THE CALCULATION OF CYCLICALLY ADJUSTED BALANCES AT BANCO DE PORTUGAL: AN UPDATE*

Transcription

1 Aricles Winer 2006 THE CALCULATION OF CYCLICALLY ADJUSTED BALANCES AT BANCO DE PORTUGAL: AN UPDATE* Cláudia Rodrigues Braz** 1. INTRODUCTION 1 In recen years, he cyclically adjused budge balance has gained relevance as one of he indicaors used in he assessmen of he underlying posiion of public finances in he European Union Member-saes. Indeed, he reform of he Sabiliy and Growh Pac increased he role of he balance adjused for cyclical effecs and emporary measures, when defining, a he prevenive arm level, he medium-erm fiscal objecive based on his variable, as well as he annual minimum convergence required for he Member-saes ha have no ye reached i. As regards he correcive arm, he fiscal adjusmen imposed o Member-saes incurring in excessive defici siuaions in he recen period has also been measured in erms of he change in he balance adjused for cyclical effecs and emporary measures. The difficulies in explaining his indicaor o he general public, he diversiy of available mehodologies, as well as he ex-pos revision of cyclically adjused balances, even if only as a resul of he updae of he fuure macroeconomic oulook, are among he facors ha have hampered is widespread use. Noneheless, alhough esimaes used in he European Union framework are hose of he European Commission, oher inernaional insiuions, such as he OECD and he IMF, have been regularly publishing values for cyclically adjused budge balances. As far as he European Sysem of Cenral Banks (ESCB) is concerned, a cyclical adjusmen mehodology was adoped in 2001, which has since hen been followed by Banco de Porugal. The values obained are regularly made public in he Banco de Porugal publicaions, namely in he Annual Repor and in he Economic Bullein. The purpose of his aricle is o updae parameers and o improve some assumpions made by Banco de Porugal in he iniial implemenaion of he ESCB cyclical adjusmen mehodology. Moreover, i applies he resuls o he analysis of public finances based on he disaggregaed framework developed wihin he scope of a projec involving some cenral banks, including Banco de Porugal. The ex is srucured as follows. Secion 2 presens an inroducory shor descripion of he main feaures of he ESCB mehodology and is applicaion o he Poruguese case. Secion 3 includes he updaes and improvemens of such mehodology, as regards boh he re-esimaion of some elasiciies of he fiscal variables wih respec o heir macroeconomic bases and he adopion of new macroeconomic bases. As a resul of hese changes, Secion 4 shows heir impac on he esimaion of he cyclically adjused balances vis-à-vis he laes values made public by Banco de Porugal. Secion 5 adaps he analysis of public finances based on he disaggregaed framework o he new procedures and includes he new resuls obained. Finally, Secion 6 draws he conclusions. * The views expressed are of he auhor and do no necessarily reflec hose of Banco de Porugal. Any error or omission are of her own responsabiliy. The auhor would like o hank Mara Abreu, Nuno Alves, Mário Ceneno, Jorge Cunha, Ana Crisina Leal, Sara Moreira and Maximiano Pinheiro for commens and suggesions. ** Economic Research Deparmen. (1) The general governmen accouns for he period underlying he preparaion of his aricle were compiled by he Naional Saisical Insiue (INE), on a 2000 Naional Accoun basis, wihin he scope of he Sepember 2006 excessive defici procedure noificaion. Daa from 1995 o 1998 are Banco de Porugal esimaes. In erms of he macroeconomic scenario, he complee se of Naional Accouns according o he new basis only covers he years from 1999 o 2002, wherefore he daa for he years from 1995 o 1998 and from 2003 o 2005 are also Banco de Porugal esimaes. Economic Bullein Banco de Porugal 75

