It is well known that real estate activity follows a seasonal pattern, meaning that the number of sales, the number of properties for sale and prices fluctuate depending on the time of year. For example, many more transactions are concluded in the spring compared to the summer. This seasonal effect has an impact on how sales statistics are analyzed, since it would be inappropriate to compare the sales of two consecutive months or quarters. For example, comparing the sales levels in July with those of June would be an incorrect method of comparison since there are always fewer sales in July. To eliminate the problem of seasonality in the interpretation of sales statistics, it is more appropriate to compare the sales results of a given month with those of the same month in the previous year. Accordingly, in the previous example, we should compare the number of sales in July 2010 with July 2009 to obtain a more representative rate of change.
Seasonal patterns can also be seen in the average price or median price of transactions. Sale prices tend to be lower in the spring since there are proportionately more first-time buyers at this time and since they normally buy less expensive properties than experienced buyers who are proportionately more active in the fall. A monthly or quarterly series on the average price or median price of properties would also show a seasonal character.
To obtain a more relevant rate of price change, the same rule as previously mentioned should be applied, i.e. rather than comparing prices in two successive periods, the prices should be compared with same period of the previous year. This is the approach used by the Québec Federation of Real Estate Boards in its publications and press releases.
Another way to address the issue of seasonality is to “seasonally adjust” the data using statistical methods. There are several ways to accomplish this, but the methods developed by Statistics Canada and used by the Canadian Real Estate Association, have served as a reference.
Only when data is seasonally adjusted can we compare two successive periods, such as July sales with June sales. It is important to understand, however, that with seasonally adjusted data, the number of transactions no longer represents real data, but is “corrected” data that takes into account seasonal variations.