riffing with Toy + NY Times about higher prices “sprinkled” through Manhattan, including in the loft market (Part 1)

a different approach to outliers
I daresay that the Sunday real estate section lead article in yesterday’s New York Times by Vivian Toy, In New York, a Sprinkling of Higher Prices, will get a lot of play on the inter tubes. It seems that she and I have been reading from the same page, though perhaps at a different angle. V-Toy finds “micro-markets” (implying that there is a category of strong performers), while Manhattan Loft Guy finds outliers (implying that there is ‘noise’ in any set of market data).

On the one hand, I love that she takes a data-driven approach, looking at specific sales in specific buildings (including, as we shall see, one loft building very familiar to regular readers of Manhattan Loft Guy). And I am very sympathetic to any attempts to learn by trying to draw useful conclusions from messy market data.

On the other hand, she quotes agents saying the kind of things that … well … agents do, in support of this generalization of hers:

… brokers say, the types of apartments that had been getting 2007 and 2008 prices will probably continue to do so.

Nice try, but I doubt very much that there are “types of apartments” that out-perform The Market [edit: in this way]. Nor do I think that she (or the quotable members of the local Manhattan branch of the Real Estate Industrial Complex) has identified any such “types of apartments”.

I actually agree that there has been “a sprinkling” of higher prices among filed deeds, some of which I have hit in man-bites-dog (or similar) blog posts. I even agree that “sprinkling” is a very good word for that. But I like “sprinkling” because it implies a random distribution, rather than a systematic pattern based on “type”.

V-Toy poses the issue thusly in her lede:

BEFORE the financial markets’ most recent drubbing, New York City’s real estate prices had been flat for the better part of a year. But over the spring and summer, prices in certain pockets of property sprinkled around Manhattan and Brooklyn had rebounded to or beyond pre-recession levels.

These micromarkets could turn up along a particular avenue or even in a specific building, and they tended to be in what brokers describe as prime locations, in neighborhoods like Greenwich Village and Chelsea in Manhattan, and Brooklyn Heights and Park Slope in Brooklyn.

They fell into a few general categories, namely: apartments in move-in condition; family-size apartments with three or more bedrooms; apartments with unusual features like helicopter-level views of Central Park; and almost-new condos in their first resale. Having one of these traits did not guarantee a record-setting price, but apartments achieving that milestone tended to have one or more of them.

Let me emphasize her very important qualifier, one that is both eminently fair and extremely powerful, in that it undercuts the thrust that there are such “micromarkets”:

Having one of these traits did not guarantee a record-setting price, but …

I mean, if the “micromarkets” theory really worked, you’d be able to predict some future (strong) price activity, right?

one difference between blogging and Journalism
Readers know that I blog essentially every day about something that just happened in The (Manhattan loft) Market. I very rarely have a point other than ‘this was interesting to me because …’. And I only occasionally look back and see how the aggregation of These Interesting Things might itself be interesting (note to self: would you please do that more often?). But the day-to-day of Interesting Things goes on, and on, and …. In contrast, Real Journalists write (it seems to me) to explain something more interesting in a more general way.

Point being: V-Toy went looking for some special micro-markets and found some data that sticks out enough to (possibly) be part of a micro-market and found some agents willing to explain that these (apparent) micro-markets actually mean something, including that they exist.

another difference
With V-Toy and the New York Times having made that contribution, they move on to other Journalistic things. While we in the blogosphere get to chew on what she left behind….

Here’s what I think: we love to have explanations, but sometimes people strain to explain things that are inexplicable. And individual data points in a market are often just that: individual data points. While The Market is ‘a trend’ or an average or mean. I happen to know a little about two of the buildings that V-Toy uses to support her premise, since they are Manhattan loft buildings I have blogged about; I also happen to think that the data from these two examples do not meaningfully support her premise.

Alas, sometimes RJs are … wrong. I share V-Toy’s interest in apparent market outliers, and admire her … courage … in trying to find a unifying theory. Let’s get down to cases.

a Manhattan Loft Guy fave
I don’t know if I would have launched like this into the metaphysics of the market had V-Toy not relied on two specific Manhattan loft buildings to support her overall thesis. (Patience, Prudence: I will get to the other one in a bit.) But how could I resist when she offers this as an example of one of her (supposed) micro-markets?

[there is a] contract for a two-bedroom at the Lion’s Head Condominium, at Avenue of the Americas and 19th Street, for $2.15 million. The apartment originally sold in 2007 for $2 million.

Does that contract really evidence a micro-market that “had rebounded to or beyond pre-recession levels”? If so, it is pretty thin evidence.

StreetEasy thinks that the loft in question (obviously, this one) closed last week, with a deed pending (at the full ask, per V-Toy). Yes, it sold last at $2mm, on a contract signed at the end of 2006, but if that one sale is part of a micro-market getting “2007 or 2008 prices”, you’d think that other sales activity in the same Manhattan loft building (121 West 19 Street) would line up the same way. Otherwise, this one contract is just … errr … this one contract.

As an aside, and not to be rude, note that this loft has only one of the ‘traits’ Toy identified as ‘tending’ but not ‘guaranteeing’ a ‘record-setting’ price (“namely: apartments in move-in condition; family-size apartments with three or more bedrooms; apartments with unusual features like helicopter-level views of Central Park; and almost-new condos in their first resale”): it is in mint condition (how special is that??), but only has 2 bedrooms (convertible 3, but so are nearly all lofts around 2,000 sq ft), has no unusual features, and has had an intervening sale since the first offering.

Don’t get me wrong: if this thing in fact cleared at $2.15mm that is a very strong sale (+57% over the 2006 initial offering sale), one that might well qualify for a man-bites-dog blog post because it is also +7.5% over that intervening sale on February 13, 2007 (4 quarters before The Peak). It is just that there is no such trend at the Lion’s Head that is apparent to me.

It will take me a while to show that, so bear with me ….

As it happens, there is a Manhattan Loft Guy post for three of the four prior sales in this building in 2011, none of which found a record-setting micro-market in this 2006 new condo loft conversion in the (now hot??) blocks just off 6th Avenue in the 20s.

my apologies, but we must revisit the Lion’s Head (again)
My basic point in my July 6, Lion’s Head loft sale hits the number at 26% over 2006, was that there had been perceptible gains in this building since the 2006 initial offerings, but that it appeared as though most of those gains would have been realized immediately, if only more initial buyers had flipped in 2006.

Skip the long indented excerpts from that post unless you crave the details to support what i just said:

Sometimes The Market provides data that show Manhattan loft longitudinal trends in just a few data points. Note that the only meaningful difference between the sales histories of two lofts that sold recently at the Lion’s Head, 121 West 19 Street, is in the number of owners.

The “1,364 sq ft” Manhattan loft #9F closed on June 9 at the ask of $1.475mm, reflecting an appreciation of 26% since it was purchased from the sponsor in the frothy days five years ago, on June 12, 2006 at $1,170,987. The seller held for 60 months to get that 26% (gross) return. I hit the last loft to sell in this condo conversion in my June 1, flipped early, Lion’s Head loft later gains only 4.8% since 2006, a story that has a similar outline but one too many owners to permit everyone to smile.

That was the “1,134 sq ft” loft #8D, which showed an appreciation of 30% from the original sponsor sale ($1,008,067 on May 8, 2006) to selling on May 11 at $1.31mm. The sellers of #9F 4 weeks ago and of #8D 8 weeks ago were almost exactly equally patient, holding through the Peak and nuclear winter. But the #8D seller on May 11 had only owned the loft since July 26, 2006 (58 months), and that made all the difference, as that seller realized only 4.8% because the owner from May 8 to July 26, 2006 took the rest. Go to that post for the story of that loft, but keep in mind that difference.


