Statistics for policymaking: Europe 2020 Charlemagne Building, Brussels 10 & 11 March 2011
1st day – morning (10/03/2011)
Opening of the conference • Walter Radermacher, Director General of Eurostat, European Commission • José Manuel Barroso, President of the European Commission • Janez Poto čnik, Commissioner for Environment, European Commission • Pier Carlo Padoan, Deputy Secretary-General and Chief Economist of the OECD • Elena Flores Gual, Director of “Policy strategy and coordination”, Economic and Financial Affairs DG, European Commission
Plenary session – Europe 2020: statistics for the next decade Chair: Walter Radermacher, Director General of Eurostat, European Commission Key note speeches on Statistics for Europe 2020: Statistics for evidence-based decision making • Didier Reynders, Vice Prime Minister and Minister of Finance of Belgium • Mihály Varga, Parliamentary Secretary, Prime Ministry of Hungary
National experiences with statistics for policy making in Europe • Bertholt Leeftink, Deputy Secretary General, Dutch Ministry of Economic Affairs, Agriculture and Innovation
Statistical agenda for the next decade with regard to using statistics for policymaking outside Europe Chair: Pieter Everaers, Director of “External cooperation, communication and key indicators”, Eurostat, European Commission • John Steven Landefeld, Director of the Bureau of Economic Analysis, US Department of Commerce, USA • Jeni Klugman, Director of Human Development Report Office, United Nations Development Programme (UNDP) - Human Development Index (HDI), USA
1st day – afternoon (10/03/2011)
Sets of indicators for policy making Chair: Marie Bohatá, Deputy Director General of Eurostat, European Commission • Hans Stol, Consultant in ICT-Business Administration-Policy Making, Research Associate University of Tilburg, the Netherlands • Aurel Schubert, Director for General Statistics, European Central Bank, Germany • Andrea Saltelli, Head of Unit of "Econometrics and Applied Statistics", Joint Research Centre, European Commission • Ales Capek, Head of Unit of “Key indicators for European policies”, Eurostat, European Commission
3 parallel sessions on statistics for: Smart growth: innovation indicators, education, digital Europe Chair: Antonio Argüeso, Director of Social and Demographic Statistics, National Statistical Institute (INE), Spain • Mikael Jern, Professor, National Center for Visual Analytics (NCVA) at Linköping University, Sweden • Lucilla Sioli & Michail Skaliotis, respectively Head of Unit of “Economic & statistical analyses”, Information Society and Media DG and Head of Unit of “Information society: tourism”, Eurostat, European Commission • Anders Hoffmann, Director of Entrepreneurship and Innovation Policy, Danish Ministry of Economic and Business Affairs • Matthieu Delescluse, Unit of "Economic analysis and indicators", Research and Innovation DG, European Commission Sustainable growth: indicators for green growth Chair: Pedro Díaz Muñoz, Director of “Sectoral and regional statistics”, Eurostat, European Commission • Robin Miège, Director of “Strategy” , Environment DG, European Commission • Lucio Gussetti, Director of "Consultative works", Committee of the Regions • Catherine Larrieu, Head of the Sustainable Development Delegation, Office of the Commissioner General for Sustainable Development, Ministry of Ecology, Sustainable Development, Transportation and Housing, France • Eric De Brabanter, Department of Environment, Ministry of Sustainable Development and Infrastructures, Luxembourg and Chairman of the OECD Working Party on Environmental Information, Luxembourg Inclusive growth: indicators on poverty and social/economic exclusion Chair: François Lequiller, Director of “National and European accounts”, Eurostat, European Commission • Antonia Carparelli, Head of Unit of “Inclusion, social policy aspects of migration, streamlining of social policies”, Employment DG, European Commission • Sir Anthony B. Atkinson, Professor, Nuffield College, Oxford, United Kingdom • Inna Steinbuka, Director of “Social and information society statistics”, Eurostat, European Commission
2nd day – morning (11/03/2011)
10.50-12.00h Roundtable Chair: Walter Radermacher, Director General of Eurostat, European Commission • Dirk Ahner, Director General of “Regional Policy” DG, European Commission • Marie Bohatá, Deputy Director General of Eurostat, European Commission • Sharon Bowles, Member of the European Parliament for South East England; Chair of the Economic and Monetary Affairs Committee • Enrico Giovannini, President of the National Statistical Institute (ISTAT), Italy • Thomas Wieser, Director General for Economic Policy and Financial Markets, Ministry of Finance, Austria
Conclusions • Walter Radermacher, Director General of Eurostat, European Commission • Olli Rehn, Commissioner for Economic and Monetary Affairs, European Commission
The conference focused on the role of statistics in policymaking, using the Europe 2020 strategy as a case in point. The trends in using statistics for politically highly sensitive purposes and the related complications for statisticians have been discussed.
The aim of the conference was to bring together policymakers and statisticians in a dialogue on how to shape the statistics underpinning policymaking. How should indicators be defined so that they provide the required information? How should policymakers and statisticians cooperate in order to get the best possible outcome on both sides? How can the statisticians ensure that statistics influence policymaking rather than being contaminated by it? How can the current trends in measuring economic, social and environmental developments be reflected in statistics used for policymaking and, in particular, in support of the Europe 2020 strategy? The conference underlined the importance of good interaction and understanding between policy makers and statisticians in order to be able to correctly measure the political targets set.
The participants in the conference were users of these statistics (European and national policymakers, analysts from the European institutions and international organisations, policy advisors responsible for the national coordination of the Europe 2020 strategy, users from media and the academic sphere) and producers of statistics (Eurostat, National Statistical Institutes).
Table of contents
Opening of the conference President José Manuel BARROSO ...... 1 Commissioner Janez POTO ČNIK ...... 6 Pier Carlo PADOAN ...... 10 Elena FLORES ...... 15
Plenary session – Europe 2020: statistics for the next decade Didier REYNDERS ...... 21
National experiences with statistics for policy making in Europe Bertholt LEEFTINK ...... 26
Statistical agenda for the next decade with regard to using statistics for policymaking outside Europe J. Steven LANDEFELD ...... 30 Jeni KLUGMAN ...... 35
Sets of indicators for policy making Hans R. STOL ...... 41 Aurel SCHUBERT ...... 52 Andrea SALTELLI ...... 60 Ales CAPEK ...... 70
Smart growth: innovation indicators, education, digital Europe Mikael JERN, Linnea STENLIDEN ...... 73 Lucilla SIOLI, Michail SKALIOTIS ...... 83 Matthieu DELESCLUSE ...... 95 Antonio ARGÜESO ...... 99
Sustainable growth: indicators for green growth Robin MIEGE ...... 101 Lucio GUSSETTI ...... 104 Catherine LARRIEU ...... 108 Eric DE BRABANTER ...... 112 Pedro DIAZ MUNOZ ...... 118
Inclusive growth: indicators on poverty and social/economic exclusion Antonia CARPARELLI ...... 121 Anthony Barnes ATKINSON ...... 130 Inna STEINBUKA ...... 134 François LEQUILLER ...... 139
Conclusions Commissioner Olli REHN ...... 140
Speech in the opening session President José Manuel BARROSO
Good morning,
It is a pleasure for me to open such an important conference where both policymakers and statisticians meet.
It is indisputable that official statistics do matter. The role of statistics in the social and economic development of our societies has now been officially acknowledged with the first World statistics day celebrated on the 20 th of October 2010, and rightly so.
Statistics also play a key role in communicating our policies. We all enjoy watching diagrams and tables when we read the papers or watch websites. Moreover, modern communication tools, such as blogs and websites, are as hungry for reliable quantitative information as they are hungry for videos: because it captures the essence in a visual and user-friendly way.
Today and tomorrow you will have in-depth discussions on the role of statistics in policy-making using the Europe 2020 strategy as a case in point.
And so intervening at the opening of these series of debates, I thought of just sharing with you a few general remarks from a policymaker's perspective.
First, I would like to say a few words on the current European context .
The economic and financial crisis has wiped out years of economic and social progress. And it has vividly exposed structural vulnerabilities within the economies and societies of individual Member States.
It has highlighted weaknesses in the implementation of our Economic and Monetary Union. And above all it has clearly shown how interdependent our economies became.
Opening Session Page 1
The crisis has amplified the need for strengthened economic coordination and enhanced surveillance in the Euro zone in particular and more generally in the European Union as a whole.
We clearly need an effective framework to deal with macroeconomic imbalances.
We need a common strategic vision on the future of our European social market model in a fast- changing, uncertain and increasingly competitive globalised world.
To address these challenges, the Commission has front-loaded all the measures and put forward new instruments that are to set the agenda for the European policy-making over the coming years: the Europe 2020 strategy for growth and jobs, the reinforced economic governance, including through the European Semester of ex-ante coordination of economic policies in the European Union, - and the regulation and supervision of financial markets.
The Europe 2020 strategy itself is an integrated and coherent approach to support smart, sustainable and inclusive growth rooted in greater coordination of policies at the national and European levels.
The structural reforms that are being implemented as part of the Strategy will contribute to stimulate a broad-based growth, create jobs and allow economies to adjust more readily to changing conditions.
But why are statistics so valuable to policy makers in addressing the crisis and laying sound foundations for an inclusive and sustainable growth?
In a nutshell: implementing the overall response to the crisis will require sound, high quality data and statistical analysis, on which decisions can be based.
As we are increasingly faced with complex questions to answer and decisions to take, we obviously need to know what is going on in our economies and our societies.
The essence of statistics is precisely to concentrate the complexity we face into meaningful and relevant information for the policy makers, but also for the citizens; to ensure transparency of decisions and of decisions making process.
Opening Session Page 2
Transparency and accountability are indeed crucial to foster mutual trust between policy makers, both in Brussels and the capitals of our Member States and the citizens.
There must be not doubt on the European determination to draw a line under the crisis. That is why we need to be providing first a credible assessment of our strengths and weaknesses as well as vulnerabilities. It is only based on such a sound diagnosis that the policy-makers can take decisive actions to address them.
Broader and enhanced surveillance of fiscal policies, but also macroeconomic policies and structural reforms must rely on high quality statistical information.
Statistics are also crucial when setting our targets and using indicators for monitoring and evaluation purposes.
In particular, measuring trends in competitiveness and following the developments in the macroeconomic imbalances within the enhanced economic governance framework will be based on a scoreboard of economic indicators.
The high quality of statistics produced under robust quality management becomes even more important when new enforcement mechanisms are foreseen in case of non-compliance by Member States with agreed targets or policy commitments taken.
Sound statistics are also a condition sine qua non for accurate forecasts and projections that also form an integral part of the recommendations made by the Commission in the framework of the economic policy coordination and fiscal surveillance in the European Union.
Statistical information that the policy-makers receive must be relevant, timely and accurate to best decide on the policy direction and the best course of action to take that should be both ambitious and realistic.
Statistics are, therefore, literally present in all parts of our global response to the crisis. And it is in this overall framework that the Commission has repeated the necessity to grant Eurostat with extended powers in the field of fiscal statistics, which was agreed by the Council in August last year. These powers allow for an in-depth review of upstream data sources used in the excessive deficit procedure.
