CEPR - Center for Economic and Policy Research


En Español

Em Português

Other Languages

Home Publications Blogs Beat the Press Underwater Homeowners Cannot Explain the Weak Recovery

Underwater Homeowners Cannot Explain the Weak Recovery

Friday, 23 November 2012 08:05

Anyone wanting to learn about the economy who talked to the nation's top economists in 2006 would have been wasting their time. Almost none of them had any clue that the collapse of the $8 trillion housing bubble was going to wreck the economy. This presumably reflects a rigid dogmatism and conformity on the part of these economists, since it should have been both very easy to recognize an unprecedented run-up in house prices as a bubble and also to understand that the collapse of the bubble, which was quite evidently driving growth, would lead to a severe downturn.

Remarkably, it seems from a Washington Post article that attributes the continuing weakness of the economy to the indebtedness of underwater homeowners, that many of the country's top economists have no better understanding of the economy today than in 2006.The claim is the drop off in consumption due to the debt burden of these homeowners explains the weakness of the recovery.

Some simple arithmetic shows the absurdity of this view. The amount of underwater equity is estimated at between $700 billion (Core Logic) and $1.1 trillion (Zillow). Suppose that we can disappear this debt through some decree, how much additional consumption would we see? If we assume that these households spend an incredibly large share of this increase in their net wealth, say 15 cents on the dollar, this would imply additional consumption of between $105 billion (Core Logic estimate) and $165 billion a year (Zillow estimate).

However we would have also destroyed the wealth of the mortgage holders. Let's assume that they just spend 2 cents on the dollar of their wealth. This would imply a net boost to demand of $91 billion to $143 billion. While this would be a helpful boost to the economy, equivalent to a government stimulus program of this size, this would hardly be sufficient to make up a shortfall in annual output that the Congressional Budget Office puts at close to $1 trillion.

Furthermore, even this gain is almost certainly a huge exaggeration of the actual effect. With 11 million homeowners underwater, the above calculation implies an increase in average annual consumption of between $9,500 and $15,000 a year. The median homeowner has an income of less than $70,000 a year. It doesn't seem likely that such a family would either have this amount of savings each year that they could instead decide to consume if they were no longer underwater in their mortgage or that they could borrow this amount on any sort of sustained basis. In short, the numbers in my calculation above almost certainly hugely overstate the economic impact of eliminating underwater mortgage debt.

In fact, there is no need to turn to implausible underwater mortgage debt explanations for the weakness of the economy. The economy is acting exactly as those who warned of the bubble predicted. We saw a sharp falloff of residential construction as we went from a near record boom, with construction exceeding more than 6.0 percent of GDP at the 2005 peak, to a bust where it fell below 2.0 percent of GDP. This meant a loss in annual demand of more than $600 billion a year.

We also saw a large falloff in consumption due to the loss of $8 trillion in housing wealth. The housing wealth effect is one of the oldest and most widely accepted concepts in economics. It is generally estimated people spend between 5 and 7 cents each year per dollar of housing wealth. This means that the collapse of the bubble would be expected to cost the economy between $400 billion and $560 billion in annual demand. 

There is no mechanism that would allow the economy to easily replace the combined loss of between $1 trillion and 1.2 trillion in demand that would be predicted from the collapse of the housing bubble. Therefore it is hard to see why anyone would feel the need to look to explanations involving the indebtedness of underwater homeowners, the whole downturn is easily and simply explained by the collapse of the bubble.

In this respect it is worth noting that, contrary to the impression given by the article, consumption remains unusually high relative to disposable income, not low.


                             Source: Bureau of Economic Analysis and author's calculations.

As can be seen consumption as a share of disposable income is well above the level of the 60s, 70s, 80s, and even the 90s prior to the point where the stock bubble led to a consumption boom in the late 90s. If anything, we should be asking why consumption is so high, not why it is low. (Adjusted disposable income refers to the statistical discrepancy in the national income accounts.) In short, the underwater homeowner story is an explanation for a mystery that does not exist.

There are a number of other points in this piece that incorrect or misleading. For example, it refers to the Obama administration's failure to address the situation of underwater homeowners as:

"a persistent and largely unaddressed problem that represents the missing link in what many economists consider the administration’s overall strong response to the recession."

