# Using Past Real Estate Seasonality To Forecast 2011 Home Sales

Posted by Joe Manausa on Monday, February 14th, 2011 at 12:29pm.

By measuring seasonality in the Tallahassee real estate market, we can forecast future home sales based upon past rates of sales.

I was wondering if it were possible to use historic real estate seasonality measurements to forecast future home sales, so today's blog will be an ad hoc experiment to forecast home sales in Tallahassee for the remainder of 2011.

To begin our simple real estate math exercise, let's first define seasonality as

A characteristic of a time series in which the data experiences regular and predictable changes which recur every calendar year. Any predictable change or pattern in a time series that recurs or repeats over a one-year period can be said to be seasonal.

When we apply this definition of seasonality to real estate, we observe that different times of the year produce different levels of home sales in the Tallahassee real estate market. For example, summer months historical produce more sales than winter months, and through long-term measurements, we can determine a factor for each month of the year.

The real estate graph above shows seasonality measurements for Tallahassee real estate. In the graph, the three-year trend is highlighted and has been calculated by averaging the seasonal measurement of each month over the past three years. For example, we can see that January (over the past three years) has seen 5.5% of the annual closings each year, while June has the most at 12.3% of all home sales in Tallahassee.

### Real Estate Forecast Using Seasonality

We can reverse engineer a forecast (time will tell how effectively) by using the seasonality figures. In our recent Home Sales Report, we found 227 homes closed for the entire month of January. Using the January factor from the graph above (5.5%), we can use simple division to forecast a total of 4111 home sales in 2011 (a number far higher than our current supply and demand estimates have produced).

The real estate graph above was created by using the three-year seasonality trend. It is important to note that it forecasts 4111 home sales for 2011 because we used the average of the past three years. The table on the right shows the seasonal factor for January in each of the past 20 years, and the corresponding forecast for 2011 using that year's January seasonal factor.

Unfortunately, this is a low-certainty forecast because our 20 year range of values would forecast (to a high level of certainty) that 2011 homes sales would fall between 2,986 and 5,088 (which of course is a very large range). But as the year progresses, our range will tighten and the forecasts that we produce will become more and more likely to occur.

So our simple, ad hoc real estate math experiment might not have provided us with a clear forecast for the 2011 Tallahassee housing market, but it did provide us with another tool that helps us analyze the Tallahassee housing market. When we combine this with all of our other tools, we gain an insight into the market and a better ability to make decisions on real estate pricing and value movement.

Joe Manausa, MBA is a 26 year veteran of real estate brokerage in Tallahassee, Florida and has owned and managed his own company since 1992. He is a daily blogger with content that focuses on real estate analytics and providing his clients with a tactical advantage in today's challenging market.

#### 3 Responses to "Using Past Real Estate Seasonality To Forecast 2011 Home Sales"

Roy Tomlinson wrote: Why don't you run month on month averages over the years and do a 95% confidence interval for the proportions listed in your data? Maybe I could get one of my Stats students to do it. You could then capture with 95% confidence the true proportion of homes sold, say, in the month of January. Confidence intervals would be useful - you could do a two proportion confidence interval to see if there is indeed a difference in the proportion of homes sold in January year to year. My guess is there isn't, which makes your predictive model more compelling. Interesting stuff. Roy

Posted on Monday, February 14th, 2011 at 6:31pm.

Joe Manausa, MBA wrote: Make it so #1.

Posted on Monday, February 14th, 2011 at 6:32pm.

Joe Manausa, MBA wrote: I actually did that a few years back in one of my blogs. I would love for you to put a student on it, let me know what data you need. Thanks Roy.

Posted on Monday, February 14th, 2011 at 6:34pm.