Manhattan High End Surging, Explaining ACRIS $5M+ Chart
A: Lets check in on the $5M+ market in Manhattan. With a quick check into pending sales for properties with a last ask of $5M+, we can see how strong this segment of the market has been. Now, where these deals close out and how far from peak we are right now is something we will have to wait for. Remember, pending sales measures listings that go from ACTIVE to CONTRACT SIGNED; they have not closed yet! When we check out our ACRIS $5M+ trends, we get a chart that is exposed to the sample size problem that I discuss here often. Let me show it to you again.
Manhattan: ACRIS $5M+ 90day Moving Avg (green line) vs PENDING SALES $5M+ (red line)

First I want you to focus on the green straight line that is stuck on a value of "2" in this chart. This is telling us that, on average, only two sales of $5M+ properties are occurring per day. The ACRIS line chart stops at end of March for a reason --> our ACRIS line charts are based on "sale date", not the filing date. Due to the lag between actual sale date and ACRIS filing date, we were forced to show ACRIS volume trends at a lag.
Looking at the red line in the chart, I can see that right now there is a pool of 150 listings in contract and pending closing. If we were to assume that all these listings will ultimately close over the next 60 days (and that is an aggressive assumption), then it would equal out to an average of 2.5 closings per day. So you see, the current green line in the chart above that shows us the daily 90-day moving average of $5M+ closings is in fact accurate.
The ACRIS $5M+ volume line chart is straight and useless because the sample size of this metric is too small! There is just not enough data coming in to make a worthwhile chart to show you. This is an ongoing challenge for me when talking to brokers about making our systems even more granular. With a real time system for tracking Manhattan real estate, like the one we built, having a large enough sample size to generate worthwhile charts to interpret is very important! I touched on this topic back in February, as we deal with the challenge of adding a 5th input into our submarket search interface: PROPERTY TYPE --> allows you to break down charts further by coop, condo, townhouse, or all.
Adding more granularity to a system that already scrubs the data increases the total combinations of charts with a sample size too small to be worthwhile to interpret. And that is what this site is all about. Guiding you to a submarkets performance without sacrificing quality due to having too small of a sample size of data to chart out. The $5M+ ACRIS chart above is an example of this - as the straight line basically tells you nothing. This reality is making us seriously think about tweaking the price points for ACRIS sales volume charts, perhaps only showing $4M+ market instead of the $5M+ market.
As for the current pace of $5M+ deals going into contract and counted as 'pending' in our systems, I think we are peaking out right now. I would expect it to be very difficult to sustain this kind of pace moving forward, especially as we get into mid June and July. All in all, it was a very active start to 2011 and we should expect a slowdown as we hit the summer.



Posted by datajunkie
Wed May 4th, 2011 10:24 AM
Noah, you can also try a smaller moving average setting. Maybe try a 30 day setting and see if this slice of the market shows more volatility? Keep up the great work!
Posted by urbandigs
Wed May 4th, 2011 11:24 AM
thx dj! Yes, we can definitely try to tweak the moving avg settings before we change anything. Good suggestion! Im more concerned about the inventory charts when we add that 5th input for charting out coops, condos, townhouses or ALL.
I tested it already, there are so many useless chart combinations due to getting too granular, that users will bump into them often and wonder if the system is broken. So we will have to re-engineer the search UI to remove this on the fly and only give users options for data choices when enough data exists to even chart the request out
Posted by leveldata
Wed May 4th, 2011 11:42 AM
Your charts and analysis are always intersting and insightful. Do you ever try to extrapolate and tie your data series to trends in the NY Metro area? For instance what percentage of these high end sales correlates to high end purchases in New Jersey or Connecticut. Very interested in data on trends and drivers of owners that exit the NYC market for suburbs.
Thanks
Posted by urbandigs
Wed May 4th, 2011 12:03 PM
leveldata - thx! hmmm, where would I get similar stats and trends on NJ and CT? I just bought a home in CT, low end though 2nd home kind of place, so Im definitely interested in these real time pending sales trends..Any sites that you like to use?