A Bimodal, Metrocensual Model of Foreclosures

I will be brief – in my discussion of foreclosures, and the extent to which they can be “explained” by either:

• Falling home prices, and/or
• Rising unemployment.

For data, I will use:

• RPX/RadarLogic home prices;
• US Census unemployment rates for 25 large metropolitan statistical areas (MSA’s); and the
• RealtyTrac 2008 U.S. Foreclosure Market Report™.

To build a simple and rough:

• bimodal [two factors, home prices and unemployment]; and
• metrocensual [metro area, determined by US Census Bureau]

model of foreclosures.

Bimodal Distribution

A “bimodal distribution” is a continuous distribution with two different modes or distinct peaks.

It most commonly arises when two different unimodal (see Figure 1, below and left) distributions - each with one peak - are mixed or combined. When this occurs, it may produce a distribution with two peaks (see Figure 1, below and right).

Figure 1: Examples of Unimodal (left) and Bimodal (right) distributions.

I’m discussing the bimodal (or two peak) distribution for one simple reason – to indelibly drive home the point that there are two distinct “causes” of the growing foreclosure crisis:

• Falling home prices, and/or
• Rising unemployment.

The recession that began in 2007, and was predicted in 2006 by an inverted yield curve (see my “Learning Curves” of 26 Oct 2008), began with falling home prices.

This produced the collapse (and eventual government assumption?) of the investor-owned US banking industry. It also produced layoffs and unemployment (see my “Size Matters” of 5 Feb 2009).

So there are at least two, repeat – two, “causes” of foreclosures.

Setting Expectations

Before continuing, please note that regional foreclosure rates are tough to model. Leslie McGranahan, for exampled, describes a 13-factor model of state foreclosure rates in the Dec 2007 issue of Profitwise - The Determinants of State Foreclosure Rates. The model explains about two-thirds of the variation in foreclosures. We’ll try and make do with one (or two!) carefully chosen factors.

Quick Review of 2008

The graph below charts 2008 foreclosure filing (default notices, auction sale notices and bank repossessions) rates as reported by RealtyTrac® for the 25 MSA’s covered by RPX/Radarlogic. I’ve sorted the data in Figure 2 to improve readability.

RealtyTrac reported that just under 2.0% of all US housing units received at least one foreclosure filing during 2008, up from 1.0% in 2007. For the 25 MSA’s in the chart, foreclosure rates ranged from a low of 0.7% for the NY MSA, to 8.9% for Las Vegas.

Figure 2: 2008 Foreclosure Rates, by MSA (Source: RealtyTrac® 15 Jan 2009.)

Impact of Home Prices and Unemployment

The two Scatterplots below depict the relation between 2008 MSA foreclosure rates and the:

• 2006-2008 home price depreciation (“HPD”); and
• 2007-2008 unemployment increase.

As expected (see Figure 3), larger home price declines were associated with higher foreclosure rates. In addition, there was a weaker relation between higher unemployment and foreclosures.

Figure 3: Relation Between Foreclosure Rates and Home Price Depreciation [top] and Unemployment [bottom].

But if you want to come up with an even “better fitting” simple foreclosure model, you can create one by using – as your explanatory variable – the product of the decline in home prices and the increase in unemployment, that is:

• Increase In Unemployment Rate x Home Price Depreciation

You get something that looks like this:

Figure 4: Relation Between Foreclosure Rates and Unemployment x Home Price Depreciation.

Why This Matters

The Administration provided additional details about its Home Affordable Modification Program on March 4th. Eligible borrowers, facing financial hardship, may temporarily:

• lower the rate on their first mortgage to as low as 2.00%; and
• extend their loan term to as long as 40 years

if their monthly mortgage and escrow payments exceed 31% of their gross monthly income.

The program is, at best, - a hair of the dog cure for the non-traditional mortgage products that contributed to today’s crisis. No private lender would be allowed to peddle a similar program in today’s market.

More than half of all HAMP borrowers will become delinquent again if the private sector cannot create real jobs with higher pay before the temporary teasers expire - after the next Presidential election.

Having demonized the financiers, our elected representatives must quickly abandon their bullying and figure out how they can best encourage private industry to create real jobs paying real wages. That will be the best way to prevent foreclosures. Then maybe we’ll all be able to sleep at night.

Comments

[user 777:]

It would be nice if the author gave a definition for Metrocensual, or at least what the hell he describes it as. It's not in the dictionary.

