The full text on this page is automatically extracted from the file linked above and may contain errors and inconsistencies.
SPECIAL ISSUE THE FEDERAL RESERVE BANK OF CHICAGO SEPTEMBER 2002 NUMBER 181a Chicago Fed Letter The household balance sheet—Too much debt? by François Velde, senior economist Are U.S. households carrying too much debt? Debt levels have grown recently to unprecedented levels. But that is nothing new, as debt has long been following an upward trend. Rising debt can be a good thing, signaling improved access to credit. For aggregates, what matters is net worth. That has risen in the late 1990s and fallen recently, but consumption had not kept up with the rise, and may not follow the fall. Much concern has surfaced recently about the balance sheet of the household sector in the U.S. economy. One common worry is that households are excessively indebted and may be forced to cut back on their consumption, with negative consequences for economic growth. The purpose of this Chicago Fed Letter is to place this concern in perspective. 1. Assets and liabilities of the household sector Record levels of debt ratio to GDP (log scale) 5 4 3 assets 2 1 0.9 0.8 0.7 0.6 0.5 0.4 liabilities 0.3 0.2 1955 SOURCE : '60 '65 '70 '75 '80 Federal Reserve Board, Z1 Release. '85 '90 To do this, it is useful to look at both sides of the aggregate balance sheet of the household sector over a long period. Since total assets and liabilities are nominal amounts, we need to scale them so as to make them somewhat comparable over time. One method is to take such factors as '95 2000 inflation and population growth explicitly into account and look at per capita constant dollars. Another way, which turns out to give the same overall picture, is to measure assets and liabilities compared with gross domestic product (GDP), as in figure 1. Figure 1 makes clear that the ratio of assets to GDP is roughly constant. This means that the value of assets grows over time at roughly the same rate as the economy. That is not true of liabilities, which have been growing steadily over the past 50 years, from 24% to 78%. As a result, the ratio of assets to liabilities has fallen from 14 to 6. It is worth emphasizing that this growth of liabilities is not a recent phenomenon, but rather a secular trend. Thus, while it is correct to say that in the fourth quarter of 2001, household debt reached unprecedented levels, that is not, in fact, very newsworthy. The fact is household debt reaches record levels every other quarter on average (44% of the time to be exact). Thus, high levels of debt are not informative for current economic conditions. We may also note that the growth in liabilities has been uneven over time, and distinguish three periods: an initial rise until 1965, then a period of stagnation until 1984, and another rise since 1984. In that last period, liabilities increased 62% relative to GDP, or 2.7% per year on average. Compared with this last figure, the growth rate of liabilities during the last five years (2.9%) is not out of the ordinary. A final observation one can make is that both ratios display little or no cyclical pattern: They do not rise or fall markedly in recessions or in expansions. Of course, this could simply mean that the numerators and denominators of transfer resources from debt. Mortgage debt grew from 40% times and circumstances in of disposable personal income in 1984 ratio to DPI (log scale) which they are less needed to 73%. 2 to those in which they are The increase in mortgage debt can be value of household more needed. Typically, real estate attributed to two factors. One is simply younger households have that the rate of home ownership has 1 less resources than older been increasing, from 64% in 1994 to 0.9 households, yet they need 0.8 68% at the end of 2001. For those con0.7 housing just as much, if not sumers who have switched from being 0.6 more. They borrow to renters to homeowners, the debt bur0.5 bring forward part of the den stemming from the mortgage mere0.4 future income they will enly replaces the rent they paid formerly. joy in their prime, and use mortgage debt 0.3 Thus, in terms of their budget, the init to buy a house now. Ancreased debt burden puts no additionother example is the use of 0.2 al pressure on their consumption. 1955 '60 '65 '70 '75 '80 '85 '90 '95 '00 debt to make up for sudSOURCE : Federal Reserve Board, Z1 Release. den falls in income and The other factor is the combination of maintain a relatively stable smaller down payments on initial mortlevel of consumption. In the real gages and increased use of home equity the ratios are moving together. Indeed, world, there are constraints when one compares real rates of growth of liabilities and GDP, there is a relation- on people’s ability to 4. Net worth of the household sector ship. But, as it turns out, the former is a borrow, due to various inratio to GDP formational and legal diffilagging indicator of the latter. In other 4.75 words, reductions of households’ liabil- culties. Lenders may be reluctant to lend because ities follow rather than precede downthey can’t be sure that borturns. Nor are such high levels related 4.25 rowers will be able or willto economic performance over longer time spans. For example, liabilities as a ing to repay. In this 3.75 context, a relative growth ratio to GDP remained about constant in indebtedness indicates from 1965 to 1984, but increased 62% that these constraints are from 1984 to 2002. Yet the average an3.25 being loosened, and that nual growth rate of GDP was the same financial intermediation is over these two periods (3.2%). becoming more efficient 2.75 Debt is good over time. In particular, an 1952 '56 '60 '64 '68 '72 '76 '80 '84 '88 '92 '96 2000 improved ability to borrow Debt has been growing steadily over SOURCE: Federal Reserve Board, Z1 Release. means that consumers can time. That’s good news. Household better plan their consumpdebt is a means for households to lines of credit, both of which represent tion and smooth it over greater access to credit. This allows time. Thus, shocks to in3. Household debt burden homeowners to use their homes as come ought to have less of an impact on consumption collateral for the purpose of smoothshare of disposable personal income 15 ing the variations in their overall conthan before. sumption, over the life cycle as well as Household liabilities, as of total for other reasons. Figure 2 shows the the first quarter of 2002, 12 value of household real estate and the are made up essentially of value of mortgage debt, each as a ratio mortgage debt (67%) and of disposable personal income. Because consumer credit (21%). 9 the scale is logarithmic, the difference consumer credit These shares have been between the two lines corresponds to relatively stable, and they the percentage of homeowners’ equi6 tend to mirror each other. ty. The two lines are almost parallel in Since the 1970s, however, mortgage debt the last five years, indicating that this the share of mortgage debt percentage has been stable. As house 3 has been growing. Since 1980 '82 '84 '86 '88 '90 '92 '94 '96 '98 2000 '02 values have increased, so has mortgage total debt itself has been SOURCE : Federal Reserve Board, Z1 Release. debt in the same proportion. growing, so has mortgage 2. Household real estate and mortgage debt 5. Nondurables and services consumption nondurables and services consumption/after-tax labor income ratio 1.14 1.10 1.06 1.02 0.98 5. 0 5 .5 6. 0 6 .5 7. 0 7 .5 SOURCE: Arguably, such a transfer might affect the consumption level of debtors and creditors differently. Suppose, to take an extreme case, that the debtor is forced to absorb the increase in debt burden one-for-one through a cutback in consumption, but that the creditor increases consumption only fractionally: The difference represents a fall in aggregate consumption. This difference, however, is not lost to the economy as a whole. It will become an increase in saving (by the creditor) and, therefore, an increase in investment. Another concern is bankruptcy. If the debtor’s situation really worsens, he might default. In principle, however, defaults actually increase consumption, since they free up debtors’ income. 8. 0 net wealth/after-tax labor income ratio Federal Reserve Board, Z1 Release. The bottom line is that consumers are indeed more indebted. I have given several reasons why this could be seen as a “good thing.” Nonetheless, a picture like figure 3, which shows the debt burden as a percentage of disposable personal income, might raise concerns. The data series provided by the Federal Reserve Board only go back to 1980. They show two major upswings in this debt burden; it now stands at 14.05%, not far below the previous peak of 14.38% in the third quarter of 1986. If interest rates rise abruptly, might consumers find this debt burden too high and then cut back on consumption? It is worth keeping in mind at this point that, in a closed economy, one person’s liability is another person’s asset. If an event or a development has a negative impact on the debtor, it has a positive impact on the creditor. What counts for aggregate consumption is the net effect. In particular, increases in interest rates are just a transfer from debtors to creditors. Debtors reduce their consumption, but creditors increase theirs. 8 .5 The other side of the sheet Ultimately, consumers base their decisions on net wealth. If liabilities increase but assets 6. Nondurables and services to household net worth increase even more, then consumption/total wealth (percent) consumers in the aggre18 gate are wealthier and can spend more. 16 14 12 10 1952 ’57 ’62 ’67 ’72 ’77 ’82 NOTE : ’87 ’92 This measures the ratio of consumption of nondurables and services to household wealth on a quarterly basis. SOURCE: Federal Reserve Board, Z1 Release. ’97 2002 Household assets at the end of the first quarter of 2002 were around $48 trillion, about 4.6 times GDP. Onethird of this is tangible, mostly real estate (28%) but also consumer durables (6%). The other twothirds are financial—bank deposits (10%), bonds (4%), non-corporate equity (10%), directly held equity (12%), pension fund and life insurance reserves (20%), mutual fund shares (6%), and others (4%). Almost half of the last three categories consists of equity, indirectly held. All told, corporate equity (directly and indirectly held) represents 27% of assets, about $13 trillion as of the