Monday, July 20, 2009

House price indices: not necessarily the same story

I compare three competing monthly home price indices: the S&P Case-Shiller Composite 20, the FHFA purchase-only index, and the LoanPerformance HPI. Over the year, the stabilization in home values is evident across the board. However, on a 3-month annualized basis, the majority vote shows a stark second-derivative improvement in home values.

The differences between the S&P Case Shiller Composite 20 and the FHFA (formerly OFHEO) purchase-only index are well known. The FHFA tracks home values of mortgages guaranteed by Fannie Mae and Freddie Mac (conforming mortgages only). The S&P Case Shiller Composite 20 does not discriminate and includes home valuwes tied to jumbo mortgages (non-conforming mortgages) as well. On the other hand, the monthly FHFA covers a broader geographic region, including all of the census regions, while the S&P Case Shiller Composite 20 covers just 20 metropolitan areas.

The monthly LoanPerformance HPI claims to be both geographically superior, building its index up from the bottom at the zip-code level. It covers all 50 states, including D.C. (see its methodology at the bottom of the page), and tracks home values of all loan types. This is the HPI used by the Fed to calculate the value of real estate assets in the Flow of Funds accounts.

The chart above illustrates the annual growth rate of each home price index, where each index is showing stabilization in home values on a Yr/Yr basis. However, over the last three months, it is a very different story.

Thi chart illustrates the 3-month annualized growth rate of each home price index. The annualized growth rate is the implied growth rate over the next year if the next 3 quarters saw the same growth rate as the latest quarter (February '09 through April '09, the latest data point).

Here, the stories diverge. The S&P Case-Shiller Composite 20 is still falling at a very quick rate, -18% annualized. However, the FHFA and LoanPerformance HPI are showing stark improvements over the last 3 months, -5% and -3% annualized growth.

If this was a majority vote, FHFA and LoanPerformance would win, and the monthly growth in home values is not as dire as suggested by the S&P Case Shiller Composite 20. There has been significant second-derivative improvement in the last three months.

Rebecca Wilder


  1. Very interesting comparison, especially on the last chart. Would the percentage of the total mortgage market each occupies be available?

  2. Hi Aunt Jane,

    That's pretty difficult to say - and not reported (from what I know of). There are always many more houses off the market than there are on the market. And furthermore, most of these home price indices do not include condo sales....

    Thanks for commenting! R

  3. Kenneth G. Smith IIJuly 20, 2009 at 5:24 PM

    Post to this comment

    Excellent point, but take it one more step.

    If someone can pay rent, that rent can be considered part of a mortgage payment. The government is providing $8,000 for first time buyers, so why can’t the government pay part of the payment and have the borrower repay the government in the future. Here is an example of how it could work.

    • Mr. and Mrs. ZZZZZ have a mortgage payment of $1,170 ($200,000 loan with 30 year payout at 5.75% interest).
    • The ZZZZ’s lose their job and can only pay $470, so the government pays the difference of $700
    • So the ZZZZ’s remain homeowners and work through their problem. It takes the ZZZZ’s 10 months to get back on their feet, the government paid out $7,000 and now the ZZZZ’s owe the government.
    • But the government says okay, you can start paying us back in seven years and the payment will be over 10 years at an interest rate of 3%.

    What the government has done is to provide assistance to the property owner (just like the bailout plans for the Financial Industry and Automotive Industry) and requires them to pay back the obligation starting in seven years. This is not a freebie, but short term assistance. Franklin Roosevelt called it Lend Lease.

    This program is not perfect, but it can assist a lot of people who want to own homes. Most importantly, it is channeled directly to the property owner, not a large corporation that has other motives besides keeping the property owner solvent. A significant benefit of this program is that payments to financial institutions will resume and cash flow will get back to normal levels, thus credit availability should improve.

    There needs to be conditions such as confirming gross income via income tax statements; confirming employment and confirming current payroll. The only group of individuals who would be excluded are those who own more than one property (there should be no break to the investor who treated real estate as a business) and cases where mortgage fraud exists in the form of straw buyers and invalid sales (properties that sold more than three times within five years and the value change was greater than 150%).

    • This total assistance would be capped at $50,000 and could run for 24 to 36 months
    • In a given year up to $25,000 could be provided.
    • The government would be releasing the funds over 12 months, thus the federal outlay would be limited.
    • The total cost of $10 million loans receiving assistance would be $250 billion per year or $500 billion in total.
    • This is much cheaper than the TARP bailout and part of this can be funded with the current $70 billion in TARP repayments.

    The greatest difficulty in implementing this program is processing and accounting. Loan Servicing companies would need to add staff (if one servicer can process 50 applications a week, 4,000 servicers would need to be hired, plus additional support staff) Wow, as many as 10,000 new jobs would be created. Add to this job creation the fact that several million homes do not go into foreclosure and more jobs are not lost due to desperate situations.

    Yes it is possible and yes it can work.

    The reason it can work is because real estate goes through cycles. If people are forced to sell at liquidation prices, everyone loses. Give property owners a chance to get back on their feet, get back to work and the whole economy starts to turn around.

    As stated earlier, this is not perfect and many will complain about the injustice. But think about the injustice of the corporate bailouts, the injustice that first time home buyers get a break, the injustice that shareholders come before the individuals who created value in the companies by buying products. One can go on and on, or we can try.

    We only fail if we do not try.

  4. May 2009 Loan Performance HPI has been released and shows further improvement in the second derivative.

    C-S is still a superior metric for measuring aggregate national housing wealth because it is value (or capitalization) weighted not count (or transaction) weighted.

    Take a simple example. A $100,000 house appreciates 5% in Lafayette, LA and a $500,000 home in Lafayette, CA depreciates by 5%. FHFA and HPI would say that home prices are flat. C-S would say that home prices are down 4%. Now extrapolate nationally.

    It is arguable that the y/y difference between HPI and FHFA is the inclusion of non-conforming transactions in the HPI data and that the y/y difference between the HPI and the C-S is the weighting difference in the C-S data.


Note: Only a member of this blog may post a comment.