2 Winer 2006 Aricles 2. THE METHODOLOGY OF THE EUROPEAN SYSTEM OF CENTRAL BANKS Currenly, Banco de Porugal calculaion of he cyclically adjused general governmen balance, regularly made public in he Economic Bullein and in he Annual Repor, is based on he mehodology agreed in he ESCB in 2001 (see Bouhevillain e al. (2001) and Neves and Sarmeno (2001)). This mehodology assumes ha he fiscal variables influenced by he economic cycle have oher-han GDP macroeconomic bases ha beer explain heir developmen. However, in order o deermine he semi-elasiciy of he budge balance wih respec o GDP, hese macroeconomic bases mus be Naional Accouns aggregaes. I is worh highlighing hree aspecs relaed o he selecion of he fiscal variables and he respecive macroeconomic bases. Firsly, i is assumed ha non-ax revenue and he differen expendiure componens, oher han unemploymen benefis, are no affeced by economic aciviy. This characerisic is common o he cyclical adjusmen mehodologies implemened by imporan inernaional insiuions, such as he European Commission, he OECD and he IMF, and resuls from he difficuly in measuring, in a sandardised manner in he differen counries and even in one counry in differen years, he impac of he economic cycle on mos public expendiure iems and non-ax revenue. Indeed, he insiuional srucure relevan for budgeary decisions, he ype of governmen, he saring poin of he fiscal posiion, iner alia, are facors deermining he response of public expendiure o he macroeconomic conex. Secondly, he macroeconomic bases are defined in real erms, which means ha he effec of acual inflaion on public accouns is no aken ino consideraion. The main difficulies in he quanificaion of he impac of he price changes on he budge balance emerge on he expendiure side, given ha i highly depends on he behaviour of budgeary auhoriies in response o inflaion deviaions from iniial forecass. Therefore, alhough in erms of ax and social conribuions revenue i would be more appropriae o consider nominal macroeconomic bases in he cyclical adjusmen, his does no happen in he mehodologies currenly implemened. However, he disaggregaed framework menioned in Secion 5 considers his issue o some exen, when assuming ha he srucural change in each of he iems of ax revenue evolves in line wih he respecive rend macroeconomic base defined in nominal erms. Finally, some iems are excluded from he fiscal variables so ha he reamen of revenue and expendiure in erms of he cyclical adjusmen is consisen. A presen, in he Poruguese case, he revenue of he personal income ax (Poruguese acronym: IRS) deriving from final wihholding schemes, which sems essenially from ineres on financial insrumens, and from public secor labour income is excluded from he receips of axes on household income. The same occurs wih acual conribuions o he civil servans pension sysem and impued conribuions, which are deduced o he oal amoun of social conribuions. Table 1 presens he fiscal variables wih a cyclical impac, as well as he respecive macroeconomic bases. Taking ino accoun hese consideraions, he cyclical componen of each of he fiscal variables is calculaed according o he following formula: R C i Xi X i R X R i, * i i X * i (1) where: R i fiscal variable i; R C i cyclical componen of fiscal variable i; X i macroeconomic base for fiscal variable i; X i rend value of he macroeconomic base for fiscal variable i; 76 Banco de Porugal Economic Bullein

