In Depth Analysis: CalculatedRisk Newsletter on Real Estate (Ad Free) Read it here.

Monday, April 08, 2013

Lawler: Goldman's Pretty Weak Argument that “National” Home Prices Last Year “Really” Increased by 3-4%

by Calculated Risk on 4/08/2013 05:38:00 PM

From housing economist Tom Lawler:

In Goldman Sach’s weekly “Mortgage Analyst” report (April 4), Goldman analysts argue that the 7-8% “national” home price growth rate in 2012 suggested by some “major” home price indexes overstated the likely “true” growth rate in US home prices, and that “national” home prices more likely increased 3-4%” last year. Here is a quote from the piece.

“(W)e argue that national indices weighted by transactions (flow weighted) rather than housing stock (stock weighted) inflate the measured growth rate. If one controls for the “weighting effect” as well as the effect of a declining share of distressed sales, national house prices more likely increased 3-4% than 7-8% in 2012.”
Interestingly, the piece includes a table showing the 2012 growth rates in a number of different “national” home price indexes, some of which include distressed sales but others that don’t; some of which are “flow” based; some of which are housing stock based; some of which are “hedonic”; and one of which (the S&P Case-Shiller “national” HPI is a “flow/stock hybrid” (flow based for Census Division HPIs, but stock based (in value) when aggregating the Census Divisions into a “national” HPI).

House Price Indexes 2011 and 2012Click on table for larger image.

What Goldman analysts don’t explain, however, is if “national” home prices adjusted for adjusted for shifting “distressed” sales shares and “stock vs. flow” weighting “really” increased by 3.4%, then why did EVERY HPI that excludes distressed (or at least foreclosure) sales and which is “stock” weighted increase by MORE than 3-4%?

The piece does, in a sloppy way, make a good point (which I’ve made many times): how one “builds up” local home price indexes to “national” home price indexes (e.g., unit vs. value stock weights, granularity of geographic HPIs, etc.) can have a substantial impact on the “national” HPI. But it’s estimate of “true” national home price growth last year is way too low.

CR Note: This is a great summary table, although, as usual, ignore the NAR's median sales price. This suggests to me that "national prices" increased about 6% to 7% in 2012, after falling about 4% in 2011.