The Vanishing Consumption Gap

There’s an often-quoted statistic that the ratio between the average incomes of the richest and poorest quintiles of Americans is 15 to 1. Earlier this year I stumbled over some research (“You Are What You Spend”) indicating that there is less to that difference than meets the eye. According to the authors, the difference in actual annual spending (as opposed to annual income) falls to 4:1, apparently because lots of people have sources of spendable cash that don’t show up as annual income (asset sales, securities not subject to capital gains taxes, insurance policy redemptions, and so forth).

But it gets better. If you adjust for size of household, the consumption ratio between richest and poorest quintiles drops to 2.1:1. They note that the average person in the middle income quintile consumes just 29% more than the average person in the lowest quintile. American spending patterns look dramatically more egalitarian than the raw numbers on income distribution would suggest. What the heck is going on here?

On these numbers, it doesn’t sound like being rich actually buys you a lot. OK, you get to save a lot of money for the future, while poor people live paycheck to paycheck and spend a significant fraction of their income on debt service. That’s significant. There are other ways to quarrel with this analysis, too; averaging over quintiles arguably underweights inequality at the extreme top and bottom ends of the distribution.

Still, that 2.1:1 figure is pretty startling. That gives us a factor of 6 difference between income spread and consumption spread over at least 80% of the income distribution, which suggests that something is operating that drastically equalizes Americans’ consumption power. And some other papers that I found by googling for “consumption inequality” actually back this up with different lines of evidence. The second and third I read noted that while U.S. wage inequality has increased significantly since 1970, consumption inequality has not. Both attribute this to the buffering effect of consumer credit on wage volatility.

I think the authors of “You Are What You Spend” are on to something more basic, though. There simply isn’t much stuff left that is so expensive that rich people can buy it but poor people can’t. This is true even for classic rich-person luxury goods. Proles buy wedding diamonds in strip malls. Caviar is no big deal; I know a woman who has carried it as backpack rations becuse the food-value-to-weight ratio was so good, which might sound like a rich-kid story until I tell you she’s a calligrapher who makes a somewhat precarious living doing piecework. And I know a non-wealthy programmer who recently bought himself an airplane.

More prosaically, look at what “poor” people have. Refrigerators. Cars. Televisions. Cell phones. Computers. I wrote about this in 2003 in an essay Mobilizing the Poor and Other Delusions. In that essay I was arguing that the goods consumption of all Americans above homeless drug addicts on the SES scale is so high as to make “poverty” a term that cannot meaningfully be applied in the U.S. Not if you’ve lived, as I have, in places where they have the real thing.

The data on consumption spreads demonstrate something else: not just that “poverty” is a silly word to use in the U.S., but that the degree of “inequality” we have here barely moves the needle off the peg on the historical misery meter.

The authors of “You Are What You Spend” explain this mainly by noticing that consumer goods have plummeted in price while rocketing in quality. Basically, everything is cheap now. Even with the recent spike in the price of oil, for example, Americans spend less on gasoline in constant dollars than they did in 1971. Food, clothing, housing, and other basics also cost drastically less than they did as recently as my teenage years, continuing a 150-year trend. Rates of home ownership are at an all-time high, and can be expected to continue increasing despite the current mortgage flap.

Heck, in 1928 food was so expensive that “a chicken in every pot” was a presidential campaign slogan. That’s right: chicken was a luxury good. If you find that bizarre and hard to believe, wake up. You just learned something about how wealthy you are.

The authors of “You Are What You Spend” get the first-order consequences right; the cheaper consumer goods get, the more the consumption gap between rich and poor will tend to vanish even if income inequality is flat or rising. Having lots of pictures of dead presidents matters less when you only need a relative few of them to live comfortably.

A more interesting question is: why is this happening, and can it be expected to continue?

At this point the authors of “You Are What You Spend” handwave vaguely in the direction of free markets. They’re not wrong — of course the reason we can buy cheap food and clothes and electronics is because of market-driven innovation, price competition rewarding efficiency of production, economies of scale, yadda yadda yadda.

But I think there’s a more fundamental and often missed cause, one that leads to strong predictions once you notice it. It’s what Bucky Fuller called “ephemeralization”, the substitution of design information for material and energy costs. Consider, for example, the difference between a computer in 1950 — a multi-ton behemoth — and a laptop today. The laptop uses design information to substitute ounces of silicon and plastic for thousands of pounds of steel, glass, and rare earths. This substitution has a ripple effect through all the transport, energy, and opportunity costs associated with producing and using computers.

This suggests that we can expect the consumption gap between rich and poor to continue closing — regardless of what income inequality does — as long as we can continue finding cleverer ways to arrange atoms. Or optimize supply chains. Or write risk-spreading financial instruments. Or, to anticipate one objection, find cheaper ways to make synthetic fuel from genetically tailored algae (algae genomes are design information, too).

The vanishing consumption gap has political consequences as well, but I’ll save those for another post.

26 comments

  1. Data like this suggests that the post-scarcity economy is not too distant, perhaps within our lifetimes. Eventually the consumption ratio hits 1:1, at which point further innovation contributes toward shortening our workday.

  2. “It’s what Bucky Fuller called “ephemeralization”, the substitution of design information for material and energy costs.”

    This I think follows from the law of diminishing marginal utility – and if it does it leads to some quite surprising predictions. The progress of technology since the Ind. Rev. could – very roughly – be described as throwing more and more materials into stuff (the age of steel), then when the additional marginal utility gained became very low then throwing more and more information/knowledge into products (the age of electronics), and then when even that started to dimish information became a lucrative product in itself (the age of software).

    The point is, there isn’t much beyond that… Instead of an accelerating technological change and the coming of the Big Transhuman Rapture (y’know, singularity etc.) it sort of looks like a decelerating technological change.

    This sounds a bit counter-intuitive as the vast majority of patents ever registered were registered in the last couple of decades, but the question is how important they are, especially from the point of view of the end-user experience. Today’s cars are more efficient, comfortable and safer than the cars of the seventies but the general experience of driving them is not so different. The Internet is great but it’s basically an instant version of the mail-order shop, library and magazine subscription – it’s nice that now the latency is measured in seconds instead of days, but I think back in the time when the printing press and the railroads were invented, and in those times were the latency reduced from anywhere between infinite and a year to a couple of days was perhaps more important than when the Internet reduced the latency from days to seconds.

    And how much more it can get? When we soon get to the point that every science, art, product etc. is available online in seconds there isn’t much further than that. And that can be extrapolated to other fields. If cars, airplanes and trains did not get radically better from the end-user point of view in the last 30 years why should they do so in the next 30-50 years? I think the law of dimishing marginal utility means we sooner or later run out of ideas on how to make anything really, significantly better and will keep on polishing and tinkering with basically the same stuff forever. And that can also mean that Malthus will have the last laugh, but I hope to be wrong on this particular one.

  3. To anticipate and summarize the gist of your future post on political consequences: the closer we get to this day, the less we give a damn about capitalism vs. socialism, and tend toward the gift economy a la Homesteading the Noosphere.

  4. >I think the law of dimishing marginal utility means we sooner or later run out of ideas on how to make anything really, significantly better and will keep on polishing and tinkering with basically the same stuff forever

    Maybe. But supposing that’s true, it’s not a problem for where I’m going with this analysis if the consumption spread effectively disappears before technological change stalls out. Which, at the rate the spread has been decreasing in recent historical time, looks extremely likely.

    >To anticipate and summarize the gist of your future post on political consequences: the closer we get to this day, the less we give a damn about capitalism vs. socialism, and tend toward the gift economy a la Homesteading the Noosphere.

    Actually, no. My projection is more complex than that. But you’ll see when I get there.

  5. Hi esr,

    A study based on income and spending doesn’t say anything about life quality : what if the minimum bearable to live is 1000$/month but for some it is extra easy to obtain and they decide not to work a lot (computer scientists, for example), and some really struggle to earn them and the term surviving is correctly employed here ?
    What i mean is you shouldn’t forget that money is _not_ a correct way to measure richness. It’s only an element, as the time needed to earn that money.

    People in cuba, france, and some other countries, have low income and low spending, included about medication, but they have access to total health care almost freely. This doesn’t show in studies in income and spending. But the same service, if it should be chosen by an american guy, would cost him a lot.
    Ergo : studies about income and spending are incomplete by definition to measure richness.

    Income include the ability to invest and so to be protected in later life, for one and his family. Spending doesn’t.
    So spending studies allow to think about consuming (consume == destroy), but not about investing. That’s only one part of people’s life.

    A provocative thought : if you need to work to live, then you are a slave. If you don’t, then you are a free man (you may like your enslavement, i do). Spending deals with consuming and do not make differences between a slave and a free man. Income deals with investing and does. That’s why income studies matter more, too.

    So the real question is : what do you want to measure ?
    If it’s spending, spending studies are good for you.
    If it’s income, income studies are good for you.
    If it’s richness, neither is enough for you.
    If it’s well-being, neither is enough for you.
    If it’s egality among people, neither is enough for you.

    Ok, i read your article another time, and then i add this : if the consumption gap declines more, then it should be interesting to think about the consequences on investing. Will more people be able to invest ? What will be the result on the globe society ? Will the investing studies matter more ?

  6. “Actually, no. My projection is more complex than that. But you’ll see when I get there.”

    I hope you can somehow include status goods as I’ve yet to see any sensible take on them. I’m thinking something among the following lines: the very nature and function of status goods is that they are consumed to express, show represent non-equality, therefore, paradoxically, when basic needs are fulfilled and thus more and more effort is put into acquiring status goods, then perceived, subjective non-equality can actually get bigger, meaning more envy, more ressentiment, more popularity for egalitarian ideas etc. etc.

  7. One additional factor making the consumption gap less than the income gap is that most wealth transfer programs are not counted as “income”. Subsidized housing, Medicare, EITC, and (IIRC) food support payments all improve the consumption of the lower fifth, but don’t show up in the income statistics at all.

    I would be very interested in knowing how education consumption fits into all of this, how they weight for education quality, and whether they disentangle education-as-learning vs. education-as-signalling. The lower two fifths can’t afford the credentials provided by an Ivy education, even if all of the actual knowledge and skills provided by an Ivy education are cheaply available either at lower-cost schools or on-line.

  8. Actually, Dave, Ivy League educations may become more affordable than the average state university, if a recent trend pans out: Harvard recently instituted free schooling for students who qualify, academically and economically, and other universities in and out of the Ivy League are moving in that direction. This is mostly in response to loud and growing complaints about their enormous endowments that go basically untaxed.

  9. I had noticed this effect as well, that the public housing townhouse was not significantly different from my rather large house, except in room size. It had all the same rooms and appliances; mine were just bigger and better. (Yeah, there are days when I think I should just go live there…) Likewise, we eat the same food and shop at XMart. The difference is slightly in quality. It’s hard to pick out the barely-making-it from the doing-really-well on the street.

    The difficulty is that it’s smoke and mirrors. It’s based on food being readily available, and affordable, which is about to collapse along with the debt market: Farmers won’t be able to get loans to buy seed and diesel fuel, which will still be priced by BigAgriculture and BigOil, so the crop will cost more to plant and harvest than the farmer can get for the crop, which means either more farm subsidies or the collapse of the American farming industry as we know it. How are *your* farming skills? I’m hoping my Amish cousins will bail me out, frankly.

    Likewise, those goods are so cheap because they’ve come from X-world countries, where the currency exchange rate meant that a reasonable wage over there was 1/10 or less than here. That’s changing as we speak, so go buy whatever you’re going to want to have in the next ten years *now*. Your money is worth more right now than it will be worth in the future, if current inflation models are to be believed.

    Fortunately, you have what will allow you to continue to survive: Good skills that transcend the basic marketplace. You will continue to find a way to trade your skills for food, clothing and shelter. The less fortunate will not be able to do that.

  10. It seems to be hard to measure this with much useful precision once the difference gets small. E.g., I see a probable slop factor of 1.2 or so in the way that high income regions are correlated with high cost-of-living regions, a factor of 1.1 or so in the underground economy stuff that isn’t covered in official consumption figures, and 1.3 or more if you try to account for the economic value of leisure time. Uncertainties of that scale bite pretty hard into a figure like 2.1:1.

