Consumer Price Indices

Transcription

1 Consumer Prce ndces Metodologcal note Te Consumer Prce ndex for te wole naton (NC) s based on te consumpton of te entre present populaton. Te Harmonsed ndex of Consumer Prces (HCP), calculated accordng to te EU regulatons n force, s used for te comparson of nflaton between Member States and as a key ndcator for te monetary polcy of te European Central Bank. Consumer prce ndces are calculated usng a caned Laspeyres formula, n wc te basket of products and te wegtng system are updated annually. Montly ndces for te current year are calculated wt reference to December of te prevous year (calculaton base) and subsequently caned over te perod cosen as a reference base n order to be able to measure prce trends over a perod of tme longer tan a year 1. Reference base year for NC and HCP Te NC ndces are expressed wt 2010=100 as a reference base year 2. Te HCP, on te oter and, are calculated and publsed wt 2005=100 as a reference base, as establsed by te Regulaton (EC) no 1708/2005 of te 20 t October Classfcaton for consumer expendture, basket of goods Te classfcaton of consumer spendng adopted for te consumer prce ndces s te nternatonal COCOP (Classfcaton of ndvdual Consumpton by Purpose) wose erarccal structure makes provson for tree levels of dsaggregaton: Dvsons, Groups and Classes. Startng from data referred to January 2011, te ndces are calculated accordng to a more detaled classfcaton sceme wc takes nto account, wt some adjustments, te proposed revson of te COCOP classfcaton currently beng dscussed n Europe for dsaggregaton levels lower tan Classes. Te classfcaton sceme, wc s adopted for te tree consumer prce ndces publsed by STAT, s dstngused by two addtonal lower levels of dsaggregaton, Product Sub-Classes and Consumpton segments. Consumpton segments are represented by a sample of products or groups of products tems, called Representatve tems. n 2015, tere are 618 representatve tems (1,441 products) for NC and 623 representatve tems (1,457 products) for te HCP. As regards NC, te ndces are released wt a level of detal tat reaces 326 consumpton segments; NC ndces by type of products (a classfcaton of goods and servces dfferent from te COCOP), by regulated and non-regulated products and by purcase frequency are also calculated and released. As regards HCP, te ndces are publsed wt a level of detal of te COCOP-HCP product classes, n accordance wt te publcaton carred out by Eurostat for te HCP of sngle EU countres and for te HCPs calculated for te EU and te EMU; furtermore, HCP ndces by specal aggregates (HCP-SA) are released. HCP-SA ndces are calculated usng te same classfcaton sceme and te same metod adopted by Eurostat (terefore dfferent from te metod used for te calculaton of NC ndces by type of products), n order to guarantee comparablty among te talan HCPs and te HCP of te oter EU countres and te HCPs for te EU and te euro area produced by Eurostat 3. 1 STAT calculates anoter ndex named Consumer Prce ndex for blue- and wte-collar worker ouseolds (FO) based on consumpton of ouseolds wose reference person s an employee. 2 Te FO ndces are expressed wt 2010=100 as a reference base year, too. 3 HCP-SA ndces ave been released startng from data referred to February Te descrpton of product classes wc are ncluded n te specal aggregates s avalable on Eurostat web ste at te followng lnk: ttp://ec.europa.eu/eurostat/ramon/nomenclatures/ndex.cfm?targeturl=lst_nom_dtl&strnom=hcp_2000&strlanguagecode=en &ntpckey=&strlayoutcode=.