Opening Session Page 3
But let me tell you here that it is already in 2005 that the Commission has put forward the proposal to strengthen the mandate of Eurostat, which was unfortunately not met with enthusiasm by the Member States at that time. I personally believe that we could have avoided many of the future problems were the extended powers to the European Statistical Authority granted earlier than last year. *** Given that in your discussions you will take the Europe 2020 strategy as a relevant policy case example, let's have now a closer look at it.
Sound statistical analysis underpinned by robust statistics will be key for steering the implementation of the Europe 2020 strategy. As you know, statistics are almost all over the Strategy starting with its headline quantitative targets that embody its overall objectives. Let me recall them for you at the beginning of your conference:
Europe 2020 is about boosting growth and employment through the following objectives at the European level: • raising the employment rate from the current 69% to 75%; • boosting spending on R&D to 3% of GDP from the current 2%; • reducing the school drop out to less than 10% from the current 15% and increase the share of people in their early 30s with a university degree or equivalent from 31% to at least 40%; • achieving the EU's 20/20/20 climate change goal; • lifting at least 20 million people out of the risk of poverty and exclusion.
These targets serve as a benchmark for the necessary policy action and at the same time as a communication device. They are also being translated by the Member States into national targets under each of the headings. This allows citizens to know why and how decisions are taken, and follow their implementation. We will only make a verifiable progress with the many initiatives under Europe 2020 if the related policy assessments are grounded in an indicator- based analysis.
Thus the appropriate choice of indicators is key to boost our understanding of the complexity of our diverse societies within the European Union, to better communicate on it, and to better respond to new policy needs as for example with the " GDP and beyond " initiative to include measurement of well being.
Opening Session Page 4
That being said, certain pre-conditions have to be respected for producing statistics of the highest quality.
Statistics are compiled within the European Statistical System based on an extended cooperation between the National Statistical Institutes across the Union and between the Institutes and Eurostat.
The data used by the Commission reflect both, the Eurostat competence but also the quality of data notified by the Member States. In case the latter were of insufficient quality, this would have a negative impact on the quality of the European statistics but would also put at stake the credibility of our policy-making.
We have seen in the context of the recent economic and financial crisis that the weaknesses in the quality of upstream public accounting data were compounded by the weaknesses in the statistical governance arrangements in place at the national level. This in turn influenced the way we were seeing the economic reality in some Member States at that particular moment.
Further strengthening of the governance of the European Statistical System will be key in this respect. The Commission is currently reflecting on ways to achieve it and we will soon present some concrete proposals in a forthcoming Communication. Obviously designing and monitoring statistical indicators for policy purposes requires close cooperation between policymakers and statisticians at national and at European levels. Policy- makers must provide the right incentives for and guarantee the professional independence of statisticians, strengthening their impartiality and objectivity. This is even more relevant in the European context to ensure trust between the Member States and between the Member States and the Commission.
Given that statistics are expected to assume an even stronger role in our democracies, this is the right moment to reflect on the challenges in the use of statistics in the policy-making.
Your conference, organised in a very timely manner, is the appropriate forum to foster such a discussion.
I wish you very fruitful deliberations and am looking forward to reading your conclusions.
I thank you for your attention.
Opening Session Page 5
Speech in the opening session Commissioner Janez POTO ČNIK
• I would like to welcome Eurostat initiative to organise this conference that brings together EU and Member States politicians and statisticians.
• Europe 2020 strategy has a clear environmental dimension and identifies the new paradigm of green growth. The Strategy and the Flagship Initiative to a Resource-efficient Europe provide the conceptual framework to define the quality of growth. Our objective is to ensure that resource-efficiency considerations are integrated in all policies. This will steer businesses, investors and consumers towards a systemic change which is necessary to achieve green and competitive economy.
• Following the adoption of the Strategy, a new governance cycle has been agreed. It introduces a number of new elements. One of them is a reinforced integrated surveillance, aimed at monitoring Member States efforts in implementing Strategy goals.
• Europe 2020 strategy has defined EU targets which will be monitored by the headline indicators, and flagship initiatives which will be monitored by indicators currently being developed by the Commission services.
• At the same time DG Environment is working on a roadmap on "Resource efficient Europe" which will include indicators and probably targets to monitor progress in resource efficiency. We are working on a set of resource efficiency indicators. They will help us to communicate the need to improve resource efficiency substantially, will track our record against the objectives of the flagship initiative, and will help us to find out where actions towards more efficient and sustainable resources use are taken most effectively.
• In order for the Commission to design, implement and monitor EU policies we need reliable data and information, and a knowledge base approach is fundamental to ensure that we set policies to tackle EU problems and challenges.
• Eurostat and the European Statistical System are one key source of data on which EU policies are built on. When it comes to environmental issues, the European Environment Agency and JRC are also much contributing.
• I'm an economist and was Head of Macroeconomic Institute of Slovenia, so I know about the importance of economic indicators such as Gross Domestic Product, Net Domestic Product, Net National Income. However to achieve truly inclusive and sustainable growth in the
Opening Session Page 6
sense of Europe 2020 we need social and environmental statistics and indicators on the same level with economic statistics, concerning scope, details and timeliness. This is true in particular seeing the increasing dependency of businesses on natural resources, including eco-system services. Therefore I fully support the statement by Mr. Radermacher last year that in the medium to long term there should be as many resources being devoted to environmental and social accounts as classical national accounts. We are working closely with Eurostat and other Commission services to narrow down this knowledge gap.
• Nowadays we have a significant amount of data available, but official statistics cannot always meet the EU data needs. This is because either they do not cover all relevant issues (such as biodiversity) or because they are too old-often 2 or 3 years. Policy makers need fresh information that reflects current problems.
• We should therefore look for additional sources of information and open the door to research institutes business organisations, space agencies and think tanks, and we invite the European Statistical System to also start a dialogue with data sources outside the statistical community.
• I know that resources in the National Statistical Institutes have been reduced e.g. in the wake of the crisis and that, social and environmental data are not considered as essential as economic and financial data. But statistics are an investment in the knowledge base, they are not administrative burden. Good statistics are cheaper than wrong decisions.
• We could think about counting the expenses for statistics together with RTD expenses as they inform us about the state of reality which is the starting point for further research and development.
• In the coming years we need to develop indicators to monitor progress on Green growth, more specifically on the Environmental good and service sector, on Green Public Procurement and on eco-innovation. We also need a common understanding how to measure subsidies that have harmful side effects on the environment and there in the medium to long term on our natural resource base. With the final aim to reshape those subsidies so that the social and economic objectives are met without the harmful side effects. And I am happy to know that some preliminary information exists, even if major progress is needed.
• We need to better measure our natural capital, its eco-system services, in physical terms and – as preliminarily tested by TEEB – in monetary terms. We expect from companies via our accounting rules to set up very detailed asset accounts to be sure that this company is
Opening Session Page 7
not at risk of collapse, that it is actually increasing wealth and not just burning money. However, our economies as a whole we have currently to run without such accounts of our natural capital and man-made wealth. Certainly such accounts would provide very valuable knowledge base for policy making to help reach the sustainable growth objective of Europe 2020.
• To get the prices for natural resources and environmental services right we need to know about extent of the existing externalities – positive and negative ones.
• Figures and indicators on emerging issues need to be good enough, but not perfect. Roughly right is better than no information at all.
• I know about the hesitation on the side of the statistical system to enter the business of valuation or monetisation of environmental damages and eco-system services. However, as the Stern report and TEEB have shown, this is vital information for policy making.
• If the statistical offices are not the right place to provide this vital information we need to discuss who would be best placed to do so.
• With our focus on measuring green growth and resource efficiency we will not forget or neglect measuring the final objectives we strive for: prosperity, quality of life and well-being for citizens while preserving the natural capital of our planet. Measuring these issues in a comprehensive way is a key mandate of the Commission's beyond GDP agenda which I have the pleasure to coordinate together with our host today, my fellow Commissioner Olli Rehn.
• From the adoption of the Communication "GDP and beyond" in 2009 and the announced report on its implementation in 2012, we are now roughly half way through the 3 years. With the first deliverable of the EC/Eurostat on environmental accounts currently debated in EP and Council.
• For designing and assessing policies detailed data are vital. For communication and political debate headline figures are also vital. Indices and small dashboards are very useful when we need to assess complex issues like environmental pressures on our environment, resource use and biodiversity. We have currently no accepted comprehensive overall figure for the environment in the concert of GDP, inflation and unemployment rate. Under the beyond GDP road map, we are developing a composite index on environmental pressures and a Sustainable Development Scoreboard.
Opening Session Page 8
• I'm very glad to see that the initiative coming from the European Commission in 2007 is gaining more and more momentum. International organisations such as OECD and World Bank are fostering measurement of societal progress. There are also important new initiatives in Member States, such as the well-being indicators in UK and a Parliamentary Committee in Germany looking into the relation of growth, prosperity and quality of life. Also with the mandate to develop indicators. I'm looking forward to the results of these efforts.
• I would be glad to co-operate with Eurostat and the European Statistical System, with the EEA and the JRC, with business and civil society organisations, with environmental and space agencies and others in view of producing a regular monitoring system on green growth and resource efficiency and societal progress beyond GDP.
Opening Session Page 9
Better statistics, better policies, better lives Pier Carlo PADOAN
President Barroso, Commissioner Poto čnik, Ms. Flores Gual, Mr. Radermacher,
Ladies and Gentlemen:
I am very pleased to attend this Conference on “Statistics for policymaking: Europe 2020.” Your kind invitation to speak at this opening session attests to the excellent relations between the European Union and the OECD.
Today, 21 EU Member States are also members of the OECD, which now gathers 34 countries overall. This obviously implies close ties between the EU and the OECD. We have common purposes in terms of economic stability, sustainable growth, structural reforms and social cohesion. Our strategic partnership has been developing over the years through constructive discussions, joint initiatives and day-to-day co-operation between the various general Directorates of the Commission and the OECD Secretariat.
This year, the OECD is celebrating its 50 th Anniversary. This is the occasion to restate the Organisation’s mission to develop best standards in public policies in order to improve the lives of citizens in OECD countries and around the world -- in other words to promote better policies for better lives . Our work encompasses nearly all areas of public policy, from macroeconomic to entrepreneurship, from fight against corruption and tax heavens to biotechnology, from health to competition, from education to development aid, from renewable energy to migration, international trade, innovation and corporate governance. In all these areas, we measure and analyse, we compare countries’ performances, we set standards and identify best practices.
Statistics are a central pillar of this policy work: they provide us with the essential foundation for our analyses, forecasting and benchmarking work; they are at the heart of our evidence-based policy recommendations. Today, I would like to highlight the key role of statistics in our response to the crisis and, more broadly, in our work on assessing the well-being of people and fostering the progress of societies.
Enhancing statistics: a major lesson from the crisis
The world is slowly recovering from the worst financial, economic and social crisis of the past seventy years. This is time for us to take stock of a number of important lessons that we have learned from the pernicious dynamics that led the world economy close to the abyss.
