Statistics Division Home
Fundamental Principles
Country Practices
Reference Materials and Links
Sign in

A Peer Review of the Swiss Statistical System

A Peer Review of the Swiss Statistical System

Neuchâtel, 2000

Published by: Swiss Federal Statistical Office (SFSO) © SFSO, 2000
Information: Carlo Malaguerra, Director General SFSO, ’Phone 032 713 60 01

Authors: Ivan P. Fellegi, Chief Statistician of Canada
Jacob Ryten, Former Assistant Chief Statistician of Canada

Obtainable from:
Swiss Federal Statistical Office
CH-2010 Neuchâtel
’Phone 032 713 60 60 / Fax 032 713 60 61
Order number: 379-0000
Price: Sfr. 50.

Series: Swiss Statistics
Field: 0 General topics
Original language: English
Translation: Statistics Canada and SFSO Translation Service
Cover picture: Prisma/Coll
Copyright: SFSO, Neuchâtel 2000
Reproduction with mention of source authorized
(except for commercial purposes)
ISBN: 3-303-00211-8


Introduction ----------------------------------------------------- 5
In lieu of an executive summary ----------------------- 7
I Setting the stage ------------------------------------------ 10
II Facts and Observations ------------------------------- 14
III Perceptions ------------------------------------------------- 36
IV Recommendations -------------------------------------- 39

Annex 1: List of Interviewees -------------------------- 45
Annex 2: Questions to External Interviewees -- 47
Annex 3: Questions to Internal Interviewees --- 51
Annex 4: Peer Review Method ---------------------- 55
Annex 5: Organigramm ------------------------------------ 60
Curriculum Vitae: Ivan P. Fellegi, Jacob Ryten - 61


Some time ago the Swiss Federal Statistical Office (SFSO) embarked on a course of change aimed at enabling it to clarify its image so as to improve its chances of success in facing the new challenges arising from the developments and demands of our society. This large-scale operation will inevitably have major repercussions on the relations between our Office and its partners, and it will entail a restructuring of the whole current Swiss statistical system. These ambitious goals which we are striving for are clearly stated in the Multi-Year Programme for Federal Statistics, 1999 to 2003, which has just been approved by the Federal Council and has now been passed on to the Federal Parliament.

In particular, the implementation of the strategic elements of this change must be based on a systematic examination of the mission of the organisations involved and the way existing structures function. During this present crucial phase, we considered it essential to draw on the experience and expertise of authoritative institutions and exemplary individuals in international public statistics. In this respect, Statistics Canada and its Managing Director, Mr Ivan P. Fellegi, stand out as pioneers. After specifying the principles and standards on which a modern statistical system should be based, Mr Fellegi and his management team have set up a highly efficient organisation and structure which have served as an inspiration to many other countries.

Against this background, the SFSO decided to commission an evaluation, or Peer Review, to be carried out jointly by Mr Ivan P. Fellegi and a former senior executive of Statistics Canada, Mr Jacob Ryten, who is currently a statistics consultant.

We set two primary goals for this commission:
- to identify and describe the strengths and weaknesses of the current Swiss statistical system and the SFSO in particular
- to draw up proposals and recommendations aimed at improving the situation.

We also agreed with the Canadian experts that their evaluation would be made public.

These aims have now been achieved. Anyone who has any interest in the functioning of Swiss public statistics can therefore consult this publication to find an extensive range of useful information. For its part, the SFSO management will adopt the experts’ recommendations. Within the context of the structural reforms concerning the SFSO, and hence the Swiss statistical system as a whole, it will implement those recommendations which lie within its responsibility, and will subject to political consideration those which affect the whole legal framework currently governing Swiss public statistics.

As far as we know, this is the first time in Switzerland - or indeed in the world - that a Peer Review has been voluntarily requested by the management of the statistical institution in question. However, we regard an undertaking of this sort as essential for any country. Indeed, because of their specific features, national statistical institutions (NSI) are organisations which can only be compared to one another; it is impossible to compare them to any other unit of public administration within the country, although this practice is unfortunately a frequent one. This applies all the more now that the international dimension of statistical information has become a reality. We have, therefore, taken a pioneering step and hope that others will follow our example.

We would like to take this opportunity of offering our very sincere thanks to Mr Ivan P. Fellegi and Mr Jacob Ryten of Statistics Canada for their immense efforts, their perspicacity and their availability. Their evaluation of the SFSO and the statistical system constitutes one of the essential foundations for the development of Swiss public statistics in the future. We would also like to thank everyone who has provided support and assistance to our Canadian experts as they worked on their evaluation.

Carlo Malaguerra
Director of the Swiss Federal Statistical Office
Neuchâtel, June 2000

In lieu of an executive summary

This report is about the Office Fédéral de la Statistique (OFS) in its role as the centrepiece, leader, and coordinator of the Swiss statistical system. It was written by two outsiders to the office, system, and indeed, country. The conditions associated with the drafting of the report were that it be public, totally independent, and especially addressed to a group of concerned and influential users of statistics. The method adopted to collect, collate and analyze the necessary information was to conduct a large number of intensive interviews, both of insiders and outsiders to the system. When boiled down, three questions were asked. In attempting to answer them findings were made that in turn gave rise to recommendations for further action. Questions, findings and recommendations are spelled out below. As is always the case with social institutions and activities, questions, findings and recommendations are intimately related, so much so that it would take daring to suggest that any one answer corresponds exclusively to any one question.

The questions are general. They apply to any national statistical system or component thereof:
- How adaptable is the system in adjusting to evolving needs?
- How effective is the system in meeting existing client needs?
- How credible is the system in terms of quality and objectivity?

Findings and recommendations are as follows:

1. The OFS is a sound organization. It is professionally competent, and in spite of evidence of occasional inappropriate political interference, scrupulous in its observance of the ethics of official statistics. We found it willing and eager to produce and interpret what is relevant and useful to the community’s decision makers.

2. Statistical organizations have a number of unique features. In order to be useful they must be credible and in order to be credible they require a very large measure of substantive and apparent independence. An apparent or substantive lack of autonomy implies a vulnerability to interference with domains that should be the exclusive prerogative of the Government Statistician, which in turn could undermine in the most serious way his enterprise, the objectivity of the staff, and the acceptance of their output by the public at large.

- We submit that the OFS, as one of several directorates within a Ministry of the Confederation, does not have enough autonomy. We recommend it should have more through a change in the character of the institution and of the law. We seized on the notion of having the OFS become an institute located in the third Kreis. The intention should be clear: take on the status that gives the Office the greatest degree of administrative and substantive independence subject to the realization that a substantial majority of its budget -- though certainly not all of it - will always have to come from the Confederation. Even if in accepting a new status, there were a demand for an initial modest reduction in resources it is our conviction that such a reduction be accepted.

We contend that the adoption of this recommendation is the most effective way of answering the third of our questions.

3. The activities of statistical agencies reflect choices among different alternatives according to a notion of social priorities and an assessment of efficient means of satisfying them. There is no known way to choose among competing activities in a purely objective manner. But the choices must be visible as must be the weighing of different alternatives and the discussion that led to the ultimate choice.

- We submit that the OFS, in spite of the undeniable progress it has made to date, requires a considerable strengthening in the methods, scope, visibility and transparency involved in the framing of its multi year plan. Our full report makes specific recommendations in this regard.

- We argue that the instruments required to make the planning activities more effective -- flexibility in moving and supplementing resources -- can be more easily accommodated within a more independent administrative status.

- We recommend that the agenda, weight accorded to, and reach of the deliberations of the Statistical Commission be expanded pari passu so as to provide the Director General with the advice and support needed to make the best possible priority choices.

We reckon that in acquiring these capabilities, the Swiss statistical system will be far better equipped to answer the first of our questions

4. For official statistics about a country’s economy, society, and environment to be effective they must be as interlinked as the subjects they measure. Which is why tradition, technical literature and ultimately the application of commonsense suggest that official statistics constitute a well-integrated body. Moreover, for statistics to be truly effective they must be communicated to their users not as the simple result of counting, sampling or balancing accounting tables, but as the intelligent sum of insights that results from quantifying and relating factors and outcomes and bringing out the unexpected as much as the predictable.

- We advise in the strongest possible terms that the OFS take active steps to improve its analytical capabilities using the most effective means at its disposal, be they training of staff or partnering with research institutions, in a regular rather than occasional enterprise. We further advise that this analytical capability be deployed systematically, that it extend to all areas served by the OFS and that the results become a model for all other members of the Swiss statistical system.

We believe that only those offices that are so equipped can truly acquit themselves when attempting to answer the second of our questions.

5. Switzerland -- like Canada -- is a confederation of members that historically have enjoyed a great deal of autonomy and who have legitimate needs for statistical information in order to pursue local economic and social policies. But the efficient development of statistical information does not know the boundaries that separate local from central government. For this reason we suggest that the organization recommended in 2. above be a federal-cantonal entity, charged with a number of substantive statistical responsibilities including collection and analysis so that the joint enterprise does not become too one-sided.

We recognize that this is general advice, which is seldom practical advice. The latter requires means, institutions, mechanisms, and codified practices. We have provided suggestions concerning all of them. We have supplemented that advice with indications on how to strengthen the cohesion of the OFS. The precise nature of what we advocate is recorded in the pages that follow.

I Setting the stage

1. Introductory

There are three kinds of critical reviews of a statistical agency. The first is ordered, directly or indirectly, by the country’s elected representatives usually because their concern has been triggered by instances of incoherence or inadequacy of the statistics. The Boskin Commission’s work on the Consumer Price Index of the United States is an example; the Moser Commission’s work on the Italian National Accounts is another. There are those reviews, which the law or its regulations prescribe at regular intervals and of which there are several examples in the OECD area. And there is a third category of which until recently no example could be found. This is an instance where there are no programme or legal obligations to conduct a review; where no politician or political body has expressed the desire or need to conduct such a review; and yet the Director General of a statistical
agency of his own accord decides that there are good reasons why such a review would be both healthy and fruitful.

The critical review of the Swiss Statistical System and more precisely of the Office Federal de la Statistique is an example -- the example -- of the last of the three. Hopefully it will not be the last of the kind. For it is far better to certify that there is no ailment than to acknowledge that the illness is terminal. The Swiss statistical system is not sick. Like all systems and institutions, there are improvements that can be made to its performance and some should even be undertaken with a sense of urgency. But structurally and basically the system shows integrity, vitality and adaptability, the three most important attributes if objective, quantitative information is to be made available to the Swiss body politic and to Swiss society.

2. Definitions, standards, and methods as well as the questions, to which this report purports to hold answers

2.1 The agents

In this report references are made to the following agents:

- The statistical system (or the system for short): involving all parts of the public sector -- Federal,cantonal, and municipal -- that have an explicit statistical mission resulting in the publication of official statistics with the customary attributes of objectivity, neutrality and impartiality;
- The Office Federal de la Statistique (OFS for short): the key institution -- responsible for the substantial majority of total official statistical output within the statistical system and the one that is being reviewed most directly;
- The Director General (who in a different setting would be called the Befehlshaber): the incumbent from whom is demanded accountability for the efficiency, effectiveness and responsiveness of the Swiss statistical system and to whom somewhat limited powers of decision and coordination are accorded;
- The Minister Responsible (or Minister for short): the elected representative and member of the Federal Council to whom the political responsibility for the conduct of the OFS is given;
- The stakeholders: those persons or institutions who have most to gain from an efficient, effective, responsive and creative statistical system, one capable of bringing to the forefront of public discussion the most relevant information. The stakeholders include public sector decision takers, researchers, legislators, international or supranational bodies whose respective mission requires a close watch of Swiss society and its economic performance, and of course the Swiss public at large as in any other socially advanced and democratic society;
- The Reviewers: the two authors of this review in their quality of collectors of information on the Swiss statistical system, appraisers of the adequacy of its institutions and advisers on what measures could be taken to increase the quality of its overall performance.