2 All ndces are publsed n.stat, te wareouse of statstcs produced by STAT, nsde te teme Prces, sub-teme Consumer prces (ttp://dat.stat.t/). n.stat, n addton to ndces at natonal level, NC ndces at provncal, regonal and macro area level and FO ndces at provncal level are publsed. Prce collecton and calculaton metod for seasonal product prce ndces Te metod for collectng and calculatng prces of seasonal products s n accordance wt Regulaton (EC) no 330/2009 of 22 nd Aprl 2009, wc sets out mnmum standards for dealng wt seasonal products n te HCP 4. Ts metod, also used for te NC 5, s appled to te product groups and classes Frut, Vegetables, Clotng and Footwear. Te European Regulaton defnes as a seasonal product one wc, durng certan perods of te year (of at least one mont), t may not be possble to purcase, or s purcased n modest or nsgnfcant volumes by consumers. t also establses tat n a gven mont seasonal products are consdered n season or out of season. On te bass of ts standard, STAT as defned a montly calendar for te wole 2015, wc establses n a gven mont wen eac specfc product belongng to te abovementoned product groups or classes must be consdered n season or out of season. Te adopton of a seasonalty calendar entals tat te local consumer prce survey s carred out only n monts n wc te product n queston s defned as n season, wle prces of out of season products wll be estmated on te bass of a metod tat s consstent wt standards contaned n te aforementoned European regulaton. Survey geograpcal bass and rate of coverage, temporal coverage Data contrbutng to te complaton of montly consumer prce ndces are tradtonally collected n two dstnct surveys: te local survey, carred out by Muncpal Offces of Statstcs, under stat supervson and coordnaton, and te central survey, carred out drectly by stat. n 2015 te geograpcal bass of te survey s made up of 80 muncpaltes (19 regonal captals and 61 provncal captals) wc partcpate n te ndces calculaton for all te representatve tems of te basket and of oter 12 muncpaltes partcpatng n te survey for a subset of products wc ncludes local tarffs (water supply, sold waste, sewerage collecton, gas for domestc use, urban transport, tax, car transfer ownersp, canteens n scools, publc day nursery, etc.) and some local servces (buldng worker, football matces, cnema, teatre sows, secondary scool educaton, canteens n unverstes etc.). Overall, te coverage of te ndex, measured n terms of resdent populaton n te provnces wt captals partcpatng n te survey for all tems n te basket, s 83.5%. Concernng te basket subset ncludng local tarffs and some local servces wose wegt on te NC basket s equal to 6.8% wt te partcpaton of te oter 12 muncpaltes, te coverage of te survey, measured n terms of provncal resdent populaton, rses to 91.9%. n te consumer prce survey, n 2015, tere are more tan 41,300 statstcal unts (ncludng outlets, enterprses and nsttutons) were te prce of at least one product s montored, as well as around 8,000 dwellngs for observng rents. 501,900 prces are sent montly to stat by Muncpal Offces of Statstcs eac mont. Prces collected eac mont drectly by stat are 95,600; among tese, about 13,000 are collected usng web scrapng tecnques for consumer electroncs products prce collecton on nternet. Te percentage of products observed drectly by stat, calculated accordng to te wegt assgned to eac product wtn te NC, s 23.1%. Prces are collected at central level for tose products (for a total of 76 representatve tems): - tat do sow no varablty along natonal terrtory or are admnstered at natonal or regonal level (.e. tobacco, telepone servces, prescrpton medcnes, magazne and oter perodcals, some transport servces suc as natonal and regonal ralway transport); - tat are tecncally too complex to be collected at terrtoral level because of contnuous tecnology canges (.e. consumer electroncs); Te HCP-SA calculaton metod s descrbed n te HCP Compendum wc s downloadable at te followng lnk: ttp://ec.europa.eu/eurostat/documents/ / /ks-ra en.pdf/59eb2c1c-da1f-472c-b191-3d0c76521f9b?verson=1.0. Back seres startng from January 2001 are publsed on.stat, te wareouse of statstcs produced by STAT, nsde te teme Prces (ttp://dat.stat.t). 4 t as been adopted startng from data referred to January t s used for FO ndces, too. 2

3 - wose consumpton s not strctly lnked to te terrtoral areas (tourst servces suc as package oldays, batng establsment etc.). Wt regard to te local survey, prce collecton s carred out n te frst ffteen workng days: - b-montly for products wc sow a strong temporal varablty of ter prces (fres frut and vegetables, fres fs; transport fuels; gas n cylnder and eatng ol); - once a mont, for te remanng products. For some goods or servces, suc as for example, water supply, town gas and natural gas, urban transport by bus and combned urban transport, tax or tckets (contrbutons to NHS) for specalst practce, servces of medcal analyss laboratores and X-ray centres and oter paramedcal servces, t s detected te prce appled te 15 t day of te mont to wc te ndex s referred. Concernng te centralzed survey, prce collecton s wdely carred out once a mont n te frst ffteen workng days. Hereafter te exceptons to te general rule: - for some goods and servces suc as for example tobacco, games of cance, medcnes, telecommuncatons servces, regonal ralway transport, wagon lts, out of town bus servces, out of town combned passenger transport, postal servces, gway tolls, car transfer ownersp, car overaul, t s detected te prce appled te 15t day of te mont to wc te ndex s referred; - tree tmes per mont, accordng an annual calendar fxed at te begnnng of te year, for natonal ralway transport; - b-montly for passenger transport by ar, passenger transport by sea and nland waterway, local daly newspapers and magaznes; - on eac day of te mont for tourstc, recreatonal and cultural servces (fun parks entrance tcket, batng establsment, sk lfts, etc.). Wegtng structure n te table 1 te wegtng structure for te year 2015 of NC and HCP s reported. TABLE 1. WEGHTS USED FOR CALCULATNG CONSUMER PRCE NDCES, BY EXPENDTURE DVSON. YEAR 2015, percentage values Expendture dvsons Wegts Food and non-alcoolc beverages Alcoolc beverages, tobacco Clotng and footwear Housng, water, electrcty, gas and oter fuels Furnsngs, ouseold equpment and routne ouseold mantenance Healt Transport Communcaton Recreaton and culture Educaton Restaurants and otels Mscellaneous goods and servces All tems NC HCP Harmonzed ndex of consumer prces at constant tax rates Te Harmonzed ndex of Consumer Prces at constant tax rates (HCP-CT) 6 s calculated as establsed by te Regulaton (EC) no 119/2013 of te 11 t February t measures te cange of prces at constant tax rates. t follows te same computaton prncples as te HCP, but s based on prces at constant tax rates. Prces at constant tax rates are estmated cancellng out te effects due to canges n taxes n te current mont compared to te tax rates system n force n December of prevous year (calculaton perod base). 6 Te HCP-CT as been released startng from data referred to Marc Back seres startng from January 2002 are publsed on.stat, nsde te teme Prces (ttp://dat.stat.t). 3

4 Te taxes consdered n te HCP-CT are tose drectly lnked to fnal consumpton. Tey are manly VAT, excse dutes and oter taxes on some specfc tems (suc as cars and nsurance). Subsdes and taxes pad on ntermedate stages (e.g. producton, transportaton) are not taken nto account. n prncple, fort te complaton of HCP-CT, all taxes sould be ncluded and kept constant; owever, due to practcal consderaton, taxes wc generate very small tax revenues may not be taken nto account. n detal, accordng to recommendatons reported n te Eurostat HCP-CT Manual, taxes wc cover less tan 2% of te total tax revenue can be excluded. On te wole, ncluded taxes must cover a mnmum of 90% total tax revenue. Terefore n te complaton of te talan HCP-CT, taxes kept constant are te followng: VAT, excse dutes on tobacco and energy tems (fuels, eatng ol, gas, electrcty, etc.), te man local surcarge on electrcty and gas, tax for te publc lablty nsurance and contrbuton to te Natonal Healt Servce for transport means nsurance. On te bass of Natonal Accounts data taxes wc cover less tan 1% of te total tax revenue are excluded and, on te wole, taxes ncluded cover almost 98% of total revenues carred out wt taxes on fnal consumpton. Te HCP-CT covers te same goods and servces as tose covered by te HCP. Te same wegt structure s appled as for te HCP (Table 1). As HCP, t as expressed 2005=100 as a reference base year. Te HCP-CT provdes a measure of te teoretcal mpact of canges of ndrect taxes on te overall HCP nflaton. t as to be empassed tat t does not provde an exact measure of ts mpact, rater an ndcaton for ts upper lmt. n effect, te dfference between HCP and HCP-CT growt rates ponts to te teoretcal mpact of tax canges on overall HCP nflaton, assumng an nstantaneous and full pass-troug of tax rate canges on te prce pad by te consumer. t as to be ponted out tat, durng te year, te talan HCP-CT may be revsed followng ntroducton of metodologcal canges requred by ndrect taxaton system canges. Data become fnal n te next year to te reference one. ndces rates of cange calculaton Hereafter formulae for te calculaton of montly, annual and annual average rates of cange for consumer prce ndces are descrbed 7. Te HCP formulae apply also to HCP-CT. Te frst expresson concerns calculaton of rates of cange between ndces n te same reference base perod: Montly rate of cange (NC, HCP) Te montly rate of cange s te current mont s ndex n respect to te prevous mont s ndex (wt one decmal place), for example: Feb, 2012 MOR Jan, 2012 ; Feb, 2012 Round ;. 1 Jan, 2012 Annual rate of cange (NC, HCP) Te annual rate of cange s te current mont s ndex n respect to te same mont s ndex a year prevously (wt one decmal place), for example: ANR Feb, 2012 Feb, 2011 ; Feb, 2012 Round ;. 1 Feb, 2011 Annual average rate of cange (NC) Te annual average rate of cange s te current annual average ndex n respect to a prevous annual average ndex (wt one decmal place), for example: 2012 AVR 2011 ;2012 Round ; Te expressons and te roundng rules descrbed for NC are also carred out for FO. 4

