MEASURING EDUCATION SERVICES AS INTANGIBLE SOCIAL INFRASTRUCTURE

Transcription

1 Working Paper Series No. 19 MEASURING EDUCATION SERVICES AS INTANGIBLE SOCIAL INFRASTRUCTURE Carol Corrado Mary O Mahony Lea Samek SPINTAN Projec: Smar Public inangibles. This projec has received funding from he European Union s Sevenh Framework Programme for research, echnological developmen and demonsraion under gran agreemen no:

2 Spinan working papers offer in advance he resuls of economic research under way in order o disseminae he oupus of he projec. Spinan s decision o publish his working paper does no imply any responsibiliy for is conen. Working papers can be downloaded free of charge from he Spinan websie hp:// Version: November 2016 Published by: Insiuo Valenciano de Invesigaciones Económicas, S.A. C/ Guardia Civil, 22 esc. 2 1º Valencia (Spain) DOI: hp://dx.medra.org/ /spintan-wp-19

3 SPINTAN Working Paper Series No. 19 MEASURING EDUCATION SERVICES AS INTANGIBLE SOCIAL INFRASTRUCTURE* Carol Corrado Mary O Mahony Lea Samek** Absrac The saring poin for his paper is ha sociey's consumpion of educaion services is he acquisiion of schooling knowledge asses whose change in value should be included in saving and ne invesmen. We esimae he nominal value of educaion services produced by he public secor by using he Jorgenson- Fraumeni lifeime income approach. Enrolmens by educaion ype are muliplied by he amoun by which lifeime earnings a ha ag sex, and educaion change wih addiional qualificaions aking accoun of he exra ime required o achieve ha addiional educaion. Implemening his approach requires a number of assumpions on esimaing wages ne of experienc aking accoun of inernaional sudens who pay for he cos of heir uiion, survival raes, he discoun rae and deflaors. The model is esimaed using daa for he UK under a range of assumpions. The raio of our preferred measure o educaion expendiures is jus under hre suggesing ha sociey obains a very high economic benefi from educaion. * This research has received funding from he European Union s Sevenh Framework Programme for research, echnological developmen and demonsraion under gran agreemen no: (SPINTAN Projec: Smar Public Inangibles). ** Carol Corrado: The Conference Board; Mary O Mahony: King s College London); Lea Samek (NIESR and King s College London.

4 1. Inroducion. The public secor produces services such as educaion and healh ha can be viewed as inangible social infrasrucure which add o invesmen, savings and wealh. Typically his is no included wihin he naional accouns framework. These services provide benefis o sociey in many forms including increasing he produciviy of workers as well as social gains such as arguably conribuing o sable democracies. This paper considers he worker produciviy aspec of educaion services hrough using he lifeime income approach pu forward by Jorgenson and Faumeni (1989, 1992a, 1992b). The nex secion ses ou he concepual framework for modeling educaion services as social infrasrucure. Secion hree ses ou he lifeime income model. We hen apply his approach o daa for he UK over he period 2002 o Secion 4 discusses he daa used and secion 5 presens resuls. 2. Framework: Educaion as Social Infrasrucure. Many sudies show ha reurns o educaion accrue o privae individuals in he form of higher wages raher han as paybacks o producers of educaion services. A fundamenal feaure of he educaional process as modelled, e.g., by Jorgenson and Fraumeni (1989; 1992a; 1992b), is he lenghy gesaion period beween he applicaion of he educaional inpus (mainly he services of eachers and he ime of heir sudens) and he emergence of human capial embodied in graduaes of educaional insiuions. In he Jorgenson and Fraumeni framework, he household invess ime and money via purchases of eacher services (eiher a cos for public insiuions in naional accouns or acual oulays in he case of privae services) o build human capial. Household producion is ou of scope for GDP as radiionally defined, and he JF approach o modelling human capial producion and invesmen is usually considered relevan for building a human capial saellie accoun and no necessarily relevan as an approach for measuring educaional oupu in headline GDP. In his paper we reconsider he uiliy of he JF approach for measuring educaional oupu. Our approach begins wih he view ha he service capaciy of a naion s educaion sysem is, in effec, social infrasrucure. In his view, spending by educaional insiuions o improve he capaciy of he educaional sysem o deliver improved eacher services would be inside he asse boundary of GDP, i.e., such spending would be considered an inangible invesmen as in Corrado, Hulen, and Sichel (2005, 2009). In oher words, a school sysem's 1