I am pretty sure I went nearly four years without hitting lofts at 121 West 19 Street, but have now hit them five times in 2011. I will try to avoid coming back until there is something different to say.

V-Toy’s example of the to-be-filed deed at $2.15mm would be a sale at $1,080/ft, but should that be something to brag about if it had not sold for (only) $2mm ($1,005/ft) in February 2007? Not based on the other 2011 sales mentioned in my July 6 post. The featured sale in that post was #9F, which closed on June 9 at the ask of $1.475mm, or $1,081/ft, and which I compared in that post to #8D, which closed on May 11 at $1.31mm, or $1,155/ft. Before you say that that sale might prove V-Toy’s recent-sprinkling-of-higher-prices-by-type thesis, note that the #8D seller in 2011 was the flip-ee in 2006 from the original 2006 buyer, and realized only a 4.8% gain from selling 3 months ago after holding for 5 years.

It is hard to argue that #2C is special (assuming it really did close recently at $1,080/ft, up 7.5% since February 2007), when #8D at $1,155/ft was up only 4.8% since May 2006.

Bottom line: I don’t see how you can use this one contract at the Lion’s Head to support V-Toy’s version of Sprinkling, once you put that contract in the context of the building’s sales history.

is this loft building (72 Mercer Street) evidence of a trend?
The other loft building V-Toy hit (remember: 2 of her specific examples are loft buildings!!) is less of a Manhattan Loft Guy fave, but I have hit this relatively small new building twice in the last two years. Here is how V-Toy used the 2008 new development 72 Mercer Street:

What defines a building or an apartment as special can vary by neighborhood. In SoHo, where smaller buildings predominate, it could be a doorman or outdoor space. In Greenwich Village, where prewar buildings are common, it could be a signature building of midcentury design.

Two months ago Richard Orenstein, an executive vice president of Halstead, sold a two-bedroom loft at 72 Mercer Street in SoHo a week after listing it. The full-floor apartment had sold for $3 million in early 2008, and it resold at the full asking price of $3.995 million.

“Things in general take longer to sell now, but special apartments sell for special prices,” Mr. Orenstein said. The SoHo condo stood out because it had a large terrace and also was in a doorman building.

I hit the very sale V-Toy featured in a blog post 3 weeks ago, August 2, velocity in the Manhattan loft market / 72 Mercer Street sale was not the fastest. where my point was more about recent velocity than the scale of appreciation since that initial offering. That the featured loft (#2W) sold on June 30, 2011 at $3.995mm and on January 3, 2008 at (only) $3mm is certainly noteworthy, but I am not at all sure that it constitutes support for a sprinkling theory based on type. (V-Toy suggests that this 33% gain from The Peak was due to #2W being a special type, a Soho doorman condo with a large terrace.)

Building history lends a little support to the main V-Toy thesis (if not by type), as another Soho doorman condo loft with a large terrace sale (#4E) that I hit in my February 6, 2010, new development 72 Mercer holds its own from peak, re-sold in January 2010 essentially flat with its original new offering price in April 2008 — with #2W two months ago a full 25% above #4E 18 months ago.

So perhaps the strong sale of #2W compared to itself in 2008 and to #4E in January 2010 (partially) supports (some of) the V-Toy thesis that some prices are up since The Peak. But I fear that the claimed ‘type’ (Soho doorman condo with a large terrace) is too small a data set to permit a proper test of the hypothesis. The implication is that Tribeca lofts with doorman and outdoor space are not the right ‘type’ to compare to 72 Mercer, or that Soho lofts with doorman but without outdoor space (or vice versa in Soho) are also not the right ‘type’ to compare to 72 Mercer.

paging Bill James
Maybe yes, maybe no. It sounds to me like correlation instead of causation. Sort of like those idiot baseball stats you sometimes see in a game broadcast, such as a pitcher being 3-0 with a low ERA and favorable strikeout-to-walk ratio in Sunday afternoon starts, compared to being 6-8 with a higher ERA and worse K/BB ratio on other days. The data set is just too small to conclude that the Sunday afternoon correlation ‘showing’ Sunday afternoon ‘success’ has something to do with it being Sunday, instead of showing that performance in any three wins by a 9-8 pitcher is going to have a lower ERA and associated pretty numbers than the rest of the guy’s resume.

