Financing official statistics: Some reflections

1.Introduction

Bill Seltzer’s 1994 paper on Politics and Statistics: Independence, Dependence or Interaction (1) includes government control over the financial resources available for official statistics as a threat. He noted that “a government…unhappy with the performance of its statistical service may endeavour to punish the service by reducing its budget or adding additional constraints to the expenditure of already budgeted resources”. He also noted that a response that promotes self-financing of official statistics to guard against this risk creates other parallel risks that “can make the agency, at least at the margin, subject to the agenda of those ministries or private sector users who have the most resources”.

I agreed with his analysis then and still do. The ability of the statistical system of a country to fulfil its duty of providing trustworthy, high quality and relevant statistics about the state of the nation to citizens can be frustrated by cuts to its budget, controls on how its budget is spent or the creation of user pays approaches to funding.

2.United Kingdom experience

I have had personal experience of the impact of financial controls exercised by United Kingdom governments on official statistics spanning over 40 years. In the United Kingdom, the official statistics budget has always been considered within the wider national public sector budget framework. In the 1980s, following a review, sponsored by the government, the budget for official statistics was sharply reduced. The quality of official statistics declined and a further review some years later, also sponsored by the government, resulted in significant increases to the budget.

In the 1990s, having had the rollercoaster experience of the 1980s, the system experimented with an approach that included encouragement of user pays solutions. In some areas the result was a skewing of effort towards ministries that were able to fund specific sets of statistics and a loss of cross-cutting statistical development. In other areas, a great deal of effort went into “marketing” activities that struggled even to cover their costs with revenue. And in those areas where there was revenue to be had, the logic of the market resulted in companies being spun off out of the world of official statistics and into the private sector.

In the 2000s, a late cut to the budget for the population census, driven by a short-term government imperative to save money across the public sector, resulted in serious challenges to the delivery of a successful census. Then, in the 2010s, a more sustained public sector austerity drive progressively put delivery of high-quality official statistics at risk. Following another government sponsored review, a multi-year budget was negotiated that included investment in system-wide infrastructure.

3.Reflections on the United Kingdom experience

So what do I conclude from these experiences?

By its nature, there is only one national statistical system in a country. The core purpose, as Winston Churchill demanded when he established the UK Central Statistical Office in 1941, is to ensure that there is a body of statistical information that can be accepted and used without question. Even where there are many producers of statistical information, the essence of the job of a national statistical system, especially the central co-ordinating statistical authority, is to integrate and assess statistics. The goal is to ensure that decision makers do not argue about the validity of the numbers and can focus on deciding the most appropriate course of action to deal with the issues at hand.

A risk with this kind of system is lack of incentives to innovate or drive efficiency in the delivery of the required services. Statistical systems need to develop strong signals from those they are seeking to serve about whether what they are providing is what is needed. To be effective, these signals need to mimic what would happen in a market environment to balance demand and supply.

The national statistical system is only worthy of funding if it is seen to be delivering public value. The work of the United Nations Economic Commission for Europe to analyse the value of official statistics (2) is essential reading for those looking to consider how to judge the level of finance that can be justified. The marketing approaches of the 1980s did provide such signals in some countries, but they were partial and risked skewing priorities. Similar concerns have been raised in countries where donor funding is significant.

A more comprehensive approach would gather evidence on how statistics are used from across the user spectrum, including the current government, and use this to drive change, for example to provide statistics more quickly, in more fine-grained forms and in a more relevant way. This evidence can then be used to justify public funding.

A critical element of the process of gathering such information, and using it to drive change, is building a relationship with users who will then themselves be strong advocates for the national statistical system when budget decisions are being made. This insight is at the heart of the approach to measuring the performance of national statistical systems as elaborated in the World Bank Statistical Performance Indicators (SPI) (3).

Whilst the concept of national statistics is, in effect, monopolistic, the delivery often is not. Many organisations make up the national statistical system. Even in those countries that are highly centralised, there are other players, for example the central bank, who have fundamental roles. In this age of the Sustainable Development Goals, a plural approach to delivery of official statistics is becoming more widespread.

Accordingly, the approach to creating incentives for good quality statistics needs to reach across the spectrum of agencies involved. One example from the UK is that improvements to official statistics about crime required incentives within the police service to collect data more consistently. These were achieved by building relationships with police forces themselves, and also police training bodies and the police regulator, who had a more direct impact on police practices at the point of data collection.

Another consequence of the special place of national statistics in society is the risk that professional independence will be misinterpreted by demanding financial autonomy. Official statistics is funded by taxpayers and must be accountable. This means that official statisticians cannot stand outside politics and must make the case for funding. This is not always comfortable or easy, especially in times of national austerity. Often it is at times when the country is itself struggling to finance its wider public service priorities that official statistics are most needed to guide action. However, it may also be necessary for the official statistics system to play its own part in cutting costs to retain public and political support.

It is also necessary for the national statistics office to demonstrate efficiency, integrity and transparency in its use of public funds. This requires an outstanding finance function within the office. Statistics offices reasonably place emphasis on employing the best professional statisticians and economists, but highly professional accountants (as well as several other key professions) are also vital to the success of the organisation.

4.Conclusions

As a state function, official statistics must be funded by the state and respect the accountability arrangements necessary for state funding to be seen as legitimate. The critical element is strong protection from partisan manipulation for political ends. The ability for the chief statistician to speak out is important, as are opportunities to demonstrate accountability to cross-party parliamentary committees or civil society.

However, this may not have much impact when there is a dominant political party, especially one which does not consider that the current body of official statistics is of great value to them or wishes to focus effort on statistics that are attuned to their interest rather than those of others. In this situation, the wider international network, becomes vital. There are many case studies, including the UK experience, that have been incorporated into wider systems. The requirements of these systems can be used both to justify resources nationally and to defend against partisan reductions to them.

At the United Nations level, the fundamental principles (4) are the core but so too are the many systems and guidelines that have been developed such as the System of National Accounts and guidelines for population censuses. The standards developed by the International Monetary Fund and the Organisation for Economic Cooperation and Development also give a basis for assessing the financial requirements of national statistical systems. In the European Union, a statistical system spanning all member states has been developed which includes sanctions for non-compliance with regulations agreed at the level of the Union. All of these international structures can act as a bulwark against unilateral action by an individual country.

My conclusion is that statistical leadership is the core attribute for ensuring that official statistics have the finance they need. Statistical leaders need the ability to make the case and be seen to be able to use finance wisely to generate the statistics needed by the nation. They also need to engage with the international community in order that their national statistics benefit from and contribute to the global statistical system.

However, strong leadership within the national statistical system may not always carry the day. Statistics is a mirror of the state. Statutory arrangements for funding, or other checks and balances, may give the state cause to think twice, and the influence of international bodies may be decisive in situations where respect for the rules of those bodies matters to national leaders. Even so, if the state does not respect the position of statistics, no protections given can hold out against its power to adjust the levels of funding available.

Nonetheless, increasingly, countries are appreciating that a well-funded statistical system is critical to their ability to make good decisions. Indeed, the indicator framework for the Sustainable Development Goals, applicable to and agreed by all nations, includes an indicator (indicator number 17.18.3) for the “number of countries with a national statistical plan that is fully funded and under implementation…” The COVID pandemic has accelerated this trend for recognition of the value of official statistics and the consequent need for commensurate state funding.