Consumer prices: final data

Size: px
Start display at page:

Download "Consumer prices: final data"

Transcription

1 16 May 2018 Consumer prices: final data April 2018 In April 2018, te Italian consumer price index for te wole nation (NIC) increased by 0.1% on montly basis and by 0.5% compared wit April 2017 (it was +0.8% in Marc 2018), confirming te flas estimate. In April, te slowdown of te growt on annual basis of All items index was mainly due to te trend reversal of prices of Regulated energy products (from +5.0% in Marc to -1.2%) and Services related to transport (from +2.5% to -0.7%), strengtened by te prices of Services related to communication (-0.7%) wic also reversed te trend from te previous mont (+0.4%). Te slowdown would ave been wider witout te acceleration of prices of Processed food including alcool (from +1.2% to +1.8%) and te reversal trend of Unprocessed food (from -0.4% to +0.7%). Inflation excluding energy and unprocessed food (core inflation) was +0.5% (down from +0.7% in Marc) as inflation excluding energy (+0.5% as in te previous mont). Te increase on montly basis of All items index was mainly due to te rises of prices of Non-regulated energy products (+1.1%), Services related to recreation, including repair and personal care (+1.0%, wic were affected by seasonal factors) and Processed food including alcool (+0.7%), almost completely offset by te decrease of prices of Regulated energy products (-5.5%). Te annual rate of cange of prices of Goods was +0.7% (te same value registered in te previous mont) and tat one of prices of Services was +0.3% (down from +0.9%). As a consequence, te inflationary gap between Services and Goods was negative and equal to -0.4 percentage points (+0.2 in te previous mont). Prices of Grocery and unprocessed food increased by 0.4% on montly basis and by 1.2% on annual basis (up from +0.4% in Marc). In April 2018, according to preliminary estimates, te Italian armonised index of consumer prices (HICP) increased by 0.5% compared wit Marc and by 0.6% wit respect to April 2017 (it was +0.9% in te previous mont). Te increase on montly basis was mainly due to te final end of te winter sales of Cloting and footwear (+3.4% compared wit Marc 2018), wic are not taken into account in te national index NIC. In April core inflation, measured by Italian HICP was +0.5% (down from +0.8% in Marc). Also inflation excluding energy, food, alcool and tobacco (+0.2%) and inflation excluding energy (+0.6%) decelerated (bot of tem down from +0.7% in te previous mont). In April 2018, te Italian armonized index of consumer prices at constant tax (HICP-CT) increased by 0.5% compared wit te previous mont and by 0.6% wit respect to April Terefore, te difference between HICP and HICP-CT growt rates 1 wic incorporates te effects of canges in indirect taxes occurred in te last twelve monts was null. ITALIAN CONSUMER PRICE INDICES. April 2018 (base 2015=100) INDICES April 2018 Mar 18 Italian consumer price index for te wole nation (NIC) Italian armonized index of consumer prices (HICP) Te difference between te rates of cange of HICP and HICP-CT represents te upper limit of te impact of canges in indirect taxes occurred in te last twelve monts on HICP, assuming teir full and instantaneous pass-troug on prices paid by consumers.

