Household Debt

I’ve said this before: Betting against the US consumer has been a losing wager. And yet when we look at the parabolic rise of non-mortgage houshold debt (via the Research dept. of the St. Louis Fed) one cannot help but imagine that it eventually will be  problematic.

The issue is one of timing.


Household Credit Market Debt Outstanding (non-Mortgage)
(billions of dollars)
1953-01-01 to 2005-01-01


UPDATE  June 11, 2005 12:02pm
Someone in comments suggested this chart merely reflects population trends. I doubt that. For one thing, it started at Zero in the mid-fifties, while the population was a bit higher than that. But more significantly is the exponential acceleration  of that curve — a true geometric gain would look like a straight line higher as the population expanded.

Not only are we becoming more indebted, but we are doing so at a more rapid pace. You can draw at least 3 separate trend lines on that chart — 1960-80; 1985-96 1999-2005 — each running at a higher rate than the prior trend.

Again, its been a losing wager betting that this ends tomorrow. However, that will not be the case forever. Whoever figures out exactly when the U.S. consumer gives up the ghost — with a Q or so — stands to make a lot of money betting on the downside, shorting retailers, banks and consumer cyclicals to name but a few.   

Household Sector: Liabilites: Household Credit Market Debt Outstanding
Board of Governors of the Federal Reserve System
Last Updated:     2005-06-09

Print Friendly, PDF & Email

What's been said:

Discussions found on the web:
  1. marshal’s ghost commented on Jun 11

    The geometric growth rate of this series makes me suspect that it is measured in nominal dollars and the magnitude of the series makes me suspect it is total housholds instead of per household.

    Population grows at a geometric rate. The nominal price level grows at a geometric rate. Even if debt per household was constant on a real basis the graph in question would quickly shoot off to infinity.

    Does this data exist on a per houshold and real dollar basis? If not, why not present it as a fraction of nominal GDP -which also grows at a geometric rate.

    I suspect household debt has increased, I just can’t tell from this graph.

    [EDITOR: Take that argument up with the St. Louis Fed — its their chart]

  2. Barry Ritholtz commented on Jun 11

    Remember, general credit cards were not available before the 1950s, and they were not truly widely held until the 1970s.

    Yes, you could get a revolving credit card from specific merchants — but there’s a huge difference between paying off a refrigerator on time and living on plastic, as we so commonly do now.

    That hockey stick growth pattern is as much a matter of historical adaptation of a new financial instrument as it is indebtedness

  3. Marshal’s ghost commented on Jun 11

    Lets try this again.

    In year 0 there are P people with $1 worth of year zero debt. Each year the population increases at rate p and inflation increases the price level at rate i.

    Hold per person debt fixed at $1 real dollar and at year t we have the following total nominal debt:


    Note that this number is growing geometrically because of the componding of inflation and population growth.

    All I’m saying is if you show us a graph of total nominal household debt over time you might want to divide by number of housholds (to get a per houshold debt) and express it in constant inflation adjusted dollars before concluding that houshold debt is growing at geometric pace.

    I’m sure houshold debt has gone up. On an inflation adjusted per houshold level it just hasn’t gone up 10-fold since 1980.

  4. Barry Ritholtz commented on Jun 11

    Feel free to pull that data and redraw the St.L Fed’s chart.

    But as long as we are picking nits, doing so would only show the MEDIAN household indebtedness — in other words, you would be building the assumption into your chart that indebtedness is smoothly distributed across income groups.

    We know thats not true, both as a percentage of net worth and as an ability to service that debt (i.e., bankruptcy filings by income group).

    Draw the same chart in deciles, and you may be astonished to see just how in the red large swaths of Mr & Mrs John Q. Public actually are.

    And from an investment perspective, you either believe they eventually crack and therefore you want to be on the short side of the trade, or that they are in fine condition, and you can therefore stay long.

    I’m not short – yet. But at some point in the future, I will be saying “Sold to you.”

    Its just that we aint quite there yet . . .

  5. Chris commented on Jun 11

    Another issue with consumer debt will be the recently past bankruptcy bill. Instead of people being able to clear their debt, and start consuming again, they will have to continue paying their creditors.

  6. Marshal’s ghost commented on Jun 11

    You have to take the data seriously. The difference between total and per capita and the difference between real and nominal matter. From your comments I assume you think the graph shows an alarming increase in the rate of debt growth. But this is nominal total data on a linear scale. Nominal data assures growth even if the real level stays fixed. A linear scale plays tricks with the eyes and makes you think the growth rate is accelerating even if its constant.

    To see this download the data and define the gross quarterly growth rate at time t as (Debt_t)/(Debt_t-1).

    If we take the average quarterly growth rate over the past 2 years you get 2.56% growth. Over the past 4 years you get 2.43% growth.