Opening Session Page 10
The crisis was not caused by a lack of statistics. Many indicators were available that pointed to stress and unsustainable tensions in our economies but unfortunately these were largely overlooked by policymakers and the general public. Nevertheless, the crisis did reveal some important data gaps and underscored the importance of timely and internationally comparable information. In particular, the crisis has heightened the need to examine closely, systematically and regularly the development of imbalances in the economy, both globally and locally. This includes developing better measures of economic slack (such as the output gap), balance sheets (of households, businesses, the financial and the government sectors), current and capital account positions, prices of financial assets, real estate and commodities, exchange rates and various inflation measures. One main reason why the crisis took us by surprise is precisely because the extent of the imbalances, and their interactions, were underestimated. Stronger surveillance efforts are thus required and this entails a strong demand for relevant statistics.
In 2008, the G-20 leaders requested that the IMF, the Financial Stability Board and other international organisations explore some of the data gaps and provide proposals to address them. The OECD has been actively involved in this effort. Our contribution has focused on the collection and analysis of quarterly sector accounts. These include timely accounts for government, financial and non-financial corporations, and households. While such data are not yet fully developed in all countries, their elaboration is critical to better understand and quickly capture issues such as household and corporation debt, asset revaluation, exposure of sectors to financial risks and the interconnectedness among sectors within an economy and across borders.
The OECD is conducting this work in close collaboration with the European Statistical System (Eurostat and the ECB). I am pleased to see that our two Organisations are increasingly sharing their data, analyses and expertise in the area of key economic and financial statistics.
But today we also observe the social consequences of the crisis: millions of people have lost their jobs, their homes, their pensions. The pace of the global economic recovery is still too slow to reintegrate many of them into the labour market and unemployment remains at unacceptable levels in the face of high public debt levels.
In this context, the massive protests of unemployed youth in some low and middle income countries are a strong warning signal. While the political environment in our countries is of course different, our weak labour markets can be a source of major social conflicts. We are currently putting a lot of efforts to develop statistics to better monitor global economic and financial developments. We also need to improve our statistical base to better monitor the social consequences of the crisis which in many countries will take a long time to unfold. For
Opening Session Page 11
example, a few weeks ago, an OECD report on Jobs for Youth showed that in most of our countries young people have been among the most affected by the crisis at a time when they were already more than twice as likely to be unemployed as the average worker. And an upcoming OECD report will also show that in many countries, households’ poverty rates have increased during the crisis. More detailed micro-data are needed to analyse the situation of at- risk and disadvantaged groups in the population and help design appropriate policies.
Beyond GDP: a stronger focus on people’s well-being
More fundamentally, the crisis has not just highlighted areas where our statistical capacity remains deficient, but it has also undermined the confidence of people in markets, public policies.and official statistics. Restoring confidence requires providing evidence of what matters to people’s everyday lives.
The idea of measuring well-being has been recognized as a priority by top European political Leaders. Among them are President Sarkozy of France, Chancellor Merkel of Germany, Prime Minister Cameron of the UK and President Türk of Slovenia and it also figures highly in the EU2020 agenda. Beyond Europe, President Lee of Korea, and the Prime Minister of Japan, Mr Kan, have also put well-being and progress at the top of their political agendas.
Much of the statistics and indicators that we routinely produce are not looking at what truly matters to people. There is today a consensus that we put too much emphasis on measuring economic production – principally through gross domestic product – and not enough on assessing people’s well-being. Too many important policy decisions are still being taken with GDP per capita as the main measurement rod.
Of course, GDP has to remain at the centre of our statistical systems. But we need to complement it with indicators that measure well-being more broadly. In addressing this challenge, we can build on the rich body of data and indicators that the OECD has produced over the years in such fields as education, health, innovation, the environment and climate change.
We can also rely on the accumulated knowledge and research that has taken place over the past ten years or so. The OECD has in fact led the international reflection on the measurement of progress. We have organised three World Forums in Palermo in 2004, Istanbul in 2007 and Busan in 2009. These events have allowed us to engage in discussions about relevant measures of well-being and progress with policy makers, statisticians, scientists, economic and social actors from more than 130 countries. The 4 th World Forum will take place in New Delhi in October 2012.
Opening Session Page 12
The research agenda has also benefitted greatly from the recommendations made by the Commission on the Measurement of Economic Performance and Social Progress , established by the French President Nicolas Sarkozy, also known as the Stiglitz, Sen, Fitoussi Commission.
Building on all these initiatives, the OECD is now stepping up its statistical work on three areas that matter for well-being and progress: material living conditions, quality of life and sustainability.
Our first effort is directed at better measuring households’ material conditions . What happens at the level of the entire economy does not necessarily tell us what happens to households’ purchasing power. For example, over the last decade, real GDP in my own country Italy grew by about 1.6% per year, whereas the real income of Italian households only rose by half that amount. Large discrepancies can also be observed in many other countries, which points to the importance of understanding and explaining why the benefits from growth have not been passed onto households.
Equally important, standard statistics on economic production do not account for the many important services that households produce at home. These include, for example, child care, cooking, care for the elderly, volunteer activities. Such activities could add between 20% and 40% to our conventional measures of GDP. Taking into account these services would give a very different picture of economic performances across countries and highlight the importance for policies of meeting new challenges, such as that of reconciling work and family life.
Most people would also agree that there is more to life than money. Indicators of quality of life should reflect health conditions, competencies, the time people spent commuting or with their families, their housing conditions, their local environment, their political participation, social connections, and the various risks that shape their feeling of security. Importantly, for each of these dimensions, we need to capture various forms of inequality in the population. We are also working on the development of guidelines on internationally comparable subjective well-being indicators. This is a huge area of work, where gaps between the statistics available and what is needed are even larger than those for economic statistics.
Lastly, can current well-being be sustained over time ? How can we ensure that our well-being today is not achieved at the expense of our children’s well-being? We are now attempting to develop better metrics of how our production and consumption patterns impact on the environment, both domestically and globally, through the development of indicators in the context of the OECD Green Growth Strategy. Taking a broader perspective, the concern on sustainability also requires developing better measures of human and social capital, and of knowledge and innovation.
Opening Session Page 13
As you can see, we have an ambitious statistical work programme ahead of us. We are working in very close collaboration with the EU on this, in particular through joint work on improved measurement of income inequalities and direct participation in the Eurostat/INSEE Sponsorship which is following-up on the recommendations of the Stiglitz-Sen-Fitoussi report at the European level. This will feed directly into the indicators work underpinning the EU2020 agenda.
On the occasion of our 50 th Anniversary this year, we will issue a report entitled How is life? , which will present for the first time a set of comparable indicators on material living conditions and quality of life for OECD countries. These indicators will mainly reflect what data currently exist but over time, as our and others’ work deliver their results, these will be incorporated in future editions of the publication.
Ladies and Gentlemen:
Let me leave you with a key message: developing better statistics is not an end in itself. It is a means to improve policies that affect the functioning of our economic system and the well-being of people living in it.
There are several economic, societal and environmental goals that the EU and the OECD share. These goals determine what we have to measure, how measurements have to be transformed into useful knowledge, and how this knowledge has to inform policy design and decision making.
I hope that this Conference will stimulate ideas on the statistical agenda for the next decade. The OECD stands ready to work closely with the European Commission to enhance the statistical basis that underpins our policies.
Thank you.
Opening Session Page 14
Speech in the opening session – Outline Elena FLORES
Summary
In her introductory speech, Elena will present the main recent policy initiative proposed for the European Union in response to the economic and financial crisis, highlighting their statistical implications.
Europe 2020, the enhanced economic governance and the coordinating framework called "European semester" contains all references to statistical indicators.
These initiatives show how the use of specific statistical indicators helps not only in defining the policy targets, but also in assessing the policies developments and in communicating progress to the general public.
At the same time, the use of statistical indicators in the conception, implementation and monitoring of policies requires data of high quality and, as a consequence, a great effort from statisticians.
Structure
1. Introduction
2. Europe 2020
3. Enhanced economic governance
4. EU semester
5. Conclusions
1. Intro
• The main initiatives proposed by the EU Commission in response to the economic and financial crisis which hit European citizens are: the Europe 2020 strategy and the enhanced economic governance, as well as the enhanced economic policy surveillance within the coordinating framework called "European semester".
• Each of these initiatives makes references to official statistics, in all their phases: conception, implementation, monitoring and assessment of policies.
Opening Session Page 15
• Translating the political objectives into statistical indicators helps to better define them, as well as to identify possible trade-offs.
• Furthermore, statistics constitute a powerful communication tool: they play a crucial role in improving transparency and understanding of the policy performance and promoting European citizens' joint ownership of the initiatives.
• Because of their relevance, statistical indicators used in the policy context need to be of high quality and easily accessible.
2. Europe 2020
• The Europe 2020 strategy put forward by the Commission sets out a vision of Europe's social market economy for the 21st century as a smart, sustainable and inclusive economy delivering high levels of employment, productivity and social cohesion.
• The Union has set five ambitious objectives - on employment, innovation, education, social inclusion and climate/energy - to be reached by 2020. The objectives are defined in very concrete terms, measured by statistical indicators:
1. Employment:
75% of people aged 20 to 64 to be employed
2. Research and development:
3% of EU GDP to be invested in R&D/Innovation (public and private combined)
3. Climate change / energy
20% greenhouse gas emissions lower than 1990
20% of energy from renewables
20% increase in energy efficiency
4. Education
Reducing school drop-out rates below 10%
At least 40% of 30-34 years old completing level education
5. Poverty/social cohesion:
At least 20 million fewer people in or at risk of poverty and social exclusion
• Each Member State adopts its own national targets in each of these areas, reflecting different starting situations and national circumstances.
Opening Session Page 16
• At national level, therefore, the interaction between policy makers and statisticians in the definition of the national targets is necessary.
3. Reinforced EU economic governance: budgetary, macro-economic and structural surveillance
• On 29 September 2011, the Commission presented a legislative package on Economic Governance, whose core elements are:
o the reinforcement of the Stability and Growth Pact (SGP),
o the extension of the surveillance to macroeconomic imbalances,
o the setting up of a range of enforcement mechanisms, including sanctions, which would kick in at an early stage.
• The elements of this approach were outlined in May 2010 and a concrete "toolbox" was presented in June. This was complemented by preparatory work and consultations with a broad range of stakeholders, in particular by the Task Force on Economic Governance chaired by President of the European Council Herman Van Rompuy. The Commission, the European Parliament and the Council have agreed that the Economic Governance Package should be adopted by summer 2011.
• The Economic governance package has several statistical implications, which I would like to briefly highlight here.
3.1 Fiscal side :
• 3 regulations reinforcing the Stability and growth pact.
o One regulation introduces new elements regarding the analysis of medium term debt position, which call for extended statistical information.
o For example: maturity structure and currency denomination of the debt; accumulated reserves and other government assets; guarantees, notably linked to financial sector.
• One directive for adequate national budgetary frameworks
o This directive refers to the quality of "upstream public finance data" which are used to compile EDP statistics. It aims at reforming national public accounting and budgetary processes and making them more harmonised (therefore comparable) among Member States.
Opening Session Page 17
o Improved information on public spending, also at local level, will be available for the surveillance tasks assigned to the Commission.