Actually, many economists did not view the administration's response to the recession as strong, pointing out at the onset that the proposed stimulus was woefully inadequate. These economists were not surprised by the subsequent weakness in the economy.

At another point it explains to readers:

"Some people reduced spending because they had lost their homes to foreclosure, damaging their ability to borrow. Others no longer could tap home-equity lines of credit. Still others, facing high monthly payments, used every extra penny to pay off debt."

This comment is extremely confused. If someone loses their home to foreclosure, then they are no longer in debt, except in the extremely rare case where a lender pursues a deficiency judgment against the homeowner. Lacking the ability to borrow because they are not homeowners or losing access to a line of credit because they have no equity are not problems of indebtedness, they are problems of lacking equity in their homes. This is the standard housing wealth effect story. No one has proposed that the government should not only eliminate the negative equity of underwater homeowners, but also give them substantial positive equity so that they can again borrow against their home. Therefore this argument has nothing to do with the underwater homeowner story that is the central theme of this piece.

The economics profession did an astounding amount of damage to the country as a result of its complete failure to see the housing bubble and the dangers it posed to the economy. Economic reporters also failed the country by not being able to exercise any independent thought to understand that the bulk of the profession was missing something important. (There were prominent economists like Robert Shiller at Yale and Paul Krugman who did warn of the bubble.) It is unfortunate that economics reporters still write pieces that rely exclusively on economists who could not see an $8 trillion housing bubble. 

By the way, we certainly should be trying to help underwater homeowners as a simple matter of fairness. We bailed out Wall Street billionaires, it seems a pretty minimal proposition to offer assistance to homeowners who bought into a bubble that all the top economists insisted did not exist.

Comments (12)Add Comment
written by JSeydl, November 23, 2012 9:28
The problem is that economists aren't interested in predicting anything useful about the real world. They're too busy speculating about how beliefs and desires inform preferences and choices, and then using those speculations to build representative-agent models based on efficient markets. Never mind if those models tell us nothing about the real world -- at least they incorporate fancy equations!

At its core, the problem is at the universities. Professors are not interested in, for example, teaching students how to pull housing data from sources like CoreLogic. I thought the graduate economics curriculum might change after 2008, but it turns out that that was wishful thinking.

written by Robert Waldmann, November 23, 2012 10:02
In the comment I define your disagreement with post reporter Zachary A. Goldfarb as a disagreement about whether reduced but still positive home equity causes a reduction of consumption on the order of between 5 and 7 cents each year per dollar of housing wealth or on the order of zero.

That is, I won't address you valid points that the recovery of consumption was roughly normal while the recovery of house construction was not.

Notably Goldfarb cites empirical estimates based on micro data. He claims there is a significant change in the slope of the home equity effect on consumption at home equity equals zero. To say he is wrong, you must convince me that the 5 to 7 % estimate is valid for a sample containing only homeowners with positive equity.

An estimate with aggregate data just does not address the claim in the article which you criticize.

Notably, in this case the Washington Post cites specific research by named economists. You, in contrast, cite what all macroeconomists know.

I am shocked to find that I call this one for The Washington Post. Your conclusion may be correct, but your reasoning is based on the assumption that all functions are linear. I don't like to be square, but that's not true.
written by jerry, November 23, 2012 11:00
Aren't a good majority of these mortgages owned by the Fed or some other government organization at this point? So why do we really care about their losses? Have the government reimburse any writedowns that are due to the private sector if we really feel like being overly generous to the vampire squids again.

I'm not sure how you can argue that the indebtedness of people with underwater mortgages is not a big issue or contributing factor to the weakness of the economy right now. People have to throw away a third of their income on artificially inflated debt - this makes them less competitive, with less money to spend into the economy, less money for retirement, etc.
The Housing Wealth Effect Estimate Are Based Almost Exclusively on Homes with Positive Equity
written by Dean, November 23, 2012 11:30

I gave a link to the google scholar pages for "housing wealth effect." There are many dozen articles that give estimates of the size of the housing wealth effect. Almost all of these estimates would be based almost entirely on people with positive equity in their homes because in times prior to the collapse of the bubble, almost no one had negative equity in their homes. Pick an article, there's a range of estimates, but I don't think you'll find any where the estimates for the wealth effect are zero or anything close to it.