5 expendiures on eacher raining is an invesmen if i increases he effeciveness of he sysem o deliver educaional services in fuure periods. 1 Bu wha abou he oupu of educaional insiuions? If an educaion sysem plays a par in producing human capial, we need a framework ha views he producion of educaion services as he producion of a socieal asse as opposed o regarding educaion services as an inpu o he producion of human capial wihin households. The basic idea is ha sociey's consumpion of educaion services is in fac he acquisiion of schooling knowledge asses, ΔE, whose change in value P ES ΔE should be included in saving and wealh even hough i is no used in curren producion (or consumed). Raher, he asses are held in invenory, wihin he school sysem, unil sudens graduae and ener he working age populaion, afer which he value is unchanged (by he shool sysem). 2 In his view, he real oupu of an educaion sysem, Q ES is he knowledge sock of his year's graduaes plus he incremen o knowledge held by sudens sill wihin he sysem, or Q ES = E Grads +ΔE InSchool. Under cerain assumpions, his implies Q ES ΔE because a any poin in ime he value of las year's graduaes is unchanged (and enrans a he lowes level are assumed o have a zero sock). The producion funcion F E for educaion services is hen given by: E (1) F K L which implies E Q, ES, ES,, ES F E (2), ES,, ES 1 E K L where E 1 is he beginning of period knowledge socks held by his year's sudens, and educaion services producion is he schooling produced incremen o hose socks. There is no depreciaion of schooling produced knowledge socks while sudens are enrolled in school. K ES and L ES are he educaion sysem's fixed capial and labor services inpus. 1 This expanded view of invesmen by educaional insiuions has been implemened in he daabase produced by he SPINTAN projec, which covers 22 EU counries, he Unied Saes, Brazil, and China. See for furher deails. 2 Noe ha his invenory view follows he logic of Ruggle's approach o accouning for consumer durables (Ruggles, 1983; see also Moulon, 2001) and he SNA's approach o he reamen of valuables. 2

6 These simple accouning relaionships are direcly relaed o he JF lifeime income approach o human capial measuremen. Some observers have suggesed ha he JF marke componen of human capial producion be used o replace he exising measures of educaion services in convenional GDP (e.g., Ervik, Holmoy, and Haegeland, 2003). Our invenory" approach is a differen adapaion of he JF model for inclusion in convenional accouns. Like he JF work, however, and as discussed in Chrisian (2014), our approach includes values, volumes, and prices as basic elemens, and in ha capaciy embraces human capial wihin he convenional boundary of he SNA. Mincer's seminal conribuion (Mincer, 1974) mapped he heory of invesmens in human capial o he empirical lieraure on he reurns o schooling. According o Mincer's model, a he end of each period of schooling, individuals (a) have a level of human capial consisen wih ha level of schooling, and (b) choose he opimal level of schooling (i.e., years in school) up o he poin ha he opporuniy cos of one more year of schooling equals foregone earnings. This implies an individual's reurn o schooling mus be commensurae wih hese foregone earnings. The Mincer framework underpins he lifeime income approach of Jorgenson Fraumeni which is discussed in he nex secion. Afer recogniion of schooling produced knowledge asses, real invesmen in naional accouns includes he ne acquisiion of knowledge capial held wihin he educaion sysem ΔE, which is equivalen o he real gross oupu of he educaion sysem. Invesmens in schooling produced knowledge asses end o be a funcion of he age srucure of a sociey, and hus a relaively sable fracion of GDP in mos advanced counries, suggesing ha he implicaions of capializing educaion as social infrasrucure for real GDP and produciviy change will largely depend on rends in he implied price index for educaion services. Nowihsanding, recogniion of schooling asses as socieal wealh packs an exra punch for ne saving and real ne expendiures (relaive o real GDP, ha is) due o he fac ha in moving from GDP o real ne expendiures, no depreciaion charge is aken. 3. The Jorgenson Fraumeni framework This secion suggess a mehod of inegraing he Jorgenson Fraumeni (1989, 1992a,b) lifeime income approach o measuring human capial wih he reamen of educaion as social infrasrucure as argued above. The Jorgenson Fraumeni framework is se ou below, followed by a discussion of concepual issues ha arise when using he framework o esimae he value of a sociey s invesmens in educaion. 3