Point being: you show me an exceptional result, and I can conjure something about the sale that is ‘different’, but that difference may or may not be related to why the sale price was exceptional. (Most pitchers are not that Sabbath directed.)

I don’t know enough about the other “apartment” buildings that V-Toy used in support of her thesis to comment much (you’re welcome! but I won’t stop), but it should be clear that I think she is straining to label a trend that is hard to describe, if it exists at all. I can also cherry pick some relevant data (which may lack context, unfairly to V-Toy, but I tried to be fair). As you can guess, I am not persuaded.

miscellaneous data points can do that to you
Maybe V-Toy should not have taken some of her agent quotes at face value.

That estate sale at Butterfield House (where the “price works out to about $1,200 per square foot, which [the agent] described as exceeding the $1,100 per square foot that was the norm in 2007 and 2008”), for example. Neither our data-base nor StreetEasy quotes a size for that unit (#10F), but the floor plan and room dimensions for #10F match very very closely with those of #3F, which is said to be “1,400 sq ft” (floor plans here and here). If #10F is also “1,400 sq ft”, that estate sale was really at $1,019/ft.

StreetEasy shows only two public sales in the building in all of 2007 and 2008, so it is hard to say what the ‘norm’ was in those years. One was #2J at $2.6mm, which we have at “2,200 sq ft”, or $1,181 sq ft; the other was #2K at $1,123/ft. Both public sales in 2007-08, in other words, were on much lower floors than the exemplar estate sale in 2011, and were $162/ft (16%) and $104/ft (10%) higher.

Non-public sales at The Butterfield in 2007-08 for which I have size data are #5G at $1.877mm for “1,700 sq ft” ($1,110/ft), #7D at $2.2mm for “1,250 sq ft” ($1,760/ft), and #7B at $1.8mm for “1,550 sq ft” ($1,161/ft). Again, all lower floors than #10F, with a very broad range from 9% to 73% higher. Not the sort of contextual data I would build a case for #10F as a revelatory sale around.

How does the City Spire story fit into V-Toy’s narrative? She said a seller “listed her two-bedroom apartment at the City Spire, a 72-story tower built on West 56th Street in the 1980s, at the beginning of the year. There was little activity at first, but she sold the place earlier this month to Italian investors for $1.625 million. She bought it for $1.48 million in 2007 and had the kitchen renovated.” The unit is #5806, and it sold at $1,366/ft with a new kitchen after selling (to the listing agent!!) for $1,246/ft in December 2007.

Other 2-bedroom 2-bath activity on high floors at the City Spire in 2007-08 suggest that the #5806 2007-buyer-turned-2011-seller may have gotten a deal in 2007, not that she sold last month (at $1,366/ft) at a huge premium to 2007-08 building trend values. #5802 sold at $1,504/ft two months before the #5806 sale. #6102 sold at $1,608/ft in September 2007. The smaller #6005 closed at $1,582/ft in March 2008. Both the larger #6304 (cleared at $2,085/ft) and the larger #5101 (cleared at $2,006/ft in July 2008) are … unhelpful.

With this building sales history as context, the exemplar July 2011 sale at the City Spire at $1,366/ft is at least 10% (and up to 53%) lower than sales of comparable units from 2007-08. One more time: how does the City Spire story fit into V-Toy’s narrative that some ‘types’ of apartments sell at a premium to pre-recession prices?

I just discovered (a) that there is a limit to how long a post can be, and (b) that this one exceeds it; CONTINUED here

© Sandy Ma

Tagged with: , , , , , , , , ,

Leave a Reply