Wednesday, February 11, 2009

The Home as ATM

Federal Reserve economists Alan Greenspan and James Kennedy produced two papers [1] [2] in recent years that analyze some trends in US mortgages. Since then, they have adjusted and updated their analysis [3]. Among other things, they looked at equity extraction – the amount of home equity that was converted into liquid funds by US homeowners. Gross equity extraction, or mortgage equity withdrawal (MEW) as it is also termed, is defined as equity extracted due to home sales (first mortgages minus the debt paid off due to the sale) plus cash out due to refinancing of first mortgages plus the change in home equity loans net of any unscheduled payments on first mortgages. Net MEW is gross MEW minus costs associated with the equity extraction such as closing costs. Net mortgage equity withdrawal is the stored value that homeowners removed from their houses and converted into cash in order to make various purchases, provide funds for investments, or to payoff non-mortgage debt.

An interesting thing happens when we compare their net MEW numbers with GDP for the US economy [4]. Because the home equity we are looking at here was liquidated and spent, it showed up in the calculation of GDP (see a caveat to this statement below). If the equity had not been extracted, however, it would not have impacted the GDP. It would still be there in the various homes as equity. Thus, we can compare real GDP growth as it has been reported [5] with real GDP growth minus net MEW. In this way, we can get a good idea of how GDP would have changed if US homeowners had not extracted equity from there homes. We can also look for any changes in the comparison over time.

The following chart presents the comparison of real GDP change and real GDP w/o net MEW.

There are several interesting points to this chart. First, the overall impression is that MEW has been a significant part of the official measure of economic growth for years. Second, we can note that MEW has had an increasing impact on real GDP since roughly 1996. From 1991 to 1995, new MEW added about 1% to the growth of GDP. But in 1996 and 1997, the impact grew to about 1.5%. After 1997, the impact of MEW on GDP grew until, by 2004, net MEW was contributing 6.5% to the change in GDP. Since 2004, net MEW has declined and with it, the GDP. The third point we can note is probably the most obvious. Had it not been for sizable amounts of equity extraction, the US economy, as measured by real GDP, would have contracted every year since the last recession. And with drops of 1.5-3.5% per year, we’re talking about a significant contraction.

There is one caveat that I’d make here. Some of the free cash generated by the equity withdrawals was used to make investments such as stock and bond purchases. While some investments directly show up in GDP (e.g., art or precious metal purchases), the purchase of financial instruments such as stocks and bonds are not counted in GDP. The investments listed by Greenspan and Kennedy are not distinguished by source so there is no way to tell what showed up in GDP and what did not. However even if we assume that none of the investments showed up in GDP, investment as a category explains no more than a moderate portion of MEW. Moreover, the percent of extracted free cash that went into investments consistently declined over the period in question. In addition, the purchase of financial instruments does have some indirect impact on GDP. All this is to say that for the purposes of this analysis, it is assumed that all of net MEW showed up in GDP. This overstates somewhat the affect of MEW on GDP and makes the chart look a little more dramatic than actual events would warrant but I don’t think the error is large. Absent net MEW, we would still be left with a falling GDP throughout this decade.

So to put it bluntly, US homeowners have been using their houses as ATMs. The equity they withdrew fostered the illusion that our economy was healthy and growing. But the spending growth was not fueled by new savings that was invested to increase production. Instead, current assets were leveraged and consumed to fuel the GDP. In some cases, the assets were realized capital gains (such as home sales) but in other cases, the gains were unrealized (such as home equity loans). And now that the housing market has been crushed, this fire hose of funds has been reduced to a trickling garden hose at most. With few exceptions, net MEW was $150B-$200B per quarter from 2005 to Q2 2007. Since then, net MEW has tanked as follows [6]:

Q3 2007: $119.3B
Q4 2007: $92.3B
Q1 2008: $51.2B
Q2 2008: $9.5B
Q3 2008: -$64.1B

And judging by the hit that the US housing market has taken, it will be some time before it's even possible for mortgage equity withdrawal to be a significant player in GDP again. Perhaps now we’ll begin to realize that leveraging and consuming assets is not the path to sustained economic growth.


[1] Estimates of Home Mortgage Originations, Repayments, and Debt On One-to-Four Family Residences, Sept. 2005,

[2] Sources and Uses of Equity Extracted from Homes, March 2007,


[4] The idea for this comparison came from John Mauldin ( There is a consistent difference between his calculation of how MEW impacts GDP and my results. I show a bigger affect of MEW on GDP for each year that our charts have in common. However the respective overall points as well as the diachronic trends displayed by our charts are the same. I have contacted Mauldin’s company to ask about the difference but have not heard back from him.

[5] GDP data retrieved from

[6] The most recent data came from the following sources:



Post a Comment

<< Home