3 Aricles Winer 2006 R i X i, elasiciy of fiscal variable i relaive o is macroeconomic base. The rend values of he macroeconomic bases are obained by applying he Hodrick-Presco filer o series exended wih forecass elaboraed by each naional cenral bank, so as o avoid a bias a he end of he sample, in which he value for he filer smoohing parameer () is 30. The esimaion of he elasiciies in he Poruguese case was chiefly based on he ax codes and is described in deail in Neves and Sarmeno (2001) One of he main advanages of he disaggregaed approach suggesed by he ESCB vis-à-vis he oher mehodologies for he calculaion of cyclically adjused balances consiss in he possibiliy of aking ino accoun he effecs of differen composiions of economic growh on public accouns. However, in view of he radiional approach in which he cyclical componen depends only on he oupu gap, i has he disadvanage of no allowing he sraighforward calculaion of he semi-elasiciy of he fiscal balance as a percenage of GDP wih respec o real GDP, defined as: FB% GDP FB % GDP, GDP r GDPr GDP r (2) Noneheless, in he mehodology adoped by he ESCB, his semi-elasiciy can be obained indirecly from he following formula: Ri FB GDP GDP * * i GDP %, r Ri, Xi Xi, GDPr The elasiciies of he macroeconomic bases wih respec o real GDP x i GDP r (3), were calculaed on he basis of a specific shock scenario, described in deail in Bouhevillain e al. (2001). This one is based on Naional Accouns ideniies and akes ino accoun six main assumpions: (i) he conribuion of ne expors o he change in real GDP is nil; (ii) public consumpion (including public wages and employmen), public invesmen and subsidies do no reac under he shock scenario; (iii) he conribuion of consumpion of fixed capial o he change in real GDP is negligible; (iv) he gross operaing surplus and he privae secor wage bill grow a he same rae; (v) privae invesmen evolves similarly o privae consumpion; (vi) he elasiciy of he labour force wih respec o real GDP is almos nil. As can be observed in Table 2, based on he resuls hen assumed for he Poruguese case (columns 1 and 2), Table 1 BUDGET VARIABLES AND RESPECTIVE MACROECONOMIC BASES IN ESCB METHODOLOGY Taxes on household income (excluding he IRS revenue from final wihholding schemes and public secor labour income) Privae secor wage bill Taxes on corporae income Gross operaing surplus Taxes on producion and impors (including European Union own receips) Privae consumpion Social conribuions (excluding acual conribuions of he civil servans pension sysem and impued conribuions) Privae secor wage bill Unemploymen benefis Number of unemployed Economic Bullein Banco de Porugal 77

4 Winer 2006 Aricles Table 2 THE SEMI-ELASTICITY OF THE FISCAL BALANCE WITH RESPECT TO GDP IN THE ESCB METHODOLOGY Elasiciy of Elasiciy of he fiscal he variable wih macroeconomic respec o he base wih macroeconomic respec base (a) o GDP (a) Elasiciy of he fiscal variable wih respec o GDP Weigh of he fiscal variable on GDP (b) Conribuion o he semi-elasiciy of he budge balance wih respec o GDP (1) (2) (3)=(1)*(2) (4) (5)=(3)*(4) Taxes on household income Macroeconomic base: privae secor wages Macroeconomic base: privae employmen Taxes on corporae income Taxes on producion and impors Social conribuions Macroeconomic base: privae secor wages Macroeconomic base: privae employmen Unemploymen benefis Toal: 0.43 Noes: (a) According o Neves and Sarmeno (2001) and o he secion abou Porugal in Bouhevillain e al. (2001). (b) The raios of he fiscal variables o GDP are based on Naional Accouns and correspond o averages for he period. In order o obain raios ha are more consisen wih he mehodology o calculae cyclically adjused balances adoped by he ESCB, IRS revenue from final wihholding schemes and public secor labour income was excluded in he case of axes on household income, acual conribuions o he civil servans pension sysem and impued conribuions were deduced from oal social conribuions, while he European Union own receips were added o axes on producion and impors. bu wih he currenly available general governmen accouns (column 4), he value obained for he semi-elasiciy of he balance wih respec o GDP is UPDATES AND IMPROVEMENTS WITHIN THE SCOPE OF THE ESCB METHODOLOGY This Secion analyses he relaionship beween he differen fiscal variables assumed o be influenced by he economic cycle and he respecive macroeconomic bases in he period. In some cases, i is suggesed o change and updae he procedures adoped by Neves and Sarmeno (2001). Concerning he esimaion of elasiciies, hree key poins are worhy of noe. Firsly, he period considered in he analysis covers chiefly one economic cycle and is herefore shor in erms of he acual number of observaions. However, he possibiliy of saring he analysis in some earlier year is hindered by he fac ha he inroducion of he main axes in he curren Poruguese ax sysem occurred chiefly unil he lae 1980s/early 1990s, as well as by he absence of a reropolaion of Naional Accouns on a 2000 basis for he period prior o Secondly, since he objecive is o deermine he cyclical elasiciies of he fiscal variables, regressions are based on series of differen axes and social conribuions adjused for he effecs of permanen or emporary discreionary measures. Noneheless, i is no possible o disregard he effecs of possible srucural changes ha may have occurred, bu, in view of he shor period of he analysis, hey are no expeced o be significan. Lasly, i is worh sressing ha he specificaion of he regressions is deemed appropriae for he purpose of he elasiciy esimaion o be used in he cyclical adjusmen of he budge balance wihin he scope of he ESCB mehodology and, as such, is no necessarily based on he explanaory variables which would allow he bes possible fi o daa observed. 78 Banco de Porugal Economic Bullein

