Skip to main content

Statistics Canada offices in Ottawa on July 21, 2010.Sean Kilpatrick/The Canadian Press

There were some curious and intriguing details behind the headlines of Statistics Canada's monthly employment report, as there always are. It's a serious shame – and a serious problem – that almost no one could see them.


The national statistical agency's website was out of commission since early Friday morning, before the 8:30 a.m. ET release of the February labour force survey. As of late afternoon, Statscan's website remained dark; the details of one of the most important economic indicators of its monthly calendar were invisible to the Canadian public all day.

(By the way, the report showed that the Canadian economy added an estimated 15,000 net new jobs in February, a bit better than economists had expected, and the unemployment rate dropped to 6.6 per cent, matching an eight-year low.)

You might recall that something like this happened before, about eight months ago, when Statscan's systems were down for more than seven hours on another jobs-report Friday. Not to mention the many, many occasions that Statscan's website has fizzled out for much briefer periods shortly after the release of major economic indicators, during moments of peak traffic scrambling for the fresh data.

At the time of this writing, we don't know what the problem was with Friday's system failure. Neither Statscan nor Shared Services Canada, the agency that oversees e-mail, data and network services across the vast breadth of the federal public service, got back to us with an explanation. Certainly past snafus have been placed at the feet of Shared Services, the $1.9-billion brainchild of the previous Conservative government that was supposed to streamline Ottawa's complex tangle of information technology, but has instead been blamed for everything from AWOL paycheques to RCMP systems failures.

The previous head of Statscan, Wayne Smith, resigned last September over Shared Services' handling of Statscan's information systems, which he said had not only become "disruptive, ineffective, slow and unaffordable," but compromised the independence and confidentiality of the statistical agency's data.

Now, I'm not here to point fingers. But the point is that these Statscan failures, while maybe not the same risk to public safety as the RCMP's problems, are more than just a nuisance to the economists and journalists who wallow in these economic numbers.

The system problems, when they arise, create inequitable distribution of information that is relied on, and more to the point traded on, by financial markets. That's a serious problem.

In the case of Friday's jobs data, instead of every market participant being able to see the same data at the same time on the same website, each was left to his or her own devices (literally and figuratively). The lucky ones had access to Bloomberg data and news terminals, the expensive yet indispensable toys of professional trading operations, where at least the basic highlights of the report would have been fairly quickly disseminated. Others could have turned to media reports from the smattering of news organizations that attended Statscan's pre-release lockup (in which reporters were given the release in advance but kept sequestered in a room, unable to communicate the information until the moment of the release time).

But if you were in need of the deeper statistical details below the surface of these quick-hit reports, good luck. Even the research departments of the big banks were scrambling, relying on friendly contacts at Statscan to e-mail to them whatever data they could.

All of which not only delayed the dissemination of this key economic data to the public and to financial markets, but also resulted in some very uneven distribution – in terms of both the timeliness and the amount of information that reached different sets of ears and eyes with an interest in the data.

And the employment data are very significant indeed to the bond and currency markets, especially now. It has become increasingly evident that the direction the Bank of Canada will take on interest rates hinges substantially on the evolution of the labour market. In its rate decision earlier this month, the central bank pointed specifically to "subdued growth in wages and hours" as key evidence of "persistent economic slack" in Canada.

And indeed, the February jobs report showed that despite the improvements in hiring and the unemployment rate, growth in wages and hours worked remained disappointing. Knowledge of this spoke volumes to any bond or currency trader placing bets on the timing of future Bank of Canada rate moves. And some traders had access to this information long before others.

That's simply unacceptable.

As long as these technology problems persist, they undermine the integrity of an independent, impartial national statistics provider. Access to critical data can't play favourites, even if it's by accident.

Rob Carrick discusses why some people should consider working past age 65 in order to have a better retirement

Report an error

Editorial code of conduct