Good government and Statistics Canada: The need for true independence

On Saturday, June 26, 2010, the Government of Canada announced its decision that the 2011 census would include only the eight questions from the traditional short-form. In effect, this cancelled the mandatory long-form census that included an additional 53 questions on a variety of demographic, social, and economic subjects. The government asked Statistics Canada to undertake a voluntary survey instead, including the original 53 questions from the long-form.

This decision did not go over well with users of census data, including provincial and municipal governments, non-government organizations, academics, the media, pollsters, and many others. According to one count, 370 organizations— representing the whole spectrum of the Canadian population—expressed their displeasure at the decision.

The government’s initial response was two-fold: they insisted that Statistics Canada had given them advice that a voluntary survey can produce as good results as a census; and they claimed that Statistics Canada and its Chief Statistician were totally supportive of the government on this issue. This was not the case. I should know; I was the Chief Statistician at the time. I resigned shortly thereafter.

It is useful to quote at length Alex Himelfarb on the reasons behind my resignation as the Chief Statistician:

Let me be clear about what this was not. This was not a public servant substituting his own judgment for that of the government or in any way being disloyal. Quite the contrary: in the face of criticism from colleagues, Statistics Canada seemed poised to implement the voluntary approach and, in the traditions of public service, Munir was and continues to be publicly silent about his advice. Nor was this an instance of a public servant fighting for turf or more resources. This is not about defending big government or public service jobs as some critics of government and public service will immediately assume. Indeed, the voluntary approach will cost more and require more people. Munir himself played a major role in the past in cost cutting and reducing the size of public service, and since becoming Chief Statistician, he has overseen cuts to surveys, cuts which the agency and some of its clients found very difficult and troubling, but which he did nonetheless and with no visible controversy. No, it was none of these things. This was about the integrity of Statistics Canada and of the public service (emphasis added). The decision to replace the long form census with a voluntary version put the Chief Statistician in a difficult position. The way the decision was handled put him in an impossible position.

These events were covered extensively by the Canadian media. The issue was taken up first by a Parliamentary Committee, then Parliament itself. It became an international news story.

But beyond the cancellation of the long-form and the resignation of the Chief Statistician, a third issue emerged which is equally important. However, it did not receive the attention it deserved. This is the issue of Statistics Canada’s independence from government interference. The following are the famous words of the Minister Responsible for Statistics Canada, Tony Clement, that appeared in an interview session with Steven Chase of the Globe and Mail on July 20, 2010:

Q: Is Statscan an independent agency? I am unclear on that.

A: It operates pursuant to legislation and it does report to a minister who is responsible and accountable to the public.

Q: So it’s not independent like [Auditor- General] Sheila Fraser? A: No. No. Q: S o it’s not arm’s length A: No. Q: Ok I was unclear on this. I think maybe I got the impression it was.

A: Sometimes some of them like to think they are—but that doesn’t make it so. They report to a minister.

Clearly, to the government of the day, Statistics Canada was not independent. This has serious implications for the quality and utility of data collected by Statistics Canada, and the public decision-making that this data supports.

This article examines the need for an independent Statistics Canada in order to ensure sound policy decision-making, informed public choice, and good government. First, it examines the importance of evidence-based decisions, and shows that this depends on high-quality data. It then goes on to look at how independence is necessary for Statistics Canada to achieve its goals, and how the current Statistics Act does not deliver that crucial separation from government. It concludes by proposing some changes to the legislation to achieve true independence.

The importance of evidence-based decision-making

Decisions based on evidence, rather than ideology, enhance the well-being of citizens both at the personal and public policy levels.

Consider monetary policy. The Bank of Canada has an inflation target and adjusts monetary policy when it believes the target will not be met to its satisfaction. Canada’s inflation outcomes, and the Bank of Canada’s role in that context, are some of the factors that have contributed to Canada’s strong economic performance in recent years, including its ability to cope with the current financial and economic crisis.

Consider corporate tax policy. Two contradictory views are often heard. On one side, the argument goes like this: lower corporate taxes increase investment that, in turn, improves productivity and creates jobs (the conflict between jobs and productivity in the short run is unfortunately forgotten in this equation). On the other side, the argument contends that corporate tax reductions transfer wealth from the poor to the rich, and this carries unacceptable social costs.

Only evidence can bridge the gap between these conflicting views to allow policy makers to follow a policy that enhances citizen well-being. This evidence could show that the outcome may depend on a range of other factors that may shift over time. Thus, it may be hard to determine a priori which of the two outcomes to expect at a point in time.

MunirSheikhIn an article I published in the April 20, 2011 Globe and Mail, I made three points based on evidence drawn from a particular time period to argue that Canadian corporate tax cuts did not produce the expected outcomes during this period: first, if we look at the actual investment performance during the 2000s in relation to the 1990s, we find that investment growth did not keep pace with profit growth by a long margin, even before taking into account the reductions in corporate tax; second, using simple illustrative calculations, every one-point reduction in the Canadian corporate tax rate was equivalent to the Government of Canada writing a $500 million cheque to the US government; and third, with the US corporate tax rate double that of Canada’s, we are nowhere in sight of US productivity growth which is the true anchor for any country’s rising living standards.

At a personal level, Canadians make decisions every day based on evidence. They look at mortgage rates before deciding whether, and where, to get a mortgage. They look at food prices to determine what to buy and how much. They look at the job market in various parts of the country to decide whether to move or not.

Now imagine all of this happening without citizens and governments paying attention to an evidence-based analysis of the issues: the Bank of Canada not interested in understanding why the inflation target is important; the federal government not realizing why it should or should not cut corporate taxes; and citizens not thinking about what high mortgage rates, high food prices, and job opportunities could do to their well-being. Without appropriate evidence-based analysis, we will all be poorer—in every sense of the word.


The importance of good data

The Consumer Price Index (CPI) provides a helpful example of what data can do for its users.

Data describe events as they unfold and thus give us information on things as they change. Every month Statistics Canada releases the CPI describing the change in consumer prices for the past month. It may show that average prices rose or fell by a particular magnitude. This monthly measurement can be compared with previous months to get a sense about inflation rates over different time periods.

Data can also be used to gain insight into a phenomenon. The detailed information contained in the CPI release can pinpoint where prices are changing most. For example, data may show that the main reason the average price rose last month was because of significant increases in auto insurance premiums. This would allow citizens to understand the reasons for an increase in their cost of living.

Data allow analysis of the reasons behind observed developments. Using other relevant data, such as the frequency and seriousness of accidents, it may be feasible to analyze the causes underlying the increase in insurance premiums. The understanding provided by this analysis can be helpful in making improvements in outcomes, such as a policy to improve highway safety that helps control insurance premiums.

The analysis made possible by data then allows the provision of a context for decision-making. The information contained in the analysis may, for example, show that the increase in insurance costs was driven by factors that may not be around permanently, in which case there may be no need for policy action. Or, this information may show otherwise.

Data help in decision-making. Indeed, it is the most important contribution data make to improve the well-being of citizens. Continuing with the CPI example, the increase in the inflation rate, along with the details of where the increased pressure may be coming from, gives the Bank of Canada the ability to relate this information to its objectives and adjust its policy levers to achieve desired results.

Data are also used to monitor progress in achieving objectives. For example, the Bank of Canada monitors progress on the inflation front by examining the core rate of inflation, which subtracts the volatile inflation components from the overall rate of inflation, in the context of its inflation targets.

Data are used as well to build systems. In the context of the CPI example, inflation is a key variable in the development of economic models that are used for a variety of purposes.

These models, built on data, can be used for forecasting and predicting. These predictions allow decision-makers to anticipate adverse events and take action. For example, models may show that the inflation rate could fall below the central bank’s target range, encouraging the bank to take preemptive corrective action.

Data are used as well for evaluation of outcomes. The evaluation exercise is helpful in determining whether or not objectives have been achieved. If yes, the data can determine whether the goal was achieved satisfactorily, and if it had the desired effect. If not, the data can likewise show why not. Such evaluations are a key to making adjustments in public decision-making.

In sum, data provide the foundation for knowing things the way they are and taking steps to making things the way they should be. In this sense, the importance of data in enhancing human well-being must not be underestimated.

The need for an independent statistical agency

In view of the 2010 Census developments and the views expressed by the Minister responsible for Statistics Canada at the time, Canada would benefit significantly by enhancing Statistics Canada’s independence through changes to the Statistics Act.

Despite problems with the existing legislation, Statistics Canada was considered among the best statistical organizations in the world. Some would have rated it as the very best.

There were three key reasons for this reputation: its centralized structure allowed it to gather data for the whole country in a cost-efficient manner; its governance and management structure allowed it to be innovative; and it enjoyed a tradition of operating at arm’s length from all governments.

The decision by the government to abolish the longform census, and the Minister’s very public description of Statistics Canada as subject to political control, have damaged this long-held tradition.

Both of these factors—the cancellation of the long-form census and the Minister’s view—have negative consequences for the quality of data Statistics Canada produces.

