ARCHIVED - Expenditure Review of Federal Public Sector - Volume One - The Analytical Report and Recommendation

• Previous Page
• Table of Contents
• Next Page

This page has been archived.

11. Recommendations on Transparency and Accountability

At present, there is little overall understanding of federal compensation. In effect, the facts are lodged in hard-to-access and internally inconsistent databases that are maintained in diverse parts of the Public Service. It is true that the Public Accounts do report a total amount spent on Personnel (standard object 01). For 2002–03 this was given as $25.1 billion.^[154] This compares closely with our reported total for 2002–03 of about $24.8 billion. However, there are few supporting details reported in either the Public Accounts or other regular government reports. Neither is there any report that sets out how these expenditures have changed from year to year, or what forces have driven the changes.

The result is that the context for discussions about federal compensation issues is fuzzy at best. In fact, there is little public discussion. The main exception occurs during periodic public service strikes, when the media rely mainly on the bargaining positions publicized by the union leading that particular strike. Unions naturally reference whatever information best supports their demands. The federal government for its part tends to be slow to offer its arguments. And like the union, the government puts out the data that most favours its position in the current negotiation. The media is left to try to make sense out of all this, since they generally do not have access to broader data that would allow them to formulate their own interpretations.

From time to time other stories emerge, nearly always without any broader perspective being offered. We hear criticisms of performance pay for executives, for example, but little about how their pay compares to the private sector nor the role pay-at-risk plays in that comparability. Unions naturally complain that their economic increases are low, but we seldom hear about other contract terms affecting wages, nor the overall trend in average salaries actually paid to employees. We hear that federal public servants enjoy on average a premium versus private sector employees, but not much is said about the limitations of studies based on macro data such as the Census or the Labour Force Survey.

If it were only the media and the public lacking a broad understanding of federal compensation, that would be a significant concern. But in fact, the government itself does not maintain a systematic overview of the levels and trends of the various components of federal compensation. As far as we can determine, this study is the first comprehensive compilation and description of the subject, at least since the Glassco Report in the early 1960s, more than 40 years ago.

The existing situation appears to result from several factors. First, as we have learned in preparing the present Report, it is inherently difficult to access the information necessary to create a comprehensive view at a meaningful level of detail.

Second, much of the analytical capacity in place within the federal government was dismantled during Program Review in the early and mid 1990s. The Pay Research Bureau (a part of the Public Service Staff Relations Board) was abolished in 1992, partly as a cost-saving measure, and partly because the employer apparently viewed the Bureau's surveys as skewed. Interestingly, the bargaining agents may have had their own reservations about the Bureau's product, albeit from a different perspective. In the Treasury Board Secretariat, the six-year suspension of collective bargaining and arbitration led to the departure of key experts, as well as the reallocation of people remaining to more urgent issues. The process of rebuilding this capacity is well launched but far from complete.^[155]

Finally, it is difficult to escape the impression that tacitly the parties have found it congenial for compensation to remain an arcane area. In such an environment of limited public scrutiny, deals can be struck more easily and difficult conflicts minimized. 

Annual Report on Compensation

Good information is the foundation of sound management. Accordingly, the single most effective step toward better management of federal compensation would be to collect, analyse and publicize the most important data routinely. An annual report on federal compensation would act as the most visible product, as well as the main driver, of an improved regime for transparency and accountability for federal compensation.

The annual report would update and track a set of key indicators, and report major changes in employment levels, collective agreements, and compensation policies and expenditures during the year. Ideally it would also provide a more in-depth analysis of the evolution of some special topic each year. The exact design of the annual report was beyond the capacity of this project but we can outline some guidelines for such a product.

The annual report will work best if it provides a succinct overview of the levels and trends in federal compensation, without becoming impenetrable in its detail and complexity. On the other hand, for those who do seek details, or whose interest is specialized, there should be easy access through the internet to more detailed data sources. The biggest conceptual challenge will be to select a limited but meaningful set of key indicators to track. These should be indicators in the true sense––that is, they should summarize important information in such a way that non-specialists can discern trends and emerging issues.

