How to Lie with Real Estate Statistics – Particularly in Greenwich

By Mark Pruner

Douglas Elliman Realty

 Too many years ago when I was taking a statistics class at college we had to read a little book called “How to Lie with Statistics”. It was the most useful book I read at Dartmouth. Its author Darrell Huff shows you how numbers can be manipulated to show what the author wants. It also shows how hard it can be to present numbers in a fair way for multiple audiences.  So this week I thought I’d look at some of the most abused numbers.

Greenwich is a very unique town for real estate only three other towns in Connecticut are comparable; Westport, New Canaan and Darien, all of whom are substantially smaller than Greenwich which makes our statistics particularly unique. The other towns do have the advantage of being more homogenous than Greenwich where Byram, Belle Haven and Banksville attract very different buyers.

While we are larger than the other towns we still only hav 600 – 700 sales per year. As a result as soon as you start looking at a particular market in a particular price range for a particular month you will likely have only a handful sales. So a few sales up or down for two consecutive months can “change” a rising market into a falling market when the reality is that there just enough of sales to say either way.

Here’s my five favorite abused numbers:

  1. Price appreciation

I don’t like to talk about price appreciation in Greenwich, because it is the most abused number out there. People calculate price appreciation in a variety of different ways and get to often much different numbers.  A lot of the national stats compare the average sales price for this month to the average sales price for the same month last year.


In Greenwich that is not a good way to look price appreciation, because our ultra-high end is so high and low end doesn’t have a lot of units. The result is that our “average” jumps around from month to month.

This jumpiness gets extreme when you look at just one area of town. I’ve seen some claims of year over year price appreciation in say Cos Cob of 50% simply because one high end house was sold this year that didn’t sell last year.


The median, the sale in the middle, is a better way to calculate appreciation, but it to is affected by changes in marketplace. When the number of sales at the high-end falls the median falls even as the houses in the middle are actually appreciating giving exactly the opposite result for those houses.

Assessment ratio

The indicator of appreciation that I like is the ratio of the sales price to the assessment. If the price is the same then that ratio should 1.42 (the reciprocal of our 70% assessment ratio). This indicator also has problems; the major one being that it only works if the Town Assessor got her assessment of the property right. This ratio also has a built-in bias, since people who have a low appraisal don’t appeal their assessment making the gain look artificial high. In addition, renovations done after the last appraisal may not be included so once again that ratio is biased towards showing greater appreciation.

  1. Square footage

With the rise of Zillow and other housing websites square footage has gotten a lot more important for rankings, price per square and pricing, but it is not a fixed number.

Field Card

Now you would think that square footage would be an easy number to calculate, but anyone who has ever done a commercial lease can tell you there is no one right answer. In Greenwich the number that you see on most listings is the square footage from the Tax Assessor’s field card.


This number is not calculated the same way that Planning and Zoning calculates square footage for FAR (floor area ratio) purposes, but agents use it that way all the time. In most cases the FAR square footage will be smaller so a house that is maxed out on FAR based on the Assessor’s number may not be, but don’t expect to be able to add another wing on.

Basements and Attics

The Assessor does not normally count underground basements in square footage just as they are not counted for FAR purposes, but walkout basements are included. Attics also usually don’t count, but do count if they are finished and the roof is over 7 feet tall.

New Construction

The other source for square footage is from an architect. Architects generally include the basement, especially since builders want to maximize the square footage they can put on a property.

Bottom if you like the house, but think there may be too much or too little square footage go see it. That’s why there are open houses. 🙂 Also don’t buy a house based on price per square foot, particularly list price per square foot, see the house.

  • Averages

I already spoke about averages in calculating appreciation, but the average itself can be manipulated. The statistics from the Greenwich MLS Flex system include houses, condos and co-ops, which I don’t think are very useful since houses and condos/co-ops are two separate markets.

I’ve also seen GMLS averages that are just houses, but they include out of town listings which obviously skew the average. Another GMLS issue is the way the neighborhoods are done. A lot of houses fall into South of the Parkway an area that runs all the way to the Post Road. Averages for this neighborhood are not that useful since it has several different zone and areas all of which are different. On the flip-side Banksville while a distinctive community is so small as to make anything less than an annual average not very useful.

  1. Shift in what’s selling

The other problem with averages, whether means or medians, is that they don’t tell you if the mix of what is selling is changing. You can’t say that house prices are down 6% if the average is down 6%. Most house prices could actually be up and only a few lower sales of the most expensive houses could make for a lower average. The best example of that is Copper Beach Farm, 50 acres on Long Island Sound, selling for $120,000,000 in 2014. That sale raised the 2014 average sale price by well over $100,000, but barely changed the median sales price.

  1. Zestimate and others

If there is one thing I hate it is Zillow’s Zestimate. The Zestimate amount can be close to the sales price, but often it isn’t and too many people trust the Zestimate. It is often wrong by hundreds of thousands of dollars. It keeps buyer’s from going to see houses they can afford and sellers who need to sell from talking to an agent, because they think they are underwater.

We onetime had a Zillow representative come to our office to try to get agents to advertise on Zillow. He agreed that the Zestimate was horrible, but that it was conversation starter with clients. I definitely don’t want to spend my first 10 minutes with a client explaining why what Zillow says is wrong.

There are several variations out there and used judiciously they can be helpful, particularly if you know enough about the market and the houses to eliminate the outliers, but if you know that you don’t need Zillow.

  1. Sales Numbers

Like square footage you’d think that counting the number of sales in town of 62,000 people with less than a 1,000 sales of all types, but it happens for a variety of reasons. The biggest reason is that most numbers you see from local agents are the sales that went through the Greenwich MLS and don’t include private sales.

Now there is an argument to be made that since these houses were sold privately they aren’t really part of the market, but they are around 13% of the sales and sometimes in some price range go as high as 25% of all sales.

  1. Flex