Monday, April 18, 2011

Housing bubble graph FAIL! (Don't trust Barry Ritholtz)

I noticed Mark Thoma posting this graph from Barry Ritholtz's blog. Barry Ritholtz attributes it to reader Steve Barry. This graph supposedly shows Robert Shiller's famous 2006 inflation-adjusted housing graph updated through 2010. Notice that as of December 31, 2010, housing prices are at roughly index value 140, way above the 1989 bubblette peak.

(Click the image to see the full-sized version.)

There's just one problem. The graph is bogus. At least it isn't updated with Robert Shiller's national home price data. You see, Robert Shiller updates his own graph (it's in his housing data Excel file) which you can see below. According to Shiller, as of December 31, 2010, housing prices were at index value 124, not 140.

(Click the image to see Robert Shiller's spreadsheet and graph.)

It's possible Steve Barry used the FHFA House Price Index [***See updates below.***] to update the graph, but the S&P/Case-Shiller National Home Price Index is considered way more reliable, and it's the data Robert Shiller uses for home prices from 1987 onward. (Why use one data set for the up side of the bubble and a different data set for the down side?)

I also question how Steve Barry came up with his "projection". It looks like he just drew a dashed line where he wants home prices to go, but that's not valid forecasting.

Update: It appears that Steve Barry is making two errors. First, he is using the S&P/Case-Shiller 20-city index rather than the national index that Robert Shiller used. Second, Robert Shiller's portion of the graph is adjusted for inflation, but Steve Barry's (after 2006) is not.

Update #2: I contacted Barry Ritholtz about the issues with the graph. He then contacted Steve Barry. It has been confirmed straight from the horse's mouth, from 2006 onward the disputed graph uses the S&P/Case-Shiller 20-city index, rather than the national index that Robert Shiller used, and it does not adjust for inflation. This graph is really making its way around the web, which is unfortunate because it is worthless.

Update #3: Despite being informed of the error, Barry Ritholtz never issued a correction on his blog.

1 comment: