Productivity statistics: Sources and methods. (10th edition)

Transcription

1 Produciviy saisics: Sources and mehods (10h ediion) 1

2 Crown copyrigh This work is licensed under he Creaive Commons Aribuion 3.0 New Zealand licence. You are free o copy, disribue, and adap he work, as long as you aribue he work o Saisics NZ and abide by he oher licence erms. Please noe you may no use any deparmenal or governmenal emblem, logo, or coa of arms in any way ha infringes any provision of he Flags, Emblems, and Names Proecion Ac Use he wording 'Saisics New Zealand' in your aribuion, no he Saisics NZ logo. Reproducion of maerial Maerial in his repor may be reproduced and published, if i does no claim o be published under governmen auhoriy and ha acknowledgemen is made of his source. Liabiliy While all care and diligence has been used in processing, analysing, and exracing daa and informaion in his publicaion, Saisics New Zealand gives no warrany i is error free and will no be liable for any loss or damage suffered by he use direcly, or indirecly, of he informaion in his publicaion. Ciaion Saisics New Zealand (2014). Produciviy saisics: Sources and mehods (10h ediion). Available from ISBN (online) Published in March 2014 by Saisics New Zealand Taauranga Aoearoa Wellingon, New Zealand Conac Saisics New Zealand Informaion Cenre: info@sas.gov.nz Phone oll-free Phone inernaional

3 Conens Tables and figures Lis of abbreviaions Purpose and summary... 7 Purpose... 7 Summary The labour series Inroducion Desirable feaures of a labour volume series Opions for measuring labour inpu The labour volume series The labour inpu index Composiion-adjused labour inpu The capial series Inroducion Capial services consrucion Capial asse ypes wihin he scope of he PIM Capial asse ypes ouside he PIM scope The indusry capial services index Esimaing indusry capial income Roading asses are excluded from indusry capial services indexes The measured-secor capial services index Produciviy esimaes Inroducion The oupu index Labour produciviy and capial produciviy Mulifacor produciviy Conribuions o growh Conribuions o labour produciviy The capial-o-labour raio Growh cycle analysis Uni labour coss Inroducion Raionale Relaing labour coss o labour produciviy Mehodology Daa sources

4 Produciviy saisics: Sources and mehods (10h ediion) 6.6 Indusry coverage Raing forward CoE using LEED Appendix A: ANZSIC06 divisions and published indusries Appendix B: Capial series Saisics NZ asse classes and ypes Glossary References and furher reading References Furher reading

5 Tables and figures Lis of ables 1 Summary of how he economy-wide labour volume series is consruced Indusry coverage under ANZSIC Capial series Saisics NZ asse classes and ypes Lis of figures 1 The wo sages of deriving measured secor capial services Composiion of labour cos

6 1 Lis of abbreviaions AES ANZIND ANZSIC BDD census CFK CoE DoL FTE GDP GEO GFKF GST HLFS HYEIS KAU LEED LVS QES MFP NKS NZIS OECD PIM PKS PPI ULC Annual Enerprise Survey Ausralian and New Zealand Indusrial Classificaion Ausralian and New Zealand Sandard Indusrial Classificaion Business Demography Daabase Census of Populaion and Dwellings consumpion of fixed capial compensaion of employees Deparmen of Labour (Now par of he Minisry of Business, Innovaion and Employmen) full-ime equivalen gross domesic produc geographic uni he aciviies of a KAU (see below) underaken a a specific locaion. Muliple GEOs are recognised where KAUs underake producion a separae permanen physical sies gross fixed capial formaion goods and services ax Household Labour Force Survey Half Yearly Employmen Informaion Survey kind-of-aciviy uni a KAU is a recognisable producion uni (normally an enerprise or a subdivision of an enerprise) a which resources are assembled o produce goods and services and for which accouning records required o esimae value added are available; ypically a facory, warehouse, shop, or office Linked Employer-Employee Daa labour volume series Quarerly Employmen Survey mulifacor produciviy ne capial sock New Zealand Income Survey Organisaion for Economic Co-operaion and Developmen perpeual invenory model producive capial sock producers price index uni labour coss 6

7 2 Purpose and summary Purpose Produciviy saisics: Sources and mehods oulines he daa sources, definiions, and mehodologies used o compile he annual labour, capial, and mulifacor produciviy (MFP) indexes. Summary Produciviy is a measure of how efficienly inpus (labour and capial) are being used in he economy o produce oupus. I is commonly defined as a raio of a volume measure of oupu o a volume measure of inpu use. Labour produciviy growh is a key deerminan of growh in gross domesic produc (GDP) per capia, he benchmark by which counries are ranked on a maerial sandardof-living basis. Growh in GDP iself can come from eiher he conribuions of labour inpu, capial inpu, or MFP. In his sense, produciviy growh becomes a driver of growh in he economy. The produciviy growh indexes cover a subse of he economy, referred o as he measured secor, and are available from he base year of 1978 (year ended March) for he former measured secor and 1996 for he measured secor. The former measured secor and measured secor consis of indusry divisions ha mee hese wo crieria: 1. Consan-price oupu measures are derived independenly of inpus in all principal working indusries for all years. 2. Boh capial and labour inpus are used in all principal working indusries. The measured secor is defined as ANZSIC06 published indusries AA1 o AA3, BB1, CC1 o CC9, DD1, EE1, FF1, GH1, GH2, II1, JJ1, KK1, LL1, MN1, MN2, RS1, and RS2. Appendix A liss he published indusries included in he measured secor. Produciviy series for he measured secor are published wih a one-year lag, while indusry produciviy series are published wih a wo-year lag. Indusry produciviy esimaes are available for he educaion and raining (PP1) and healh care and social assisance (QQ1) indusries. They are no included in he measured secor, which reains comparabiliy wih he Ausralian Bureau of Saisics produciviy series, and mainains focus on he marke secor of he economy. 1 We have adoped an index-number approach o calculae he produciviy series, based on guidelines published by he Organisaion for Economic Co-operaion and Developmen (OECD) in he Produciviy Manual (OECD, 2001b). Apar from he benefi of aligning wih an inernaional consensus on good pracice, following he OECD guidelines should, over ime, allow us o beer compare produciviy saisics beween New Zealand and oher counries. 1 See Saisics New Zealand (2013a) for more mehodological deail for hese indusries. 7

8 3 The labour series 3.1 Inroducion This chaper deails he mehods used in consrucing: (a) a labour volume series (LVS) a indusry level, and (b) a labour inpu index for he measured secor. The LVS measures he quaniy of labour inpu over ime. Hours paid is he uni of labour volume in he LVS. The series is based on hours paid per week for a given quarer. For daa on couns and hours for employees and working proprieors across indusries, he LVS uses he: Quarerly Employmen Survey (QES) Business Demography Daabase (BDD) Household Labour Force Survey (HLFS) Census of Populaion and Dwellings (census) Linked Employer-Employee Daa (LEED) Deparmen of Labour Quarerly Employmen Survey (DoL QES) and Half Yearly Employmen Informaion Survey (DoL HYEIS). 2 These daa are combined wih daa on worker s compensaion o consruc a labour inpu index for use in compiling produciviy saisics. The LVS is based on esimaes of paid hours for all employed people engaged in producing goods and services in New Zealand. There are four componens. Componen 1 includes employees in indusries covered by he QES. From 1987 o 2000, annual BDD couns of employees are inerpolaed using quarerly movemens in employee coun from he QES. This series is hen muliplied by quarerly esimaes of average weekly hours (including overime) from he QES o achieve a quarerly series for paid hours. From 2000 onwards, he mehodology is similar, excep ha annual LEED couns of employees replace he BDD couns. Componens 2 and 3 are esimaed in he same manner, using a combinaion of household survey daa, esablishmen survey daa, census daa, and LEED. Componen 2 covers working proprieors, and componen 3 covers employees ou of scope of he QES. For boh of hese componens, he hours daa are benchmarked using oals from he census. Unil 2000, job couns are also benchmarked from he census, bu from 2000 on, LEED oals provide he benchmarks. From 1988 onwards, movemens beween hours benchmarks are inerpolaed using quarerly esimaes of change from he HLFS, applying he same mehodology used in he inerpolaion of componen 1. Over his period, movemens beween job coun benchmarks are also inerpolaed using quarerly HLFS esimaes of change. Componen 4 covers non-civilian employees in he armed forces. For he enire 1988 o 2012 series, quarerly employee couns for his componen are sourced from he Minisry of Defence. An average 40 working hours a week is assumed and muliplied by he employmen coun o ge a quarerly series of paid hours. The four componens are hen summed o derive indusry oals, and March annual averages of he series are aken. This basic mehodology is complicaed by a number of seps ha make daases consisen over ime. More deail on he four componens is given in secion The Deparmen of Labour is now par of he Minisry of Business, Innovaion and Employmen 8

9 Produciviy saisics: Sources and mehods (10h ediion) The labour inpu index for produciviy measuremen in he measured secor is formed using a Tornqvis index formula ha weighs he annual indusry averages of paid hours by heir relaive shares of compensaion of employees, he laer sourced from he naional accouns. 3.2 Desirable feaures of a labour volume series Key requiremens: consisen and compaible wih he oupu series (ie he gross domesic produc (GDP) series), in erms of indusry deail and coverage, and insiuional breakdown accurae annual esimaes (long-erm rends are of primary imporance, wih secondary imporance aached o quarerly paerns) a ime series sufficienly long o allow he analysis of business cycles a coninuous measure, raher han one based on a poin in ime. 3.3 Opions for measuring labour inpu The following opions for measuring labour inpu are lised in ascending order of concepual correcness. Employmen coun The mos sraighforward measure of labour inpu is a coun of he labour force. This is usually he easies measure o obain, bu does have some drawbacks. One issue wih his opion is ha workers are given he same weigh regardless of wheher hey work full or par ime. A change in produciviy is implied if here is a change in oupu accompanied by a change in he mix of full-ime and par-ime employmen bu he employmen coun is held consan. This occurs because he volume of oupu changes bu measured labour volume has no. Full-ime equivalen A measure of he full-ime equivalen (FTE) labour force is preferred o a sraigh coun measure as i akes ino accoun he mix of full-ime and par-ime employmen. This measure can be consruced when an indicaor of full-ime saus is available. I makes an assumpion abou he relaive inpu of full-ime o par-ime workers. The mos common assumpion is ha a par-ime worker has a weigh of one-half ha of a full-ime worker. An increase in produciviy is implied if here is an increase in oupu accompanied by a change in he relaive inpu beween full-ime and par-ime workers bu he FTE coun, based on a fixed relaiviy, is held consan. Again, his is because he volume of oupu changes bu measured labour volume has no. Hours paid An improvemen on he FTE volume measure is a measure of hours paid. This does no require an assumpion abou he relaive inpu of full-ime o par-ime workers bu i is sill deficien. Ofen workers are paid for a se number of hours, bu i is common for workers o change he number of hours from week o week. An increase in produciviy is implied if here is increased oupu as a resul of more hours worked bu hours paid is unchanged. Noice ha he drawbacks of each of he above mehods can be generalised as no fully accouning for changes in hours worked. Hours of work/acual hours Hours of work or hours worked is a more accurae measure of labour volume han hours paid. However, i reas hours worked by all individuals as equal, regardless of heir 9

10 Produciviy saisics: Sources and mehods (10h ediion) qualiy. An increase in produciviy is implied if here is increased oupu as a resul of changes in he disribuion of hours over skill caegories, bu hours worked is consan and he average produciviy of workers in each skill caegory is unchanged. Composiion-adjused hours worked The mos represenaive measure of labour volume is composiion-adjused (ofen referred o as qualiy-adjused ) hours worked because differences among workers are explicily recognised. To accoun for differences in skill composiion, more weigh should be given o unis of relaively higher skill in aggregaing unis of labour inpu. In doing his, changes in labour composiion ha affec oupu will be refleced in changes in labour conribuion, no as a change in produciviy. An increase in he average qualiy of labour implies ha a composiion-adjused measure will rise faser han an unadjused one. 3.4 The labour volume series Daa sources Daa from esablishmens (or firms), household-based surveys, and adminisraive axaion sources are combined o creae a series ha covers he economy. This is he preferred approach menioned in he Produciviy manual (OECD, 2001b). The manual saes ha naional accouns daa on employmen ha makes use of he comparaive advanages of differen saisical sources should generally be preferred over employmen daa from any single source. Quarerly Employmen Survey (QES) The QES is a sample survey of New Zealand enerprises wih paid employees. I is designed o esimae average hourly and weekly earnings, average weekly hours paid, and he number of filled jobs for he pay week ending on or immediaely before he 20h of he middle monh of he quarer. QES daa are available from he March 1989 quarer onwards. Business Demography Daabase (BDD) BDD daa provide a snapsho of he Business Frame in February, and conain informaion (such as employee job coun) for all unis on he frame. The BDD does no provide qualiy esimaes of he number of working proprieors. The prime objecive of he Business Frame is o provide an up-o-dae and accuraely classified regiser or lis of he survey populaion for Saisics NZ s economic and financial survey programme. The BDD allows analysis by variables ha include region, indusry, insiuional secor, business ype, degree of overseas ownership, and employmen levels. BDD daa are available from 1987 o 2003, bu are used only unil 2000 in he LVS. Household Labour Force Survey (HLFS) The HLFS is a quarerly sample survey ha produces esimaes for he civilian noninsiuionalised usually residen New Zealand populaion aged 15 years and over. Saisics are produced for he employed, unemployed, and hose no in he labour force. Employmen esimaes include couns of employed people and hours of work. HLFS daa are available from he March 1986 quarer onwards. Census of Populaion and Dwellings The Census of Populaion and Dwellings is a five-yearly census of he populaion of New Zealand and is dwellings. Individuals in he household are included in census esimaes of he labour force if hey are 15 years or older. The census produces esimaes of employed people and hours of work. I is he primary source of informaion on he size, composiion, disribuion, economic aciviies, and sae of well-being of he populaion. The daa used for he LVS correspond o published esimaes. For he purposes of he LVS, census daa are available for he March quarers of 1976, 1981, 1986, 1991, 1996, 2001, and

11 Produciviy saisics: Sources and mehods (10h ediion) Census and labour hours non-response The census daa include employees and working proprieors who don record any labour hours. The effec of his non-response is ha labour hours from census daa are undersaed. Tha is, labour hours should be higher. The curren mehodology has been revised o adjus for non-response, increasing he labour hours. The adjusmen involves separaely applying he average hours of employee and working-proprieor respondens a he indusry classificaion level, o he employee and working-proprieor nonrespondens. Employees and working proprieors are separae, disinc groups and have differen characerisics. For example, employees ypically work (on average) fewer labour hours han working proprieors. Noe ha unusually high labour hours recorded by respondens influence average hours upwards. These high hours are generally ascribed o working proprieors and affec indusries wih a relaively higher proporion of working proprieors han employees. Linked Employer-Employee Daa (LEED) LEED uses longiudinal informaion from exising adminisraive axaion sources, and he Business Frame, o provide a range of informaion on he dynamics of he New Zealand labour marke. LEED provides informaion on all axpayers, and is herefore very close o a census of he working populaion. Due o differences in ax forms, here are differences beween he LEED informaion available for wage and salary earners and ha available for working proprieors. LEED informaion is used solely for job couns in he LVS, and is available from he June 1999 quarer onwards. Adminisraive daa from he Minisry of Defence Quarerly couns of non-civilian employees in he armed forces are available from he Minisry of Defence. For he LVS, his adminisraive daa are available from 1978 onwards. Deparmen of Labour Quarerly Employmen Informaion Survey (DoL QES) and Half Yearly Employmen Informaion Survey (DoL HYEIS) For he LVS, he DoL employmen daa are available from 1978 o The DoL HYEIS was a full-coverage semi-annual survey unil 1980, when i was renamed he DoL QES, and provided quarerly sample daa he February quarer was full coverage. Minor changes o he coverage of he survey were also made in 1980, and i became a posal quesionnaire. Firm-based daa Firm-based daa was seleced as he primary source for he LVS. Indusry daa from esablishmen surveys are more robus, have a beer oupu link, and greaer poenial for generaing a longer ime series of reasonable qualiy Hours paid he measure of labour inpu in he labour volume series Toals for hours are calculaed on an hours paid basis raher han hours of work basis. Hours paid is he number of ordinary and overime hours for which an employee is paid. I excludes unpaid overime bu may include some hours ha are no acually worked, such as paid leave and sauory holidays. While hours of work is he preferred measure of labour inpu, only hours paid is available from he QES. The selecion of hours paid over hours of work has mean rading off he heoreical preference of hours of work wih robus indusry daa from he QES, BDD, and LEED. Two main reasons are behind choosing hours paid as he measure of labour inpu. 11

12 Produciviy saisics: Sources and mehods (10h ediion) Greaer confidence in indusry esimaes of hours paid Robus indusry daa and good alignmen wih oupu and capial daa are imporan for producing indusry measures of produciviy. Acual hours worked, derived from household survey daa, would have been seleced as he measure of labour inpu if he sole objecive was a measured-secor (or economywide level) series wihou any indusry splis. A he measured-secor level, he annual change in acual hours worked (esimaed from he HLFS) is as saisically robus as hours paid (esimaed from he QES). However, here are some caveas: A he indusry level, he hours paid measure is more robus han he hours worked measure. We have more confidence in aligning indusry labour inpus wih corresponding indusry oupus using hours paid daa. Given ha he mehodology we use o calculae he labour volume index is based on aggregaed indusry-level daa, i is desirable o have good alignmen of indusry inpu and oupus. Using esablishmen surveys ensures consisency wih he oher componen series of produciviy. The QES and he BDD largely survey he same enerprises ha are covered by he daa sources feeding ino he oupu and capial series. The annual change in hours paid a he aggregae level is no significanly differen from he annual change in acual hours worked. For produciviy purposes, he main ineres is in he annual growh of he volume series. I is assumed ha he annual growh in hours paid is a good proxy for he growh in acual hours worked. HLFS indusry daa can be volaile. We don know how much of any quarerly movemen in esimaed indusry oals of acual hours worked is affeced by real flucuaions in work as a resul of holidays or exernal facors. Qualiy of responses and coding can creae volailiy, bu his impac on change esimaes has ye o be fully invesigaed. Furher work required o model indusry oals for acual hours worked Curren opions for modelling indusry oals of acual hours worked are based on HLFS daa and are no suiable, given concerns abou he robusness of HLFS indusry oals and insufficien undersanding of facors affecing quarerly indusry movemens. More work is required o invesigae possible auxiliary daa sources ha can be used for modelling acual hours worked Componens in he series From 1988 o 2000, esablishmen survey daa (from he QES and BDD) is he primary daa source for he LVS. Daa for coverage gaps (working proprieors and employees in indusries no covered by he QES and BDD) come from household survey daa (census and he HLFS). As a resul, he LVS combines people-coun daa from he HLFS and census, wih job-coun daa from he QES and BDD. There is a concepual difference in ha he HLFS couns people, whereas he laer wo sources coun jobs. The HLFS collecs limied informaion on muliple-job holding, and indusry informaion relaes o a responden s main job only. This is an issue regardless of how he series is compiled, as here is no workable alernaive o using household survey daa o fill gaps in he QES coverage. From 2000 onwards, all job-coun and people-coun benchmarks are sourced from LEED. BDD is no used from his ime, and census daa are used only o provide hours benchmarks. In he fuure, i is anicipaed ha LEED will coninue o provide he job- and people-coun benchmarks. To ensure consisency wih he indusry capial and oupu series, he source daa for he LVS has been aligned wih he indusry definiions (referred o as ANZIND or ANZSIC working indusries) used for measuring New Zealand s gross domesic produc (GDP). The indusry alignmen is performed on daa (pos-1987) from he QES and BDD a he 12

13 Produciviy saisics: Sources and mehods (10h ediion) kind-of-aciviy uni (KAU) level raher han a he enerprise or geographical locaion level. This is in line wih he level used in he calculaion of oupu (GDP). LEED, which is available a he geographic uni (GEO) level, has been aggregaed up o he KAU level o mainain consisency wih he BDD and QES. Differen mehodologies are used o calculae four componens in an economy-wide LVS. Componen 1 Employee jobs in indusries covered by he QES (sourced from esablishmen or firm-based daa, and adminisraive axaion daa) This componen of he quarerly hours paid series is consruced by firs esimaing an employee job-coun series. From 1978 o 1980, job couns are obained from full-coverage semi-annual daa from he DoL HYEIS. Linear inerpolaion is used o conver he biannual daa o quarerly. For 1980 o 1988, he full-coverage February quarer acs as a benchmark, wih he May, Augus, and November quarers being inerpolaed using growh raes a he ANZSIC working-indusry-by-secor level. From 1989 o 2000, quarerly movemens in he QES 3 job coun are used o inerpolae beween annual employee job-coun oals from he BDD, a he ANZSIC workingindusry-by-secor level. ANZSIC working indusries (or ANZIND) are defined by six-digi ANZSIC codes. From 2000 onwards, LEED provides he annual employee job-coun oals, a he ANZSIC working-indusry level. LEED monhly poin-in-ime daa (from he 15h of each monh) on he couns of jobs, and people in jobs, are used. Job couns are combined wih quarerly job-level hours-paid daa from he QES o calculae he oal hours paid in each indusry. There are a few excepions o his sandard mehodology where oher daa sources are judged o produce more suiable labour volume esimaes. The quarerly hours paid series is hen compiled as he produc of he inerpolaed jobcoun series and he average weekly hours paid per job calculaed from he QES. (The average weekly hours paid per job from he QES is calculaed by dividing oal hours paid per week, as recorded in he survey, by he survey employee job coun.) Quarerly oals of hours paid are no furher benchmarked, as hours daa are no colleced for he BDD or LEED. Componen 2 Working proprieors (sourced from household survey daa, and adminisraive axaion daa) Refer o componen 3. Componen 3 Employees in indusries excluded from QES (sourced from household survey daa, and adminisraive axaion daa) Throughou he LVS, usual hours worked (from he census) provide a benchmark o esimae hours paid for componens 2 and 3. This measure is he number of hours he responden usually works in employmen during a paricular reference period, even if hey did no in fac do so during he survey reference period because of emporary absences (eg due o sickness or holidays). The QES measure of hours paid is differen in concep o usual hours of work measured by he HLFS and census. I is assumed, a leas for he relevan indusries, ha hese are close enough o allow hem o be reaed as measuring he same hing. 3 QES esimaes of change are calculaed a he ANZIND level. 13

