Services producer price indices for Marke research and public opinion polling Indusry descripion for SNI group 74.13 SPPI repor no 24 Ulf Johansson Services producer price indices, Price Saisics Uni, Saisics Sweden December 2007
Foreword Services price indices for Marke research and public opinion polling Privae services producion has made up a large share of Sweden's gross domesic produc (GDP) for a long ime, bu relevan price indices were no developed o any grea exen before he year 2000. This can in par be explained hisorically; commodiies have previously made up a dominan par of he economy. Economic saisics have also been focused on commodiies. Anoher reason is he fairly complicaed naure of measuring price rends in services. As par of he process o improve economic saisics, producer price indices for services, Services producer price indices (SPPIs), were developed a few years ago. SPPIs are developed for he mos common services in indusries where eiher he Naional Accouns or he EU require hem o be. This repor is one of a series of SPPI repors ha discuss he rend in services producer price indices in differen indusries. The aim of hese repors is boh o give a descripion of he saisics for exernal users and o provide a basis for producing new services price indices. I also gives a general descripion of how services price producer price indices can be calculaed. Acknowledgemen Thanks o our daa providers - privae individuals, companies, auhoriies and organisaions - Saisics Sweden is able o provide reliable and up-o-dae saisics ha saisfy sociey's need for informaion. 2
Conens Foreword...2 1 Summary...4 2 Background...5 2.1 Condiions for saisical work in he SPPI field...6 2.2 The Naional Accouns' requiremen for price indices...6 3 Indusry descripion...7 3.1 Srucure of he sub-indusry and saisical classificaion...7 3.2 Key raios and facs abou he indusry...9 3.3 The marke, services and price formaion...10 4 Tes survey srucure...12 5.1 Sampling procedure...12 4.2 Daa collecion procedure...12 4.3 Evaluaion of he es survey...12 4.4 Tes resuls and conclusions...12 5 Design of fuure surveys...13 5.1 Sampling procedure...13 5.2 Daa collecion procedure...13 5.3 Price measuremen mehods...13 5.4 Assumpions, weighs and index calculaions...14 5.5 Non-response managemen...15 5.6 Qualiy adjusmen...15 3
1 Summary This repor discusses he developmen of price indices for services in SNI 74.13 Marke research and public opinion polling and also includes a descripion of he indusry. In he marke research and public opinion polling indusry, he price is normally deermined by he coss for he saff needed o provide he service. Oher coss, such as elecommunicaion coss, regiser purchases and any necessary invesmen coss mus hen be added. The services are ofen bough as a package, in which all relaed services are included. Because he services are ofen unique, "Model pricing" is a suiable mehod for measuring he price rend. A number of ypeservices have been creaed in consulaion wih he indusry. These are repriced every quarer. If he service is no provided during a quarer, he company is asked o esimae he price. When seing he price, he coss for manpower, common coss and he gross profi margin should all be aken ino consideraion. "Model pricing" can be difficul for some companies and he alernaive is for hem o submi hourly prices for a number of saff caegories. The advanage of collecing hourly prices is ha hey can be found in mos business adminisraion sysems and are herefore easy for he companies o submi. The disadvanage of hourly prices is ha hey do no incorporae produciviy improvemens. 4
2 Background The Swedish service indusry's value added amouns o abou 62 percen of he business secor's oal value added. Calculaed as a share of GDP, he business secor's service producion consiues abou 50 percen. For mos of he 1990s, his share was jus under 40 percen of GDP and rose gradually owards he end of he cenury. As service producion's share of GDP has grown, he need for correc volume calculaions of service producion has also increased. To calculae rends in individual indusries, value added mus be purged from inflaion. Price rends for differen producs and services mus be measured and indices calculaed before he curren prices can be convered ino fixed prices. In order o cover mos of he groups of services in Sweden's naional accouns and o fulfil inernaional requiremens, new indices for he service indusries are coninuously developed. The services producer price index (SPPI) is a producer price index for services ha aims o measure price rends in indusries ha produce business services. There are currenly SPPIs for jus over 40 percen of he business secor's services producion. The majoriy of he SPPIs produced oday have been developed since 1999. SPPIs for rens (for business premises), hoel services and domesic air ravel began o be developed during he 1990s. A join Eurosa-OECD working group was esablished in 2002 in order o develop a common mehod for European services price indices and o provide pracical help o counries in he process of developing hem. A manual wih, among oher hings, pracical descripions of SPPI for differen indusries was developed. 1 The manual refers o he exising SPPI manual for index heory and 1 Mehodological guide for developing producer price indices for services. Websie:hp://www.oecd.org 5
