how many Days on the Market? no simple math in the Manhattan loft market

fuzzy math makes things less clear
It seems as though there should be a simple answer to the question “how long has this loft been on the market?” But not in this town.
Rumor has it that out there in America there are standard definitions for “Days On Market” and that this information is broadly available to real estate licensees (and perhaps to buyers). As a result, in most real estate markets there are wonderful and historical data about how long it takes something to sell, which can be very useful macro data.
Manhattan data are weak, skewed and suspect
Not so in Manhattan, for at least two reasons. First, what data exist are not widely shared (about when a loft or apartment was first offered for sale, when it went into contract and when it sold). Second, the time lag between a binding contract and eventual closing ranges anywhere from easily 90 days for a coop apartment purchase to as long as 150 days if the Board approval process is difficult, and can be anywhere from 30 to 60 days for a simple condo purchase. This post-contact skew makes it very hard to compare the data we do have.
And there is a third problem – agents can manipulate the data. Or – just make mistakes.
when does 92 = 150? (should be ‘never’)
The latest Halstead Q3 report illustrated the problem of standardization. Halstead says the average Days On Market for Manhattan coops and condos was 92 days, compared to 68 days a year ago. Halstead does not define Days On Market (page 3 of the report) except to exclude new development (which makes sense) and units listed over 9 months (which does not make sense).
Miller Samuel, on the other hand, has been reporting Days On Market for many years. They use “from last list date” to – presumably – closing, which the press reports as 150 days for the same quarter that Halstead counted as 92 days on market. At that magnitude of difference, that’s a difference in definition not a rounding error.
different data from different sources
I suspect that this difference has to do with Miller Samuel being an appraiser (with easy access to public information such as when an apartment closes) while Halstead probably uses agent reports of when things go to contract I can’t think of another reason the two reports should be so radically different.
Miller Samuel has the benefit of using a data point with a long history, so it is especially interesting from a macro perspective to see that the 150 days for Q3 06 compares with the “historical low” described by Miller Samuel about five quarters ago of (I recall) 98 Days On Market.
Halstead reported Days On Market for the first time in this last report –though they provide one year’s historical comparison (as noted, they counted 68 Days On Market as rolling 5-quarter low).
Miller Samuel won’t have “official” information about contract dates, but Halstead would through the Real Estate Board of New York for agents self-reporting about their transactions. This is probably the same source that the NY Times uses for its “Residential Sales” section each Thursday and Sunday (“most recent listing to the dales agreement”). But that date can be pretty squishy.
squishy data in the NY Times
199 Bowery #11D was featured in the NY Times Residential Sales 10.8 as selling after 10 weeks on market at full ask BUT it was really listed on Jan 19 at $960k, then dropped on February 17 to $925k, then went into contract on April 25 (the ten weeks being from Feb 17 to April 25), and then closed July 7. In this case, the $35k price change re-set the clock the Times uses by dropping four weeks.
Similarly, NY Times Residential Sales for last Sunday shows that loft #5B at 257 W 17 St sold in 6 weeks at $1,232,500 (off the $1.265mm ask) BUT it actually had been on the market since January 15 (starting at $1.595mm, then reduced on March 2 to $1.395mm, and then finally on June 6 to $1.265mm) before going into contract on June 28 and closing on August 23. So the Times reported that this property took “only” six weeks to sell – which is a time period that does not make sense from any of the price changes. This contract took 24 weeks from the first offering, 16 weeks from the first price change, but only three weeks from the last price change. Since it is usually the listing agent that provides the information to the NY Times, I assume the listing agent just made a mistake. (Cynical people might have another hypothesis.)
Here’s a mistake in the other direction. NY Times Residential Sales for last Sunday reports that loft #4 at 40 White St sold above ask in 15 weeks (at $1.027mm; ask was $995k) BUT it really took only four weeks to go into contract. According to our inter-broker data sharing network it was first offered for sale on May 10 and went into contract on June 6, and closed no August 24. So in this case, the agent must have reported the time between the first offering and the closing date.
what’s the point?
That the data are messy.
That only Miller Samuel uses a standard that permits long historical comparisons.
And that the data are messy.
© Sandy Mattingly 2006
Tagged with: , , , , , ,

Leave a Reply