end of 2001. Liabilities total about $8 trillion, leaving net worth at $40 trillion, about 3.8 times GDP. The share of equity in total assets has varied considerably over time, averaging 18%, and ranging from around 10% in the mid-1980s to a peak of 36% in 1999. In recent months, the market value of equity has fallen considerably. As a rule of thumb, one can divide the S&P 500 index by 40 to get the percentage share of equity in total assets, or by 33 for the effect on net worth. Figure 4 shows net worth relative to GDP, up to the end of March 2002. Using the rule of thumb, that ratio, which then stood at 3.84, would be around 3.57 at the end of July. Substantial variations in the value of equity inevitably bring up the matter of wealth effects. Consumption is usually thought to be based on consumers’ perceptions of their long-term wealth, both human (as manifested through labor income) and nonhuman (or net worth). It is natural to think that consumption on one hand and labor income and net worth on the other Michael H. Moskow, President; William C. Hunter, Senior Vice President and Director of Research; Douglas Evanoff, Vice President, financial studies; Charles Evans, Vice President, macroeconomic policy research; Daniel Sullivan, Vice President, microeconomic policy research; William Testa, Vice President, regional programs and Economics Editor; Helen O’D. Koshy, Editor; Kathryn Moran, Associate Editor. Chicago Fed Letter is published monthly by the Research Department of the Federal Reserve Bank of Chicago. The views expressed are the authors’ and are not necessarily those of the Federal Reserve Bank of Chicago or the Federal Reserve System. Articles may be reprinted if the source is credited and the Research Department is provided with copies of the reprints. Chicago Fed Letter is available without charge from the Public Information Center, Federal Reserve Bank of Chicago, P.O. Box 834, Chicago, Illinois 60690-0834, tel. 312-322-5111 or fax 312-322-5515. Chicago Fed Letter and other Bank publications are available on the World Wide Web at http:// www.chicagofed.org. ISSN 0895-0164 should remain roughly in line over time. Researchers have found such a relation in the long run. Steindel and Ludvigson estimated the coefficient on net worth to be about 4.6%, based on data up to 1997.1 Using the most recently available data (up to the first quarter of 2002), the same methodology yields an estimate of 2.9%. How can one explain this change? The answer is in two parts. First, it is imprecisely estimated in the data. Second, recent years have shown consumption to be abnormally low given the long-run historical relation between consumption and wealth. As a result, the relationship between consumption and wealth weakens if recent numbers are included. This departure can be illustrated by plotting the ratio of consumption to income against the ratio of wealth to income, as shown in figure 5.2 The last 22 observations (corresponding to the period 1996 to present) are plotted in black. The large upswing in the wealth– income ratio (rightward movement in the figure) was not accompanied by a corresponding increase in the consumption–income ratio (upward movement). One way to read the figure is that, relative to their historical behavior toward their total wealth, consumers have ignored the run-up in their nonhuman wealth of the late 1990s. A simpler but more vivid way to make this point is simply to ignore income and compare consumption and wealth directly, as shown in figure 6. The historical mean for this ratio since 1952 is 15.3%. The recent falls in wealth, which were accompanied by only a mild slowdown in consumption growth, brought the ratio to 15.1% as of the first quarter of 2002. Figure 6 provides a quick way to assess the import of further falls in net worth. Using the rule of thumb given above, a 200-point fall in the S&P 500, for example, would reduce net worth by 6%. If consumption remained at the same level, it would raise the consumption–wealth ratio to 16%, a number that can be compared with the historical record. Conclusion Overall, neither the assets side nor the liabilities side of household balance sheets seems grossly out of line with the past. The reduction in wealth due to equity price movements appears to have brought wealth back in line with consumption. On the liabilities side, secular increases in debt should be placed in perspective against even greater increases in assets; this trend can be seen as evidence of improved intermediation, hence reduced dependence of consumption on income. The degree to which financial distress can affect aggregate consumption depends on the comparison between debtor and creditor behavior, and it is not obvious that high debt burdens (such as they presently are) signal an imminent curtailment of consumption. 1 Sydney Ludvigson and Charles Steindel, 1999, “How important is the stock market effect on consumption?,” Economic Policy Review, Federal Reserve Bank of New York, July, Vol. 5, No. 2. 2 To be consistent with Ludvigson and Steindel, I plot the same measures of consumption (nondurables and services) and income (after-tax labor income). Using total consumption or disposable income yields similar pictures.