3.2 Macroeconomic side: the excessive imbalance procedure
• One Regulation on the prevention and correction of macroeconomic imbalances. This regulation introduces an alert mechanism based on a scoreboard composed of a limited number of statistical indicators. For countries whose economic imbalances triggered the alert mechanism, an in-depth analysis of their origins will be carried out.
o Both the alert mechanism and the in-depth analysis will be based on statistical indicators.
o Neither the scoreboard nor the detailed indicators used in the in-depth analysis will be included in legal acts.
o The translation of policy objectives into the economic concept and then to the corresponding statistical indicator is not as straight as one can imagine. An example is given by the indicator used for assessing current external balance, where a set of indicators from different statistical sources can be used.
o The statistical indicators to be included in the scoreboard are being the subject of discussions in several fora, with the participation of economic analysts and statisticians: these discussions have shown the importance of the interaction between policy makers, analysts and statisticians. Such interaction will lead to very well grounded choices.
• One Regulation on enforcement measure to correct excessive macro economic imbalances in the euro area (no statistical implications).
• In both the public finance and macroeconomic excessive imbalances contexts, the use of statistical indicators guarantees transparency of the process and provides objectives grounds for policy decisions.
4. The European semester o The Europe 202 strategy and the budgetary and macro-economic policies are coordinated through a new architecture, called the "European semester" and approved by the Member States in September 2010. o This new cycle has several stages:
Opening Session Page 18
• It starts each year in January with the Annual Growth Survey, in which the Commission provides an analysis of the EU economic challenges and policy responses, on the basis of:
the progress on Europe 2020 targets;
a macro-economic report;
the joint employment report
which is to be discussed by Council formations and the European Parliament ahead of the Spring meeting of the European Council in March.
• At the Spring Council, Heads of State and Government identify the main challenges facing the EU and give strategic advice on policies.
• In April, taking this guidance into account, Member States send to the Commission:
their medium-term budgetary strategies through Stability and Convergence Programmes
the National Reform Programmes, including the national targets under the Europe 2020 strategy.
• Based on the Commission's assessment, the Council will issue country-specific guidance. This will be particularly timely for the finalisation of Member States' draft budgets for the following year.
5. Conclusions o The recent initiatives that I briefly presented have shown once again how statistical indicators constitute a fundamental tool to identify, elaborate, monitor and communicate policies. o As such, the statistics used for policy making should be of high quality: comparable, reliable, timely and easily accessible. o Statistics are a public good which help to:
o Provide objective ground for analysis and policy decisions
o guarantee transparency of surveillance and enforcement actions;
o improve "public appropriation" of policies.
Opening Session Page 19
o This important role requires flexibility and adaptation from statisticians, mainly to respond to new demands, e.g. financial data, statistics on public and private sectors, productivity and competitiveness indicators. o A specific effort should also be made with respect to the publication of both the data and the methodologies applied for their production. This will improve the mutual understanding and cooperation between policy-makers, analysts, statisticians and the citizens.
Opening Session Page 20
Didier REYNDERS
Dear President,
Dear Director-General,
Ladies and Gentlemen,
First of all, I would like to thank you for your invitation to attend this conference, during which the role of statistics as a decisive factor in decision-making related to EU 2020 will be studied. First, I will briefly discuss the role and importance of statistics. Next, I will tackle some questions about the role of the GDP. Finally, I will discuss some questions about the other indicators in the EU 2020 Strategy.
If we want statistics to be useful, we must first set quantified objectives. It is only because the EU 2020 Strategy, elaborated by the President of the European Commission, Mr José Manuel Barroso, has set realistic objectives based on figures, that indicators play a key role in this field. Statistics are therefore an integral part of the Europe 2020 Strategy. The headline indicators measure the progress made by the EU and the Member States towards achieving the five headline targets of the strategy.
We have also seen some months ago, at the triggering of the sovereign crisis in the Euro Zone, how much sound and reliable public statistics were of the utmost importance.
Consequently, our State and - more broadly speaking - all stakeholders in our society need statistic indicators when policies are set, but also when they are tested or subjected to a final evaluation. Since our economies have become even more complex, there has been a significantly greater need for transparency, for the development of IT and the improvement of the population’s education, as well as for reliable and harmonized statistical data.
Indicator-based evaluation of policies has become a democratic requirement to ensure the transparency and objectivity of the results obtained by politicians. Democratically, it’s essential for the citizens to have appropriate statistics at their disposal to evaluate the political decisions and, then, the politicians. Adequate statistics are therefore indispensable. The indicators, as set in the EU 2020 Strategy, enable us to measure and evaluate the social and economic facts.
Decision-making should be based on an efficient, comparable and reliable statistical system. The European Statistical System was built up gradually, on the basis of the national institutes and Eurostat, with the objective of providing comparable statistics at EU level. The European Statistical System is efficient and offers high-quality services. In June 2010, the EU Finance
Opening Session Page 21
Ministers drew their conclusions from the Greek crisis, which had seriously harmed the European Statistical System’s credibility, by granting Eurostat auditing powers.
EU 2020 is the new strategy for the European Union to stimulate a smart, sustainable and inclusive growth, which is interrelated and mutually reinforcing. The European Union has set five EU headline targets, on employment, innovation, education, social inclusion and energy, to be reached by 2020.
In the EU 2020 Strategy, the Gross Domestic Product is an essential criterion to measure economic development, and understandably so. The GDP was introduced after the Great Depression of 1929 and is the sum of the added values of all economic activities of a monetary nature.
The GDP is universally recognized and accepted as a key indicator to make comparisons. It also provides a solid basis for sound economic policy decisions. It is an important advantage that the GDP is objective and therefore not open to interpretation. By increasing our growth, the European Union will meet tomorrow’s challenges, mainly the ageing of the population and the risk of declining. An increase in GDP will contribute to full employment and will enable us to guarantee the sustainability of our economic and social model.
However, no indicator is perfect or sufficient in itself to describe the complexity of our society. The GDP is an indicator of our economic activity, but does fully not measure our population’s well-being. The GDP keeps its relevance, but it would be logical to have a set of indicators at our disposal which would complement the more traditional statistics, without actually replacing them. The GDP does not sufficiently take into account the non-market sector, the distribution of income or the over-use of environmental capital. I will tackle this issue later on.
I would like to quote the example of a car’s dashboard. Initially, the driver only got information on the car’s speed. Nowadays, he sees in one glance extra information, such as the oil level, the engine speed, the petrol level or the temperature. The dashboard has to remain user- friendly and sparing with information and it must meet the stakeholders’ pragmatic needs, without causing administrative or bureaucratic overload to the economic and social actors.
I can conclude by saying that we need more than one indicator, but that an overabundance of indicators should be avoided, because it would prevent us from efficiently monitoring our public policies.
I would like emphasize the role which Eurostat has played in this field by anticipating the improvement of measurement at EU level, as well as the work done by the European Commission in its communication « the GDP and beyond ».
Opening Session Page 22
Consequently, the GDP indicator should be complemented to surpass its limitations, which are from now on well-known in academic circles. These limitations were emphasized in the “Stiglitz- Sen-Fitoussi Report” on the Measurement of Economic Performance and Social Progress, which was published in September 2009.
The GDP does not take into account the loss of environmental capital and the use of non- renewable resources. Intergenerational and intra-generational equity is another issue. The distribution of income is not sufficiently taken into account and the intergenerational equity, for instance in the case of pension financing, is not taken into consideration. However, intergenerational equity, by controlling public expenditure, is essential to the quality of our public policies.
The five headline targets include economic, environmental and social indicators. The chosen indicators have several positive dimensions: feasibility by looking at timeliness and coverage and technical soundness, comprising overall accuracy and comparability (over time and across countries).
- The employment rate is the work-related indicator, with the objective of having 75 % of the population aged 20-64 employed. This traditional and solid indicator is better than the unemployment rate, because it offers a better basis for comparison and is less biased. Contrary to the Lisbon Strategy, the lower limit has been raised to 20 years, to better take into account the reality of the labour market. The upper limit was maintained, although legislative changes have been observed in some EU Member States. Even though senior-specific indicators are no longer included, it is essential to concentrate our efforts on keeping employed persons aged 55 and older.
- Innovation is the second headline target, with the objective of investing 3 % of the EU’s GDP in R&D. This traditional indicator is a major driver of economic growth. However, while implementing policies, the efficiency and effectiveness of the - mainly public - expenses should be thoroughly analyzed. The European deficit in R&D investments, if compared to other large economic entities, leads to insufficient growth and a lack of employment creation (the indicators’ objectives are interrelated). This deficit is mainly due to a lack of private investments in R&D. Would it not be better to include the fiscal expenses contributing to innovation or private research, while at the same time improving the conditions of applied research in Europe, as well as the link between universities and spin-offs?
- There are three environmental indicators, namely the greenhouse gas emissions (base year 1990), the share of renewable in gross final energy consumption and the energy intensity of the economic system. These three indicators are complementary, as they measure at the same
Opening Session Page 23
time the absolute emission threshold, the origin of the energy and the energetic efficiency. We can deduce from these indicators our capacity to maintain our standards of living, by checking whether the environmental capital stock will still be passed on to the next generations.
The economic growth and the environmental objectives are not conflicting. As the EU 2020 Strategy points out: « Resource efficient Europe to help decouple economic growth from the use of resources, support the shift toward a low carbon economy ». What we need, is more growth while consuming fewer resources. An increase in energetic efficiency will even lead to more competitiveness and, thus, to more growth.
A rapid data transmission is crucial in this context. Economic data are provided after a few weeks, while it often takes several years to obtain environmental information. I’m perfectly aware of the fact that this demand is very difficult from a technical point of view.
The two indicators related to education and training are of course essential. The share of early school leavers should be under 10 % and at least 40 % of the persons aged 30-34 should have completed a tertiary or equivalent education.
I would like to make two remarks on this subject:
- By selecting the cohort aged 30-34 in 2020, we include persons who have already begun higher education. The policies which are now being implemented will have effect in 2025, or even in 2030, on the cohort aged 30-34.
- I would also like to emphasize the importance of finding a balance between labour supply and demand. There must be a link between university education programmes and labour market needs. As regards the 40 % educated persons aged 30-34, the share of scientists and engineers has to increase significantly and more generally, the employability of students has to be improved. Once they are educated, they will find in the European Union a sufficiently attractive and competitive environment to keep living there.
Finally, the EU 2020 Strategy aims at preventing 20 million persons from being exposed to the risk of poverty and exclusion. With a view to their efficiency and effectiveness, social policies must be quantified and evaluated on the basis of realistic objectives. The traditional poverty rate indicator rather measured inequality in the distribution of income, than poverty itself. By complementing this indicator with the persons living in households with very low work transfers and severely materially deprived persons, we obtain a composite indicator which better measures the poverty risk.
Opening Session Page 24
However, we must make sure to have a simple, comparable and reliable indicator in the end. This indicator anyhow shows that policies related to work programmes and growth strategy are important in the anti-poverty strategy.
Before I conclude, I would like to make some remarks:
• The indicators of the EU 2020 Strategy have to be analyzed in parallel with the ones related to competitiveness and budget (deficit and debt ratio). They will have to be included in the European Semester and in the six proposals of the European Commission to reinforce economic governance by strengthening the Stability and Growth Pact and by improving macro-economic surveillance in the European Union. The budget and competitiveness indicators have to play a key role in the economic, social and environmental roadmap which is being elaborated.
