Bailout

Lawler: Home Sales in May: A Look at the Data

By Anonymous

Housing economist Tom Lawler is taking the under. The following is from Tom:

While most (though by no means all) of the country appears to have experienced a sharp gain in existing home sales (closed) this May vs. a year ago, the nationwide increase does not appear to have been as high as the surge in pending sales (related to the expiring tax credit) in March and April would have suggested. Last year?s ?comp,? of course, was pretty low: the NAR estimates that existing home sales last May ran at a seasonally adjusted annual rate of 4.75 million, and unadjusted sales were estimated at 447,000. While this May?s ?seasonal factor? should be lower than last May?s (meaning flat unadjusted sales would produce a seasonally adjusted increase), I estimate that unadjusted sales this May vs. last May would have to be up about 19.2% for seasonally adjusted sales to be flat to April. Obviously sales in many areas of the country were up a lot more than that, but in some large states sales showed much smaller gains, and a few saw declines.

Dataquick estimates that new and resale home sales in California were up 4.9%, with the ?small? gain mainly resulting from a decline in foreclosure sales. While this stat is based on deeds recorded and includes new sales, MLS data I?ve seen aren?t too far off. The Michigan Association of Realtors reported a YOY decline in sales in May, with the drop coming many in foreclosure sales in very distressed areas. Sales in both Las Vegas and Phoenix were also down from a year ago, with both areas seeing sizable declines in foreclosure sales.

Many other states saw YOY sales gains that were decent, but hardly explosive. Based on realtor data from North Texas, Houston, San Antonio, Austin, Lubbock, and Wichita Falls, it appears as if Texas sales increased about 19.2%. Other states with ?so-so? YOY sales gains include Iowa (20.8% statewide), Colorado (18% for Denver and Colorado Springs combined), Birmingham (16.8%), Indiana (14% Indianapolis and Fort Wayne combined), Minneapolis-St-Paul (10.7%), Greater Northern Virginia (6.2%), and quite a few (but not all) Florida markets showed very modest gains.

To be sure, many areas of the country saw sizable increases, but interestingly the vast bulk of these areas saw smaller YOY sales gains in May than in April.

When I add everything I have up, and make estimates for areas where I couldn?t find any reliable data, I come up with an estimated seasonally adjusted annual rate for existing home sales that is much smaller than I would have expected a few weeks ago ? something in the range of 5.83 to 5.84 million, which would translate into an unadjusted YOY sales gain of around 20.5 to 20.6%, and would be a boatload under consensus.

Of course, my regional tracking ? which until the last few months has easily produced a better-than-consensus estimate, of late has been low to the downside. More troubling (to me, at least), I?ve had trouble ?reconciling? to the NAR data even after getting state/local realtor sales data not available until the day of the existing home sales report (or later in some cases!).

Nevertheless, the ?raw? data I?ve seen so far suggests that existing home sales in May will come in well under ?consensus? ? for reasons that are unclear, and until recently to my surprise.

On the new home sales front (to be reported by the Census Bureau), anecdotal reports from a wide range of builders, as well as from some local realtors, suggest that new home sales as defined by the CD probably ran a bit lower than last May, when sales (based on contracts signed/deposits taken) ran at a SAAR of 367 k. I?m ?guessing? that the preliminary number for May will be somewhere around 330 k, which would be a monthly decline from April of 34.5% (which is probably close to the monthly decline we?ll see in the pending home sale index scheduled to be release by the NAR on July 1.

CR Note: This was from housing economist Tom Lawler. Note the caveats, but clearly the data suggests a downside surprise.

Full post as published by Calculated Risk on June 21, 2010 (boomark / email).