at the 99.99th, 99.9th, and 99th percentile dropped sharply over these years. The source is the World Top Incomes Database, available at http://topincomes.parisschoolofeconomics.eu/ [accessed Oct. 24, 2014].

21 It may seem surprising that the share of housing in gross assets declined very little between 2007 and 2010, given the steep drop in housing prices, but the prices of other assets also fell over this period, particularly those of stocks and business equity.

22 One possible explanation for this finding is that the least educated group is also the oldest group, who probably bought homes in the more distant past. This fact could explain their low incidence of negative home equity.

23 In 2007, the average house value was $207,600 and the average mortgage debt was $72,400, resulting in an average home equity of $135,200. If house prices decline by 24 percent and mortgage debt remains fixed, then average home equity falls to $85,400, for a decline of 37 percent.

24 This assumes that the prices of “other assets” remain unchanged.

25 In particular, it is assumed that there are no systematic differences in returns of, for example, stocks by wealth class. Though work on this issue is limited, there is one paper, in particular – Feldstein and Yitzhaki (1982) -- that found that high income investors received a higher rate of return on their investments than low income ones. However, the study, based on income tax returns, relied exclusively on capital gains realized on corporate stock.

26 An earlier analysis was conducted by the author for the 1969-1975 period in the U.S. See Wolff (1979) for details.  

27 The residual group, American Indians and Asians, is excluded here because of its small sample size.

28 It should be stressed that the unit of observation is the household, which includes both families (two or more related individuals living together), as well as single adults. As is widely known, the share of female-headed households among African-Americans is much higher than that among whites. This difference partly accounts for the relatively lower income and wealth among African-American households.

29 One important reason for the wealth gap is differences in inheritances. According to my calculations from the SCF data, 24 percent of white households in 1998 reported receiving an inheritance in 1998 or earlier, compared to 11 percent of black households, and the average bequest among white inheritors was $115,000 (present value in 1998), compared to $32,000 among black inheritors. Thus, inheritance differences appear to play a vital role in explaining the large wealth gap, particularly in light of the fact that black families appear to save more than white families at similar income levels (see, for example, Blau and Graham, 1990; Oliver and Shapiro, 1997; and Gittleman and Wolff, 2004).

30 There was almost no change in the relative home ownership rates of the two groups – both experienced moderate losses – while the share of households with non-positive net worth actually increased more in relative terms for white households than black ones.

31 This pattern may also be partly a cohort effect since 401(k) plans and other defined contribution plans were not widely introduced into the workplace until after 1989.

32 According to Wolff (forthcoming), net inheritance flows for middle class households are quite small on an annual basis.

33 Results are different for mean net worth. In this case, mean wealth fell by 16 percent between 2007 and 2010. Based on the annual rate of return for all households, it would have fallen by 20 percent. This was offset by a positive annual savings rate of 1.2 percent on initial mean wealth.

34 Results are similar for mean net worth. Mean wealth showed a slight increase from 2010 and 2013. Based on the annual rate of return for all households, mean net worth should have increased by 20 percent. This was offset by an annual dissavings rate of 6.6 percent on initial mean wealth.

35 Saez and Zucman (2014) also reported a substantial dissavings rate for the bottom 90 percent of the wealth distribution over years 2007 to 2010 – of the order of 5 percent on income. Their data series ends in 2011