please look at the numbers for negative equity and use whatever assumptions you want about its impact on spending. I'm using arithmetic and logic, if you want to bankers bad names, it's fine by me, but I think it helps explain the state of the economy.
It's not about the balance sheet
written by ts, November 23, 2012 11:42
It's about the income statement. Underwater homeowners have difficulty refinancing, and therefore are often still paying mortgage rates from the mid-2000s. Include taxes and insurance, and quite often homeowners are saddled with monthly payments of 40-50% or more of their disposable income (not to mention the plethora of interest only/option ARMs that have reset). This has a huge effect on how much discretionary spending a household has. Most households in the U.S. do not have that much discretionary spending to begin with - 20% of income or more is a big number (see any Consumer Expenditure Report by BLS)

The second issue is that after every recession, economic agents need to repair the financial damage done. 401k's and other assets had to be cashed in, credit card balances increased, people borrowed from family members. Often, a job loser who finds employment is working for 70-80% of their previous salary. If the new employment itself is tenuous, it is unlikely the individual is likely to go on a spending spree.

These are the main reasons why recoveries in general are gradual and not as fast and steep as the recessions, and why this recovery in general has been slower than most.
written by skeptonomist, November 23, 2012 2:58
The Google search that Dean refers to actually seems to have about as many papers which discount the housing-wealth effect as affirm it. The influences on savings and spending rates is a complex and controversial subject and I don't think the dogmatism that Dean applies to it - attributing the level of savings to wealth effects only - is justified.

Stock-market wealth was not always correlated with savings rate. There was a peak of stock-market wealth in the 60's which had no effect on savings rate:


There was also a huge housing boom with corresponding increase in housing wealth after WW II up to around 1970 and this seems to have had no effect on savings rate, which just increased steadily until around 1980. Presumably the extremely high savings rate during WW II resulted in wealth in savings bonds - where is the effect of this factored in?

Psychologically, wealth effects seem dubious. I doubt if very many homeowners, apart from those few who are actively speculating, keep close track on their equity and adjust their spending accordingly. People who actively speculate in the stock market may adjust spending according to perceived wealth, but such people are a small fraction of the population.

As shown in the graph, revolving or credit-card debt has increased as savings rate has gone down - credit cards hardly existed before the 60's. It is also hard to believe that the decrease in real wages in the lower part of the income spectrum since about 1970 has not affected savings rates for most consumers. Have these and other influences really been ruled out conclusively, or is the emphasis on wealth effects just faddism and oversimplification? Probably many things are correlated with overall societal optimism, but it is difficult to put a number on this. Dean is right to call attention to asset bubbles, but they don't explain everything.
What about the income effect?
written by Blissex, November 23, 2012 6:30
The housing equity effect on consumption is the usual conservative propaganda to boost the political strength of the rentiers.
Just ask yourself: what is the effect on spending of earned INCOME? not wealth, of earned INCOME?

Isn't that likely to be massive?

If most consumers feel poor and don't have money to spend, is that because their income from work has been stagnant or declining, or is it because house prices have been stagnant or declining?

What about policies that boost wages instead of house prices to boost demand and employment?

Perhaps they are politically impossible, because 70% of voters are middle aged or retired wannabe rentiers with small properties, and dream of ever higher house prices for themselves, and ever lower wages for everybody else.

They are fools who will pay dearly for their "F*ck YOU! I got mine" attitude, but in the meantime...
Housing data
written by Lrellok, November 23, 2012 6:31
Prof. Baker
While i agree with your core argument, i am afraid that i must disagree with some of your statistics. The MEAN household income is less then 70k, the MEDIAN is around 50k, with a 3/5 mean of $49,842 in 2011. http://www.census.gov/hhes/www...household/
While i fully concur with the idea that household spending as a portion of income is to high, i feel strongly that this is not a spending problem but a income problem. If HH median income where 70k, households could spend at current levels and save nearly 30%. Where household to save merely 20% ( a four fold increase from today) we would see a groundswell in consumer spending.
I would like to take a moment to inform you (and the economics community at large) that i have found MAJOR discrepancies in the BEA's statistics on income. Both the census and the Social Security peg mean personal income at around $40,000, however if you divide BEA's total wages by their Full + Part time labor force, you end up with a mean of $48602 (I will not even go into OECD calling mean wages at $54000, which is frankly absurd). Could you possibly write an article or post on this? As discussions of inequality and wage rigidity move forward, have accurate stats is going to be critical, it does not help to have means and medians all over the map.
asset bubbles, credit bubble, the LBO of America etc.
written by Blissex, November 24, 2012 7:01
«Dean is right to call attention to asset bubbles, but they don't explain everything.»

