The Most Useless of Real Estate Statistics

Real Estate analysts and statisticians love to publish numbers, but the questions you have to ask is, How helpful are the numbers? What did they really tell you? Are there better numbers that could be reported that tell us more?

The most frequently published number is a Year Over Year (YoY) number that show last month's performance compared to the same month one year earlier. On the surface, this seems like a number that might be helpful benchmark and tell us how the market is doing. And if we were in a market that had a daily observable variation of data, 30 days of information would be a significant amount of data and might tell us something about a trend or coming trend. However, real estate is not something that is measurable or affected by daily sales. In fact, I would argue that many products and many industries can not look at daily sales data and determine a trend line from 30 days of sales data--most certainly not with real estate.

There are some seasonal businesses that have exceptions to this data rule and I will guess that the first analysts who started running real estate trends came from this business background--but it makes no sense for real estate. Take retail sales as an example. Certain times of the year show retail sales spikes that are measured during specific retail events, like the weeks leading up to Christmas and the days surrounding Black Friday. December sales are a well known benchmark on retail sales and a quantifiable data point that can be understood and measured. Same is true of Black Friday (weekend) and other holidays to a lesser degree for certain products--like appliance sales during a few select national holidays. These seasonal events can be defined, measured and compared to previous years with some reasonable predictive quality. By looking at these trends, analysts can often see which way the wind is blowing and take a temperature of consumer confidence and have a better idea of things to follow.

Now to compare this to our business. Real Estate has some "traditional" cycles with low periods during the holidays (the opposite of retail sales) and one high season cycles during the summer, but what we see most often reported in real estate statistics is ONE MONTH this year compared to ONE MONTH last year. This is a completely misleading, and devoid of useful information. One month of data in real estate tells us almost NOTHING without know what happened during the preceding two months or the following months. The value of a Real Estate statistic depends on a bigger picture than 30 days of sales data (1) because real estate is not a one day purchase--a transaction typically takes at least 30 days (2) because the number of confounding factors that affect real estate business are much bigger than a single event, and include a whole host of geographical factors, economic factors, local momentum and basic business factors such as supply and demand which can be dependent upon many factors beyond the control of Real Estate business people. And like all businesses, there is an expected ebb and flow that has to be smoothed out when looking at the numbers, else the only trend you will find is that there are no dependable trends to forecast the future. With a little smoothing out of the months, trend lines can emerge in a local market to better see what may be happening while it is happening.

Real Estate does have some traditional peaks, but even when we look at the peak summer season (during school vacation), there are two different data humps that have to be examined in most markets and these two stand-alone points are a minimum of 60 days in length (x 2), not 30 days. We have the first spike in sales that occurs in the May/June block and the second spike in sales that occurs in July/August. One, as school lets out and parents are trying to move before the July 4th holiday, and one, as school is preparing to open--for families who missed the first of the summer rush. Again, preparing data-sets for one month during any part of the summer will be a misleading roller coaster and not helpful in understanding the trend in the Year Over Year analysis.

When analyst compare one month to the same month one year prior, inevitably what you see and read are reports of a lumpy and inconsistent market. Inconsistencies are not trends. They are confusion. Confusion created by persons who don't understand the value or the harm of statistics. However, if analysts took the time to provide rolling averages over a 90 or 120 days look at the market, your can begin to see the trend lines very quickly and much more reliably. The variations of market influences are quickly mitigated and real market analysis is available to even the casual observer.

I am not statistician, but I look at the kind of information that is reported today for real estate and can easily take them apart. When one month the numbers are up, the next month the numbers are down and the next month the numbers are up again, this is not a report on the health of the market, it is a report on the lack of sophistication in our understanding of our tools to measure the market and report on that data. It is time that we, as an industry, demand better reporting of data and information from those who would claim to be industry analysts and reporters of data. The data is there; now is the time to learn how to report that data in a responsible and helpful manner to investors and home buyers alike.  

Tweet

Subscribe to receive free email updates: