Commentary on Bill C-36, An Act to Amend the Statistics Act

… an Atlas blog post

Paul G. Thomas, 17 February 2017

1 Introduction

With the introduction in Parliament of Bill C-36 on December 7, 2016 the Liberal government was acting on the commitment made during the 2015 federal election to strengthen the independence of Statistics Canada (SC).

On January 28 2017 debate on second reading of the bill commenced in the House of Commons. After approval in principle by a vote in the Commons, the bill will be referred to the House of Commons’ Standing Committee on Industry, Science and Technology for detailed study and possible amendment.

This commentary offers an assessment of those components of the bill that seek to strike a clearer and more appropriate balance between, one the one hand, accountability and responsiveness of SC to the government, Parliament and the public and, on the other hand, greater protection for the independence, impartiality and professionalism of SC.

Simply calling for a better balance between accountability and independence for SC is not enough. It is necessary to go beyond such vague declarations to describe in concrete terms the statutory provisions, structural arrangements, ongoing procedures and accountability mechanisms that will be put in place to achieve such a balance.

The integrating theme of this commentary is that Bill C-36 is an important first step towards modernizing and clarifying the governance framework for SC. There are parts of the bill that are worth analysis and debate. Some additions to the bill should be considered. While an amended Statistics Act can set a legal framework, it is the working relationships that arise from its provisions that will shape to a great extent how accountability and independence operate in practice.

The discussion on this theme will proceed as follows. Section 2 will briefly clarify some of the key terms used in the debates over independence and accountability. Section 3 will describe briefly the institutional status of SC within the constellation of organizations that comprise the Government of Canada and what that status has meant for the balance between independence and accountability. Section 4 will discuss the statutory mandate of SC and the mechanism of policy directives to clarify and update that mandate over time. Section 5 will discuss the key role of the Chief Statistician (CS) as the leader of agency, in particular how that official is appointed, the term of the appointment and the basis for removal. Section 6 discusses the new requirement for an annual report to Parliament. Section 7will discuss the proposed replacement of the National Statistical Council (NSC) with much smaller National Statistics Advisory Council. Section 8 offers a brief word on the relationship of Parliament with SC. Section 9 will offer some thoughts on non-legislative ways that an appropriate, clear and constructive balance between independence and accountability can be achieved.

Bold passages in the text are used to highlight recommended changes to Bill C-36.

2 A Brief Terminological Digression

Commentators seldom take the time to define key terms like accountability and independence; instead they presume we all agree on the meaning and practical significance of such terms. For this reason, it is worth clarifying briefly how the terms are being used in this commentary.

An insistence on greater accountability has become an automatic, emotional rallying cry when something goes wrong within government. As a consequence, the meaning of the term has become highly elastic and lost whatever precision of meaning it ever had. My preference is for a narrower, more precise definition.

Accountability should be used to refer to a formal relationship among actors or institutions that is maintained by a process of interaction. There are five components to an accountability relationship:

  • The delegation of a task or set of responsibilities, ideally based upon agreed expectations;
  • The provision of the appropriate authority, resources and a generally supportive environment to allow for the performance of the task;
  • The obligation by the accountable individual or institution to answer for the performance of the task based upon valid information;
  • The duty of the authorizing body to monitor performance and to require corrective action when problems arise;
  • The potential for the use of rewards and sanctions based on performance.

Breakdowns in accountability can arise because of inadequacies on one or more of these five components of an accountability relationship.

Numerous sources and of accountability operate within our political and administrative systems, such as: constitutional, legal, political, managerial, hierarchical, financial, performance based and contractual.

Confusion arises when accountability is used synonymously with such terms as responsibility, answerability, responsiveness and transparency. These concepts are related to, but do not express the full meaning of the concept of accountability.

Central to accountability is the notion of control. There is both a positive and a negative meaning of control. In a positive sense we say someone is in control when they are granted sufficient authority, resources and support to enable them to achieve the outputs/outcomes for which they are being held accountable. The negative meaning involves responsible and accountable actors being under control in the sense that they are subject to direction, scrutiny and rewards/sanctions for abuses, mistakes and shortcomings.

Independence is a related, elusive word. It implies some measure of autonomy or freedom to pursue a course of action without direct interference from other actors. Independence is achieved through the design of a set of legal, institutional and procedural arrangements that create a zone of autonomy and related accountability requirements. The trick is to find the right balance between independence and accountability.

