Consumer Price Index. Base 2016

Transcription

1 INSTITUTO NACIONAL DE ESTADÍSTICA Consumer Prce Index. Base 2016 Methodology Subdrectorate General for Industral on Prces and Household Budgets Madrd, Aprl

2 Index 1. Introducton 2. Indcator defnton 3. Indcator scope Temporal scope Base perod Reference perod of the prces Reference perod of the weghtngs Populaton scope Geographcal scope Consumpton feld Shoppng basket Functonal breakdown of the ndces Geographcal breakdown of the ndces 4. Sample desgn Selecton of muncpaltes Selecton of commercal areas and establshments Selecton of artcles Specfcatons of the artcles Number of observatons 5. General calculaton method Elementary ndces Weghts 2

3 5.3 - Aggregate ndces Calculaton of Rate Changes Monthly Rates Changes Accumulated Rate Varaton Annual Rates Varaton Calculaton of contrbutons Monthly contrbutons Accumulated contrbutons 6. Type of artcles Crtera for the classfcaton of artcles Seasonal tems Defnton Methodology Calculaton of ndces Provncally collected tems wth tarff prces Rentals for housng Centralsed collecton artcles 7. Collecton of the prces Frequency and prce collecton perods Prce reductons Crtera and types of prce reductons Feldwork organsaton 3

4 8. Data processng Data sendng Edton and valdaton of prces Qualty changes Defnton Qualty adjustment methods 9. Seres lnkng 10. Harmonsed Consumer Prce Index, base 2015 Annexes Annexe I. Calculaton of grouped ndces Annexe II. Hstorcal Evoluton 4

5 1. Introducton The operaton of the change n the Consumer Prce Index (CPI) System manly conssts of revsng and updatng each of ts components and determnng the best optons for achevng a representatve and precse ndcator that adapts to economc trends. Untl the entry nto force of base 2001, the CPI-based ts calculaton on what s called a fxed-base system, whose man characterstc s that both the composton of the shoppng basket and ts weghtngs are unaltered for the entre tme that the base s used. The changes n the base were carred out every eght or nne years because such was the perodcty of the Basc Household Budget Survey (BHBS), the source used for the preparaton of the weghtngs and of the shoppng basket. Due to ths, the only way to be able to collect the changes n consumer behavour and for the CPI to adapt to these trends was to wat untl the followng change n base. Evdently, n some cases, the perod was excessvely long. As of 1997, the two household budget surveys that had coexsted (one contnuous and quarterly, and another, carred out every eght or nne years) were substtuted by a sngle, quarterly one, whch provded nformaton that was closer to the basc survey, as per the breakdown level. Ths survey, called the Household Budget Contnuous Survey (HBCS), provded the nformaton necessary for the updatng of the weghtngs, as well as the renewal of the composton of the shoppng basket n the base change of CPI In addton, t made possble the permanent updatng of sad weghtngs and the revson of the shoppng basket, whch was an mprovement on the changes of the CPI System. Thus, wth the CPI, base 2001, a new Calculaton system began, whose most mportant characterstcs are ts dynamsm and ts contemporary nature. It s a more current CPI, gven that t permanently revses ts methodologcal system, for the purpose of mprovng t. To ths end, drect contact s establshed wth the dfferent academc forums and natonal and nternatonal producng bodes. It s also, however, a more dynamc CPI than ts predecessors n that t annual revses the weghtngs for certan functonal breakdown levels, and ncludes, n the shortest tme possble, any change detected n the market components, be t the appearance of new products, changes n the consumpton structure or the sample of muncpaltes or establshments. In addtons, t establshes the base changes every fve years, carryng out a complete revson of the methodology and the sample and the updatng of weghtngs at all breakdown levels. As a result of ths new scheme of functonng, n January 2017, the Consumer Prce Index System entered nto force, wth reference base n the year Ths System substtutes the CPI, whch wth base 2011, was n force untl December The CPI, base 2016, mantans the man characterstcs of the CPI, base 2011, and the same as ths, wll annual revew the weghtngs for a certan functonal breakdown level. To carry out ths updatng, t wll use the nformaton provded by the Household Budget Survey (HBS) base In 1999, the CPI Workng Group was created, at the heart of the Hgh Councl on Statstcs, comprsed of representatves from the Mnstry of Economy and Tax, Bank of Span, consumer assocatons, unverstes, trade unons, OECD, Chambers of Commerce, among others. The Group meets perodcally for the purpose of the INE 5

6 nformng on the most relevant aspects of the preparaton of the new prce systems and the annual updates. The CPI, base 2016, counts wth the assstance of the Workng Group, whch guarantees the ample support of the man economc agents. In addton, the methodology of the new System was analysed by the Hgh Councl on Statstcs and was the object of study by the Permanent Commsson of sad Councl untl t was approved n the Plenary of the Hgh Councl on Statstcs. On the other hand, snce 1997, the INE also prepares the Harmonsed Consumer Prce Index (HCPI) followng the consensual norms n the heart of the European Unon (EU), wth the purpose of obtanng nflaton ndces wth a hgh degree of comparablty among the member States. The methodology appled n the natonal CPI s, for the most part, the same as that of the HCPI. It only dffers n some respects where the methodology proposed by European regulaton, wth the objectve of comparablty between countres, was not adequate to meet the objectve of the natonal CPI. The man characterstcs of the consumer prce ndex, base 2016, as well as those aspects n whch the HCPI dffers from ths one, are collected n ths methodology. The INE would lke to express ts grattude for the nterest shown by all those bodes and professonals who have partcpated n the dfferent workng groups that were formed for the change n the CPI System, and lkewse, thank the cooperaton of all those respondents who provded the data necessary to comple the CPI wthn the establshed terms. 6