14 Produciviy saisics: Sources and mehods (10h ediion) From 1978 o 1986, for indusries wihin he scope of he DoL employmen surveys, census benchmarks for working-proprieor couns are inerpolaed using movemens in he DoL employmen survey daa. This daa is limied o working proprieors who employ wo or more saff, implicily assuming ha his group is represenaive of all working proprieors. Working-proprieor hours are inerpolaed using movemens in employee hours from he DoL surveys. For agriculural sub-indusries ouside he scope of he DoL survey, census daa for 1976, 1981, and 1986 are used o creae an annual working-proprieor average-hours series. This series is creaed by using linear inerpolaion beween census benchmarks. This series is hen combined wih Agriculure Census employmen daa o ge an annual series of agriculure hours worked. For oher indusries where robus daa are no available from he DoL surveys or Agriculure Census, linear inerpolaion beween census benchmarks is used o creae a quarerly series of labour inpu. From 1986 o 2000, he quarerly series for employmen, and hours paid for he above wo componens, are derived by inerpolaing beween five-digi ANZSIC oals from he census, using quarerly esimaes of change from he HLFS a he published one-digi ANZSIC level. The resuling five-digi ANZSIC-level series are hen aligned o ANZIND for he relevan indusries. Usual hours worked (from he HLFS) is used as a proxy for hours paid. From 2000 onwards, he series for employmen is sourced from LEED for boh componens 2 and 3. Consisen wih he earlier period, hours paid are derived by inerpolaing beween five-digi ANZSIC oals from he census, using quarerly esimaes of change from he HLFS a he published one-digi ANZSIC level. Monhly poin-in-ime LEED daa on secondary jobs are also used, combined wih HLFS daa on hours worked in secondary jobs. These daa are no disaggregaed by worker ype or indusry. A small proporion is added o he employmen coun o accoun for unpaid family work. These daa are sourced from he HLFS. For working proprieors, quarerly proporions from he HLFS and QES are used o add seasonaliy o he LEED couns, while keeping he annual average couns unchanged. Componen 4 Non-civilian employees in he armed forces (sourced from adminisraive daa) Quarerly employmen couns come from he Minisry of Defence. The Minisry does no provide hours paid daa. In esimaing oal hours paid for non-civilian employees in he armed forces, 40 hours is assumed o be he average weekly oal. 14

15 Produciviy saisics: Sources and mehods (10h ediion) Table 1 1 Summary of how he economy-wide labour volume series is consruced Summary of how he economy-wide labour volume series is consruced (1) (2) Indusry Daa source for employee coun Daa source for employee hours Daa source for workingproprieor coun Daa source for workingproprieor hours DoL QES DoL QES DoL QES Census/DoL QES Census/DoL indusries (3) QES Agriculure Services o agriculure; huning and ramping Commercial fishing Residenial propery operaors Privae households employing saff Non-civilian defence saff Foreign governmen represenaion Inernaional sea ranspor Census/Agriculure Census Census Census/Agriculure Census Census Census Census Census Census Census Census Census Census Census Census DoL QES Census/DoL QES Census Census DoL QES DoL QES Adminisraive 40 hours/week daa Excluded Excluded Noe: For foonoes and symbols, see end of able. Table 1 coninues nex page 15

16 Produciviy saisics: Sources and mehods (10h ediion) Table 1 coninued Indusry Daa source for employee coun Daa source for employee hours Daa source for workingproprieor coun QES indusries BDD (4) /QES jobs QES paid hours coun Agriculure Services o agriculure Commercial fishing Residenial propery operaors Privae households employing saff Non-civilian defence saff Foreign governmen represenaion Inernaional sea ranspor coun coun coun coun coun usual hours usual hours usual hours usual hours usual hours coun coun coun coun coun Adminisraive 40 hours/week daa Excluded Excluded Noe: For foonoes and symbols, see end of able. Daa source for workingproprieor hours usual hours usual hours usual hours usual hours usual hours usual hours Table 1 coninues nex page 16

17 Produciviy saisics: Sources and mehods (10h ediion) Table 1 coninued 2000 onwards (5) Indusry Daa source for employee coun Daa source for employee hours Daa source for workingproprieor coun Daa source for workingproprieor hours QES indusries LEED QES paid hours LEED (6) usual hours Agriculure LEED usual hours Services o agriculure Commercial fishing Residenial propery operaors Privae households employing saff Non-civilian defence saff Foreign governmen represenaion Inernaional sea ranspor LEED LEED LEED LEED usual hours usual hours usual hours usual hours LEED (6) LEED (6) LEED (6) LEED (6) LEED (6) Adminisraive 40 hours/week daa Excluded Excluded usual hours usual hours usual hours usual hours usual hours 1. Exclusions from he series include inernaional sea ranspor (people working in his indusry are working abroad) and foreign governmen represenaion (embassies are deemed o be island saes and economies of heir paricular home counry). 2. Daa sourced from he census and Agriculure Census are linked o daa in All oher DoL QES-based employee daa are linked o BDD/QES in The DoL QES sared in Before his, DoL HYEIS daa were used. 4. Annual BDD employee-coun benchmarks are incorporaed ino he series from 1987 for mos indusries, and from 1988 for he remainder. 5. Where LEED daa are unavailable due o imeliness, labour couns are raed forward using movemens in he QES and HLFS. 6. LEED couns for working proprieors are based on annual daa, supplemened by daa from HLFS and QES. Symbol: no applicable 17

18 Produciviy saisics: Sources and mehods (10h ediion) Oher issues HLFS sample redesigns The HLFS sample is redesigned periodically (approximaely every five years) based on daa from he populaion census. When he sample is redesigned he sample sraificaion may be changed, all primary sampling unis (PSUs) are re-seleced, and weighs are adjused. Noe ha his differs from a sample rebase, where weighs are adjused in order o benchmark he oal working-age populaion o he census working-age populaion. HLFS sample redesigns occurred in , , and For he firs wo redesigns, he new sample was phased in over four quarers, wih wo roaion groups from he new sample roaed in and wo roaion groups from he old sample roaed ou in each of he four quarers. In he mos recen redesign, he new sample was phased in over eigh quarers. The objecive for inroducing he sample over four or eigh quarers is o smooh he ransiion beween samples. This is why any variaion in he series due o a redesign is no obvious when viewing a ime series of HLFS daa. During he phasing in of he new sample, movemens repored in he HLFS will be due no only o changes in he labour marke bu also o any differences beween he new and old sample. Any analysis of HLFS daa during hese redesign periods should consider he sample redesign. In general, he sample redesign does no affec he general labour marke rends. However, in several changes made o he design of he HLFS sample may have increased he measured growh in employmen during his period, 4 including he sraificaion basis changing from urban/rural area o regional council area. Oher changes were as follows: The number of sraa reduced from 194 more variable-sized sraa o 122 more even-sized sraa. Proporionaely fewer rural PSUs were seleced. Proporionaely more PSUs were seleced from sraa wih relaively high concenraions of Māori. Targeing Māori in he new HLFS design improved he accuracy of Māori saisics. The number of PSUs reduced from 2,400 o 1,752. The average number of households seleced in each PSU increased from 6.5 o 9.0. QES and casual employmen The colleced daa relae o paid employees of all ages bu casual labour is no well covered. This is paricularly relevan during periods where environmenal facors affec he composiion of he labour force. For example, one such environmenal change occurred when he Employmen Conracs Ac 1991 came ino force. One objecive of he Ac was o increase flexibiliy in he labour marke; i is quie likely ha along wih his increased flexibiliy came an increase in he number of casual workers. Daa abou unpaid labour are no colleced by he QES. In he majoriy of businesses, daa are exraced from he payroll sysem. Individuals wih more han one paid job can be couned a each workplace if hey are no in an indusry excluded from he QES, or in non-economically significan enerprises. Esimaed couns of working proprieors wih employees are produced, bu he qualiy is no good, and daa on hours are no colleced for his group. 4 Some of hese changes were inroduced due o he availabiliy of new informaion or daa ha were useful for sample design for insance, regional councils were creaed in

19 Produciviy saisics: Sources and mehods (10h ediion) Creaing a coninuous LVS Various daa sources have been used o compile a long-erm coninuous series. The key linking period for daases is 2000 as LEED daa are inroduced as he sole daa source for couns of employees, and as he benchmark for working-proprieor couns. The linking is done a he ANZSIC working-indusry level, meaning ha published indusry and measured-secor aggregaes will also be linked o creae a coninuous series. The renal, hiring, and real esae services (LL1); professional, scienific, and echnical services (MN1); adminisraive and suppor services (MN2); and oher services (RS2) indusries are included in he expanded measured secor from 1996 onwards. No aemp is made o creae a coninuous LVS a his poin, so here is a break in he series. Timeliness of LEED The LEED series is he preferred daa source for employmen and working-proprieor couns. However, due o he way daa are colleced, he series is no as imely as oher survey-based daa sources. Because of his, he series is raed forward o calculae he laes period s labour couns when necessary. This is done by using movemens in QES for employees in indusries wihin is scope, and using HLFS daa for all oher indusries employees and working proprieors. This QES and HLFS daa are esed for robusness, and adjusmens are made where necessary. There is generally a 14-monh lag following he quarerly reference period for LEED employmen daa, and a 20-monh delay in LEED annual working-proprieor daa becoming available. This means ha he laes quarer of employee daa is no imely enough, and neiher is he laes year of mos LEED working-proprieor daa. LEED is also subjec o revision, due o lae ax forms being processed. This primarily affecs he working-proprieor daa. To accoun for his, he working-proprieor series is raed up, using previous years undercoun proporions daa unil he LEED daa are no longer provisional Qualiy assurance of he indusry labour volume series As a qualiy assurance measure, several coherency adjusmens are made o he employee coun and hours series ha feed ino he measured-secor LVS. The main daa sources in consrucing he LVS are sourced independenly of he esimaes of compensaion of employees (CoE) from he naional accouns. CoE esimaes are primarily derived from he Annual Enerprise Survey (AES), while LVS esimaes are compiled using a number of sources, as discussed above. Curren-price CoE esimaes are deflaed using he QES average hourly earnings measure o provide an implici LVS. This provides a benchmark for comparing agains he LVS a an indusry level. Adjusmens are made o he indusry daa, based on alernaive labour daa sources in he years where he LVS shows a significanly differen movemen o he deflaed CoE series. 3.5 The labour inpu index For produciviy purposes, he main ineres is in he annual growh rae of labour inpu for he measured secor. To derive his rae, indusry annual-average hours paid per week are firs calculaed for he March year for each indusry. These annual values of indusry labour volume are he simple average of he four quarerly values. A chain-linked Tornqvis index is hen used o calculae he measured-secor labour inpu index. The quaniy relaives in he index are wo-period raios of annual indusry labour volumes. Indusry wo-period mean shares of measured-secor nominal labour income form he exponenial weighs. Assuming a posiive correlaion beween indusry labour incomes and skill levels, his indusry weighing regime goes some way owards qualiy adjusing he labour volume index. Qualiy adjusing in his manner means ha he series capures, o some degree, 19

20 Produciviy saisics: Sources and mehods (10h ediion) changes in he qualiy (ie skill level) of labour used in producing oupu. This implies ha an increase in he aggregae level of oupu due o an increase in he skill level of he labour force will only parially show up as an increase in labour produciviy. If he labour measure did no have his degree of implici qualiy adjusmen, he increase in oupu would be fully refleced as an increase in labour produciviy. The formulas for he Tornqvis index and is weighs are given in he equaions below: Equaion 3.1 L L ( 1) i i( 1) Equaion 3.2 L L i Wl i Wl i 1 2 i Yl i Yl i + i Yl i( 1) Yl i( 1) Where L measured secor volume of labour in period L i volume of labour in indusry i, period Yl i curren price labour income in indusry i, period. A Tornqvis index is used o consruc he measured-secor labour inpu index. This is consisen wih he formulaions of he measured secor capial services series and he composie inpu series used in calculaing mulifacor produciviy Scope of he labour inpu index he measured secor The labour inpu index covers he subse of he economy ha is referred o as he measured secor. This includes ANZSIC06 published indusries AA1 o AA3, BB1, CC1 o CC9, DD1, EE1, FF1, GH1, GH2, II1, JJ1, KK1, LL1, MN1, MN2, RS1,and RS2. Indusries include: agriculure foresry and fishing, and services o agriculure, foresry, and fishing mining food, beverage, and obacco manufacuring exile, leaher, clohing, and foowear manufacuring wood and paper producs manufacuring prining peroleum, chemical, polymer, and rubber producs manufacuring non-meallic mineral producs manufacuring meal producs manufacuring ranspor equipmen, machinery, and equipmen manufacuring furniure and oher manufacuring elecriciy, gas, waer, and wase services consrucion wholesale rade 20

21 Produciviy saisics: Sources and mehods (10h ediion) reail rade accommodaion and food services ranspor, posal, and warehousing informaion media and elecommunicaions financial and insurance services renal, hiring, and real esae services professional, scienific, and echnical services adminisraive and suppor services ars and recreaion services oher services. The measured secor excludes indusries ha are problemaic for produciviy measuremen. Excluded indusries are hose for which esimaes of value added a consan prices are derived largely using inpu mehods, such as number of employees. This mainly comprises governmen non-marke indusries whose services are provided free or a nominal charges, such as governmen adminisraion and defence. The LVS is consruced a he ANZSIC working-indusry level for all indusries (excep he exclusions noed in able 1). I can be aggregaed o he published indusry, measuredsecor, or oal economy-wide level. For esimaed employmen in he measured secor (on average): componen 1 conribues abou 72 percen componens 2 and 3 (combined) conribue abou 28 percen. In erms of hours paid in he measured secor (on average): 5 componen 1 conribues abou 69 percen componens 2 and 3 (combined) conribue abou 31 percen Esimaing indusry labour income To calculae he measured secor labour inpu index, i is necessary o derive esimaes of indusry labour income in curren prices. The annual March-year naional accouns include esimaes of indusry gross operaing surplus, compensaion of employees, and ne indirec axes on producion. However, gross operaing surplus includes he ne mixed income received by working proprieors and hose who are self-employed. This ne mixed income mus be separaed ino is wo componens he income flows ha resul from he use of labour and capial. The previously calculaed employee and working-proprieor hours paid series, along wih he naional accouns compensaion-of-employees daa, are used o allocae he indusry ne mixed income o he primary facors. The labour income of working proprieors is impued as: Equaion 3.3 Ylwp i COE i WPhp TEhp i i 5 Componen 4 is par of he governmen adminisraion and defence indusry and is herefore excluded from he measured secor. Componen 4 conribues less han 1 percen o oal-economy hours paid. 21

22 Produciviy saisics: Sources and mehods (10h ediion) Where Ylwp i labour income of working proprieors in indusry i, period COE i compensaion of paid employees in indusry i, period WPhp i working proprieor hours paid in indusry i, period TEhp i oal hours for paid employees in indusry i, period. The labour income of working proprieors for indusry i, as defined in equaion 3.3, has a ceiling; i canno be greaer han he ne mixed income of ha indusry. In such cases, he ne mixed income apporioned o capial will be equal o zero. A share of ne axes on producion is apporioned o labour income as well. This is based on he relaive share of labour income o he sum of labour and capial income. Toal ne axes on producion aribuable o labour are given by: Equaion 3.4 NTl i COE COE i i + Ylwp + GOS i i NT i Where NTl i ne axes on producion aribued o labour income in indusry i, period NT i ne axes on producion in indusry i, period. GOS i gross operaing surplus in indusry i, period. So ha curren price indusry labour income is given by: Equaion 3.5 Yl COE + Ylwp + NTl i i i i Where Yl i labour income in indusry i, period. 3.6 Composiion-adjused labour inpu As noed, he heoreically ideal approach for measuring labour is hrough composiion adjusmen. A composiion-adjused labour inpu measure no only capures changes in he quaniy of labour being uilised, bu also he skill level of labour. The approach firs calculaes an index of he skill level of workers, based on movemens in he disribuion of proxies for skill, educaion, and esimaed work experience, and hen applies he movemens of his series o he unweighed LVS mos recenly published. The resul of his is a labour inpu series ha accouns for changes in boh he qualiy and quaniy of labour. The HLFS is he main daa source for composiion-adjusmen, providing a breakdown of he employed by he variables of ineres. The New Zealand Income Survey (NZIS) is an annual supplemen o he HLFS. NZIS provides average incomes for each caegory of labour, which provides informaion on he relaive incomes. These income shares are used in an approach similar o ha for compiling he labour inpu index, as specified in he Tornqvis index specificaion in secion 3.5. See Saisics New Zealand (2008) for more informaion on he composiion-adjused labour inpu series. 22

23 4 The capial series 4.1 Inroducion This secion deails he capial services series. This inpu series, along wih he labour inpu series and consan-price value-added daa, is used in forming produciviy esimaes. 4.2 Capial services consrucion Consrucing esimaes of capial produciviy and mulifacor produciviy (MFP) using an index number echnique requires an index represening he flow of capial services. Capial services are generaed by using capial asses over a specified period of ime (year ended March in his case). Capial asses include 24 of he 26 produced asses, which are esimaed in he naional accouns using a perpeual invenory mehod (PIM). 6 The PIM accumulaes hese asses invesmen flows and applies reiremen, efficiency, and discoun parameers o derive esimaes of producive capial sock (PKS), ne capial sock (NKS), and consumpion of fixed capial (CFK). In addiion o hese asses, capial services esimaes encompass land (for all indusries excep fishing), and invenories in seleced indusries. Livesock (in agriculure) and imber (in foresry) are included in he invenories esimaes. See appendix B for a full lis of he asses wihin he scope of he Saisics NZ capial services series. The consrucion of he capial services index for he measured secor has wo disinc seps. For each indusry wihin he measured secor, a capial services index is calculaed by aggregaing he services provided by he capial asses. This is he process of horizonal aggregaion (see figure 1). Secondly, hese indusry indexes are aggregaed o he measured-secor level, which is verical aggregaion. Figure 1 1 The wo sages of deriving measured secor capial services The wo sages of deriving measured secor capial services Asses 1 2 N Indusry CS 1a CS 2a CS na A K a CS 1b CS 2b CS nb B K b K MS In figure 1, CS xy represens he capial services generaed by asse x in indusry y. The indusry capial services index is denoed by K y and K MS represens he measured-secor capial services index. 6 Cenral governmen roading and local governmen roading are excluded. Refer o secion 4.7 for furher deail. 23

24 Produciviy saisics: Sources and mehods (10h ediion) The capial services indexes are calculaed a he aggregae (wo-characer) indusry level, as published in he naional accouns. Wihin he measured secor 25 published indusries are formed from 103 working indusries. Appendix A has a full lis of published indusries. Esimaes of capial services cover he privae and governmen secors. This coverage is consisen wih he labour and oupu series. 4.3 Capial asse ypes wihin he scope of he PIM Produciviy sudies require a measure of he PKS. An asse s producive capial sock is is gross capial sock adjused for he decline in is efficiency. PKS a consan prices represens sandardised efficiency unis and can be inerpreed as a measure of he poenial capial services ha he asse can conribue o he producion process. 7 For each asse ype, he Saisics NZ PIM specifies a mean expeced useful life, a reiremen funcion based on a disribuion abou his life, and a paern of (hyperbolic) efficiency decline. These parameers and gross fixed capial formaion (GFKF) in consan prices are used o esimae an asse ype s PKS in consan prices. Considering one asse ype of a single vinage, a hyperbolic age-efficiency funcion is specified as: Equaion 4.1 E M A 0 M ba ( A M) Where E he efficiency of he asse a ime, relaive o is efficiency when new M he asse life, as per an assumed reiremen disribuion A he age of he asse of he given vinage, a ime b he efficiency reducion parameer. To aggregae he efficiency unis of asses wih differen lives (M), i is necessary o consruc he weighed average of he age efficiency erms. The weighs are he relaive frequencies of asse lives, as deermined by he Winfrey probabiliy funcion. Equaion 4.2 M max R( A ) M M A P( M) M ba A ( 0 A M) Where R(A ) relaive efficiency of asses of age A P(M) relaive frequency of asses wih life M M max maximum useful life of he asse, as deermined by he Winfrey funcion. The real producive capial sock of an asse ype wihin a given indusry is hen calculaed as he sum of he relaive efficiency of each vinage, weighed by is GFKF: 7 The producive capial sock measures a flow of poenial capial services in one period, which can be generaed by he asses in exisence. For his reason, Measuring capial (OECD, 2001a) does no use he erm producive capial sock. 24

25 Produciviy saisics: Sources and mehods (10h ediion) Equaion 4.3 M max PKS R( A). GFKF A 0 ( A) Where PKS producive capial sock of he asse ype in period, in consan prices GFKF gross fixed capial formaion in consan prices. The criical inpu o he PIM is curren-price GFKF. GFKF daa a a deailed asse level are esimaed by working indusry, secor of ownership (privae, local, or cenral governmen), and marke group (marke or non-marke enerprise) as par of he annual naional accouns supply-use balancing process. The supply-side esimaes are sourced from daa on impors and domesic producion AES sales figures and quarerly surveys. Demand-side esimaes are obained from AES GFKF daa, direc enquiry, he financial accouns of large corporaes, and governmen-secor financial accouns. For he years in which he demand and supply esimaes have ye o be reconciled hrough supply-use balancing, economy-wide GFKF daa by asse ype and secor are obained from quarerly GDP esimaes. Where specific informaion is no ye available, indusry and marke group allocaions are made according o he proporions from he laes balanced year. Daa for all years ener he PIM by 26 asse ypes, by secor, by marke group, and by working indusry. From his poin, consan-price GFKF esimaes are derived using price index deflaion. Age-price and age-efficiency funcions are hen used o form esimaes of PKS, NKS, and CFK. The AES covers approximaely 90 percen of New Zealand s economically significan enerprises. An enerprise is deemed o be economically significan if i mees a leas one of he following crieria: has annual GST expenses or sales of more han $30,000 has a rolling mean employmen coun of more han wo over any 12-monh period is in a GST-exemp indusry (excep residenial propery leasing and renal) is par of a group of enerprises is a new GST regisraion ha is compulsory, special, or forced is GST regisered and involved in agriculure or foresry. Indusries excluded from he AES include residenial propery operaors no elsewhere classified, and foreign governmen represenaion, neiher of which are in he Saisics NZ definiion of he measured secor. Religious organisaions, and privae households employing saff are also excluded from AES. The elemenal level a which PIM oupus are produced is he working-indusry-by-asseype level. For a single asse ype, esimaes of hese oupus a he published-indusry level are he sums of he corresponding working-indusry level values. This is so for boh curren and consan prices. The asse price indexes employed in he PIM for deflaion (and reflaion) are specified by asse ype and by secor (privae and governmen), bu are invarian o he indusry of use. 8 However, he implici price deflaors ha he PIM produces (a he working-indusry- 8 Asse price indexes are used in deflaing curren price GFKF and he subsequen reflaion of consan price NKS, PKS, and CFK esimaes. The majoriy of hese price indexes are provided by Saisics NZ s quarerly capial goods price index. 25