5 Annual average rate of cange (HCP) For te HCP, n a dfferent way compared to NC, te annual average rate of cange s obtaned drectly from te montly ndces and terefore t s based on te unrounded annual average ndces. Ts metod, appled n complance wt Eurostat, guarantees nternatonal comparablty of data. For example: AVR 2011;2012 Round Jan, 2012 Jan, 2011 Feb, 2012 Feb, Dec, 2012 Dec, ;. 1 Te followng expresson descrbes te calculaton of montly rate of cange between ndces expressed n dfferent reference base year; t can be also used for te calculaton of te annual rate of cange and te annual average rate of cange: Montly rate of cange - ndces expressed n dfferent reference base year MOR m, Xt X X t, 1 n, j; n Round C X ; X 1 C X 1; X 2... C X 2; X ;. 1 X1 t t t t m, j X were 1 m, j s te ndex, wt one decmal place, of te mont m year j, expressed n te more remote X reference base X 1, t n, s te ndex, wt one decmal place, of te mont n year, expressed n te more recent reference base X t, and C( X ; X 1) wt =2,,t are te splcng coeffcents between contguous reference bases. Tese coeffcents are equal to te annual average ndex of te year correspondng to te new reference base expressed n te prevous base, dvded by 100. Tey are as many as base canges ave been carred out durng te consdered perod. Flas estmates of HCP: accuracy and computaton metodology Flas estmate of talan HCP (and NC) are usually publsed on te last workng day of te reference mont accordng to te Eurostat release calendar of HCP Flas estmate for euro area. Fnal data are generally publsed around 13 days later. Te am of te nflaton flas estmates s to provde a tmely nformaton on nflaton, predctng as accurately as possble te fnal HCP (and NC) annual rate of cange released about two weeks later. Te analyss of ter revsons represents an mportant tool to evaluate te correct balancng between te two qualty dmensons, tmelness and accuracy. Totally n lne wt te Eurostat Statstcs Explaned on nflaton metodology of te euro area flas estmate, ts secton analyses te accuracy of te talan HCP flas estmates and descrbes te metodology used n ter computaton. Accuracy of flas estmates Table 2 compares te flas estmates and te fnal HCP annual rates for te same reference mont. Over te last trteen monts, te maxmum dfference between te flas estmate and te fnal estmate of te annual rate of cange of te All tems HCP was 0.1. Over te same perod, wt reference to te man specal aggregates, te maxmum dfferences between te flas estmate and te fnal estmate of te annual rate of cange concerned Energy (0.5 n Aprl 2015) and Non energy ndustral goods (0.4 and 0.3 respectvely recorded n September 2014 and n January 2015). Te gest dfferences for Non energy ndustral goods togeter wt te gest frequency of revsons (10 monts out of 13 monts) are manly due to te sales dynamcs of Clotng and footwear, for wc te partal nformaton avalable as a ger mpact on te flas estmate and terefore t turns out to be less accurate. 5

6 TABLE 2. FLASH ESTMATES AND HCP ANNUAL RATES FOR THE ALL-TEMS AND MAN SPECAL AGGREGATES. SEPTEMBER 2014-SEPTEMBER 2015, percentage values (Base 2005=100) Specal aggregates Sep-14 Oct-14 Nov-14 Dec-14 Jan-15 Feb-15 Mar-15 Apr-15 May-15 Jun-15 Jul-15 Aug-15 Sep-15 Food ncludng alcool and tobacco: Processed food (ncludng alcool, tobacco) Processed food Energy Non energy ndustral goods Servces All-tems All tems excludng energy and unprocessed food (Core nflaton) All tems excludng energy, food, alcool and tobacco All tems excludng energy Flas , HCP , Flas HCP Flas HCP Flas HCP Flas HCP Flas HCP Flas , HCP Flas HCP Flas HCP Flas HCP Te Mean Absolute Devaton (MAD) provdes anoter way to measure accuracy. t s calculated as te average of te absolute dfferences between te flas estmate and te fnal HCP annual rate over te last trteen monts. Fgure 1 sows te MAD for te all-tem ndex and te man specal aggregates. Over te last trteen monts, te Non energy ndustral goods component as recorded te gest MAD (0.162 percentage ponts). FGURE 1. MEAN ABSOLUTE DEVATON BETWEEN FLASH ESTMATES AND HCP ANNUAL RATES. SEPTEMBER SEPTEMBER 2015, percentage ponts Food ncludng alcool and tobacco Processed food (ncludng alcool, tobacco) Unprocessed food Energy Non energy ndustral goods Servces All-tems All tems excludng energy and unprocessed food (Core nflaton) All tems excludng energy, food, alcool and tobacco All tems excludng energy Te drecton of nflaton s correctly predcted f bot te flas estmate and te fnal one sow ncreasng (declnng or no cangng) annual rates of cange wt respect to tose ones calculated n te prevous mont. Tere are tree possble outcomes for te comparson of te drecton of nflaton: 6