2.2 The standards

While any infringement of the Fundamental Principles of Official Statistics would be considered as a major blot in the OFS’s copybook, the code is only used as a background rather than as an operational yardstick to assess the quality of the Swiss statistical office’s performance. In practice, and at the request of the Director General, the conceptual framework for assessment was provided by a text authored by one of the Reviewers (I. P. Fellegi: «Characteristics of an Effective Statistical System» (Footnote: Originally delivered as the Morris Hansen Lecture, Washington Statistical Society, October 1995. The full text of the lecture can be obtained from either the OFS or from Statistics Canada.) heretofore referred to as «Effective System»).

«Effective System» asks three fundamental questions against which the performance of the Canadian statistical office is measured:

- How adaptable is the system in adjusting ... to evolving needs?
- How effective is the system in [meeting existing] client needs?
- How credible is the system in terms of ... quality and ... objectivity?

This review of the Swiss statistical system proposes to answer the three questions largely by an assessment of the following elements:

- The solidity of the legal and institutional environment (the law, the statistics council, the institutional setting, the position of the Director General);
- The trustworthiness of the quality attributes that accompany the OFS’s products (sound survey methodology, absolute confidentiality protection afforded to individual data, respect for the privacy of respondents);
- The masse de manoeuvre at the disposal of the OFS’s Director General (budgetary authority, personnel mobility, access to sources of authority and information, analytic feedback); and
- The adequacy of the instruments developed by the OFS in the pursuit of more effective management and better client service (information on costs, transparent planning process, analytic outputs, advisory committees, access given to users). The headings mentioned in the paragraph above are taken from «Effective System».

2.3 The Questions

The section below elaborates on the questions that concerned the Reviewers under each of the elements.

1. Solidity of legal and institutional environment

- Does the legal basis provided to the OFS make it easier to collect and compile information that is objective and relevant to the concerns expressed by all levels of Swiss society? Where such information is not collected by the OFS, does the legal basis provide a sound basis to allow it to function as chief statistical coordinator of the Swiss statistical system? Does the coordination function extend to coordination between Federal and cantonal activities?
- Is the Statistics Commission an effective body to assist the Director General in determining priorities, coordinating activities within different parts and levels of Government (Federal Ministries and cantonal authorities)?

2. Trustworthiness and completeness of the OFS’s (and of the rest of the system’s) output

- Are the OFS’s statistics believable? Are there instances of undue interference by non-statistical authorities in the production of the statistics and the form of their release?
- How does the statistical output of Switzerland compare to that of other countries of similar size and at a similar level of development?
- Does the OFS treat the documents in its custody with absolute confidentiality protection? Does the OFS show respect for the respondents’ privacy?

3. The tools at the Disposal of the Director General

- Can the Director General move his staff around in either a planned or an ad hoc form in order to better respond to shifts in priorities? Can the Director General order a realignment of the budget as an alternative means to respond to changes in the ranking of priorities?
- Does the Director General have access to those circles or persons from whom he can best glean the direction in which public policy issues are moving?

4. The adequacy of the instruments developed by the OFS

- Does the Director General have sufficient information to assess the OFS’s cost structure? Do staff members realize the cost of their statistical activities? Has there ever been an organization-wide drive to lower cost by increasing productivity?
- Does the staff understand how plans and priorities are made? Is there an advisory network that helps make those plans?
- Can the management of the OFS rally the staff, inspire, and lead them confident that the human resource policies in place provide the assurance and motivation required?

2.4 The Method

The Reviewers relied on the following sources of information -- extensive documentation prepared by the OFS to assist them in the review; a series of eleven interviews with key users of statistics (see list of names and institutional affiliation in Annex I); and sixteen interviews (the names of the interviewees are also included in Annex I) with officers at OFS ranging from the Director General to the majority of the Section Chiefs (two levels down from the DG). (Footnote 1) The reader is referred to Annexes II and III for a list of the questions that were asked from each group of interviewees.

Footnote (1): A detailed description of the method adopted was provided to the 48th Plenary Session of the United Nations Conference of European Statisticians as a paper transmitted by the Swiss Federal Statistical Office: Peer Review as an Essential Part of the Restructuring of National Statistical Services --Switzerland’s Experience. The relevant part of the paper is reproduced in as Appendix to the present Review.

The drift of the questions asked from the external interviewees was to establish a picture of the system’s adaptability (to new circumstances, changing needs and interests and so on) and effectiveness (capacity to meet current demand). This involved discussing the law, the Statistics Commission, the form in which priorities are decided upon, and the apparent effectiveness of the Director General.

II Facts and Observations

1. Standing

1.1 The OFS in an international context

Place Switzerland and the OFS on an international comparative scale. Switzerland falls within a group of OECD countries of approximately the same population (Austria and Sweden all with more than 7 and less than 9 million inhabitants) and GDP per capita (between 23 and 28 thousand dollars after correction for purchasing power disparities). At the lower end of the group there is Denmark, Finland and Norway all close to 5 million inhabitants and also with similar GDPs per capita. At the high end of the population scale there is Belgium Hungary, Greece, Portugal and the Czech Republic with ten million or so each but, with the exception of Belgium, much lower GDPs per capita.

All these countries with the exception of Belgium have statistical offices with 500 to 1000 permanent employees. Most of Belgium’s economic statistics are compiled by the Central Bank. While some of the other offices are considerably larger than the Swiss, most are droipping in size. The experience of the Swiss office is the opposite. At 470 or so authorized posts it has more than doubled in size over the last fifteen years.(Footnote 2)

Footnote (2): Note that a considerable part of this growth was achieved through transfers from other directorates, whose statistical functions were moved over to the OFS.

In the end we took as our comparators the six countries (Footnote 3) that were closest to Switzerland combining GDP per capita and population.

Footnote (3) Austria, Belgium, Denmark, Finland, Norway, and Sweden

More precise comparisons should not be attempted unless they are carried out in the greatest of detail. Firstly, none of these offices have strictly speaking the same agenda. Some engage in a bigger variety of undertakings than others simply because there are fewer institutions outside the statistical office with the power, budget, and ability to undertake statistical surveys. Others -- the Swiss may well constitute the outstanding example -- do comparatively little field work out of their own work force and prefer to commission specialized marketing companies to have them look after the administration of questionnaires and the initial phases of data collection. And still others -- principally the Danes -- have the tradition, the political green light, and the necessary methods to make extensive use of administrative files and even to link them in ways which would not be acceptable everywhere. Whatever the merits of such practices they explain some of the differences in staff and in product range among seemingly similar statistical offices.

There are two relatively easy ways of making crude comparisons between the variety and detail of statistics produced in Switzerland and in the comparator countries. One is by placing side by side the Statistical yearbooks of the countries concerned. The other is to select an international statistical bulletin that covers a wide variety of statistics and examine how the Swiss page fares relatively to the others’.

1. 2 Results of a comparison of Yearbooks

A cursory comparison using Sweden as a comparator suggests that with the exception of a handful of domains in which we know Switzerland to be particularly interested -- tourism, agriculture -- the Swedish statistical output exceeds the Swiss by a margin in terms of detail and variety of subjects broached. While the comparison with the other countries improves the results for Switzerland the overall impression is that on matters of structural data there is some catching up in order to reach comparable levels.

1.3 Results of a comparison of country pages in the OECD’s Main Economic Indicators

Switzerland’s performance as measured by the frequency and apparent detail of its statistics of short-term economic change as featured by the OECD is not outstanding. It would appear that on the whole and with the exception of the financial series where Switzerland has a lead, in certain key instances the other economic statistics are not as developed as those for say Austria or Sweden. In fact we tried out a smallscale experiment. We compared the country pages in the OECD’s monthly Main Economic Indicators (the organization’s flagship statistical publication) for Austria, Belgium, Denmark, Finland, Norway, Sweden and Switzerland. The purpose was to find out what were the most important gaps in the Swiss product line relatively to similar countries and how well Switzerland did in terms of timeliness. Using arbitrary weights (two negative points for every month behind the month of publication of the fastest country in the sample for a particular subject; one negative point for quarterly series where the majority had monthly versions; and forty eight negative points for each subject for which a particular country did not have at least one short term indicator), the following were the results:

Table 1. Ranking of comparator countries according to the completeness and timeliness of their main economic indicators

1. Sweden -------13
2. Austria ---------14
3. Finland ---------68
4. Switzerland --69
5. Denmark ------70
6. Belgium --------71
7. Norway --------75

Source: OECD, Main Economic Indicators December 1999

The subjects compared included the quarterly national accounts, the index of industrial production, selected indicators of manufacturing performance (typically new orders), selected indicators of construction activity (permits issued), retail trade, the rate of unemployment, labour compensation, job vacancies, consumer prices and foreign trade.

Our reaction to the line up of Swiss short-term economic statistics is twofold. Firstly, the supply side is more developed than the demand side. This does not seem to us a natural development from the point of view of support to policymakers. The reasons are that demand (final consumption at home and abroad, purchases of machinery and equipment by the domestic corporate sector, government expenditure on goods and services) is notoriously more volatile and more responsive to borderline changes in tax rates or in rates of interest. Accordingly, up to date and detailed information on the demand structure should be a pressing need for policymakers engaged in advising on monetary and fiscal conditions. (Footnote 4)

Footnote (4): A number of key users assured us that their overwhelming concern was to get as advanced a set of demand indicators on consumption and investment as feasible.

Secondly, the labour market seems to have less than its normal complement of statistics. Thus, the labour force survey at this stage does not produce monthly or even quarterly statistics; and there are no short-term statistics on changes in the cost of labour in spite of the attention devoted by policymakers to price inflation.

We are in no position to comment on matters of reliability except for the fact that there were no major doubts emitted by the external interviewees on the soundness of the figures (Footnote 5) . But we did find out that at this stage, even if the OFS, and in general the Swiss statistical system, were commanded to do so, it could not comply fully with the minimum European list of short-term economic statistics required from all members of the Union.

Footnote (5): As is the case with most statistical offices there is folklore in the memory of experienced users that includes examples of embarrassing errors, revisions and so on. We did not think that there was anything systemic or particularly noteworthy that would make the OFS stand out.

By and large, our interviewees were explicit on the fact that they did not consider it advisable for Switzerland to fall behind EU countries in matters of statistical availability -- variety of series and timeliness thereof. We can understand the concern. As capital markets are successively liberalized and venture capital becomes increasingly mobile no one country deems it prudent to have less information on its economic conditions than its neighbours whether or not it is attracting foreign capital at any particular moment.

But notwithstanding the current state of affairs, the OFS has come a long way. Twenty years ago, the OFS had a much smaller range of outputs. Today, it accounts for a substantial majority of total printed Swiss statistical output and its range of statistics -- using the same sources -- is far wider than it ever used to be.

2. General characteristics of an effective system: analytical output, relations to other levels of government; the multi year plan

2.1 Nature of statistical output

We reviewed a sample of the OFS output -- both of the paper and electronic varieties. We were favourably impressed by the accessibility, professional presentation, and the accompanying documentation on methods. We ascertained that the publication of key statistics was preceded by press releases, which attempted to highlight the major findings. We also gained a clear impression that the output of the Swiss statistical system was based on sound and modern statistical approaches.

But there is one significant reservation that we wish to express under the heading of «nature of output». We found little evidence of a strong analytic presence. Indeed this could reflect the fact that in building the professional competence of OFS staff there has not yet been a conscious emphasis to establish a capable analytic facility. In our view, as outlined below, this is a shortcoming largely because an analytically capable statistical agency can tap into a unique source of strength.