    (None of this takes much away from the main point that 4:1 or 2.1:1 is very different from 15:1. Or from the consequence that now that everyone appreciates how cherry-picking quintile statistics gives wildly different results, people who present their favorite cherry as political ammo without even mentioning competing ways to slice the data should lose a lot of credibility. At least assuming they hadn’t already lost all credibility by goofiness like using un-inflation-adjusted statistics, or using statistics for changes in nominal wages without mentioning changes in fringe benefits…)

    Roxanne: I think you’re being too pessimistic. We are multiple doublings of per-capita GDP above basic sustenance. We could triple the price of multiple basic inputs (including perennial anxiety favorites like energy, various foodstocks, or fresh water) without being driven into “you will continue to survive … good skills … but the less fortunate will not be able” territory. (That’s “we” in the reasonably-developed world. The bottom billion could take it in the chin, but only in part because of the fundamental technical tradeoffs imposed by physical reality, and more because of the dysfunctional institutions that are their political reality.) And I judge it unlikely that any of those prices will stay tripled very long (in the absence of some huge political disaster, like WW3). In theory a tripled price is a very effective incentive both for conservation and for increased production, and lo, in historical fact, fear of sustained tripled prices seems far more common than actual sustained tripled prices.

  11. >A study based on income and spending doesn’t say anything about life quality

    Worrying about “life quality” is a luxury of the rich. By the time asking that question is meaningful, you already live better that 99.9% of humans ever have.

    >People in cuba, france, and some other countries, have low income and low spending, included about medication, but they have access to total health care almost freely.

    This sentence is wrong in at least two ways; France is a G8 country and not “low income” by any measure, and the wonderfulness of Cuban health care is a myth. See this or this.

    I could refute the rest of your claims point by point, but the premise of your whole post is silly. Only those fantastically wealthy by any historical standard can afford to even dream of making the kinds of distinctions you claim are crucial. Even now, most of the world’s population would consider you a spoiled rich kid for pooh-poohing the consumption convergence in favor of worrying about whether people have enough to invest – and they’d be absolutely right.

    Fast-forward a decade or two and people with your lack of perspective will be worrying, in an equally silly way, over the fact that while almost everyone has enough to be part of the investor class, some people can’t afford to choose lower rates to invest only in ethically-approved corporations.

  12. >The difficulty is that it’s smoke and mirrors. It’s based on food being readily available, and affordable, which is about to collapse along with the debt market: Farmers won’t be able to get loans to buy seed and diesel fuel, which will still be priced by BigAgriculture and BigOil, so the crop will cost more to plant and harvest than the farmer can get for the crop, which means either more farm subsidies or the collapse of the American farming industry as we know it

    Roxanne, I know you’re smarter than this; what happened to your judgment and ability to check facts? Small family farmers are a huge emotional deal in the U.S. but they’ve been declining as a productive factor for decades now; according to Wikipedia, they account for just 27% of production today. The other 73%, whether they’re labeled “family” or not, are agribusinesses that fund their seed and fuel purchases from previous revenue, not loans. It has to be this way, because risks from weather and the volatility of produce prices are high enough to make farming on loan unsustainable for for than a handful of years, even when we’re not having a debt crisis.

    Even if the debt crisis wiped out 27% of our farming capacity in one year, what of it? A 30% increase in food prices would still not push them above the level of 1970 (see for example Rising Food Prices). And that capacity would not stay out of the game; rising prices would be exactly the signal required to induce investors (and the 73% of sustainable agribusinesses) to buy the acreage and put it back in production. Life goes on.

    There is no reason to expect the debt crisis to cause more than a small and transient bobble in the downward trend of food prices – did you think there had never been one before? If you’d wanted to make up a really scary story, you should have invoked the rather more serious problem of rising energy prices putting a hard floor under cost per unit. That pain could turn out to be real (though short term; oil is going to cease to be a bottleneck within a decade, I think).

  13. Shenpen, there are at least two possible steps after information mastery. One is to explore the augmentation of the human itself. Human body augmentation isn’t even science-fiction, human mind augmentation is but still seems likely even if we never quite get to the “living entirely in the computer” extreme.

