On Friday, the S&P 500 was down 0.2% and Toronto was down 0.1%.
Costco was up 2.0% to $241.46 after releasing very strong same-store sales figures for August. I listened to a recorded call regarding the results and it was not clear to me how much of the increase was due to price and how much due to increased traffic through the stores. In any case, the results were very strong and also indicate strength in the U.S. economy (were most of the stores are located). Costco continues to trade at a high P/E multiple based on expectations of continued strong growth.
On Friday, Statistics Canada released the latest jobs report which saw a notable decrease of 52,000 jobs in Canada “after two months of increases”. To my mind, it is almost irresponsible for the financial press (who should know better) to report these results without (seemingly ever) highlighting the inherent statistical and other potential sources of error in the numbers. For political polls we usually hear something like the results are considered accurate by plus or minus 3% 19 times out of 20. And that would only be the statistical sampling error, biases in political survey questions could introduce far larger errors.
For the jobs data, Statistics Canada does not highlight the error level (it probably should) but does provide a link that indicates that the statistical error alone is plus or minus about 58,000 19 times out of 20. AND, it acknowledges that there are many other sources of errors.
To my mind, it is misleading to speak of 52,000 jobs lost in August as fact given the size of the statistical errors and the other errors.
At best, the jobs survey should be interpreted as indicating a trend based on some months of data. A single month’s results like August are simply too volatile to rely on. And in this case the poor August results look more like a correction to big job gains reported in July than any true loss in August.
And note too, that the figures are seasonally adjusted. The apparent job losses cannot be explained by normal seasonal patterns of students returning to school and such, as that is already accounted for (in an estimated manner).
So, how many jobs did the Canadian economy actually create or lose in August (on a seasonally adjusted basis)? The answer is no one knows, we only have statistical estimates. Yes the point estimate was a loss of 52,000 jobs from July to August. Not good. But the error level in the numbers for both July and August is too high to be very sure that there was ANY loss in August. Trends in job gains or losses need at least several months to be confirmed.