Analytic capacity in the statistical office contributes to the effectiveness of the office in the following ways:

- Analytic outputs help the public understand the significance of statistical findings
- indeed shed light on public issues in an objective manner;
- The ability to shed light on issues immeasurably enhances the public profile of the statistical agency as a key source of relevant information;
- A high public profile contributes to the statistical agency’s ability to successfully defend itself against improper political interference;
- Analytic work carried out inside the statistical agency gives rise to an acute awareness of users’ perspectives. Such awareness, in turn supplements importantly the outcomes of formal consultations. As a result, internal analysts become important «allies» of external user groups;
- Analysts working within the statistical agency make major conceptual contributions to the development of new statistical survey instruments; and
- Analytic personnel represent a natural link to key client communities- policy analysts working in other departments (at the federal or cantonal levels) and the academic community.

Note on the analytic role of the quarterly national accounts

The reviewers were surprised at the apparent soft-pedalling of the quarterly estimates of the national accounts. Normally the accounting framework is the analytic framework that integrates all economic statistics of short-term change. Indeed, no major OECD country uses alternative frameworks. There are several reasons why it is important that the accounting framework be used and a number of them affect mostly the supply side of the statistics:

- It is ideally suited to judge the relative reliability of the basic statistics that go into the making of the accounting estimates and to make judgments about the opportunity and the need to fill in missing detail;
- There is no equally well-suited means to discipline efforts made to improve the timeliness of collection and processing. In Canada the first great step forward in creating a system of basic economic statistics was taken by subordinating a timeliness drive to the needs of timely quarterly national accounting estimates;
- There is no comparable means to ensure system consistency among basic statistics than to force them to confirm to a system of accounting identities (for example, the trivial notion that the change in the output of the Swiss watch making industry should equal the change in consumption of said output plus the change in imports minus the change in exports plus the change in stocks -- assuming of course that the watch making industry only produces
watches is extremely powerful to examine the reliability of production, trade and stock statistics);
- It is the most effective way of providing the compilers of basic economic statistics with feedback on the reliability of their estimates because they can examine the behaviour of their statistics once placed in the framework of the national accounting identities.

These arguments suggest that analysis of short term economic statistics is necessary but cannot proceed without certain perquisites and that there are interesting spin-offs to the supply side even if initially, serious users do not use these estimates fully.

2.2 Institutional arrangements regarding other levels of government

Cantons and communes play a major role in the Swiss system of government. Little needs to be said here to underline their importance. And yet, their de facto statistical capacity varies enormously: some cantons have no statistical office at all, others have some vestigial functions, and yet others are hosts to significant statistical operations. Furthermore, there is no formally legislated or negotiated statement of the respective roles and responsibilities of the federal and cantonal statistical offices, nor are there legally binding tools to harmonize their work.

The apparent gap in defining and legislating statistical roles and responsibilities has very significant implications for the design of an effective statistical system. On the one hand, substantial weight must be given to the cantonal authorities’ needs for statistical information. On the other hand, by the very nature of their authority, cantons have developed administrative records, which can take on central importance for the overall statistical system. This is certainly the case in respect of taxation data, as well as with health and education related information. The former, suitably exploited, can both enhance the reliability and detail of existing statistical outputs, as well as contribute to significant reductions in reporting burdens. The latter form the basis of a nationally comparable information system on the basis of which cantons can assess the relative effectiveness of their respective arrangements for the delivery of health and education services.

Under the current Director General, there has been a conscious recognition of the need for complementarity between the federal and cantonal statistical offices. The OFS has initiated a regular series of meetings with cantonal offices, sharing of plans, and coordination of work. However, the OFS has no formal means to influence the work of these offices, cannot help create offices in cantons that do not already have them, and has no direct means to respond to the statistical information needs of cantons.

2.3 Functional requirements of a statistical system

For a statistical system to be recognized as a «system» it must be endowed with a number of key characteristics. These include:

- Effective means of informing itself about the statistical needs of its clients;
- Capacity to «filter» those needs and interpret them so that they can be placed in an operational setting and assigned a priority rank;
- Ability to effectively formulate plans to respond to those needs that are judged to be of key priority; and
- Means of adapting to unforeseen requirements.

While the OFS has significant capabilities related to all these characteristics, some notable shortcomings subsist. They include:

Means of gathering information on client needs.

It is crucial to the effectiveness of a statistical system to run a systematic programme of liaison with each major client group. Thus:

- With respect to the federal client departments, there is an annual visit to them by the Director General, but we believe this is insufficient for the establishment of a solid relationship. It should probably be supplemented by an organized programme of regular bilateral meetings to explore each other’s emerging programmes and recognize existing constraints;
Status and standing

For solid bilateral relationships to develop it is preferable if both sides are approximately equal in terms of their standing within the government bureaucracy. While the Reviewers are not in a position to engage in a careful examination whether this is indeed the case, what rough evidence they have suggests that it is not. For example, a simple comparison of budget per capita does not show the OFS in a particularly comfortable position. Very few departments appear to have a proportionately lower budget. The Reviewers recognize that without detailed figures and more time than is available to allocate to the subject, the budgets compared are strictly speaking not comparable. Some include significant sums of grant money or sums earmarked for capital purchases and accordingly show distorted ratios of budget to personnel. It nevertheless caught their attention that only ministerial secretariats (which one presumes would consist mostly of salaries), the computer services office and the bureau of price competitiveness compare in budget per capita to the OFS. Given the latter’s substantial non-salary expenditures (even if it does not conduct surveys, it needs to pay the survey services purchased), it would appear that the salary levels of the OFS’s staff are lower than average. If this is indeed the case, the development of what would be regarded as a truly useful relationship may require some rebalancing if it is not to get off to a poor start. (Footnote 6)

Footnote (6) The point about salary comparability appears to be systemic and caught the eye of the United Nations statistical establishment some twenty years ago as a structural weakness which concerned governments should attempt to remedy (see United Nations: Handbook of Statistical Organization, New York 1979)

- With respect to the cantonal authorities, systematic meetings with senior policy
makers (jointly with the federal authorities, where applicable) in each of the
areas of primary cantonal jurisdiction might be considered. Such meetings work
well in Canada in such areas as health, education, and justice statistics; and
- With respect to the academic community a small core of substantive analysts
within the statistical office should be given the task of ensuring permanent

Capacity to filter client needs.

Once there is a comprehensive system of «listening posts», (antennes in the French context) it is essential to have an effective ability for setting priorities. While there is no magic formula to achieve this task, some insights are offered.

Firstly, there is a need for a planning system, which brings together the «signals» received. We have examined the multi-year plan for the Swiss Statistical System but were left with a number of questions. For example, it did not provide us with evidence that the full range of client needs were systematically explored nor that there was a formal process of considering their respective merits. We do not advocate, nor do we believe it possible, to spell out the criteria of choice particularly if the choices are to be applied mechanically. But we do believe that a formal exploration of the costs and benefits of alternative courses of action must take place -- as an input to what is, unavoidably and ultimately, a decision taken on subjective grounds.

Secondly, there is a need for a broadly based Statistical Commission of «eminent people» (Commission de notables), actively engaged in providing social feedback to the OFS on its plans, discussing social developments, and how those might impact on the statistical system. We do not necessarily recommend a decision making body, but rather one whose moral influence makes itself felt largely as a result of the weight of authority of its individual members. It is, of course, essential for such a commission to engage the interests of its members. Unless that happens, attendance at its meetings will be poor -- as indeed, we are told, is often the case with the current Statistical Commission. This requires careful preparation, attention paid to the choice of agenda, as well as evidence that the Commission’s views are treated thoughtfully and respectfully. A Commission structured along those lines can become the arch-defender of the integrity of the statistical system and of its immunity to political interference.

Thirdly, the statistical agency itself must be engaged in judging emerging governmental priorities and broad social needs. This is achieved through active participation by agency staff in senior federal and cantonal committees, and by a conscious and unceasing effort to assess (through formal programme evaluations and through an on-going analytic programme) the relevance of its outputs.

Fourthly, while the OFS makes use of advisory committees of external (Swiss) experts, the practice does not yet extend to all subjects (7) . We suggest that the OFS should examine which additional areas could benefit from regular external expert advice (on both methods and priorities).

Footnote (7): We do notice though that when used, the level of intervention is very high as in the case of the expertise on the Swiss Consumer Price Index and the so-called Boskin bias.

Ability to formulate plans.

Paraphrasing what one of us wrote at a very early stage in the fact-finding part of the mission:

«It would appear that the OFS has much to learn before it becomes truly client oriented. [In spite of the multi-year plan] the setting of priorities is still far from transparent [and until it becomes much more so, it will be difficult for the implicit priorities to be widely accepted]. Nor are the choices made associated with detailed cost estimates (comptabilité analytique)- at least not in the mind of influential observers and stakeholders. Accordingly there is an impression, which could be right or wrong but is not substantiated by factual evidence, that the OFS could expand its output if only its internal resources were reallocated more efficiently. Until this impression is either dispelled or acted upon the OFS will not be ready to engage in the kind of budget supplementing we believe it needs. Actually some government agencies would be eager to engage in joint venturing and indeed have the money and the will to do so.»

«Client orientation», in the sense of interest in and concern with specific needs of specific clients or client groups, is not yet entrenched in the OFS’s culture. This was asserted quite explicitly in the course of one of our internal interviews. We would argue that being conscious of this gap is a major step forward along the path to improvement. It has been argued by heads of statistical agencies that the notion of «client orientation» is the single most powerful agent of change in the wake of which all other necessary changes (budgetary, human resource policy etc.) can take place.

Today, some twenty interviews and four months later there is not much we would modify in those early remarks. In summary, we see the following shortcomings:

1. The OFS does not yet have a transparent procedure for the consideration of priorities;
2. The OFS does not yet have the tools that would allow its Director General to move people and financial resources from one application to another flexibly but explicitly; and
3. Possibly because the major tools for adaptation are missing, the OFS is not yet in a position either through the Statistical Commission or through intensive bilateral contacts to maintain the kind of dialogue that well informed, helpful and supportive stakeholders require.

If the right conditions were present there would be a better focus to key users’ critical comments. In fact they might start arguing not so much about matters of choice, allocation, institutional effectiveness and so on but rather whether the envelope of resources that the government puts aside for the development of statistical information is sufficient.

One of the most important consequences of planning transparency is that it fosters greater goodwill on the part of enlightened users to bolster OFS’s budget every time they needed something, which by common agreement was outside the boundaries of «core statistics».

The multi-year plan

A multi year plan consists of the following elements:

- A long term statement (say five to ten years) of broad objectives usually accompanied by an analysis of the environment, the capabilities, the threats and the strategic moves required to improve chances to attain the objectives;
- A short term (one to three years) statement of moves designed to reach a set of corresponding objectives because they in turn are milestones required to achieve the longer term goals, usually accompanied by an appropriate allocation of costs involved, units responsible, means of monitoring progress and form in which the results can be certified.

The OFS, apparently one of the first and still few among Swiss federal departments, has put forward a multi year plan which attempts to do mostly the functions of the first of the elements but also includes some of the features of the shorter run component.

The short run element comprises mostly the ability to do three things:

- to reallocate a given budget so as to ensure that no resource is idle, that all resources work in a coherent fashion, and that the work tracks by and large the shifts in demand for information;
- to reassign people in a manner such that their motivation and effectiveness are jointly maximized; and
- to use appropriate mechanisms to ensure that the provisions laid out in the plan are complied with in the form in which they were foreseen and that there is a regular appraisal of how effective they are.

These actions cannot be taken up in a world where there are no matching institutions to carry out each one of them. Firstly, there is a need to bring the senior staff (Abteilungschefs and Geschäftsleitung) of OFS together into a formal planning session -- probably once a year -- during which identified needs are explored in depth. These needs would normally include consideration of recommended improvements to existing programmes, new programme initiatives, and as an obligatory feature the lowest priorities within the existing work programme. Such a planning session would probably result in an improvement of the plan itself. But just as important, it would make a major contribution to the development of a shared vision by the agency’s staff and a far greater commitment to its successful implementation.

Secondly, the consideration of possible trade-offs within an existing budget can only be considered intelligently if there is an effective system that identifies and tracks the cost of each existing statistical programme, as well as a well-established capacity to estimate the costs of proposed new programmes and modifications.

Thirdly, it must be possible to reallocate resources. In turn, this depends on two conditions being satisfied. Within broad guidelines the allocation of resources to specific projects ought to be left to the discretion of the OFS. Secondly, the OFS must have sufficient flexibility to reassign its staff from lower to higher priority projects. This implies that much more attention be paid to a systematic programme of human resource development: training, staff development, developmental movement of staff within the agency, and so on.
Fourthly, given the significant amount of decentralization of the Swiss statistical system, there must be a means whereby the components of the system outside the formal authority of the Director General of the OFS are influenced to move in the direction of coordination and harmonization.

Ability to adjust to unforeseen needs or needs outside the «core» requirements of a statistical system.

It is essential for the statistical system to have the flexibility to adapt itself to client needs. A planning system like the one outlined above would go far in that direction. However, client needs tend to exceed the resources that are made available to the statistical agency but leaving them unmet may well be an inefficient social response to legitimate requirements for information. Accordingly, we believe that allowing the statistical agency to accept commissioned work results in social benefits. It meets user needs; it leads to a much greater client service orientation; it forces the statistical system to become more adaptable and agile; and it places new statistical information in the public domain in conditions where its priority has been «market tested».

The flexibility required to make this possible presumes a number of prerequisites:

- It must be legally and administratively possible for the statistical office to accept and retain payment for services rendered;
- The office must develop the operational capacity for such work. While this is not easy, it brings major rewards. The rewards are not limited to the sight of satisfied clients, but include the development of client oriented staff, highly conscious of cost, efficiency, and turn-around times; and
- Explicit policies must be drafted to rule on the type of work that the statistical office would undertake, and the conditions in which it would do so. These are spelled out in more detail in «Effective System».

3. Structural characteristics: members of the statistical system, structure of the OFS, Staff – morale, professionalism and cohesiveness. Legal basis

3.1 The Swiss statistical system: who comprises it?

The answer is found in the regulations appended to the Law on Federal Statistics. These list all the authorized censuses, surveys and other inquiries resulting in statistical information as well as the government agencies responsible for carrying them out. (8)

Footnote (8): Federal Department of Foreign Affairs: Directorate of Political Affairs. Swiss Agency for Development and Cooperation. Federal Office for Culture. Federal Office for Public Health. Federal Social Insurance Office. Federal Office for Justice. Federal Office of Police. Federal Office for Private Insurance. Federal Office for Regional Planning. Federal Office for Sport. Federal Finance Administration. Federal Tax Administration. Federal Customs Administration. State Secretariat for Economic Affairs. Federal Office for Vocational Training and Technology. Federal Agricultural Office. Federal Veterinary Office. Federal Housing Office. Federal Office for Civil Aviation. Federal Office for Water Management and Geology. Federal Office of Energy. Federal Office for Highways. Federal Office for the Environment, Forestry and the Landscape. Swiss National Bank. Institute for Business Cycle Research.

In order to be listed, each of these offices has to conduct at least one survey designed to obtain statistics directly or indirectly or to manage at least one statistical application. It is these offices that the OFS is requested by Law to coordinate and it is about their statistics and applications that the OFS is to provide advice on methods.

In addition to these members of the system the following should be added:

1) The Cantonal statistical authorities,
2) The communal statistical authorities,
3) The officially funded research institutions.

While the factual question of who comprises the system can be easily answered we are left with a far more difficult question: are the existing institutions and mechanisms sufficient to ensure the effective coordination of a system as complex as the one implied by the lists above?

3.2 The OFS relatively to others

With 471 approved posts, the OFS represents approximately two and a half per cent of the manpower employed in the Swiss federal government. We estimated this proportion by applying several adjustments based on the criterion that we were after the proportion of civil servants with approximately the same characteristics. Accordingly, the comparison excludes inter alia the armed forces (staff and the militia); the field personnel employed by the Department of Justice and Police; the diplomatic and consular services; and the personnel employed by the railways. But it does include the numbers employed in taxation and customs work, which are as considerable as some of the categories excluded. Among the remaining agencies only a very small number are any larger than the OFS. The Bureau of Refugees is marginally bigger and the Office of Agriculture is more than twice the size. (9) In terms of people working from a desk in the same surroundings day-in and day-out (as opposed to non-office fieldwork) the OFS is the third largest Directorate in the Confederation.

Footnote (9): Figures drawn from The Swiss Confederation, A Brief Guide, Berne 1999

3.3 The OFS’s staff structure

There is no indication that the vacancy rate in the OFS’s authorized complement is particularly high or that it varies substantially from one period to another. Almost half of the incumbents are professional; the rest are support personnel. These proportions make the OFS rank as one of the more professional among offices in OECD countries. (10)

Footnote (10): Thirty years ago less than twenty per cent of what was a very small office (114) were professional.

The organizational structure (see Annex IV) is relatively flat. There are three levels of management: one deputy director doubling up as one of five divisional directors, and under thirty section heads (Sektionsschefs). The control spans for each of the levels are not out of the ordinary at five per director and less than ten per section head.

Neither staff turnover nor the staff structure by subject is noteworthy for good or bad although the recent initiative to centralize the government’s computer systems may create an internal imbalance.

Statistics and Informatics

A recent initiative taken by the government consists in centralizing into a single administrative unit for each Ministry, all computer systems analysts. The OFS is of course involved in this matter as computing is an essential activity for any statistical office not to mention one in a very advanced country. The Reviewers were somewhat perplexed by this move. Whereas both experienced or were aware of similar initiatives in other parts of the world, both assumed that technology had overtaken the need to centralize software specialists. Both would find it counter productive -- in the sense that it would be an inefficient move - were such an initiative to be taken in Canada and applied to Statistics
Canada. But since neither was confident of fully understanding the deeper reasons that led the Swiss government to take this step, they decided to restrict their reaction to signalling their surprise and alerting the readers to the fact that elsewhere, comparable moves have been tried and met with failure. The ability to recruit does not seem to be impaired either although the move from Bern to Neuchatel will obviously produce changes in the medium term. The Reviewers did not inquire into possible difficulties of recruitment in the medium term and none of the interviewees brought up the subject.

3.4 Policy instruments

While this is an area where there is a serious deficiency, the OFS is taking very active steps to correct it. The matter is as follows. Right now the policy instruments are the Law, its regulations, and the office’s folklore. Not that the significance of the latter should in any way be minimized but it is not sufficient. The OFS has grown rapidly and so have its attributions and generally speaking its importance. This growth has not yet been matched by a comparable development in instruments necessary to ensure coherence and consistency in the staff’s reaction to similar situations at different times or when experienced by different parts of the organization. While the small size of the office -- in an absolute sense -- its high proportion of professionally trained staff and the existence of experienced cadres all contribute to attenuating the risks of major policy inconsistencies, those risks are there and could result in diminished efficiency, less than optimal effectiveness and an embarrassment for the management of the office at an unwelcome time.

Fortunately, the management of the OFS has taken the initiative and is in the midst of correcting the situation through its project PRAXIS (11) one of the expected results of which is to produce an accepted Policy Manual. Indeed, one of the Reviewers is actively engaged in advising the managers of this process. The process however is not and cannot be instantaneous. For it to be truly effective it has to go through the stages of making all parts of the office conscious of the need for such an instrument, agreed on its contents and capable of developing internal monitoring mechanisms so that in addition to having policies, those policies are generally complied with in a sensible and consistent fashion.

Footnote (11): The authorities of the Swiss Federal Statistical Office launched a project (PRAXIS) designed to recommend steps to restructure the office principally in the wake of the introduction off the new Federal Statistics Act (1993); the rising demand for office output with seemingly no matching expansion in resources and in general because of the expected effects of changes in technology and in the office’s scale of operations.

3.5 Policies on human resources

This is an area in which there are weaknesses. For example, there are no agency wide standards on recruitment, induction of newly arrived personnel, deployment, rotation, and so on. The absence of policies dealing with recruitment and retention, training and career development as well as planned job rotation is of course a deterrent to an efficient deployment of resources as well as weakening the instruments required to ensure staff cohesiveness. The Reviewers were reminded of a situation they faced in their own office some fifteen years ago when personnel policies were the exclusive prerogative of the Personnel Department but consisted mostly in applying government wide regulations rather than finding the right mix of incentives and deterrents to attract, develop, and retain the best human resources available with their budget (12) .

Footnote (12): Statistics Canada, Training and Development at Statistics Canada, Statistics Canada Training Institute, March 1995. Ryten, J., Management Training and Development in Statistics Canada, United Nations, Economic and Social Council, Statistical Commission, 28th session, March 1995.

3.6 Cohesiveness

When asked about the strongest impressions that resulted from an extended stay in Statistics Canada, a visiting head of a national Statistical Agency reacted immediately by saying it was the strong corporate consciousness shown by all members of the staff with whom he had interacted. The reviewers experienced similar feelings when visiting the Australian Bureau of Statistics and it is a general view held by visitors to INSEE. Partly these institutions derive a benefit from their size (several thousands each); the fact that they have been large statistical offices for comparatively long periods of time; and that there is an immediate recognition in their respective societies of their importance and mission. But another reason is that there has been a conscious investment on the part of the management of those institutions in building up among the staff a strong feeling of corporate belonging as well as shared views regarding the directions in which their respective agencies are heading.

In comparison, the staff of the OFS does not strike the same immediate note. This can be explained in a variety of ways not least among them is that it is only recently that the entire staff have been housed in one dedicated building; that the suddenness of its growth spurt has stood in the way of creating a shared corporate consciousness; and that in part, management has not been sufficiently aggressive in creating the necessary bases for this feeling to develop.

We submit that a combination of human resource and planning initiatives might correct the situation in the short run. The measures we advocate are spelled out in greater detail in the section on recommendations but for now they consist in engaging the staff’s good will and enthusiasm in developing an integrated human resource policy designed to liven up people’s career in the OFS and ensure that their contribution is maximized. This includes a mixture of training and career advising -- not necessarily by personnel specialists -- but rather by their own line supervisors creating a cascade effect.

Equally important, among the measures that should be envisaged, is a collective venture in planning the agency’s near and mid-term future through the development of a strategic overview, an operational plan, codified procedures for planning and so on.

Initiatives taken by the OFS

In connection with its desire for restructuring, the OFS has taken a major initiative to provide itself with a sound basis to restructure its activities and organization. The initiative rests on several projects:
- An umbrella project (PRAXIS) that takes stock of, rationalizes and codifies the office’s procedures and processes as part of assisting in drawing up plans for restructuring the office together with a subordinate
- project to take stock of and reorganize its product line in order to maximize its effectiveness (PRODIMA); and
- a project to provide users (inside and outside the office) with a computer accessible databank, complete with supporting methodological information and with other ancillary information necessary to locate and assess data (CODAM).

3.7 Morale

We found it to be good. Perhaps the most important note struck was the general esteem in which the Director General is held. The two complaints most often heard but rather mildly expressed were pleas for greater transparency in the process of budgetary allocation and for a steadier hand in keeping priorities from sudden changes.

3.8 Professionalism

By this attribute is meant a set of qualities such as openness, familiarity with best practices and a permanent striving to implement them, a thorough knowledge of the environment both in the sense of understanding what is expected and knowing what is feasible. The Reviewers found the staff to be imbued with those qualities that distinguish the professional from all others and were particularly grateful to see their questions answered in an open, objective, and dispassionate way.

4. Legal provisions

4.1 The Federal Statistics Act

In making these remarks the purpose is not to suggest that the Act upon which the activity of the OFS is founded should be amended as a matter of urgency. Rather the intention is to list the surprises that the Reviewers find in the text of law and regulations (ordonnances). It also has the purpose of making readers of the Review aware of features of the Act that strike us as a possible impediment as the OFS and its management go about making their role more effective.