    The other is using your easy manufacturing and easy energy (fission for certain, here’s hoping for some sort of fusion) to start spreading out into space. This harnesses both the extensive manufacturing capabilities we’ll have to build self-sustaining colonies, and the ability of that manufacturing to collect lots more energy than we have now.

    You’re basically saying we’re at the end of history and that all we’ll be is middle-class or leisure-class modern-day humans forever, but that’s unlikely in every way. You can (and perhaps should) mock specific predictions about the nature of a putative Singularity, but in general, the prediction that the future is going to be just like today is the worst possible one. One way or another, we are in a transitional period of some sort. I don’t know where it’s going, perhaps even to extinction (though I personally doubt that), but one thing for sure is that it’s not going to stay here.

  14. I noticed on the graph that taxes make up a large proportion of the upper quintile’s gap between consumption and income. The majority of taxes go to fund transfer payments. If what Dave says is true, that transfer payments are not counted in income but are counted in consumption, doesn’t this suggest that redistributionism is working?

  15. My Linux setup, if measured in the cost of “equivalent” proprietary software, would cost well into the thousands. Perhaps ephemeralization is making life too cheap to meter.

  16. Even if the debt crisis wiped out 27% of our farming capacity in one year, what of it? A 30% increase in food prices would still not push them above the level of 1970 (see for example Rising Food Prices).

    It doesn’t affect the main thrust of your argument, but of course a 30% decrease in output does not necessarily result in a 30% increase in prices. If we were all on the edge of starving a 30% output decrease would cause prices to become astronomical – clearly that isn’t the case now, but I suspect elastistity in the food market is such that it would be a more than 30% increase short-term (nobody wants to change their eating habits, and we can afford to pay more), and less longer-term (say, shifting from beef to pork to increase efficiency of feed grain use).

    This is a constant irritant to me in discussions of the oil market; a 5% change in production might well result in a quite dramatic short-term price change (but very little long-term).

  17. >but of course a 30% decrease in output does not necessarily result in a 30% increase in prices.

    You’re right. I was aware of that point, but chose to rhetorically ignore it because it’s so easy to show that even the maximum 30% long-term increase would not be disaster territory.

  18. Is any consideration given in this study to the plausible hypothesis that higher quality items (usually more expensive) tend to last longer than cheaper items. Therefore the annual cost of buying quality goods works out to not that much more than purchasing cheaper goods.

    To quote from Terry Pratchett:

    “The reason the rich were so rich, Vimes reasoned, was because they managed to spend less money.
    Take boots, for example. He earned thirty-eight dollars a month plus allowances. A really good pair of leather boots cost fifty dollars. But an affordable pair of boots, which were sort of OK for a season or two and then leaked like hell when the cardboard gave out, cost about ten dollars. Those were the kind of boots Vimes always bought, and wore until the soles were so thin that he could tell where he was in the city on a foggy night by the feel of the cobbles.
    But the thing was that good boots lasted for years and years. A man who could afford fifty dollars had a pair of boots that’d still be keeping his feet dry in ten years’ time, while a poor man who could only afford cheap boots would have spent a hundred dollars on boots in the same time and would still have wet feet.”

  19. >Is any consideration given in this study to the plausible hypothesis that higher quality items (usually more expensive) tend to last longer than cheaper items. Therefore the annual cost of buying quality goods works out to not that much more than purchasing cheaper goods.

    That would tend to flatten consumption. The trouble is that it only works for durable goods in a certain range of lifespans. Clothes, yes, certainly; in a rather Vimes-like pattern, I myself buy expensive Rockports rather than cheap off-brand sneakers partly because they last twice as long. And Pratchett is right, this is an upper-middle-class and upper-class habit that I absorbed with my mother’s milk (which is why I chuckled to myself when I first read that passage in situ).

    But it doesn’t work for food, because expensive food doesn’t give you more meals than off-brand versions of the same stuff. Nor for houses; yes, an expensive house may last longer but you don’t generally collect a cost benefit from that, while the serious maintenance stuff like needing to replace your roof shingles or gutters or central air unit will happen at about the same frequency as on a cheaper one.

  20. Eric,

    I think you and all those who commented above missed a fundamental point. A 2:1 ratio is not sustainable and does not make any sense. After all, why should we work so hard if we can spend half the money the rich are spending while being poor?