pracice. 2 SPPI developmen and producion work is currenly regulaed by Council Regulaion (EC) 3. 2.1 Condiions for saisical work in he SPPI field SPPI is a producer price index ha aims o show he average price rend in he producer chain for services supplied by domesic producers. The index figure is for one quarer and he price daa is o be an average per quarer. The measuremen should refer o ransacion prices, i.e. he prices he purchaser acually pays afer discouns. When developing a services price index for a new indusry, specific indusry-relaed price measuremen problems mus be solved. For example, i is no always obvious how o define he service. Anoher problem is ha he services supplied are ofen ailor-made for a specific cusomer and are only supplied once. This makes i difficul o find comparable services o follow over ime. These problems are some of he hings discussed in his repor for he indusry in quesion: For which services are he prices o be measured? Wha does he indusry's price formaion look like? And in wha way can we include he prices in a way ha reflecs he price rend wihou placing oo much of a burden on daa providers? 2.2 Requiremen for price indices in he naional accouns As a resul of he Sabiliy Pac in Maasrich, he EU esablished crieria for he reporing principles for Member Sae naional accouns in fixed prices. For fixed price calculaions, i is recommended ha producion and inpu consumpion be deflaed separaely using approved, qualiy-adjused price indices. This 2 Producer Price Index Manual: Theory and pracice. Websie:hp://www.oecd.org 3 1165/98 concerning shor-erm saisics. 6
means ha he indices are only o reflec changes ha can be assigned o price changes whils changes regarding quaniy, qualiy or composiion of he produc/service should no influence he indices. As long as hese recommendaions are followed, he index will be considered an A-mehod for fixed price calculaion. All EU Member Saes have ranked heir mehods, under he supervision of Eurosa, as A, B or C mehods. As from 2006, no C- mehods may be used in Member Sae naional accouns unless dispensaion has been graned. For a more deailed descripion, please refer o he European Sysem of Naional Accouns, ENS 1995, Chaper 10. 3 Descripion of he indusry 3.1 Srucure of he sub-indusry and saisical classificaion According o he Swedish Sandard for Indusry Classificaion 2002 (SNI 2002), Marke research and public opinion polling is included in he group Legal, accouning, book-keeping and audiing aciviies; ax consulancy; marke research and public opinion polling; business and managemen consulancy; holdings (SNI 74.1) which comes under Oher business aciviies (SNI 74). Table 1: SNI 74 Oher business services SNI Descripion Legal, accouning, book-keeping and audiing aciviies; ax consulancy; marke research and public opinion polling; 74.1 business and managemen consulancy; holdings Archiecural and engineering aciviies and relaed echnical 74.2 consulancy 74.3 Technical esing and analysis 74.4 Adverising 74.5 Labour recruimen and provision of personnel 74.6 Invesigaion and securiy aciviies 74.7 Indusrial cleaning 74.8 Miscellaneous business services n.e.c. Source: SNI 2002: Swedish Sandard for Indusry Classificaion 2002. 7
Table 2: Sub-groups in SNI 74.1 Legal, accouning, book-keeping and audiing aciviies; ax consulancy; marke research and public opinion polling; business and managemen consulancy; holdings SNI Descripion 74.11 Legal aciviies 74.12 Accouning, book-keeping and audiing aciviies, ax consulancy 74.13 Marke research and public opinion polling 74.14 Business and managemen consulancy aciviies 74.15 Managemen aciviies of holding companies Source: SNI 2002: Swedish Sandard for Indusry Classificaion 2002. Sandard for Swedish Produc Classificaion (SPIN 2002) is a sandard linking producs o aciviies (indusries) in SNI 2002. SPIN 2002 is he Swedish equivalen of he EU's Classificaion of Producs by Aciviy (CPA). Table 3: Deailed groups in 74.130 Marke research and public opinion polling Indusry SPIN 2002 CPA 2002 Marke research 74.130.01 74.13.11 Public opinion polling 74.130.02 74.13.12 Source: SPIN 2002: srucure, SCB/ES. 8
