[Abridged from "Survey Says..." by Karen Weaver and Samir Bhatt, securitized assets research, Credit Suisse First Boston]
Every three years, the Federal Reserve Board conducts the Survey of Consumer Finances, in which over 4,000 randomly selected U.S. households are interviewed about family assets, liabilities and income. A summary and interpretation of the latest survey was just made available and can be referenced on the Federal Reserve Board's website (http://www.bog.frb.fed.us).
The time period covered by this survey, 1995 to 1998, was a period of economic expansion, posting cumulative growth rates of 5% in median home prices, 18.8% in nominal GDP and 76% in the S&P 500. At the same time, mortgage rates declined from 7.4% to 6.9% and unemployment fell from 5.7% to 4.5%.
Against such an auspicious backdrop, not surprisingly, the U.S. household's financial health is generally quite strong though we see some indications of over-leverage, to wit:
*Household net worth is rising;
*Household financial leverage, as measured by debt-to-total assets, is declining;
*As measured by debt-to-total income, household financial leverage is increasing; and
*The trend toward overleveraging (debt-to-income>40%) in lower income tiers continues, but the 1990's also brought a notable jump in 40%+ debt to income ratios for middle America, i.e. the $25,000-$49,999 income tier.
The 1995-1998 survey shows continued growth in both mean and median income levels, to $53,100 and $33,400, respectively, in 1998. Inflation-Adjusted, household incomes now top levels reached in the 1980's expansion.
The survey also includes the percentage and families in each income tier. The breakdown was relatively stable form 1995-1998, although the < $10,000 income cohort shrunk, from 15% of families in 1989 to 13% in 1998, its impact on portfolios is even smaller than its 13% share of families might suggest, since balances owed by this cohort would be disproportionately smaller than their numbers. This < $10,000 cohort also often includes students, retirees, part-time workers, etc. (keep in mind that the survey's definition of "family" means that an individual may be counted as a family). The unusual financial circumstances of many in this cohort (e.g., virtually no living costs in the case of some students, significant assets in the case of many retirees), hinders analyses of this cohort's financial condition. This, coupled with the group's relatively small share in most consumer credit portfolios, leads us to suggest that ABS analysts should generally be less concerned about this income stratum.
Accordingly, we find statements such as, "nearly a third of all families earning less than $10,000 a year allocate more than 40% of that money toward paying debts," (Wall Street Journal, 1/19/2000) to be factually correct, but hyperbolic and potentially misleading.
As of 1998 family annual income was distributed as follows: 13% earn < $10,000 annually; 25% earn between $10,000-$24,999; 28% earn between $25,000-$49,999; 25% between $50-$99,999 and 9% earn $100,000 or more.
Household net worth (simply, total assets net of liabilities) rose quite sharply since the last survey, with median net worth posting a 17.6% increase between 1995 and 1998. Notably, median household net worth now exceeds the high-water-mark of the last expansion (1989) by a hefty 20% ($59,700 to $71,600). As with any data series, averages (and medians) can lie. Indeed, median net worth rose predominantly in income tiers over $25,000, and that rise was modest for all but, those earning in excess of $100,000 per year.
It is noteworthy, as well, that the net worth improvement is not due to a reduction in liabilities, or increases in (relatively stable) assets such as CD's, bonds or residential real estate (though homeownership rates did reach an all-time high of 66.2% in this 1998 survey). Rather, because these net worth improvements have been equity-market-driven, one might question the "quality" of this balance sheet improvement.
