Have you wondered what seasonally adjusted statistics are and why RAHB puts them into monthly stats releases? Here’s the scoop (with thanks to CREA Chief Economist Gregory Klump and CREA Senior Economist Shaun Cathcart for their help) .
Whenever you read in the news about GDP growth, or how many jobs were created last month, consumer spending, housing starts, etc., it’s likely you are reading seasonally adjusted data. By looking at seasonally adjusted data, trends can be more easily identified than they can be with raw, or actual, data.
Seasonal adjustment puts monthly data on an apples-to-apples basis with all other monthly data. You can compare any month’s raw data (i.e. actual results) to the raw data of same month in any other year, but that’s not a perfect comparison because:
(a) it does not account for the trading day effect – that is, more or fewer weekend vs. weekdays in a month compared to the same month in other years will impact how many transactions can be processed;
(b) a comparison to a year ago tells you nothing about what’s happened between then and now; and
(c) any aberrant data point – say, a surge in demand in one month – will give you a big increase at the time but will often result in a big decline in the same month a year later.
Seasonal adjustment overcomes these issues. The combined adjustment to the raw (actual) data includes adjustments for the seasonal factor (for example, slow months like December and January get big upward adjustments while April and May get adjusted down), the trading day effect (more or fewer weekend vs. weekdays in a month compared to the same month in other years), and the Easter effect (an adjustment for the Easter holiday based on whether it falls in March or April of a given year).
Seasonally adjusted stats are obtained by feeding raw data into a complicated computer model which is, according to Gregory Klump, “yawn-inducing stuff to all but statisticians”. RAHB receives its seasonally adjusted stats every month from CREA, which has contracted with Statistics Canada to produce them from our monthly raw data.
If seasonally adjusted stats are so great, why does RAHB also report on actual numbers of transaction through our MLS® system?
Reporting actual data is the standard – we do it, CREA does it, even real estate associations using the new Home Price Index do it. People like to know what the actual numbers are, and the numbers do have value. We do try to compare our actual data in a way that makes sense – for instance, we look at the 10-year average for the month to see how the numbers stack up over the longer term, and if there’s something out of whack, we will make a note of it.
Seasonally adjusted data is most useful for tracking month-to-month trends. Because there are seasonally slow months and busy months, comparing one against the other doesn’t really tell you anything useful if you look only at the actual data. That is why RAHB does not compare one month against the previous month in our stats releases. Seasonally adjusting the data to make months comparable to each other will show if sales have been trending up or down or holding steady from month to month.