2 TABLE 1. ITALIAN CONSUMER PRICE INDEX FOR THE WHOLE NATION (NIC), BY ECOICOP DIVISION. April 2018, weigts, indices and percentage canges (base 2015=100) EXPENDITURE DIVISIONS Weigts Indices Mar 18 Mar-18 Food and non-alcoolic beverages 165, Alcoolic beverages. tobacco 30, Cloting and footwear 72, Housing, water, electricity, gas and oter fuels 107, Furnisings, ouseold equipment and routine ouseold maintenance 71, Healt 84, Transport 146, Communication 25, Recreation and culture 77, Education 9, Restaurants and otels 117, Miscellaneous goods and services 91, ALL ITEMS 1,000, TABLE 2. ITALIAN CONSUMER PRICE INDEX FOR THE WHOLE NATION (NIC), BY TYPES OF PRODUCT. April 2018, weigts, indices and percentage canges (base 2015=100) SPECIAL AGGREGATES Weigts Indices Mar-18 Mar-18 Food including alcool: 175, Processed food including alcool 105, Unprocessed food 69, Energy: 88, Regulated energy products 43, Non-regulated energy products 45, Tobacco 20, Non energy industrial goods: 255, Durable goods 88, Non-durable goods 64, Semi-durable goods 102, Goods 539, Services related to ousing 74, Services related to communication 19, Services related to recreation, including repair and personal care 178, Services related to transport 77, Services - miscellaneous 111, Services 460, ALL ITEMS 1,000, All items excluding energy and unprocessed food (Core inflation) 841, All items excluding energy, food, alcool and tobacco 715, All items excluding energy 911, Grocery and unprocessed food 197,

3 TABLE 3. ITALIAN HARMONIZED CONSUMER PRICE INDEX (HICP), BY ECOICOP DIVISION. April 2018, weigts, indices and percentage canges (base 2015=100) EXPENDITURE DIVISIONS Weigts Indices Mar 18 Mar-18 Food and non-alcoolic beverages 175, Alcoolic beverages, tobacco 32, Cloting and footwear 83, Housing, water, electricity, gas and oter fuels 114, Furnisings, ouseold equipment and routine ouseold maintenance 75, Healt 42, Transport 155, Communication 26, Recreation and culture 60, Education 10, Restaurants and otels 124, Miscellaneous goods and services 97, ALL ITEMS 1,000, All items at constant tax rates 1,000, TABLE 4. ITALIAN HARMONIZED CONSUMER PRICE INDEX (HICP), BY SPECIAL AGGREGATES. April 2018, weigts, indices and percentage canges (base 2015=100) SPECIAL AGGREGATES Weigts Indices Mar 18 Mar-18 Food, alcool and tobacco: 208, Processed food (including alcool and tobacco) 116, Unprocessed food 92, Energy: 94, Electricity, gas, solid fuels and eat energy 50, Liquid fuels and fuels and lubricants for personal transport equipment 44, Non-energy industrial goods: 264, Durable goods 84, Non-durable goods 63, Semi-durable goods 116, Goods 567, Services related to ousing 79, Services related to communication 26, Services related to recreation, including repairs and personal care 167, Services related to transport 81, Services - miscellaneous 77, Services 432, ALL ITEMS 1,000, All items excluding energy and unprocessed food (Core inflation) 813, All items excluding energy, food, alcool and tobacco 697, All items excluding energy 905,

4 TABLE 5. REVISIONS OF CONSUMER PRICE INDICES. April 2018, indices and percentage canges (base 2015=100) Italian consumer price index for te wole nation (NIC) Italian armonized index of consumer prices (HICP) Flas estimates Final data INDICES RATES OF CHANGE% INDICES RATES OF CHANGE % April 2018 Mar-18 April 2018 Mar For more details please refer to te Italian version Date of previous release: 30 April 2018 Date of next release: 31 May 2018 Contact person: Rosabel Ricci (rosabel.ricci@istat.it) Istat Italian National Institute of Statistics Via Cesare Balbo Rome, Italy pone