The scope of the annual report should be the same as for this Report: namely, the core public service––for which the Treasury Board is the employer–, the separate employers, the Canadian Forces, the regular and civilian members of the RCMP, as well as special groups such as federally appointed judges, parliamentarians, employees of Parliament, ministerial staff, and students. We excluded federal business enterprises and other Crown corporations on the grounds that a relatively small portion of their personnel costs are financed through parliamentary appropriations, and they establish their own personnel policies. All forms of compensation should be included, reported on the basis of cost to the employer, rather than value to the employee.^[156]

These key indicators would include total spending on salaries and total compensation, employee population, average salaries, changes in the composition of the workforce, reclassifications, performance pay, value of negotiated or managed salary increases, pension contributions, other paid compensation, insurance and health expenditures, and contributions to statutory programs.

Total spending on salaries and total compensation

These are the most aggregate measures of compensation trends. We note in Volume Two that salaries and wages in the federal public sector (excluding federal business enterprises and other Crown corporations) grew from $12.5 billion in 1997–98 to $17.9 billion in 2002–03, an increase of 43%. Total compensation grew faster, rising from $16.6 billion in 1997–98 to $24.8 billion in 2002–03, a growth of 49%. These amounts should be reported in both current and constant dollars.

Employee population

An important factor in interpreting changes in the salary mass and total compensation spending is change in the number of federal employees. In reporting on growth or decline in this indicator, the Annual Report should summarize the major sources of change. It should highlight both the aggregate of Treasury Board decisions to authorize hiring more staff––or reducing staff complements––and total changes arising from departmental decisions to reallocate from non-salary to salary budgets. In this way, the evolution of the size of the public service can be tracked and debated.

Average salaries

This is a particularly interesting indicator, although not without controversy. Average salary is a gross measure resulting from dividing the salary mass by the employee population. During the period from 1997–98 to 2002–03, average salaries rose considerably faster than inflation. Although this was partly a result of economic increases that exceeded the cumulative effect of inflation, it was also a result of additional changes in salary structures, e.g. the addition or deletion of salary increments within a particular pay band, as well as changes in the composition of the federal workforce.

This indicator could be criticized as overstating changes in federal salaries, because an important component in the recent growth in average salaries resulted from shifts in workforce composition. The fact remains, however, that tracking the average cost of public service salaries is intuitively important, and usefully consolidates the several effects of disparate influences, some of which may merit deeper scrutiny.

Changes in the composition of the workforce

As we describe in Chapter 4, the makeup of the federal public service changed markedly over the 1990s and the early 2000s. Relatively highly paid groups such as Computer Services (CS), Law (LA) and Economics, Sociology and Statistics (ES) more than doubled in size over slightly more than a decade. By contrast, more operational groups such as Secretarial, Stenographic and Typing (ST), General Services (GS), General Labour and Trades (GL), and Clerical and Regulatory (CR) have declined by substantial proportions––from 82% to 28%––over the same period. Tracking such changes is essential to understanding federal compensation.

Reclassifications

We noted earlier in this Report that there are sound reasons to reclassify positions. However, it is also important to keep an eye on trends in this area, and to conduct appropriate audits to ensure that the system is managed well. The incremental cost of reclassifications should be reported as well.

Performance pay

The total expenditure and per capita value of performance pay paid to eligible federal public employees should be summarized. The proportion receiving the normal amount (in the case of Executives at levels 1 to 3, for example, up to 7%),^[157] and the proportion receiving more should also be reported.

Value of negotiated or managed salary increases

For most employees, collective agreements determine how their salaries will change. The Annual Report should aggregate the cumulative average salary increases arising from collective bargaining. This should include both the well-known economic increases, and the average value of other less publicized changes such as the addition or deletion of increments within a salary band. An appendix should provide this information for all collective agreements concluded within the year. The Annual Report should comment briefly on the rationale for changes that are outside the general pattern.

Pension contributions

The most significant area of compensation expenditure after salaries is pension contributions. The total amount contributed by the government and by the plan members should be reported for each part of the federal public sector, as well as the ratio of these contributions.

This information is currently reported in the Public Accounts and in the annual reports of the main pension plans, but its inclusion in an overall Annual Report on Federal Public Sector Compensation would bring this important aspect of compensation into its proper context.

Other paid compensation

The Annual Report should summarize the value of such other paid compensation as allowances, overtime, severance pay, payments in lieu of leave, and pay equity payments. Any important changes in the level or nature of any of these components should be noted.