The survey data reveal some important trends for 1998 versus 1995 in the type and size of borrowings, to wit:
*A 42.3% increase in the median household balance, for all types of debt combined, to $33,300 in 1998 from $23,400 in 1995;
*In credit cards, the percentage of respondents carrying a balance dropped 3.2 points, to 44.1% among those who have a balance, the median amount owed was steady ($1,600 versus $1,700) among those with bank cards (with balances), limits rose from $8,000 to $9,500;
*In installment lending (largely, autos and student loans) the median amount owed rose 36%, to $8,700 this likely reflects increasing expenditures on autos; and
*In mortgage lending, mortgage debt rose 12.9%, though median home prices rose only 5.4% over the 1995-1998 period, supporting evidence of higher LTV ratios the hue-and-cry over home equity debt stealing market share from credit cards may be overstated. For example, the survey indicated that 41% of homeowners interviewed in 1998 had refinanced at some point, but only 26% of those took out cash, and only 20.8% of that was used for debt consolidation (41% x 26% x 20.8% = a mere 2.2%). Moreover, only 7% of households have equity lines, with 63.7% of those drawing on the lines (though this is up from 5% and 56% in 1995).
There are two vantage points from which to view household leverage, each with its own merits. The debt-to-assets ratio is what we call a "balance sheet approach," because it measures leverage by comparing the principal amount of debt outstanding to total assets. The debt-to-income ratio, which compares debt service obligations (i.e., principal and interest payments) to pre-tax income, is what we refer to as an "income statement approach."
As a general matter, we find the income statement approach to assessing household leverage more "on point" for ABS investors. The debt-to-income ratio more specifically addresses capacity to meet payment obligations, accounts for the term of the debt (i.e., the difference between owing $50,000 on a three-year loan versus a 30-year mortgage) and, because income is easier to measure than asset value, may invite less estimation error. That being said, debt-to-asset measures have macroeconomic implications, policy implications and, importantly for ABS investors, are often better predictors of bankruptcy filings. (Consumers are more resourceful about finding ways to repay debt if they have assets to lose in a bankruptcy proceeding.)
The good news is that the ratio of household debt-to-assets, finally drifted downward in 1998, to 14.4% after rising in the past three surveys (12.4% in 1989, 14.7% in 1992 and 14.7% in 1995).
Of course, as we've said, the improvement in this measure of consumer leverage largely reflects the strength of equity markets. If we recast the series by backing out equity holdings, debt to total assets rose to 18.4% in 1998, up from 17.2% in 1995, 16.3% in 1992 and 13.5% in 1989.
The debt-to-income ratios also show increasing leverage. Debt-to-income ratios (debt payments to gross income) over the past four surveys have risen from 12.7% in 1989 to 14.1% in 1992, 13.5% in 1995, and ended 1998 at 14.5%. Increased borrowing is even more pronounced in the data series that measures only that subset of households who do have debt where debt-to-income rose from 16.1% in 1995 to 17.6% in 1998.
A common benchmark for "overleverage" is debt-to-income ratios over 40%. Households above the 40% level merit concern, in that they are less able to shoulder interruptions or reductions in incomes (or the corollary, increased expenses) and continue servicing debt. The percentage of families who are overleveraged is growing. The picture broken down by income tier indicates that this overleveraged group is largest in the under $25,000 cohorts, but growing across income tiers, with the most striking increases in cohorts up to $50,000.
If the economy turns and enters a recession, the data indicate that the impact on consumer credit would likely be more severe than the 1991 recession, all else equal.
In toto, the survey is a positive one for the consumer. Incomes and net worth grew nicely, and, by one measure, leverage is down. Moreover, frankly, even a rise in leverage (as debt-to-income shows) is a fairly rational choice for the consumer faced with today's extraordinarily robust economy.
We find it useful to look at the consumer credit environment by income tier. As discussed, the < $10,000 income group is less important to investors. The $10,000$24,999 group concerns us, especially since nearly 20% of them are over-leveraged, higher than any other group, and they are the only group (other than < $10,000) that posted a decline in net worth from 1995 to 1998 - and 60% of this income cohort admit that they don't save (i.e., in survey terms, expenditures exceed income). This group has fewer assets to weather a storm, less wiggle room if fortunes turn, and is likely to be younger and less educated than average.
We recommend that investors address their exposure to this "at risk" group ($1024,999 income) by avoiding the risk (favor portfolios with higher income borrowers, and/or homeowners) mitigating the risk (shorter maturities, better structures. More credit enhancement, etc,) or, all else equal, require incremental yield.