Independence and impartiality are often used synonymously, but they are best seen as related phenomenon. While independence is mainly structural and procedural, impartiality is more of a cultural phenomenon. Independence can support impartiality, but someone can be independent without necessarily being impartial with respect to some matters.

Impartiality refers to a state of mind involving a commitment to the tasks of the organization and to responsible behavior based on the values of honesty, objectivity, integrity and professionalism. The leader of the organization should embody impartiality in her behavior and seek to embed the value in the culture of the organization.

Credibility is another important value. It refers to both the reality and the perception that an individual or institution is honest, trustworthy, competent, inspiring and courageous in defending core values in those defining moments when they are being challenged by circumstances or decisions.

In summary, accountability and independence are contentious, multi-dimensional notions that can be dynamic in the meaning over time. There is no one formula for balancing independence and accountability in all circumstances. Instead, how these values are applied in practice will depend on the nature of the organization and the context in which it operates.

This brings us to the role and nature of SC, which is not simply another department of government.

3 The Distinctive Tasks of SC

Under the classification scheme for organizations that comprise the Government of Canada, Statistics Canada is listed among the statutory agencies, not as a regular department of government. The distinctive features of such agencies are that they have a narrower mandate than most departments and usually operate at some distance from government, with varying degrees of autonomy. The status assigned to SC reflects the highly important and sensitive nature of the task of producing national statistics upon which society depends for greater understanding and better -informed decision-making. Society must have confidence in the soundness of the statistics and the objectivity of analysis applied to them.

SC is governed by the Statistics Act, which establishes the agency and its mandate. Passed in 1985 the Act is highly detailed and prescriptive in its content in ways that do not reflect its institutional status as a semi-independent part of government. For example, Section 22 of the current Act directs the Chief Statistician to produce statistics on a long, detailed list of subjects. It should not require an amendment to the Act to allow the Chief Statistician to add to or delete from the subject matters about which statistics will be gathered.

The Backgrounder to the bill released by the Minister for Innovation, Science and Economic Development Canada (ISEDC) declared that the proposed changes to the statistics Act would align the legislation with international standards promoted by the United Nations and the OECD.

I recommend that the bill be amended to declare in a preamble the Government of Canada’s adherence to United Nation’s Fundamental Principles of Official Statistics. Such a preamble would serve both a contextual and constructive role in the interpretation of the statute. It would confirm the spirit of the law and help with the interpretation of any ambiguity therein. It would provide a foundation for the development within SC of a culture of independence, impartiality and objectivity in the production and publication of official statistics. It would provide a basis for discussions and negotiation between the Chief Statistician and government officials when issues of independence arise.

On reading the current Statistics Act one would assume that the agency is under the continuous supervision and direction of the responsible minister for almost everything that it does. In practice, in what has become an informal, unwritten convention, ministers have not normally become involved with the methodological and operational aspects of producing and publicizing official statistics. This stance of non-interference has been appropriate to ensure both the quality of the statistics and the credibility of the data/information/ knowledge presented by SC. The unfortunate episode of the minister ordering the cancellation of the mandatory household survey illustrated the danger of relying completely upon unwritten conventions regarding respect for the operational independence of SC.

The Statistics Act also needs to be revised to grant SC the flexibility needed to adapt its statistical programming when it operates in a more complicated, dynamic and demanding environment than existed decades ago when the Act was last amended.

As discussed below, Bill C-36 goes part way towards securing greater independence for SC, but there are aspects of the proposed governance arrangements that could be improved.

4 Updating and Refining the Mandate of SC

A revised Statistics Act that is less detailed and prescriptive will mean more autonomy for SC in interpreting and applying its general principles. To achieve an appropriate balance between independence and accountability, Bill C-36 provides authority for the cabinet on the recommendation of the responsible minister to issue binding directives to the Chief Statistician on two dimensions of the activities of SC.

The Backgrounder to the bill declared that the government was basing the proposed new governance arrangements on a split between policy, which is rightly belongs to responsible and accountable ministers, and operational matters related to the gathering, production and publication of statistics, which are best left to the professional judgments of the experts in SC.

The proposed division of authority is the right one, but Bill C-36 does not fully uphold the distinction between the two domains of decision-making.