26 Produciviy saisics: Sources and mehods (10h ediion) by-asse-ype level) are indusry as well as asse specific. The indusry deflaor is influenced by he relaive conribuions of he secor s capial socks o he indusry s overall capial sock. To be consisen wih he elemenal level of analysis, our calculaion of capial services employs hese implici price deflaors as he asse price indexes. A comprehensive descripion of Saisics NZ s PIM is provided in he manual Measuring capial sock in he New Zealand economy (Saisics NZ, 2013b). 4.4 Capial asse ypes ouside he PIM scope Land Esimaes of producive land sock are produced for he enire measured secor. Esimaes of agriculural land use, by hecares, come from Saisics NZ agriculural producion censuses and surveys. For years where agriculural producion censuses and surveys are no held, land-volume daa from Quoable Value Ld are used o inerpolae. The volume esimae for foresry land use, by hecares, comes from he Minisry of Agriculure and Foresry s Naional Exoic Foresry Descripion repors. Esimaes of residenial, commercial, indusrial, mining, and oher non-agriculural land volumes come from Quoable Value Ld annually. Residenial, commercial, indusrial, mining, and oher non-agriculural land are included in he capial inpu series from 1996 onwards. 9 Curren-price esimaes of he sock of land are calculaed for all indusries by muliplying heir respecive volumes (hecares) by a price per hecare. Average per hecare prices are esimaed by using valuaion and volume daa from Quoable Value Ld. The resuling curren-price series include he value of land improvemen expendiure, hereby creaing a double coun wih he PIM asse ype land improvemens. To address his, currenprice producive capial sock esimaes of land improvemens, sourced from he PIM, are deduced from he curren-price land values. These curren-price values, ne of land improvemens, are hen divided by land volumes o esablish implici price deflaors. Volume esimaes of land for produciviy analysis are derived by quaniy revaluaion using 1995/96 prices. These volumes are boh he land s producive capial sock and ne capial sock, as land is regarded as a non-depreciable asse Invenories Invenories are included wihin he scope of he capial asses. Price and volume esimaes of invenories are obained from he naional accouns for seleced indusries in he measured secor. They are included from 1978 for he agriculure and foresry indusries, and from 1987 for he manufacuring; wholesale rade; reail rade; and accommodaion, cafés, and resaurans indusries. The daes reflec he availabiliy of source daa. Livesock and imber socks, used wihin he agriculure and foresry indusries, respecively, are now sourced from he naional accouns as a par of he invenories series from 1980 onwards. Before 1980, movemens are linked o he naional accouns-sourced series. 4.5 The indusry capial services index The capial services index for an indusry is a chain-linked Tornqvis index of he capial services provided by he asse ypes employed in ha indusry. A Tornqvis index is used, 9 Esimaes of he producive land sock for he educaion and raining, and healh care and social assisance indusries are compiled in he same fashion. 26

27 Produciviy saisics: Sources and mehods (10h ediion) o mainain consisency wih he consrucion of he inpu index for MFP esimaes, and he labour volume index. Because movemens in he volume of capial services provided by an asse are unobservable, hey are proxied by movemens in he asse s producive capial sock (ie he poenial volume of capial-service flows based on he asse s physical sock and ageefficiency funcion). This is effecively an assumpion of iner-emporal consan capaciy uilisaion. Assuming consan capaciy uilisaion is a simplificaion ha ignores ha raes vary over business cycles. The Tornqvis index is consruced as he geomeric mean of asses wo-period producive-capial sock raios (represening capial services raios) weighed exponenially by each asse s mean wo-period share of he indusry s value of capial coss. An asse s cos of capial is is user cos (renal price) muliplied by is flow of capial services. 10 Equaion 4.4 KC u ij ij PKS ij Where KC ij cos of capial services for asse j in indusry i, period u ij user cos of asse j in indusry i, period PKS ij producive capial sock of asse ype j in indusry i, period. Indusry capial cos is he sum of each asse s cos of capial services: Equaion 4.5 KC i j u ij PKS ij Where KC i cos of capial services in indusry i, period. So ha he capial services index is represened by equaion 4.6 and is associaed weighs by equaion 4.7: Equaion 4.6 C PKS k i ij i C i j PKS ( 1) ij( 1) Equaion 4.7 Wij 1 KC Wij 2 KC ij i KC + KC ij( 1) i( 1) Where k i capial services index for indusry i in period C i volume of capial services produced by indusry i in period. 10 Capial service flows are assumed o be proporional o he level of he producive capial sock. 27

28 Produciviy saisics: Sources and mehods (10h ediion) The user cos is he cos of using a capial good for a specified period (in his case, one year). I is analogous o he wage rae in he LVS. In he conex of hin or non-exisen renal markes, he user cos is approximaed by he implici renal ha owners of capial pay o hemselves. We adop a user cos ha is deermined by four facors: he asse price, as new, relaive o is base-period price a real rae of reurn he asse s rae of economic depreciaion he effecive rae of non-income ax on producion. This user cos is given by equaion 4.8: Equaion 4.8 u p ( i + d ) + p ij ij ij ij x i Where p ij he price index of new capial asse j in indusry i, period i he real rae of reurn (se a 4 percen) d ij he rae of economic depreciaion of asse j in indusry i, period x i he average non-income ax rae on producion for indusry i, period. The asse price erm, p ij, is he implici price index ha resuls from dividing an asse s ne capial sock in curren prices by is consan-price equivalen. Equaion 4.4 shows ha he cos of capial services for a single asse is equal o is user cos muliplied by he volume of capial services, represened by PKS. In pracice, PKS (he capial-services volume) is measured in consan dollars, hence he user cos (he capial-services price) is calculaed as a price relaive, so he resuling produc is measured in dollar unis. The user cos in he above formulaion is effecively uni free. Using he asse price erm p ij resuls in he cos of capial services being expressed in he prices of period. The rae of economic depreciaion, which combines he devaluing effecs of efficiency loss and ageing on an asse, is he raio of he consumpion of fixed capial o is producive capial sock: Equaion 4.9 CFK d ij PKS ij ij Where CFK ij consumpion of fixed capial of asse j in indusry i, period PKS ij producive capial sock of asse j in indusry i, a end of period. Boh CFK ij and PKS ij values are in consan prices. The rae of economic depreciaion (equaion 4.9) is expressed wih PKS ij as he denominaor in order o derive he correc cos of capial services (refer o equaion 4.4 in which PKS ij appears as he volume measure). Non-income business axes are axes on producion ha are assignable o capial inpus. The parameer x i is he raio of hese axes a he indusry level o he indusry s consan-price producive capial sock. 28

29 Produciviy saisics: Sources and mehods (10h ediion) Equaion 4.10 x i NTk PKS i i Where NTk i ne axes on producion aribued o capial in indusry i, period PKS i producive-capial sock in indusry i, a end of period Boh NTK i and PKS i values are in curren prices. The allocaion of axes and subsidies on producion o capial is elaboraed on in secion 4.6. The remaining erm in he user cos equaion 4.8 is he real rae of reurn required by firms o effecively cover heir financing coss. Inernaional work a he OECD, in which exogenous real raes of reurn were used for capial services measuremen a he oaleconomy level, showed ha in he 18 counries examined, long-run averages of real ineres raes moved beween 3 and 5 percen per year, depending on he counry (OECD, 2009). Work done a Saisics NZ during he developmen of he PIM of capial sock found ha for , he mean long-run real ineres rae (represened by long-erm ineres raes on New Zealand Governmen sock, deflaed by he consumers price index) was 4 percen. This rae is used in he PIM o discoun he composie ageprice funcion o derive he composie age-efficiency funcion. To ensure inernal consisency, he same 4 percen real rae is used in calculaing he user cos of capial. 11 The user cos of capial, and in paricular he ineres rae, could be derived in a number of ways. A good summary of hese differen approaches is provided in Measuring capial: OECD manual (2009). One alernaive adoped by many saisical agencies is o derive he rae of reurn endogenously. Under his approach, capial income is equaed wih he value of capial services (see equaion 4.5) o give: Equaion 4.11 Yk u i j ij PKS ij Where Yk i curren price capial income in indusry i, period. Using he user cos equaion 4.8 and solving for he real rae of reurn gives: i i Yk i Equaion PKS ( p d + p. x ) j j p ij ij ij PKS ij ij ij i While he endogenous approach has some heoreical appeal, and is by definiion selfbalancing, here are a number of drawbacks o his approach: Where asse coverage is incomplee, he rae of reurn will be biased upwards. 11 See MacGibbon (2010) for more deail on Saisics NZ s approach o measuring he user cos of capial. 12 Noe ha equaion 4.12 resuls in a real rae of reurn. Including a capial gains erm on he righ-hand side of he equaion would resul in a nominal rae of reurn. 29

30 Produciviy saisics: Sources and mehods (10h ediion) An endogenous approach assumes perfec foresigh on he par of economic agens (as i is an ex-pos rae). The approach assumes he absence of economic rens. There is increased likelihood of negaive user coss where here is eiher low capial income or significan capial gains. Afer assessing alernaive mehods for calculaing he user coss, o be used in weighing movemens in volumes of asses wihin he indusry capial services indexes, he exogenous real rae (4 percen) was seleced as he mos appropriae for New Zealand circumsances. In a few cases he user cos can be negaive or zero, which leads o compuaional problems. In boh hese cases, a user cos of is used, o ensure a non-zero and non-negaive user cos of capial. In pracice, his is a rare even and he impac of removing his reamen, which is based on pragmaism, has been quanified as negligible. 4.6 Esimaing indusry capial income Capial income is assumed o be he sum of gross operaing surplus (afer he labour income of working proprieors is deduced as defined in equaion 3.3) and ne axes on producion ha are aribuable o capial. Gross operaing surplus and ne axes on producion are sourced from he curren-price indusry producion accouns published in he naional accouns. These producion accouns are available for years in which he indusries are balanced hrough an inpuoupu analysis. 13 Adjused gross operaing surplus (ie gross operaing surplus less he labour income of working proprieors) 14 is given by: Equaion 4.13 AGOS i GOS i Ylwp i Where AGOS i adjused gross operaing surplus in indusry i, period GOS i gross operaing surplus in indusry i, period Ylwp i labour income of working proprieors in indusry i, period. Indusry capial income is herefore calculaed as he sum of adjused gross operaing surplus plus a share of ne axes on producion, which is added so he capial and labour incomes sum o indusry gross value added. This is based on he relaive share of capial income o he sum of labour and capial income. Indusry capial income is esimaed as: Indusry capial income, indusry capial service weighs, and facor income shares are raed forward for non-balanced years. 14 Labour income of working proprieors is sourced from he enrepreneurial income published in he insiuional secor accouns. 15 For he educaion and raining, and healh care and social assisance indusries, capial cos (as defined by equaion 4.5) is used as he facor weigh for capial raher han capial income. This is because, as non-marke dominaed indusries, gross operaing surplus is negligible hus leading o a poenial bias in using income shares. See secion 6 of Saisics New Zealand (2013a). 30

31 Produciviy saisics: Sources and mehods (10h ediion) Equaion 4.14 Yk AGOS + NTk i i i Where NTk i ne axes on producion aribued o capial income in indusry i, period. And: Equaion 4.15 NTk i Where AGOSi NTi COEi + GOSi NT i ne axes on producion in indusry i, period COE i compensaion of paid employees in indusry i, period. Noe: NTk i is used as he numeraor in calculaing he average rae of non-income ax on producion (see equaion 4.12). 4.7 Roading asses are excluded from indusry capial services indexes Of he 33 asses ha are poenially wihin he scope of capial as defined for he produciviy measures (refer appendix B), only 31 conribue o he measured-secor capial services index. The wo excluded asses are cenral governmen roading and local governmen roading. These are owned by ANZSIC published indusries local governmen adminisraion (OO1) and cenral governmen adminisraion, defence, and public safey (OO2), which are ouside he measured secor. No indusry wihin he measured secor bears he cos of using hese asses. They are essenially public goods in he sense ha hey are non-excludable and non-rival in consumpion. The roading nework, like oher public goods, does no represen a capial asse wih an allocable user cos. Raher, he nework enhances he use of oher ranspor equipmen asses and i is absorbed ino he calculaed MFP residual. 4.8 The measured-secor capial services index The capial services index for he measured secor is formed by a Tornqvis aggregaion of he consiuen indusries capial services indexes. The aggregaion weighs are he mean wo-period indusry shares of measured secor curren-price capial income. Equaion 4.16 summarises his aggregaion: Equaion 4.16 K k i i Wki ( ) Where K he measured secor capial services index, period. This is equivalen o: 31

