53rd Conference of European Statisticians, 13 to 15 June 2005 Seminar on Sustainable Development Session I: The role of official statistics Discussant: Heinrich Brüngger, UNECE

Main elements of the Swedish paper NSO has been involved in many indicator initiatives SD indicators focusing on: –Resource productivity –Distributional aspects –Adaptability –Legacy to future (e.g. avoidance of debts) SD indicators presented separately (no aggregation or specification of trade-offs)

Swedish paper (ctd.) NSO have advantages in working with SD measurement: –Multidimensional –Best overview of existing data; infrastructure for processing data in many ways –Framework of official statistics ensures quality of data sources and outputs –Dissemination infrastructure –Can contribute to policy analysis (methodological know-how within NSO)

Swedish paper (ctd.) New challenges –Development of indicators for fuzzy concepts –Integration of various dimensions SDIs should vary between regions of the world and even countries of the same region

Main elements of the EUROSTAT paper No single normative interpretation of SDI possible, no shared understanding of the concrete concept of SD Indicators pyramid with three tiers, based on existing data (including from private sources): –SD strategy –Policy development –Implementation and analysis

EUROSTAT paper (ctd.) Some indicators use a modelling approach Policy analysis requires integration of various dimensions using interdisciplinary approaches and cooperation with research Role of official statistics (OS): –Development of new concepts and new tools, especially of frameworks for integrating various dimensions –Bring statistical aspects into the debate in a proactive way

EUROSTAT paper (ctd.) Role of official statistics (ctd.) –Better provision of metadata together with the results –EUROSTAT is very restrictive in producing composite indicators, in spite of high demand (no scientific foundation) –SDI should be core business for OS –OS has to engage in forecasting and model- building to be able to respond to “what – if” questions

Questions to the authors of the two invited papers Division of work between policy level (PL) and official statisticians (OS) 1.Which of three processes to SD indicators describes your case best: a.Between 90 and 100% of indicators defined at PL; tasks of OS are limited to the fine- tuning of specifications and above all to the provision of numbers

Questions to authors (ctd.) Question 1 continued b.Around 50% of indicators defined at PL, only description of components for the rest; tasks of OS like in a. for 50%, methodological development of indicators for the rest c.Only components of SD are defined at policy level; decisions about definitions of indicators and data collection completely left to OS

Questions to authors (ctd.) 2.Who made decisions about the labels of the indicators (terminology): PL or OS? 3.How was the relevance of individual indicators for sustainability ascertained so that a move in one indicator over time can be interpreted as a ceteris paribus improvement or decrease of SD? Dissemination

Questions to authors (ctd.) 4.How are the SD indicators disseminated: -as results of national official statistics resp. Community statistics like GDP or CPI; or -as result of a special tabulation responding to user-defined specifications? 5.How is the limited responsibility of OS indicated in the disseminated products? 6.What type of comments, and by whom, accompany the release of SD indicators?

Questions to the floor 1.Are indicators defined at the PL to be considered automatically as standards of official statistics (concerning definitions and terminology)? 2.In the case OS does not consider indicators defined at PL good enough, what can we do to avoid that our role remains permanently limited to data collection?

Questions to the floor (ctd.) 3.Is the use of the same indicator as part of different indicator sets, under different labels, something we should encourage or not? 4.How can we best indicate the limited responsibility/control of OS over SDI in the dissemination of results? 5.How far should the responsibility of OS go with respect to:

Questions to the floor (ctd.) -the use of subjective weights for aggregating different indicators? -the conversion of physical indicators into values in cases where there are no observable market prices (or where market prices cover only a very small and specific part that cannot be generalised)? -forecasting- or scenario-based indicators of SD that are different from stock, asset or debt concepts?