Yesterday, the folks over at online data tracker RealtyTrac gave me an exclusive on their monthly foreclosure numbers, because the numbers were just SO out of hand (and because I positively begged them). Foreclosure activity nationwide is up 19% from April to May and up 90% from May 2006 to May 2007. That’s big, and let me tell you, all the media outlets jumped on the bandwagon. ABC World News Tonight even led with the story.

But out in Newark, NJ, a reporter at the Newark Star Ledger newspaper was throwing a big wrench in the data. The paper reported yesterday that RealtyTrac is revising its formula, and that could, “dramatically decrease the number of foreclosures in any given state, according to RealtyTrac.”

So all the bosses at NBC Nightly News, who are about to put me on the big show to report these numbers that I got, start filling my Blackberry with worries. Are these numbers from Irvine, California-based RealtyTrac to be believed? Are they in fact hyping the numbers because part of their services include listing houses for sale, and buyers looking to get a deal on properties in distress--pay $49.95 a month to subscribe to RealtyTrac?

Here’s the deal. RealtyTrac gets its total number of foreclosures--176,137 in May--by adding up “default notices, auction sale notices and bank repossessions,” according to their press release. Does that mean that some properties get counted multiple times? You bet it does.

Were there 176,137 foreclosures in May? Nope. A house can go through several permutations: first the default, then the bank repossession, then the auction, and in some states this can go on for many months with many different filings; therefore, many of these homes are counted multiple times. So do I trust that 176 thousand number? Nope. But I do trust the percentage increases.

If RealtyTrac has been counting the homes in the same way for many years, then the monthly and yearly increases are in fact apples to apples, and are in fact relevant. Are they perfect? Of course not, but they are credible in showing real movement in foreclosure activity. Make sense?

As for the change that RealtyTrac is implementing, well I furiously emailed Rick Sharga, the Veep of marketing at the company, worried that they were suddenly going to announce their previous data were bunk. He wrote back, “Just to put your mind further at ease, we’re NOT changing the way we get the data, or count the filings. We’re just digging deeper into the data we already get in order to provide one more metric (# of individual homes affected)." In other words, they are reacting to the criticism and offering a new number that will show only the specific number of homes involved, and that will clearly be less than 176,137.

Does this clarify the issue at all? Hope so. I still stand by our reporting the RealtyTrac numbers because an increase of 90% even in “foreclosure activity” is news. Tomorrow we get the quarterly delinquency data from the Mortgage Bankers Association, which by the way widely criticizes RealtyTrac, so look for it around 10 am.

We’ve been saying that the subprime ripples were only just beginning, and this proves the premise.

Questions?  Comments?