Wall Street Journal oversells a Manhattan 4 bedroom “boom”

[UPDATE: note the way The Journal acknowledged the error and then edited to make the article even less coherent: December 19, Manhattan Loft Guy, 1 – Wall Street Journal, 0, as WSJ posts 4 bedroom correction]

(mis) adventures in arithmetic
When I saw the Wall Street Journal-sourced Curbed item about a ‘trend’ of large Manhattan families snubbing the suburbs to remain in Manhattan I had three immediate thoughts: 

1. I’ve read this Manhattan residential real estate story before;  
   
2. “nearly doubl[ing]” the number of 4-bedroom sales in Manhattan from 136 in 2005 to 270 in 2009 is true, but a tiny enough percentage of overall Manhattan sales to (yet) be a reliable trend; and
   
3. there’s got to be something very wrong with the Census data purporting to show that the market share of 4-bedroom New York City apartments went from 1% in 2005 to 12% in 2008.

Having now looked at the U.S. Census Bureau’s New York City Housing and Vacancy Surveys for 2008 and 2005, I can report that the Wall Street Journal article is very very wrong about the data there.

a thought experiment
Before we go on, now would be a good time to re-read the article by Melanie Lefkowitz, More at Home in 4 Bedrooms, but when you do, skip over this troublesome paragraph:

According to the U.S. Census’s New York City Housing and Vacancy Survey, apartments with four or more bedrooms represented more than 12% of owner-occupied units in New York City in 2008, up from 1% in 2005.

What you are left with is a trend report based on one consumer (who happens to be a real estate agent), two developers, and The Miller. That would be three sources of anecdotes and one source of data, with that single data point about the market share of 4-bedroom sales. Much as I love The Miller, that is a pretty thin article — if you read it without that Census data — and one that you have read before.

Here is the nugget relying on The Miller:

The number of four-bedroom apartments sold in Manhattan has nearly doubled over the past five years, to 270 in 2009 from 136 in 2005, according to the appraisal firm Miller Samuel Inc. Sales of large apartments began rising in the 1990s, only to decline sharply after the economic crash of 2001. But in the past five years they have resumed their upward trend, with sales of apartments with four or more bedrooms rising by 29% in the third quarter from the second quarter.

I think that this is over-reliance on a single year’s sales to anchor a trend, and I suspect the explanation for the spike may be that these few very large condo developments are selling out in 2009 and 2010. If a large portion of the large apartments were offered for sale in 2009 and 2010 as new developments, it is likely that the market share of large apartments will regress to a mean next year when these new developments are absorbed.

While there is a story there (one that the New York Times reported not too long ago, in fact), it is properly focused on developer reaction to the willingness of very wealthy large families to buy or combine apartments into 4 or more bedrooms. How familiar does this the money quote from the June 25 New York Times Large Apartments Are the Rage in New York City by Vivian Toy, sound?

Large family apartments have always been a small segment of the housing market, but the number has grown as developers have started to build more of them. Within the last year, developers of several new projects have responded to the demand by combining units to produce as many as seven bedrooms, seven and a half bathrooms and close to 6,000 square feet.

(Forgive me, but I can’t help but notice how many of the same people are quoted in the Toy article in June and the Lefkowitz article today, including that same consumer/real estate agent [with still more pictures of her combined apartment!]; but I digress….)

Let me pin down some Miller numbers before moving on to the really bizarre error in the Wall Street Journal article. The full year-by-year market share report by apartment size is in the Miller Samuel 10 Year Manhattan report, 2000 – 2009 (page 5, which is 6 of 61), but here are a few selected years, with my calculated percentages:

Yes, “a lot” more 4+ bedroom apartments sold in 2009 than in any previous year, and The Miller’s chart in the WSJ article says that trend continued in 2010. But if we are talking about this because 150 extra families bought super-large apartments two years in a row, that’s not terribly interesting.

how did they do this??
It would be interesting if there were a complementary highly credible source for very large numbers regarding Manhattan (and city-wide) residential real estate patterns, such as (say) the United States Census Bureau. If there is a Journal editor who insisted that Lefkowitz use some US Census numbers if available, it appears that the editor did not ask her to re-check her arithmetic. Once again, here is the WSJ money quote about the Census data:

According to the U.S. Census’s New York City Housing and Vacancy Survey, apartments with four or more bedrooms represented more than 12% of owner-occupied units in New York City in 2008, up from 1% in 2005.

Uhhh … no.

The percentage of 4+ bedroom owner-occupied housing units in New York City (not to be confused with NYC “apartments”, or especially with “Manhattan apartments”) was 12.67% in the 2008 survey (129,120 out of 1,019,345), but it was 12.28% in the 2005 survey (124,090 out of 1,010,370). You don’t have to trust Manhattan Loft Guy on this; here are the reconstructed tables (the original source links to the full tables are in each “Table 29”):

2008 NEW YORK CITY HOUSING AND VACANCY SURVEY
ALL OCCUPIED HOUSING UNITS BY TENURE     
SERIES IB – TABLE 29

2005 NEW YORK CITY HOUSING AND VACANCY SURVEY
ALL OCCUPIED HOUSING UNITS BY TENURE
SERIES IB – TABLE 29

is it newsworthy that the real trend involves outer boroughs?
My strong guess is that the large number of “owner occupied” 4+ bedroom units are outside Manhattan, in the thousands upon thousands of single-family and multi-family houses that are (especially) in Queens and Brooklyn. If you agree that using the data for coops and condos is a fair proxy for “apartments”, note the tiny percentages that result for 4+ bedroom units on a city-wide basis:

If there is a trend in these Census numbers, it is backwards: the percentage of 4+ bedroom coops and condos in the overall housing stock of the city declined (slightly) from 2005 to 2008. Not exactly consistent with the thrust of the Lefkowitz article. (Note: I am not saying this is meaningful; I am just playing with the same data the WSJ used.)

fun fact: true trend (to me)
By the way, I happen to believe that the trend is true: more families that formerly would have moved to the ‘burbs as any kids reached school age are staying in Manhattan (especially, families with multiple children). I think this has been going on at least since our kids were school age, which would be the late 1980s. (This one of my all-time favorite Manhattan Loft Guy posts has a bunch of links to Tribeca population trends, and the influx of “three-foot people”: my March 12, Quote For The Day, 2000 edition.)

I just think that it is hard to prove this trend. I especially think that the use of Census data in today’s Wall Street Journal article is a very bad way to prove this trend.

© Sandy Mattingly 2010