5 Consumer Price Indices Metodological note Te Consumer Price Index for te wole nation (NIC) is based on te consumption of te entire present population. Te Harmonised index of Consumer Prices (HICP), calculated according to te EU regulations in force, is used for te comparison of inflation between Member States and as a key indicator for te monetary policy of te European Central Bank. Consumer price indices are calculated using a cained Laspeyres formula, in wic te basket of products and te weigting system are updated annually. Montly indices for te current year are calculated wit reference to December of te previous year (calculation base) and subsequently cained over te period cosen as a reference base in order to be able to measure price trends over a period of time longer tan a year 2. Reference base year for NIC and HICP Te NIC indices are expressed wit 2015=100 as a reference base year 3. Te HICP are calculated and publised wit 2015=100 as a reference base, as establised by te EU Regulation 2015/2010 of te European Commission of 11 November Classification for consumer expenditure, basket of goods Classification of consumption so far used for HICP, NIC and FOI is te international classification ECOICOP (European Classification of Individual Consumption by Purpose), wose ierarcical structure as 4 levels of disaggregation: Divisions, Groups, Classes of product and Subclasses of product. Since te final data of January 2016, Istat as been adopted te classification ECOICOP, annexed to te new European framework regulation on armonised indices of consumer prices and te ouse price index, (2016/792), tat introduced an additional level of detail, te subclasses of product. Te 2018 basket for te Italian consumer price index for te wole nation (NIC) and for blue and witecollar ouseolds (FOI) is made up of 1,489 elementary products, wic are grouped into 920 products and into 404 product aggregates (tey were 1,481 in 2017, grouped into 920 products and 405 product aggregates). TABLE 1. CLASSIFICATION NIC AND FOI INDICES. Year 2018 Year expenditure divisions 43 product groups 102 product classes 230 product sub-classes 303 consumption segments 404 product aggregates 920 products 1,489 elementary products 2 ISTAT calculates anoter index named Consumer Price Index for blue - and wite-collar worker ouseolds (FOI) based on consumption of ouseolds wose reference person is an employee. 3 Te FOI indices are expressed wit 2015=100 as a reference base year, too. 5

6 Te 2018 basket for te Italian armonized index of consumer prices (HICP) is made up of 1,506 elementary products, wic are grouped into 923 products and ten into 408 product aggregates (tey were 1,498 in 2017, grouped into 923 products and 409 product aggregates) 4. Segments of consumption are te most disaggregated level for wic NIC indices referring to te entire national territory are disseminated. For HICP indices, te level of detail of te dissemination is tat of te product classes (te dissemination of HICP subclass indices is expected to start in 2018). FOI national indices are disseminated at level of expenditure divisions. At local level (geograpical area, region, province), NIC indices are publised up to te product groups and FOI indices, just at provincial level, up to te divisions. Furtermore, HICP indices by special aggregates (HICP-SA) are released. HICP-SA indices are calculated using te same classification sceme and metod adopted by Eurostat (terefore different from te metod used for te calculation of NIC indices by types of product), in order to guarantee comparability among te Italian HICPs and te HICP of te oter EU countries and te HICPs for te EU and te euro area produced by Eurostat 5. All indices and data are available and publised on Istat data wareouse, I.Stat, inside te teme Prices and subteme Consumer prices. In addition to indices at national level, NIC indices at provincial, regional and macro area level and FOI indices at provincial level are publised too. Price collection and calculation metod for seasonal product price indices Te metod for collecting and calculating prices of seasonal products is in accordance wit Regulation (EC) no 330/2009 of 22 nd April 2009, wic sets out minimum standards for dealing wit seasonal products in te HICP 6. Tis metod, also used for te NIC 7, is applied to te product groups and classes Fruit, Vegetables, Cloting and Footwear. Te European Regulation defines as seasonal product tat one consumers may not purcase in certain periods of te year (at least one mont), or tey may purcase in modest or insignificant volumes. It also establises tat in a given mont seasonal products are considered in season or out of season. On te basis of tis standard, Istat defines a montly calendar for te wole year, wic establises, in a given mont, wen eac specific product belonging to te above mentioned product groups or classes must be considered in season or out of season. Te adoption of a seasonality calendar entails tat te local consumer price survey is carried out only in monts wen te product in question is defined as in season, wile prices of out of season products will be estimated on te basis of a metod tat is consistent wit standards contained in te aforementioned European Regulation. 4 Te difference between te two baskets is due to two elements: on one and in te HICP basket (but not in te NIC/FOI one), contribution to te NHS for parmaceutical products, specialist practices and services of medical analysis (six items) are included; on te oter and in te NIC/FOI basket (but not in te HICP one), Games of cance are included. 5 HICP-SA indices ave been released starting from data referred to February Te description of product classes wic are included in te special aggregates is available on Eurostat web site at te following link: ttp://ec.europa.eu/eurostat/ramon/nomenclatures/index.cfm?targeturl=lst_nom_dtl&strnom=hicp_2000&strlanguagecode=e N&IntPcKey=&StrLayoutCode=. Te HICP-SA calculation metod is described in te HICP Compendium wic is downloadable at te following link: ttp://ec.europa.eu/eurostat/documents/ / /ks-ra en.pdf/59eb2c1c-da1f-472c-b191-3d0c76521f9b?version=1.0. Back series starting from January 2001 are publised on I.Stat, te wareouse of Istat statistics, inside te teme Prices. 6 It as been adopted starting from data referred to January It is used for FOI indices, too. 6

