Funding of official statistics

1.Historical overview

Discussions about the funding of NSOs and official statistics are not new. I became very much aware of such discussions at international level when I joined the fellowship of official statisticians some 35 years ago and I dare say that they were not new even at that time. In the mid-1980’s, NSOs in Europe were predominantly funded through the annual central government budgets of the individual countries. In the latter part of the 1980’s, central governments in some European countries, among them some Nordic countries, were increasingly requesting their NSOs to economise on their operations or even curtail some of their activities in order to reduce the need for central government funding of the NSOs. These claims were made despite the fact that user demand, even government demand, for official statistics was increasing. In Europe, the economic and political cooperation made increasing claims on official statistics; in particular statistics of ever greater detail and higher frequency. This was particularly challenging for the NSOs that were facing claims from their governments on using less funds for their activities. Their response was to look for new or increased revenues to fund their regular activities, even to find revenues to increase and enhance their activities.

This gave rise to the preoccupation of many NSOs to market their products, their statistical output [1]. On the one hand, some of the NSOs sought to earn revenues by selling their products. On the other hand, the NSOs found it necessary to market their products, both linked to their revenue making efforts but no less to aid in the prioritisation and planning of their activities. Thus, marketing of official statistics became the focus of many NSOs in the 1990s, in Europe and elsewhere, such as in Australia, New Zealand, Canada, even the US. This new wave happened concomitantly with the great strides that were made in digital technologies at this time. Originally, I believe the digitisation of the statistical production and of dissemination in particular was seen as aiding the efforts to earn extra revenues by the NSOs. Hence, several statistical agencies spent considerable effort in both marketing their statistical outputs and in devising rules and models for the pricing of the digital products. Marketing and pricing became topics at international meetings. Thus, both these topics received substantial attention at the IAOS conferences in Reykjavik in 1996 and in Aquascalientes in 1998. However, with the great advances made in digitising the production and dissemination of official statistics in the course of the 1990’s, the proliferation of digital products and the great reduction in the cost of digital dissemination, the pricing efforts soon became basically meaningless and were largely abandoned.

2.Principles and codes

Another main event in our history, the establishment of the UN Fundamental Principles of Official Statistics [2] in the early 1990’s and their increasing acceptance and applicability in the following years, also had a bearing on the question of funding of official statistics and the efforts to earn additional revenues by the NSOs. The Fundamental Principles re-awakened and underscored the the long-held view that official statistics were a public good to be shared by all interests. Hence, the status of official statistics as a public good was underpinned by the Fundamental Principles and their insistence on the statistics being an integral part of the democratic process in countries. Acting on those principles and given the futility of earning any major revenues by selling their outputs, some European NSOs around the year 2000 adopted the policy of free and open access to the riches of their websites.

The stance of the Fundamental Principles was further strengthened by the establishment of codes of practice of official statistics in Europe since around 2005. These and similar codes have come to guide statistical practices in many countries since then, emphasising professional independence, transparency, integrity and equal access, and specifying adequate government funding as one of the cornerstones of high-quality official statistics. The Fundamental Principles and the codes of practice and the efforts to monitor their compliance have also taken the discussion on funding one step further in the sense that it is no longer accepted that NSOs and similar agencies that are publicly funded can operate side activities to generate extra revenues without regarding or placing the ensuing statistics in the public domain. In Europe, this has also been supported by competition arguments and rules, stipulating that agencies which are publicly funded cannot use such funds to help generate products that are then sold in the market in competition with private producers.

3.Official statistics should be publicly funded

As I see it and having mainly Europe as my point of reference, the events of the last 35 years or so have brought us back full circle as regards the funding of official statistics, from predominantly public funding in the 1980’s, through the efforts of marketing, pricing and raising market funds in the course of the 1990’s, gradually to the current situation that the public good of official statistics should be publicly funded. If that principle is accepted and practiced, we are then left with the issues of how to arrange such funding in a way that allows full efficiency to be brought to the statistical processes and does not interfere with the professional independence and integrity of the NSOs.