Sure, there is also energy prices and demographics, but when there have many so many so enormous asset price bubbles sponsored by so many governments in so many countries they may not explain everything, but they are CENTRAL.

What all these asset price bubbles have in common is a colossal credit bubble, well obvious in this "the most important chart in the world":


That's debt (BOTH private and public) to GDP, with two dramatic trend changes in 1980-1981 and 1994-1995.

If you show that graph to someone in a business school, not a department of "Economics", they immediately understand that's a typical financial profile of a company after an LBO by asset strippers: loaded with debt, to allow for massive equity extraction via dividends (e.g. tax cuts) and insider transactions (e.g. Medicare part D).
shifting losses to suckers has worked!
written by Blissex, November 24, 2012 7:19
«a good majority of these mortgages owned by the Fed or some other government organization at this point? So why do we really care about their losses?»

That was exactly the point, to transfer the losses on mortgages from the rich who own large financial institutions to the taxpayers, who are only the middle class and the poor.

Also because a lot of mortgages are to relatively wealthy speculators, as someone who has assets nominally worth hundreds of thousands of dollars obviously is far richer than most middle class and poor families, and some of the biggest growth in crazy mortgages was to wannabe lords of the manor to buy McMansions.

Conversely, if one looks at mortgages to poor Americans, usually colored, they tend to be tiny (a few dozen thousand dollars), and to be current (not in default), as the (mostly colored) poor both tend to be more prudent with money and are less favoured by lenders than the imprudent "aspirational" upper and middle classes who feel entitled to the lifestyle of the lord of the manor.
Credit Card debt peaked at 10 percent of disposable income
written by Dean, November 24, 2012 8:19

Credit card peaked at around 10 percent of disposable income in the late 90s http://www.federalreserve.gov/...fault.htm. It was stable or falling as the saving rate fell to zero in the last decade. I don't think there was a single year where the rise in credit card debt was even 1 percentage point of disposable income. In other words, you can't explain much of the change in saving rates with credit card debt, it just isn't large enough.

As far as the wealth effect in the post-war years, housing wealth was actually depressed in 1960s so it is not surprising that we didn't see a big consumption boom kicked off by the relatively short-lived stock boom of the late 60s. Also, savings did rise in the 70s, when the market tumbled, and then began to fall in the late 80s when the market started to exceed historic PE ratios. There are other factors, but the basic numbers fit the wealth effect story quite well.
Good criticism
written by bakho, November 24, 2012 9:33
Good criticism.
However, your post has too many interconnected ideas, and I don't think the message comes through. I see the takehome as follows:

During the bubble, both increased building and increases in leveraging were fueling demand.
After the crash, neither was contributing to demand.
The demand (in the short term) needs to be replaced and BigG is a good candidate to do so through stimulus.

This message gets muddled with fixing the housing market (a separate issue).
Re-inflating the bubble and return to leveraging houses at bubble prices is a fool's errand. However, allowing the housing market to clear would be helpful for a variety of reasons.

In terms of demand, Distribution of wealth and income matters and it matters a lot.
Your statement, "However we would have also destroyed the wealth of the mortgage holders. Let's assume that they just spend 2 cents on the dollar of their wealth" misses part of the equation. Paying down debt = savings. Thus, underwater homeowners are in a "forced-savings" program. Forgiveness would relieve the forced savings which is far greater than the demand from the mortgage holders. If, in fact, Uncle Sam is the mortgage holder (as some propose) then we get an increase in BigG spending today with higher revenues in the future when the economy recovers. The article you criticize makes a muddle (as most popular articles do) of the rationale for clearing the housing market, but there are good reasons to do so that are not enumerated.

Of course, clearing the housing market is independent of the need for more stimulus. But one form of stimulus could both clear the housing market and increase private demand. Certainly other forms of stimulus would be welcome and are a good idea.

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.