Economists are predicting better news in Friday's revised jobs report based on the view that everything from housing starts to retail sales suggest the Canadian economy is not as bleak as first suggested in last week's erroneous Labour Force Survey.
The agency took the highly unusual move this week of pulling its monthly job report after finding an error. The nature and size of the mistake won't be revealed until Friday morning when Statistics Canada releases its corrected report.
The report that was pulled showed Canada lost 59,700 full-time jobs in July, while 60,000 part-time jobs were created, for an overall flat market.
Economists had been projecting 20,000 overall job gains in July and were dusting off those projections ahead of Friday's release.
National Bank chief economist Stéfane Marion is among the more optimistic forecasters, with an expectation that July will actually show a net gain of 30,000 jobs. He argues the original Statistics Canada report was out of sync with the strength shown in other economic indicators.
One of the most glaring examples was that last Friday's report said the construction industry lost 42,200 jobs in July, which was at odds with a Canada Mortgage and Housing Corp. report showing housing starts beat expectations in July.
"That was mind-boggling," he said. "These are numbers we don't see outside recessions. I just don't believe we're in a recession, so therefore I would expect the full-time employment numbers to be [higher]."
The Labour Force Survey is arguably the agency's most important report. It produces the national and regional unemployment rates and provides key information as to the performance of various sectors and regions of the economy.
Former agency employees say there is a detailed system of checks and balances that is meant to catch errors. Statistics Canada has said it is not aware of any previous occasion where it has pulled its Labour Force Survey. The only explanation it has provided is that the error was connected to a new processing system that was launched in July.
One issue economists are struggling with is whether to give the pulled report any weight at all or dismiss it completely.
CIBC's chief economist Avery Shenfeld put out two projections Thursday. If the pulled Statscan report is rejected out of hand, then CIBC reverts to its original expectation of a 20,000 job gain. On the other hand, with the assumption that the Statscan report was "partly" right, the bank would cut its forecast to 10,000 new jobs.
"Combining these two scenarios, then, logic suggests that we should indeed be expecting an upwards revision," wrote Mr. Shenfeld in a research note. However the economist cautioned that it is hard to make such estimates given that the agency has not provided any hint as to the magnitude of its mistake.
As a former employee of Statistics Canada, Toronto-Dominion Bank chief economist Craig Alexander expressed some sympathy for officials at the agency.
"This was my true worst nightmare when I worked there," said Mr. Alexander, who held several positions at the agency between 1992 and 1996. "I can appreciate that things things can happen. On the other hand, oh my goodness, the Labour Force Survey? It was absolutely the wrong survey to have an error on."
The mistake builds on broader frustration with the appearance of high volatility in the survey's month-to-month numbers. Part of that may be a mix of a low-growth economy and the survey's wide margin of error. The effect is that observers now focus more on trends rather than the latest monthly report.
"I find it very frustrating," Mr. Alexander said. "The volatility in the numbers is such that I don't put a lot of weight on the last data point any more. I focus on three month, six month, 12 month averages because it's too volatile."