7 Survey geograpical basis, rate of coverage and frequency of data collection Data contributing to te calculation of montly consumer price indices are traditionally collected using different sources: te local survey, carried out by municipal statistics offices, under Istat supervision and coordination; te central survey carried out directly by Istat or troug different data providers; te scanner data; te administrative sources. In 2018 te weigt of te products exclusively collected by te local survey is equal to 60.7% despite of tose products wic are collected by central survey, wose weigt is 23.9%. In addition to tese two ways te acquisition of scanner data wit regard to te distribution cannel of ypermarkets and supermarkets (for 55.4%), and local survey wit regard to oter types of points of sale (for te remaining 44.6% of grocery) is used for grocery products. Finally, an administrative source is used: te database of fuel prices of Ministry of Economic Development wose weigt is equal to 3.9%. Local survey In 2018 te geograpical basis of te survey is made up of 79 municipalities wic contribute to te indices calculation of all te product aggregates included in te basket - and of oter 17 municipalities 8 participating in te survey for a subset of products wic includes local tariffs (water supply, solid waste, sewerage collection, gas for domestic use, urban transport, taxi, car transfer ownersip, canteens in scools, public day nursery, etc.) and some local services (sport events, cinemas, teatre sows, secondary scool education, canteens in universities, etc.). For te wole basket, te coverage of te index in terms of resident population in te provinces wose cief towns take part in te survey is 83.2%. Te participation of provincial cief towns is total for six regions (Valle D Aosta, Trentino-Alto Adige, Veneto, Friuli-Venezia Giulia, Emilia-Romagna and Umbria) but it is still incomplete for te oters, in particular in Puglia (40.8%), Abruzzo (47.7%) and Sardegna (56.0%). Starting from December 2017, Campobasso, regional cief town of Molise as extended te survey to te wole basket improving te coverage of te survey, wic in 2018 goes back to including all Italian regions. At te macro-area level, coverage is total in te Nort-East; it is equal to 89.4% in te Nort-West, 83.3% in te Centre, 65.7% in te Sout and 75.3% in te Islands. Concerning te basket subset including local tariffs and some local services - wose weigt on te NIC basket is equal to 6.1% - wit te participation of te oter 17 municipalities te coverage of te survey. measured in terms of provincial resident population rises to 92.5%. Te participation becomes total for 13 regions and it is stable in te remaining regions. Central survey In 2018 prices/quotes collected eac mont directly by Istat are more tan 153,000, of wic: 152,700 via web, also using web scraping tecniques, or collecting data from different providers. Te main data providers for te central price data collection are te following: Italian Customs Agency, for Tobacco products and games of cance; Italian Association of Concessionaries Higways and Tunnels (Aiscat), for motorway tolls; Farmadati, for parmaceutical products; Italian Association of Publisers (AIE), for prices of scool books; Specialized magazine Quattroruote for prices quotes of cars and second and cars; Sanguinetti Editore, for prices of cars, motorcycles and motorbikes, caravans and motoromes; GfK Italia S.r.l., for information on a large sub-set of tecnical consumer goods; 8 Tey are Asti, Cieti, Fermo, Foggia, Frosinone, Isernia, L Aquila, Matera, Monza, Prato, Ragusa, Salerno, Savona, Termoli, Vasto, Verbania and Vibo Valentia. 7