7 3.1 The Jorgenson Fraumeni (JF) framework Lifeime income We begin by absracing from non marke aciviies, employmen oucomes and labour force dropous and simply assume ha any suden enrolled in school will, in he following year if hey leave educaion, earn he marke wage corresponding o ha level of educaion. The JF framework calculaes he values of human capial socks based on lifeime incomes by sex (s), age (a) and educaion level (e). Their original papers calculae his for all persons in he populaion. A more common approach is o calculae he sock only for he working populaion, e.g. Gu and Wong (2010), Wei (2004). Le: pop = populaion y = curren marke income li = lifeime income δ = he discoun rae g = average income growh senr = he enrolmen rae sr = he survival rae. The JF framework calculaes lifeime income by s, a and e for essenially wo groups. Assume no one of age 35 and above is enrolled in educaion. The firs group, for hose aged 35 and over, is he mos sraighforward. The simples assumpion is o say ha lifeime income is 0 beyond some ag say 80. For hose aged 80, lifeime income (li) in year is jus curren labour income. (3) li s,a80, y s,a80, For hose aged 79 lifeime income is curren labour marke income plus discouned fuure income of hose aged 80 wih he same educaion and gender, condiional on survival: (4) li s,a79, y s,a79, sr s,a80, 1 g 1 y s,a80, 4

8 In general he lifeime income of hose aged 35+ is given by: 1 g 1 (5) li y sr li a 35 s, a, s, a, s, a1, s, a1, This valuaion for individual i a ime is he value of curren income plus he income of hose one year older of he same ag sex and educaional aainmen imes growh in income discouned o he presen, plus he income of hose wo years older and so on up o age 80. I herefore assumes ha he bes esimae of a person's income nex year is ha earned his year by a similar person who is one year older. The naure of he income growh erm, g, is discussed furher below. For persons aged beween 5 and 34, lifeime income akes accoun of if hey are enrolled in educaion or no. For hese age groups: (6) li s,a, y s,a, 1 g sr s,a1, 1 senr s,a,li s,a1,e1, (1 senr s,a, )li s,a1, 5 a 35 Thus, if a person aged a is enrolled in educaion level heir lifeime income depends on ha for a person one year older wih level e+1. If he same individual is no enrolled in educaion heir lifeime income depends on ha for an individual one year older wih educaion level e. Finally lifeime income for hose aged 0 o 4 is calculaed he same way as for hose aged 35 and over excep ha earnings are zero and educaion is se a he lowes level. Value of human capial. The oal value of he human capial sock in year can be calculaed by summing he lifeime earnings by s, a and e: (7) HC pops, a, lis, a, s a e Noe if he working populaion is used as he weighing facor in (5) hen hose enrolled in compulsory educaion (usually aged 5 15) no longer feaure. This is an issue for calculaing he oupu of he educaion secor as discussed below. Chrisian (2010) defines ne invesmen in human capial (NIH) as he effec of changes from year o year in he size and disribuion of populaions. This is given by: (8) NIH li, ( pops, a, 1 pops, a, ) s a e s, a, e 5

