Wednesday, April 18, 2012

Affordable Housing as Source of Bubble

...policies to push home ownership weren't responsible...
Working papers are the place to go for recent research; journals are passe, too far behind the times to be my first resort.
A recent paper by Rubén Hernández-Murillo, Andra C. Ghent and Michael T. Owyang is a good example. Furthermore, as with most working papers, it's freely downloadable, though you'll need to have a subscription when it comes out as a standard journal article some years hence. Here I provide the title, abstract and a link. Of course (well, not all papers are that way) it's written with economists in mind and so even the abstract is heavy on jargon -- though their one-word summary is clear and says it all. Anyway, I'll translate...
Did affordable housing legislation contribute to the subprime securities boom?
No. In this paper we use a regression discontinuity approach to investigate whether affordable housing policies influenced origination or affected prices of subprime mortgages. We use merged loan-level data on non-prime securitized mortgages with individual- and neighborhood-level data for California and Florida. We find no evidence that lenders increased subprime originations or altered pricing around the discrete eligibility cutoffs for the Government Sponsored Enterprises (GSEs) affordable housing goals or the Community Reinvestment Act. Our results indicate that the extensive purchases of risky private-label mortgage-backed securities by the GSEs were not due to affordable housing mandates.
Let's crunch the data before and after policy changes to see whether the number of sub-prime mortgages changes, or the terms and conditions changed. We put in variables known to affect the mortgage market, which otherwise might obscure (or exaggerate) the effects of policy. Furthermore, we'll use really detailed local data to make sure we're comparing apples to apples, and don't by accident compare a region with lots of subprime mortgages to one with few. Finally, looking at those who all along would have qualified for subprime mortgages doesn't tell us much, or rather may obscure what is going on. For example, the number of such people may have increased reflecting changes in demographics and job markets. So we'll look at people around the margin, and not in the middle. The result: before and after are the same. Policy didn't do it.
Now that's what those who have "eyeballed" the data have long maintained. But we worked really hard to see if any of the reasons given that "eyeballing" was wrong held water. They didn't. No smoking gun. No evidence, period.
OK, maybe the authors wouldn't go quite so far ... but that's how I as an economist read the subtext of the abstract.
...Mike Smitka...
Link to pdf: (don't click if you don't want the actual paper!)

Has Unemployment Really Fallen?

...retirement is not the answer...
The Financial Times has a good April 17, 2012 article looking at the sources of continued high unemployment and declining labor force participation.
On the unemployment side, there are structural arguments—workers live in the wrong places, have the wrong skills—and cyclical arguments. The article finds a bit of both, with the suggestion that the latter is pretty strong.
A drop in participation can potentially mask far higher underlying slack in labor markets, as people give up looking for jobs, or return to school (and stop looking for jobs), or retire (and, duh, stop looking for jobs). But to be counted as "unemployed" in our monthly labor force survey you have to be actually job hunting, not merely job dreaming. My own calculations show participation has dropped 2.2%, using the average going into the Great Recession and the average of the past 6 months. If all those would in fact have temporarily given up on finding work and so are counted as unemployed, that would push the current rate from 8.1% to 11.5%.
Other countries use similar methods while differing in details, such as how recently you have to have undertaken concrete job search efforts. Japan uses two weeks, the US four, and so unsurprisingly Japanese unemployment tends to look low from a US perspective.
Technically they are NILFs, Not In the Labor Force. But how many of those who have officially exited the labor market retired according to plan? The higher that number, the less that this shift in participation matters.
A block of the baby boomers have now hit age 62 and thus can potentially begin collecting their Social Security pension. Leaving aside the possibility that many had no intention of retiring, how much of the bump in NILFs is due to people age 62 and above? I'm doing such calculations for Japan and will post those results in due course to my Japan and Economics blog. Unfortunately, comparable age-specific data are not publicly available for the US. The gist of the FT article is that (unfortunately) intended retirement is not the cause.
The recession, in other words, remains really bad. The graph below thus projects the pace of recovery in employment using pre-recession employment-to-population ratios. At the current pace, we'll be back to par in 2016, assuming the the rapid pace of recent months continues unabated for the next four years.
...Mike Smitka...
Note: Another component of the drop in participation is due to those who qualify for disability pensions and actually "retire" and start claiming them. While some of that is because of the onset of disability, some is a function of the difficulty for an older individual to land another job. leading them to drop out of the labor force (become a NILF) and claim a pension. Herman Schwartz, a political scientist at UVA, has looked at this in the European context, and finds strong evidence that in hard times (or when unfavorable adjustments are made in the rules of regular retirement pensions) those who claim disability increases, indeed increases a lot. Of course that means that people aren't naturally welfare cheats, to use current US labels: lots of people who could retire due to age or disability in fact don't. Most people prefer work to just sitting around...