The current Act reflects profound amendments to the previous legal arrangements. It was passed in 1992 and was supplemented by regulations in the course of 1993. In contrast, previous arrangements were time honoured. The laws on the census of population went back to 1860 and those on the collection of general statistical information to 1873. There was a quilt of special provision passed during a period of 100 years or so and dealing with special powers to conduct a particular inquiry. The new law brought arrangements up-to-date and, together with a body of regulations, has made it possible for the OFS to work much more effectively.

The law is modern in the sense that it reflects the values embodied in the Fundamental Principles of Official Statistics (13) such as the obligatory public nature of the output, its firm grounding in statistics as an objective discipline, and the recognition of the international nature of national statistics.

Footnote (13): Decision C (47) of the Annual Report of the Economic Commission for Europe, 1991-1992.

There are however a number of features and gaps that strike the Reviewers as susceptible to become impediments to the smooth functioning of the statistical system. As a background there are three issues that must be reckoned with:

- The statistical information function does not lend itself easily to being allocated to different institutions and to different levels of government. In order to ensure that there is a statistical system there has to be coherence in planning, production, and dissemination. If for reasons which are not examined here the system is distributed among different institutions -- different departments, Confederation and Cantons and communes and so on -- there must be a very strong central institution endowed with considerable powers of coordination to maintain cohesion and ensure integration;

- The role of coordination must be vested in a person -- the titular head of the statistical system -- for it is impossible to spell out how an institution coordinates other institutions;

- Where there are different levels of government and structurally different demands placed upon the statistical system there has to be a body of law that defines the obligations of the entire system and the conditions in which they can enter into exchanges of individual data; and

- There are unavoidable lags in the development of data to meet rapidly evolving needs. To reduce those lags, the statistical authorities must have the greatest freedom of movement in their choice of instruments and methods and consequently the law must be as non-prescriptive as circumstances and the political environment will allow.

These considerations are not sufficiently taken care of in the legislation. The following are selected examples of the neglect of one of the two considerations or another:

- The first nine articles define the role, principles, and obligations relating to statistical collection and compilation without once mentioning the institution or person in charge of ensuring that all statistical activities are imbued with the full range of necessary attributes;

- It is not until the 10th article that mention is made of the OFS but even then no reference is made to its Director General (who for that matter is not mentioned in the legislation at all) and yet in the preceding articles mention is made of activities that can hardly take place in the absence of a dedicated Office and its titular head;

- While defining the OFS as the central statistical organ of Confederation, the Law does not go into a definition either of how that «centrality» is assured or the powers that accrue from its status as «central». Moreover the Act implies «centrality » in relation to other Federal agents whereas true «centrality» must comprise the cantonal and communal statistical agencies irrespective of size;

- The Act designates the OFS as the coordinator of federal statistics (Article 10) and as the guarantor of national and international comparability through the data bases which it creates and manages, as the statistical system’s multi-year planner but it leaves power in the hands of the Federal Council to designate some other body in the Confederation to carry out a specific collection of even a statistique de synthèse. The implication is that the OFS is the default data collector and synthesizer but there is no principle that would bind the Federal Council to assign to it a statistical task in all cases unless there were compelling reasons to do otherwise;

- In spite of its responsibilities as a coordinator, the OFS does no more than advise (Article 11) other federal producers of statistics and by virtue of another principle -- that of respect for the respondent -- is asked to place at the disposal of others its data collections insofar as by doing so it does not infringe any laws;

- While article 12 enjoins other producers of statistics to «consult» the OFS, it
does not spell out the accountability of the recipients of said advice should they
choose to ignore it and not show cause. Rather the emphasis is on ensuring that
it gets consulted;

- Article 13 lays down the creation of a Federal Statistical Commission. While emphasis is placed on the Commission’s wide representation, its only attribute is to advise. The article does not lay down what is the weight of the recommendations issued to other Federal producers of statistics;

- Section 4 of article 18 gives the Federal Council very extensive powers. In fact the Council could intervene for reasons which it and it alone would consider impérieuses and limit or, one presumes, suspend the right to publish certain results. It is difficult to comment on this power without seeing the jurisprudence relating to it but at the same time it is a power that comes dangerously close to colliding with the provisions of the Fundamental Principles; and

- The regulations prescribe on average ten essential attributes for each inquiry, census or survey. Thus, the department responsible, the definition of the survey, its objective, the type and method, the target respondents, the character of response, the date and periodicity, the partners if any and particular circumstances if applicable are spelled out in the regulations accompanying the Law. This strikes the Reviewers as over-prescription. The essence of flexible response is to use existing vehicles and adapt them to changing circumstances. That possibility is denied if it requires previous amendment to the regulations. And if the latter can be amended instantaneously the question is raised why burden them with excess detail?

It is possible that the provisions of the Law and its accompanying regulations have the kind of text and detail that made them go through the legislature successfully and that a simpler text with less prescription might have fared badly at the hands of legislators. Be that as it may, in this Review we record the warning that:

- vaguely specified attributes relating to the power and prerogatives of institutions with respect to statistics (the Commission, the Federal Council) may lead to a less than effective performance or to one in which the objectivity and impartiality that are lodestones for any statistical activity may not always be honoured. This warning applies to the articles on the Commission and to the powers of the OFS relatively to other members of the statistical system;

- excessively specified statistical activities in law or regulations (listing in detail all surveys and other inquiries together with their attributes) may well prevent the head of the statistical office from exercising his professional judgment in using available resources in the best possible way to provide the Confederation with the information it requires. In fact there is an even subtler risk. All matters of substance regarding statistics that must go through the Federal Council require party political consensus which in turn subjects what is approved and in which form to the vagaries of political considerations. We certainly endorse the principle that the government should have a major role in determining statistical development priorities -- but we believe that this should stop well before it gets to the articulation of particular surveys that might be needed, let alone their methodologies;

- not specified matters can give rise to a mixture of serious embarrassment or outright inefficiency depending on the luck of the draw. The fact that the Act does not go into the rights and obligations, prerequisites and accountability of the non-Federal members of the statistical system and in particular of the cantonal authorities could well become an impediment to making the system work well.

The issue of the Act is brought up again in the section on recommendations.

4.2 Protection of individual data

The Law mentions the obligation to protect individual records from unauthorized inspection and to apply all possible measures to prevent the office from engaging in inadvertent disclosure. The Reviewers found absolutely nothing in the course of their interviews that would make them suspect that legal provisions were being violated. The staff of the Office seemed to be totally attuned to the idea of confidential protection applied to individual data (Datenschutz) irrespective of their level or place of work.

4.3 Obligation of reporting

The issue under this heading is double-barrelled. Firstly, there is the question of fact, whether or not the majority of inquiries undertaken by the Swiss statistical system is obligatory. Secondly there is the question of whether there are significant differences in response rate between optional and compulsory inquiries. Both are questions of fact and on the first, the Reviewers are satisfied that for a large majority of inquiries and certainly for all key inquiries, the Act declares them to be compulsory. No reference was made in the course of the interviews to a systematic campaign of «statistical disobedience». The Reviewers therefore consider that this is not a priority matter.

4.4 Public Compliance with the Act

Apparently, except in those rare cases where there are written protocols -- in which case the exchange of information proceeds like clockwork -- the OFS appears to have difficulties in making its requests to other public bodies stick.

The matter of accessing administrative data holdings is inevitably a complicated matter and it would appear that, notwithstanding what the Law says, the agencies concerned regard the OFS’ s requests askance whether or not meeting them implies making marginal changes to collected information. A strong bilateral relation could probably correct difficulties of this nature and would be more efficient than trying to duplicate inquiries for which there is a satisfactory administrative surrogate.

But not all is difficulty or obstruction. There are opportunities as well. We noted as an encouraging development the protocol signed between OFS and BUWAL (Federal Office for the Environment, Forestry and the Landscape.) This protocol delineates the areas of responsibility assumed respectively by the two agencies in the pursuit of joint objectives and in the context of a joint statistical venture. While we realize that in many ways this joint venture was the first of a kind, we believe that it creates a precedent: a way was found to fund jointly a common enterprise, as well as to decide on how confidential information would be handled in the framework of the Law and on an ongoing basis.

4.5 Access to official information holdings

The matter of access to holdings of administrative records is featured both in the Act and in the provisional charter of OFS principles so it is not necessary to go over familiar ground. Let us remind ourselves that intensive use of administrative data has the pronounced external effect of removing burden from potentially irritated respondents and keeping the stock of respondents’ goodwill towards the statistical agency relatively intact. It has another beneficial effect. Through more intensive use and more continuous testing, administrative data collections themselves can improve in quality and coherence.

There are, as is the case in most matters, negative effects. For example, providers of individualized information may find it threatening to see government departments -- no matter what the safeguards -- with the power and the knowledge to engage in multiple matching of information that concerns them as individuals much more than it concerns government. They may find it inadmissible if they have not been forewarned that such usage was possible.

Dependence on government records collected for administrative purposes creates other vulnerabilities. Unlike statistics where continuity and comparability over time are almost absolute virtues, administrative data collections are usually in a process of ongoing improvement and often in a process of simplification. If the statistical agency has no say in how this process is managed it will be continually surprised to its disadvantage.

The remedies to these vulnerabilities are on the one hand to build a series of fences and safeguards so that the public at large does not fear the Government’s inordinate ability to match individual records. For example, where records belonging to different registers have been matched, the public should be given some guarantee that once the purpose of the matching has been served the linked records will be destroyed. Furthermore, the statistical agency should have an explicit policy that it will only undertake record linkage for statistical purposes (implying that the results of the linked data are exclusively disseminated in the form of statistical analyses). In instances where there is genuine doubt whether the contribution to the public good is commensurate with the intrusion into privacy involved in the matching operation, the public should have the right to know that broad consultations
with interested groups have taken place. However these and other necessary institutional mechanisms do not appear to exist.

On the other hand, the statistical agency must strive to get into a position where its recommendations are taken into consideration whenever other departments find themselves planning changes, improvements and simplifications.

4.6 Autonomy and interference

This is a delicate area. The context in which the Reviewers approached it was to test the range of arguments that prompted some to argue for a change in institutional and legal status for the OFS. A selected group of internal interviewees were asked whether they knew of instances where any of the following pressures had been applied:

- A political body argued that statistical findings should not be published or else that the publication should be deferred to a later date;
- A political body argued that figures in a press release should be altered in order to «better reflect» reality; and
- A political body argued that words in a press release should be modified in order to suit the convenience of a particular policy.

Examples of all three attempts were produced. The details are of less interest than the conclusions to be drawn from instances, which are recorded in the collective memory of the OFS’s staff. Two points are of particular importance:

- Even though the incidents described for the benefit of the Reviewers were few and tended to be on the margin, the staff reacted to them with strong disapproval; and
- In its current condition the OFS is and could continue to be vulnerable to undesirable interference even though inspired by the soundest of motives.

The fact that there have been some such incidents in the past could be a pretext to question the reliability of the figures produced or the objectivity of the analysis imposed on them. (14) Furthermore, the fact that there is widespread awareness within the OFS of past incidents of this nature might, over time, lead to some staff attempting to second guess political reactions to statistical findings or analyses -- a potentially pernicious possibility. The past occurrence of instances of political interference, even though rare, constitute therefore additional reasons for seeking a more independent position for the OFS, one that places it totally above suspicion in the public’s mind.

Footnote (14): The OFS has taken the initiative of producing an internal charter of principles presumably with the intention of ensuring (a) wide diffusion to basic standards of conduct; (b) protection against undue interference with its activities; and (c) means of promoting consistency of behaviour among the staff. Currently, the charter reflects a mixture of principles and objectives and it is to be hoped that in the course of time a streamlined version is given official approval, and possibly added to the Law. As an example of the Principles the following are the first ten:
- All statistics produced by the Swiss statistical system (hereinafter referred to as "official statistics") are relevant
- The statistics produced by the Swiss statistical system are public
- The statistics produced by the Swiss statistical system are made available to all parties in accordance with their needs
- The methods used to produce official statistics are documented and available to all
- All official comments on published statistics by the members of the statistical system are impartial
- No user is favoured by being informed of statistical results ahead of others
- Official statistics are timely
- Official statistics form part of the nation’s collective memory
- The methods of compilation of official statistics are based on science
- Official statisticians act independently in their choice of methods


1. How do outsiders view official statistics?

We will not hide behind the excuse that the number of outsiders we saw was too small to answer the heading of this section. We regard it as improbable that a group as diverse as one made up by two bankers, a legislator, a newspaperman, two senior civil servants, an advisor to industry, and an academic should coalesce and provide us with a false answer on a matter as vital as their perception of the OFS. Thus we hold their answers to be a useable reflection of what a much larger sample of concerned and well-informed users would have said had we the time and the feeling of need to consult a more extensive group.

We heard marginal criticisms of the Swiss system: about its sense of priorities; the apparent lack of timeliness; the less than perfect coherence among all parts of the system and so on. Similar criticisms would be entertained (and indeed be valid) by any well-informed group of users about any reputable statistical system. The fundamental finding though is that not one among the interviewees had any doubt about either the professionalism with which the statistics are compiled or about the fact that the OFS was the competent body to compile them. Moreover, there was widespread agreement that under current management the Office has made noteworthy progress and become an important contributor to the information that the Swiss public wishes to have about itself, its social organization, and economy. It is in this light that the strictures mentioned below should be read.

2. What changes would outsiders applaud?

Users with a social bent would like to see more and better social statistics -- in particular better statistics on entitlements, education, on health delivery, and on population health conditions. Users with an economic bent deplore the fact that there are not enough data on short term changes in global demand; that the level of detail available for data on value added is insufficient for a thorough analysis of industrial conditions; that there are not enough data on modern services (a standard complaint all over the industrialized world); and that the data on retail sales are derived from a survey that is in bad need of being overhauled.

No user complained seriously about what we perceived as the lack of robustness in the situation regarding the quarterly national accounts. (15) Indeed, when confronted directly with the question about the possible need to strengthen the quarterly national accounts the reaction was that at best the quarterlies are reference figures to be looked up occasionally but not to use seriously in any economic analysis leading to the formulation of, say, monetary policy.

Footnote (15): Under the current situation, whereas the OFS compiles the annual national accounts, one of the Directorates in the Ministry of Economic Affairs compiles the quarterly. The consultations involved a few high-powered users who claimed that for their use, the current situation was adequate and did not merit high-level intervention.

Outsiders did not attack the existing figures as such but they felt that the institutional position of the OFS needed strengthening. For presentation purposes we divided their views into three groups under the headings:

1. Strengthening the OFS’s position as such -- in which direction and for which underlying reasons;
2. Strengthening the Statistical Commission so that it can become truly effective;
3. Improving the effectiveness of the multi-year plan.

3. Strengthening the OFS’s position

The Director General of OFS has called into question the current status of the OFS -- an agency with the character of a Direction Générale within a Government Department – and speculated whether it should change its status to that of a more independent agency (in the so-called Third Circle). Similar concerns were voiced by a number of interviewees who feared that currently the OFS was much too exposed to understandable political pressures arising in its Ministère de tutelle. Whether those fears were borne out by evidence or not was of less concern than the fact that there could be a strong perception of political interference given the reporting relations between the head of the OFS and the Minister. The damage that such perceptions could inflict on the credibility of the figures was such that those who upheld the view that a change was desirable were not overly deterred by its cost. (16)

Footnote (16): The precedent chosen is the Patent Office, which is self-financing and accordingly has very considerable financial autonomy, contrary to the situation of the OFS.

There are other categories of damage, equally pernicious. One that comes about in a subtle way starts out by affecting those members of the staff who are more prone to succumb to pressure. They are the ones who if pushed might contaminate data by attempting to second guess what politicians are after. Members of the OFS’s staff, who also felt that steps should be taken to reduce the vulnerability of the OFS, preferably by changing its institutional status, entertained similar views regarding the risks associated with the current situation.

However, there was no consensus on two matters: whether there should be an effort to fulfill the conditions that would turn the OFS into a Third Circle Institute straight away; whether there should be a substantial period of transition; or whether the OFS should manoeuvre into an intermediate position for a few years before changing its status once again.

4. Strengthening the role of the Statistical Commission

Both external and internal interviewees owned up that they felt the Statistical Commission was not an effective body. Their views were formed by the fact that the Commission passed resolutions and wrote letters but neither seemed to have much effect on the status of the OFS, its programme or its budget. One of the critics interviewed in the course of this Review went so far as to state that the call for a representative Commission (rather than an effective body) meant that each member took an excessively narrow view of his or her interests, which made convergence on any one issue virtually impossible. The advice given was to (a) review carefully the body’s composition and the effectiveness of the membership (no one questioned their competence) and (b) ensure that the status of the Commission is looked at in a new light so that its Chairman has greater access to and influence on decision takers in Government and in Parliament.

5. Improving the role of the Multi-year Plan

There is much to be proud of regarding the multi-year plan. Indeed, its very existence was upheld by one of the interviewees as a model in the Department for Home Affairs. But the plan has also been criticized on at least two grounds. First, there is the question of it being repeatedly ignored by Parliament. The legislative assembly does not feel committed because a particular project takes longer than one parliamentary year to run from start to finish. And secondly there is the problem created by a Parliament that repeatedly showed its independence by legislating, sometimes in considerable detail, changes to the budget.


Legal and institutional arrangements

1. Over a period, the length of which we are in no position to estimate but one that ought to allow due preparation, the status of the OFS should change to that of an independent Institute. Whether the best location is in the so-called Third Circle is something that the Reviewers do not feel competent to judge. There are too many nuances involved and they have more to do with Swiss Public Administration that with the running of a statistical agency. From the point of view of the Peer Review, it is of key importance that the head of a statistical agency, the Government’s Chief Statistician, be given the status and standing equivalent to that of the Directors of other Institutes to which substantial administrative and professional independence was given. As head of the Institute, the Government Statistician would be

- placed at the summit of the Swiss statistical system and recognized as guarantor of its integrity (freedom from political interference, protection of the confidentiality of individually identified records, professional integrity);
- designated as the interlocutor, at the highest levels of governments, on behalf of key client needs; and
- chosen for being a person with standing and influence in respect of (typically marginal) changes to administrative record systems that would materially enhance their usefulness for statistical purposes.

At the same time, thought should be given to the possibility that the new entity should have a federal-cantonal character with the explicit mandate of meeting the core statistical information requirements of both the federal and cantonal governments, as well as the non-governmental sector. Without getting involved in the actual design of such an entity, the following are general indications about the elements that ought to be considered in its make up. The administration of the entity itself should have considerable freedom to make the most effective use of a budget composed of:

- a federal contribution annually voted by Parliament;
- contributions from the cantons whose statistical information needs are served; and
- funds received from federal or cantonal clients, as well as from academic institutions, for specific contracted work.

The Government Statistician (Director General of such an «Institute») should report to a Board composed of senior federal, cantonal, business and academic persons. Initially the critical mass of elements the new entity should consist of, includes:

- the OFS; and
- the statistical offices of those cantons that have a significant capacity for current statistical work.

But the spirit of the enterprise should not be exclusive. In other words, any cantonal authority that gets into a position where it can demonstrate that it has adopted statistical legislation and it abides by the binding provisions of confidentiality should be allowed entry.

2. Even if the first of our recommendations does not lead to the creation of a new organizational entity, arrangements should be made to strengthen the formal mechanisms of cooperation between the OFS and the cantons. In addition, we recommend to the relevant authorities that:

- legally mandated access to administrative records maintained by cantons should be provided to the OFS for statistical purposes;
- the legislative framework under which the OFS and cantonal statistical offices operate should be harmonized.

In particular:

- the cantonal offices should have legally binding confidentiality protection of individually identifiable records (comparable to that of the OFS) ;
- it should be possible for the OFS to exchange identifiable records for statistical purposes with those cantonal statistical offices that have confidentiality protection comparable to that of the OFS; and
- the legal role of the OFS (or the new entity, if the first of our recommendations is acted upon) as a coordinator of the Swiss national statistical system should be articulated.

3. The OFS (17) should be given the administrative flexibility to undertake contract work for clients and to incorporate the proceeds in its own budget. This is a key element of the flexibility needed for it to become more responsive to client needs.

Footnote (17): For ease of reference, the central statistical agency is referred to as OFS -- whether or not the first recommendation results in a change of status and name.

4. The Statistical Commission should be reconstituted and strengthened. In particular

- the Commission’s members should be selected primarily on the basis of individual eminence, rather than on representational principles;
- the Commission should be given an explicit mandate to provide a watching brief over the non-political character of the OFS itself, and its freedom from political interference;
- the Commission’s members should be appointed by the Federal Council, on the advice of the Government Statistician, in order to attach to their advice a maximum of prestige and weight; and
- the Commission’s agenda should be carefully structured to reflect the key importance attached to its advice. For example, it might be asked to advise in respect of the needs of the Swiss national statistical system to move closer to European Union standards, the priorities for such an undertaking, the potential benefits of strengthening the demand side of economic statistics, the conside rations in respect of the future evolution of social statistics (for example, the proposed Observatoire in the health domain) and so on.

Client relations

5. We recommend that a great deal of emphasis be placed by the OFS (with the support of the Federal Council) on strengthening its client relations, particularly in respect of federal and cantonal governments. In particular:

- The status of the Government Statistician should be elevated to facilitate his direct interaction with the most senior levels of governments;
- On-going bilateral relations with key federal client departments should be strengthened. The need for formal protocols governing those relations should be explored -- not as a substitute for strong working level relations, but as a framework to facilitate their evolution.
- Where commissions are formed to discuss sectoral policies (transportation, forestry, tourism for example) involving cantonal and Confederation authorities, the Government Statistician should be asked to take part as a matter of course. Invariably, the agenda for such commissions includes either the use of information or else a demand for new information. In either case the active participation of the Government Statistician is helpful to the discussion and is an important aid to forward planning.
- An administrative mechanism should be established whereby the statistical office could recover funds from clients for contracted work, and use these funds for the purpose. (18)
- The OFS should carry out a systematic examination of the need for external expert advisory committees in fields in which it does not already have the benefit of such advice.

Footnote (18): This recommendation has already been made above but it is reiterated because of its importance under the heading of strengthening client relations.

Internal management issues

6. The momentum developed by the OFS in creating and using the multi year plan should be carried forward preferably with the addition of a number of explicit statements. They include:

- a formal evaluation of new high priority statistical requirements of external clients and their estimated cost;
- improvements needed for existing programs and infrastructure and their estimated cost;
- an enumeration of the lowest priority programs in each area (from which, if necessary, resources can be diverted to higher priority programs) and the estimated resources which would be freed up were such programmes to be dropped; and
- a statement of the implications of not acting on proposed new projects and of diverting resources from identified lower priority projects.

The preparation of such a plan should be based on high-level priorities articulated by the Government Statistician who in turn ought to base his assessment on his continuing and wide-ranging contacts with client groups. For transparency to prevail, the senior staff of the OFS should take active part in those planning discussions, the conclusions of which should be documented and widely shared internally and externally. (19)

Footnote (19): The Reviewers recognize that the current management of the OFS has taken active steps to move in this direction -- both in terms of planning and in terms of broadening participation in the accompanying discussions. The recommendation is in the sense of doing more of it, more widely, more intensively, and with as much systematic documentation including costs as possible.

7. Projects should be approved only if the respective managers are given the resources required to carry them out. These resources should not be the subject of subsequent changes during the year without a re-negotiation with the respective managers of the task for which the funds were initially provided.

8. In order to provide the cost elements needed for such a planning system, all costs of all statistical projects should be tracked on an on-going basis. Managers should have current access to these project costs for management purposes.

9. We recommend increased attention to human resource development issues.

- There should be a planned training program corresponding to the skills and knowledge expected from the professional and technical staff of the OFS. Whatever the configuration of the programme it should include elements designed to sharpen the analytical skills of the staff. Advice should be provided to staff regarding the type of knowledge needed at different levels of their career; and
- There should be a formal program that encourages and facilitates the rotation of staff within the office. Such a program would make a major contribution to staff development, to a much greater ability to reallocate resources from lower priority activities to higher ones, and to the intensifying of institutional cohesion. (20)

Footnote (20): Statistics Canada, Training and Development at Statistics Canada, Statistics Canada Training Institute, March 1995. Ryten, J., Management Training and Development in Statistics Canada,United Nations, Economic and Social Council, Statistical Commission, 28th session, March 1995.

Output-related issues

10. We recommend that the OFS set about developing gradually but systematically a strong capacity to use its statistical outputs as a basis for analytical pieces on social and economic matters. But in doing so it should also develop those mechanisms that are required to ensure that the analyses are strictly non-political, in other words that they refrain from normative judgments, from critical remarks that could be construed as political criticism and from advocacy.

I. P. Fellegi
J. Ryten
Ottawa, 9 May 2000

Annex 1: List of interviewees

NAME ---------------------------------------- TITLE

List of outsiders interviewed

DREIFUSS Ruth ------------------------ Minister Responsible
ANTILLE GAILLARD Gabrielle --- University of Geneva
KAUFMANN Claudia ------------------ Secretary General, Home Affairs
KAPPELER Beat ------------------------ Journalist
PARAVICINI Gian Antonio --------- Canton of Lucerne
RICH Georg ------------------------------- Swiss National Bank
ROTH Jean-Pierre --------------------- Swiss National Bank
ROCH Philippe --------------------------- Environment and Forestry
SAURER Peter -------------------------- Finance Administration
STRAHM Rudolf ------------------------ National Councillor
WALSER Rudolf ------------------------ Secretary, Vorort

List of insiders interviewed

MALAGUERRA, Carlo --------------- Director General
BUHMANN Birgitte -------------------- Labour Force
BUSCHER Marco ---------------------- Population
COTTER Stéphane -------------------- Vital Statistics
GILOMEN Heinz ------------------------ Society and Education
GROSSENBACHER Armin -------- Information
HAUG Werner --------------------------- Population
HANNI Werner -------------------------- Central Service
HERZIG Felix ---------------------------- Deputy Director
KAMMERMANN Michel ------------ Spatial Economics
KOCH Dieter ----------------------------- Prices
MEIER Ruth ----------------------------- National Accounts
SUAREZ DE MIGUEL Raul ------ International Relations


Annex 2: Questions to External Interviewees


External interviews are divided into the following three sections:

1. Matters of fact in which the identity of the interviewee (and possible biases) is ascertained; the nature of his concerns, how they are translated into needs for quantitative information; and how those needs are communicated to the statistical office.

2. Matters of opinion in which there is a probe into how the interviewee views and rates the past performance and future prospects of the statistical agency in the light of his experience. This is the heart of the interview and serves to establish how the BFS (and for that matter the rest of the statistical system) has responded to and indeed anticipated demand - the speed and the quality, the detail, the ancillary information provided etc. The questioning is aimed at establishing whether in the interviewee’s mind the system has succeeded in convincing outsiders of the neutrality and objectivity of its information.

3. Matters of advice in which the interviewee is asked to share with the interviewers his thoughts on how to improve the serviceability of the statistical system, through what measures, budgetary, institutional and personal and attempt to see the order in which such reforms should be introduced - according to the interviewee.

In what follows the headings are explained and so is the direction in which the interviewee is steered. Loaded words are avoided. For example, there is no mention of the word «integrated». And yet it is important to find out whether perceived lack of integration is a determining factor in the opinion in which the BFS and the rest of the system is held. Nor is there any mention of the word «analysis». And yet it is important to know from the interviewees what is their attitude towards explicit analysis. These impressions can only be gauged during the interview by leading the interviewees without actually placing words in their mouth.

Matters of fact

1. Identity of interviewee

The first part of the interview is to find out who is the interviewee and how he fits into the hierarchical structure of the organization to which he belongs. This implies finding out how far is he from the top; what is the level in the statistical organization that he would define as his counterpart; how long has he had dealings with the statistical office; with which other statistical organisations does he have regular or significant dealings etc. The intention is to gather enough information to attach a weight to the answers he will give to subsequent questions.

2. Issues that concern the interviewee

Next are the questions that make the interviewee tick. For example, is he concerned above all with finding out what are the quarterly national accounting growth rates? Does he follow the labour market and immigration patterns? Are his interests short term? Doe he operate in the mode of a researcher? Is he an intermediary, acting on behalf of someone else (a Minister’s assistant) etc? Is he someone that is more concerned with methods that with results? Is he someone that is more concerned with issues relating to good governance, integrity, viability etc. of the statistical rather than with issues that are strictly speaking statistical? The intention is to steer the remaining questions to those matters that are of interest to the interviewee.

3. The interviewee and the BFS - demand (1)

It is important to ascertain something more specific about what the interviewee expects from the BFS. This is done along two axes: one relating to quality in the widest sense of the word and the other to the precise nature of the issues of interest. More explicitly: on issues of quality is the demand mostly concerned with timeliness or accuracy or detail of supporting information. On issues of nature of information: are they micro economic; regional or cantonal as opposed to national; structural as opposed to conjunctural etc. The purpose is to construct a composite picture of what the interviewee considers the ideal output from his point of view - in a static framework.

4. The interviewee and the BFS - demand (2)

In addition to «static» demand, it is necessary to find out what in the interviewee’s mind is the ideal capacity of innovation and of proactive supply that a body such as the BDS should have.

5. The BFS’s response- in the eyes of the interviewee

After finding out who the interviewee is, in what mode he operates and what he would like to see the BFS (and the rest of the statistical system) produce, it is necessary to ascertain how the BFS’s actual supply is seen both in itself and in the context of the total supply of the statistical system. While the purpose is not to constitute a catalogue of sins, a balanced picture of the interviewees’ requirements should include a reference to how many and in which cases they are not met.

6. Mechanisms of adjustment- in the eyes of the interviewee

If the supply of statistical information is not ideal - as one supposes it is not, at least as a point of departure - the next question is how has the interviewee attempted to bring into balance supply and demand. Specifically, what he has done, using what channels of communication, how persistently and with what rates of success. The interviewee’s awareness of the levers at his disposal and the frequency with which he used them are weighting factors when we come to assess his opinions and advice.

Matters of opinion

This section is divided into three parts: institutional personal, and professional. The first is to determine adequacy, the second suitability and the third, competence.

1. Adequacy

The question is whether in the eyes of the interviewee, the institutional set-up is adequate to produce useable statistics or is the inadequacy of the set-up a factor in preventing supply from rising to the expectations of demand. There are several aspects of the institutional set-up. They are divided into four categories: institutions vis-à-vis respondents; institutions vis-à-vis the user community; institutions that bind together the statistics producing agencies into one system capable of producing timely and integrated statistics; and institutions that allow for a correct relationship between the head of BFS and the political community. The knowledge of the interviewee will play a determining role in the direction taken by this part of the interview.

2. Suitability

Personalities cannot be ignored. They may or may not be suitable. What must be known is whether in the opinion of the interviewee the current set of senior people is suitable, why and why not. And if the interviewee could replace them (or had to replace them) what attributes would he deem to be the most important and why. There is also the matter of interaction. If in the opinion of the interviewee there should be an institutional change are the current people best suited to bring it about or to live with it once implemented - and if not, why not.

3. Competence

Is the current leadership of the BFS (and of the other agents who are members of the Swiss statistical system) competent? If they are not, what is lacking? Is there any evidence that they are or are not recruiting young professionals in their own image? Does this mean that the future will have all the shortcomings of the past?

4. Recommendations

Outsiders cannot be expected to provide detailed recommendations about the organization and internal workings of a statistical office unless of course they are alumni of a statistical office. But they can - and very often do - entertain opinions about how the office, or the statistical system could be made more effective if only any one of the following, for example, were altered:

- The programme priorities;
- The system if there is one to make priority choices explicit;
- The board of administration (or Statistical Commission);
- The relations with the executive and legislative arms of government;
- Relations with peer organizations abroad or in the international arena.

The purpose of this section of the interviews is to elicit from the interviewees their view as to what they would do or what they think should be done in order to improve the effectiveness of the system.

Annex 3: Questions to Internal Interviewees

Facts: the process

1. Who are the people you do business with (alone or through intermediaries)
2. How often do you see them?
3. How do you know what they want?
4. How do you know they are happy with what you gave them?
5. How do you tell them that you need more resources to comply with their wishes?
6. Do you discuss joint financing?
7. How do you go about persuading your management that they should provide with resources to comply with your interlocutors’ wishes?
8. How much discretion do you have over the resources that were given to you?
9. How do you cost your new outputs?
10. How do you mobilize parts of the infra structure which you may need?
11. What do you do if you made a gross mistake in your cost estimates?
12. When was the last time you had to attempt to increase your budget to meet new demand?
13. As far as you know do your colleagues proceed in the same fashion?
14. How are your resource problems settled? Bilaterally? Multilaterally?

Facts: the pressures

1. Do your users ask you for advance information (ahead of anyone else)?
2. If they do what do you do?
3. Do your users inquire into unpublished or unpublishable details?
4. What do you if the quality of the results appears unacceptable?
5. Do you consult with colleagues to find out if your procedure is consistent with theirs?

Facts: the product

1. Once complied with, what do you do with your output: hand it over; take your time to analyze it; verify whether it discloses what it should not; examine it for consistency with other statistical indicators; or a bit of everything? Give examples.
2. Are there any limitations on what you can do with your statistical outputs? Are they imposed by others explicitly or by yourself? Give examples if relevant.
3. Is there a process that forces you to have your outputs reviewed by others? If so who are the others?
4. If there is no obligatory process, do you take the initiative?
5. Is the rule the same for your colleagues?
6. Give relevant examples or review procedures -- method and outcome.

Opinions: on your work

1. Are you troubled by the gap between what you think is expected of you and what you believe you can do with the resources that were given to you?
2. Does the gap (if there is a gap) affect the reliability of what you produce? The timeliness? The amount of publishable detail?
3. Are you the only one who holds that view or do your colleagues and your Director share it?

Opinions: on your office

1. What do you think of the way resources are allocated?
2. What do you think of the way institutional consistency is promoted?
3. What do you think of the manner conflicts within the institution are solved?
4. What do you think of the manner in which external conflict (difference d’opinion) is handled?

(In all cases: tell us what is the procedure and give examples where it works inadequately -- if that supports your opinion)

Opinions on your office in the context of the statistical system

1. Your office is perhaps the most important element of the Swiss statistical system but not the only one. Is this to be deplored or applauded? Why?
2. If deplored, is the office going about it the right way to increase its relative importance?
3. Give examples of missed opportunities -- if any?

For meeting with all section heads

1. Have you had collective meetings in the past?
2. Give examples of the topics you discussed
3. What do you regard as the greatest threat facing your office?
4. If faced with the obligation to shrink what should it do less of?
5. Does the office have an opportunity to:
-- Expand
-- Diversify
-- Hit the headlines
-- Become more influential in the long run
Give examples
6. Is the office going about exploring its opportunities in the right way? If not, why not and what is the right way?

Annex 4: Peer Review Method


Of course, this paper could have been far more interesting if it included the Reviewers findings. But that is the subject of their Report. Rather, having discussed the motivation of the Review, the comments for now are limited exclusively to a description of process - the criteria, objectives and methods that were used in the Review.