Insurance and health expenditures

Total expenditures in the areas of life insurance, disability insurance, and health and dental care plans should be reported. Because of its materiality, the per capita expenditure for the Public Service Health Care Plan (PSHCP) should be reported. Each year, details could be given on changes in one of the other plans, as well as more detail every few years on the components of PSHCP costs.

Contributions to statutory programs

Total spending in this area should be reported, along with the explanation of any material changes.

In accordance with the Government's general approach to financial reporting, the financial information in the proposed Annual Report should be presented on an accrual basis, with details on current expenditures as appropriate. This will facilitate meaningful comparisons between fiscal years.

The same information should be reported for groups whose salaries are determined without collective bargaining, including executives; Governor-in-Council appointees, including Crown corporation heads; other unrepresented employees such as Personnel Officers (PE); members of the Canadian Forces; regular and civilian members of the RCMP; federally appointed judges; and parliamentarians.

For employees, it would be useful to prepare a simplified version tailored to their own situation. Treasury Board Secretariat currently provides a very informative personalized annual report on their benefits. This could be augmented to report on the relevant employer contributions on employees' behalf, and on broader trends in federal compensation.

There is certainly a danger that the proposed annual report would invite selective reading and unjust attacks on federal public sector compensation levels and policies. All transparency entails the risk of misuse of the information provided. However, a full presentation that is updated annually is itself the best defence against unbalanced critiques. Accordingly, our first and most important recommendation is this:

Possible budgetary framework

Aside from reporting on compensation outcomes, the Government should consider expressing a broad view in the federal budget on the Government's intended spending track on total compensation as an important contribution to public debate.^[158] The expected increase or decrease in spending on compensation should be forecasted, as is done for other key spending categories. The amount presented should be the aggregate of increases expected because of changes in the level and composition of employment in the federal public sector, changes in salary levels for all purposes, and changes in spending on all other aspects of compensation from pensions to health insurance. 

At present, the Government does establish a Compensation Reserve within the fiscal framework.^[159] This Reserve does not provide for the additional costs of expanding the size of the federal public sector, only the cost of increased salaries or benefits. As statutory expenditures, pension costs are also managed separately from the Compensation Reserve. The amount included in the Reserve is kept secret on the grounds that negotiators should not show their hand in advance of concluding an agreement with the public sector unions. Even for unrepresented groups like the Canadian Forces or the regular and civilian members of the RCMP, it might be feared that it would distort the discussion between the Treasury Board Secretariat and the Department of National Defence and the RCMP for notional compensation allocations to be made public.

Of course, publishing details on expectations about negotiation outcomes would be foolish. What we suggest in this Report is a macro forecast of total changes in personnel spending, including the impact of changes in the number and types of people employed. While revealing such a forecast is not without risk, it has the following advantages:

• Implicitly, such an envelope emphasizes the point that there are trade-offs between hiring more people and paying the existing people more, as well as between salaries and other benefits.
• As such, setting forth what amounts to an intended compensation envelope will help to discipline the conversation about compensation.
• In order to establish and defend the forecast, the government will need to maintain a strategic capacity to manage compensation as a whole, as we propose later in this chapter.
• Most importantly, there will be a need for coherent coordination of the various processes that result in changes in compensation expenditures. 

The proposed forecast can contribute to an ongoing system of transparency and accountability regarding the government's management of compensation but it should not be seen as a rigid straitjacket on sensible management of changing conditions. The forecast can be compared to the results published in the Annual Report on Federal Compensation recommended in the previous section of this Chapter. There will normally be a sound rationale for the difference, for example, significantly different-than-expected inflation changes or labour market conditions, and this should be provided to the public. In effect, setting out a proposed compensation spending envelope and then reporting on the results obtained would create a planning and accountability cycle for federal public sector compensation.

As a result of this reasoning, we recommend:

Managing compensation data

As a practical matter, the main obstacle to producing a reliable Annual Report is the unsatisfactory state of data relating to federal compensation. Several issues became evident during the course of this study: inconsistent definitions, incompatible data and reporting, and incompleteness.

Definitional problems abound. Most notably, we wrestled with four major issues, including establishing public sector domains, interpreting components of compensation, timeframes, and reporting periods.