9 This in urn can be broken down ino various componens such as birhs, deahs, ne invesmen from educaion of persons enrolled in school and depreciaion and aging of persons no enrolled in school. In measuring he nominal value of educaion as social infrasrucure we concenrae on he porion of he populaion enrolled in educaion. The erm corresponding o hose enrolled in school is herefore given by: (9) 1 g NIH ( enr) ( enrs, a, 1 enrs, a, ) srs, a1, lis, a1, e*, s a e 1 where enr are school enrolmens, and li s, a1, e*, [ s, a1, s, a1, e1, s, a1, s, a1, senr li (1 senr ) li ] via equaion (6), as hese persons are enrolled in educaion heir curren marke income is zero. The value of educaional services (VES) can be esimaed by rewriing equaion (9) as (Chrisian, 2010): (10) VES enr li li ) s a e s, a, e( s, a1, e1, s, a, Enrolmens are muliplied by he amoun by which lifeime earnings a ha ag sex, and educaion change wih he addiion of one exra year of educaion and he one exra year of age required o achieve ha addiional educaion. 3.2 Valuing ne Invesmen in human capial for persons enrolled in educaion. There are a number of issues o resolve in order o value equaion (7). These include he aribuion of lifeime earnings o educaion, he naure of he income growh erm g and he survival probabiliies sr. Aribuion Wha is he income of a person one year older wih he same educaion level capuring? In Mincer's canonical wage equaion, in which individual j's wage is a reurn o human capial, here are wo key erms, one a reurn o schooling and he oher a reurn o work experienc suggesing HC j = E j +LX j where HC j is individual j's oal human capial and LX j is he porion acquired hrough work, i.e., labor marke, experience. From he poin of view of he schooling sysem, his suggess schooling produced knowledge asses can be defined as he presen discouned value of expeced wages of graduaes upon enry o he labor marke, i.e., when he reurn o experience is virually nil. Then he income sream arising from educaion services should be consan a he graduaion earnings hrough ime. In ha 6

10 case he lifeime income sream only depends on how long he person is in he workforce afer graduaion. The oher exreme is o assume ha all fuure labour income is aribuable o he level of educaional aainmen of he individual. This amouns o using he full JF calculaion however, in our conex i is difficul o jusify his assumpion. (This assumpion is embedded in previous work such as by Chrisian (2010) and Gu and Wong, 2010). A pracical soluion migh be o derive he wages on graduaion as a T year average from he poin of graduaion. This could be jusified by assuming some degree of asymmeric informaion whereby firms do no pay he full marginal produc immediaely in case he worker urns ou o be a lemon. T could be se a say 3 years. Anoher approach is o use Mincer regressions, conrolling for oher influences such as experience his was he mehod used by O Mahony and Sevens (2009) and O Mahony e al. (2012). This mehod also allows for direc modelling of he probabiliy of employmen. However his mehod also leads o difficul economeric issues, mosly relaing o idenifying he difference beween age and experience. This mehod is no pursued furher in his paper. The calculaions should also ake accoun of he opporuniy coss of saying in educaion beyond he age of compulsory educaion. However hese foregone earnings are likely o be small relaive o lifeime earnings. Finally we need o ake accoun of foreign sudens. Survival raes If we concenrae on he working populaion hen sr akes accoun of boh moraliy and reiremen. These in urn can be calculaed using life ables and age specific reiremen raes. Arguably survival raes should also depend on he probabiliy ha a person is employed (and no unemployed or exied he labour force). If we ignore employmen probabiliies we are esimaing he poenial human capial only adjusing for permanen exis such as deah and end of working life reiremen. This would be equivalen o ignoring uilisaion raes for physical capial. We deal wih his by muliplying curren income by employmen raes, as is sandard in calculaions of human capial sock for he working populaion (Jones and Fender, 2010). Growh in income and he discoun rae Consrucing values for equaion (10) requires assumpion abou he growh in income (g) and he discoun rae (δ). A relevan quesion in our conex is, does he g ha deermines income growh include produciviy and/or inflaion gains. In oher words, are nominal holding period gains o schooling par of he value of human capial? I seems ha somehing of he sor mus be here if g is, say 2 or 3 percen as in he human capial measuremen lieraur and hus par of he nominal 7

11 change in human capial may be in fac be a holding period (i.e., capial) gain in a naional accouning sens e.g., as in he oal change in he value of schooling produces asses is given by (11) where from before P ES E is he acquisiion value of schooling produced human capial, and he second erm on he RHS is he holding gain (where oher changes in volume and higher order erms are ignored). Looking a his equaion makes i abundanly clear ha he value of school sysem producion is he firs erm on he RHS. The second erm is no included as per he usual exclusion of asse valuaion changes from GDP. In his case i makes sense o se g=0 if, as argued abov changes in individual s income afer graduaion mosly reflec experience and raining which again suggess a zero value for g. On he oher if educaion effeciveness needs ime o maur especially perhaps for universiy graduaes, and i is hough desirable o ake a T year average as discussed abov hen seing g>0 is likely necessary. In he esimaes below we se g equal o 1%, which is half he usual assumpion employed in Human capial sock calculaions (Jones and Fender,2010; Chrisian (2010); Gu and Wong, 2010). In addiion we need o assume a value for δ. In he JF framework his is he annual discoun rae o consruc he presen value of he fuure income sream bu is no discussed in any deail in ha lieraure. Theoreically, his should be a rae of ime preferenc which in his case would be a social rae. An empirical sraegy for esimaing he social rae of ime preference (SRTP) for a counry is se ou in he OECD capial manual; updaed SRTP esimaes for each SPINTAN counry are repored in Corrado and Jaeger (2015). Based on he laer work, in his paper we se δ equal o 2%, again lower han commonly assumed. Educaion progression The UK daa are available by ype of qualificaion raher han years of educaion, divided ino 4 groups GCSE or equivalens (he ypical exam qualificaion aained usually a age 16), A level or equivalens (he ypical exam qualificaion for hose who say on a school, usually aained a age 18), furher educaion (FE pos secondary bu below eriary, ypically vocaional qualificaions ha can eiher be a follow on from GCSE or someimes from A level) and Higher Educaion (HE eriary educaion leading o degrees or equivalens). This means ha assumpions need o be made o implemen equaion (10) in regard o progression across differen ypes of qualificaions. We aggregae all sudens up o age 16 and compare heir li wih he li of someone aged 17 who has an A level. FE are compared wih GCSE for hose aged up o 18 and wih A levels for older sudens. HE is compared o A level raher han FE as mos sudens go o Universiy following A levels raher han progression via FE qualificaions). 8

12 Foreign Sudens The knowledge asses of graduaes exiing he counry needs o be excluded in his calculaion if he probabilisic full resource cos of he annual educaion of foreign sudens is charged o hem (i.e. heir charges reflec he coss of heir educaion discouned by he probabiliy hey ener he domesic labor force). In his way P ES reains is inerpreaion as he domesic price of schooling produced domesic knowledge asses because he cos incurred in producing a foreign graduae is fully offse in revenues, which are subracions from nonmarke producion values esimaed on he basis of producion coss. Deflaors These calculaions are in nominal values. Real educaion oupu can be esimaed as weighed enrolmens, wih weighs equal o he presen value of he lifeime reurn o an addiional year in educaion. For example Gu and Wong (2010) esimae a volume index of educaion oupu as: (12) lnq lnq 1 v lnenr s,a, lnenr s,a,1 s,a,e Where v is he share of individuals wih s, a in he oal value of invesmen in educaion, averaged over year 1 and. The price index of educaion services (P ES ) can hen be esimaed by dividing he nominal value of educaion services by he volume index of educaion services. Chrisian (2012) also discusses he alernaive of measuring real ne invesmen in educaion by deflaing nominal ne invesmen in educaion by he consumer price index. This he erms an oucome based measure as i capures he amoun of goods and services ha could be consumed by he educaion services raher han he amoun produced, i.e., i capures he opporuniy cos of foregoing curren consumpion for invesmens in schooling. A hird alernaive is o divide P ES E by he number of school sysem graduaes in he workforce (aged < 35). Ineresingly Gu and Wong cie Diewer (2008) as showing ha valuing oupu a average coss in measuring oupu and produciviy growh is a second bes opion while he bes opion would be o use final demand prices o value oupu. The use of he final demand prices should correspond o he [lifeime] income based approach in he conex of educaion services. 9

