Wednesday, July 29, 2009

Those cheery US housing numbers, reconsidered

A debate is brewing between über housing blogger Calculated Risk and the equally anonymous Free exchange blog. The crux of the debate is whether the May numbers for US house prices, as represented by the Case Shiller index, are as bullish a sign as commentators (read: the media) seem to think.

Here, for instance, is how the FT reported the numbers:

US home price rise hints at stability

US house prices showed their first monthly gain in three years in May offering another sign that the stricken residential real estate market is stabilising.

And the WSJ:

Home Prices Rise Across U.S.

Home prices in major U.S. cities registered the first monthly gain in nearly three years, according to a new report that provided fresh evidence that the severe U.S. housing downturn could be easing.

Even David Rosenberg thought he’d spied green shoots. But Calculated Risk was having none of it:

…the Case-Shiller report today really bothered me. To be more accurate, the reporting on the Case-Shiller report bothers me. As I mentioned earlier today, there is a strong seasonal component to house prices, and although the seasonally adjusted Case-Shiller index was down (Case-Shiller was reported as up by the media) - I don’t think the seasonal factor accurately captures the recent swings in the NSA data.

I have no crystal ball - and maybe prices have bottomed - but this potentially means a negative surprise for the market later this year - perhaps when the October or November Case-Shiller data is released (October will be released near the end of December). If exuberance builds about house prices, and the market receives a negative surprise, be careful. Just something to watch later this year (I will post about house prices, but I will not mention the possible impact on the stock market in future posts)

Free exchange was not quite so persuaded, arguing thus:

…the first step in [CR’s argument] is that all of the writers out there trumpeting rising home prices are paying attention to the not seasonally adjusted data. This is true, but not that damning; it’s not as if the seasonally adjusted data show a sharply different picture.


Which then leads Calculated Risk to argue that the seasonal adjustments are insufficient. The evidence? Well, a chart of the NSA and SA data together shows that the SA data series hasn’t been smothing seasonal spikes in the last decade as completely as it did during the 1990s (although the SA series is pretty smooth between 2005 and 2008; it basically points straight down). He therefore concludes that seasonal factors aren’t adequately being taken into account, and that prices will surprise to the downside in the fall.

Free exchange cites to main reasons to doubt CR’s pessimism:

The first is that I don’t understand why the seasonal factors shaping housing prices would have intensified in recent years. If anything, it seems as though the typical seasonal pattern should have weakened, as more market participants bought or sold out of necessity.

And the second point is that I don’t see how you judge the importance of cyclical factors-which are hugely important-relative to normal trends. The recession began in the winter, then leveled off last summer (at which point the economy was actually notching an increase in output), then imploded in the fall and entered free fall this past winter. There’s nothing seasonal about the cycle, that’s just how the timing worked out. But that timing has obviously had a major influence on housing prices.

Conclusion? We’ll see.

In the interim, Rosenberg has backpedalled a bit from his green shoots call, saying in a note on Wednesday:

On a seasonally adjusted basis, and that is really the only way to look at the data sequentially, 8 of the 20 major cities still posted house price declines in May (though the front page of the WSJ chose not to mention that fact — Home Prices Rise Across U.S.— it screams). At issue is whether (i) home prices have indeed stopped falling, (ii) whether this is just noise in what is still a fundamental downtrend, or (iii) there are some methodological problems with the data.

Indeed, Rosenberg contends that the index itself is flawed, because the price data:

are likely not accurately taking into account the mix of actual home sales because it has come to our attention that an increasing number of high-end homes are now entering the foreclosure sales process, which are skewing these price indices higher after a prolonged period when the data were being pulled down by the preponderance of ever-cheaper subprime units hitting the auction market.
The Gluskin Sheff chief economist notes that at the beginning of the year when prices were sliding more than 2 per cent per month,

half the sales were coming from foreclosure auctions and many of these were low-end units; now the foreclosure share is down to 30% and more of these are higher-end homes seeing as prime-based mortgage default rates are now rising faster than subprime.

And, he says:

the banks are sitting on a record number of foreclosed units that have yet to hit the market (don’t forget that the government-imposed moratorium just terminated). And once these homes flood the market, we may well get another big leg down in the price data — the shadow inventory of homes is very troubling and is not counted in the official data from the National Association of Realtors or the Commerce Department. To date, the sales of foreclosed units have been rather limited and we are hearing that there has been bidding wars for the well-priced properties at these auctions — these bidding wars may dry up as more supply comes onto the market, which is the major point.

The downgrade pressure on house prices from fundamental factors like inventory levels is significant, Rosenberg argues - the bottom is a ways off yet.

Related links: A comment on seasonal adjustments - Calculated Risk

No comments:

Post a Comment

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