    Between 1995-2004 average quarterly growth was: 2.07%
    Between 1985-1994 average quarterly growth was: 2.06%
    Between 1985-1994 average quarterly growth was: 2.06%
    Between 1975-1984 average quarterly growth was: 2.74%
    Between 1965-1974 average quarterly growth was: 1.99%
    Between 1965-1974 average quarterly growth was: 1.99%
    Between 1955-1964 average quarterly growth was: 2.42%

    For the entire sample the average quarterly growth rate was 2.274% with a standard deviation of .815%.
    So the most recent quarter’s 2.33% growth rate is about ¾ of a standard deviation above the mean, the average growth rate of the past 2 years is 1 standard deviation above the average mean and the growth rate of the past 4 years is about 4/5 of a standard deviation above the mean.

    As for parsing the data into income decals etc. I suspect your too pessimistic about the plight of the lower income distribution as well. The BLS reports household expenditure by income quintile as well as household occupation, renter versus owner, mortgage versus non mortgage etc. The data is available at

    Search for average household expenditure by income quintile and after tax income by income quintile.

    If you find after tax income and expenditure by earnings quintile you will find that the poorest households are taking on a lot of debt but the poorest households have always taken on a lot of debt and the current difference between income and expenditure is lower then at any time in the past 20 years!

    In 1984 poorest 20% of households had a total expenditure of 250% of their after tax income.
    In 1994 poorest 20% of households had a total expenditure of 150% of their after tax income.
    In 2004 poorest 20% of households had a total expenditure of 124% of their after tax income.

    In 1984 richest 20% of households had a total expenditure of 84% of their after tax income.
    In 1994 richest 20% of households had a total expenditure of 74% of their after tax income.
    In 2004 richest 20% of households had a total expenditure of 69% of their after tax income.

    So household balance sheets have gotten better at the same time that total debt has skyrocketed! How could this be? Its because total debt is expressed as a nominal total value and not as a per capita real value. In short, the growth rate in households + the growth rate in inflation has outpaced the growth rate in borrowing so if we express debt as a per capita fraction of income it has fallen.

    None of this means we can’t get a credit crunch. After all poor households are spending more then they earn. But they have always spent more then they earn.

  7. touche commented on Jun 11

    Here’s consumer debt divided by GDP and disposable personal income:

    Year Q GDP DPI
    2000 1 68.3% 93.1%
    2000 2 68.5% 94.2%
    2000 3 69.7% 94.7%
    2000 4 70.5% 95.9%
    2001 1 71.5% 96.9%
    2001 2 72.4% 99.0%
    2001 3 73.8% 98.1%
    2001 4 74.7% 101.5%
    2002 1 75.8% 101.2%
    2002 2 76.6% 101.9%
    2002 3 77.4% 104.0%
    2002 4 78.9% 106.3%
    2003 1 80.1% 107.9%
    2003 2 81.4% 109.7%
    2003 3 81.6% 109.7%
    2003 4 82.1% 111.2%
    2004 1 82.9% 112.7%
    2004 2 83.5% 113.8%
    2004 3 85.0% 116.1%
    2004 4 85.7% 115.0%
    2005 1 86.2% 117.4%

    What a debt binge!

  8. RichL commented on Jun 11

    Two observations on the growth in credit card receivables.

    1. Have you been getting the same zero interest rate loan offers that I’ve been receiving unbidden in the mail? Perhaps some of the growth is a rational reaction to teaser rates.

    2. More ominously, interest rates on the existing debt are far higher than economic growth. For the less creditworthy, high-teens compounding interest rates are quite powerful. While the rolling (unpaid) loan gathers no loss, it does increase in magnitude at astounding rates.

  9. confused commented on Jun 11

    Where would be the information,”bankrupcy per thousand”. or something similer?

  10. touche commented on Jun 11

    BTW, the caption “Household Credit Market Debt Outstanding (non-Mortgage)” should be changed to drop the “non-Mortgage” as it includes both mortgage and “consumer” debt, the latter consisting mainly of auto loans and credit cards.

  11. james commented on Jun 12

    Are there any incentives to save? How many times must we be taxed when we engage in productive activities? So americans are already spending their future earnings in lieu of saving to avoid the taxes they’d pay on their savings to acquire the goods they desire. The low interest rate environment is like a carry trade for the consumer. Load up now, avoid taxation on your savings and finance now. Afterall, the interest payments will be lower than their taxes on that money if they saved. This is especially prudent for high income earners. Why save? Why pay taxes on savings to acquire the good they desire? The interest payments will be less than the taxes generated from savings.

  12. jck commented on Jun 12

    The chart is meaningless,it is in nominal $ and doesn’t adjust for population growth…on the same site look for Household Financial Obligations as a percent of Disposable Personal Income and Household Debt Service Payments as a Percent of Disposable Personal Income.As

  13. Barry Ritholtz commented on Jun 12

    If you think the St. Louis Fed chart is meaningless, then fear not — You can vote with your dollars.

    On the other hand, if you think it has ominous portents, then await the appropriate opportunity to bet the other way.

    Oh, and have a look at these charts showing the change in houshold liabilities.

