Monday, July 24, 2017

Housing: Part 244 - Moral hazard in securitizations

Here is an interesting working paper from John Krainer and Elizabeth Laderman at the Federal Reserve Bank of San Francisco. (pdf)

The abstract:
We compare the ex ante observable risk characteristics, the default performance, and the pricing of securitized mortgage loans and mortgage loans retained by the original lender. We find that privately securitized fixed and adjustable-rate mortgages are riskier ex ante than lender retained loans or loans securitized through the government sponsored agencies. We do not find any evidence of differential loan performance for privately securitized fixed-rate mortgages. However, we do find evidence that privately securitized adjustable-rate mortgages performed worse than retained mortgages, even after controlling for a large number of risk factors. Despite the higher measures of ex ante risk, the loan rates on privately securitized adjustable-rate mortgages were lower than for retained mortgages.
There is a lot of talk about how the securitization boom created moral hazard, how you used to go to the bank and get a mortgage and the bank giving you the loan was the one you would be paying, but now they sell the mortgage off and it gets sliced and diced among a million investors who have no idea how good of a credit risk you are, and so the bankers are willing to underwrite garbage, and this caused the whole system to collapse in a heap of greed.

Great story.  Very plausible.  It just didn't happen.  There was no significant increase in securitization as a whole during the housing boom.  The private securitization boom largely happened after prices peaked.  There is a brief period - lasting months really - between the rise of private securitization and the implementation of Fed rate hikes where some markets appeared to have some extra price appreciation.

In the Krainer and Laderman paper, which is based on a large data set from California, from 2000 to 2007, they find some evidence that privately securitized mortgages tended to have somewhat riskier characteristics, but it really doesn't amount to that much - certainly not enough to be the defining development of a massive bubble.  They report: "Indeed, observable risk factors such as current LTV and the subprime and documentation status are estimated to have a much much larger effect on the default hazard than investor type." (page 23)

Here is a graph of defaults, from the paper.  The blue line is retained mortgages, the red line is privately securitized mortgages, and the black line is mortgages securitized with the GSEs.  The GSEs had default rates much lower than either retained or privately securitized mortgages.  And, while privately securitized mortgages defaulted at a faster rate, by the end of the period analyzed here, default rates of retained and privately securitized loans were similar.

This doesn't look very promising for the "moral hazard" view.

In fact, I wonder if one effect of the GSEs is to provide a securitization market for the least risky loans, which would, ironically, leave the banks worse off, because there are many mortgages which can be reasonably made which don't happen to conform to GSE standards.  Those loans could be reasonable, and yet, they will inevitably be more risky than conforming loans.  What we have here is the opposite of moral hazard.  The banks are left with a riskier retained portfolio because of the GSEs.

Might we speculate that this would have led to constrained lending among more marginal borrowers, and private securitization provided an outlet for banks to realign their retained portfolios to be more like the neutral portfolios they would have been in a market without the GSEs?

--------------------------
Richard K. Green and Susan M. Wachter have a paper with some international comparisons.  Now, clearly, if you limit access to liquidity for a given asset, the market price of that asset will decline.  This is conventional financial theory.  Where I part ways which so much analysis about the housing bubble is the universal knee-jerk reaction to blame growing mortgages and expanding credit for rising prices, with little accounting for scale, alternative hypotheses, etc.

Green and Wachter briefly describe the mortgage markets of several nations.

United Kingdom, Canada and Japan do not have significant securitization markets.  Two bubble markets and one market that did not have a bubble.

The US, Denmark, and Germany have significant securitization markets.  Two bubble markets and one market that didn't have a bubble.

The funny thing is, some of these countries tend to have strict down payment requirements while others do not.  Some have recourse loans others do not.  Some tend to have fixed rate.  Others do not.  Some have balloon payments, others have long amortizations.  Some allow pre-payments without penalties.  Others do not.

Now, there are some correlations, like countries with higher down payments appear to tend to have lower prices.  But, there is no obvious pattern with these mortgage market characteristics and housing bubbles.  This was true within the US, too.  There were some areas where private securitizations might have tweaked the market a bit, but for the most part the US market has relatively uniform characteristics.  Yet, homes in a few cities became much, much more expensive than homes in most cities.

Everyone understands that those high prices are triggered by supply constraints.  Yet, in all of these countries, from the US, to Canada, to Australia, debates rage about how to curb lending or keep out foreign investors, etc., as if the mortgage market is the causal factor, and prices are rising, unmoored from rational rental value, because of it.  Yet, if we look at Germany and Japan compared to the UK or Australia.  Or, if we look at Dallas vs. San Francisco.  Clearly supply constraints are a universal factor.  Strange that there are many articles about all that China money flowing to Australia or San Francisco, or Vancouver.  Why are they flying over Seoul and Tokyo to get to those other cities?

(This NAR report suggests that there are as many buyers in places like Florida and Texas as there are in California and New York.)

I don't know about Japan, but Germany and Switzerland appear to have tax policies that don't favor homeowners.  In the US, I find that the monetary value of the GSE subsidy is basically a rounding error compared to the income tax benefits.  So, of the three major sources of price inflation, the order of importance seems to be - supply, taxes, and lending policies.  Securitization is a single factor within the factor that is of least importance.

6 comments:

  1. Another great post.

    Property zoning causes property appreciation if demand for property is growing.

    A city/region may have demand from local industry and residents for property, or it may come from outside the city too, from people or investors who want in.

    There does seem to be a constant--nations with chronic current account trade deficits and property zoning see the most property appreciation.

    ReplyDelete
  2. Add on: I should say "tight property zoning."

    ReplyDelete
  3. This comment has been removed by the author.

    ReplyDelete
  4. OT but interesting

    https://beatricecherrier.wordpress.com/2017/02/20/how-not-to-screw-up-your-economic-expertise-lessons-from-the-kennedy-tax-cut-grandmaster-walter-heller/

    The above post is a fun if long read. But for purposes here, the first few paragraphs are remarkable.

    Macroeconomists are wondering why the public has become disenchanted with them.

    The words "property zoning" are absent from the self-analysis,

    In other words, property zoning is simply not on the table as an macroeconomic topic, despite the fact that high housing costs are eviscerating living standards throughout much of the developed world. Dampening real output. Putting strains on commuters that do not show up in GDP.

    So the middle-class gets boxed out housing in Great Britain and large parts of the coasts in the US. Not a topic.

    Real median earnings for adult U.S males are less now than in 1979 (the key word is "median").

    So macroeconomists jibber-jabber about the need for tight money to prevent asset bubbles…in property.

    A grim chuckle, no?

    ReplyDelete
  5. OT but worth a thought:

    Bullet trains.

    Tokyo not only has "loose" zoning, but bullet trains.

    The trains must mitigate property appreciation in-city to some extent.

    I think. Maybe not. The trains may increase the value of commercial real estate. Interesting topic.

    ReplyDelete
    Replies
    1. Good point. Self-driving cars and busses in the US might also create new dynamics.

      I think you are going overboard on the zoning complaints. I see a rising interest in zoning and the supply problem.

      Delete