CEPR - Center for Economic and Policy Research


En Español

Em Português

Other Languages

Home Publications Blogs Beat the Press The Problem Is Not Debt: Consumption Is High Not Low

The Problem Is Not Debt: Consumption Is High Not Low

Thursday, 24 July 2014 04:27

Economists and economic reporters continually try to make the problem of the weak economy and prolonged downturn appear more complicated than it is. After all, if it is very simple then these people would look foolish for not having seen it coming and figuring out a way around this catastrophe. Fortunately for us, if unfortunate for them, it is simple.

One of the efforts to make it more complex than necessary is to assign an outsized role to the debt associated with the collapse of house prices. This is the argument that we heard on Morning Edition this morning. The argument is that when house prices plunged after the housing bubble burst in 2007, homeowners were left with large amounts of debt, pushing many of them underwater. This debt supposed discouraged them from spending, leading to a sharp falloff in consumption.

There is a big problem with this story. Consumption is not low, it is actually still quite high. The graph below shows consumption as a share of GDP. It is actually higher than during the bubble years and essentially at an all-time peak. That makes it a bit hard to explain the downturn by weak consumption. (Some folks may recall hand wringing about inadequate savings for retirement, as in this NYT column by Gene Sperling yesterday. Too little savings and too little consumption are 180 degree opposite problems, sort of like being too heavy and too thin.)


There would be a modest decline in consumption from the peak bubble years if it was shown as a share of disposable income (tax collections are lower today than in 2004-2007), but it would stiill be unusually high by this measure. The basic story is straightforward. The run-up in house prices created by the bubble created $8 trillion in housing bubble wealth. Standard estimates of the housing wealth effect suggest that this would increase annual consumption by 5-7 percent of this amount, or $400 billion to $560 billion a year. This would have been equal to 3-4 percent of GDP.

When the bubble burst this wealth effect went into reverse, with people cutting back consumption in line with their loss of wealth. The consumption share of GDP did not fall both because GDP fell and also there was a sharp drop in tax collection due to both tax cuts and simply the drop in income. The fact that people had debt may have made some difference, but it was really secondary to the loss of wealth. If a homeowner owed $100,000 on a home whose price dropped from $300,000 to $200,000 (leaving them with $100,000 in equity), we would expect them to cut back annual consumption on average by between $5,000 and $7,000.

If the homeowner owed $250,000 on this house, so the drop in price left them $50,000 underwater, then their decline in annual consumption may be somewhat greater, but this difference would have a relatively modest impact on the economy as a whole. To see why this almost certainly has to be the case, consider that the median income for homeowners is around $70,000. How much do we think their consumption could have fallen from bubble peaks? If we say they fell by an additional $3,500 because of being indebted (beyond the housing wealth effect) and multiply by 10 million underwater homeowners, that gets us $35 billion a year. If we plug in a multiplier of 1.5 that gets us to $52.5 billion a year or a bit over 0.3 percentage points of GDP. That won't explain much of the downturn.

Again, the story of the downturn is simple, too bad the economists missed it.


Note: Typos corrected.



I have a few quick points after reading the comments. First, the wealth effect is based on equity net of mortgage debt. This is people's wealth. If the argument is that people's debt rose relative to their equity when the bubble burst, then this is simply a wealth effect story. For there to be a point to the argument, there has to be some importance to the fact that people actually had negative equity. And to see the significance of the numbers I used, suppose the underwater homeowners' consumption had been on average $3,500 higher each year since 2008. We are now in the 7th year, so in this case, that would have translated into a cumulative increase in spending of close to $24,000. If we assume that these people have been spending close to all of their income, this means a cumulative increase in debt of close to $24,000. Even if these people were marginally above water, as opposed to underwater, in their mortgages, who would have lent them an additional $24k? I just don't see this one as making any sense.

There were some questions raised about whether we "need" more consumption. We need more demand to get to full employment (unless we redivide work), which could come in part or entirely through consumption. But the question is not whether more consumption, or as I wrote the other day more investment, would be good for the economy. The question is whether to believe that the economy would generate more consumption or investment. The answer in this case and the previous case with investment is no.

The basic point is that we have an enormous demand gap created by the trade deficit. The current deficit amounts to $500 billion in annual demand (3 percent of GDP) that is going elsewhere rather than creating growth and jobs in the United States. There is no magical process through which the economy will replace this demand. We can do it with large amounts of government spending, but this is blocked politically. That leaves getting the trade deficit down, most obviously by lowering the value of the dollar, as the only route back to full employment or something like it.




Comments (26)Add Comment
written by Last Mover, July 24, 2014 6:25

This is pretty much a Dean Baker original. Not even Paul Krugman pushes this one, highlighting the drag of debt overhang instead.