32 Produciviy saisics: Sources and mehods (10h ediion) Equaion 4.17 C C C C ( 1) i i( 1) i Wk i Where C he volume of capial services in he measured secor, period. And: Equaion 4.18 Wk i 1 2 i Yk i Yk i + i Yk i( 1) Yk i( 1) Equaions 4.16 and 4.17 provide an index represening proporionae movemens in measured-secor capial services. Assigning a value of 1000 o he base year volume of capial services, and using hese movemens, produces he capial services index. 32

33 5 Produciviy esimaes 5.1 Inroducion The purpose of his secion is o ouline he mehods for obaining: indexes of labour and capial produciviy an index of mulifacor produciviy (MFP) a decomposiion of oupu growh ino growh in labour volume, growh in capial services, and MFP growh a decomposiion of labour produciviy growh ino capial deepening, and growh in MFP. 5.2 The oupu index For he produciviy measures, oupu is defined as value added a consan prices. Annual value added for he measured secor is derived by following he same procedures used o derive consan-price GDP (a chained Laspeyres volume index of he consanprice value added of he indusries ha comprise he measured secor). The resuling chained volume series are re-expressed as an index wih an expression base of Labour produciviy and capial produciviy Given indexes of labour volume and value added, i is possible o calculae labour produciviy for he measured secor as: Equaion 5.1 V LP L Where LP an index of labour produciviy in period V an index of oupu (value added) in period L an index of labour volume inpu in period. Similarly, a capial produciviy index is found by dividing he oupu index by he capial services index (K ): Equaion 5.2 KP Where V K KP an index of capial produciviy in period K an index of capial services inpu in period. These single facor (or parial) indexes reflec he combined influence of boh primary facors. For example, labour produciviy can be enhanced hrough greaer levels of capial accumulaion. For his reason, any aemp o rack echnological progress or efficiency gains mus explicily conrol for changes in he levels of boh capial and labour. The measuremen of MFP growh aemps o address his issue. 33

34 Produciviy saisics: Sources and mehods (10h ediion) 5.4 Mulifacor produciviy Esimaing MFP requires an index represening oal inpus. This inpu index is calculaed using he volume indexes of labour and capial. Capial and labour are assumed o conribue o oupu according o he convenional Solow producion funcion: Equaion 5.3 Q A f ( K, L ) Where Q real oupu, period A he observed MFP residual, period f(k, L ) a producion funcion capuring inpus of labour and capial volumes. The residual erm represens all facors ha conribue o oupu, oher han he volume of inpus. An increase in is value is commonly referred o as echnical change or efficiency growh. However, i is more accuraely inerpreed as some combinaion of echnological progress, efficiency gain, deviaion from consan reurns o scale, unobserved change in capaciy uilisaion, or deparure from economy-wide long-run equilibrium. The residual also capures he impac of unobserved inpus on producion. Indeed, Abramoviz (1956) refers o MFP as a measure of our ignorance. The producion funcion of equaion 5.3 assumes Hicks neural echnical progress. This means ha efficiency gains are exernal o he wo primary facors of producion, as represened by he residual erm being disembodied from he inerior of he producion funcion equaion. A Hicks neural specificaion implies ha changes in he residual erm do no affec he relaive marginal produciviies of he primary facors. From he OECD (2001b), Saisics NZ has adoped he index number approach in a producion heoreic framework. The growh accouning echnique examines how much of an observed rae of change of an indusry s [or economy s] oupu can be explained by he rae of change of combined inpus. Thus he growh accouning approach evaluaes he MFP growh residually. Consider equaion 5.4: Equaion 5.4 V A I Where A MFP in period I a composie inpu index value in period. This is an index-number approach inerpreaion of MFP in which he residual is calculaed as he raio of he oupu index o he inpu index. Equaion 5.4 is equivalen o: Equaion 5.5 V A I, which is similar o he above producion funcion. The selecion of inpu index is herefore deermined by he assumed underlying producion funcion. In line wih he recommendaion in he OECD manual and following Ausralian Bureau of Saisics pracice, we have adoped a Tornqvis inpu index. 34

35 Produciviy saisics: Sources and mehods (10h ediion) Consrucing he inpu series wih a Tornqvis index is consisen wih he exisence of a ranscendenal logarihmic (ranslog) producion funcion. The ranslog producion funcion is a flexible aggregaor ha does no impose he resricion of consan elasiciy of subsiuion beween he inpus. The Tornqvis index is a symmerical index i aribues equal imporance o he facor inpu weighs of he curren and immediaely previous periods. I is also described as a superlaive index because i is exac o a flexible-form producion funcion. A Tornqvis index is used o aggregae he primary facors o a composie inpu index. I is calculaed as he geomeric mean of he wo-period inpu raios, weighed by he respecive wo-period oupu elasiciies. The oupu elasiciies are unobservable and i is necessary o invoke wo assumpions o proceed. The firs assumpion is of consan reurns-o-scale echnology his imposes he resricion ha, in any one period, he oupu elasiciies sum o uniy. The second assumpion is he exisence of perfecly compeiive markes compeiive inpu markes imply ha he real paymens o capial and labour are equal o heir respecive marginal producs. Wih hese assumpions, i can be shown ha oupu elasiciies are equal o heir respecive facor-income shares. 16 Hence, in he Tornqvis index, he facor weighs are mean wo-period shares of oal-facor income. The inpu index is given in equaion 5.6, while equaions 5.7 and 5.8 specify he facor weighs: Equaion 5.6 I I K K Wk Wl L L ( 1) ( 1) ( 1) Where I a composie inpu index value for he measured secor, period. Equaion 5.7 Wl 1 Yl 2 Yl + Yk + Yl Yl ( 1) ( 1) + Yk ( 1) Equaion 5.8 Wk 1 Yk 2 Yl + Yk + Yl Yk ( 1) ( 1) + Yk ( 1) Where Wl mean wo-period proporion of labour income o oal-facor income WK mean wo-period proporion of capial income o oal-facor income Yl measured-secor labour income, period Yk measured-secor capial income, period. 11 The wo assumpions ensure ha he value of oupu is equal o he sum of he inpu values (ie o he sum of he facor-income shares). This condiion is needed o derive he reurn o capial endogenously. Refer o equaion 4.15, which solves for he rae of reurn, i i, given he level of capial income Yk i. 35

36 Produciviy saisics: Sources and mehods (10h ediion) Measured-secor labour income is he sum of he consiuen indusries esimaed labour incomes. Labour income is calculaed as he sum of compensaion of employees, he labour componen of working-proprieor income, and a proporion of ne axes on producion. Equaion 5.9 Yl Yl i i Where Yl i labour income in indusry i, period. Measured-secor capial income is he sum of each indusry s esimaed capial income. Capial income is calculaed as gross operaing surplus less he labour income of working proprieors and a proporion of ne axes on producion. Equaion 5.10 Yk Yk i i 5.5 Conribuions o growh Growh accouning is he decomposiion of oupu change ino is conribuing facors. For a given producion funcion, changes in oupu are due o changes in he volumes of he primary inpu facors and/or changes in he MFP erm. The growh accouning equaion is consruced by subsiuing equaion 5.6 ino a raio form of equaion 5.5. Equaion 5.11 V V A A I I ( 1) ( 1) ( 1) Equaion 5.12 V V A A K K Wk Wl L ( 1) ( 1) ( 1) ( 1) L In his expression, he growh rae in oupu has been decomposed ino he growh rae of MFP and he growh rae in labour and capial inpus, each weighed by heir relaive shares of oal-facor income. In pracice, he change in value added ha is no explained by he conribuions of labour and capial is aribued residually o MFP. 5.6 Conribuions o labour produciviy The growh accouning framework can also be used o decompose he rae of change of labour produciviy. Change in labour produciviy is measured as he difference beween he rae of change of oupu growh and he rae of change of labour inpu growh. From equaion 5.1, an index of labour produciviy in period was defined as: V LP L In is raio form, we have: 36

37 Produciviy saisics: Sources and mehods (10h ediion) Equaion 5.13 V V L L ( 1) ( 1) A A I I L L ( 1) ( 1) ( 1) A A K K Wk L L Wl ( 1) ( 1) ( 1) ( 1) L L As Wk 1 Wl, we have: A A ( 1) K K ( 1) L L ( 1) Wk Wk A A K L K L ( 1) ( 1) ( 1) Wk Equaion 5.13 shows ha he change in labour produciviy can be decomposed ino he change in MFP and he change in he capial o labour raio, weighed by he capial income share of oal-facor income (see secion 5.7 for explanaion). 5.7 The capial-o-labour raio The capial-o-labour raio is expressed in an index form and is calculaed by dividing he capial services index by he labour volume index: Equaion 5.14 KLr K L Where KLr capial-o-labour raio for he measured secor, period. This raio s ime profile indicaes wheher he measured secor is experiencing capial deepening or shallowing. 5.8 Growh cycle analysis Breaking he series down ino growh cycles allows for more meaningful comparisons beween sub-periods. Year-by-year comparisons can be problemaic, due o issues such as he variaion of capaciy uilisaion over cycles. We invesigaed a number of univariae filers o generae cycles over he series, and ulimaely he Hodrick-Presco filer was seen as he mos appropriae approach (see Saisics NZ, 2007, for more deail abou he mehodology for calculaing growh cycles). The final growh cycles seleced also ook ino accoun he economic evens hroughou he series, which jusified he cycles generaed from he filer. 37

38 6 Uni labour coss 6.1 Inroducion Uni labour coss (ULC) measure he average cos incurred in he employmen of labour per uni of oupu. ULC is calculaed as he raio of labour coss o oupu. Equivalenly, i is he raio of labour coss per hour o labour produciviy (oupu per hour). Tha is, ULC describes he relaionship beween labour compensaion and labour produciviy. In paricular, growh in ULC indicaes ha average labour coss are rising faser han labour produciviy. There are wo ypes of ULC measures. The firs is nominal ULC and reflecs general price increases, as labour coss are expressed in nominal erms. The second is real ULC, which removes he effec of general cos increases. We produce he following ULC esimaes: nominal ULC: level, index, and percenage change real ULC: level, index, and percenage change. These esimaes are available for he following levels of aggregaion: oal economy measured secor measured secor excluding agriculure, foresry, and fishing goods-producing indusries service secor indusries published indusry level. 6.2 Raionale Expanding Saisics NZ labour cos measures We currenly produce average earnings from he QES, compensaion of employees (CoE) from he naional accouns, and he labour cos index (LCI). Each reflecs differen conceps. Measures of average earnings and CoE are designed o measure he average level of wages or labour cos per employee. The LCI is inended o represen employers expendiure on he benefis graned o heir employees as compensaion for an hour of labour. These measures are useful for undersanding developmens in he labour marke, bu hey do no provide a direc undersanding of he growh in labour cos relaive o growh in oupu or changes in inflaion. Nominal ULC measures he average cos of labour per uni of oupu, while real ULC adjuss for inflaion. 6.3 Relaing labour coss o labour produciviy Under he neo-classical economic model of he labour marke, i is assumed ha growh in produciviy leads o growh in business income or in naional income. Any increase in income is disribued o eiher owners of he labour hrough wages, or o owners of capial hrough operaing surplus, wih operaing surplus being an economic equivalen of profi. Many facors wihin he economy can disor he link beween wages and produciviy, such as he inabiliy of workers o move freely beween jobs, or growh in labour coss due o wage bargaining power. 38