Section 4.1(1) of Bill C-36 provides for directives on the methods, procedures and operations of SC issued by the cabinet on the recommendation of the minister responsible for SC. Such directives must be tabled in the House of Commons within 15 days of being issued. Unlike directives that relate to broad policy matters (see discussion below), there is no provision in Section 4.1(1) for the Chief Statistician to make his objections public before implementing the directive on technical/operational matters.

Section 4.1(1) contradicts the policy/administration split that supposedly provides the foundation for the division of authority between SC and the government. The assumption seems to be that governments cannot be completely indifferent to issues related to what types of information are collected and how they are collected. The Background document implies that this use of the directive power would be exceptional, occurring only when the cabinet determined that certain technical/operational actions involved the “national interest.”

The dispute over where policy ends and operations begin was central to the controversy over the cancellation of the mandatory long form census when the responsible minister misrepresented the advice provided by the Chief Statistician that the quality of the data from a voluntary survey could not be guaranteed. In that case the minister tried to engage in a “backdoor directive” that would escape parliamentary and public scrutiny, but the resignation in protest by the Chief Statistician brought the disagreement between SC and the government into the public domain.

Under Section 4.1(1) political involvement in the technical and operational domain could still occur, but such actions would become public and potentially become the subject of parliamentary and public debate. Entangling the Chief Statistician in political debates is not desirable; even less desirable would be another resignation by the head of SC after the two most recent departures. If the Chief Statistician acquiesced in a directive that went against his professional judgment, and that of the executive of the agency, he would lose credibility with his peers.

There is no easy way to draw a clear line in legislation about what constitutes policy versus operations. There are sensitive issues related to the Census, population numbers, employment figures etc., where governments have legitimate concerns about how official numbers are produced and published. It could be argued that potential political risks of appearing to interfere with the operational/technical aspects of SC will deter future governments from using the directive power provided in Section 4.1(1). Given past history, however, I favour greater assurance that governments will have to boast and confess in advance that a technical matter involves a national interest, not a short-term political interest. In order to ensure that the directive power over operational/technical matters is used as a last resort in highly exceptional circumstances I recommend as follows:

Section 4.1(1) should be amended to require that before a directive respecting an operational/technical matter comes into force it must be tabled in Parliament and be subject to a 60 day notice and comment period.  

The current provision of the Statistics Act that allows the cabinet to decide the content of the Census of Population that takes place every ten years remains in place under Bill C-36. Retention of this provision means in principle that the cabinet could reduce the scope of the census to what amounts to a truncated short- form version. It is unclear whether Section 4.1(1) would allow the Chief Statistician to challenge publically such a cabinet decision and thereby oblige the government to defend its decision. If the Chief Statistician is not allowed to disagree publicly and his professional conscience did not allow him to execute the cabinet decision he would have to resign.

Section 4.1(1) should be amended to clarify that the Chief Statistician has the authority to disagree publicly should the cabinet modify the scope and content of the Census of Population in a manner that contradicts the advice of SC.

Section 4.2(1) provides that the minister alone, without the approval of cabinet, may issue directives on statistical programs. Unlike directives related to operational /technical matters, the clause provides the Chief Statistician with the authority to require that a policy directive be made in writing and be made public. There are several concerns and questions related to the wording and intended application of this clause.

Legally the minister is responsible for SC and politically he is answerable before Parliament for the affairs of the agency. Providing him with policy direction authority makes sense therefore. Requiring cabinet approval might be cumbersome and time consuming if the policy matters are narrow in scope. On the other hand it might be advisable that such additions or subtractions form the programming of the national statistical agency should at least be included in an appendix to a cabinet agenda for automatic concurrence, unless one or more ministers wished to bring them forward for discussion. Some statistical programs are of particular interest to different parts of the country and a concurrence arrangement would allow ministers from different regions to raise concerns about changes to statistical programs.

Policy directives should not be so broad as to amount to an amendment to the Statistics Act, which can only be changed after a debate and vote in Parliament. The provision that the Chief Statistician could require that a directive be made public might guard against excessively wide changes to the mandate of SC.

