Mothra v. Godzilla, or the epic battle over Manhattan sales volume reports
Terra takes on The Matrix
I commented last Friday about the discrepancy between the number of coop and condo sales in Manhattan in the first quarter of 2008 as reported by 3 major firms (April 11, Q1 Manhattan market reports highlight high prices + major caveats), and that the holding company for Halstead and Brown Harris was "throwing some smack" about 600 sales "missing" from the Miller Samuel report for PruDE.
The title of The Miller’s response to this tempest (on his Matrix blog yesterday) sums up his attitude about it: [Grandstanding] Rational Markets, Irrational Public Relations. Let’s just say that The Miller is pissed. Here is what The Miller said to the NY Times by way of grudging defense of his report (which was paraphrased in Josh Barbanel’s column yesterday ):
I have consistently compiled data from both public record and reliable industry sources using the same methodology since 1994. These numbers represent the overall Manhattan market and reporting discrepancies could be for a number of reasons, including timing of when the data was polled and which neighborhoods are included in the report. I stand by my numbers and the methodology used to compile them.
too much Inside Baseball?
If the topic of the minutiae in reporting Manhattan real estate data is too geeky, too detailed, then stop reading here. (I don’t know the provenance of the term Inside Baseball, but it is used to separate the geeks and Sabre-metricians who measure and count everything from those who just enjoy the game, without slide rules or their modern equivalents; noting wrong with either approach, they’re just different.) I am about to launch into a long but meandering and too-confusing analysis, which is the best I can do at blog-speed today. (Before we get into a Hillary-as-RockyI-or-as-RockyIII thing, I have no idea if Godzilla beat Mothra in the movie, or if there were sequels; my point is that we are all getting crushed.)
Anybody who has paid any attention to the periodic reports from the major real estate firms about residential sales activity knows that they are always different. While this is not news in the sense of being new, it is news in the sense that someone should figure out and explain why (though we are so far from that world). Barbanel describes (and quotes) from Greg Heym’s ‘internal’ Halstead and BHS memo, with Heym saying:
[the Miller Samuel] report was incorrect and “missing almost 600 sales, which were available for anyone to view” in the city’s online records.
As The Miller says on his blog, he does not describe other firm’s reports as "incorrect" when they differ from his reported data. He also says that "there are many other more important issues to deal with these days" (in his comments), but I happen to agree with Heym and Barbanel that this data point is very important for market psychology: were there 34.3% fewer buyers closing Q108 than Q107 as The Miller reports, or only 1% fewer as Terra reports?
up close to the Madding Crowd
I gather that out in America local Realtor® boards report actual sales data on periodic real time bases without much difficulty or controversy. It is still baffling to me that we can’t do this in Manhattan, though it is clear that we cannot do this in Manhattan.
The Miller has his methodology and I have no reason to doubt that he has used the same method since 1994 — long before coop sales were publicly reported. One of the great things about his reports has always been that they are consistent, and reported in a track-able historical manner. Indeed, on his firm’s website, one can slice and dice a lot of different data sets from past years. He has been as transparent as anyone — indeed, more transparent than anyone — for years.
That is not to say that Terra’s reports are less valid, but to acknowledge that they are different and that it is not so easy to compare their data over time.
Take the 2,857 Q108 sales reported by Terra, a result described as in Heym’s memo as "available for anyone to view" in City records (meow!). Terra says that is off 1% year-over-year, so the Q107 base line was about 2,571 (2,857 – 286 = 2,571). Stay with me for a minute, please….
That current 2,857 number seems to include all of Manhattan, as the constant footnote in the Q108 Halstead report says Terra numbers have included Northern Manhattan since Q307, and that current comparative data has been adjusted to reflect this change but that the reports from earlier quarters have not been changed. However, the Q107 Halstead report (i.e., before Northern Manhattan was included) reports 2,657 sales (see p7 of 8 in the report). This year-old number without Northern Manhattan (2,657) is larger than the entire Manhattan number implied for that same year-old number as the base line for measuring change from this year’s quarter (2,857 less 1%).
If you follow that, you’ll see that the data comparison — on Terra’s terms — breaks down. If you can’t follow that, it’s my fault.
Let’s try this another way…. Remember that Heym says his number (2,857) is "available for anyone to view" in City records. Barbanel in the Times went to City records. He found that "a review of closing documents filed by last Wednesday, nine days after the end of the quarter, showed that the number of sales was roughly flat compared with the same quarter a year ago." That statement is (1) seemingly consistent with Terra’s thrust, and (2) different from The Miller’s conclusion about the drop in volume. But Barbanel went on to report
"They [Q108 sales] fell, but by less than 1 percent, or a decline of 11 sales out of nearly 3,500 sales reported in April 2007."
1% sounds like Terra, but nearly 3,500 sales sounds like The Miller. Did Terra under count Q107 by 800+ sales?? According to Barbanel’s review of City records, yes. Evidently, it is harder to simply "count" on ACRIS than Terra implies.
playing nice together
The sad fact is that it ridiculously difficult to measure Manhattan real estate and that apparently professional people come up with different answers. Personally, I like to use The Miller analysis for trends because he is the most transparent, has been doing it the longest, and presents the easiest stuff to work with. I can’t say if it makes a difference that he is an appraiser, while Heym works for a holding company that owns 2 residential real estate brokerage firms.
The sadder fact is that a critically interesting question (was sales volume flat or did it drop a lot?) has different answers depending on who you talk to. We will have to wait for more data to get a handle on this, even though that future data is guaranteed to differ in different reports.
© Sandy Mattingly 2008