Composition of federal public sector domains

In this Report, we have distinguished six domains:

1. the core public service domain, including the main departments and agencies as listed in Part I of Schedule I of the Financial Administration Act, for which the Treasury Board is the employer;
2. the separate employers, as listed in Part II of the FAA Schedule I;
3. the members of the Canadian Forces;
4. the regular and civilian members of the RCMP;
5. the federal business enterprises and other Crown corporations; and
6. a residual domain comprising federally appointed judges, parliamentarians, employees of Parliament, ministerial staff and students.

We chose these categories because they correspond to important distinctions in governance, and consequently in compensation policies and practices. We found, however, that the exact list of what organizations fell into which domain was hard to confirm.

Interpretation of compensation components

While this Report covers all components of federal compensation, from salaries to overtime, allowances, pension contributions, insurance plans, health plans, and leave entitlements, it is difficult to report consistently on this basis. Detail on some domains is available in the Public Accounts system or in reports derived from payroll information produced by Public Works and Government Services. For the Canadian Forces and the RCMP, internal reporting systems are the sources of important details. All of this is summarized in reports by Statistics Canada. However, terminology and interpretations are not reliably consistent across these sources.

Similar issues arise in the area of reporting on the level of federal employment.

Dealing with the issue of estimation

One of the simplest questions turns out to be hugely complex. How much was spent on a given component, or in relation to a particular group, in a specific period of time? Similarly difficult is the question: How many people were employed in a particular group or organization either over a set period, or at a given time? The fundamental problem is that the public service is in constant flux. People are hired and resign every day. Some work part time. Salaries or benefits are paid, and then periodically corrected at various times covering various periods. The general practice of paying salaries every two weeks means that pay periods seldom correspond exactly to reporting periods such as months or years.

These examples are sufficient to illustrate why every compensation and employment aggregate number is in fact an estimate, based on some particular method of approximating the actual total. For example, one way to estimate the employee population over a given period is to take the populations at various points and average them. Even this is harder than it seems, in that at any given point people with the legal status of employee may fall into various categories such as working full time, working part time, on leave with pay, on sick leave, on leave without pay, and so on.

Such complexities require each number that must be estimated to be determined by a particular method of estimation. In this environment, what is most important is to define and apply consistently a reasonable estimation method.

Defining the reporting period

A final matter for standardization or at least greater clarity is that of reporting period. Some data is reported by fiscal year and some by calendar year. While differences between these time periods are normally small, and trends over time quite similar,^[160] it would be less confusing to use one reporting period, likely the fiscal year since that is the basis of government accounting.

Need to clarify responsibility

To a large degree, the issue of incompatible data and reporting arises from fragmented responsibilities for managing aspects of the compensation system. In the absence of a culture of collaboration, definitional confusion flourishes. There are numerous circumstances where more than one organization reports on what appears to be the same topic, but with conflicting results. A notable case that arose during the preparation of this Report related to reclassifications in the core public service domain. The Public Service Commission maintained data on this, based on its staffing reports; as well, the Human Resources Management Modernization Branch of the Public Service Human Resources Management Agency assembled information from its records on position classifications. Initially, the results were wildly different, varying for 2002–03 from about 4,800 to around 8,800. After some difficult joint work to reconcile their figures, the two groups concurred in blessing a total of nearly 6,700, as we reported in Chapter 4.^[161]

For all data elements of significance, it is critical to agree which organization is responsible to define the element, and to report consistently. Other organizations with an interest in a specific element should have a means to register their views with the lead agency, but should then incorporate the result in their own work.

In a real sense, incompleteness is also an issue rooted in definitional concerns. On various topics, we were unable to locate reliably similar data for different organizations.

In the absence of an agreed meaningful format for reporting on compensation levels, it is not surprising that some information was hard to locate. There are protocols governing the attribution of expenditures in the Public Accounts. However, these are designed with financial accounting in mind, rather than supporting an analysis and understanding of compensation issues. Once definitions are regularized to support the recommended Annual Report on Federal Compensation, the financial reporting systems feeding the Public Accounts would need to be aligned.