39 Produciviy saisics: Sources and mehods (10h ediion) Hence, if here is an increase in average labour coss bu also an equal increase in labour produciviy, here will be no change in he ULC. In conras, in he same scenario, here would be an increase in he LCI, average earnings, and average CoE. 6.4 Mehodology Nominal ULC are he raios of average employee coss o average labour produciviy. Average employee cos is calculaed as: Average employee coss employee labour coss employee hours where employee labour coss CoE + specific axes aribuable o labour. Employee hours refer o hours paid by employees and come from he LVS. Average labour produciviy is calculaed as: GDP kp Average labour produciviy oal hours where GDP kp refers o chain-volume consan price GDP. Toal hours refers o oal hours paid by employees and self-employed; hey come from he LVS. Nominal ULC is calculaed as: average employee coss ULC average labour produciviy Average employee coss in he numeraor concerns employees, bu he average labour produciviy denominaor has a broader scope and concerns employees plus he selfemployed. To make he scope consisen, we ake a pragmaic approach and assume ha average employee coss and average self-employed employmens coss are he same. This approach is used by Canada, Japan, France, Germany, Norway, Sweden, and he Unied Saes. Nominal ULC, by subsiuing in he previous relaionships, becomes: employee labour coss employee hours ULC GDP kp oal hours employee labour coss oal hours employee hours GDP kp oal hours employee labour coss employee hours GDP kp Nominal ULC reflecs general price increases, as labour coss are expressed in nominal erms. Real ULC (RULC) eliminaes he effecs of general price increases by deflaing nominal ULC wih he GDP deflaor. The GDP deflaor is given by: GDP deflaor GDP cp GDP kp 39

40 Produciviy saisics: Sources and mehods (10h ediion) RULC is calculaed as: ULC RULC GDP deflaor oal hours employee labour coss employee hours GDP kp GDP cp GDP kp oal hours employee labour coss employee hours GDP kp GDP kp GDP cp oal hours employee labour coss employee hours GDP cp where GDP cp refers o curren price GDP. 6.5 Daa sources Composiion of labour cos The choice of an appropriae variable for labour cos is dependen on wha is pracical and concepually consisen. Available opions include average earnings from he QES or LEED, and average CoE. CoE has been seleced as he mos appropriae choice as i is consisen wih he naional accouns framework under which oupu or value added GDP is produced. Inernaionally, i is generally recognised as a good proxy o he concepual ideal of oal labour cos, due o similariy of coverage. CoE consiss of wages and salaries, as well as cash and in-kind paymens, and employers social conribuions. CoE is defined in he Sysem of Naional Accouns 1993 (SNA93) as he oal remuneraion, in cash or in kind, payable by an enerprise o an employee in reurn for work done by he laer during he accouning period (p164). Labour coss are calculaed as CoE plus payroll axes minus employmen subsidies. Training and recruimen coss are excluded because hey are no specifically colleced in he AES. This means ha he cos of labour is slighly lower han oal cos o employmen. The impac is ha he level of ULC will be lower. CoE also excludes he labour coss relaed o self-employed people. In he absence of beer daa, an assumpion is made ha he self-employed earn, on average, he same as employees. This assumpion is in line wih inernaional recommendaions and should no affec he overall inerpreabiliy of oal economy ULC. However, in indusries where he proporion of self-employed is high, he resuls should be inerpreed wih some degree of cauion. Figure 2 presens a schemaic view of he ideal coverage of he labour cos componen of ULC. 40

41 Produciviy saisics: Sources and mehods (10h ediion) Figure 2 2 Composiion of labour cos Composiion of labour cos Labour coss Compensaion of employees Oher labourrelaed coss Wages and salaries in cash and in kind Employers social conribuions Wages and salaries in cash Wages and salaries in kind Conribuions by employers o secure social benefis for heir employees Ordinary ime and overime paymens Taxable allowances Commissions, grauiies, and ips Income ax Regular bonuses Regular paymens under profisharing schemes Ad hoc bonuses and incenive paymens Ad hoc paymens under profisharing schemes Benefis such as free or subsidised accommodaion, ravel, food, or moor vehicle. Employers acual social conribuion (o KiwiSaver and superannuaion funds) as well as employers impued social conribuion 1 Provision for severance, erminaion, and redundancy paymens 2 Paymens for illness/maerniy leave 2 ACC levies Ne axes on producion relaed o labour 3 Training cos 4 Recruimen cos 4 1. Some employers provide social benefis hemselves direcly o heir employees ou of heir own resources wihou involving an insurance enerprise ec for he purpose. 2. In pracice, i can be difficul o separae hese paymens from oher salary and wage paymens, in which case hey are grouped wih he laer. 3. Includes fringe benefi ax, employer superannuaion conribuion ax, less payroll subsidies, ax credi on superannuaion, and ax credi for research and developmen. 4. Training and recruimen coss are excluded because hey are no specifically colleced in he AES. Source: Sysem of Naional Accouns

42 Produciviy saisics: Sources and mehods (10h ediion) Toal and employee hours The recommended mehod is o use hours paid, as his more closely reflecs he cos incurred in he employmen of labour. The LVS used in Saisics NZ s produciviy series is based on hours paid and is available for his purpose. The LVS is based on hours paid per week for a given quarer. The LVS uses QES, BDD, HLFS, census, and LEED for daa on couns and hours for employees and he selfemployed across indusries. Employees and he self-employed are separae, disinc groups and have differen characerisics. For example, employees ypically work (on average) fewer labour hours han he self-employed. Unusually high labour hours end o be ascribed o he self-employed. The ULC calculaion raes up employed hours o allow for CoE for self-employed labour. The raing up ends o affec indusries wih relaively higher proporions of he self-employed compared wih employees, eg: agriculure consrucion renal, hiring, and real esae services professional, scienific, and echnical services oher services Gross domesic produc GDP is a frequenly used measure of economic aciviy. I describes, in a single figure and wih no double couning, all oupu or producion carried ou by all enerprises, governmen and non-profi insiuions, and households in New Zealand during any given ime. SNA93 defines he scope of oupu and producion measured in he naional accouns. Aciviies ha are difficul o rack down, such as illegal aciviies and unpaid work, are excluded from he scope of he naional accouns. Consan and curren price GDP daa, based on he producion approach, were used in consrucing ULC. The producion approach calculaes wha each separae producer adds o he value of final oupu (value added) by deducing inermediae consumpion from gross oupu. Value added is summed for all he producers. In pracice, he componens of value added are also independenly esimaed. This is he appropriae measure for ULC as i has a producion by indusry dimension. Ideally, for ULC, GDP should be measured a basic prices. Excluding indusries from he oal economy while measuring oupu a basic prices is challenging. However, he aggregaes required can be readily compiled using chain-volume value added, which uses producer prices Indusry coverage Indusry coverage is based on ANZSIC06 described in able 2. The measured secor is based on he indusries covered by Saisics NZ s produciviy series. I excludes he following indusries: ownership of owner-occupied dwellings local governmen adminisraion 17 The producer s price is he amoun received by he producer from he purchaser for a uni of goods or a service produced, minus any VAT, or similar deducible ax, invoiced o he purchaser. I excludes any ranspor charges invoiced separaely by he producer. The basic price is obained from he producer price by deducing any oher axes while adding any subsidies. 42

43 Produciviy saisics: Sources and mehods (10h ediion) cenral governmen adminisraion, defence, and public safey educaion and raining healh care and social assisance. Uni labour coss canno be readily compiled for indusries wih significan non-marke componens. Real ULC (employee labour coss over curren price GDP) are ofen preferred as hey accoun for boh inflaion and produciviy. However, when curren price GDP is based on he sum of inpu coss (as is he case for non-marke providers), his leads o a raio exhibiing minimal change. This is a saisical arefac raher han a precise represenaion of he labour marke. Nominal ULCs (employee labour coss over consan price GDP) will be biased in level erms as he sum of inpu coss approach o curren price GDP is required for he base year value of consan price GDP. 6.7 Raing forward CoE using LEED To derive esimaes for main aggregaes in pos-balanced years, indusry-level CoE daa are raed forward using he growh in LEED oal gross earnings. The raed-forward CoE value is hen balanced wih he provisional esimae for oal economy CoE. This enables ULC saisics o be imelier. Nominal ULC for he main aggregae (measured secor and measured secor excluding agriculure) are available wih a one-year lag. Goodsproducing and service-secor ULC are available wih a wo-year lag. This is because he indusry-level labour volumes are no available. Real ULC esimaes are available unil he las balanced year of he naional accouns. 43

44 Appendix A: ANZSIC06 divisions and published indusries Table 2 2 Indusry coverage under ANZSIC06 Indusry coverage under ANZSIC06 ANZSIC06 indusry Measured secor (4))(5) Primary secor AAZ Agriculure, foresry, and fishing (1) BB1 Mining Goods-producing secor CCZ Manufacuring (1) DD1 Elecriciy, gas, waer, and wase services EE1 Consrucion Service secor (2) FF1 Wholesale rade GH1 Reail rade GH2 Accommodaion and food services II1 Transpor, posal, and warehousing JJ1 Informaion media and elecommunicaions KK1 Financial and insurance services LL1 Renal, hiring, and real esae services (3) MN1 Professional, scienific, and echnical services (3) MN2 Adminisraive and suppor services (3) RS1 Ars and recreaion services RS2 Oher services (3) Indusries excluded from he measured secor LL2 Owner-occupied propery operaion OO1 Local governmen adminisraion OO2 Cenral governmen adminisraion, defence, and public safey PP1 Educaion and raining (6) QQ1 Healh care and social assisance (6) 1. Series for wo agriculural sub-indusries and nine manufacuring sub-indusries are also available. 2. Service secor differs from naional accouns service secor. 3. No included in he former measured secor. Indusry series available from Former measured secor includes ANZSIC06 indusries AA1 KK1 and RS1. Series available from Measured secor includes ANZSIC06 indusries AA1 MN2, RS1, and RS2. Series available from No included in he measured secor. Indusry series available from Source: Saisics New Zealand 44

45 Appendix B: Capial series Saisics NZ asse classes and ypes Table 3 3 Capial series Saisics NZ asse classes and ypes Capial series Saisics NZ asse classes and ypes Asse class Inangible asses Land improvemen Non-residenial building Oher consrucion Plan, machinery, and equipmen Residenial building Transpor equipmen Land Invenories Asse ype Mineral exploraion Oher exploraion Sofware Land improvemens Transfer coss on land Non-residenial buildings Transfer coss on non-residenial buildings Power consrucion Rail consrucion Transfer coss on consrucion Oher consrucion no elsewhere classified Compuers Elecronic equipmen Elecrical equipmen Furniure General purpose equipmen Heavy machinery Residenial buildings Transfer coss on residenial buildings Aircraf Buses Rail Road vehicles (excluding buses) Ships Agriculural and foresry land Residenial land Commercial land Indusrial land Mining land Oher non-agriculural land Invenories (includes livesock and imber) 45