Nick

Mar 13 08:51 AM

[Ira Artman:]

User 68420 - Thanks for reading, and sorry that you missed the definition which appears early on in the piece. I've reprinted it below (or see above), reformatted for emphasis. - Ira

"To build a simple and rough:

bimodal
[two factors, home prices and unemployment]; and

metrocensual
[metro area, determined by US Census Bureau]

model of foreclosures."

Mar 13 09:46 AM

[cynewulf:]

I liked it. watch out banks

Mar 13 09:58 AM

[John Preston:]

So, maybe the bulk of the problem is not "stupid" home buyers who were delusional or liars....maybe something else is at hand....like falling household income....thanks, Ira!

Mortgage lending offered, and borrowers accepted, too much monthly mortgage debt...income was over-leveraged...by ALL lenders, distorting the housing market.

The decline in household income, which is at the heart of much of the current crisis, was exacerbated by the over-leveraged lending practices of the recent past, and left the American household no wiggle room...and, byh extension, the economy had no wiggle room!



Ira wrote: More than half of all HAMP borrowers will become delinquent again if the private sector cannot create real jobs with higher pay before the temporary teasers expire - after the next Presidential election.

Having demonized the financiers, our elected representatives must quickly abandon their bullying and figure out how they can best encourage private industry to create real jobs paying real wages. That will be the best way to prevent foreclosures. Then maybe we’ll all be able to sleep at night.

Mar 13 10:54 AM

[kotika98:]

ok, so where is that bimodal distribution used in your model?

Second, you first have two scatterplots where you are trying to predict unemployment and then home depreciation using the foreclosure rate, and then conclude with a graph where you do it the other way around, trying now to predict the foreclosure rate by using the HDP and unemployment. So which quantity are you trying to "predict"?

Third, you are using monthly data for the period 2006-2008, where all three quantities happened to be steadily going up. So, regressions are bound to produce a high R^2 no matter what. I would be much more impressed if data from the last 25 years was combined, and i would trust the regressions much more even if you used yearly data instead of monthly.

Mar 13 01:43 PM

[Socialism c...:]

You need to get through your head one thing: foreclosure is NOT a crisis, nor are declining home values: they are measures of lack of qualification to have bought, and affordability, respectively. If the current owner is foreclosed, let him buy again a couple years from now, at the newly lowered pricing. The lower prices are a BOON for would-be buyers who were WISE enough to live within their means to this point. QUIT the DAMN social engineering!!!

Mar 13 05:36 PM

[Socialism c...:]

One more thing...if you wanna talk about declines in real income, look at income tax policy and how much income that takes from middle class on up. Look also at the gouging that property taxation is, and how much of a monthly mortgage payment it now represents!! Your model is wholly insufficient!!

Mar 13 05:38 PM

[Teutonic Kn...:]

Mr.Artman,

Thank you for your elegant footwork on your "Bimodal Model". It certainly came across with a big and somewhat intimating title.

BTW, the name Bimodal strikes my nerve as previously I came across a similar Bimodal Model that was pretty sophisticated (and very hard to understand DARPA PhD-level mathematics) from a renowned Cornell professor some years ago.

In any case, I appreciate the bottom line that you are advocating - the real effective means of combating foreclosure lies in the creating productive jobs that garnered competitive goods and services.

At this point, however, it just occurred to me the recent story of that young man with his newly-minted MBA who commanded an $80K starting salary on Wall Street but just got laid off. The poor guy had to resort to selling himself at public transportation crossings and, luckily, landed some alternative work.

Oh, well, as a boring lowly engineer having practiced the trade for several decades, I could tell you that it would take tens of years for one to reach that kind of salary (subject to being appraised fully satisfactorily annually too).

Now you see what I mean to say: For the average person, and I consider myself average, there is really no such thing as a SHORT CUT in life. The only way to achieving a modest means of "success" remains that old fashion way - Hard Work!

Only the private sector could create meaningful, lasting, and innovative jobs.

Mar 15 08:40 PM

[Arnold Semmons:]

Can I use this information to determine when would be the optimum time to buy a house (at the bottom of the market)? I would love advice such as... "as soon as unemployment starts to drop, it typically takes 15 months before housing prices start back up."

Mar 31 05:32 PM

[Ira Artman:]

Arnold S:
Thanks for reading.

Perhaps. But think - as noted by today's release of S&P Case Shiller, as wonderfully highlighted at SoldAtTheTops' Blog:

paper-money.blogspot.c...

the scale of current decline blows away past declines, so optimization/market timing must use history of much weaker corrections. Think that will make it tough. - Ira

Mar 31 06:19 PM