7 2. Indcator defnton The Consumer Prce Index, publshed monthly, has the objectve of measurng the evoluton of the prce levels of consumer goods and servces acqured by the households resdent n Span. System Base 2016 uses the defnton of consumpton expendture of the FBS: " consumpton expendture s the monetary flow that the household and each of ts members spend on the payment of certan goods and servces, amed at the household tself or to be transferred free of charge to other households or nsttutons". It has, therefore, elmnated from the CPI consumpton feld the consumpton categores defned n the HBS such as the value of the goods receved n knd, wthn the concept of self-consumpton, self-supply, salary n knd, free or subsdsed food and rent of the dwellng n whch the household resdes, when the household s the owner of the same or has t granted free of charge or at a low prce by other households or nsttutons. Some non-consumpton taxes have also been abolshed from the pont of vew of the FBS. However, ths conceptual framework does not always conform to the defnton of expendture on fnal consumpton of households establshed n the European Account system (EAS-95), whch governs the natonal accounts of all EU Member States. The objectve of the FBS s to measure the level and the evoluton of the expendture carred out by households n goods and servces ntended for consumpton, rrespectve of the monetary benefts that may be perceved as subsequent, the result of that expendture (va ndemntes or subsdes) and the sector wth whch the transacton was made (households or companes) In the CPI, base 2016, some of the estmates of expendture from the FBS have been adjusted to adapt the CPI to the conceptual requrements of the EAS. These adjustments affect the CPI calculaton as follows: - Gross spendng vs. net spendng. The FBS ncludes the total expendture made by households, rrespectve of the possble monetary counterparts that t may have perceved (except for Health and Educaton, where expendture refers to the actual effect, after subsdes). However, n accountng terms, the expendture made must deduct the amounts receved by the household, so that only the net expendture of subsdes and counter benefts s consdered. Ths change affects nsurance n the CPI (that nclude health nsurance, housng nsurance, and vehcle nsurance). The expendture used so far to estmate ts weghtng s the one that was really carred out by households, regardless of how they may have perceved any amount of compensaton. Accordng to the conceptual framework of the EAS, the amount of the benefts must be detracted from the gross expendture. - Transactons between agents belongng to the same economc sector. The FBS consders the expenses made n all transactons made by prvate consumers, regardless of the sector wth whom the operaton was carred out. The EAS affects n that the transactons consdered n the defnton must have occurred between the household sector and that of companes or other nsttutons. Transactons between households should not be taken nto account n estmatng expendture. 7

8 The CPI parcel that must be adjusted due to ths conceptual change s that of used cars snce the expendture from the FBS ncludes purchases made to companes and other households. - Expanson of the consumpton feld. Gamblng s a servce that s lkely to belong to the consumer feld of the CPI. However, for fundamental methodologcal reasons, t has not yet been ncorporated nto the CPI (for the same reason, t s also excluded from the HCPI consumpton feld). However, an mportant part of household spendng s a payment to servce provders, so ther ncorporaton nto the consumer feld of the CPI s justfed. The precson wth whch ths short-term ndcator measures the evoluton of prce level depends on two qualtes that every CPI must have: representatveness and tme comparablty. The degree of CPI representatveness s determned by the adaptaton of ths ndcator to the economc realty of that tme; thus, the varaton rate calculated from the CPI wll more closely approxmate the evoluton of the whole of the prces of the economy, the more the elements selected for measurement adapt to the behavoural trends of consumers. To acheve ths, the artcles selected to form part of the shoppng basket must be the most frequently consumed by most of the populaton, the establshments from the sample must be the most vsted, and the relatve mportance of each artcle n the shoppng basket must answer to household consumpton trends. The better the selecton of these elements, the more representatve the ndcator wll be. Furthermore, the CPI s an ndcator that only makes sense when comparsons n tme are establshed; n fact, an ndex number barely has any meanng f a comparson s not establshed wth ndces from other perods, to obtan the correspondng varaton rates (ths could be one month, one year, or any other perod). To ths end, the other qualty that may be attrbuted to a CPI s tme comparablty, that s, the need for the elements that defne the CPI to reman stable over tme, except, logcally, the prces collected monthly. Ths makes t possble for any varaton n the CPI to be due only to changes n the prces of the selected artcles, and not to any change n the methodologcal content of ths ndcator. The applcatons of the CPI are numerous and of great mportance n the economc, legal and socal areas. Worth notng among these s ts use as a measurement of nflaton. It s also appled n the revson of real estate rental contracts, as a reference n wage negotaton, n the establshment of pensons, n the updatng of nsurance premums and other types of contract, and as a deflator n the Natonal Accounts. 8

9 3. Indcator scope 3.1 Tme scope BASE PERIOD The base perod, or reference perod, of the ndex, s that for whch all ndces are made equal to 100. Ths s normally an annual perod. In the new system, the arthmetc average of the twelve publshed monthly ndces of the year 2016, n base 2011, s equal to 100. The perod to whch the ndex refers, therefore, s the year 2016, and t s usually called CPI base Ths means that all ndces publshed wll refer to ths year REFERENCE PERIOD OF THE PRICES Ths s the perod wth whose prces the current prces are compared, that s to say, the perod chosen for the calculaton of the basc ndces. Wth the calculaton formula used for CPI base lnked Laspeyres - the reference perod of the prces vares each year, and t s the month of December of the year mmedately before the year consdered REFERENCE PERIOD OF THE WEIGHTINGS The reference perod of the weghtngs s that to whch the weghtngs are servng as the structure of the System refer. For the year 2017, the calculaton of the weghtngs was carred out based on the data from the HBS, whch provdes the basc nformaton on the expendture of famles on consumpton goods and servces correspondng to the year For the adaptaton of the expendture to the concept of "net weghts" n nsurance, the elmnaton of the purchase-sales between ndvduals of new cars and the ncluson of gamblng expendture, the nformaton provded by the Natonal Accountng and the supply sources of these sectors has been used. In addton, to correct the lag produced between ths reference perod of the weghtngs and the reference perod of the prces (December of the year 2016), the weghtngs have been updated, usng nformaton on prce and quantty evoluton, from the CPI and other sources. Thus, the reference perod of the weghtngs s December 2016, durng the year And n subsequent years t s the month of December pror to each year, gven that the weghtngs wll be updated annually, usng the annual nformaton of the HBS, and nformaton from other sources, such as the evoluton of prvate consumpton from the Natonal Accounts, the evoluton of CPI prces and the nformaton provded by sources on the supply of the dfferent sectors. 9

