Finally! Empirical Evidence About How the Foreclosure Crisis is Impacting the Recovery

Finally, a paper gets to the bottom of the relationship between foreclosures and the recession.  But first, are foreclosures a good or bad thing?

This is the question the administration has been dealing with since the beginning of the crisis.  As we can see from ProPublica’s recent story by Paul Kiel and Olga Pierce, Dems: Obama Broke Pledge to Force Banks to Help Homeowners, Summers and Geithner were reported to not see any advantage to cramdown and other serious foreclosure relief efforts.  Democrats go on the record saying that during the cramdown debate Treasury Secretary Tim Geithner “was really dismissive as to the utility of” cramdown, and Larry Summers “was not supportive of this” to any degree, even though President Obama campaigned on it.

What would it mean to foreclosures to be a good or bad thing? There are three levels to thing about it. The first is the individual event level (with borrowers/lenders), another at the residential sector level, and the third is at the macroeconomic level. We are going to stick to the macroeconomic level.

Foreclosures could be a good thing for the macroeconomy. By allowing for a quick deleveraging of a consumer, who loses a massive debt burden and gains new spending power by renting a cheaper residence, it can increase consumer spending.  Or it could be a bad thing for the macroeconomy. It could amplify the effects of financial frictions, like those seen in a Kiyotaki-Moore model, by causing fire sales of residential properties, reducing the value of other properties causing retrenchment by other leveraged players, which leads to reduced consumption through balance sheet mechanisms, which spirals.

The hard part is finding a proper instrument for this analysis. Foreclosures are correlated with many other bad housing situations which are related to overall delinquencies, so it is tough to separate correlation from causation. Luckily, there’s a new paper, Foreclosures, House Prices, and the Real Economy, by Atif Mian, Amir Sufi and Francesco Trebbi. Here’s the voxeu summary (h/t Free Exchange). They use judicial foreclosures, where some states make it harder to foreclose than other states, as an instrument to statistically analysis the impact of foreclosures while controlling for all the other variables that could impact this study.  From the voxeu study (my bold):

Using this instrumental variable approach, we find that foreclosures have a substantial effect on house prices. Our state-level baseline estimate suggests that a one standard deviation increase in foreclosures in 2008 and 2009 leads house price growth to be two-thirds of a standard deviation lower over the same period….Our estimate of the effect of foreclosures on house price growth is robust to extensive controls for demographics and income differences across states…..
We then turn to residential investment and durable consumption. Employing a similar two stage least squares estimation strategy, we find that a one standard deviation increase in foreclosures per homeowner leads to a two-thirds of a standard deviation decrease in permits for new residential construction. Further, a one standard deviation increase in foreclosures leads to a two-thirds of a standard deviation decline in auto sales. Our estimates are robust to controls for demographics and income.
We use our microeconomic estimates to quantify the aggregate effects of foreclosure on the macroeconomy. Our estimates suggest that foreclosures were responsible for 15% to 30% of the decline in residential investment from 2007 to 2009 and 20% to 40% of the decline in auto sales over the same period….

But our estimates suggest that foreclosures lead to more abrupt declines in these outcomes than would be observed in the absence of foreclosures, and these declines are likely to be more painful in the midst of a severe recession. This is consistent with the amplification mechanisms emphasised in Kiyotaki and Moore (1997) and Krishnamurthy (2003). We believe that these results demonstrate a direct connection between a financial friction – forced sales induced by foreclosures – and a reduction in residential investment and durable consumption during and after the recession of 2007 to 2009.

Our estimates of the effect of foreclosures on residential investment and auto sales can partially explain both the magnitude and length of the recession of 2007 to 2009. For example, the sharp rise in foreclosures began relatively late in the recession and continues into 2010. If we combine this fact with the finding in Leamer (2007) that residential investment is among the most powerful components leading the US out of recession, it is possible to argue that foreclosures have likely contributed to the length of the recession and sluggishness of the recovery. Similar arguments apply to our findings on auto sales.

Here are two graphs that get to their point.

As you can see from this graph foreclosures are correlated with a decrease in investment and durable sales growth.  (Notice the giant outlier of Washington DC in the right chart – lobbying as full-employment program!)   But how can we see if this is foreclosures themselves doing this work, or foreclosures actually reflecting housing delinquencies, unemployment, and other bad economic indicators?   This is where they use judicial foreclosures as a mechanism to analyze foreclosures controlling for everything else:

Notice how pronounced the difference becomes in the middle of the recession, when balance sheets are the most weak and fire sales the most severe.

There needs to be more investigation on the actual transmission mechanisms of how this is happening.  Is it abandoned properties generating precautionary savings on the part of neighbors?   Cascading effects of the fire sales?   Either way, this is a major empirical win for those who have said the broken servicing model accelerating foreclosures and the lack of any serious response by the administration would be a problem for the recovery and the future of our country.

This entry was posted in Uncategorized. Bookmark the permalink.

10 Responses to Finally! Empirical Evidence About How the Foreclosure Crisis is Impacting the Recovery

  1. Mike Easterly says:

    What I don’t understand is, when did it become in the banks’ interest to foreclose? A year ago, they were bending over backwards not to have to write down the value of mortgages ( Has the ‘extend and pretend’ era ended? My gut tells me that banks (and their regulators) still have plenty to hide. Am I missing something?

  2. Milton Recht says:

    There are two issues that I believe the paper leaves unanswered that may undo its conclusions.

    Mike K, if you look at your published foreclosure chart in your October 10, 2010 blog of US foreclosures from 1979 onward, you see that US foreclosure rates are upward sloping and increasing with a sharp dip during the recent housing bubble and a sharp rise when the housing bubble burst. This post 1979 upward foreclosure trend, which was masked by the bubble, should be removed from the data. Once the trend is removed there may not be any statistical causality left. I do not believe the authors de-trended the data.

    Second, as I mentioned on Kling’s blog when he mentioned this research:

    “Foreclosed homes are sold “as is,” cannot always be inspected prior to purchase and the purchaser is responsible for removing any residing tenant, previous owner or renter. A foreclosed home can be in terrible condition and require a sizable investment to make the house livable. Foreclosed abandoned homes can have their plumbing fixtures and appliances removed, walls, doors and windows broken, etc.

    Foreclosed homes sell at a discount to non-foreclosed homes. Bloomberg reported, RealtyTrac compared the prices of foreclosed and non-foreclosed sales and found a 27 percent discount for foreclosures, .

    The major home price averages combine foreclosed and non-foreclosed homes in their averages and indices.

    Non-judicial foreclosure states have more foreclosure sales than judicial foreclosure states. Judicial foreclosure states will have fewer homes selling at an average 27 percent discount and naturally (mathematically) will reflect higher average sales prices than non-judicial sale states, which will have a greater volume of homes selling at the discount.

    The behavioral characteristics of occupants may be different in judicial sale states versus non-judicial sale states. The length of time of owner/borrower occupancy post default is likely longer in judicial sale states because of the longer legal process and occupants likely have an interest in keeping things in better condition while they live there. Foreclosures in judicial states can take two years. The discounted price of foreclosed home sales may differ in judicial sales versus non-judicial sales states.

    The question is whether the study put in enough controls for occupant behavioral differences in judicial and non-judicial states and whether the volume effect of the inherent discount of foreclosed homes on the home price index averages was adequately controlled for in the major home price indices and averages used in the study.”

    So, the question is is the data compromised by the increasing post 1979 US foreclosure rate and by the housing price indices not correcting for the unequal weighting of lower prices of foreclosed homes due to property destruction in states with more foreclosure sales.

  3. Mark T says:

    I am not sold on the empirical win because I am not sold on the instrument in question although I agree it is thoughtful research and worth considering. A suitable instrument (x) should affect the final variable (z) only through its effect on an intermediate variable (y). Emphasize “only”. But it is not clear to me that this paper proves a causal effect between x and y, which leaves the impact on z uncertain. I see the correlation but there could be an outside variable or variables (p, q etc) that are driving both x and y. This paper does not tell me that.

    Milton Recht has identified some other questions in his analysis and I agree with them.

  4. Mark T says:

    Just to finish my thought, the implication of this kind of study is that the counterfactual, e.g., no foreclosures, would play out in the opposite direction with more or less the forecast degree of effect. And that is where I become less than convinced. If you hypothesize no foreclosures, would residential investment and auto sales have been more favorable? Maybe, maybe not. It might be that there is a common problem to all of these variables, i.e., the credit market freezing up. Would a no foreclosure rule have ameliorated that (freed up either credit or defaulting homeowners’ cash?) or exacerbated it (frightened lenders into even greater panic by adding political risk to the mix and frozen the credit markets even more)? That is the counterfactual and so I am not sold on this thesis. What happened in the Great Depression when states froze foreclosures?

    This is not to disagree per se with the thesis, just to say , I think there is more work to be done before the question is resolved.

  5. Pingback: This Just In – Bridget Magnus Shows the World as Seen from 4'11"

  6. copy editor says:

    What would it mean to foreclosures to be a good or bad thing? There are three levels to thing about it.

    let me THINK about that THING

  7. Pingback: Thursday Reads | The Big Picture

  8. Pingback: Romney Tells State With Country’s Highest Foreclosure Rate ‘Don’t Try And Stop The Foreclosure Process’ | Greediocracy

  9. Pingback: Foreclosed for Business | Bear Market Investments

  10. Pingback: VIDEO: Five GOP Candidates, Zero Solutions For The Housing Crisis | Greediocracy

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s