3.2 Key raios and facs abou he indusry This secion is based on daa from he Saisics Sweden Business Regiser. The daa have been processed o he exen ha companies have been included if heir aciviies in he area amoun o a leas 55 percen of heir oal urnover. Table 4: srucure by size caegory broken down by number of employees in SNI 74.13 Marke research and public opinion polling. Number of employees Number of companies % Number of employees % Turnover, (SEK housands) % 0 302 64% 0% 105 573 5% 1-4 102 21% 185 7% 162 347 8% 5-9 22 5% 149 6% 147 993 7% 10-19 22 5% 304 12% 227 593 11% 20-49 12 3% 338 13% 275 369 13% 50-99 9 2% 578 23% 367 762 18% 100-6 1% 1 007 39% 795 057 38% Toal 475 100% 2 561 100% 2 081 694 100% Source: Saisics Sweden Business Regiser According o he Business Regiser, here are nearly 500 companies employing abou 2 500 people in indusry 74.13. Toal urnover amouned o jus under SEK 2.1 billion. The larges share (64%) of he companies are sole proprieorships (companies wih 0 employees) bu hey were only responsible for abou SEK 0.1 billion ou of a oal urnover of SEK 2.1 billion. There are 15 companies in he wo larges size caegories wih more han 50 employees and hey are responsible for 56 percen of he indusry's oal urnover. Sixy-wo percen of he oal number of employees in he indusry can be found in hese companies. 9
3.3 The marke, services and price formaion Informaion in his secion is based on visis o a number of companies and o he marke research indusry's rade organisaion. The rade organisaion is called he Associaion of Swedish Marke Research Companies (Sveriges Marknadsundersökningsföreag - SMIF). Mos of he large companies are members of SMIF. The services offered by he companies in he indusry are mosly research and analysis of areas such as rademarks, media, coworkers, public opinion and consumpion. The research can be divided up uno qualiaive and quaniaive research surveys. Qualiaive surveys can be e.g. group discussions and in-deph inerviews. Quaniaive surveys are performed on he elephone, inerne or by pos. The services in he indusry are ofen bough as a package by he cusomer. This package includes he differen service componens, e.g. daa collecion and associaed repors wih analysis and resuls. The oal price of a service is normally calculaed afer he ime i will ake for he differen personnel caegories o provide he service has been esimaed. Oher coss, such as elecommunicaion coss, regiser purchases and any necessary invesmen coss mus hen be added. Price are generally adjused once a year and he price level of he services can vary depending on he cusomer. The indusry is inernaional and almos all he larger companies belong o a mulinaional group. Only aciviies in Sweden are of ineres o Saisics Sweden when measuring prices. Previously, mos companies performed he enire marke research producion chain hemselves, from planning and fieldwork o repors and presenaion of he resuls. This has changed o a cerain exen in recen years in ha some companies have begun specialising in daa collecion and oher companies work more wih analysis and compiling repors. Produciviy has developed srongly in he indusry in recen years hanks o new IT and elephony echnology. I is difficul o predic how produciviy will develop in he fuure. New ools and programs are currenly being developed which creae pars of he repors auomaically. This will speed up he process somewha. How price rends could be measured was discussed on he company visis. Since he services are ofen priced as a package, a number of companies fel ha price movemens in he indusry 10
could be racked by following a number of represenaive services over ime. This is normally referred o as "Model pricing".4. In consulaion wih he companies, Saisics Sweden has esablished a number of ype-services which are repriced every quarer. If daa providers have no performed he service in one quarer, hey are asked o make an esimae. The coss of manpower, common coss and profi margins shall all be aken ino consideraion. The reason why hus price collecion mehod is suiable for his paricular indusry is ha exacly he same services are seldom repeaed. Indeed, he indusry is dominaed by unique services. Some companies find "Model pricing" difficul since i is in many cases a maer of negoiaing he price wih he cusomer when seing he price of a service. The mehod also requires a considerable amoun of work for daa providers. One opion hey have is o submi hourly prices for a number of personnel caegories. The advanage of hourly prices is ha he daa can be found in mos business adminisraion sysems and are herefore easy for he companies o submi. The disadvanage of using hourly prices is ha his price measuremen mehod has a endency no o include improvemens in produciviy and insead shows hem as price increases. For example, if more efficien working mehods lead o more work being performed per hour and a subsequen increase in price, i can be difficul o deermine wha is a price rise and wha is an increase in qualiy. 4 See he "Mehodological guide for developing producer price indices for services" from he "Join OECD/Eurosa ask force" for a review of differen price measuremen mehods. 11