10 Ths annual revson of weghtngs wll be performed for certan geographcal and functonal breakdown levels, usng the nformaton avalable closest to the moment of the revson. In addton, every fve years, a base change wll be carred out, n whch the weghtngs wll be updated for all functonal and geographcal breakdown levels. 3.2 Populaton scope The populaton of the ndex or reference stratum s the populaton group whose consumpton expendture structure serves as the bass for the selecton of representatve artcles and the calculaton of ther weghtngs. In CPI base 2016, the reference stratum ncludes all the populaton that resdes n famly dwellngs n Span; excludng, therefore, the expendture of the persons who resde n group dwellngs or nsttutons (convents, nursng homes, prsons, etc.) and the expendture of non-resdents. 3.3 Geographcal scope The geographcal scope of the research s comprsed of the entre natonal terrtory. 3.4 Consumpton feld It s the group of goods and servces that the households of the reference stratum use for consumpton; therefore, the expendture on nvestment goods, mputed rentals and expenses subsdsed by the publc admnstratons are not consdered. Nor are some taxes not consdered consumpton n the feld of consumpton from the pont of vew of the EAS-95. The FBS, as well as the CPI, base 2016, have been adapted to the new Internatonal Classfcaton of Consumpton n the European Unon (EU), called ECOICOP (European classfcaton of assumpton by purpose). Each consumpton dvson of the HBS s represented by one or more artcles n the CPI, n such a way that the evoluton of the prces of these artcles represents the evoluton of all of the elements comprsng sad dvson. The adaptaton to the ECOICOP from the COICOP used so far mples, n many cases, to dsaggregate subclasses (maxmum level of dsaggregaton at present) n two or more plots. In some cases, ths change assumes that the new subclasses wll contan fewer representatve artcles as a result of the dstrbuton of artcles that together prevously represented a subclass, and n other cases, t would requre the ncorporaton of new artcles to represent htherto non-exstent subclasses. 10

11 SHOPPING BASKET The group of goods and servces selected n the CPI, whose prce evoluton represents that of all those that comprse the ECOICOP parcel to whch they belong. The selecton of artcles comprsng the shoppng basket has been performed based on CPI base 2011, and data from HBS The crteron for determnng whch parcel should be ncluded s to take nto account all those that exceed a mnmum threshold of total spendng. Once the parcels of expendture that are represented n the ndex were determned, the artcles that made up the basket of the purchase of the base 2011 were revewed, ncreasng, decreasng or mantanng the artcles of each plot, dependng on the weghtng of the parcel, the number of artcles and prces representatve of t and the varablty of the prces of such tems. Thus, the total number of artcles comprsng the shoppng basket of CPI base 2016 s 479. For each of the artcles, a descrpton or specfcaton s prepared, for the purpose of facltatng ts dentfcaton on the part of the surveyor and permttng the correct collecton of the prces. These specfcatons take nto consderaton the partculartes of each regon. 3.5 Functonal breakdown of the ndces The CPI base 2016 adapts completely to the ECOICOP nternatonal consumpton classfcaton. The artcles n the shoppng basket are aggregated nto subclasses, these subclasses nto classes, subsequently nto subgroups, and fnally, the subgroups nto groups. The functonal structure of the CPI s extended to all levels (except the groups), wth respect to the CPI, base 2011, due to the mplantaton of the ECOICOP. Snce January 2017, ths structure conssts of 12 groups, 43 sub-groups, 101 classes and 219 subclasses. In addton, the 57 headngs and 29 specal groups exstng n CPI base 2016 are mantaned. 11

12 CPI 2011 artcles are dstrbuted n the 12 groups as follows: Number of artcles, CPI Base 2016 Groups Number of artcles 01 Food and non-alcoholc beverages Alcoholc beverages and tobacco Clothng and footwear Housng Furnshngs, household equpment and routne mantenance of the house Health Transport Communcaton 5 09 Lesure and culture Educaton 7 11 Restaurants, cafes and hotels Other goods and servces 37 TOTAL Geographcal breakdown of the ndces CPI base 2016 publshes the ndces for the dfferent geographcal and functonal breakdown levels that were publshed n CPI base The followng table collects the breakdown for whch data s publshed monthly. INDEX Natonal Autonomous Provnce Communty General X X X Groups X X X Subgroups X X X Classes X Subclasses X Headngs X X Specal Groups X X 12

13 4. Sample desgn As s most European Unon (EU) countres, the sample desgn of the prces that ntervene n the calculaton of the CPI s ntentonal, and therefore t s a nonprobablstc desgn, gven the characterstcs of the target populaton of the study. To obtan sgnfcant ndcators at all functonal and geographcal breakdown levels for whch the CPI s publshed, a sample selecton process has been structured n three large sectons, each of whch has the objectve of selectng the dfferent components of the same. These are as follows: - Selecton of muncpaltes. - Selecton of commercal areas and establshments. - Selecton of artcles. 4.1 Selecton of muncpaltes The selecton of the muncpaltes that are a part of the new CPI System has been carred out, takng nto consderaton not only demographc crtera but also geographcal representatveness. The offcal populaton data that has been used to make the selecton of muncpaltes s that obtaned from the revson of the Muncpal Regster of Inhabtants on 1 January The geographcal coverage crteron was manly based on the populaton of the group of selected muncpaltes. In ths way, the muncpaltes selected must cover 30% of the populaton of the provnce and 50% of the populaton of the Autonomous Communty. Wth ths crteron, the muncpaltes were selected by sze, untl the requrement was fulflled, wthout consderng the geographcal dstrbuton of the same wthn the provnce. In addton, ths departure crteron s completed as follows: - geographcal representatveness: t s mportant for the muncpaltes n the sample to be dstrbuted throughout the entre provnce, avodng a concentraton n certan populaton centres. - populaton representatveness: representatveness of small muncpaltes s emphassed; n prevous bases, when the selecton crteron was exclusvely populaton, part of the populaton resdng n smaller muncpaltes were excluded. - shoppng basket representatveness: all muncpaltes must have artcles of all groups; for ths reason, a reduced basket has been compled from the total basket, n whch basc consumpton goods have been ncluded. Wth ths, the representatveness of the CPI has ncreased consderably. Thus, the sample of muncpaltes obtaned wth the crtera above conssts of 177 (the 52 provncal captals and 125 non-captal muncpaltes). In 97 of these 177 muncpaltes, prces were collected from the entre shoppng basket of artcles, n 44 of them, prces were collected n the Food shoppng basket (groups 01 13

14 and 02), and part of the rest of the shoppng basket, and n the 36 remanng muncpaltes prces were collected from a reduced part of the shoppng basket (comprsed of 48% of the artcles). It s mportant to hghlght that, n practce, populaton percentages hgher than those ndcated are covered, granted that some establshments ncluded n the sample, such as hypermarkets, shoppng centres, repar workshops or furnture stores, are found outsde of the muncpaltes or n borderng muncpaltes, due to whch the real populaton represented n the ndex s larger than the theoretcal populaton. 4.2 Selecton of commercal areas and establshments For the selecton of the number of establshments, CPI base 2016 has used as ts startng pont the sample from CPI base 2011 and has studed the exstng network of establshments avalable n each provnce, payng specal attenton to the dfferent types and characterstcs of sad establshments. In addton, the representaton of commercal realty and the evoluton of the prces n all areas have been taken nto account. In ths selecton process of commercal areas and establshments, the partcpaton of the personnel of INE provncal delegatons s essental, because they are the man experts n the commercal relatons of ther provnce. As an overall crteron, the number of establshments that, on a monthly bass, nform on the prces of an artcle, was calculated dependng on the weghtng of the artcle n the ndex, and on the varablty of ts prces: the more weghtng and/or varablty of prces, the more number of establshments shall be selected. Lkewse, for the calculaton of the number of establshments, a mnmum number was establshed for each artcle n each provnce, dependng on the type of artcle and the type of collecton of sad artcle. For the selecton of the types of establshment, the dstrbuton of sales percentages by type of establshment (hypermarkets, supermarkets, markets and specalsed stores) were taken nto consderaton, dependng on each artcle. To ths end, we counted on nformaton from several sources, among these beng the Annual Trade Survey (INE) and the Mnstry of Agrculture, Food and the Envronment. Specal attenton s pad to shoppng centres, hypermarkets and supermarkets, gven ther mportance n terms of sales volume. In many cases, the stuaton n sad centres, as well as the presence of markets, condtons the creaton of "commercal areas", whch are explctly defned, n each muncpalty of the sample, for pershable food artcles (meat, fsh, fresh fruts and vegetables), and mplctly so for the rest of the artcles. In the defnton of the commercal areas, we begn wth the hypothess that the populaton that purchases n sad commercal area has homogeneous behavour and habts as regards consumpton. These commercal areas have been delmted wth the ad of the provncal delegatons. 14

15 For the pershable food artcles, three types of commercal areas were defned, based on the sze of the muncpalty and the number of establshments susceptble of beng selected for the dfferent types of artcles consdered. In addton, these artcles were classfed nto two large groups, consderng the varablty of the prces that they present and the weght that they have n the shoppng basket. Ths classfcaton determnes the number of establshments n whch prces are collected, accordng to the type of commercal area and the type of artcle n queston. For the rest of artcles, even though a strct delmtaton has not been carred out n commercal areas, the selecton of establshments s carred out complyng wth the representatveness objectve: the establshment sample must represent, wth the evoluton of the prces of the artcles sold n them, all the establshments of the town. Based on the aforementoned premses, the personnel of the provncal delegatons of the INE carry out the selecton of the respondent establshments, so as to comply wth the followng basc norms: - The sample must represent all of the commercal areas and the dfferent types of exstng establshments. - The establshments must have the most frequent and massve flow of publc n the town, and/or the greatest sales volume. - The establshments must be representatve of all types of artcles for whch nformaton s collected. - In each establshment, no more than one prce may be collected for the same artcle on the same day. - An establshment may not concentrate an mportant number of observatons of prces of dfferent artcles. The objectve of ths prce polcy s to try to avod one sngle establshment from condtonng the evoluton of the ndex. - Excluded from the sample are those establshments wth access restrcted to a sector of the populaton, such as cooperatves, company stores or smlar establshments. Itnerant street vendng and door-to-door sales are not consdered ether. - -The establshments selected must offer suffcent guarantees of contnuty n the sale of artcles for whch prces are collected, gven that ths sample remans fxed over tme, except n the case of closng, change of actvty, loss of representatveness as regards consumpton, or no longer sellng the artcle for whch prces were collected. In these cases, the establshment wll be substtuted by another that meets the necessary requrements to belong to the sample. All of these crtera are dependent on other aspects, such as collecton costs, the wllngness of the respondent to cooperate and relablty of the data suppled. However, 15

16 Law 4/1990 establshes the oblgaton of provdng the necessary data for the complaton of these Statstcs. In the CPI base 2016, prces are collected n approxmately 33,000 establshments, dstrbuted throughout the natonal terrtory. 4.3 Selecton of artcles To select the artcles that are representatve of the expendture dvsons of the HBS, dfferent bodes, busness assocatons, manufacturers, traders and establshments were consulted, whch provded nformaton on those artcles that best represented the dfferent dvsons, n accordance wth the followng selecton crtera: - The evoluton of the prces of the artcles selected must be smlar to that of the rest of the artcles of the dvson that they represent. - The artcles must be habtually consumed by the populaton. - They must have prces that are easly observable. - They must offer reasonable guarantees of contnuty on the market. Thus, n CPI base 2016, the shoppng basket s comprsed of 479 artcles SPECIFICATION OF ARTICLES Once the artcles are selected, t s necessary to determne the specfcatons that defne them. The preparaton of these specfcatons allows for comparson over tme of equal or equvalent-qualty artcles, for the purpose of measurng real varatons of prces and not those caused by a dfferent qualty of the artcles. For ths reason, the determnng factors of the prces must be specfed for each artcle, among whch are: the unt of measure, type of packagng, sze, composton, shape and dmensons. In order to determne the characterstcs of each tem and determne the specfcatons, we also obtaned nformaton from numerous bodes, professonal assocatons and companes; n ths way, t was possble to comple some general specfcatons that subsequently were adapted, n each provnce, to the specfc products selected n each establshment. Thus, when the nformaton collecton s carred out on the part of the agents, the artcles are dentfed n a detaled manner. The specfcatons of each artcle reman fxed over tme, so long as they reman representatve of the consumpton of the area, and are therefore modfed, when the artcle ceases to be sold or ceases to be representatve of the consumpton of a town. The tmes n whch t s necessary to change a product, the provncal delegatons must nclude a detaled specfcaton of the new product selected as the substtute of the pror; ths wll allow the calculaton of a lnkng coeffcent that assures varatons n the 16

17 ndex exclusvely caused by prce varatons and not but changes n the characterstcs of the artcle. 4.4 Number of observatons As prevously stated, the number of observatons used for the calculaton of the ndex depends on the type of artcle n queston, as well as on the establshments that have been selected n each of the provnces. Sectons 6 and 7 on "Types of artcles" and "Prce collecton" ndcate the frequency of prce collecton that ultmately determnes the sze of the sample of prces that are collected on a monthly bass. The total number of prces processed each month amounts to approxmately 220,

18 5. General calculaton method The formula used to calculate the ndces of CPI base 2016, s the lnked Laspeyres formula, whch was frst used n CPI base The overall ndex correspondng to month m of year t s mathematcally expressed as follows: Smlarly, t can be expressed as: 0 I t LE = k t p q k 1 k = 1 p q k 1 k 1 where: 0 I t LE I = t k = 1 p k = k-1 k 1 p k k p k 1 k 1 p q p k 1 k 1 p q and k 1 = p t k = 1 k 1 k 1 k 1 k 1 = W k-1 k-1 p q q I k W k 1 As may be observed, a lnked ndex establshes comparsons between the current perod (t) and the base perod (0) but consderng the ntermedate stuatons (k). In CPI base 2016, the ntermedate stuatons consdered correspondng to the months of December of every year. Thus, the ndex n base 2016 for month m of year t, s obtaned as the product of the ndces as follows: 16II mmmm GG = II dddddd(tt 1) 16 GG dddddd17 = II dddddd16 dddddd16ii GG 16 GG 100 where: dddddd(tt 1)II GG mmmm 100 dddddd(tt 2)II GG dddddd(tt 1) 100 = dddddd(tt 1)II GG mmmm 100 II GG mmmm 16 s the overall ndex, n base 2016, of month m of year t. II GG mmmm dddddd(tt 1) s the overall ndex, referrng to December of year (t-1), of month m of year t. 18

19 The man nconvenence of lnked ndces s the lack of addtvty. Ths makes t mpossble to obtan the ndex of any aggregate as the weghted average of the ndces of the aggregates that comprse t. Thus, for example, the overall ndex cannot be calculated as the weghted average of all of the ndces of the twelve groups. 5.1 Basc ndces A basc aggregate s the consumpton component wth the lowest aggregaton level for whch ndces are obtaned, and n whose calculaton weghtngs do not ntervene; the ndces of these aggregates are called basc ndces. The Spansh CPI calculates a basc ndex for each artcle of the shoppng basket n each of the provnces, and therefore, the basc aggregate s the artcle-provnce. Basc aggregate ndex s obtaned as the quotent of the average prce of sad basc aggregate n the current perod and the average prce n the reference perod of the prces, that s, December of the prevous year: where: mt dc( t 1) I dc( t 1) I mt P P mt = 100 dc( t 1) s the ndex, referrng to December of year (t-1), of basc aggregate, n month m of year t. mt P s the average prce of basc aggregate, n month m of year t. dc( t 1) P s the average prce of basc aggregate, n December of year (t-1). At the same tme, the average prce of aggregate, n perod (m,t), geometrcal average of the prces collected n sad perod: mt P, s the smple n mt mt mt n P = P, j j= 1 (1) where: mt P, j s the prce of basc aggregate collected n establshment j, n perod (m,t). n mt mt s the number of processed prces of basc aggregate, n perod (m,t). The geometrc average grants the same mportance to the varatons of all prces, regardless of ther level. 19