• As the Sofia Memorandum states: “There is a growing societal and political demand to measure progress, well-being and sustainable development in a more comprehensive way”. I agree that statistics will take better account of the complexity of our societies.
• In a strategy such as the EU 2020 Strategy, we need objective, comparable and reliable indicators. Which should not restrain us from making progress in the well-being indicators, based on the citizens’ subjective perception, to measure the quality of life in a more multidimensional way.
• I will underline the importance of good interaction and understanding between policymakers and statisticians in order to be able to correctly measure the political target set.
I can conclude by stating that analysts are always inclined to measure a society on the basis of economic and social statistics: these data are important and revealing, but other statistics, which are much more difficult to obtain, are as important and reveal as much about a society: does a society have more memories than dreams or more dreams than memories? I’m sure that thanks to the EU 2020 Strategy, which must be read in parallel with the European Semester, the Competitiveness Pact, the Small Business Act and the Single European Act, the European Union has more dreams than memories.
Thank you for your attention.
Opening Session Page 25
National experiences with statistics for policy making in Europe Bertholt LEEFTINK
Plenary Session Page 26
Plenary Session Page 27
Plenary Session Page 28
Plenary Session Page 29
The role of statistics in the United States' Economic Future J. Steven LANDEFELD
Plenary Session Page 30
Plenary Session Page 31
Plenary Session Page 32
Plenary Session Page 33
Plenary Session Page 34
Measures of well-being: The HDI and related indices Jeni KLUGMAN
Plenary Session Page 35
Plenary Session Page 36
Plenary Session Page 37
Plenary Session Page 38
Plenary Session Page 39
Plenary Session Page 40
The contribution of statistics to the improvement of transparency in the policymaking process Hans R. STOL
1. Introduction
I start this paper with three statements. These statements are scientifically proven to be correct in my PhD thesis (Stol, 2009). I mention each of the three statements followed by a short explanation.
Statement 1. The policymaking process (PMP) covers the whole process, from decision making up and until the evaluation of the implementation of the policy.
Policymaking (PM) is the process by which governments translate their political vision into programmes and actions to deliver "outcomes"- desired changes in the real world (Cabinet Office, 1999). This definition of policymaking may suggest that policymaking ends with programmes instead of changes in the real world. Policymaking always intends to change aspects or parts of the reality. 1 Therefore, considering the effectiveness of the policymaking process means that we must take into account (1) the process of setting and directing the course of action pursued by government, (2) the realisation of the actions in the real world, (3) the measurement of the results, and (4) the evaluation of the policy.
Statement 2. Transparency inheres verifiable and measurable observations of changes in reality due to policymaking.
A transparent PMP is a process by which a full, accurate and timely disclosure is given on the content of the process. Verifiability of the PMP is the possibility to ascertain the correctness of all the information about the steps in the PMP. The effects of the PMP should be measurable (if this is not the case: verifiability and transparency are out of question). This means that the following situations in reality should at least be measurable: (1) the reality that is relevant for the realisation of the objectives at t+1 versus the expected relevant reality at t, (2) the implementation of the objectives versus the original objectives, (3) the relation and the expected relation between the target of the policymaking process (at t) and its environment versus the actual relation at t+1. Verifiability as well as measurability are preconditions for full, accurate and timely disclosure on the content of the PMP (=transparency).
Statement 3. Transparency in policymaking improves the quality of the policymaking process
Transparency in policymaking is not only important because of the populism: people "force" nowadays their representatives in government to explain clearly (1) what they want, (2) how they want to realize this, (3) how they want to measure the results of their intentions and (4) how they want to explain the results versus their intention. Transparency is also a characteristic of the policymaking process that is closely related to the quality of the policymaking process (Stol, 2009). The objective of policymaking is achieving the desired changes in the real world. So, the quality of policymaking is as high as possible if all the desired changes in the real world are achieved. The effectiveness of the policymaking process is related to its quality and its transparency.
The following sections implicitly build on these statements.
2. A framework for the policy-making process
1 ) Althaus (2007) formulates this as follows: "Public policy is ultimately about achieving objectives. It is a means to an end. Policy is a course of action by government designed to attain certain results."
Sets of indicators for policy making Page 41
This section shows the framework for the PMP (Stol, 2009) 2. Implementation of the framework enables the transparency of the PMP. A framework for the PMP process should support the quality of the policy-making process. Transparency and verifiability in the PMP are necessary to optimize the quality in decision making. This means in the first place that the policy process should be structured so that all steps can be clearly defined. Although people may argue that every individual policy decision is different because of different situations, objectives, and circumstances (see, e.g., the approach by Davies, 2001), there is always (on a certain level of abstraction) the possibility to structure the process (see, e.g., Easton, 1979; Checkland 1999; de Leeuw, 1982a). This is not only useful to transfer knowledge about politics (Dorey, 2005) but also to make the process transparent and measurable. The framework for PM provides the elements and structure for a PMP, so that individual cases can be judged and analyzed, and, even more, can be designed so that the process gains in transparency and verifiability. In the framework, the Evidence-based Policy Making (EBPM) is considered as a controlling system
ENVIRONMENT
JUSTIFICATION: INTENDED INDICATOR VALUES related to LOBBYING INFLUENCING MEASURED INDICATOR VALUES
EVIDENCE BASED INDICATOR INPUT DECISION VALUES MAKING
OUTPUT RESULTS IN
INTENDED OBJECTIVES/ INTENDED MODELLING / OBJECTIVES/ DERIVING INDICATOR- INPUT INDICATOR- ANALYSING INDICATORS VALUES INDICATORS VALUES
DERIVING INPUT
EBPM MEASURES TO MEASURES TO CHANGE CONTROL SYSTEM: CHANGE MEASURED REALITY MEASURED REALITY VALUES ACCORDING VALUES ACCORDING OBJECTIVES OBJECTIVES
RESULTS IN REPRESENTS
REALISING OBSERVATION REALISING OBJECTIVES OF EFFECTS OBJECTIVES
OBJECT SYSTEM OF INTEREST over the part of reality that is the objective of the PMP. We call the part of reality that is the target of the PMP: the Object System of Interest.
Figure 1 : Framework for Policy Making Process
The EBPM control system has three main relations with its environment.
(1) Actions directed toward its environment in order to obtain the best possible policy decision (high quality and acceptance).
2 ) In Stol (2009) this framework is called "Framework for Evidence-based Policy Making", in order to avoid confusion with the Framework of Easton (1979).
Sets of indicators for policy making Page 42
(2) Receiving reactions and lobbies of interested parties in its environment.
(3) Producing justifications towards its environment for the policy measures achieved.
The EBPM control system itself consists of seven interconnected parts.
(1) The Evidence-based Decision Making (EBDM), in which the policy decision is made after interaction with the environment.
(2) The determination of the objectives and the kinds of indicator to be measured to clarify the results of the policy measures compared with the objectives.
(3) The derivation of the values of the indicators as representation for the achievement of the objectives..
(4) The determination of the measures to change the relevant part of reality according to the objectives.
(5) The measurement of the results of the policy measures.
(6) The analysis of the results of the measurement versus the intended values of the indicators.
(7) Producing indicator values as representation of the results achieved by the implementation of the policy measures.
Evidence-based Decision Making is the part in which the politicians arrive at the policy decision. Although each decision-making process is different and its progression depends on many factors, one may recognise similar elements in all of them. In the context of this paper the following three similar elements are relevant.
(1) The policy decisions refer to an Object System of Interest. The complexity of the Object System of Interest may differ and may lead to differences in (a) the scope of the decision and (b) the organisation of the Change Process.
(2) The kinds of decision process can be categorised varying from rational to more or less random. The most appropriate category for a unique decision process depends on its characteristics (routine decision, complexity Object System of Interest, available information for decision making, see also Stol (2009)).
(3) The content of the policy decision refers to a part of the real world. This means that for each policy decision an interaction exists with the environment that will be influenced by the decision (lobbying, interest groups, etc.). This is a factor that is often used by politicians to escape from EBPM. Their argument is that it is never possible to convince every participant in the Society involved in the decision making. A negative side effect of taking too much time for achieving a most satisfying decision is that the Object System of Interest can be changed and the issue under discussion is no longer actual.
Considering the divers options for a decision, information of former outcomes of PMP's plays a role. According to the literature on EBPM this is considered an important asset to achieving the best possible decision.
Sets of indicators for policy making Page 43
In the framework of the PMP we emphasise the significance of the indicator values that have been produced and analysed as result of former policy decision in the relevant area. These figures may be used for (1) new decisions in the relevant area, (2) continuing the policy measures, and (3) changing the former policy measures.
3. Statistics and transparency in the PMP
This section explains how transparency in policymaking is related to statistics. Two quotes from Wikipedia concerning "Official Statistics" (30-11-2010) are listed below:
(1) "Official statistics are statistics published by government agencies or other public bodies such as international organizations."
(2) "Official statistics should be objective and easily accessible and produced on a continuing basis so that measurement of change is possible. Official statistics: • result from the collection and processing of data into statistical information by the government institution responsible for that subject-matter domain. They are then disseminated to stakeholders and the general public. Statistical information allows users to draw a relevant, reliable and accurate picture of the development of the country, compare differences between countries and changes over time. They enable stakeholders and decision makers to be well informed and develop policies for addressing actual development challenges. • make information on development accessible to the public and therefore assist in the accountability of public decision-making. One use of official statistics is to measure the impact of public policies and highlight the need for development."
The descriptions on Wikipedia show how statistics are related to public policymaking and how statistics need understanding of the reality in order to provide unambiguous information about reality. This section demonstrates the importance of statistics in improving the transparency in public policymaking. Before doing so, we will explain some of the principles of statistics: the meaning of data, meta data and information.
In this paper the following definitions are used for data, message, and information (see, e.g., Stol, 2009, Sundgren, 1975, Stol, 1990, and IFIP, 1998).
Data is a structure intentionally arranged into a medium to represent by this structure a part of reality.
A message is a reference to a part of reality as interpreted by the receiver of the data.
Information is new knowledge derived from (1) the old knowledge and (2) the knowledge implied by the absorption of the message.
In semiotics and linguistics a distinction is made between the syntactical, semantical, and pragmatical levels of communication (see, e.g., http://en.wikipedia.org/wiki/Semiotics, Chomsky, 1975, Nielen, 1976, and, Stamper, 1996 3).
Syntax is the study of principles and rules for constructing sentences in natural language. Syntax is about symbols and a combination of symbols.
3 ) In this context the Semiotic Framework by Stamper (1996) is worth mentioning. In this framework, Stamper distinguishes the IT platform and the human information functions. In the IT platform the framework presents the following three levels (1) physical world, (2)empirics, and (3) syntactics; in the human information function we see (1) semantics, (2) pragmatics, and (3) social world are recognised.
Sets of indicators for policy making Page 44
Semantics is the study of interpretation of signs within particular circumstances and contexts. Semantics is about the references to the real world; it gives the relation between the symbols and the objects to which they refer.
