Every now and again, I perform a refresh on a project I commenced shortly after moving to South Florida and becoming a realtor. Oftentimes I hear realtors discuss relative property values in terms of the statistic “price per square foot.” Perhaps you have heard such discussions yourself – or even participated in them.

My questions in the process are twofold. First, since literally everybody is using this measure, it must be a pretty good predictor of value… Isn’t it?

And second, if it’s really not that great a predictor, is it possible to construct something better?

As always, we start with the data. I looked at all single family home sales in Wilton Manors in the 12 months ending April 30, 2019. Then I analyzed the relationship between sale price and square footage.

The result is the chart included here. The dots are all the actual values. You can see there is somewhat of an upward trend with the dots, which is what you would expect. A larger home would generally cost more to buy.

I then (well my software truth be told) put in a line to show the best fit among all the dots. That’s the thick red line here.

You should notice something. Yes, the line slopes up and to the right, just like the dots. However, there is a pretty wide variation in the distances of some of the dots from that best-fit line. And what that means is very simple – price per square foot is not a “one size fits all” measure. It is only a ballpark number and should always be used and presented as such.

Of course, this doesn’t happen, and I probably won’t change things in everyday practice with just one SFGN column. But I can still try. (I am correct, after all.)

That red line has an R-squared (a statistical measure of goodness of fit) of 54 percent. (100 percent would be perfect.)

To wrap up this week’s episode: Well, if you were going to use just one measure to make an assessment of value, price per foot would be the best one to use. But it’s really not all that great. So the follow-up question is: Can we come up with something even better – a calculation that will move that R-squared thingie closer to 100 percent? And what factors are most important?

Tune in next time for the answer.

*James Oaksun, Florida's Real Estate Geek(SM), is Broker-Owner of New Realty Concepts in Oakland Park. In addition to having degrees from Dartmouth and Cornell, he is a Graduate of the Realtor Institute (GRI).*