What the reviewers did

What is it that users, wherever they are, wish to be sure of relatively to a statistical office? Essentially of three things:

I. that the office calculates its numbers correctly (for example, when it says that the economy grew at 3.5 per cent in the last quarter or that 4 in 100 members of the labour force did not find work in the week to which the survey relates, these numbers could not be improved upon if someone else undertook to estimate them);

II. that the office does not waste resources measuring things that few wish to know but rather that what it does is helpful to settle public controversy, to assist those in authority to come to decisions, and to provide a sound basis for insights into how the economy and society work; and

III. that the office uses the most appropriate methods to perform each of its calculations and that as a result it produces the best results one can hope for with its given budget and moreover that as recognized methods improve, the office appropriates them in order to do better with the same budget.

It is possible, albeit with some difficulty to ascertain whether what a statistical office does is generally speaking relevant and helpful. And one can track whether an office keeps abreast of the most recommended methods for each of the applications with which it is concerned. Unfortunately, it is most difficult to invite outsiders to verify that the statistics estimated are correct and indeed the best one can hope for, with the methods at play.

But there is an indirect way of answering all questions, even the first one. If a sample of the staff appears to be competent and well motivated, and if the basic mechanisms exist, those that make it possible for a statistical office to work as it should, the probability must be high that everything else is also in good order of repair. Any Peer Review fighting limited time would concentrate on the indirect method in order to pass comment on the the office reviewed. If the findings suggested that the basics were sound, the Review would have achieved its objectives particularly if it also suggested improvements in existing procedures and mechanisms were in order.

Getting the political level involved

It is not sufficient to receive terms of reference in this case from the Director of the Statistical Office. For a Review to be effective the level above must be engaged in at least three ways. Firstly, it must be aware and supportive of the effort, which means committed to act on the advice received. Secondly, an opportunity must be created for those at the level above to communicate their concerns and establish the basis of a dialogue with the reviewers. And lastly, once the Report is delivered, the findings must be made public in addition to engaging the political level. A failure to do so would imply that the process lacked a fundamental element to its integrity - the capacity to cause beneficial change.

The standards

Some are fundamental and others are appropriate to the occasion; some are ideal and others are practical; some rest on attitudes and beliefs and others on mechanisms and devices to promote certain goals. In the case of this Review the practical standard used was provided by the analytical description that one of the Reviewers (I. P. Fellegi) gave of the workings of the Canadian statistical system. (1) Allowing for differences in legal constraints, administrative practice, scale of operations and matters of public concern, the Swiss statistical system was reviewed to find out whether it had appropriate mechanisms to deal with those problems that are present in any statistical office. For example, were changes in users concerns detected and acted upon, was the machinery required to list priorities and rank them effective and understandable to the staff, was there an ongoing pressure to become more efficient, were there means of ensuring that resources were reassigned to those activities which were likely to yield the greatest payoff and so on.

Footnote (1): I. P. Fellegi: Characteristics of an Effective Statistical System, 1995 Morris Hansen Lecture, Washington Statistical Society.

Standards such as the Fundamental Principles of Official Statistics were in the background as the Reviewers made practical attempts to find out whether the staff could perform objectively, impartially and with the required neutrality.

Of the opinion in which the system -- and by implication, the office -- is held

The reviewers interviewed a sample of concerned users of statistical information, chosen because together they accounted for critical aspects of Swiss public life. Thus there was a lawmaker and a banker, a public sector economist and a counterpart from the private sector, a policy analyst from another government department, a statistician from the cantonal authorities, and an academic distinguished by contributions to quantitative economics. Some of the interviewed were intimately aware of the workings of the Office of Federal Statistics through their membership of the Statistical Commission that advises the Minister responsible.

The interviews were extensive and focused on the opinions held about the office’s capacity to provide an effective service, its adaptability, the reliability in which its data outputs are held, the quality and concern of the staff, the assessment of the leadership provided by the director, and on the views held about the institutional adequacy of the system. Annex I to this paper includes the outline of the interview. While this outline was followed in spirit in all cases, it was inevitably adapted to the circumstances, interests, and perspective of the interviewee.

Interviewees were asked to recommend changes to the system - additions, modifications or redistributions of responsibilities - to the agents within the system and to their interaction with government and with the rest of society. There were no restrictions placed on the advice given, that is to say it could range from modifications in the law, in existing institutions or in the behaviour of the OFS and of other members of the Swiss statistical system.

Of the morale, motivation and competence of the OFS’s staff

Selected members of the staff were subjected to interviews of the same depth as the panel of outsiders. Of course, the questions - more importantly, the concerns of the Reviewers - were different (the actual interview outline is provided as Annex II to this paper). The reviewers were looking for four key subjective elements:

I. whether the staff had a shared sense of commitment, purpose, and direction;
II. whether the staff realized that it operated in a service organization which by definition should be sensitive to users wishes;
III. whether the staff were subject to undue pressures that might prevent it from fully respecting the Fundamental Principles; and
IV. whether the staff felt sufficiently involved in the process that allowed the office to adjust to changes in user demand - in other words, to the way in which resources were reassigned to reflect changing priorities.

Under ideal circumstances the sample of those interviewed should reflect however roughly the staff’s hierarchical structure as well as its variety of disciplines and subjects of interest. In actual practice striving for a truly representative sample would have resulted in much too long a process. Besides, at the point when the opinions solicited get to converge there is probably not much more additional information that the process can yield.

Of finance and personnel

The most precious resource of a statistical office is its staff. But to acquire, develop and keep staff the office must pursue an active personnel policy. The adaptability of a statistical office is vastly improved if the means to move resources and to price inputs and outputs are provided by its systems of financial management. This is why personnel finance and planning are the nerve centers to which very special attention must be paid in the course of the Review. There is not much that one can say ex-ante about these matters other than that law and regulation must not constitute an obstacle to rational decisions. Rather they must perform as helpful instruments to effective management. Naturally, a statistical office is part of a national public administration and is not independent enough to set its own practices in such matters as compensation, revenue management, hiring and firing and so on. The Review process was designed to keep separate what is imposed - while recognizing the resulting constraints - from the office’s practice and to concentrate on the latter to see how much room for maneuver there is left for the office’s management.

Group interviews and confidentiality rules

Bilateral interviews are essential because the interviewed can speak freely particularly if - as was the case - they are offered ironclad confidentiality protection. Nonetheless, there are limits imposed by the format. It shows no interaction among the staff; it is not sharp enough to help detect rivalries or antagonisms; and it does not show up the potential for collective action. To a limited extent, the Peer Review made use of the two techniques even though most of the effort went into bilateral interviews. There was however one occasion in which the reviewers met the third level of management as a group and provided them with three subjects of common interest (see Annex III) for discussion.

The last act

The findings must be delivered to someone. The point about a self started Review is not that they be delivered to the Director of the Office but rather to his political audience on the one hand and to the society in which he operates on the other. The report must therefore be made public. The act of making it public which also entails that it be taken into consideration is made incomparably easier if the Report is delivered to the same group of people who were made aware of the Process at the outset.


The process is viable. There are no doubt other forms of carrying it out but the one adopted in this case appeared to be efficient. Frankness and openness are indispensable requirements. A staff uneasy about communicating concerns to the interviewers would make the process incomparably more difficult. An unconcerned public with the issue of sound statistical information would have made the process impossible.

Annex 5: Organigramm

--- deleted for technical reasons by UNSD ---


Ivan P. Fellegi
Chief Statistician of Canada

Born in Szeged, Hungary on June 22, 1935

University of Budapest, Hungary B.Sc. (Math.) 1956
Carleton University, Ottawa M.Sc. (Math.) 1958
Ph.D. (Math.Stat.) 1961

Statistics Canada
1957-62 Statistician
1962-65 Chief, Sampling Research and Consultation Section
1965-71 Director, Sampling and Survey Research Staff
1971-73 Director General, Methodology and Systems Branch
1973-78 Assistant Chief Statistician, Statistical Services Field
Government of the United States
1978-79 Seconded to President Carter’s Commission on the Reorganization of the U.S. Statistical System
Statistics Canada
1979-84 Assistant Chief Statistician, Social Statistics Field
1984-85 Deputy Chief Statistician
1985-present Chief Statistician of Canada

Recipient of the Order of Canada (Member in 1992, Officer in 1999)
Recipient of La Médaille de la ville de Paris (échelon Vermeil), 1989
Awarded an Honorary Doctorate of Law by Simon Fraser University, Vancouver, 1995
Awarded an Honorary Doctorate of Law by McMaster University, Hamilton, 1997
Awarded the Gold Medal by the Statistical Society of Canada, 1997
Awarded the Robert Schuman Medal by the European Community, 1997
Honorary Member, International Statistical Institute, 1993
President, International Statistical Institute, 1987-1989
President Elect, International Statistical Institute, 1985-1987
President, International Association of Survey Statisticians, 1985-1987
President, Statistical Society of Canada, 1982
Vice-President, International Statistical Institute, 1977-1981
Chair, Board of Governors, Carleton University, 1995-1997
Chair, Conference of European Statisticians, 1993-1997
Member, Board of Governors, Carleton University, 1989-1992
Honorary Fellow, Royal Statistical Society
Fellow, American Statistical Association
Fellow, American Association for the Advancement of Science

Jacob Ryten
Assistant Chief Statistician of Canada

Jacob Ryten was Assistant Chief Statistician of Canada until the end of 1997, when he retired from the Canadian Public Service. He served in his former position, which ranked at the level of an Assistant Deputy Minister, since September 1985. Before then he held positions as Director and Director General in the Canadian statistical office.

Mr. Ryten is Portuguese born and British educated. He got first and postgraduate degrees (1957 and 1959) in Economics and Statistics from the London School of Economics. He is also a graduate of the Canadian National Defence College (1978-79) and a past President of the Inter-American Statistical Institute (1993-1996).

During his career, Mr. Ryten served with both the United Nations (1959-62 and 1979-1984) and the OECD (1962-1969). In his second assignment with the United Nations, he was posted as Advisor to the Ecuadorian Government and subsequently advised on matters relating to national accounts and taxation several Latin American, African and Asian governments. In addition to his current work in Switzerland, Mr. Ryten is currently a part time consultant to the Statistical Office of the United Nations, to the Argentinean and Chilean Ministries of Finance, to the Spanish and Israeli Statistical Institutes and has kept teaching responsibilities in the Training Centre of his former office in Canada.

Jacob Ryten has authored a number of technical papers on economic statistics and is currently working on a book on statistical organization commissioned by the United Nations. He is the holder of the 1997 award of the Most Distinguished Executive Career in the Canadian Public Service.

Jacob Ryten lives in Cirencester, Gloucestershire in the UK.

Swiss Federal Statistical Office
Espace de l’Europe 10
2010 Neuchâtel
Tel. 032 713 60 11
Fax 032 713 60 12

The management of the Swiss Federal Statistical Office (SFSO) has commissioned two experts from one of the world’s best statistical agencies to conduct a «Peer Review» of Swiss statistics. This is intended to achieve the following goals:

- to identify, and describe in more detail, the strengths and weaknesses of the Swiss statistical system in general and the SFSO in particular, by making comparisons with recognised standards
- to benefit from the experience and know-how of a statistical agency which is internationally recognised as an authority and a reference
- to draw up proposals and recommendations to improve the present situation.

The expert report was compiled in the spring of 2000 by two statistics specialists: Ivan P. Fellegi, Director-General of Statistics Canada, and Jacob Ryten, Deputy Director-General of Statistics Canada (retired). The expert report includes publication of the results on 16.6.2000.

The voluntary Peer Review of Swiss statistics by these Canadian specialists is an international «first». In no way is it a step which has been initiated externally because of suspected malfunctions in the system: on the contrary, it is based on our own wish to have the system’s strengths and weaknesses highlighted by an external view.

Back to top | Statistics Division Home | Contact Us | Search | Site Map
Copyright © United Nations, 2014