Allowing the Chief Statistician to insist on a written, public directive would provide him with leverage in discussions with the minister over what areas of statistical programming should be added or dropped, in light perhaps of the budgetary circumstances of the agency. Presumably the policy directive would only be issued as a last resort when there is an ongoing impasse between the minister and the Chief Statistician with respect to the range of programming conducted by the agency. Making such a disagreement public could cause political problems for the government so this is another reason why there should be an opportunity for cabinet to have its say. 

Section 4.2 (1) should be amended to require that any policy directive issued by the minister be placed on the cabinet agenda for potential discussion. No policy directive should amount to an indirect amendment of the Statistics Act.

5 Appointment, Tenure and Removal of the Chief Statistician

Another factor that affects the balance between independence and accountability involves the appointment process for the Chief Statistician. Appointed by the Governor in Council on the recommendation of the Prime Minister, the Chief Statistician currently serves for an indefinite period “at pleasure”, which means in principle that he can be dismissed at any time. In practice such a dismissal has never occurred.

As part of the roster of deputy ministers, the Chief Statistician participates in the regular meetings of that group hosted by the Clerk of the Privy Council and Secretary to Cabinet and with officials of the central agencies who serve the Prime Minister and the cabinet. It is important that the Chief Statistician remain within this community of senior public servants so that he can be aware of the overall priorities of the government and the statistical needs of the policy departments. However, governments have not -and should not- treat the position of head of SC as another departmental assignment that is part of the regular rotation of deputy ministers. Given the specialized and sensitive nature of the tasks of SC, there needs to be appropriate background knowledge and skills and a measure of continuity in the office of the Chief Statistician.

Bill C-36 provides that the Chief Statistician will be appointed by the Governor in Council for a five-year renewable term and will hold office on the basis of good behavior, not at pleasure. Under the bill he could only be removed for cause.

The Backgrounder to the bill states that appointment will be made following an “open, transparent, merit-based selection process in accordance with the government’s new approach to Governor in Council appointments.”

The descriptions of the government’s new approach found on the government’s website are not very enlightening about the process and selection criteria followed for such appointments. The website states that the selection process will vary with the nature of the position and the scope of its responsibilities. A customized approach for the appointment of the Chief Statistician makes sense but if the process and criteria are hidden there will not be sufficient assurance that the best candidates have been considered and that the final choice is qualified and committed to uphold the core values of the agency. I recommend a more independent and transparent process:

Bill C-36 should be amended to provide for an advisory appointment panel of three distinguished persons with appropriate background knowledge to conduct recruitment activity and to review applications/nominations for the position of Chief Statistician. The panel will recommend one name and place two alternative names before the Prime Minister for possible recommendation to the cabinet. If the Prime minister finds none of the nominees suitable he could nominate his own choice, but would be required to give reasons for not accepting someone from the list provided by the panel. 

The adoption of this recommendation might delay the appointment of a Chief Statistician somewhat, but with the adoption of fixed terms it will be possible to plan for anticipated vacancies. Media reports indicate that there is already a backlog of unfilled Governor in Council appointments pending. Bill C-36 introduces greater timeliness into the appointment of a Chief Statistician by a provision limiting the term of an interim Chief Statistician to six months.

The provision for a five- year would strengthen the independence of SC and bolster the willingness of the Chief Statistician to defend the integrity of the agency and its operations. The potential for renewal allows for continuity of leadership at the top of the agency. There is no perfect length of term, but allowing for up to ten years would enable the Chief Statistician to lead the agency through a decennial census. Renewal would provide an opportunity to deal with the possibility of leadership problems at the top of the agency.

The provision that the Chief Statistician serves on the basis of good behavior and is only removable for cause means if a government sought to replace a Chief Statistician it would have to demonstrate incompetence or some other serious problem and the action to remove would in all likelihood become the subject of parliamentary and public controversy.

The amendment that I propose would accomplish a better balance between independence, competence and professionalism in the position of Chief Statistician with an appropriate degree of loyalty and responsiveness to the government of the day.

Beyond the legalities and mechanics of appointment the degree of impartiality, courage and credibility of the Chief Statistician will depend on the character, values and capacity for reflection possessed by the incumbent.

6 Annual Report form the Chief Statistician

Bill C-36 requires that the Chief Statistician file an annual report on the performance of the agency with the responsible minister and this report will become a separate part of a report from the minister tabled annually in Parliament.