13 Educaion services and educaion expendiures Wha is he relaionship beween he nominal value of invesmen given by equaion (10) and expendiures on educaion as currenly measured in naional accouns (i.e., educaion coss)? I could be a measure of rae of reurn, or effeciveness of he school sysem, i.e. (13) where γ ( 1+rae of reurn) can be equal o, greaer han, or less han one. We usually hink of effeciveness as a correcion for qualiy, bu here i is more like a rae of reurn. If γ is greaer han 1 i can be inerpreed as a measure of socieies reurn from invesing in educaion and compared o reurns from invesmen in oher asses. If γ is less han on hen one could say ha here is a penaly exaced from sociey due o he resources of he school sysem no being used effecively or, pu differenly, due o he labour marke no using schooling produced human capial effecively (i.e., when here is long erm unemploymen). The poenial policy relevance of γ suggess ha he assumpions used o derive equaion (10) he reamen of employmen probabiliies, he use of a T period average for wages, and choice of discoun rae need o be concepually valid and empirically well undersood. Given he large number of assumpions required γ is bes compared over ime or across counries raher han puing oo much weigh on is absolue value. Noe furher ha if he LHS of equaion (13) replaces educaion expendiures in inangibles augmened growh accouns, he conribuion of he educaion services secor o produciviy growh is boosed (or diminished) direcly by γ. 4. Daa sources We use sandard daa sources o carry ou he compuaions described above for he UK. These were: The Labour Force Surveys and Annual Populaion Survey for earnings, populaion and employmen raes by gender, ag and qualificaion. Enrolmen raes from Educaion saisics, his uses boh published daa and unpublished abulaions from HESA for foreign sudens Life Tables for survival probabiliies Educaion expendiures from COFOG ables. 10

14 We exclude enrolmens of par ime sudens in FE and hose aged greaer han 21 as hese sudens are likely o be aking courses ha are paid for by he individuals hemselves or by heir employers. In he case of foreign sudens we disinguish beween EU and non EU sudens only he laer are considered foreign. Below we show a varian where we exclude all foreign sudens by his definiion. This will underesimae educaion services o he exen ha some of hese sudens remain and work in he UK pos graduaion. Agains his some EU naionals do reurn o heir naive counries. There are no reliable daa available on foreign naionals working in he UK cross classified by if hey were educaed in he UK or abroad. 5. Resuls I is useful firs o look a enrolmen raes o ge an idea of he composiion of he UK educaion secor. Char 1a shows he oal numbers and he division by hree groups, school, furher educaion (FE) and higher educaion (HE). School is by far he larges group, reflecing ha pupils ypically spend years in his form of educaion whereas hey spend only 3 4 years in higher educaion and abou wo years in FE. Char 1b shows he growh raes, indexed a 2002=100. This shows a sligh upward rend in aggregae. This is he resul of wo opposing rends generally downward rend in school enrolmens a leas o 2012 and increases in boh FE and HE. The laer shows a dip in 2014 as a consequence of he inroducion of full cos fees for mos universiy programmes. FE is much more volaile and suggess a financial crisis impac wih high growh raes during he crisis period and some fall off afer ha. 11

15 Char 1. Enrolmens in UK Educaion, a. Numbers enrolled School FE HE Toal b. Index 2002= School FE HE Toal

16 Char 2 shows he growh in foreign compared o domesic HE sudens in he period under sudy. This illusraes ha much of he growh in his secor in recen years has been in he inernaional marke wih foreign sudens in 2014 comprising nearly 20% of he suden populaion, from 13% in Char 2. Domesic and Inernaional Sudens in Higher Educaion, UK, domesic 140 foreign Table 1 shows he resuls for 2013 under a number of scenarios, boh he nominal value of educaion services and he raio of ha value o nominal expendiures on educaion. The firs row shows he resuls when here are no adjusmens for aribuion. This sugges a high raio of educaion oupus o expendiures in he UK, and higher for similar exercises for he US where he raio is abou 3 (Chrisian, 2014). When we accoun for aribuion however, he nominal values decline by 30%. Similarly, removing foreign sudens reduces his by abou 15%. Taken ogeher he wo adjusmens lead o nominal values of educaion services ha are abou 60% of he unadjused values. Wih hese adjusmens he rae of reurn from educaional services goes down o 170% which is sill very high. Therefore on his measure sociey obains a very high economic benefi from educaion. 13