46 Glossary ANZSIC96 Ausralian and New Zealand Sandard Indusrial Classificaion ANZSIC06 Ausralian and New Zealand Sandard Indusrial Classificaion Average annual growh rae reflecs he average change in a variable across a period of ime. Raes are calculaed as geomeric means, which ake accoun of he compounding of growh raes over ime. Arihmeic averages give higher growh raes and would lead o a differen index figure for he laes year when applied o he base year. Capaciy uilisaion he relaionship beween acual and poenial use of an inpu. Capaciy uilisaion is high when acual oupu is close o poenial oupu because he mos use is being made of labour and capial. In he produciviy measures we produce, i is assumed ha capial and labour are uilised a a consan rae over ime. Capial-o-labour raio he capial inpu index divided by he labour inpu index. Capial deepening posiive growh in he capial-o-labour raio. See also conribuion of capial deepening. Capial income he par of he cos of producing gross domesic produc ha consiss of gross paymens o capial. I represens he value added by capial in producion, and is equivalen o he gross operaing surplus, less he labour income of working proprieors, plus he capial proporion of axes less subsidies on producion. Capial produciviy measured as a raio of oupu o capial inpu. The raio is derived by dividing he index of he chain volume measure of GDP by an index of capial services. Capial services he amoun of 'service' each asse provides during a period. For each asse, he services provided in a period are direcly proporional o he asse's producive capial value in ha ime. As an asse ages and is efficiency declines, so does he producive capial value and he service he asse provides. Capial services is he appropriae measure of capial inpu in producion analysis. Capial shallowing a decline in he capial-o-labour raio. Chain volume measures annually-reweighed chain Laspeyres volume indexes referenced o he curren-price values in a chosen reference year (ie he year when he quarerly chain volume measures sum o he curren-price annual values). Chain Laspeyres volume measures are compiled by linking ogeher (compounding) movemens in volumes, calculaed using he average prices of he previous financial year, and applying he compounded movemens o he curren-price esimaes of he reference year. Compensaion of employees he oal remuneraion, in cash or in kind, payable by an enerprise o an employee in reurn for work done by he employee during he accouning period. I has wo sub-componens: wages and salaries, and employers' social conribuions. Compensaion of employees is no payable for unpaid work underaken volunarily, including he work done by members of a household wihin an unincorporaed enerprise owned by he same household. Compensaion of employees excludes any axes payable by he employer on he wage and salary bill (eg payroll ax, fringe benefis ax). Conribuion of capial deepening he growh in he capial-o-labour raio, weighed by capial s share of oal income. Given ha capial s share of oal income is always less han 100 percen, he conribuion of capial deepening is always less han he growh in capial deepening. I is used for growh accouning for labour produciviy. 46

47 Produciviy saisics: Sources and mehods (10h ediion) Conribuion of capial inpu he growh in he capial inpu index, weighed by capial s share of oal income. Given ha capial s share of oal income is always less han 100 percen, he conribuion of capial inpu is always less han he growh in capial inpu. I is used for growh accouning for oupu. Conribuion of labour inpu he growh in he labour inpu index, weighed by labour s share of oal income. Given ha labour s share of oal income is always less han 100 percen, he conribuion of labour inpu is always less han he growh in labour inpu. I is used for growh accouning for oupu. Gross domesic produc (GDP) he oal marke value of goods and services produced in New Zealand wihin a given period, afer deducing he cos of goods and services used up in he process of producion, bu before deducing allowances for he consumpion of fixed capial. Thus, GDP is 'a marke prices'. I is equivalen o gross naional expendiure plus expors of goods and services less impors of goods and services. Gross mixed income he surplus due o owners of unincorporaed businesses. I is ofen referred o as profi, alhough only a subse of oal coss are subraced from he oupu of unincorporaed businesses o calculae i. Gross mixed income is spli and allocaed o capial and labour as facors of producion. Growh accouning decomposes he growh rae of an indusry s oupu ino he par due o he increase in facors of producion (labour and capial), and he par ha canno be accouned for by changes in labour and capial uilisaion. This residual growh in oupu ha canno be accouned for is known as mulifacor produciviy (ie he exen o which an indusry is geing more oupu from he same amoun of inpus). Growh cycle he span of years beween he peak of one cycle and he peak of a following cycle. Peaks are deermined using saisical echniques, and are chosen o represen high poins in capaciy uilisaion of he economy. Produciviy is bes analysed over growh cycles, as annual movemens can be volaile and don usually represen rue changes o he underlying producion funcion. Index an index series is a simple way of expressing, wih reference o a base ime period, he change in some variable from a given poin in ime o anoher poin in ime. Invenories a class of produced non-financial asses consising of: socks of oupus ha are sill held by he unis ha produced hem before being furher processed, sold, delivered o oher unis, or used in any oher ways; and socks of producs acquired from oher unis ha are inended o be used for inermediae consumpion or for resale wihou furher processing. Labour income he par of he cos of producing GDP ha consiss of gross paymens o labour. I represens he value added by labour in producion, and is equivalen o compensaion of employees, plus he labour income of working proprieors, plus he labour proporion of axes, minus subsidies on producion. Labour inpu index an index of he weighed number of hours paid in he measured secor. I is creaed by weighing ogeher he indusry-level labour volume series using labour income weighs. Labour produciviy is measured as a raio of oupu o labour inpu. Labour produciviy esimaes are indexes of real GDP per hour paid. Labour produciviy reflecs he conribuion of labour o changes in produc per labour uni, bu is also influenced by he conribuion of capial and oher facors affecing producion. Labour volume series (LVS) an esimae of he oal number of hours paid in paid employmen per week for he whole economy or a given indusry over ime. 47

48 Produciviy saisics: Sources and mehods (10h ediion) Measured secor he indusry coverage of produciviy saisics is defined as he measured secor. I consiss of indusries for which esimaes of inpus and oupus are independenly derived in volume erms. Indusries for which real value added in he naional accouns is largely measured using inpu mehods, such as number of employees, are excluded. These are mainly governmen non-marke indusries ha provide services (eg adminisraion, healh, and educaion) free or a nominal charges. Mulifacor produciviy (MFP) esimaes are indexes of real GDP per combined uni of labour and capial. They are derived by dividing chain-volume esimaes of marke secor GDP by a combined measure of hours paid and capial services. An increase in MFP is referred o as echnical change or efficiency growh. However, i is more accuraely inerpreed as a combinaion of echnological progress, efficiency gain, deviaion from consan reurns o scale, unobserved change in capaciy uilisaion, or deparure from economy-wide long-run equilibrium. MFP is essenially a residual, and so also capures he impac of unobserved inpus on producion. Oupu chain-volume value added. Annual value added for he measured secor is derived following he same procedures used o derive consan-price GDP (ie as a chained Laspeyres volume index of he consan-price value added of he indusries making up he measured secor). The resuling chained volume series are re-expressed as an index wih an expression base of 1000 in he March 1978 year. Producive capial sock a measure of producive capaciy ha forms he basis for he measure of capial services. Producive capial sock esimaes are derived as he wriendown value of each asse as is efficiency declines, due o age. This sock is measured in unis of sandardised efficiency. Renal prices also referred o as he user cos of capial. I is he uni cos for he use of an asse for one period. Tha is, he price for employing or obaining one uni of capial services. The renal price of an asse is deermined by is price index when new, is rae of economic depreciaion, he average ax rae on producion wihin he indusry, and an exogenous real rae of reurn (se a 4 percen). Toal income he par of he cos of producing GDP ha consiss of gross paymens o facors of producion (labour and capial). I represens he value added by hese facors in he process of producion and is equivalen o curren-price GDP. Uni labour coss (ULC) he raio of labour coss o oupu. I is also he raio of labour coss per uni of oupu o labour produciviy (oupu per hour). There are wo ypes: he firs is nominal ULC, as labour coss are expressed in nominal erms; and he second is real ULC, which removes he effec of general cos increases. 48

49 References and furher reading References Abramoviz, M (1956). Resource and oupu rends in he Unied Saes since American Economic Review, 46, pp5 23. MacGibbon, N (2010). Exogenous versus endogenous raes of reurn: The user cos of capial in Saisics New Zealand's mulifacor produciviy measures (Saisics New Zealand Working Paper No 10 03). Available from Organisaion for Economic Co-operaion and Developmen (2001a). OECD manual: Measuring capial; measuremen of capial socks, consumpion of fixed capial and capial. Paris, France: OECD. Organisaion for Economic Co-operaion and Developmen (2001b). OECD manual: Measuring produciviy; measuremen of aggregae and indusry-level produciviy growh. Paris, France: OECD. Organisaion for Economic Co-operaion and Developmen (2009). Measuring capial: OECD manual (second ediion). Available from Saisics New Zealand (2007). Exracing growh cycles from produciviy indexes. Available from Saisics New Zealand (2008). Accouning for changes in labour composiion in he measuremen of labour produciviy. Available from Saisics New Zealand (2013a). Educaion and healh indusry produciviy: Available from Saisics New Zealand (2013b). Measuring capial sock in he New Zealand economy. Available from Furher reading Allen, RC, & Diewer, WE (1993). Direc versus implici superlaive index number formulae. In WE Diewer & AO Nakamura (Eds), Essays in index heory: Vol 1. Elsevier Science Publishers BV. Aspden, C (1990). Esimaes of mulifacor produciviy, Ausralia (Occasional paper, caalogue No ). Canberra: Ausralian Bureau of Saisics. Ausralian Bureau of Saisics (2000). Produciviy measures. Ausralian naional accouns: Conceps, sources and mehods, chaper 27. Canberra: Ausralian Bureau of Saisics. Black, M, Guy, M, & McLellan, N (2003). Produciviy in New Zealand 1988 o 2002 (New Zealand Treasury working paper 03/06). Wellingon: The Treasury. Available from Diewer, E, & Lawrence, D (1999). Measuring New Zealand s produciviy (New Zealand Treasury working paper 99/5). Wellingon: The Treasury. Available from Diewer, WE, & Nakamura, AO (2003). Index number conceps, measures and decomposiions of produciviy growh. Journal of Produciviy Analysis, 19, pp

50 Produciviy saisics: Sources and mehods (10h ediion) Hicks, JR (1932). The heory of wages. London, UK: Macmillan. Hulen, CR, Dean, ER, & Harper, MJ (Eds) (2001). New developmens in produciviy analysis (Naional Bureau of Economic Research Sudies in Income and Wealh, 63). Chicago: The Universiy of Chicago Press. Available from Iner-Secrearia Working Group on Naional Accouns (1993). Sysem of Naional Accouns Available from hp://unsas.un.org. Jorgenson, DW, & Griliches, Z (1967). The explanaion of produciviy change. Review of Economic Sudies, 34, pp Mawson, P, Carlaw, KI, & McLellan, N (2003). Produciviy measuremen: Alernaive approaches and esimaes (New Zealand Treasury working paper 03/12). Wellingon: The Treasury. Available from McLellan, N (2004). Measuring produciviy using he index number approach: An inroducion (New Zealand Treasury working paper 04/05). Wellingon: The Treasury. Available from Organisaion for Economic Co-operaion and Developmen (2003). OECD capial services esimaes: Mehodology and a firs se of resuls (OECD Saisics Working Paper 2003/6). Paris, France Schreyer, P (2004, July). Measuring mulifacor produciviy when raes of reurn are exogenous. Paper presened a he SSHRC Inernaional Conference on Index Number Theory and he Measuremen of Prices and Produciviy, Vancouver. Schreyer, P, Diewer, WE, & Harper, A (2005, July). Cos of capial services and he Naional Accouns, Updae of he 1993 SNA. Issue No. 15, Issues paper for he July 2005 AEG Meeing. Solow, RM (1957). Technical change and he aggregae producion funcion. Review of Economics and Saisics, 39, pp