90613 International Comparison Program [00.01] Report from the INAG Global Office 5th Technical Advisory Group Meeting April 18-19, 2011 Washington DC Table of Contents I. Summary of INAG Proceedings ..................................................................................................... 3 II. Policy Issue ..................................................................................................................................... 4 A. The Board’s request .................................................................................................................... 4 B. The previous TAG and Board decisions ..................................................................................... 4 III. INAG Terms of Reference .......................................................................................................... 4 IV. Issues for discussion ................................................................................................................... 6 2 Report from the INAG I. Summary of INAG Proceedings The first meeting of the 2011 ICP National Accounts Working Group (INAG) was held on February 16-17, 2011, at the World Bank, Washington DC. The purpose of the INAG is to support the Global Office in ensuring that GDP expenditures compiled by the countries participating in the current round of ICP comply as fully as possible with the SNA-93 so that they are as comparable as possible across the countries. Mindful that the reliability of converting national GDPs into a common currency using PPPs depends on the accuracy and comparability of GDP expenditure values at the different levels of the ICP classification, the INAG agreed on a 5-point strategy to support the countries in the compilation of their GDP expenditure statistics in a consistent and comparable manner. The 5 strategic points are summarized below. Identify problematic areas – An initial priority list of major problematic areas that emerged from the 2005 round of the ICP in countries’ national accounts was established. It includes FISIM, Actual and imputed rents for dwellings, Government consumption/occupations, Consumption of fixed capital for government, Own account production and consumption (including subsistence agriculture) and Final expenditures of non-profit institutions serving households. This list will be supplemented with inputs from Regional coordinators and finalized by March 2011. Refer policy issues to the ICP Executive Board - The ICP requirement for the maximum consistency possible among economies’ GDP data may result in some economies changing their official GDP levels as they work to remedy shortcomings in their national accounts. The INAG deliberations were carried out within the ICP Executive Board requirement that, as was the case in the 2005 ICP, the national accounts to be used for 2011 ICP purposes are the official national accounts of the economies. An important policy issue that has not been considered to date for either the 2005 or the 2011 ICP rounds is about the revision of PPPs in 2015 or beyond when official GDP estimates for 2011 have become final in most countries. Some countries have revised their GDP estimates for 2005 and these changes have not been incorporated in an update of the 2005 ICP results. This issue will be referred to the Executive Board following technical discussions in the TAG to agree on the recommendations to be submitted to the Board to inform its decisions. Develop a model reporting document for expenditure data and metadata – The Global Office will develop by August 2011 a country case study that will show how basic heading expenditure values can be estimated in practice. Each country will then estimate their expenditures, learning from that Model Report on Expenditure Statistics (MORES). Countries will also use the National Accounts Form 1 which remains obligatory, whereas the other 4 forms are seen as useful tools that economies may use if they choose to do so. However, all economies are expected to report detailed metadata on how they derived their basic heading expenditures for 2011, whether or not they use the Global Office form 2 for that purpose. Build capacity on the compilation of GDP expenditures - Countries will be trained on the National Accounts tasks to be carried out and all related approaches and methods. This will be done through technical assistance missions, on a needs basis, and regional workshops that will address: basic concepts; the estimation and reporting of expenditure 3 values and metadata; the review of selected problem areas; the treatment of comparison- resistant services; and data validation. The estimation of basic heading expenditure values will be carried out according to 5 approaches (direct estimation; extrapolation from recent year; borrowing per capita quantity; borrowing category’s structure; and using expert opinion.) About 4 workshops should be organized per region over the program life span. Validate expenditure data for quality, consistency and comparability – The validation procedures will be implemented at country, regional and global level. They will consist in basic checks, comparisons of values reported with international data bases, as well as comparisons of per capita real expenditure values. II. Policy Issue A. The Board’s request Following a brief report on the above made to the Executive Board at their February 2011 meeting, the Board resolved that: “The Global Office should submit to the Board, by April 2011, an INAG report on the critical policy issue of the impact of GDP exhaustiveness and SNA-compliant approaches of treating some expenditure areas on countries’ GDP levels, given that the ICP 2005 requirement of using exclusively official GDP figures should be carried over for ICP 2011.” The underlying issue is that some Board members understood INAG discussion on components of the national accounts identified as being understated in many economies as an attempt for ICP to depart from official GDP figures by promoting the compilation of different GDP estimates especially for ICP purposes. B. The previous TAG and Board decisions It’s worth recalling that: At its first meeting held on October 1-2, 2009 in Washington DC, the ICP 2011 TAG pronounced itself on the issue of official GDP as reported in the minutes: “40. The TAG agreed that the ICP results should be based on the GDP estimates provided to the Global Office by each country, although countries should be asked to explain any discrepancies between the national accounts data supplied for the ICP and the equivalent data held in the databases of international organisations.” The recommendation was echoed by the Executive Board at their Washington, DC meeting on October 14, 2009 and included in the meeting minutes as follows: “1- Members welcomed the NA framework for the ICP. They indicated that there will be no problem if some countries used SNA 1993 and others used SNA 2008, because the differences between the two versions are not major. The final GDP numbers and breakdowns used in computing ICP results must be those submitted by the countries.” III. INAG Terms of Reference The main functions of the INAG are: 4  to the extent possible, ensure that GDP for the reference year is forecast/estimated