    Debt is a problem, and its a growing problem. You can deny this all you want. My concern isnt the IF this is an issue, but rather the WHEN it finally impacts equities.

    Feel free to deny it. Someone’s gotta be on the other side of my trade . . .

  14. James Cameron commented on Jun 12

    Just a note regarding the comment that US population growth has been exponential over the years in question (1950, on). Many populations, particularly in Europe, are falling. Growth in the US has been more dynamic, but is certainly less than exponential. Here:

    My guess is the fed data – which is not particularly useful as presented – hasn’t been adjusted for inflation AND reflects greater consumer debt. Afterall, credit cards, lines of credit on the home, etc have increasingly become a much greater part of the social fabric over this period. No surprises the curve moves up in some fashion. Perhaps a more important issue than the actual shape of the curve for individual or household debt, is the ability of consumers to service it.

  15. seamus commented on Jun 14

    touche, thanks for doing the work. That’s quite a chart!

  16. omarthecat commented on Jun 16

    There is a fairly straightforward explanation for the rise in household indebtedness that has generally gone unrecognized: the rapidly increasing sophistication of the consumer finance industry. Innovations in the industry combined with low interest rates and new secondary markets for consumer debt have given consumers the equivalent of financial fast food: immediate gratification with payment due later. If we are foolish enough to assume consumers are financially responsible, we should look elsewhere for the ballooning debt loads. A more pragmatic view suggests this accounts for the lion’s share of the parabola.

  17. Mark T commented on Jun 16

    surely the big long run structural change has been home ownership. The increase in debt reflects the fact that people pay off a mortgage instead of renting. Their cash flow could theoretically look identical but their balance sheet is radically altered – both assets and liabilities. The great advantage the US consumer has however is that he can fix his rent for long periods of time. Here in the UK there is no such option and every move by the Bank of England either puts money in the consumer’s pocket or, more recently, takes it out. This is not to say that there haven’t been recent problems with floating rate debt and the froth in the housing flipping market, nor that consumer non mortgage debt does not present a possible problem. It’s just a question of scale. Debt service costs and their volatility are v important. If you are worried about US you should be really worried about the UK!

  18. David Bennett commented on Jun 16


    There were significany taxes on profits from interest and much higher taxes on capital gains back when savings rates were higher. So this isn”t the explanation.

    Marshall’s Ghost:

    Population growth was more rapid in the fiifties, I believe inflation was similar to what we’ve had in the last decade. Also income was lower and much of that population growth was kids requiring increased expences unmlike today’s growth based on immigrants who are productive when they hit society.

    I notice you don’t plug in numbers to check your “truths.” Off the top of my head I can tell you that overall increases in this debt exceeed GNP growth, GNP is roughly 10 times what it was 30 years ago and while it’s difficult to get exact numbers, there is definitely an increase, an increase which continues through prosperous periods as well as poor. We should have seen the sharpest increases in the seventies, we don’t.

    The notion that the poor borrowing more than they can afford and always doing so is questionable. The poor tend to borrow from check cashing places, as with just about everything else they pay more, the loans are very short term and limited. They simply don’t explain the increase in numbers. And *if* your claim that the poorest householdsspent 240% of their income ion 1980 is true, simple common sense would tell you that this came from government and other sources, not private credit which is what’s being measured. I will tell you a secret, private lenders do not give money to peple unless they profit. In your imaginery world a family making $5,000 in 1984 could borrow $12,000 that year, the borrow some more the next and keep on borrowing year after year and just joyously rack up huge debts, but it ain’t happening. They are borrowing at the check cashing places and paying month by month, often slowly getting sucked in deeply.

    Nor does it make economic sense that an increasingly wealthy population would take on short term debt at frequently absurd rates. Some of this may be zero percent financing for one year, but huge portions on credit cards at 10 to 20%.

  19. Matt Miller commented on Oct 9

    Why consider debt in a sole context ? …I would think it would be good to express household debt with a overlay of both annual income and total assets.

  20. Barry Ritholtz commented on Oct 9

    Total assets fluctuate — consider assets in 1999/2000, when equity prcies topped.

    Home prices have appreciated dramatically; if prices pull back even modestly, while debt stays the same, then this ratio (debt to equity) spikes.

  21. brock easton commented on Feb 7

    1)What is the household debt over the past 5 years.?

    2)Analyse the behaviour of banks and other financial institutions, and the ease with which menoey is made available to consumers by them?

  22. Eoin commented on Apr 10

    This data (contrary to the title inserted on the graph) actually seems to include mortgage debt — mortgage debt at the end of 2006 stands at $10 trillion and other consumer credit at $2.4 trillion.

    While that is a fundamental reporting error (on your part), the overall argument is still valid as evidenced by that cumulative debt against other key economic measures.

  23. Barry Ritholtz commented on Apr 10

    I am under he impression this is credit debt )non securitized), not asset backed debt such as a mortgage.

    If you can find information showing this includes mortgage debt, it would be appreciated . . .

Read this next.

Posted Under