It puts Keynesianism itself front and center as the explanation and solution. There really is a giant hole in aggregate demand which will not close anytime soon on its own from consumption, the main source of demand.

The economy is stuck in stagnation equilibrium far below its potential despite a slow crawl upwards. Debt is not holding it back. Only fiscal spending or a reduction in the trade deficit can push it to potential.

The sock puppets can't get to this explanation because the power of debt fearmongering has such a death grip on them they have to believe the economy would bounce back but for all that debt.

Otherwise they would have to explain the economy under Keynesianism as the answer and that is totally off limits in their indoctrinated austerian mindset.
The Last Piece
written by Larry Signor, July 24, 2014 6:42
...of the puzzle. I agree with Last Mover. Perhaps, one day, we will call it the "Baker Rule", i.e. the Bubble Circle. If we accept Deans explanation (which I do) of the recent depression, the reverse wealth affect should have been anticipated. [Sometimes I wonder why I waste my time reading anyone besides Bernstein and Baker.]
Banks' shattered balance sheets contributed.
written by Ralph Musgrave, July 24, 2014 6:48

The discovery by banks that the value of their assets (i.e. mortgages) was less than they thought also contributed, didn't it? It induced them to lend less I suspect.
Our Secular Stagnation is Politically Induced
written by Robert Salzberg, July 24, 2014 8:05
Our national infrastructure deficit is over $2 trillion or around 13% of GDP. If you factor in a standard multiplier, fixing our infrastructure would increase GDP by 2% for nearly a decade before we got caught up. An increase of 2% of GDP over baseline would mostly bring us to full-employment and give wages a chance to have a moderate rise, which would bump up inflation a bit, which would close our trade deficit a bit. The virtuous circle we have all been hoping for is just as plain and simple as Dr. Baker's noting that we're spending more not less than average.

Real Secular Stagnation would mean our economy isn't producing enough work. We've got plenty of work, our real deficit is in policy sanity.
ye olde denominator problem
written by rob urie, July 24, 2014 8:18
One problem with interpreting ratios is understanding what is driving the relationship. Look at PCE without GDP to see what I mean. http://research.stlouisfed.org/fred2/series/PCE The point about scaling values to give them context is a good one---except when doing so clutters the context. For more look at Ken Rogoff's misunderstanding of what drove the debt / gdp ratio when he was selling austerity.
Good point, but...
written by Dave, July 24, 2014 8:38
I'm not sure the semantics of this are agreed upon. 'Debt burden' can very easily be taken to mean the opposite of the 'wealth effect' if we look at all homeowners with a mortgage rather than only those with underwater mortgages.

I don't see 'debt burden' as referring to only those underwater. I would guess that the 'wealth effect' spending is overestimated for those who had no mortgages, and underestimated for those with mortgages.

It would be informative to see a breakdown of consumption during the bubble years of those with mortgaged homes vs. those with homes that were free and clear. I don't think most people with homes that were free and clear were participating in the wealth-effect spending nearly as much as those with mortgages.

If the primary effect on consumption was from those with mortgages, then it is nearly equivalent to call the loss of wealth-effect spending a loss of spending due to debt overhang. Perhaps debt overhang implies underwater to some people, but I think that is incorrect.
A lot of consumption was monetized with fiat currency
written by Dave, July 24, 2014 8:48
I know this will rub some people the wrong way, but when we talk about debt overhang, we have to look at what really happened in the bubble years. Homes became overvalued, and mortgages increased, monetizing the bubble values using fiat currency. When the bubble popped, those loans using fiat currency, that is, currency created based upon loans on the collateral of bubble wealth, created a lot of outstanding debt to the central bank that had no collateral.

To see why this is true, it is important to understand the mechanism of much of the mortgage debt accumulated during the bubble years. Because of the end of Glass Steagall, banks were free to use their deposits, including money market funds, to create fiat loans for buying mortgages. They shuffled it around to fool each other so that everyone looked solvent. They were solvent if the bubble wealth remained, but they took huge risks based upon the ability to use deposits as a basis for creating more currency to buy mortgages.

I'm not sure why this was never analyzed properly in public. It is almost as of nobody even realizes the degree to which the elimination of Glass Steagall contributed to this. They assume it was all fraud, but it wasn't. Perhaps we can make the argument that the elimination of the GS act was a fraudulent political maneuver, but that is another issue.
How are wealth-effect spending calculations made?
written by Dave, July 24, 2014 8:59
This all makes me wonder where the history behind the wealth-effect spending calculations originates. Could it be that our calculations have always ignored the fact that most wealth-effect spending is carried out by a minority of the people? That the aggregate calculation based upon 8 trillion uses very low percentages over the entire population instead of much higher percentages over a lower percentage of the population?