in compliance with the 1993 SNA requirements;  propose practical methods of splitting GDP into 155 basic headings;  propose practical methods of ensuring consistency between the estimated expenditure values and prices collected in the ICP surveys; and  to the extent possible, maximize consistency between data collected for comparison-resistant areas and GDP expenditures. The 2005 ICP showed that countries fall into three broad groups as far as their national accounts are concerned: (i) those with comprehensive expenditure-based estimates of GDP; (ii) those with comprehensive estimates of production-based GDP but with only rudimentary expenditure-based estimates; (iii) those having no GDP estimates for the reference year (or perhaps out of date and very poor quality accounts, typically production-based). The first group includes all countries in the OECD-Eurostat region plus some of the more statistically-advanced countries in other regions. Many countries in Asia, West Asia, Latin America and the Caribbean fall into the second group. The third group consists of many African countries as well as most Pacific Islands. The second and third groups include many countries that do not compile their accounts in accordance with the 1993 SNA, which the Executive Board has endorsed as the preferred basis for the national accounts GDP estimates for the 2011 ICP. The ICP has provided the impetus for many countries outside the OECD-Eurostat region to produce expenditure-based estimates for the first time (or at least for the first time since the 2005 ICP) while others are in the process of improving their expenditure-based estimates and making them more consistent with their official estimates of GDP (production-based). Regional agencies have been actively supporting countries to compile the expenditure-based estimates of GDP required for the ICP. One example is the Asian Development Bank which commenced a program in 2008 for Asian countries to produce supply-use tables which provide the basis for better estimating the 155 basic heading values. Another is the program initiated in the Caribbean to produce expenditure-based estimates. A third is the assistance being provided to the Pacific Island countries by the Australian Bureau of Statistics. In the ICP, the comparability of GDP requires its estimates to be “exhaustive”, which means all economic activities have to be included. From the 2005 ICP, the following components have been identified as being poorly measured in many countries’ accounts: the non-observed economy; subsistence production; rents; FISIM; consumption of fixed capital for government; parts of government final consumption expenditure (e.g. local government expenditures); final expenditures of non-profit institutions serving households; and basic headings that include sensitive data (e.g. defence, and illegal activities). It is not possible to quantify the impact of the shortcomings in countries’ national accounts due to these components because it differs so much from one country to another and no data are available on the omissions. However, the impact could be significant. For example, the estimates of rents in the 2005 ICP ranged from less than 1% of GDP to 5 17% of GDP, which is an implausible range. It is clear that many countries either ignore the imputed rents component or estimate it very poorly. The potential impact on comparisons between countries is obvious from the range quoted above. In an investigation a decade or so ago, Eurostat found that the coverage of subsistence production was very poor in the accounts of many European countries and there is no reason to believe that the coverage of the accounts in other regions would be any better. Regional coordinators are assisting many countries to split their GDP estimates into the 155 basic headings. The main aim of the various programs run by regional agencies is to obtain more comparable expenditure-based national accounts estimates for the ICP, particularly at the basic heading level. As noted above, in many cases this involves assisting them to produce expenditure estimates because their only estimate of GDP is production-based. Shortfalls and omissions in countries’ accounts are being uncovered during this process. As a result, deriving the basic headings may reveal shortcomings that could result in some countries changing their official GDP levels provided the shortcomings have been identified early enough (probably by about mid 2012 or thereabouts). Any improvements that countries choose to make to their GDP estimates for 2011 will also require relevant adjustments to be made to the historical time series. Some countries may display reluctance for various reasons such as the amount of work involved, limited capacity, limited skills, political issues associated with revisions to data for earlier years, etc. Whatever happens, the ICP will be using each country’s official GDP estimates for 2011. The 2005 ICP policy was that the national accounts to be used for ICP purposes are the official national accounts published by a country and the 2011 round is based on the same policy. The Global Office is also encouraging countries to provide metadata concerning the methods used to compile their basic heading values. It will not only help the Global Office to understand the reasons for apparent inconsistencies in basic heading data but will also assist countries in recognising shortcomings in their estimates. Many countries clearly have poor estimates for some basic headings, with imputed rents and those related to subsistence production potentially having the largest impact on the comparability of the ICP real expenditures. Methods to improve these and some other problematical estimates are discussed as part of the national accounts workshops. Apart from improving the ICP comparisons, any countries that use this opportunity to upgrade these components will also improve their time series national accounts. The INAG recommended that countries should use the Eurostat “tabular approach to exhaustiveness” because it provides a very useful framework for assessing the completeness of GDP estimates. The INAG meeting agreed that issues related to the accuracy of GDP (particularly gaps) should be submitted to the TAG. It is not possible to quantify the impact of omissions in any country’s accounts because it would be necessary to fully compile their accounts in order to measure the shortfalls. However, the INAG considered that the components identified above are those most likely to result in inconsistencies between countries’ GDP estimates and in their real expenditures at the basic heading level. IV. Issues for discussion Although the TAG has agreed since its first meeting in October 2009 (which was prior to the first Executive Board meeting) that only official GDP figures will be used in the 2011 6 ICP (a policy reaffirmed by the INAG), there is still concern amongst Executive Board members that the Global Office intends to pursue a different policy, hence the Board’s request.  What are the TAG views on the issue?  How does the TAG want to formulate its response to the Board?  What are the TAG’s views on incorporating countries’ revisions to GDP made after the 2011 results have been published in 2013? 7