Pragmatics is the study of the ability of the of natural language speakers to communicate more than which is explicitly stated. Pragmatics is about the intended meaning in the communicated language; it gives the relation between the symbols, the use, and the effects.
In Table 1, the above defined concepts are related to each other in a matrix. For most of the cells different sets of metadata are relevant for the combination of the level and concept symbolized by the cell. Each cell is identified by a number and a character.
We will call this matrix: The Metadata Matrix 4. The Metadata Matrix provides a frame of reference for the definition of the kinds of metadata to be used in the communication in the PMP.
A: data B: message C: information concept level
I : syntax IA: Structure put in the IB: Structure in the IC: Structure in the medium to represent brains brains the data
II: semantics IIA: Reference to reality IIB: Reference to part IIC: Reference to part as meant by the sender of reality as of reality in the con text of the data interpreted by the of the frame of receiver reference
III: pragmatics IIIA: Choice of the IIIB: The meaning of IIIC: What is the effect medium and structure the data for the of the new knowledge so that the receiver of receiver on the receiver of the the data can interpret data them
Table1: The Metadata Matrix.
This Section proceeds with using the Metadata Matrix in order to show which kinds of metadata are relevant for (1) the definition of the objectives and the related indicators and indicator values (see parts 2 and 3 of the framework for the PMP), (2) the measurement of the results (part 5), and (3) the dissemination of the intended values of the indicators versus the realised ones (justification, part 6 and 7). When we apply the Metadata Matrix on the output process (i.e. the dissemination of the results) , the user of the output is the “receiver”. The statistical system in its role as producer of the indicators (or better indicator/value combinations) is the sender of the data. Examples of questions to be answered on the metadata are in this consideration: What kind of metadata do we need, to make sure that:
(1) the data refer to the part of reality meant by the Statistical System?
(2) the choice of the medium and structure of the data meet the requirements of the receiver?
4 ) A Dutch version of this Metadata Matrix is published in Stol (1995a) and in English in Stol (2002b). In these publications the physical level was the fourth level referring to the media used in the communication process. Since this level does not contribute to the metadata models presented in the context of this thesis, I have omitted the physical level.
Sets of indicators for policy making Page 45
(3) the meaning of the data for the receiver is the same as meant by the Statistical System 5?
(4) the data once interpreted by the receiver refer to the part of reality as meant by the Statistical System?
Normally speaking we want to provide “unambiguous and objective” indicators, so that we should not bother on the effect of the data (cell IIIC in the Metadata Matrix) 6.
The metadata to be considered for the input process (the process providing the observations in the Object System of Interest) are briefly discussed below, in particular as far as E-Datacapture is concerned. This means in terms of the matrix that: (1) The sender of the data is the respondent 7, (2) the receiver of the data is the (automated) part of the Statistical System that registers the data so that they can be used in the further process, leading to the indicators. In Table 2 we show the relation between the cells defined in the Metadata Matrix and the parts of the E-Datacapture system in which they are addressed.
Cell in the Metadata Matrix Statistical e-Datacapture system
IA : Structure put in medium to Questionnaire form. represent the data
IIA: Reference to reality as Questionnaire responses, link with data in information meant by sender of the data system of data provider.
IIIA: Choice of medium and The software for the e-questionnaires needs to be available structure so that the receiver of for a great part of the respondents; this means that media the data can interpret them used are accessible for all of them, as well as the data formats used. (E.g., Internet and XML.)
IIB : Reference to part of reality The definition of the questions and support files used for as interpreted by the receiver explanation of the expected answer (code systems, nomenclatures).
IIIB : The meaning of the data for Availability of electronic links between the questions and the receiver the content of the business administration, may assure the data receiver that the data filled in are according the bookkeeping.
IIC : Reference to part of reality Question answers received and stored in the environment in the context of the frame of of the data collector. reference
IIIC : What is the effect of the The effect of information for the receiver of the data (the new knowledge on the receiver Indicator System) may lead action from the receiver (e.g. of the data asking for explanation of unexpected differences with former messages).
Table 2: Metadata of data-capture system.
We clarified that the Statistical System plays an important role in the framework for the PMP. Applying the framework is a precondition for obtaining transparency in the PMP. After the policy
5 ) The definition of the meaning of the data is a well know challenge for statisticians: how to add the proper metadata in order to transfer the data as they are meant. 6 ) Transparency requires producing unambiguous data without considering the pragmatical effect of the information. 7 ) In practice, the respondents can be households and enterprises.
Sets of indicators for policy making Page 46
decision making the objectives should be determined as follows: (1) the objectives must refer to a part of reality, (2) the objectives must be measurable, and (3) the objectives must be verifiable. The consequences of these requirements are:
(1) The objectives should be formulated in terms of objects (or object types) and their intended values, after the realisation of the policy. The values are attributes of the objects (or object types) and include a reference to the time on which the values should be achieved.
(2) The values of the objects (including time) must be measurable. This means that before the objectives are determined, the method for measuring the realisation of the objectives can and must be established.
(3) The values of the objects must be verifiable. It must be possible to ascertain the correctness of the information about the realisation of the objectives.
In the Statistical System we use the concept of "indicator" as item to address the progress of situations or processes. It is important to be aware that the concept of indicator is identical to the concept of the formulation of objectives as has been described above. The requirements for the specification of the objectives regard directly the use of indicators. Indicators are the same, or equivalent with, the terms in which the objectives are specified. A statistical indicator is defined as "a data element that represents statistical data for a specified time, place and other characteristics". 8
If the indicators are defined as required, it is not difficult to measure the outcome of the policymaking process. The measurement will be according the basic rules, preferably determined at the moment that the objectives are established. It is clear that this also implies that the population should be known at the moment of the definition of the objectives as well as on the moment of measurement. Depending on the subject of the policy different kinds of measurement methods may be used, varying from survey to the use of administrative data. To avoid administrative burden on the respondents Statistical Offices often use administrative data, provided for other purposes than the policy subject at hand. One should be aware of differences in the metadata by which these administrative data are determined. Quite often the data seem to refer to the same object or object group, but de facto refer to other objects or object groups.
The last step in the statistical process in the dissemination of the results of the PMP. The metadata matrix and the basic rules for transparency tell us that (1) the indicator/value should refer to the part of reality addressed by the policymaking, (2) the choice of the medium and structure of the indicator/value should meet the requirements of the receiver, (3) the meaning of the indicator/value for the receiver should be the same as meant by the Statistical System, and (4) the data once interpreted by the receiver should refer to the part of reality as meant by the Statistical System.
Finally, a few words on the use of data from earlier policymaking processes, to improve the quality of new policy decisions. Literature on Evidence-based Policy Making, show a variety of opinions on the pro's and con's of using data of earlier policymaking experiences for new policymaking (see e.g. Dorey, 2005, Cabinet Office, 1999, Pawson, 2006). In some cases, like health care (McCaughey, 2010) it is clear that results of former experiences are worthwhile to prelude upon. In situation when policymaking addresses a new situation (other circumstances, other issues) it is not always clear how evidence can be obtained. In these cases one should be very careful and not use data that are not in line with the policy that is at hand. The use of EBPM in a broad sense (Stol, 2009) forces politicians to think about their objectives in terms of
8 ) Economic Commission for Europe of the United Nations (UNECE), "Terminology on Statistical Metadata", Conference of European Statisticians Statistical Standards and Studies, No. 53, Geneva, 2000.
Sets of indicators for policy making Page 47
objects and object groups (measurable objectives), and by doing so the quality of their decisions may improve. Since they are related to reality, measurable and transparent.
4. Guidelines for transparent public policymaking
The framework for the PMP provides the means to make the PMP transparent and measurable. However, there are circumstances in which the framework can be applied with more success than in other situations. The EBPM control system must fulfill the following three requirements for effective control of the Object System of Interest ( Stol, 2009, de Leeuw, 1982).
(1) The objectives must be unambiguous and clearly formulated in terms of situations to be obtained in the Object System of Interest.
(2) Full knowledge of the Object System of Interest. This knowledge covers the current situation and the desired situation but also the way to obtain the desired situation and the status of the situation in the Object System of Interest during the Change Process.
(3) Diversity in controlling measures for an effective control of the Object System of Interest:
(a) Adaptive measures to change the organization of the Change Process.
(b) Goal changing measure to change (parts of) the objectives.
In the specification of the requirements for effective control of the Change Process in the Object System of Interest the distinction between the role and responsibilities of the Political System 9 and the government is important. The responsibility for the full control over the Change Process is shared by the Political System and the government. In the division of this task the Political System is in charge as soon as deviations occur in comparison with the original measures defined.
Although special circumstances can be assumed leading to declining evidence-based policymaking, it is hardly possible to raise arguments for denying the fact that consequences of policy decisions can be measured to verify the promises made by the politicians. (Stol 2009, p. 36)
Non-rational behavior in the evidence-based decision making process (EBDM) does not mean that the framework for the PMP is not effective. The reason is that all other parts of the framework can be executed in the flow of the proposed one. Indeed, one could argue that rational decision making leads to a better quality of the decision itself (cf. Cabinet Office, 1999). This may be true, but it does not necessary mean that the execution of the policy measures following the decision is transparent and measurable. The Garbage Can model (March and Olsen, 1979) seams to occur frequently (Koundouraki, 2007). When the decision process takes place according to the Garbage Can model the policy decision is a more or less a coincidental combination of (1) solutions, (2) problems to be solved, (3) people, and (4) choice opportunities.
As the decision process is isolated in the framework the only prerequisite for a successful use, is that the output of the decision process can be transformed in concrete objectives to be realised in an (well defined) Object System of Interest. In the day-to-day practice of public policymaking we miss sometimes this relation between the policy decision and the objective in terms of intended changes in the reality. If this relation cannot be formalized, it is not possible to
9 ) The Political System consists of politicians and the part of the government that may be organized in the same institution as the part of the government that is responsible for the control of the execution of the Change Process.
Sets of indicators for policy making Page 48
measure the outcome of the policy measures. In that case the use of Official Statistics is not is not a serious option, nor the intention of being transparent in the PMP.
We must understand that in the Political System people are chosen based on their position in a political party and the promises (in the context of the persuasiveness of their arguments) of their program. These politicians should be aware of the need of transparency and measurability of the PMP. Sometimes this may mean that they explicitly have to admit that their promises cannot be realised. The reasons for this may be (1) compromising with other politicians, (2) other circumstances than anticipated, or (3) consciously promising more than they can substantiate. When the politicians prefer to hide their deviation of the promises, they will tend to make the PMP less transparent and verifiable so that the deviations are not so visible.
We mention three circumstances that influence the effectiveness of the use of the EBPM framework: the complexity of the Object System of Interest, the complexity of the organization, and unpredictable events.
(1) The complexity of the Object System of Interest and elapsed time from the start of the Change Process until the achievement of the outcome, play an important role in the possibility to control the Change Process in an effective way (Bouchard and Caroll, 2003). Although measures can be taken to lower the complexity of the Object System, this will not always be sufficient. In some cases it is not possible to measure the effect of the policy measures because of the number of variables and uncertain behavior of the Object System in its environment.
(2) The organization in the Political System and the organization of the Change Process may be quite complex cf. Bouchard and Caroll (2003) . There are various measures to adapt the organization of the Policy System and the Change Process to the specific policy issue. Nevertheless these measures may not always work, especially when different political issues require conflicting organizations.