The question to me is whether our theory on this was thorough enough to be applied directly housing bubble wealth via mortgages. Was the theory based upon this situation, or was it based upon other situations? If the same, did they understand the breakout of consumption or just aggregate it all?
We're talking private debt overhang, not public
written by Dave, July 24, 2014 9:08
Of course, it depends upon which audience we're addressing whether it matters what we call this. As first commenter suggests, some people will read debt overhang as meaning the US has too much debt, so we need austerity. Of course, this interpretation is ignorant, because this wealth effect or debt overhang issue is all about private debt rather than public debt. If existing financial capital isn't used for direct investment, it has to be lent to others as indirect investment or consumption or else we have a recession. If we have excess financial capital, then we have to convert it to debt. It makes no sense to say the overall debt levels in the US are too high. It just doesn't have any meaning.

So either we create public debt with that excess financing or we will have a continued high unemployment rate. This really isn't open for opinion. These are facts.
written by skeptonomist, July 24, 2014 9:09
Dean keeps saying that things are "normal" with reference to GDP - yesterday it was investment and today it is consumption. But since GDP is lower than where we would like it to be, this means that investment and consumption are also lower. If you want to see where these things should be - hopefully on the long-term trend that existed in the past - then you have to look at (per-capital) values not divided by GDP. When we do this for investment:


it is clear that residential, non-residential and total investment are still well below the long-term trend. And bringing residential investment back to its trend would not nearly bring total investment back up to its trend. Consumption (demand) shows a similar thing as pointed out by rob urie - this is why it is often said that demand has been lacking (Dean himself ranks demand as the top three problems).
Consumer spending is 70% of the economy
written by Bill H, July 24, 2014 9:21
You can even compress the horizontal scale of your graph to make it look more dramatic, but in the last ten years it's gone from 67% to 69% of the GDP.

The economist's love affair os relating everything to its percentage of the GDP turns everything into gibberish. It does manage to pour oil on troubled waters very often, as Dean does here, because the primary purpose of economists is to assure the public that everything is fine so that they will spend money and reelect the same politicians.

But that makes one wonder why any stimulus spending is needed, or why any of the safety net spending claims of "it will increase consumer demand" have any validity, because according to Dean we don't need to increase consumer demand. It would seem that those claims should be ignored and the bills calling for those programs should be rejected unless better arguments in their favor can be found.
written by skeptonomist, July 24, 2014 9:35
Dean keeps claiming that the housing wealth effect is important - supposedly people spend much more when they have positive equity - but now when NPR appeals to a kind of reverse wealth effect he insists they are wrong. Is housing wealth important or not? (I think not so much as many other influences)
written by Dave, July 24, 2014 9:37
We're getting somewhere with this discussion. Dean has made very important points here that nobody else has made. You are also making good points.

It is a shame that current blogging paradigms usually result in the end of a discussion just as it is getting started...
Bubble Trouble
written by Larry Signor, July 24, 2014 9:56
This FRED graph helps explain...something.

We are indeed getting somewhere
written by Bill H, July 24, 2014 10:05
Especially when Dean adds his point, which he has made before, that the trade deficit is wiping out the usefulness of consumer demand. This is a point which all to few economists make. As consumer demand (a plus to GDP) rises so does imports, which are negative to GDP.
We're not connecting on wealth effect
written by Dave, July 24, 2014 12:00
I'll put together a much better explanation of what I'm talking about and post it on my own blog in a couple of days. I'm bad at data gathering, partially because I don't have good tools, and partially because I don't have much practice. I'll give it a shot.

I understand the point about the meaning of equity over debt, and how that translates to wealth effect spending. The point I'm trying to make is that a good portion of that $8 trillion was owned by people without mortgages at all, or by people who did not increase their debt levels despite the new wealth. In these cases, the claim of wealth effect spending is purely a psychological phenomenon, and I think much weaker than the spending effect by those who had large mortgages and perhaps even increased their mortgage levels based upon bubble wealth. My contention is that the actual wealth effect spending was mostly debt fueled. I've seen data to this effect.

This isn't an airtight case, of course. But the basis of my thesis is that we can infer a much better connection to spending in those who increased their debt under the bubble, and even for those who were highly leveraged to begin with, but who's leverage percentage decreased as the bubble grew, probably fueling psychological spending, only to be wiped out when the bubble burst.

I'll get this into numbers and graphs. I think it is a vital point, and a bridge between thew viewpoints of Dean and PK, which I think are almost identical if we get the semantics straightened out.

I'll post a link when I get there.

@Larry and @Bill
written by Dave, July 24, 2014 4:57
Yes. The mortgage debt looks like a big problem, doesn't it? I calculated this once before and found about 3 trillion in debt on phantom wealth. In other words, there was about 3 trillion owed to the Fed on nonexistent collateral when we hit bottom.

