Domus Analytics

New Construction and 0 DOM

April 19, 2021 08:55 PM Comment(s) By Amy

An adventure with New Construction and DOM

So we got this email...

It all started with a client email, as these things tend to do.  Bob Bemis, the MLS Director of the Park City Board of REALTORS┬«, sent out one of his well-done weekly market reports.  Bob has a long MLS history, and he's a big data nerd like us.  His report was titled "The Law of Unintended Consequences", and he examined how New Construction contracts were impacting Days on Market (DOM).  Coincidentally, we'd been talking a lot about DOM internally, as you'll see in the next few blogs, so we reached out to chat.  Here's a summary of what Bob wrote:


Part of the Clear Cooperation policy adopted last year requires developments to report their Pending listings when they go under contract, even if the specific units being purchased were not listed.  (Developers typically put in just a sampling of models as listings but have more inventory to sell than what appears in the MLS.)  Thanks to the developers that are doing just that, we are seeing a growing number of listings being Added and Pended on the same day.  But the DOM and CDOM averages are thrown way off because of these entries.  In just the past week, 23% of all Pendings were 0 or 1 DOM.  Agents should look at those averages closely, and look at the underlying data when doing a market analysis where such new developments are prevalent.  DOM and CDOM are likely very inaccurate as a result.


Based on our discussions and Bob's additional insight, we said "What do you think about excluding New Construction 0 DOM listings from DOM calculations?"  While it turned out to be a simple adjustment, the analysis showed us a lot more than just 0 DOM listings making their impact on the calculations.


New Construction and DOM

We have a large amount of data from groups across the country.  While this issue came up first in Park City, we did our analysis by checking a large number of New Construction listings across all sorts of different markets to understand what was going on with DOM overall.  We saw three unique things:

  • Most of the New Constructions listings were, in fact, 0 or 1 DOM.
  • Model homes would eventually sell, resulting in a very large DOM, often a year or longer.  Yep, that full-furnished corner unit with beautiful staging and every amenity possible finally sold after 2 1/2 years.  That data point is just totally different from a standard listing!  It just doesn't represent the traditional DOM.
  • Spec and custom home builders can choose to list their property on the MLS any time they want.  Some list when breaking ground, others after framing, and some wait until a CO.  We couldn't find any consistency on the listings.  This lack of consistency introduces a very high variance on the DOM value.  In analyzing many New Construction listings, we found the data was all over the place and very noisy, and didn't represent a consistent methodology.  Or, as we learned in school: "Garbage In, Garbage Out."
The conclusion here was that most every listing that was tagged as New Construction really didn't have DOM data the way we think about DOM data.  There was so much internal variance that it just didn't make sense to include it in the calculation.

Conclusion

To make a long story short (something Ed rarely does...), Bob thought removing New Construction from DOM was a great idea.  So we implemented it in his data.  The result, as expected, was a stabilization of the DOM results.  By removing all the noise, the calculations now feel more aligned to the market experience.  One of the capabilities of the Domus online dashboards, unlike PDF reports, is that we also recast all history to reflect the new calculations.  So the month-over-month, year-over-year, and the long time-series charts all make sense together without showing big jumps or gaps due to a new methodology.

Going forward, this is our new standard for handing Days on Market with New Construction.  Especially for DOM, New Construction is just different than resale.  The goal of stats is to accurately measure and communicate the behavior of the market.  A traditional resale DOM calculation doesn't make sense for New Construction, which Clear Cooperation exposed for us quite clearly, and what people are expecting when they think of DOM.  As the policy, market behavior, and data change underneath us, we continue to change right along with everything.