(3) Unpredictable events may happen or unknown effects may occur that change the Object System of Interest and/or its relevant environment during the Change Process. Taleb (2004, 2007) shows many examples of unexpected events and unknown effects of (policy) measures. Contingency planning may be effective to minimise the consequences of these events. 10
An optimal use of Official Statistics and the Statistical System in transparency of the PMP may be obtained if the following conditions are fulfilled:
(1) The Statistical System is involved in the definition of the objectives and the indicators.
(2) The Statistical System uses the same principles for the definition of the indicators mentioned in (1), as it does for the indicators on measurement.
(3) The Statistical System uses unambiguous and identical metadata for the definition of the indicators to be realised as well as for the definition of the outcomes of the PM.
(4) The Statistical System uses metadata models in which all references between reality, addressed by the PM, and the indicators are transparently shown.
10 ) Contingency planning to avoid disturbances in the Object System of Interest may be quite expensive. Compare the organisation of the Olympic Games in China where every possible circumstance has been envisaged in the preparation of the Games. The expenses have been never so high (about 25 billion euro) and probably will not be exceeded in the future. For instance, the budget for the next games in London is 13 billion euro.
Sets of indicators for policy making Page 49
Politicians and statisticians can fruitfull cooperate by using common metadata models on (1) objectives, (2) measurement, (3) dissemination of figures and (4) justification of PM. In Stol (2009, Section 6.3) examples of the metadata models on objectives, measurement and presentation (dessemination and justification) are proposed, to be used as guidelines for the cooperation between politicians and statisticians.
Sets of indicators for policy making Page 50
Literature
Althaus, C., Bridgman, P., and Davis, G. (2007). The Australian Policy handbook. Fourth edition. Allen & Unwin, Crows Nest, Australia. Bouchard, G. and Caroll, B. (2003). One Size Does not Fit All: A Contingency Approach to Policy Making. Paper presented in CPSA meeting, Halifax, Nova Scotia. http://www.cpsa- acsp.ca/paper-2003/bouchard.pdf Cabinet Office (1999). Modernising Government. Cm 4310, London: Stationary Office, UK. Checkland, P. (1999). Systems Thinking, Systems Practice . Wiley, New York, NY. ISBN 0 471 986062. Chomsky, N. (1975). Reflections on language. Panteon Books, New York, NY. Davies, H., Nutley, S. M., and Smith, P. (2001). What works?: Evidence-Based Policy and Practice in Public Services . Policy Press, Bristol, UK. Dorey, P. (2005). Policy-making in Britain . SAGE Publications. London UK, ISBN 0-7619-4903- 8 Easton, D. (1979). A Framework for Political Analysis. University of Chicago Press, Chicago. IFIP (1998). FRISCO A Framework of Information System Concepts (Eds. Falkenberg, E.D., Hesse, W., Lindgren, P., Nilsson, B.E., Han Oei, J.L., Rolland, C., Stamper, R.K., Van Assche, F.J.M., Verrijn- Stuart, A.A., and Voss, K.). International Federation for Information Processing. (ftp://ftp.leidenuniv.nl/pub/rul/fri-full.zip) Koundouraki, E. (2007). The Information Cycle in the European Commission’s policy-making process. Information Research, Vol. 12, No. 4, http://informationr.net/ir/12-4/colis/colis07.html Leeuw, A.C.J. de (1982b). Organisaties: management, Analyse, Ontwerp en verandering ,. Van Gorcum, Assen/Maastricht, the Netherlands. March, J.G., and Olsen, J.P. (1979). Ambiguity and choice in organizations . Universitetsforlaget, Bergen, Norway. McCaughey and Bruning (2010) Rationality versus reality: the challenges of evidence-based decision making for health policy makers , Implementation Science 2010 Nielen, G.C. (1976). De bedoeling van informatie voor mens en organisatie. Samsom, Alphen aan den Rijn, NL. Pawson, R. (2006). Evidence Based Policy, a Realist Perspective. Sage Publications, London, UK. ISBN 10 1 4129 1059 5 Stamper, R.K. (1996). Signs, Information, Norms and Systems, Signs of Work, (Holmqvist, P., Andersen, P.B., Klein, H. and Posner, R. (Eds.), Walter de Gruyter, Berlin, Germany. Stol, H.R. (1990). Informatieplanning in de praktijk . Samsom, Alphen a/d Rijn, Netherlands. ISBN 90 14 03644 2. Stol, H.R. (1995). Silentium en de informatiemaatschappij, Organisatie van de informa- tievoorziening: verschuivende en transparante grenzen , (ed. Mantelaers, dr. P.A.H.M., e.o.), Klu- wer, Deventer, NL, ISBN 90 267 2292 3 Stol, H.R. (2002). Metadata in data capture. Integration of Statistics in the Information Society Proceedings Statistical conference SORS , November 2002, Radenci, Slovenia. Stol, H.R. (2009). A Framework for Evidence-based Policy Making using IT, a Systems Approach. Eburon, Delft, Netherlands, ISBN 978-9059722927 Sundgren, B. (1975). Theory of data bases. Petrocelli/Charter, New York ASA, ISBN 0-88405- 307-5 Taleb, N.N. (2004). Fooled by randomness . Random House, New York, NY. ISBN 0-8129-7521-9. Taleb, N.N. (2007). The Black Swan, “the impact of the highly improbable” . Random House, New York, NY. ISBN 1400063515.
Sets of indicators for policy making Page 51
Setting the right priorities for European statistics: a contribution from a user and producer
Aurel SCHUBERT 1
The European Central Bank (ECB) conducts the euro area monetary policy with the primary objective of maintaining price stability over the medium term. It also contributes to other ESCB tasks such as the conduct of foreign exchange operations, the oversight of payment systems and the stability of the financial system. The performance of all these tasks requires a large amount of high quality statistical information. At the ECB, the statistical information is used to assess economic, monetary and financial developments in the euro area (and increasingly also in its individual Member States), to feed into the ECB macro-economic projection exercises, to assess the transmission of monetary policy decisions of the ECB to the overall economy and to analyse financial integration and financial stability, among others. A major part of the statistics used by the ECB are produced by the ECB itself, with the assistance of the EU national central banks (NCBs), which, together with the ECB, form the European System of Central Banks (ESCB). The ECB also provides the statistical support to the European Systemic Risk Board (ESRB), the newly created independent EU body responsible for the macro-prudential oversight of the overall financial system within the EU.
However, the ECB does not only use its own statistics. For the performance of its tasks, the ECB also relies heavily on statistics produced by the European Statistical System (ESS). This paper recalls some key priorities of the ECB for ESS statistics and how these statistics are used for evidence-based decision making at the ECB. It concludes that the high demand for more quantitative information as a basis for policy decisions represents good news for statisticians but also a major challenge and a major responsibility, as statisticians have to ensure that the demanded figures are “ right for the purpose ”, but also “ right on time ”.
Priority statistics for policy-making: do merits outweigh the cost?
Being a heavy user of ESS statistics, the ECB very much appreciates the current efforts of the ESS to keep European statistics focused on policy needs of highest relevance in Europe. As experience has shown, an early and close dialogue with the key users of ESS statistics is indispensable to get an informed view on the relevance and, hence, on the merits of such statistics . This is in essence what Principle 11 of the European Statistics Code of Practice (Relevance) says. For the ECB, this is one of the most important statistical quality principles.
Sets of indicators for policy making Page 52
Indeed, the ECB is fully reliant on a sound statistical basis in order to properly fulfil its tasks in an increasingly complex economic environment. This aspect will be further elaborated below.
Focusing on statistics to address the policy needs of higher relevance in Europe and re- prioritisation of statistical requirements might, however, not be enough in the current circumstances of statistical offices. An efficient priority-setting in the ESS, focusing on policy relevant statistics, goes hand-in-hand with the so-called “European approach to ESS statistics”, as spelt out in the related communication of the European Commission 2. New production methods of statistics will require investments in a modern statistical function, in particular in IT and human skills, but at the same time they will yield in the medium-term important efficiency gains and cost savings to statistical compilers . These savings are indispensable in particular in periods of resource constraints. The new production methods of statistics should be seen as an investment in the future of statistics and in its essential role in evidence-based policy making.
More modern business processes may also bring very significant gains to respondents . These include a better use of data contained in administrative sources and in business registers that are reported only once, ideally extracted directly from respondents’ systems. The approaches described above, as well as an enhanced storage (e.g. data warehousing) and an enhanced dissemination of data, may have also various advantages from the users’ perspective . For example: i) the reliance on micro-data may accommodate more flexibly and more promptly emerging ( ad hoc and regular) user needs, may reduce the timeliness of new statistical information and may increase the consistency and integration of the various data sets. We will return to this aspect later in the paper. ii) a modern data dissemination increases the accessibility of data and, thereby, increases the users’ awareness of which statistics are available and how they can be best used.
Inevitably, the current circumstances of statistical offices seem to indicate that new and more cost effective solutions alone might not be sufficient to free enough resources to accommodate growing users’ requirements. It might be required to consider, in addition, substantial simplifications and/or discontinuation of statistics that are either very costly to produce or less or no longer relevant for policy-making. Undisputed is that, when setting statistical priorities in the context of a rational use of resources, it is essential to look at the cost and at the merits of
1 The contribution of Alda Morais to an earlier version of this paper is gratefully acknowledged.
Sets of indicators for policy making Page 53
statistics in conjunction, and to give special attention to those statistics whose merits clearly outweigh their cost. And this balance will change over time, thus requiring constant monitoring.
Policy use of statistics: shaping the statistics to further increase their merits
One of the themes of this conference is whether the use of statistics should influence the processes used for disclosing statistics. As with the question on the merits vs. cost above, the answer seems to be affirmative if the relevance of their policy use does justify it. In this way, it might be possible to obtain substantial improvements in the relevance of (existing statistics) by introducing gradual changes in the processes used to produce such statistics and in the standardisation of information.
The forthcoming Excessive Imbalance Procedure (EIP) is a clear example of an important new policy use of (existing) statistics (to monitor sources of macroeconomic imbalances and ensuring appropriate corrective action when necessary). In this case, all EU countries will matter, and not only the euro area countries, and there is a need for individual country data. Already for the preventive part of the EIP, the system will be based on a scoreboard composed of a limited number of statistical indicators. While a final decision on the indicators to be included in the scoreboard is still outstanding (including their detailed definition, source, legal basis, frequency, timeliness, etc.), it is clear that an early and intense discussion between the users of these data and the statisticians will determine the “fitness for purpose” of the scoreboard: i) by choosing the most adequate indicators to meet the policy needs (in terms of their cross- country comparability, length and timeliness, for example); and ii) by ensuring that the indicators are sufficiently robust, i.e. based on a sound statistical framework and compiled in accordance with the principles laid out in the European statistics Code of Practice, rather than on an undue degree of estimation.