That is the wealth effect/debt overhang problem I'm talking about. 3 trillion in spending over a period of about 8-10 years or about 2-3% of GDP in wealth-effect spending that was due to new debt on phantom wealth. This is how I calculate the wealth-effect spending. There may have been some more by those who didn't increase their debt, but the overhang, this 3 trillion is the overlap between what PK speaks of and Dean speaks of.

Yes, the trade deficit is probably the underlying cause that required the bubble to fill demand. Nevertheless, we also have the debt overhang issue to deal with now.

written by skeptonomist, July 24, 2014 5:28
Most people don't regulate their spending - or their credit-card balance - by consulting Zillow or a real-estate agent as to the value of their house, but a housing wealth effect makes sense to the extent that there is spending based on reverse mortgages. Does the variation in reverse mortgages account for the variation in debt or consumption? The same question could be asked of NPR if they claim a reverse wealth effect.
written by Dave, July 24, 2014 5:39
The wealth-effect, debt-fueled spending we're talking about here was caused mostly by 1) cash out refinancing and 2) turnover of housing using ever-increasing mortgages.

When people who generally live paycheck to paycheck find themselves with cash on hand, it tends to get spent. They didn't perceive it as a drop in net worth because it had collateral to offset it. Banks didn't see it as a drop in net worth, and even Greenspan didn't see it that way. In fact, Greenspan explicitly encouraged people to look at it this way.

One last thought
written by Dave, July 24, 2014 9:01
I think Dean is trying to emphasize that the trade deficit needs acknowledgement from experts in trade. I agree with that. I also can see that consumption looks pretty 'normal' at this point, but that is subject. I'd guess a good portion of the debt overhang was eliminated by foreclosures, but that at least half of that 3 trillion is still weighing in to some degree. Perhaps 1% of GDP or so. The rest is probably still the remaining trade deficit as Dean is saying.

I wasn't trying to undermine the point on trade, but just point out that debt overhang and trade deficit are likely both contributing to low demand right now.
consumer liabilities
written by Squeezed Turnip, July 24, 2014 10:34
The numbers indicate the aggregate dropped a little then stabilized, recently rising a little, but far below any historical rates of increase:

Bottom line: Dean makes a good point.
@ Squeezed Turnip
written by Larry Signor, July 25, 2014 1:58
YOY change does not tell us about the aggregate effect of household liabilities. This would be the important facet of any debt overhang, not the growth of debt.
consumption as percentage of household equity
written by Squeezed Turnip, July 25, 2014 5:20
personal consumption appears to be supported by credit liabilities not real estate equities. Did the spike in liabilities hobble the economy?
General abstractions
written by Jurriaan Bendien, July 25, 2014 5:54
Dean Baker is a level-headed analyst of the economic malaise. However, a slogan such as "consumption is high not low" is deceptive. Aggregate consumption expenditure does not distinguish between different classes of consumers, and does not identify the real rate at which their consumption expenditure actually grows. Incurring debt as such is not a bad thing in itself, I'm sure, but at a certain level of debt, repayment requirements set a limit (1) on the amount of disposable income that can be spent on purchasing more goods and services, (2) the ability to incur additional debt. The macroeconomic significance of that is, that if additional spending power is somehow distributed to consumers, as a stimulus measure, this does not have the same "multiplier effect" that it had in Keynes's day.
Consumption is low
written by Juan Deshawn Arafat, July 25, 2014 1:30
Consumption is less cyclical than other parts of GDP, so the fact that C is not low as a share of GDP is no surprise. But I think you are using the wrong denominator. As a share of potential GDP, consumption is more than a full point below where it was in 2007. Basically low C has dragged down Y and I.
Confusion on the housing wealth effect
written by Dean, July 26, 2014 10:27

It doesn't matter that "most people" don't adjust their spending based on the Zillow estimate of the house price. All that matters is that some people do. There are people, and it was widely written about at the time, who take advantage of the run-up in their house price to borrow against the new wealth. They might take a vacation, buy a new car, remodel their house or who knows what. There clearly were people doing this in 2004-07, it was a major industry and there was much writing on the topic.

Since some people spent more due to their housing wealth, and no one spent less, then we have a positive housing wealth effect, end of story. Like I always, economics is simple -- it's just economists who want to make ti difficult so they can pretend they actually know something.

Write comment

(Only one link allowed per comment)

This content has been locked. You can no longer post any comments.


Support this blog, donate
Combined Federal Campaign #79613

About Beat the Press

Dean Baker is co-director of the Center for Economic and Policy Research in Washington, D.C. He is the author of several books, his latest being The End of Loser Liberalism: Making Markets Progressive. Read more about Dean.