Friday, November 07, 2008

David Stiff of Fiserv Writes (important update for S&P Case Shiller Discussion)

He writes:

I am responsible for quantitative research at Fiserv Lending Solutions, the company that calculates the S&P/Case-Shiller indexes.

Your recent blog entry "What Do House Price Indices Currently Mean?", contains some incorrect information regarding the S&P/C-S indexes.

The Case-Shiller model does not include a foreclosure dummy. In general, sales of bank-owned (REO) properties are included in the repeat sales pairs used to estimate the indexes if they occur at least 6 months after a previous arms-length transaction. (Note: Bank repossessions of properties that are recorded at deed offices are not included in the repeat sales pairs, because they are not arms-length transactions.) I am not sure how this misconception about a foreclosure dummy started -- I need to talk to Chip Case about the details of this discussion at the Berkeley/UCLA conference.

If you don't mind, could you please modify your entry to indicate that there are no foreclosure dummy versions of the S&P/C-S indexes? This misconception is generating a lot of confusion for our index customers. Thanks.

A less consequential misunderstanding -- the Los Angeles S&P/C-S index only covers Los Angeles and Orange counties. Data from Riverside and San Bernardino counties are not included in the S&P/C-S index for Los Angeles.

1 comment:

Anonymous said...

Apologies for going wildly off topic here, but...

remember back last year sometime when I popped in and made a cutesy remark about Krugman and Volker riding in on their Indians and whistling at the townies?

It's a dream come true.