8 about 400 quotes directly provided by insurance companies wic refer to protection against most risks connected to property, suc as fire, teft and oter damages and are used for te Housing insurance services price index compilation. Te percentage of products observed directly by Istat calculated according to te weigt assigned to eac product witin te NIC is 23.9% in 2018 (23.6% in te previous year). Concerning te central price collection te main canges in 2018 refer to te survey design for national rail transport and passenger transport by air price collection and te use of a new data source for te index compilation of a large sub-set of tecnical consumer goods (te new data base, supplied by GfK Italia S.r.l., wit price data wic refer to more tan a million of purcases made on bot e-commerce sites and pysical stores, per mont). Scanner data Starting from January 2018 Istat introduces scanner data of grocery products (excluding fres food) in te production process of estimation of inflation. Tis innovation concerns 79 indices of aggregate of products belonging to 5 ECOICOP Divisions (01, 02, 05, 09, 12). In agreement wit retail trade cains (RTCs) and wit te collaboration of te Association of modern distribution and Nielsen, scanner data for 1,781 outlets (510 ypermarkets and 1,271 supermarkets) of te main 16 RTCs covering te entire national territory are montly collected by Istat on a weekly basis at item code level. For te selection of te sample of outlets a probabilistic design was implemented. Outlets were stratified according to provinces (107), cains (16) and outlet-types (ypermarket, supermarket) in 888 strata. Probabilities of selection were assigned to eac outlet based on te corresponding turnover value. Concerning te selection of te sample of items, a static approac tat mimics traditional price collection metod as been adopted. Specifically, a cut off sample of barcodes (GTINs) as been selected witin eac outlet/aggregate of products (covering 40% of turnover but selecting no more tan te first 30 GTINs in terms of turnover). Te products selected in December are kept fixed during te following year. A tank of potentially replacing outlets (258) and GTINs (until a coverage of 60% of turnover witin eac outlet/aggregate) as been detected in order to better manage te possible replacements during About 1,370,000 price quotes are collected eac week to estimate inflation. For eac GTIN, prices are calculated taking into account turnover and quantities (weekly price=weekly turnover/weekly quantities). Montly prices are calculated wit aritmetic mean of weekly prices weigted wit quantities. Scanner data (SD) indices of aggregate of products are calculated at outlet level as unweigted Jevons index (geometric mean) of GTINs elementary indices. Provincial SD indices of aggregate of products are calculated wit weigted aritmetic mean of outlet indices using sampling weigts. Finally, for eac aggregate of products, SD indices and indices referred to oter cannels of retail trade distribution are aggregated wit weigted aritmetic mean using expenditure weigts. Administrative sources In 2018 automotive fuels price indices (te weigt on te basket is 3.9%) are calculated using te data base supplied by te Ministry of Economic Development tat collects prices for tese products. More tan 63,000 price quotes are montly used to estimate inflation and tey come from about 13,240 fuel stations on te territory, tat is 65.9% of te ones active and present in Ministry database. Te 13,240 fuel stations cover te entire national territory and tey are located in te different geograpical areas as it follows: 3,500 in te Nort-West; 3,100 in te Nort-East; 2,900 in te Centre; almost 2,400 in te Sout and about 1,300 in te Islands. 8

