Bank’s Bad Math – Home Price Indexing Responsible

Posted on September 30th, 2008 in Daily Mortgage/Housing News - The Real Story, Mr Mortgage's Personal Opinions/Research

The Case-Shiller Index below shows exactly why all raters and banks modeling systems have been constantly incorrect. This is why values of securities are thought to be higher in value then the street is willing to pay. There is a market for mortgage securities of all kinds, just not at the level at which the owners are willing to accept.

The HPI, Home Price Index, variable is the math that has screwed up the raters and banks from day one.  Everyone who touches a mortgage loan or security has a very elaborate modeling system that requires some variables to be entered by the user. One is the Home Price Index.

Very simplistically, the lower the rate of depreciation plugged into the model, the more it says the security or loan is worth.  Multiple choices for Home Price Indexes including Case-Shiller, OFHEO and proprietary modeling that covers large geographic regions are partially responsible for the valuation mess.

For example, below shows the San Francisco MSA down 24.8% year-over-year.  It consists of 5 counties and 4.4 million people. Within the SF MSA you have Tiburon down 5% and Antioch down 70%.  This presents problems because in many cases, especially with subprime loans, loan program types are clustered and localized. For example, there are no subprime loans in Tiburon, they are all in Antioch.

When you apply the true 70% house price depreciation figure across the subprime securities, the models flip out and the next thing you know, subprime securities are worth 15 cents, which is exactly where they are trading for today. The collateral support just is not there any longer, as house prices have fallen so hard in the past year. Even Case-Shiller has it wrong…both the CA Assoc of Realtors and DataQuick agree that the CA median home price has fallen some 40%.  The distressed securities market has it right, the banks and raters models have it wrong.

This is the exact same reason why the banks will have a tough time coming from all of the higher grade mortgage paper they still hold on balance sheet, such as Prime and Jumbo Prime as it defaults.  Banks hold this paper at face value or above when in reality it is worth 50% of that or less as it defaults. Recovery rates in harder hit states like CA where most of the loans lie are from 20-50 cents on the original face value. This goes across even Prime Fannie/Freddie paper.  Remember, ‘its a solvency problem, not a liquidity problem’.

The Fix…

Instead of doing away with mark-to-market they should be mandating that all banks and raters use a standardized modeling system in which Home Price indexing was done on a zip code or smaller basis. This way everybody’s securities and loans are priced using the same valuation methods and metrics.  You would have instant transparency and can actually build a bustling market for these assets. I can tell you there are buyers for these assets right now, but at market value.  The reason the banks say there is no market is because they refuse to sell them at market value. -Best Mr Mortgage

Table courest of Big Pcture Blog.

Other Related Mr Mortgage Posts

17 Responses to “Bank’s Bad Math – Home Price Indexing Responsible”

  1. If you did your homework, you would have found the following:

    A good number of financial institutions already avail themselves to zip code-level metrics to update mortgage collateral values/LTVs. Case-Shiller sells zip code level data (S&P doesn’t publish this).

    Neither of the two largest ratings agencies took this approach with their ABS/MBS ratings. In fact, S&P’s ratings model baked-in OFHEO indexes. This is problematic on a few levels – not only are those indexes not available below MSA geographies, OFHEO measures are biased based on appraisal data they include (and thus paint an overly optimistic picture of market trends), and unlike Shiller’s index, OFHEO doesn’t include the most volatile (jumbo & subprime) housing segments – the very collateral behind the private-label MBSs they rated!

  2. MM, great analysis as usual. Zip codes would be by far the best way to approach this question of more accurate valuation. But of course, the banks arent very good at change or the RE market.

  3. Doesn’t the government plan on turning mark to model into the same thing as mark to market with our 700 billion dollars? Why not the 6% that M Lynch got?

  4. We Cannot Handle The Truth

    What we have here is a failure to communicate.

    The facts are that the debt levels of the U.S.A., U.K., Ireland, Spain, Italy, Australia, New Zealand, Canada, and many other European and ‘advanced’ economies are too large to be serviced by the income these economies generate.

    There are only two ways out of this:
    1.) Debt destruction through default leading to a severe recession or depression after which normal and stable growth can resume.
    2.) Postponing the problem by the creation of new debt with inflationary results and the eventual necessity of having the above occur latter.

    We must pass through debt destruction now or take on more debt and go through a much larger debt destruction latter.

    That is the choice. All the other posturing and spin is bull feces.

    Paulson and Bernanke and Cox and Bush and Wall Street and the House Leadership and the Senate Leadership, all hoped they could ‘contain’ the problem until after November 4th.

    The only rational solution is the Irish Solution, that is protect depositors and thus stop the bank runs, and let investors take a bath.

    Dow 8,000 will happen sooner or latter. It is better sooner than latter.

  5. Be it a house, a car, complex mortgage backed securities, or the shirt off my back; the value of anything is what an able buyer is willing to pay for it. The markets have spoken, the banks don’t like what they are hearing. It is no different than home sellers who won’t let go of 2005 and 2006 prices so their homes just sit there. I find it hard to believe that anyone can say with a straight face that taxpayers won’t be losing money by buying stuff that is above what the market will buy it for. All we hear is that hopefully it will be worth more years down the road and if it is we might even make money. I’ve got a word for that: speculation, over leveraged speculation to be exact. That is precisely what got us to this point. The reason the Great Depression dragged on for a decade was because of government’s repeated failed attempts at stopping the market from clearing.