There is nothing wrong with this new requirement. At present SC is part of the ministerial portfolio of the Minister for Innovation, Science and Technology and the Chief Statistician currently reports to the minister within the administrative structure of the department, in other words through another deputy minister. The proposal for an annual report might give the Chief Statistician more direct access to the minister and encourage the minister, who has a wide scope of responsibilities, to pay more attention to the activities of SC.

Adding another reporting obligation to the Chief Statistician will involve additional staff time and cost. There are already numerous accountability devices that involve scrutiny of the performance of the agency. There is the annual spending review conducted through the Treasury Board /Treasury Board Secretariat. There is the annual review of the performance of the Chief statistician. SC presents the government and Parliament with Estimates of spending, a Report on Plans and Priorities, a Departmental Performance Report and a report connected to the management Accountability Framework. Some consolidation of reporting obligations would probably make sense.

7 The Canadian Statistics Advisory Council

Section 8.1 of Bill C-36 provides for the creation of the Canadian Statistics Advisory Council (CSAC). The CSAC is advisory to both the minister and to the Chief Statistician. It will advise on “ the overall quality of the national statistical system, including relevance, accuracy, accessibility and timeliness” of the statistics.

The CSAC will be comprised of ten members, including a chairperson to be appointed by the Governor in Council and to serve at pleasure. The Chairperson will be appointed for a five-year term with eligibility for reappointment for three years. The other members will be appointed for three -year terms, with eligibility for a second term of three years.

There are several issues related to this proposal to create a new advisory body.

The CSAC would replace the existing National Statistics Council (NSC). A brief history of the NSC is required to understand what is involved with its replacement.

The NSC was established in 1985 by the government of the day acting on the basis of an Order in –Council. This method of creation means that the NSC is currently not mentioned in the Statistics Act. It also means that the existence and role of the NSC could be changed through another cabinet order without reference to Parliament.

The formal mandate of the NSC is very brief: ”to advise the Chief Statistician in setting priorities and rationalizing statistical programs.” This wording reflects the context of budgetary austerity and downsizing that existed at the time of its creation. Over time the understanding of the role of the council evolved. It is now seen to be at the apex of a set of consultative relationships that SC has established to promote awareness, understanding and responsiveness to changes within Canadian society and the world.

Originally the minister responsible for SC appointed members of the NSC. Later this changed to appointment by the Chief Statistician. There are no formal representational rules involved with the selection of members. There are usually about 40 members who come from all parts of the country, have diverse educational and occupational backgrounds and represent a range of opinions. Many members are experienced users of the data from SC.

The Council meets twice annually to hear presentations from program professionals within SC and sometimes from outside experts. Except for expenses, the members of the NSC are not paid for their services.

It is also important to note that the council is advisory to the Chief Statistician. It is not a decision-making body like a board of directors of a crown corporation. Formal votes are almost never taken at the NSC. Council members can raise concerns based on their experiences in different fields of activity and this is valuable in terms of promoting within the agency awareness and responsiveness to outside influences. The Chief Statistician is also free to seek advice on specialized topics from working groups created by the NSC.

The proposal to replace the NSC with a much smaller CSAC of just ten members seems to be based on confusion over the most effective structure of an advisory body compared to a governing board. The NSC was never meant to be a governing board like those that direct crown corporations. It did not take decisions on the content of statistical programming. As paid appointees of the government there is the risk that the ten members of the CSAC will lapse into a governing approach to their duties.

Moving to a smaller advisory council seems to be driven by considerations of productivity and costs. However, narrow efficiency must be balanced with adequate representation of perspectives that should guide national policy on official statistics. A ten- member council will not be large enough to allow for adequate representation of the diversities of the country and provide a wide enough range of expertise. The savings from the smaller council will not be great, especially because the ten members of the CSAC will be paid whereas the 40 members of NSC are only reimbursed for expenses.

Section 8 of Bill C-36 should be amended to provide for 20 members, including a chairperson, appointed by the Governor in Council to reflect the wide range of interests served by the statistical programming of SC. Recruitment should be done on the basis of an open process of application and nomination. 

