Seasonal Adjustment of the Harmonised Index of Consumer Prices - PowerPoint PPT Presentation

Seasonal Adjustment of the Harmonised Index of Consumer Prices (HICP) – Main Findings of an ESCB Task Force Martin Eiglsperger European Central Bank – DG Statistics Meeting of the Ottawa Group, Wellington, 4 - 6 May 2011 The views expressed in this presentation are those of the author and do not necessarily reflect the views of the European Central Bank.

2 Consumer Price Index Manual (ILO et al.) • Chapter 13: “Publication, Dissemination and User Relations” • “Seasonal adjustment and smoothing of the index” • Chapter 22: “The Treatment of Seasonal Products” • Para. 22.7: “… rolling year indices can be regarded as seasonally adjusted price indices …” • Para. 22.61: “more sophisticated methods” of seasonal adjustment, but rolling year indices as targets  Central Banks’ views on seasonal adjustment of consumer price indices might be somewhat different

3 Current practices in the EU EU statistical institutes: headline figure  year-on-year • rate of change, no seasonal adjustment of HICPs • Seasonal adjustment of (non-harmonised) consumer price indices: INSEE (France), Office for National Statistics (UK) and Deutsche Bundesbank (Germany), some other national central banks (in particular Banca d’Italia) for internal purposes  no particular role in the EU, different to the US and Canada • Seasonally adjusted HICPs for the euro area: ECB  published on the website, in the Monthly Bulletin …

4 ECB inflation analysis and forecasting Key breakdown: “main components” • Unprocessed food (meat, fish, fruit and vegetables) • Processed food (incl. alcohol and tobacco) • Industrial goods (excl. energy) • Services • Energy  Forecasted seasonal factors  Seasonally adjusted month-on-month rates of change  Base effects in year-on-year rates of change

5 Seasonality in HICP main components mainly present in • HICPs for unprocessed food, primarily due to seasonality in prices for fresh fruit and vegetables • HICPs for industrial goods, primarily due to sales prices for clothing and footwear but also in • HICPs for services which may be also affected by calendar effects (dating of Easter) via its holiday-related price indices while weak and/ or “erratic” seasonality is found in • HICPs for processed food (alcohol and tobacco!) • HICP for energy

6 Seasonal adjustment of euro area HICP Current ECB practice • Seasonal adjustment of HICP main components except energy • No seasonal adjustment of HICP for energy, since identifiable seasonality is currently not present • Seasonally adjusted total HICP obtained by aggregating seasonally adjusted main component indices + unadjusted energy price index • Normal HICP weights are used

7 Euro area HICP Month-on-month percentage changes unadjusted seasonally adjusted 1.5 1.0 0.5 0.0 -0.5 -1.0 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010

8 Task Force: members and mandate “Task Force on Seasonal Adjustment of National HICPs” • Members: eight EU national central banks, two statistical institutes + Eurostat as observer, chair and secretariat: ECB • Mandate: • Review and specification of existing guidelines • Treatment of seasonal breaks • Treatment of changes in VAT and excise duties • Calendar effects, e.g. in price indices for package holidays • Aggregation of seasonally adjusted component indices

9 Recommendations and guidelines

10 Seasonal breaks Euro area HICP – industrial goods: index levels unadjusted seasonally adjusted (whole series) seasonally adjusted (split factors Dec.2000/Jan.2001) 100.0 99.0 98.0 97.0 96.0 95.0 94.0 1999 2000 2001 2002

11 Taxes and excise duties Euro area HICP – Processed food excl. alcohol and tobacco: month-on-month percentage changes unadjusted seasonally adjusted 2.0 1.5 1.0 0.5 0.0 -0.5 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010

12 Calendar effects Euro area HICP – Services: month-on-month percentage changes (Easter in March: 2002, 2005 and 2008) unadjusted seasonally adjusted 1.5 1.0 0.5 0.0 -0.5 -1.0 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010

13 Aggregation (1): Issue Which weights to use for aggregating seasonally adjusted component indices? • HICP: chain-linked index, weights at prices of previous year’s December • Trend levels of seasonally adjusted HICP and unadjusted HICP should not differ systematically • Current approaches: Bureau of Labor Statistics (USA), Office for National Statistics (UK) und ECB  Use of “normal” weights

14 Aggr. (2): HICP index number concept • HCPI: Chain-linked Laspeyres-type (“Lowe”) index • Link period: December • Weights refer to a certain weight reference period, expressed in December prices  “Price-updating” of expenditure shares • EU national statistical institutes conduct either annually updating weights (quantities) or every five years (or so)

15 Aggregation (3): Example – components Example by J. Walschots (Eurostat): Unadjusted indices, Price-updating of weights to December prices Component 1 Component 2 Component 3 Total 300 250 200 150 100 50 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

16 Aggregation (4): Example – first approach Example by J. Walschots (Eurostat): Seasonally adjusted indices, price-updating with unadjusted components Component 1 Component 2 Component 3 Total 300 250 200 150 100 50 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

17 Aggregation (5): Example – second appr. Example by J. Walschots (Eurostat): Seasonally adjusted indices, price-updating with seasonally adj. components Component 1 Component 2 Component 3 Total 300 250 200 150 100 50 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

18 Aggregation (6): Example – Trend level Annual average in the year 16 • Unadjusted total index: 160.1358 • Indirectly seasonally adjusted total index, price- updating of expenditure shares with unadjusted component indices: 140.6291 • Indirectly seasonally adjusted total index, price- updating of expenditure shares with seasonally adjusted component indices: 160.1358