First, it will affect the long-form survey data.

It is a statistical fact that a voluntary survey cannot hope to act as a substitute for a mandatory census. A voluntary survey will inevitably result in uneven response rates from different population groups and different geographic areas. Increasing the sample size cannot offset this problem. If there is a bias in the original sample, that bias will be magnified in a bigger sample if it continues to mimic the properties of the earlier, compromised sample. Suggesting that a voluntary survey with a larger sample size can replace a mandatory census is like saying that if you take a wrong turn, you should drive faster in the wrong direction to get to your destination. With a voluntary survey, many data users who depend on the long-form census—including the federal government—will lose the data quality they need.

Second, to the extent that the long-form census data provide a benchmark for other Statistics Canada surveys, the quality of data from these other surveys will also deteriorate.

If the government persists in the view that Statistics Canada is not independent, data quality—or at least the perception of data quality—will further suffer. Keep in mind that the vast majority of data users are not in a position to determine for themselves how good or bad the data are. They use them, and base their decisions on them, only to the extent that they can trust the organization that produces them. Trust in the organization depends on two critical factors: how good the organization is in technical matters, and how independent it is from government control to produce data that accurately reflect reality. Take away the independence of and trust in the organization, and even good data become less useful.

The Statistics Act and suggested changes

Section 7 of the Statistics Act gives the Minister wide-ranging powers on technical matters:

The Minister may, by order, prescribe such rules, instructions, schedules and forms as the Minister deems requisite for conducting the work and business of Statistics Canada, the collecting, compiling and publishing of statistics and other information and the taking of any census authorized by this Act…

Section 8 of the Act defines voluntary surveys as a technical matter, and states:

The Minister may, by order, authorize the obtaining, for a particular purpose, of information, other than information for a census of population or agriculture, on a voluntary basis…

On the issue of what questions should be asked in a census to collect data to satisfy the most important data needs of the country, the Act gives the authority to the government generally, in Section 21:

The Governor in Council shall, by order, prescribe the questions to be asked in any census taken by Statistics Canada…

Many other examples of the government and Minister’s role in managing Statistics Canada can be found in the Act. Overall, it is clear that despite its long tradition of independent action, the Statistics Act gives little protection to the autonomy of Statistics Canada.

There are three broad areas where I would suggest a change in the existing Act.

First, the Act should give the authority for all technical and methodological matters to the Chief Statistician. David Dodge, Mel Cappe, Alex Himelfarb and Ivan Fellegi—four eminent former senior public servants—also advocated for this proposal in a September 2010 letter addressed to the Prime Minister of Canada.

Second, give the authority to determine census questions to the Chief Statistician, as is the case in Australia. The census is a constitutional responsibility. However, given the way the Act currently reads, the government of the day can control the contents of the census. In the extreme, a government may ask as little as one question in the census, on any topic, to meet the constitutional requirement. There should obviously be a process where the Chief Statistician must base his/her decision about what questions to ask in the census on input received from citizens and data users. Statistics Canada currently follows a strict process that, as events have shown, can be overridden by political imperatives. A citizenand researcher-driven model, or some other independent mechanism, should be protected in law.

Third, given the nature of the responsibilities of the Chief Statistician (particularly if the law is amended as I have suggested), and the fact that Statistics Canada is a department of the government, the current mechanism for the appointment of the Chief Statistician should be replaced. At present, the Prime Minister appoints the Chief Statistician at his or her pleasure. Ivan Fellegi, a former Chief Statistician, has proposed the establishment of a committee of senior former public servants to put forward a list of appropriate candidates for the consideration of the Prime Minister. I support this proposal.

Concluding remarks

In this article, I have argued that evidence-based decision- making is essential to the enhancement of the well-being of citizens. Such decision-making is obviously not feasible without good evidence. We are fortunate in Canada to have Statistics Canada as our data collection agency. However, comments from the Minister Responsible for Statistics Canada in the context of 2010 census developments have brought to the fore a serious problem in the Statistics Act regarding the independence of the agency.

Trust in Statistics Canada is crucial for evidence-based decision-making. To ensure that that trust in the agency is not put at risk, we must amend the Act to enshrine in law what previously was a strong tradition of independence and autonomy. The cancellation of the long-form census highlights how this independence is currently vulnerable. To ensure the best outcomes for the citizens of Canada, we need to protect Statistics Canada from outside interference.

Munir A. Sheik is a Distinguished Fellow and Adjunct Professor in the School of Policy Studies at Queen’s University. From 2008-2010, he was Chief Statistician of Canada.