Reliable, consistent and complete data are indispensable to a realistic system of transparency and accountability in this as in any field. It is in the nature of complex data collection that we will always encounter errors and inconsistencies. What is essential is that we take responsible steps to minimize these risks, and provide a capacity for longitudinal tracking that can highlight issues of potential concern. Consequently, our third set of recommendations addresses this area:

Reporting on comparability

Positioning federal compensation within the larger Canadian labour market is not an easy topic, as should be evident from a review of Chapter 6 of this Volume. Despite the conceptual and practical challenges of assessing comparability, however, no system of transparency and accountability for federal compensation can be credible without addressing this perspective directly.

The Public Service Modernization Act, passed in November 2003,^[163] provides for a compensation research and analysis service to be created within the new Public Service Labour Relations Board (PSLRB). This service will "conduct compensation surveys, compile, analyze and make the information available to the parties,^[164] and in specific circumstances to the public, and conduct market-based compensation research." An Advisory Board will assist the PSLRB Chairperson in shaping the standards and analytical methods of the new service.

Logically, the new PSLRB service should be the main public source of comparability information relating to the core public service and unionized separate employers. Getting to the stage where this service functions satisfactorily, however, will not be easy.

The most difficult issue will be to find survey and analytical methods that are credible to both the federal employers and federal public service unions, as well as to external critics. Although it is usually said that the former Pay Research Bureau (PRB), which had a similar mandate, was abolished in 1992 for cost reduction reasons, oral history suggests that a lack of confidence in the objectivity of PRB's survey methods was at least as important in deciding to close the service. The PRB's policy to survey only large, unionized employers, for example, was viewed, by the employer at least, as skewing the results unduly.

Over the past several years, the Treasury Board Secretariat and the federal public service unions have worked together under the aegis of the Joint Compensation Advisory Committee of the National Joint Council to undertake a pilot project on comparability surveys. In this context, Statistics Canada was engaged to design a survey and to conduct pilot testing work in Ontario on behalf of the parties. Statistics Canada and the JCAC have started to analyze the initial results of this pilot survey with the objective of developing a methodology approved by all parties which could be used for national compensation surveys. The responsibility for pursuing this project has recently been transferred to the PSLRB. Until the project is completed, it is difficult to evaluate this experiment in collaboration. However, it is encouraging at least in that the parties were able to agree on an approach that satisfied Statistics Canada's professional standards for surveying.

Despite this positive start, no one should underestimate the difficulties of compiling comparability data routinely on a wide range of jobs across the whole country. The issues to resolve range from the political, such as deciding which categories of employers to select as comparators; to the methodological, such as how to select comparable jobs; to the practical, such as how to convince private sector employers to provide the needed data. None of these issues is easy. The challenges only multiply when the task is expanded––as it must be, if comparisons are to be truly useful––from comparing salaries and wages to a total compensation approach that would compare the full compensation package.

It is essential that federal employers^[165] bring their top leaders to the table for the decisions on the main policies and analytical methods of the new PSLRB compensation research service. Substantial analytical and conceptual energy needs to be applied to establishing a sound approach from the start. Tempting as it might seem to be to concede on disputed points in order to accelerate the process, it is better to struggle over concepts and methods at the beginning, rather than after the data has been compiled and reported. There may have been a tendency in the past to adopt an apparently pragmatic approach on various joint studies undertaken in regard to particular groups. As we described in Chapter 6, however, the resulting studies too often carried little credibility with the employer but appeared to have been endorsed by both parties. The resulting confusion is not conducive to effective negotiations.

That we can get where we need to go is beyond doubt, provided the parties focus the necessary attention and common sense. The example of the Institut de la Statistique du Québec (itself a descendant of sorts of the defunct federal Pay Research Bureau) shows that it is possible to create a credible and professional service. But getting both the government employer and unions to rely on the results has proven difficult for the Institut as well. It will be critical for the PSLRB to resist pressures from either party to "negotiate reality."

Based on the time required to develop and undertake the National Joint Council pilot surveys, we can expect it to take several years to phase in a suitable cycle of comparability reporting. In the meantime, there will be pressure to say something meaningful about how federal public sector compensation compares with that offered by other employers across the country. We can expect organizations such as the Canadian Federation of Independent Business to publish every five years their overview of compensation in the public sector versus the private sector, based on the quinquennial Census results. As we have shown in Chapter 6, such studies are so macro as to be of doubtful use, except to track apparent trends.