My own view on that issue is rather straightforward; the government of a country should allocate sufficient funds on a regular basis to the national agency or agencies that are charged with the responsibility of providing official statistics for the country. Of course, this is a very simplistic statement that needs to be refined. What is meant by sufficient funds and how is the level of funding to be determined? Obviously, that will depend much on the development level of the given country. A developed and rich country can afford a larger volume of official statistics and with greater frequency than a poor developing country. Nonetheless, the developing country will need some minimum official statistics as guided by international requirements. If that is basically fulfilled, we are still left with the question of how to organise the funding mechanism.

Another issue is that the funding should not be very uneven from year to year. There may, of course, be exceptions to this as for instance in the case of decennial censuses and other major events or undertakings. Preferably, the NSO should carry out its activities in accordance with annual plans that are fairly stable, based on rolling multi-year plans that ensure stability and progress in the operations of the official statistical agencies.

4.Current and alternative funding practices

My experience from working in a wide variety of countries is that funding mechanisms are quite varied. In some countries, the funding processes and practices are basically neutral as regards the professional independence of the NSO and allow prioritisation and efficiencies to be practiced by the NSO. This is when the NSO and other official statistical agencies are provided with funds on a regular basis, mainly annually according to some mutual agreement on the long-term regular provision of the statistics, and that the NSO and the other agencies are entrusted with the task of allocating the funds to the different statistical tasks without external interference. This of course, presupposes that all normal financial controls are in place and applied.

Perhaps a main question to be asked is what authority should determine the size and the regularity of the funding as well as any kind of stipulations concerning the use of the public funds. My opinion is that funding to NSOs and similar agencies should be determined as part of the normal regular processes for allocating funds to public institutions, i.e. through the central government budget. In most countries, the budget is proposed by the government to the parliament/legislative assembly, to be debated and adjusted there, and finally approved by parliament. In this way, the need for funds for official statistics is considered alongside the different other needs for which public funding is required. I fail to see that any other body than that debating and approving the central government budget is qualified to carry out the task of deciding on the funding for statistics. While I fully support the notion of professional independence of official statistics, I still find it necessary that their funding needs are weighed, measured and decided within the framework of regular funding of public services and administration.

5.Necessary conditions for public funding of official statistics

However, it is not sufficient that the budget for official statistics is determined by the legislative assembly as part of the overall central government budget. In order for that to work properly a few conditions must be fulfilled.

First, the allocation to the NSO should be decided as a total or a lump sum to be used in accordance with the annual and multi-annual programmes for official statistics. While fully based on activity planning and financial planning, the final appropriation should not be specified in any detail, such as broken down between different domains or tasks and certainly not between cost components, such as salaries and other cost. Such breakdowns or detailed allocations are contrary to the professional independence of the official statistical agencies and severely reduce the flexibility of the statistical managers to ensure efficiencies in the processes as well as reacting to changes in circumstances and needs for undertaking new surveys or analysis.

Second, the central government budget allocation to official statistics should cover at least a whole year of operation. In many countries, budgetary appropriations are split into monthly or quarterly instalments. If that is done, the plan for the instalments needs to be determined in advance and be fully transparent.

Third, the NSOs or similar official statistical agencies should be fully independent and responsible for their spending and financial operations. They must obviously be subject to all normal budgetary controls and scrutiny, but their operations should not be subject to any kind of fiscal restrictions or daily supervision by a superior authority. In some countries, even some so-called developed ones, ministries of finance or similar bodies are charged with supervising the detailed spending of ministries and institutions and may interfere with regular operations. Such practices can be very detrimental to the operations and should not occur.

In some countries where I have worked, the central government budget is legislated in such a way that it may specify the appropriation to the NSO (and other official statistical agencies) for a given year, but the total appropriation may still be reduced within the year at the discretion of the government. This applies in particular to developing countries. If applied, as often happens in some countries, this is likely to play havoc with the regular statistical programme as well an undermining statistical planning and financial responsibility.

Most of what I have discussed earlier applies equally to developed and developing countries. However, many developing countries are heavily dependent on funding of their statistics by international agencies and other development partners. In many instances, such funding is linked to specific surveys or projects that are carried out by the NSOs at some specific intervals, e.g. every other year, every third year etc. Several developing countries carry out both annual planning and rolling multi-year planning where the surveys demanded and funded by international development partners are included. The recommended practice is that in these plans a distinction is drawn between the regular basic operations of the NSO that should be funded by the government and the various surveys funded by development partners. This practice encourages government to provide the official statistics with regular funding thereby enhancing the professional independence and responsibility of the NSO.