9 Frequency of data collection Wit regard to te local survey, price collection is carried out in te first fifteen working days: bi-montly for products wic sow a strong temporal variability of teir prices (fres fruit and vegetables, fres fis; gas in cylinder and eating oil); one a mont, for te remaining products. For some goods or services, suc as for example, water supply, town gas and natural gas, urban transport by bus and combined urban transport, taxi or tickets (contributions to NHS) for specialist practice, services of medical analysis laboratories and X-ray centres and oter paramedical services, it is detected te price applied te 15t day of te mont to wic te index is referred. collecting tree prices for mont for te otel bedroom referring to te first tree Saturday of te mont; Concerning te centralized survey, price collection is widely carried out once a mont in te first fifteen working days. Hereafter te exceptions to te general rule: twice a mont, according an annual calendar fixed at te beginning of te year, for national railway transport; bi-montly for passenger transport by air, passenger transport by sea and inland waterway and magazines; from te 9 t to te 15 t day of eac mont for daily newspapers; on eac day of te mont for touristic, recreational and cultural services (fun parks entrance ticket, bating establisment, ski lifts, etc.); twice a mont, for tecnical consumer goods by GfK Italia S.r.l.; concerning te data base supplied by te Ministry of Economic Development, automotive fuel prices applied on te first and te tent working day of eac mont are used to compile consumer price indices. about te grocery products for ypermarkets and supermarkets te average weekly prices are collected, troug scanner data, in te first tree full weeks of referring mont. Weigting structure In Table 1 te weigting structure for te year 2018 of NIC and HICP is reported. TABLE 1. WEIGHTS USED FOR CALCULATING CONSUMER PRICE INDICES. BY EXPENDITURE DIVISION. Year 2018, percentage values EXPENDITURE DIVISIONS WEIGHTS Food and non-alcoolic beverages Alcoolic beverages. tobacco Cloting and footwear Housing. water. electricity. gas and oter fuels Furnisings. ouseold equipment and routine ouseold maintenance Healt Transport Communication Recreation and culture Education Restaurants and otels Miscellaneous goods and services All items NIC HICP 9

10 Harmonized index of consumer prices at constant tax rates Te Harmonized Index of Consumer Prices at constant tax rates (HICP-CT) 9 is calculated as establised by te Regulation (EC) no 119/2013 of te 11 t February It measures te cange of prices at constant tax rates. It follows te same computation principles as te HICP, but it is based on prices at constant tax rates. Prices at constant tax rates are estimated cancelling out te effects due to canges in taxes in te current mont compared to te tax rates system in force in December of previous year (calculation period base). Te taxes considered in te HICP-CT are tose directly linked to final consumption. Tey are mainly VAT, excise duties and oter taxes on some specific items (suc as cars and insurance). Subsidies and taxes paid on intermediate stages (e.g. production, transportation) are not taken into account. In principle, for te compilation of HICP-CT, all taxes sould be included and kept constant; owever, due to practical consideration, taxes wic generate very small tax revenues may not be taken into account. In detail, according to recommendations reported in te Eurostat HICP-CT Manual, taxes wic cover less tan 2% of te total tax revenue can be excluded. On te wole, included taxes must cover a minimum of 90% total tax revenue. Terefore in te compilation of te Italian HICP-CT, taxes kept constant are te following: VAT, excise duties on tobacco and energy items (fuels, eating oil, gas, electricity, etc.), te main local surcarge on electricity and gas, tax for te public liability insurance and contribution to te National Healt Service for transport means insurance. On te basis of National Accounts data taxes wic cover less tan 1% of te total tax revenue are excluded and, on te wole, taxes included cover almost 98% of total revenues carried out wit taxes on final consumption. Te HICP-CT covers te same goods and services as tose covered by te HICP. Te same weigt structure is applied as for te HICP (Table 1). As HICP, it as expressed in 2015=100 as a reference base year. Te HICP-CT provides a measure of te teoretical impact of canges of indirect taxes on te overall HICP inflation. It as to be empasised tat it does not provide an exact measure of tis impact, rater an indication for its upper limit. In effect, te difference between HICP and HICP-CT growt rates points to te teoretical impact of tax canges on overall HICP inflation, assuming an instantaneous and full passtroug of tax rate canges on te price paid by te consumer. It as to be pointed out tat, during te year, te Italian HICP-CT may be revised following introduction of metodological canges required by indirect taxation system canges. Data become final in te next year to te reference one. Indices rates of cange calculation Hereafter formulae for te calculation of montly, annual and annual average rates of cange for consumer price indices are described 10. Te HICP formulae apply also to HICP-CT. Te first expression concerns calculation of rates of cange between indices in te same reference base period: Montly rate of cange (NIC, HICP) Te montly rate of cange is te current mont s index in respect to te previous mont s index (wit one decimal place), for example: MOR I I Feb, 2012 Jan, 2012 ;I Feb, 2012 Round ;. 1 I Jan, 2012 Annual rate of cange (NIC, HICP) Te annual rate of cange is te current mont s index in respect to te same mont s index a year previously (wit one decimal place), for example: ANR I Feb, 2012 I Feb, 2011 ;I Feb, 2012 Round ;. 1 I Feb, Te HICP-CT as been released starting from data referred to Marc Back series starting from January 2002 are publised on I.Stat, inside te teme Prices ttp://dati.istat.it. 10 Te expressions and te rounding rules described for NIC are also carried out for FOI. 10