Rather than leave the field entirely to such private assessments, either the Treasury Board, or indeed the PSLRB, should consider publishing its own macro overview of comparability every two or three years, drawing on the data that is available. It could make sense for the Treasury Board, perhaps in partnership with the public service unions in the National Joint Council, to commission studies on aspects of comparability, for example in the areas of pensions, benefits and leave during the period before the PSLRB is ready to undertake them.

Our remarks here apply mainly to the core public service and the unionized separate employers. For non-unionized and unrepresented groups, it is important to conduct external comparative studies. The long-established comparative studies relating to executives should continue. However, it is timely to review the methodology to confirm its continuing soundness. In particular, care should be taken to see whether solid comparisons at levels above the first level of the Executive (EX) group would support salary increases for these employees. We comment further on executive compensation later in this section. Specialized non-unionized groups such as Personnel (PE) officers should be compared to the private sector on a regular cycle.

For other groups, it is equally important to publish comparability data. The existing RCMP Pay Council reports appear basically sound, although a methodology review would be sensible now that about a decade has passed since the first report. For the Canadian Forces, the existing total compensation comparison method with the core public service needs redesign. It is onerous without being convincing either for Canadian Forces management and members, or for the Treasury Board Secretariat. As we describe below in Recommendation 13.4, a Canadian Forces Compensation Advisory Committee could assist the Treasury Board in deciding on fair compensation for our Armed Forces. An early task of the new panel could be to recommend a streamlined and credible method for assessing the comparability of military pay to the core public service, since direct comparisons to the external labour market seem no more likely to be compelling now than they were in the past.

Before concluding with this topic, however, we need to consider two further dimensions of comparability that are both controversial and conceptually difficult to incorporate into an assessment of comparability. The first topic is relative job security. The general impression is that public servants are essentially set for life once they become indeterminate employees, but that private sector employees are subject to losing their job with little warning or recourse. Both aspects of this picture are exaggerated, but still apparently more true than not.

In the federal public service, the Program Review staff reductions in the mid 1990s led to many thousands of unplanned early retirements and other job losses. However, the approaches employed––the Early Retirement Incentive and the Early Departure Incentive described in Volume 2––to implement the job cuts were generous both in financial terms, and in permitting those who wanted to leave to substitute for those whose positions were being eliminated, but who wanted to stay. It is also true that it is relatively rare for public servants to be fired, with the greatest number of involuntary departures resulting from dismissal while on probation.^[166]

In the private sector, job security varies from weak––merely the few weeks' notice of lay-off required by labour standards legislation––to strong, depending on the employer. Unionized employees usually have the greatest job security, and the best compensation in the case of lay-off. But even the best union contracts generally do permit lay-offs, and the media regularly report hundreds or even thousands losing their jobs as plants close.

The second topic is that of the relative attractiveness of public service jobs. In theory, this can be measured in two ways: how many people apply for job vacancies open to the general public, and the rate of voluntary departures. On both points the federal public service appears to fare very well compared to the private sector. Externally advertised positions in the federal public service often attract hundreds of applications, although the proportion that is qualified has not so far been documented systematically. The rate of attrition (not including retirements) was low at about 1.5% in 2002-03, for example. Voluntary departures from private employers are of course highly variable, but it has been suggested to be, in the aggregate, in the range of 10%-15%.^[167]

Both relative job security and relative attractiveness of an employer have value to employees, even if they are difficult to quantify. At a minimum, the overall position of the federal public service in these two areas compared with the Canadian private sector and the broader public sector needs to be documented and reported. If the assessment presented here to the effect that the federal public service enjoys a relatively favourable position in these areas is accurate, this implies that federal salaries and benefits for the same work could reasonably be somewhat lower than in the private sector. This would of course be very difficult to negotiate thorough collective bargaining, but it needs at least to be discussed frankly and openly.

In summary, our recommendations on reporting in regard to comparability are these:

This first set of four recommendations is both the most important and the most feasible to implement. Reporting annually on key trends in federal public sector compensation, ensuring that the underlying data reflects clear definitions and estimation methods, and providing analytically sound information on how the federal public sector compares with the larger Canadian labour market will provide everyone with the context needed to manage federal compensation properly. However, implementing our recommendations on transparency and accountability implies, indeed requires, changes to how we manage compensation in the federal public sector. Our next set of recommendations addresses this topic.

 

• Previous Page
• Table of Contents
• Next Page