To some extent, ensuring the “fitness for purpose” of the scoreboard is an exercise that can be compared with the choice of indicators that have been used by the ECB in the ECB Convergence Reports. Successive waves of enlargement of the euro area (from initially 11 to 17 member countries today) has meant long lead times in the preparatory work for both the ECB and the countries joining the euro area. High quality national statistics are crucial as part of
2 “ Communication from the Commission to the European Parliament and the Council on the production method of EU statistics: a vision for the next decade”, Brussels, 10 August 2009.
Sets of indicators for policy making Page 54
EU Member States obligations to prepare for joining the European and Monetary Union (EMU) because they are also used in the assessment of their economic convergence with the EMU, which is largely based on statistical indicators.
Another example of an important policy use of statistics may be the decision-making at the ECB. Following the first Governing Council meeting in the month (when monetary decisions on the key interest rates are normally taken) the ECB President delivers his “Introductory statement” (published on the ECB website). In these statements, the President gives several examples of how statistics are used by the Governing Council to assess the economic situation and the stance of the monetary policy. From these and other ECB statements and publications, it is notable that the ECB attaches the highest importance to statistics and considers these particularly important for the well-functioning of the Economic and Monetary Union and for the financial stability in the European Union. This is particularly valid at the moment, when the strengthening of the statistical basis is seen as essential in Europe (and globally) to overcome the ongoing economic and financial crisis as soon as possible and in order to avoid a recurrence of this crisis.
Obviously, statistics lose a great part – if not all – of their relevance if they are late for decision making. If data is made available, not when they are needed to underpin the decisions, but only after the policy decisions have already been taken. (This should be without prejudice to the usefulness of quick information, e.g. obtained from opinion surveys to bridge the time needed to compile statistics.) A typical example of the “fitness for purpose” of statistics is the case of the quarterly national accounts. These should ideally be made available by 30, 60 and 90 calendar days after the reference period, in order to feed, in a timely manner, the monetary policy deliberations of the ECB Governing Council, which take place at the beginning of each month.
The important concept here is the term “fitness for purpose”. In the case of those indicators where timeliness is the most important quality feature, the use of more flash estimates could be considered, with various advantages to both users and compilers. For example, GDP flash estimates at t+30 days may be volatile at national level. In this case, however, only the results for the euro area aggregate would be used by the Governing Council, and these aggregated - and timely - figures would be expected to be of sufficiently good quality for the decision making process. Besides, the results of the GDP flash estimates at t+30 days are expected to improve substantially in quality over time, as it has happened with the release of GDP at t+45. Therefore, the ECB is happy to support and contribute, as needed, to the planned ESS study assessing the feasibility of producing such flash estimates.
Sets of indicators for policy making Page 55
The plans to advance the release of integrated euro area accounts from the current timetable (t+120) to around t+90 is yet another example of how statistics can see their merits for policy- making substantially enhanced, if they are “right on time”. These plans will allow advancing the provision of the related briefing by 1 month. In order to achieve this ambitious target, a major joint effort is needed from Member States, not only on the part of the ESS (in securing timely national contributions for quarterly non-financial sector accounts), but also on the side of the ESCB (for timely data on the financial accounts and the euro area balance of payments). Again, the outputs are expected to improve substantially in quality over time, and experience has shown that sharing best practices among Member States is instrumental in achieving such further quality improvements.
The additional effort for advancing the release of integrated euro area accounts is already reflected in the plan of the ESCB Statistics Committee, as agreed at its December 2010 meeting, to achieve a transmission deadline of t+80 for the balance of payments / international investment position by 2019 following a step-wise approach. It is, therefore, essential that the transmission deadlines for the quarterly non-financial sector accounts, quarterly government financial accounts and government debt also move in a synchronised manner. The ECB would consider prudent to make allowances for this synchronisation already in the ESA 2010 Regulation, which will be discussed at the political level within the next few weeks and months.
Last, but not least, a new policy use of (mostly existing) statistics might stem from the European Systemic Risk Board (ESRB). The ESRB will determine and analyse the information relevant and necessary to perform its wide ranging tasks. The specific ESRB data needs are still under examination but most of the new data requirements are expected to relate to financial (stability) statistics and to the detection of imbalances (in the inter-linkages and exposures of sectors and sub-sectors, countries, market segments, etc.). This will be first and foremost in the financial sector (thus, based on data to be provided by supervisory authorities or by national central banks), but also in the other economic sectors.
Without prejudice to the legal framework of the ESS in the field of statistics, ESRB data requirements might also need to encompass some enhancements in the data provided by the ESS, in order to better feed the macro-prudential analysis.
Sets of indicators for policy making Page 56
From the integrated quarterly financial and non-financial accounts, a wide range of indicators derived from the quarterly financial and non-financial accounts 3 are already used in the internal briefing material, the so-called White Book, which is provided to the ESRB on a regular basis by the ECB.
In order to address future ESRB data needs, these accounts might need to be extended, from the euro area to the whole EU and each of its Member States. In the case of financial stability, also the small Member States matter. As the economic and financial crisis has shown, shocks in small countries of the euro area can occasionally have significant effects on the whole area. Therefore, any ESS assessment of priorities should consider that raising the reporting thresholds beyond certain limits may have adverse effects on the availability of country data for national non-financial sector accounts, as this might drastically limit the number of countries actually reporting. This might turn out as saving on the wrong place.
As already mentioned in the context of the use of ESS statistics for monetary policy, in order to remain relevant for ESRB purposes, the timeliness of these accounts might also need to be advanced. Decomposing value changes in the EU financial and non-financial accounts might also bring additional value added to policy-makers, by allowing to distinguish real and price developments. Finally, the list of Principal European Economic Indicators (PEEIs) might need to be extended, namely with respect to property prices indices, in order to provide information not only on residential property, but also on commercial property, a factor that played an important role form the current financial crisis.
The repeated use of policy relevant statistics, e.g. in the ECB Convergence Reports, in the ECB President’s Introductory Statements, in Scoreboards, in ECB speeches and other public statements can be seen as an additional, new “ quasi policy tool ” of monetary policy makers. One could even argue that the traditional “ open market policy ” of central banking has been supplemented by an “ open databank policy ”, using data to influence public opinion and – ultimately – policy makers.
The future of statistics: anticipating future needs
In order to stay relevant, statistics should proactively adapt to the new needs of evolving societies. A case in point in this respect is the plan to improve the linkages between economic, social and environmental statistics in the context of the Commission communication on “GDP
3 Household debt to disposable income; household debt to total financial assets; non-financial corporations’ debt to value added and leverage ratio; household savings rate; household investment rate; non-financial corporations’ debt
Sets of indicators for policy making Page 57
and beyond” and of the Stiglitz-Sen-Fitoussi Report. Many of the initiatives in this domain, in particular as regards the recommendations in the first part of the Report (i.e. under “classical GDP issues”) are already addressed, for some time now, by the ESS and/or the ESCB, for example: i) in the joint compilation the quarterly integrated euro area sector accounts since 2007, ii) in the development of the forthcoming Eurosystem survey to collect data on household finance and consumption. These two examples of important European initiatives show the merits of proactively investing in achieving more integrated statistics, in exploring new data sources, in reutilising existing data and in presenting them in a consistent integrated framework. Priority will continue being given over the coming years to improving the completeness, geographical coverage, consistency and timeliness of the euro area accounts and to further complement them with distributional information on indebtedness, wealth and consumption of the household sector.
The financial and economic crisis confirmed that statistics need to be proactive and not simply reactive to events. The need to monitor and understand economic and financial developments and shocks in almost real time points to the potential of – and actually need for - disaggregated and detailed information. Indeed, the demand for detailed information on all sectors (including on non-financial sectors) and the growing requests for individual EU country data (in addition to the euro area aggregates) are some of the prospective data demands for financial stability purposes and to better inform the policy discussions on shocks and on the transmission mechanism of monetary policy.
The ECB experience has shown that increased use of micro databases and registers maintains the traditional statistical aggregated series but adds much in terms of flexibility for a better statistical response to these new challenges. Moreover, data need to be reported only once and the reporting burden is kept to a minimum, as the same source data can be used – and reused - more flexibly to derive various (micro and macro) statistics to meet new data needs. This presupposes an adequate confidentiality regime and a proper identification and classification of entities and instruments, based on modern data standardisation and matching principles that are a conditio sine qua non to the processing of large amounts of data.
In the case of the ESS, several similar avenues may bring better responses to users, as already hinted in the first part of this paper. These avenues may encompass the setting up of the necessary infrastructure to fully exploit and re-use data contained in administrative sources and in business registers, the integration of the various statistics (via the sharing of data and
to GDP; non-financial corporations’ leverage ratio; non-financial corporations investment share, etc.
Sets of indicators for policy making Page 58
processes) and the extracting of data directly from respondents’ systems, while ensuring that appropriate levels of statistical confidentiality remain.
In sum, standards and common tools for the efficient and secure exchange of data and metadata, be it among the ESS or between the ESS and the ESCB, are highly effective in bringing flexibility into the statistical systems. A case in point is certainly the plan for the interoperability of the Register of Institutions and Assets Database (RIAD) and the EuroGroup Register (EGR), which we expect to deliver promising results.
Concluding remarks: “A little bit late” is simply “too late”
The complexity of the modern world no longer allows conducting policy-making “ by experience or feeling ” but requires instead a sound empirical basis. The final decisions might still require a “judgement call” (“ the art of central banking ”), but this has to be based on models that create a sound and reliable basis for this final “judgement call”.
The demand for more quantitative information as a basis for policy decisions is undoubtedly good news for statisticians. But it also represents a major challenge and a major responsibility, as statisticians have to ensure that the demanded figures are “ right for the purpose ”, but also “right on time ”. Indeed, policy-making requires a timeliness pattern that is much more demanding than the pattern accepted for several other purposes (e.g. for analysis), and “ a little bit late ” is actually “ too late ”, as it makes statistics lose most, if not all, of their policy relevance.
I am convinced that statisticians have many good reasons to be proud of the impressive statistical achievements since the start of the preparation of the EMU, despite the tight statistical resources. The coordination of the two systems (the ESS and the ESCB) has been instrumental to this. A few examples of this success are the joint euro area sector accounts, the joint work on quality frameworks and the coordinated update of the umbrella Council regulations. On this occasion, I would like to conclude by praising in particular the work done in the context of the “Committee on monetary, financial and balance of payments statistics” (CMFB). The CMFB just had its 20 th birthday two weeks ago (the Council Decision establishing the CMFB dates back to 25 February 1991). The achievements of the CMFB in these 2 decades have confirmed its strategic and unique role as a body for the coordination, joint work, discussion and decision between the ESS and the ESCB. This is a solid base on which we can build our common statistical future, a future of relevant and timely statistics, for the respective policy purposes.
Sets of indicators for policy making Page 59
Indicators for EU policies: business as usual?
Andrea SALTELLI
Bruxelles, March 10-11 2011 2
Role of statistics-based knowledge in the making of EU policy
Based on: Saltelli, A➡, D Hombres, B➡, Jesinghaus, J➡, Manca, A➡, Mascherini, M➡, Nardo, M➡, and Saisana, M➡, 2010, Indicators for EU Policies➡ Business as Usual? Social Indicators Research , doidoidoi:doi ::: 10➡1007/s11205-010-9678-4➡
The problem with Lisbon
Bruxelles, March 10-11 2011 3