14. Further work
The EG DNA was launched in 2011 and has done a lot of work to develop a harmonised methodology and to explore specific methodological issues that may pose challenges in the compilation of distributional results in line with national accounts. However, there are still some areas where more work is needed. This chapter provides an overview of the main areas for further work.
Many countries are already compiling results according to the DNA methodology, but data are still missing for specific countries. In this regard, it is expected that more countries will start compiling the relevant results in the coming years in view of the new G20 Data Gaps Initiative (IMF, 2023[1]). Furthermore, the OECD and Eurostat have started working on the development of so-called centralised approaches to compile results for missing countries. These approaches rely on either publicly available micro data or micro data available to the international organisations, which are then combined with publicly available national accounts totals to arrive at DNA results. Of course, the results are sub-optimal in comparison with results that could be compiled by national statistical offices, as these would normally have more data sets at their disposal, have better knowledge of the underlying data, and have better insights in country specific circumstances that may be of relevance in imputing for missing items and in allocating any micro-macro gaps. However, the approaches would still prove useful to compile estimates for those countries for which data is not (yet) available.
In the work, Eurostat is focusing on EU countries, using data from the EU Survey on Income and Living Conditions (EU-SILC)1 and the EU Household Budget Survey (HBS)2 as their main inputs (Eurostat (2022[2])). The OECD is focusing on non-EU OECD member countries, relying on data available from the Luxembourg Income Study (LIS)3 (Zwijnenburg et al., forthcoming (2024[3])). Results for EU countries are already available and have been included in the database of Eurostat and the OECD. Results for non-EU countries are expected to become available in the course of 2023. At the same time, both institutions continue to further improve their centralised approaches, by obtaining feedback on the results and by finetuning specific elements in the methodology.
Timeliness of data is an important quality characteristic and comes at a premium in periods of rapid and important changes in the economy (such as the Great Financial Crisis and the Covid pandemic). Given the complexity of deriving DNA estimates and the time lag for many of the underlying micro data sources, DNA estimates currently suffer from relatively long time lags, often only becoming available a couple of years after the reference year. Given the huge user demand for timely data, it is important to develop nowcasting techniques to reduce the existing time lags and to ensure full relevance of the DNA results. In this regard, the new G20 Data Gaps Initiative (IMF, 2023[1]) includes the ambition for G20 economies to publish annual distributional results within 18 months after the reference period, by the end of 2026. It may be explored whether even more timely estimates may be feasible, dependent on the development of reliable nowcasting techniques.
There are already various initiatives, both at the national and international level, to nowcast more timely distributional information (see for example Office for National Statistics (2020[4]), Statistics Canada (2021[5]), Blanchet, Saez and Zucman (2022[6]) and Eurostat (2022[7])). These efforts will provide a useful starting point to explore how nowcasting techniques may be used to compile more timely DNA results.
Whereas the DNA work currently focuses on breakdowns by income quintiles, there is a large user demand for more granular breakdowns, e.g. results broken down by income decile (see the ambition for distributional results in the new G20 Data Gaps Initiative (IMF, 2023[1])) and/or percentile. The latter is deemed particularly relevant to obtain more insights into the bottom and top end of the distribution. Furthermore, in addition to breakdowns by household type and by main source of income, there is demand for breakdowns according to other types of socio-demographic characteristics, such as age and gender. It is important for the work to explore possibilities to publish at these more granular levels of detail, with the templates already including the possibilities to report data at these more granular levels.
The possibility to publish more detailed results will largely depend on the quality of the results. In this regard, it needs to be borne in mind that the allocation of micro-macro gaps and of any items for which micro data is lacking may lead to some margins of error surrounding the results, dependent on the information available to properly allocate the amounts to the relevant underlying households. These margins of error may prevent publishing at more granular levels of detail, particularly when they hamper a proper analysis of the trends. This means that more work will be needed to improve the linking between micro and macro items, to reduce gaps between the micro and macro results, to further finetune the guidance to impute for missing items and/or parts of the population, and to improve the linking across different data sources. This will reduce the margins of error surrounding the results, providing the possibility to publish results at more granular levels of detail. Furthermore, it may involve the development of sensitivity analyses to assess the sensitivity of the results to alternative assumptions for bridging micro-macro gaps and to deal with missing elements.
There is also a clear user request for more frequent results. In that regard, several countries currently only compile results every couple of years, often dependent on the availability of the underlying micro data. On the other hand, some countries are already compiling the results on an annual basis, with some even exploring publication at a quarterly frequency. It will be relevant to assess whether specific methodological guidance can be developed to assist countries in compiling results at a higher frequency. This may involve the development of interpolation techniques to assist countries to compile results for those years for which micro data may be missing. Furthermore, it may involve the development of nowcasting techniques for those reporting periods for which micro data may not yet be available.
Whereas distributional information on income, consumption and saving aligned to macroeconomic totals are essential to obtain a better understanding of how household groups are faring, it is important to broaden this work to include the wealth dimension, to provide insights into the three main dimensions of material well-being, i.e. income, consumption, and wealth. This allows policymakers to have a comprehensive overview of the economic situation of different household groups and to better attune policies to their specific needs. Furthermore, it provides compilers with the opportunity to cross-check results across income, consumption and wealth, adding to the quality of the overall results.
The Expert Group on Distributional Financial Accounts (EG DFA) has already done extensive work on developing distributional financial and non-financial balance sheets for the household sector in the euro area and EU economies, and several countries have started to develop distributional results on wealth at the national level. Furthermore, in view of the specific recommendation on distributional wealth results in the new G20 Data Gaps Initiative (IMF, 2023[1]), the OECD launched a new Expert Group on Distribution of Household Wealth (EG DHW) early 2023. This group will develop internationally harmonised templates and methodology for the compilation of household distributional wealth results on wealth in line with national accounts totals, leveraging off the work already done in this area by the EG DFA. The ambition is to have G20 economies compile distributional wealth results at decile level on an annual frequency by the end of 2026.
The EG DNA will continue its efforts in the coming years to work on these specific issues (as well as on other issues that may arise in the coming years). It is expected that this will feed into future updates of this Handbook.
References
[6] Blanchet, T., E. Saez and G. Zucman (2022), “Real-Time Inequality”, NBER Working Paper Series, No. 30229, https://github.com/thomasblanchet/real-time-inequality (accessed on 30 November 2022).
[7] Eurostat (2022), Early estimates of income inequalities, https://ec.europa.eu/eurostat/statisticsexplained/ (accessed on 30 November 2022).
[2] Eurostat (2022), Eurostat Centralised Exercise - Methodological note.
[1] IMF (2023), G20 Data Gaps Initiative, https://www.imf.org/en/News/Seminars/Conferences/g20-data-gaps-initiative (accessed on 14 March 2023).
[4] Office for National Statistics, U. (2020), Provisional estimates of income inequality in the UK for the financial year ending 2020, https://www.ons.gov.uk/peoplepopulationandcommunity/personalandhouseholdfinances/incomeandwealth/bulletins/householdincomeinequalityfinancial/financialyearending2020provisional (accessed on 30 November 2022).
[5] Statistics Canada (2021), Distributions of household economic accounts for income, consumption and saving of Canadian households, fourth quarter 2020, https://www150.statcan.gc.ca/n1/daily-quotidien/210907/dq210907a-eng.htm (accessed on 30 November 2022).
[3] Zwijnenburg, J. et al. (2024), “OECD centralised approach to calculate distributional results in line with national accounts totals - Methodology and results”.