4 Tes survey srucure 4.1 Sampling procedure The sample consised of wo large and hree medium-sized companies in he indusry, seleced from he Business Regiser. Visis were carried ou a all he companies in order o find ou he services hey offer and he various possible ways of measuring heir prices. All he companies visied were cooperaive and pu forward many good poins and suggesions as o how he form could be consruced. 4.2 Collecion procedure A conac person was seleced for each company during he visi. Informaion and forms in Excel forma were sen o he five companies by email and he companies reurned he compleed files o Saisics Sweden in he same way. 4.3 Evaluaion of he es survey Four ou of he five companies have replied and filled in he form. Three of he companies used he Model pricing mehod and i worked well. Four companies submied hourly prices. One company only submied hourly prices as i fel ha he mehod was beer suied o is aciviies. 4.4 Tes resuls and conclusions The pricing mehods used in he es survey will be used in he regular survey. Some minor adjusmens need o be made o he form. 12
5 Design of fuure surveys 5.1 Sampling procedure Service price indices are produced once a quarer. The sample of companies is drawn wih he help of a PPS sample once a year from companies classified in SNI 74.13 in he Business Regiser. In a PPS sample, he sampling probabiliies are proporional o size 5. The sample will be drawn wih regard o he companies' urnovers. A cu-off poin of 10 employees is used o reduce he daa providers' workload. The size of he sample will cover 20 of he indusry's 475 companies. 5.2 Daa collecion procedure The companies will receive user-ids and passwords by pos so ha hey can submi he daa via he SPPI websie. The aim is for all of hem o submi daa in his way. However, for hose who do no wish o submi daa online, a paper form will be devised which he companies can fill in and reurn in a samp-addressed envelope. All he daa will be sored in a daabase. 5.3 Price measuremen mehods The companies may submi prices for 1-6 pre-defined services ha are represenaive of heir aciviies in he indusry. The services are pre-defined by Saisics Sweden as regards a number of pricedeermining facors, including he number of inerviewees, how many quesions and he ype of survey being performed. When i is no possible o submi prices for services or when hese do no represen he company's aciviies, he company may submi he average hourly price for 1-6 personnel caegories. The company chooses hese caegories iself. 5 See he "Mehodological guide for developing producer price indices for services" from he "Join OECD/Eurosa ask force" for a review of PPS samples. 13
5.4 Assumpions, weighs and index calculaions Sep 1 is o calculae an index for each company: For all he companies, a geomeric mean is calculaed for he price change of he services in each company. An index for each company is calculaed by: I m 0, a = i= 1 p p a, i 0 a, i 1/ m Where I 0,a = index wih base ime 0 for company a, a ime p a, i = price of service i, company a a ime 0 p a,i = price for service i, company a, a base ime m = number of services in company a The nex sep is o geomerically weigh an overall index for he indusry from he various company indices. The larges companies are weighed based on heir urnover and he smaller companies share he remaining weigh. The oal index for SNI 74.13 is given by: 14
I 0, T = n ( I 0, a ) a= 1 wa Where I 0,T = oal index wih base ime 0 for he indusry a ime I 0,a = index wih base ime 0 for company a, a ime w a = weigh for company a n = number of companies in he survey for he indusry 5.5 Non-response managemen Should non-response occur, he price rend for he company will be esimaed based on he average rend for he oher companies in he survey. 5.6 Qualiy adjusmen If qualiy adjusmens are necessary, hey will be carried ou in consulaion wih he companies in quesion 6. 6 See he "Mehodological guide for developing producer price indices for services" from he "Join OECD/Eurosa ask force" for a review of qualiy adjusmen mehods. 15