The ECB’s requirements for consumer and house price statistics

1. The ECB’s requirements for consumer and house price statistics José Marín Arcas* (ECB, DG Economics) Reviewing the business architecture of consumer price statistics Luxembourg, 15 and 16 October 2009 * Assistance from Adrian Page (ECB, DG Statistics) is gratefully acknowledged.

2. Overview • Primary objective of the Eurosystem • How are consumer price statistics used by the ECB? • What are the ECB requirements? • What work remains ahead? • What use could the ECB make of more detailed price data? • Conclusion

3. Primary objective of the Eurosystem Article 105 of the Treaty establishing the European Community: “1. The primary object of the ESCB [Eurosystem] shall be to maintain price stability.” Governing Council of the ECB in 1998: “Price stability shall be defined as a year-on-year increase in the Harmonised Index of Consumer Prices (HICP) for the euro area of below 2%. Price stability is to be maintained over the medium term.” Governing Council of the ECB in 2003 (clarification): The Governing Council aims to maintain inflation rates at levels below, but close to, 2% over the medium term.

4. The ECB’s use of consumer price data • The HICP is a crucial input into monetary policy decisions and yardstick against which its success in achieving this objective can be measured. • Thorough analysis is made of every release to identify the key drivers behind changes in the overall index • use of breakdowns both by product and by country • analysis combined with driving forces of inflation: commodity, import, producer prices, opinion survey data, wages, profits information on government measures such as indirect taxes and administered prices • Forecasting of HICP and main components (with further disaggregation of food and energy) over a short-term horizon, HICP and HICP excluding energy over medium-term horizon • For countries outside euro area, assessment of inflation convergence as input to regular Convergence Reports

5. What are the ECB’s requirements? • The key requirements of the ECB for the HICP are: • Monthly frequency • Good timeliness • Flash estimate at t-1 • Full release at t+15 days • High degree of accuracy and precision • Given the crucial role in monetary policy stance, requirements for the HICP are higher than for any other economic statistic • High degree ofharmonisation across countries • Important both for understanding the euro area index, but also for convergence assessments for non-euro area countries

6. What work remains ahead? • Extend coverage of the flash estimate • For the ECB, flash estimate is as important as the full HICP release => arrives in time for the Governing Council meetings where interest rates are set • Breakdown of the flash into 5 main components would greatly facilitate analysis for the Governing Council meeting: • Processed, unprocessed food, non-energy industrial goods, energy, services • Quality adjustment and sampling • Current practices erode comparability and accuracy of the HICPs • Some examples:

7. What work remains ahead? – QA HICP subindex for computers (January 1999 = 100)

8. What work remains ahead? – sampling HICP subindices for package holidays in selected countries

9. What work remains ahead? – housing Treatment of owner occupied housing (OOH) in the HICP has been discussed and researched for over 10 years Irrespective of the final outcome, all economic, monetary and financial stabilitypolicy makers can agree that a high frequency, reliable and timely house price index is urgently needed The current OOH project conducted by the consumer price departments in all EU statistical officesis the most promising initiative in this area Crucial that these departments can focus sufficient resources on this project, before embarking on ambitious new ventures.

10. What work remains ahead? - summary • Therefore the most important priorities are: • breakdown of the flash estimate • harmonisation of quality adjustment and sampling • house price index for all EU countries & a solution for OOH • Aside from the long list of items on the already ambitious HICP work programme: • HICP-CT; HICP-AP; treatment of seasonal items; compliance monitoring etc. • Main topic of the conference is detailed prices: • Also long standing ECB request, but different to the idea of having average price levels…

11. Potential use of more detailed price data? • Detailed COICOP subindices (at 5/6 digit level) • Such detail is already published in many countries (BE, DE, US). • Would improve understanding and forecasting of HICP developments • Example 1: a breakdown of the liquid fuel item into petrol and diesel would allow for a more precise assessment of the impact of oil price changes on HICP inflation • Example 2: during the BSE and foot-and-mouth crises in 2001, detailed CPI information on the different meat items (pork, beef, poultry) helped but better results could have been achieved if detailed euro area information were available • Quick win: targeted increase in details for selected subindices such as in food and energy

12. Potential use of more detailed price data? • Analysis on price setting behaviour • During 2003-4, Eurosystem research project: “Inflation Persistence Network” used CPI micro data to do much valuable research on price setting in the euro area • Analysis of frequency and size of price changes • Requires monthly time series of actual price observations for specific products in specific outlets (i.e. notaverage prices) • Does not require comparability across regions or countries. • “Simply” a question of access to existing data, e.g. via an HICP research database.

13. Potential use of more detailed price data? • Communication • Most likely use by central banks of the sort of data requested by DG-SANCO (i.e. comparable average price levels) maybe not for analytical purposes but rather for communication with the general public. • Many European citizens remain convinced of a significant inflationary impact of the introduction of euro notes and coins. Evidence in the form of aggregated indices is abstract and unconvincing for many. • A limited set of average or typical prices for some micro basket of products may be useful for communication on policy outcomes. • There would be no need for these prices to be fully comparable, but they should be crucially refer back to periods before the euro cash changeover for the euro area countries. • Highly disaggregated price indices, could be an alternative to price levels.

14. Conclusions • Still long way to go with key priorities of the HICP and the development of harmonised house price indices • There is limited need for average price level data by monetary policy makers (communications a possible exception). • Broad consensus amongst NSIs that reliable and comparable average price level data would require significant resources. • Embarking on such a new and ambitious venture should not be considered unless sufficient new resources are provided to NSIs.