11 Annual average rate of cange (NIC) Te annual average rate of cange is te current annual average index in respect to a previous annual average index (wit one decimal place), for example: AVR I 2012 I2011 ;I2012 Round ;. 1 I2011 Annual average rate of cange (HICP) For te HICP, in a different way compared to NIC, te annual average rate of cange is obtained directly from te montly indices and terefore it is based on te unrounded annual average indices. Tis metod, applied in compliance wit Eurostat, guarantees international comparability of data. For example: AVR I 2011 ;I 2012 I Jan, 2012 I Feb, I Dec, ;. I Jan, 2011 I Feb, I Dec, 2011 Round 1 Te following expression describes te calculation of montly rate of cange between indices expressed in different reference base year; it can be also used for te calculation of te annual rate of cange and te annual average rate of cange: Montly rate of cange - Indices expressed in different reference base year X, j; X t I I X t ;. 1 1 MOR I m I n, n, Round CX t ; X t 1 CX t1; X t2... CX 2; X1 X1 m, j X were I 1 m, j is te index, wit one decimal place, of te mont m year j, expressed in te more remote X reference base X 1, I t n, is te index, wit one decimal place, of te mont n year, expressed in te more recent reference base X t, and C( X i ; X i 1) wit i=2..t are te splicing coefficients between contiguous reference bases. Tese coefficients are equal to te annual average index of te year corresponding to te new reference base expressed in te previous base, divided by 100. Tey are as many as base canges ave been carried out during te considered period. Flas estimates of HICP: accuracy and computation metodology Flas estimate of Italian HICP (and NIC) are usually publised on te last working day of te reference mont according to te Eurostat release calendar of HICP flas estimate for euro area. Final data are generally publised around 13 days later. Te aim of te inflation flas estimates is to provide a timely information on inflation, predicting as accurately as possible te final HICP (and NIC) annual rate of cange released about two weeks later. Te analysis of teir revisions represents an important tool to evaluate te correct balancing between te two quality dimensions, timeliness and accuracy. Totally in line wit te Eurostat Statistics Explained on Inflation metodology of te euro area flas estimate, tis section analyses te accuracy of te Italian HICP flas estimates and describes te metodology used in teir computation. 11

