Multiple Regression Analysis in the CMA

Real Estate Agent with Raving Real Estate

Any fellow economist agents here?  I've always wanted to incorporate MRA (Multiple Regression Analysis) into my CMAs, or at least my valuation studies.  Both for functional use of pricing and also for basic accuracy studies to see how well it might work.

However, I am in a non-disclosure state (sold prices are not disclosed) and of course the agents and appraisers have access to these sold prices but no outside 3rd party does.  This means we would have to create an MRA model manually, particular to our market and then manually add in data points.  Also, many of the data points would not be accounted for on the basic MLS data sheet. 

A good MRA model for real estate might have over 100 variables, including many dummy variables.  So hypothetically, one would have to go to each of these houses and check all these extra features out to create enough variables to make it accurate and give it a good p value, t value, or whatever test you are using to attain statistical viability.


Now to the question:  Does anyone here USE MRA?  If so, how?  How many variables?  Where do you get them?  Any ideas on how this could be successfully utilized for the purposes of pricing?  I ask that last one as I KNOW it has been used for economic studies.  For example, one of my professors when I was in school was working on a model for at least a year to try to build a model that valued trees.  I was always interested to read the results and find out where the point of diminishing returns happened (what number of trees per square foot).  And which trees were a liability on average, which were an asset and so on. 

This is just one example, but it reveals how interesting MRA could make a CMA and how much more in depth it could get.  Call me an egg head, call me a silly economist with a real estate license, or call me and tell me what I yearn to know:  How can one use MRA in real estate brokerage with limited data and especially in a non-disclosure state?

Comments (1)


Brendan, I have many of the same questions.  My 2 cents: I think the best way to use MRA is as a marketing tool.  RE markets are inefficent and that creates the opportunity.  If it was as easy as building a model and filling in blanks anybody could do it.  The differentiation between you and your competition will be in the quality of the assumptions you input into your MRA to increase the accuracy.  Over a period of time you can build a tract record that becomes a powerful marketing tool. 

Sep 26, 2008 03:09 AM