20 5.2 Weghtngs The weghtngs that ntervene n the calculaton of the aggregate ndces come from the HBS. Ths survey provdes estmates of the expendture on consumpton products made by the households resdent n famly dwellngs n Span. The consumpton classfcaton that the HBS (ECOICOP) uses conssts of a seres of expendture dvsons, most of whch nclude goods and servces ncluded n the consumpton feld of the CPI. For the calculaton of the weghtngs of the artcles comprsng the shoppng basket of CPI base 2016, the breakdown of these dvsons has been necessary to obtan more detaled nformaton. To ths end, we have counted on the collaboraton of dfferent bodes, assocatons, manufacturers and traders. The data used n the calculaton of the weghtngs, used durng the year 2017, are those correspondng to the year In addton, to correct the lag produced between the reference perod of the weghtngs and the reference perod of the prces (December of the year 2016), the weghtngs have been updated, usng nformaton on prce and quantty evoluton, from the CPI and other sources. In ths way, the reference perod of the weghtngs, used durng the year 2017, s December The weghtngs for each artcle represent the relaton between the expendture on the dvsons represented by sad artcle and the total expendture on all dvsons covered by the ndex: W = gasto realzado en las parcelas representadas por el artículo gasto total These weghtngs are dfferent n each of the geographcal aggregatons (provnces, Autonomous Communtes, and the total natonal), and from them, the weghtngs of the dfferent functonal aggregatons are obtaned. Thus, the weghtng of functonal aggregate A s obtaned as the sum of the weghtngs of the artcles that comprse sad aggregaton: W = A A The annual weghtngs updates, whch wll be carred out n CPI base 2016, wll be done wth the latest avalable nformaton from the HBS. W 20

21 5.3 Aggregate ndces As prevously mentoned, the basc ndces refer to December of the mmedately prevous year. In turn, the weghtngs used for the calculaton of the aggregatons also refer to December of the prevous year, thereby mantanng coherence wth the reference prces. The calculaton of the aggregate ndces s descrbed below. Functonal aggregatons wthn a provnce The ndex, referrng to December of the prevous year, of any functonal aggregaton A n a provnce p, s obtaned as the aggregaton of the basc ndces of the artcles belongng to sad aggregaton wth the weghtngs applcable n year t. Its mathematcal expresson s as follows: where: I mt dc( t 1) p, W dc( t 1) p, I mt mt = I W dc( t 1) Ap, dc( t 1) p, dc( t 1) A s the ndex, referrng to December t-1, of artcle n provnce p, n month m of year t, s the weghtng (so much per one), referred to December of year (t- 1), of artcle n provnce p, wthn Aggregaton A, that s: p, W dc( t 1) p, gasto realzado en el artículo dentro de la provnca p = gasto realzado en la agregacón funconal A dentro de la provnca p Once the aggregate ndces are calculated as detaled above, t s necessary to lnk them. These ndces are those whch are fnally dssemnated and provde contnuty for the seres publshed n base For any functonal aggregaton A, the ndex n base 2016 n provnce p s calculated as follows: 16II mmmm AA,pp = II dddddd(tt 1) 16 AA,pp dddddd(tt 1) 100 mmmm II AA,pp 21

22 Geographcal aggregatons of a functonal aggregaton In the same way as the prevous case, the calculaton of the ndex of a geographcal aggregaton R greater than the provnce, for a specfc functonal groupng A s calculated as follows: where: I mt dc( t 1) Ap, W mt mt I = I dc( t 1) ( 1) dc( t 1) AR, dc t Ap, Ap, p R s the ndex, referred to December of year (t-1), of the functonal aggregaton A n provnce p, n month m of year t. W dc( t 1) Ap, dc( t 1) s the weghtng (so much per one), referrng to December of (t-1), of functonal groupng A n provnce p, that s, W Ap gasto realzado en la agregacón funconal A dentro de la provnca p = gasto realzado en la agregacón funconal A dentro de la agregacón geográfca R As wth the functonal aggregatons, once the aggregate ndces are calculated, t s necessary to lnk them. For any functonal aggregaton A, the lnked ndex, n base 2016, n regon R, n month m of year t, s: 16II mmmm AA,RR = II dddddd(tt 1) 16 AA,RR dddddd(tt 1) 100 mmmm II AA,RR 5.4 Calculaton of varaton rates MONTHLY VARIATION RATE The monthly varaton rate of an ndex n perod (m, t) s calculated as the quotent between the ndex from the current month m and the ndex from the prevous month (m-1), accordng to the followng formula: VV mmmm/(mm 1)tt = 16II 16II mmmm (mm 1)tt = dddddd(tt 1)II mmmm dddddd(tt 1) II(mm 1)tt 22

23 where: VV mmmm/(mm 1)tt s the monthly varaton rate, n month m of year t. II mmmm 16 s the ndex, n base 2016, n month m of year t. II mmmm dddddd(tt 1) s the ndex, referred to December of the prevous year, n month m of year t. In other words, the monthly changes can be calculated wth the publshed ndces, n base 2016, or wth the unlnked ndces (referred to December of the prevous year) ACCUMULATED VARIATION RATES The accumulated varaton rate (year-to-date) s calculated as the quotent between the ndex from the current month and the ndex from December of the prevous year: VV mmmm/dddddd(tt 1) = 16II 16II mmmm dddddd(tt 1) = dddddd(tt 1)II mmmm = dddddd(tt 1)II mmmm dddddd(tt 1) IIdddddd(tt 1) where: VV mmmm/dddddd(tt 1) s the accumulated varaton rate, n month m of year t. 16 II mmmm s the ndex, n base 2016, n month m of year t. dddddd(tt 1) II mmmm s the ndex, referred to December of the prevous year, n month m of year t. That s to say, the accumulated varaton rates can be calculated wth the publshed ndces, n base 2016, or wth the unlnked ndces (referrng to December of the prevous year). 23