12 Accuracy of flas estimates Table 2 compares te final HICP annual rates of cange and te flas estimates for te same reference mont. Over te last tirteen monts, te maximum difference between te final HICP all items annual rates of cange and te flas estimate all items was -0.2 bot in February and Marc Over te same period, wit reference to te main special aggregates, te maximum differences between final HICP annual rates of cange and te flas estimates concerned Food, including alcool and tobacco (-0.7 in Marc 2018), Processed food (including alcool, tobacco) (-1.1 in Marc 2018), Unprocessed food (-0.3 in Marc 2018), Energy (+0.7 in January 2018) and Non energy industrial goods (-0.7 in February 2018). Te igest frequency of revisions for Non energy industrial goods (8 monts out of 13) are mainly due to te seasonal sales dynamics of Cloting and footwear, for wic te partial information available as a iger impact on te flas estimate and terefore it turns out to be less accurate. TABLE 2. FLASH ESTIMATES AND HICP ANNUAL RATES FOR THE ALL-ITEMS AND MAIN SPECIAL AGGREGATES April April 2018, percentage values (base 2015=100) Special aggregates May-17 Jun-17 Jul-17 Aug-17 Sep-17 Oct-17 Nov-17 Dec-17 Jan-18 Feb-18 Mar-18 Food including alcool and tobacco: Processed food (including alcool, tobacco) Unprocessed food Energy Non energy industrial goods Services All-items All items excluding energy and unprocessed food (Core inflation) All items excluding energy, food, alcool and tobacco All items excluding energy Flas ,0 HICP Flas HICP Flas HICP Flas HICP Flas HICP Flas HICP Flas , HICP , Flas HICP Flas HICP Flas HICP Te Mean Absolute Deviation (MAD) provides anoter way to measure accuracy. It is calculated as te average of te absolute differences between te final HICP annual rates of cange and te flas estimates over te last tirteen monts. Figure 1 sows te MAD for te all-item index and te main special aggregates. Over te last tirteen monts Processed food (including alcool, tobacco) (0.154 percentage points), Food including alcool and tobacco and Non energy industrial goods (0.108 percentage points bot) and ave recorded te igest MADs. 12

13 FIGURE 1. MEAN ABSOLUTE DEVIATION BETWEEN FLASH ESTIMATES AND HICP ANNUAL RATES. April April 2018, percentage points Computation metodology of flas estimates For te Italian HICP (and NIC) flas estimate compilation, eac mont. - prices collected at local level by 60 municipalities (out of 79) are used. Out of tese municipalities, tere are te 37 municipalities wic calculate te preliminary local consumer price indices and publis tem independently, at te same time of Istat national CPI and HICP release. Data collected by te oter 17 municipalities participating in te survey for a subset of products (local tariffs and some local services) are not used; tese data are used for te compilation of final indices; - all prices collected directly by ISTAT (via internet and oter sources) are used. As soon as indices are calculated for aggregate products for wic prices are collected directly by ISTAT, product aggregate indices for te municipalities, wic participate in te flas estimate of inflation rate, are compiled. For te oter municipalities, wic do not participate in te flas estimation, product aggregate indices are generally 11 calculated applying to te indices of te previous mont, te montly rate of cange of te regional product aggregate indices. Te latter are calculated using data of municipalities wic participate in te flas estimate, as follows: R I m, a i i R ir i I i m,a were i I is te elementary index of product aggregate at municipality level i of te reference mont m of year a and i is equal to te sare of resident population in te municipality i of region R on te i ir total resident population of te region. m, a 11 For some product aggregates among oters, rents and local tariffs suc as water supply, solid waste, sewerage collection, urban transport services by road for te municipalities tat do not participate in te flas estimation, indices are estimated by carrying forward te price of te previous mont. Te adoption of tis different estimation tecnique is due to te fact tat te evolution of prices in te oter municipalities of te same region is not considered a satisfactory proxy. 13

14 As soon as product aggregate indices of all municipalities are compiled, regional and, ten, national indices are calculated (by product aggregates, by upper aggregates and for all items). If all municipalities of a certain region are not included in te flas estimate, te product aggregate indices of tis region are calculated applying to te indices of te previous mont, te montly rate of cange of national product aggregate indices. Te latter are calculated using data of regions wic participate in te flas estimate, as follows: were m,a R I 20 m, a R I 20 R I R1 R R1 is elementary index of product aggregate at regional level of te reference mont (m) of year (a) and R is equal to te sare of ouseold consumption expenditure for te product 20 R R1 aggregate in te region R on te national ouseold consumption expenditure for te same product aggregate. Once product aggregate indices of all regions are compiled, national indices are calculated (by product aggregates, by upper aggregates and for all items). m, a 14