  6. Dear fedwatcher,

    As Ed McMahon would say: You are correct, Sir.

    2 points for you


  7. SEC changes the rules. You decide if this is good or not?

  8. Matt Dubuque

    The interbank lending market remains in a coma threatening us all, irrespective of what the Dow Jones Industrial Average may be doing on a short-term basis.

    As Senior Economist Gordon Sellon of the Kansas City Federal Reserve discussed in his seminal paper “Monetary Policy and the Zero Bound: Policy Options When Short-Term Rates Reach Zero” published in the Fourth Quarter 2003 edition of the Kansas City Federal Reserve’s “Economic Review”, the Federal Reserve should now consider under taking “twist” operations in the open market.

    That paper is available here at the bottom of the link:

    A “twist” operation by the Federal Reserve in the current context would consist of the Fed SELLING 3-month Treasury Bills while simultaneously PURCHASING 5-year Treasury Notes. Such operations, applied judiciously, would affect the term structure of various markets in a positive way.

    Such “twist” operations are not without precedent. It was performed during the Kennedy Administration:

    I urge people to STUDY Sellon’s critical paper at the link provided above to grasp some of the subtleties involved here before engaging in uninformed knee-jerk criticism.

    This is not a cure-all, but it is clear that it should be on the short list of our policy options.

    Matt Dubuque

  9. The CDO structures have a lot of indirection that makes it fuzzy on marking the values correctly for each payout tier. Many CDOs were made out of pure junk so first they should not be considered A paper in anyway.

    A bigger problem is that just because a $500K mortgage was issued in area X a year ago, it could easily be down more than the home price index. There was appraisal fraud just to make loans happen (or else the appraiser could be blacklisted for future work). The 25% and 75% percentiles move at different speeds than the median house price index.

    For this to work, there would have to be a lot of normalization of data which would take a lot of due diligence that was skipped when the mortgage was issued.

    These assets need to be priced lower than current market value because we all know loan defaulting and house depreciation will continue, but also their initial values are essentially indeterminate.

  10. This gets at the whole point of Paulson Plan.

    These securities can be underwritten and valued by rational parties. But there is NO BID today because all the opportunistic capital is waiting for the next shoe to drop. The only bid is predatory – like Lone Star taking Merrill to the cleaners.

    If Treasury has $700BN to BID with, hungry money will come off the sidelines. If the Feds are paying too little, others will overbid. If the Feds are getting everything they’re overpaying (and if managed by folks like Gross this won’t happen). Once markets are again being made, the issue goes away. In all likelihood the taxpayers spend a fraction of the $700BN, but Paulson needs a big number so everyone knows he’s real.

  11. The reason the banks don’t sell the assets, is because they can’t sell the assets at those prices. It is called insolvent.

    Balance Sheet Before

    Assets Liabilities/Equity

    1000 88

    Balance Sheet After (BK)

    Assets Liabilities/Equity

    80 88
    0 or (-)

  12. Wonder at what price this bailout plan will buy these insolvent assets? 50% price drop needs 100% gain just to break even. Taxpayers are in this investement for long haul.

  13. MM, We do indeed use local level data in valuing these mortgages. We go one better and typically get random drive-by appraisals (experienced appraisers working for us) to audit the values. This is not really that hard to get “as of today” values on these properties. The real killer is trying to figure out how much further the asset prices are going to fall, and what impact that will have on loan performance. You have already illustrated in many articles the propensity for people to walk away from negative equity positions even if they can affort the payments. The recession, or whatever you want to call it, is the real wild card in ALL of these models.

  14. “Walking away” needs to become much more expensive and difficult.

    As it is now it’s too “rational” and as a result we have our neighbors’ failure to abide by the terms of their contracts taking all of the rest of us down.

  15. As correctly noted by the first poster, many folks do employ zip level models. But you do have a point that many do not.

    There are two other errors in your post:

    1) It is not the median value that matters, it is the same sale repeat index that matters. You compare median sales value to repeat sales indices. This doesn’t make sense.

    2) You first make the point that the subprime loans are in Antioch, not in Tiburon but then claim that GSE paper is also in trouble. Well, your same observation is true for the GSE loans compared to the subprime loans, they are not in the same place.

  16. GSE paper – in Antioch values are down 70% and it is a heavier subprime region. However, subprime may only consist of 20% of all loans. The rest are Alt-A, Prime and Jumbo Prime. By virtue of values dropping 70% from the peak and being back at 2001 or prior levels, those higher grade loans just because multiple grades lower. Especially given the the higher grade paper in recent years allows 50%+ debt to income ratios. When at least 50% of your gross income is going into your home that is now several hundred thousand dollars underwater, walking away makes good financial sense. Thats why we see defaults in real time of the highest grade ‘a’ paper loans surging in these areas. It is the negative equity effect in living color.

  17. Some more facts:

    The GSE book is performing at around 1% seriously delinquent

    It’s not just the percent of loans that are subprime vs prime, it’s when those loans were originated. Nationally, the GSE share became very small in 2006 & 2007 and non-agency subprime and AltA loans became about half the market. So, the GSE loans have suffered much less house price decline than the non-agency loans. In Antioch, for instance, my guess is that close to 80% of all the 2006 & 2007 loans are non-agency in Antioch.

    Finally, close to 80% of GSE loans had a 20% downpayment.

Leave a Reply

XHTML: You can use these tags: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>