24 5.4.3 ANNUAL VARIATION RATE The annual varaton rate s calculated as the quotent between the ndces publshed n the current month and from the same month of the prevous year, both n base 2011: where: mmmm VVmm(tt 1) = 16II 16II mmmm mm(tt 1) VV mmmm/mm(tt 1) s the annual varaton rate, n month m of year t 16 II mmmm s the ndex, n base 2016, n month m of year t. In the case of annual varatons, these cannot be calculated wth the ndces referrng to December of the prevous year, as occurs wth the monthly and accumulated varatons. 5.5 Calculaton of effects MONTHLY EFFECTS The effect of a monthly varaton rate of an artcle or aggregate n the overall ndex s defned as the part of the monthly varaton of the overall ndex that corresponds to sad artcle or aggregate. Therefore, the sum of the monthly effects of all of the artcles n the shoppng basket s equal to the monthly varaton rate of the overall ndex. In other words, the effect that the monthly prce varaton of an artcle or aggregate has n the monthly varaton of the overall ndex s the varaton that ths ndex would have experenced f all of the prces of the rest of the artcles had remaned stable that month. The formula of the monthly effect of a specfc artcle (or aggregate), n month m of year t, s as follows: where: dc( t 1) I mt R ( m 1) t dc( t 1) I G I I W I mt ( m 1) t mt /( m 1) t dc( t 1) dc( t 1) ( m 1) t dc( t 1) dc( t 1) G = 100 s the ndex, referrng to December of year (t-1), of artcle, n month m of year t s the overall ndex, referrng to December of year (t-1), n month (m-1) of year t. 24

25 W dc( t 1) s the weghtng, referred to December of year (t-1), of artcle, n so much per one. Developng the prevous formula, we obtan an alternatve way of calculatng the effects through the varaton rates: R I I W mt ( m 1) t mt /( m 1) t dc( t 1) dc( t 1) = 100 = ( m 1) t dc( t 1) dc( t 1) I G mt ( m 1) t ( m 1) t dc( t 1) I dc( t 1) I dc( t 1) I ( m 1) t ( m 1) t W dc( t 1) dc( t 1) I G dc( t 1) I mt ( m 1) t ( m 1) t dc( t 1) I dc( t 1) I dc( t 1) 100 ( m 1) t W I dc( t 1) ( m 1) t dct ( 1) I dct ( 1) I G ( m 1) t mt/ ( m 1) t dc( t 1) W I dc( t 1) ( m 1) t dc( t 1) G = 100 = = V = I Therefore, the monthly effect of a specfc artcle, s the product of ts monthly varaton rate ( V mt, / ( m 1) t ) by ts weghtng ( dc( t 1) W of the artcle and the overall ndex of the prevous month ( ) and by the quotent between the ndex I ( m 1) t ( m 1) t dc( t 1) dc( t 1) As prevously mentoned, the sum of the monthly effects of all of the artcles that comprse the shoppng basket of the CPI s equal to the monthly varaton of the overall ndex. Ths s demonstrated as follows: / I G ). 25

26 I I W I mt ( m 1) t mt /( m 1) t dc( t 1) dc( t 1) R ( m 1) t dc( t 1) dc( t 1) G = 100 = I W I W = 100 = = mt ( m 1) t dc( t 1) dc( t 1) dc( t 1) dc( t 1) ( m 1) t dc( t 1) G I I mt ( m 1) t dc( t 1) G dc( t 1) G mt /( m 1) t 100 = ( m 1) t V G dc( t 1) G I I ACCUMULATED EFFECTS The effect of the varaton for the year-to-date (or accumulated varaton) of an artcle or aggregate n the overall ndex, represents the accumulated varaton that the overall ndex would experence f the rest of the artcles were not to experence any prce varaton for the year-to-date. In other words, t s the part of the accumulated varaton due to sad artcle or aggregate. The formula of the accumulated effect of a specfc artcle (or aggregate), n month m of year t, s as follows: R = I I W 100 = I I W W mt / dc( t 1) mt dc( t 1) dc( t 1) dc( t 1) dc( t 1) dc( t 1) dc( t 1) G mt ( dc( t 1) ) mt dc( t 1) 100 = 100 = 100 dc( t 1) I dc( t 1) 100 where: dc( t 1) I W mt s the ndex, referred to December of year (t-1), of artcle, n month m of year t dc( t 1) s the weghtng, referrng to December of year (t-1), of artcle, so much per one. Usng the alternatve formula for calculatng effects that were developed n the prevous secton, we obtan that the accumulated effect s equal to the product of the accumulated varaton by the weghtng: 26

27 R I I W mt dc( t 1) mt / dc( t 1) dc( t 1) dc( t 1) dc( t 1) dc( t 1) dc( t 1) I G dc( t 1) mt / dc( t 1) dc( t 1) V W I dc( t 1) dc( t 1) dc( t 1) G = 100 = = = 100 = V W = V 100 mt / dc( t 1) mt / dc( t 1) dc( t 1) dc( t 1) In the case of the accumulated effects, we also verfy that ther sum s equal to the accumulated varaton rate, developng the same steps as n the prevous secton. I W 27

28 6. Types of artcle 6.1 Crtera for the classfcaton of artcles The operatons ncluded n the calculaton process of the CPI, from the collecton of prces to the calculaton of the ndces, are dfferent, dependng on the partculartes of each artcle of the shoppng basket. Thus, the perodcty for prce collecton depends on the frequency wth whch they vary. The means of collectng them also dffer, dependng on the geographcal homogenety of the prces and on the avalablty of the same. Lastly, accordng to the characterstcs of each artcle, the calculaton method of the ndces s dfferent. The followng table shows the dfferent types of artcle, accordng to the crtera used n ther classfcaton: Crtera Types of artcle seasonal pershable groups 01 y 02 monthly not seasonal Perodcty for prce collecton monthly Place of collecton and recordng of prces groups 03 to 12 provnces central servces non-pershable monthly quarterly quarterly quarterly for sales seasonal whch are centrally collected Calculaton method of basc ndces of provncal collecton wth prepared prce rental of housng Perodcty of prce collecton wthout specal treatment Accordng to the perodcty and frequency of the prce collecton, a frst way of classfyng the artcles s establshed. Thus, two types of artcle are consdered: monthly and quarterly artcles. 28