7 - te flas estmate correctly predcts te drecton of nflaton, so te predcted rse, declne or no cange n nflaton s confrmed by fnal data (denoted by ); - te flas estmate wrongly predcts te drecton of nflaton, namely t predcts an ncrease wen tere s a decrease or vce versa (denoted by ); - te flas estmate ponts to an ncrease or a decrease but te fnal annual rate of cange remans uncanged; or te flas estmate predcts no cange n nflaton but te fnal fgure ponts to an ncrease or a decrease (denoted by ). Over te last trteen monts, te flas estmate accurately predcted te nflaton s drecton n 113 out of 130 estmates. TABLE 3. FLASH ESTMATE PREDCTON CAPACTY OF THE DRECTON OF NFLATON MEASURED BY HCP. SEPTEMBER 2014-SEPTEMBER 2015 Specal Aggregates Sep-14 Oct-14 Nov-14 Dec-14 Jan-15 Feb-15 Mar-15 Apr-15 May-15 Jun-15 Jul-15 Aug-15 Sep-15 Food ncludng alcool and tobacco: Processed food (ncludng alcool, tobacco) Unprocessed food Energy Non energy ndustral goods Servces All-tems All tems excludng energy and unprocessed food (Core nflaton) All tems excludng energy, food, alcool and tobacco All tems excludng energy Computaton metodology of flas estmates For te talan HCP (and NC) flas estmate complaton, eac mont, - prces collected at local level by around 60 muncpaltes (out of 80) are used. Out of tese muncpaltes, tere are te 37 muncpaltes wc calculate te prelmnary local consumer prce ndces and publs tem ndependently, at te same tme of stat natonal CP and HCP release. Data collected by te oter 12 muncpaltes partcpatng n te survey for a subset of products (local tarffs and some local servces) are not used; tese data are used for te complaton of fnal ndces; - all prces collected drectly by STAT (va nternet and oter sources) are used. Tese prces refer to 76 representatve tems wc cover 21.4% (accordng to ter wegts) of te talan HCP basket (23.1% of te NC one). As soon as ndces are calculated for representatve tems for wc prces are collected drectly by STAT, representatve tem ndces for te muncpaltes, wc partcpate n te flas estmate of nflaton rate, are compled. For te oter muncpaltes, wc do not partcpate n te flas estmaton, representatve tem ndces are generally 8 calculated applyng to te ndces of te prevous mont, te montly rate of cange of te regonal representatve tem ndces. Te latter are calculated usng data of muncpaltes wc partcpate n te flas estmate, as follows: 8 For some representatve tems among oters, rents and local tarffs suc as water supply, sold waste, sewerage collecton, urban transport servces by road for te muncpaltes tat do not partcpate n te flas estmaton, ndces are estmated by carryng forward te prce of te prevous mont. Te adopton of ts dfferent estmaton tecnque s due to te fact tat te evoluton of prces n te oter muncpaltes of te same regon s not consdered a satsfactory proxy. 7

8 R m, a R R m, a m,a were s te elementary ndex of representatve tem at muncpalty level of te reference mont m of year a and s equal to te sare of resdent populaton n te muncpalty of regon R on te total R resdent populaton of te regon. As soon as representatve tem ndces of all muncpaltes are compled, regonal and, ten, natonal ndces are calculated (by representatve tems, by upper aggregates and for all tems). f all muncpaltes of a certan regon are not ncluded n te flas estmate, te representatve tem ndces of ts regon are calculated applyng to te ndces of te prevous mont, te montly rate of cange of natonal representatve tem ndces. Te latter are calculated usng data of regons wc partcpate n te flas estmate, as follows: were m,a R 20 m, a R m, a 20 R R 1 R R 1 s elementary ndex of representatve tem at regonal level of te reference mont (m) of R year (a) and s equal to te sare of ouseold consumpton expendture for te representatve 20 R R 1 tem n te regon R on te natonal ouseold consumpton expendture for te same representatve tem. Once representatve tem ndces of all regons are compled, natonal ndces are calculated (by representatve tems, by upper aggregates and for all tems). 8