17 Value of educaional services Raio o Expendiures. A. Baseline: (including employmen propensiy) 368, B. A + adjusmen for aribuion 258, C. A+ adjusmen o remove foreign sudens 317, D. A+ adjusmens for aribuion & removing foreign sudens 218, E.Baseline wih: g=0.02, d= , Finally in his secion we presen ime series for he raio of oupus o expendiur shown in char 3 his uses he figures adjused for boh aribuion and he removal of foreign sudens. Here he resuls are no so sanguine as hey show a downward rend. Underlying his is he reducion in school enrolmens which coincided wih an increase in expendiures in ha secor. Char 3. Raio of he value of educaional services o expendiures, UK

18 6. Conclusions Using a lifeime income framework his paper esimaed values for educaion services ha far exceed expendiures for he UK in 2013, alhough here is some suggesion ha he raio of educaion services values o expendiures has been declining over he pas decade or so, largely due o declining enrolmens in schools coinciding wih increased expendiure. Ideally we would wan o use separae deflaors for oupu and spending o consider real raios. This is he nex sep in he analysis. I would also be useful o compare wih oher counries. 15

19 References Chrisian, Michael (2010), Human Capial Accouning in he Unied Saes, Survey of Curren Business 90(6), pp Chrisian, Michael (2012), Human Capial Accouning in he Unied Saes: Conex, Measuremen, and Applicaion, paper presened o he CRIW conferenc Boson. Corrado, Carol, Jonahan Haskel and Cecilia Jona Lasinio (2015). Public Inangibles: The Public Secor and Economic Growh in he SNA, paper presened a IARIW workshop, Paris France (April). Available a: hps:// Corrado, Carol and Kirsen Jaeger (2015), The Social Rae of Time Preference as he Reurn on Public Asses, SPINTAN deliverable D1.6 (May). Diewer, Erwin (2008), The Measuremen of Nonmarke Secor Oupus and Inpus Using Cos Weighs. Discussion Paper 08 03, Deparmen of Economics, Universiy of Briish Columbia, Vancouver, B.C. Canada. Jones, R. and V. Fender (2010), Human Capial Esimaes, 2010, Office for Naional Saisics, UK. Jorgenson, D. W. and B. M. Fraumeni (1989). The accumulaion of human and nonhuman capial, In R. Lipsey and H. Tice (Eds.), The measuremen of saving, invesmen, and wealh, Volume 52 of NBER Conference on Research on Income and Wealh, pp Chicago: Universiy of Chicago Press. Jorgenson, D. W. and B. M. Fraumeni (1992a). Invesmen in educaion U.S. economic growh. Scandinavian Journal of Economics 94 (supplemen), Jorgenson, D. W. and B. M. Fraumeni (1992b). The oupu of he educaion secor. In Z. Griliches (Ed.), Oupu measuremen in he service secors, Number 56 in NBER Sudies in Income and Wealh, pp Chicago: Universiy of Chicago Press. Gu, Wulong and Ambrose Wong (2010), Invesmen in Human Capial and he Oupu of he Educaion Secor in Canada, Paper presened o he IARIW conferenc S Gallen. O Mahony, Mary. and Philip A. Sevens (2009). Oupu and Produciviy Growh in he Educaion Secor: Comparisons for he US and UK, Journal of Produciviy Analysis, 31: O Mahony, Mary, José Manuel Pasor, Fei Peng, Lorenzo Serrano and Laura Hernández (2012), Oupu growh in he pos compulsory educaion secor: he European experience, INDICSER discussion paper No