[My thanks to everyone who forwarded me this one- and there were a lot of you!]

According to yesterday’s Arizona Republic, Jay Butler’s April home sales report was wrong:

Butler compiles a report each month on home-resale transactions in Maricopa County.

The report said home resales were up 15 percent compared with the same month in 2007, the first year-over-year increase since July 2005.

That conflicts with a report released Monday by the Arizona Regional Multiple Listing Service indicating a 12 percent decrease in home sales in the same period.

The reason is Butler’s report does not differentiate between "trustee sales," in which banks take over properties from borrowers in default, and routine home resales.

More than one-third of the sales reported by Butler for April, or 2,025 of the 5,585 total, were trustee sales.

When real-estate consultant Scott Smith saw Butler’s latest report, Smith said he knew something was wrong with the numbers. Smith, who owns a real-estate services firm and tracks area home sales "on a daily basis," said Butler’s April sales figures were simply too high.

"After checking the data several times . . . there is no doubt that Mr. Butler made a big mistake," Smith said.

Butler responded to Smith’s comments:

Butler said he agrees that trustee sales should not be lumped in with routine resales and would be reported separately from now on.

The market has changed so rapidly, he said, that the methodology he once relied on for accurate sales data suddenly has become obsolete.

Until recently, Butler said, trustee sales represented a very small portion of overall sales activity and often involved an actual sale, such as at a foreclosure auction, which is why he has always included them.

But as the foreclosure rate began to climb in late 2007, more and more cases involved lenders simply assuming ownership of the home, still considered a trustee sale and still included in Butler’s reports.

So what does this mean for home sales?

Butler’s revised figure would be 3,565 sales, he said.

That changes sales from +15% YOY to -27% YOY in April.

What was not mentioned in the article was how this can potentially affect the median price.  Property generally goes back to the bank at or near the amount owed.  It is likely then that the median price has also been overstated by the transfer of such a significant number of properties at more than their market value.

Butler is not the only one to have had problems with his methodology in a declining market.  RealtyTrac’s numbers have been called into question for the same issue, and the New Home Sales sales numbers released by the Commerce Department are overstated as cancellations are not figured in.  My objection to Butler’s report isn’t so much the sales number, as this statement from his report:

For the first time since July 2005, the local resale housing market has posted year-over-year improvement. April had 5,585 recorded sales in contrast to 4,855 sales for a year ago and 4,335 sales in March 2008. Given the improvement, the basic question is whether this is the first sign of the much anticipated recovery or merely a blip in a continuing weak market.

Obviously these numbers didn’t demonstrate "improvement" or "recovery" To imply that they were is to go beyond "spin" to "misleading".

As indicated by Butler, changing conditions can mean that old methodology no longer adequately reflects current realities.  Accounting for this can be problematic.  Do you change methodology and lose your "apples to apples" comparison, or do you risk further skewing your comparisons by changing methodology?  If Butler chose to keep the old methodology, that’s reasonable, but he should have done what he said he would do in the future- indicate how many sales were trustee sales.  To imply that these numbers might have been a result of market improvement when he knew otherwise is academically dishonest.

Kudos to Smith and the Republic for catching this one.  Here’s hoping we see better from Butler in the future.