29 Monthly collected artcles The prces of the monthly artcles are observed every month n all of the establshments of the sample, va personal vst, n most of the cases. In general, each establshment s vsted once a month, except those n whch the prces of pershable artcles are collected, whch, due to the hgher frequency wth whch ther prces vary, the ntervewer vsts twce or three tmes a month, dependng on the muncpalty. Unprocessed fresh food products are ncluded n ths category and suffer perodcal fluctuatons n ther prces, as well as constant changes n qualty. Included n these products are seasonal artcles (some fresh fruts and vegetables), that are only sold specfc months of the year; the prces for these artcles are only collected n the months n whch they are avalable. Wthn the monthly collected artcles, worth notng are the seasonal artcles, whch are those whose consumpton takes place only some months of the year, snce the rest of the year, they are not for sale. Moreover, these artcles are charactersed because, unlke fresh fruts and vegetables, all of those belongng to the same season have the same sale scheme n the market. In the CPI, clothng and footwear artcles are consdered to be seasonal, when they have two defned seasons (sprng-summer and autumn-wnter). Prce collecton for these artcles s carred out once a month durng the season n whch they are sold; and the processng of the prces durng the months n whch the artcle ceases to be sold s the repetton of the last collected prce, n such a way that there s no prce varaton when the artcle s not avalable. Quarterly collected artcles Quarterly artcles are those whose prces have a rather stable behavour, that s to say, they usually do not experence many prce varatons over tme (household applances, furnture, repar servces, restoraton, etc.). Quarterly collecton allows for ncreasng the number of prces collected wth the same cost. The processng of these prces conssts of dvdng the sample of selected establshments nto three sub-samples, n such a way that, each month, the establshments of only one of them are vsted, and the last prce collected n the establshments of the other two sub-samples s repeated. Ths acheves that, every month, there are establshments that nform on the prces of these artcles. In addton, n the case that more than half of the prces collected n a month vary, the followng month, nformaton wll be requested n the rest of the establshments. Wth the ncluson of dscounted prces, a new category of artcles from wthn the quarterly artcles was establshed: quarterly dscounts. These are the artcles that, despte fulfllng the requrement of prce stablty common to quarterly artcles, show sgnfcant varatons n typcal sales perods (household applances, furnture, bed lnens, etc.). For ths reason, the prce collecton durng the sales months s carred out vstng all of the establshments n the sample, and not only those of the sub-sample correspondng to that month. 29

30 Place of prce collecton and recordng Bearng n mnd the place where the prces are collected and recorded, we can dstngush between provncally and centrally collected artcles. The prces of the former are collected n each provnce, va personal vst, by means of telephone, fax or the nternet, through the offcal bulletns of the Autonomous Communtes or provnces, and are recorded n the provncal delegatons. On the other hand, the montorng of the prces of the centrally-collected artcles s carred out by the Central Servces of the INE. Included as part of ths type of artcle are those goods and servces that have one or several of the followng characterstcs: - ther prces are the same n a broad geographcal area, - ther prces are subject to rates publshed n the BOE, - there are few companes that sell the artcle, - a perfectly defned drectory of respondents s avalable, - they are artcles wth regular changes n qualty (such as technology artcles), whch nvolves dffculty n carryng out qualty adjustments; as they are collected n a centralsed way, the processng of these adjustments are homogensed. Calculaton method of basc ndces Regardng the calculaton method, there are certan groups of artcles whose basc ndces are obtaned dfferently than the general formula descrbed n secton 5.1. Accordng to the general calculaton formula, the basc ndex of any artcle of the shoppng basket s obtaned as the smple average of the prces collected, wthout consderng any type of weghtng. The excepton to ths general rule s consttuted by the calculaton formula of the centralsed collected artcle and the artcles wth a prepared prce; n both cases, the basc ndex s calculated takng nto account a set of representatve varetes or modaltes of each artcle, weghted approprately by the expendture made on each one of them. The weghtng of each modalty s obtaned from the expendture made by consumers. In ths way, we mantan coherence wth the general weghtng structure of the shoppng basket. Another group of artcles whch, due to ther characterstcs, receve specal treatment, are the seasonal artcles. Because of the perodcal oscllatons n prces and quanttes, the ndces of fresh fruts and vegetables are calculated n a dfferent manner, that consders the producton and sales calendars of each one. Fnally, and due to the specal characterstcs of the market, housng rental also receves a dfferent treatment from the general CPI formula. 30

31 The followng sectons descrbe these specal calculaton methods n detal: 6.2 Seasonal artcles DEFINITION Seasonal artcles are those that suffer perodcal fluctuatons, n both prces and quanttes consumed over the year, even dsappearng from the market durng some perod. Seasonal artcles are ncluded n the ndex of most countres. However, there are notable dfferences n the processng methods, as well as n the catalogung crtera of certan artcles as seasonal. The products that present, n all countres, a hgher seasonal profle, are fresh fruts and vegetables, although some also nclude fresh fsh, certan meats, soft drnks, ce cream, clothng and footwear artcles, plants, flowers and sportng goods. In the CPI base 2016, only those fresh fruts and vegetables that are not avalable for sale at some pont are consdered season artcles METHODOLOGY In base 2011, the calculaton of the ndces for the sub-classes of seasonal artcles was carred out n two stages. In each one of these stages, an ndex was obtaned: basc ndces for each seasonal artcle; drect and moble ndces for each seasonal sub-class. Wth the entry nto force of the new CPI base 2016, the calculaton formula s changed to assmlate t to that of the rest of the artcles of the basket, and the prce of off-season products wll be estmated wth the average varaton n the prces of the rest of the subclass products that are avalable, proceedng n the same way as wth the other tems n the shoppng cart when there s a lack of prce. Thus, these tems wll have prces every month of the year, as well as annual fxed weghts, and ther elementary ndexes can be calculated as well as those of the rest of the tems n the shoppng basket. 31