Having members of the CSAC appointed by government and making the CSAC advisory to both the minister and the Chief Statistician could lead to confusion over the role of members. Are members meant to serve as the “eyes and ears” of the minister providing a window into the ongoing activities of SC? Is the CSAC expected to take collective positions on issues related to the mandate of SC? When the Council advises the minister, will it indicate the range of opinions on an issue? What happens when there is a disagreement between the Chief Statistician and the CSAC over what advice should be provided to the minister? Are such disagreements most likely to arise on sensitive issues that involve both political and technical considerations? Would both streams of advice be made public? What happens to the constructive working relationship between the CSAC and the Chief Statistician in the aftermath of a public disagreement? I do not know whether it is appropriate and how such potential issues might be handled in legislation. Perhaps the best mechanism would be an informal protocol agreed to by the minister, the CSAC and the Chief Statistician.

8 Parliament and SC

The Chief Statistician is answerable before Parliament and its committees, but is not formally accountable to those bodies because only ministers of the crown have responsibility and accountability for appointment and removal of the head of the agency. Immediately after appointment and on a periodic basis the Chief Statistician will appear before the House of Commons Standing Committee on Industry, Science and Technology but the committee cannot issue binding directions.

Under Bill C-36 the requirement for an annual report on the state of the statistical system will likely lead to appearances, either together or individually, by the minister, the Chairperson of CSAC and the Chief Statistician. There will need to be protocols developed about how these interactions will be handled. For example, will it be appropriate for MPs to question what advice was given and accepted or not accepted by the Chairperson of CSAC and/or the Chief Statistician.

9 Some Non-legislative proposals for Balancing Independence and Accountability

The legal, structural and procedural components of BillC-36 will provide the framework for a set of formal accountability relationships and related interactive processes. The bill clarifies and strengthens the independence of SC while retaining the appropriate right of elected, responsible and accountable ministers to set policy directions and to oversee the activities of SC. I have recommended a number of changes to the bill that would refine the sections of the bill that allow for binding policy directives to be issued to the Chief Statistician.

Beyond the changes that I have recommended above, I would recommend as a practical operational matter that a matrix be developed that would indicate the range of key decisions involved with the operations of SC (e.g., policy, plans, budgets, staffing, technical matters, etc.) and that for each type of decision a designation of primary or shared responsibility be made. This will not eliminate potential questions of who is accountable for what, but it will provide guidance for how different types of decisions should be handled.

For a responsible head of SC to be held fairly accountable for the performance of the agency, the accountability framework must involve the appropriate assignment of authority, the provision of adequate resources and a generally supportive environment. The centralization in Shared Services Canada of IT services on which SC so heavily relies illustrates how broader administrative changes within government can adversely affect the performance of SC. In effect SC has lost control over the budgetary investment in its technological capacity as a time when the demands for greater volume, sophistication and security of computer systems is urgent.

I have read and heard how the new agreement between the two organizations has supposedly brought greater assurance of timeliness, reliability and cost effective service from Shared Services Canada. I think that it needs to be made clearer what performance criteria will be used to measure the quality of service provided by Shared Services Canada, what incentives are embedded in the agreement to promote efficiency and responsiveness to SC, what conflict resolution mechanisms are provided for and how accountability issues will be handled should a breakdown of some kind occur.

Independence and accountability are not just about structures and processes; they also involve individuals adhering to professional values, together with the creation of strong, shared cultures within organizations. SC already has a strong, shared culture of professionalism and objectivity. However, as the agency moves into new areas of programming and there is a generational turnover in the ranks of its managers and employees, the adoption of a specialized code of conduct on the principles and processes of national statistics gathering and publication could be a useful to reinforce and update professional norms of behavior. Other national statistical agencies have adopted such a code.

Finally, how independence works in practice will depend greatly on the character and integrity of the Chief Statistician. He or she should be strongly committed to core values of the agency, embody those values in his daily behavior and strive to embed those values deeply into the culture of the organization. An internalized subjective sense of responsibility and a concern for one’s professional reputation should be a more important motivation than formal accountability processes. This is why the selection process for the Chief Statistician is so important.


Author: Paul G. Thomas is Professor Emeritus and Senior Scholar at the Department of Political Studies, University of Manitoba. Professor Thomas has served on the National Statistic Council since 1996 but the opinions expressed here are strictly his own. Professor Thomas welcomes comment on this post and can be reached at

Page created by: Ian Clark, last modified 29 March 2017.

Image: Voices-Voix,, at accessed 29 March 2017.