StatsCan, or StatsCan’t?

Posted on May 3, 2012 in Governance Delivery System

Source: — Authors: – business – Deep cuts prompt fears statistics agency could lose its global reputation for quality
May 2, 2012.   By Louise Egan, Reuters

Just two months ago, economists and policy wonks were cheering the news that Statistics Canada, the much lauded government statistics office, had eliminated fees for its online databanks, making millions of figures available for free.

Now the quantity of that data is under threat from the biggest budget cuts in recent memory.

Experts fear the quality will fall as well, hurting StatsCan’s global reputation and compromising data that shapes government financial and social policy, as well as business investment.

Nearly half of the agency’s 5,700 staff have received the layoff notices that are being issued in many federal departments as the government seeks to eliminate the budget deficit by 2016.

The final job cuts will be far smaller than that, but there is bound to be pain.

“Government departments will see the volume and detail of information available sharply reduced,” Chief Statistician Wayne Smith said last month in a sombre private message that spoke of 2012 as “a year of sacrifice.”

In a video of his address obtained by Reuters, Smith said spending cuts at StatsCan would be much deeper than the $33.9 million — or about seven per cent — outlined in the federal budget due to an “unprecedented” loss of an additional $20 million in income from other government departments.

Three-quarters of the savings would come from cutting programs, meaning fewer surveys, less data and less analysis.

Right now, StatsCan employs an army of experts to conduct 350 surveys that range from the market-movers like employment to curiosities like dried egg stocks and bee-keeping.

StatsCan isn’t saying what it will cut. But the Conservative government says the impact will be minimal at StatsCan and across the entire federal bureaucracy.

“Seventy per cent of our staff reductions have been in back office and administrative positions. They do not affect front line at all,” said Tony Clement, the minister who oversaw the process for deciding where and how much to cut spending.

“We’ve made it clear in all of these cases that for the services that Canadians depend upon, they have been red-circled.”

That’s a promise that rings hollow with many of StatsCan’s present and former staff, among them Ivan Fellegi, who aspired to be a poet in his native Hungary before fleeing to Canada and joining StatsCan in 1957. He was the agency’s chief statistician from 1985 to 2008 and is now retired.

“Everything that is going to be cut is going to hurt a lot,” said Fellegi, who still has an office in StatsCan that he visits twice a week. “A lot of the bureau’s budget is basically not touchable.”

The latest cuts are seen as a huge setback for an institution viewed as a global benchmark for unbiased, accurate social and economic data and research since it was created almost a century ago.

Canada spends an average $530 million, or about $16 per capita on government statistics, a tiny fraction of its $276 billion annual budget. In Australia, where the main statistics office shares StatsCan’s top international billing, the cost is some $17 a head, but that excludes farm and commodities data which, unlike in Canada, is gathered by a separate agency.

Experts say it’s very difficult to compare countries’ spending on official statistics because many have decentralized their data-gathering activities.

StatsCan, which boasts the second lowest revision rate to GDP numbers of the G7 countries, is one of the most compliant with standards from the International Monetary Fund and the Organization for Economic Cooperation and Development (OECD), along with Australia and the United States.

But StatsCan faced howls of protest last year for eliminating — at government behest — the compulsory long-form census that supplemented the basic census form with questions on ethnicity, income and housing. It was replaced with a shorter, voluntary survey.

Critics say the new methodology breaks the historical statistics chain, making comparisons harder. They warn the less than complete numbers will have a ripple effect on other data, including employment, that may not be apparent for years.

Chief Statistician Munir Sheikh quit in protest, and chief economic analyst Philip Cross later followed suit.

StatsCan was already reeling from a 2010 budget freeze. And the decision to stop charging fees robbed it of at least $2 million in outside revenue. Unions estimate some 700 jobs will vanish, ending a 25-year no layoffs policy instituted by Fellegi.

Former staffers said the new cuts could not only affect the quality of data and damage its reputation, but risk financial market volatility. The sources asked not to be identified because they are not authorized to speak, or due to the sensitivity of the matter.

StatsCan likely won’t tinker with the three top-tier indicators — the labour force survey, consumer price index and the dozens of “national accounts” surveys that produce gross domestic product estimates. But beyond that little is clear.

“Fiscal and monetary programs will not be affected,” StatsCan said when asked if economic indicators would be cut.

Getting rid of the composite leading indicator is seen as a no-brainer as it needs an update to remain relevant. The agency will stop publishing seasonally-adjusted data for new motor vehicle sales.

Other items for the axe could include social surveys, data processing and analytical work, former staffers said.

StatsCan said its quality controls would remain in place. “We maintain analytical capacity to verify the accuracy and relevance of the statistics it produces, assist users in interpreting the data, and develop relevant concepts for the production of statistics,” the agency said.

It’s true that StatsCan’s plight appears mild compared to that of Britain’s Office for National Statistics, which will be cut by 17.4 per cent in real terms over four years.

The Australian Bureau of Statistics also faced cost pressures in 2008. It reduced the sample size of its employment survey after cuts, only to have its funding restored because of the negative impact on markets.

Kimberly Zieschang, head of the IMF statistics department’s real sector division, said belt-tightening governments should note that economic indicators offer a big bang for their buck, as they are essential for sound fiscal and monetary policy. And Paul Schreyer, deputy director of the OECD’s statistics directorate said statistics agencies need to analyse data as well as collect it in order to spot errors.

The thought of cuts to the relatively small team of analysts at StatsCan pains Fellegi, who built up a team of experts to popularize the numbers and shed light on key issues of the day. “I spent decades building up those areas because they were weaknesses before,” he said.

The reliability of data was on the minds of economists at National Bank recently after jobs data showed a precipitous drop in fourth-quarter employment in the province of Quebec. Taken at face value, the figures would have signalled a severe recession. The bank asked StatsCan to review its methodology.

StatsCan defended its methods and said all such surveys are subject to sampling variability, advising users to watch for trends rather than monthly changes.

Markets have put a premium on quality data since the global financial crisis. For now Canada is on one end of the credibility spectrum while countries like Greece and Argentina, who have fudged their data, are on the other extreme.

Some 200 academics who gathered in Ottawa last month to celebrate StatsCan’s newly free data want to make sure things stay that way. Their cheerleading was a polite way of telling the government to lay off statistics.

“This conference represents a broad-based effort on the part of the Canadian academic community and others who see high quality official statistics as an important pillar of a modern market economy and democracy,” said Joseph Doucet, dean of the Alberta School of Business and an adviser to the provincial government on energy and regulatory policies.


< >

Tags: , , , ,

This entry was posted on Thursday, May 3rd, 2012 at 9:55 am and is filed under Governance Delivery System. You can follow any responses to this entry through the RSS 2.0 feed. You can skip to the end and leave a response. Pinging is currently not allowed.

One Response to “StatsCan, or StatsCan’t?”

  1. Claims Tony Clement, the minister who oversaw deciding where and how much to cut spending: “Seventy percent of our staff reductions have been in back office and administrative positions. They do not affect frontline at all.”
    Clement regularly quotes Statistics Canada numbers. He would maintains that the numbers that matter can be produced with roughly half the people at Statistics Canada. Those of us who know how those numbers are made know there’s no way we get even just good quality economic numbers for Canada with half the staff at StatsCan. And some of us think that good quality health, education and other demographic numbers and environmental numbers and the numbers needed for all sorts of ongoing research in Canada matter too. .
    Sign the virtual card of support for those at Statistics Canada at if you care about this.


Leave a Reply