    In the last 15 years, a very large fraction of the consumption binge of the bottom two quintiles was made possible by the reckless use of borrowed money from reckless lenders. Borrowing against inflated home equity, mortgages with no down payment and teaser rates, multiple loaded credit cards, auto loans, negative saving rate, … The party is over! The hangover will be long lasting, and the spending gap will have to move closer to the income gap in the future.

    The ultimate irony here is that the authors of the article mentioned in your post are respectively senior vice president and chief economist, and senior economics writer at the Federal Reserve Bank of Dallas. The fact that they misinterpreted the data so badly is absolutely staggering, and so is the fact that they still work for the same employer. No wonder this country is on the verge of bankruptcy.

  21. >The party is over! The hangover will be long lasting, and the spending gap will have to move closer to the income gap in the future.

    You’re certainly right, the collapse of easy credit will tend to widen the gap. There are still forces pushing in the other direction, though; markets are still seeking production efficiency, designs are still ephemeralizing, production scale is still increasing. These trends have been going on for much longer than 15 years and show no signs of abating. Many of the gains are locked in and untouchable by a credit collapse; a polo shirt from Nordstrom’s is still pretty hard to tell from the equivalent thing at WalMart.

    So, which force will dominate? My best guess is the gap will widen for a while while we cope with the mess we’ve made, then resume the long-term narrowing tendency. The big question is: how much will it widen by?

    The other interesting question is how much it has to widen before economic envy actually becomes a force in politics again. I don’t know the answer.

  22. In the long run it’s the wealth gap that matters, not the income gap or the consumption gap, because it’s wealth that buys power and freedom.

  23. http://www.energybulletin.net/node/46564

    American Mainstream Perspective

    The vast majority of our population, mainstream America, maintains the belief that we are on the road to the promise land—perpetual economic growth and prosperity enabled by unlimited natural resources. Moreover, they believe that our American way of life is a birthright, our destiny; “the American way of life is not negotiable”[4].

    Concerned Citizen Perspective

    A small but growing minority of concerned citizens, the informed few, understands that we are actually on the highway to hell—the road to societal collapse. They implore us to slow down—to “conserve natural resources”, to “reduce our impact on the environment”, to “balance our budgets”… But we dare not do anything “too drastic”; it wouldn’t be “socially acceptable”.

    Reality

    The reality is that we are running flat out on the highway to hell, and that societal collapse is imminent—possibly within 5 years, probably within 15 years, and almost certainly within 25 years. Our only rational course of action is to “get off” the highway—to transition quickly and beginning immediately to a sustainable lifestyle paradigm. The consequences associated with “getting off” will be very painful—significant reductions in our population level and material living standards—but they pale in comparison to the consequences associated with “staying on”.

  24. yeah, that “chicken in every pot” is a WONDERFUL perspective-creator

    even over and above the very valid observations made in the referenced article, there’s a rather interesting graph in The Economist [2008.10.25, p.40] which lays out 9 countries’ Income Deciles’ averages. well worth hitting the library for. a couple of minor implications arise, but the MAJOR one is that you should knock off the top 2 deciles and the bottom decile, if you want to consider a country’s norm unskewed by dollar-unit extremes. the remaining deciles are spectacularly evenly and consistently spread, common to all 9 countries, and while the US shows greater disparity in income than the very solidly (un)spread EU countries, it spreads UP rather than DOWN — ie, in absolute terms, even the 2nd & 3rd bottom deciles are on a par with the EU equivalents, but 4th and higher are noticeably better off. to put that another way: the US has the same evenness in interdecile spreads internally (and hence impliedly fairness) as do the wealthier EU countries, but given that its non-pathological starting point is equal to its EU equivalents, this simply implies its absolute scale (wealth) is larger.

    more apocryphally, it is noticeable that americans on the mocked bottom rungs of income have profoundly better quality of life than their european equivalents. they eat far more “expensively”, they have access to public services that are the preserve of the EUK rich, and their scummy little houses&trailers are several orders of magnitude better in terms of quality and size than the EUK median house/flat. most brits DREAM of owning properties that average americans would regard as an admission they’d failed in life.

Leave a comment

Your email address will not be published. Required fields are marked *