5 Aricles Winer Taxes on household income The esimaion of he elasiciy of axes on household income wih respec o he privae secor wage bill is described in deail in Neves and Sarmeno (2001) and was based on daa made available by he ax adminisraion relaive o a sample of axpayers in The values obained (1.69 for privae secor wages and 1.0 for privae employmen) will no be updaed in his aricle. However, in spie of he discreionary measures aken in recen years, which had an impac on he srucure and progressiviy of he ax, he esimaed elasiciy is a reasonably good proxy for he behaviour of his ax revenue in he period, paricularly in he mos recen years. Char 1 presens he growh raes of revenue from axes on household income, resuling chiefly from he IRS, excluding he revenue from final wihholding schemes and from public secor labour income. This series was adjused aking ino accoun he esimaed impac on IRS of he main emporary or permanen discreionary measures implemened from 2000 o Worhy of noe among hese are: he effec of he ax rae cus and he inroducion of a new bracke in 2001, as a resul of he income ax reform in 2000; he above-inflaion updae of he ax wihholding ables in 2002, wih an opposie impac on ax revenue in 2002 and 2003; he reducion in he ax raes included in he 2005 Budge, only parly refleced in he updae of he ax wihholding ables in ha year; he exraordinary selemen of ax arrears in lae 2002 and early 2003; he securiisaion of ax credis in Char 1 also includes a series consruced from he esimaed elasiciy and he developmen of he respecive macroeconomic base, according o he following formula: e RTaxes hous inc Wpriv UIRS E * priv U IRS E priv (4) where: R e Taxes hous. inc. Esimaed growh rae of axes on household income (excluding he IRS derived from final wihholding schemes and public secor labour income); W priv Growh rae of he privae secor wage bill, ne of social conribuions; U IRS Annual updae of he IRS brackes and oher parameers ha usually coincide wih he updae of he ax wihholding ables; E priv Growh rae of privae employmen. Equaion (4) indicaes ha he effec of progressiviy, via he IRS srucure, will only be refleced in he collecion of hese axes when he developmen of wages differs from he updae of he IRS brackes and oher parameers. In effec, his should accoun for a significan par of he very marked difference beween he wo series presened in Char 1 in 1997 and 1998, when he IRS brackes and oher parameers were updaed in line wih expeced inflaion bu privae secor wages grew significanly above such values. In 1996, he opposie effec may be explained by he selemen of deb arrears wihin he scope of he Maeus plan. Finally, he divergence of he wo series in 2005, wih a more favourable behaviour of revenue from axes on household income han esimaed aking ino accoun he elasiciy, may parly illusrae he impac increased effeciveness of ax adminisraion in he collecion of hese axes. (2) IRS changes inroduced from 1995 o 1999 were no explicily considered in his analysis due o difficulies in he quanificaion of heir effecs. Economic Bullein Banco de Porugal 79

6 Winer 2006 Aricles Char 1 GROWTH RATES OF TAXES ON HOUSEHOLD INCOME 2 "Adjused" acual series (96-99 no "adjused") "Consruced" series elasiciy-based 15.0 Per cen Taxes on corporae income Taxes on corporae income, mainly composed by he Corporae Income Tax (Poruguese acronym: IRC) show a very volaile behaviour over ime, which makes exremely difficul o selec an appropriae macroeconomic base o model is developmen. Indeed, in Porugal, IRC revenue is highly dependen on he developmen of profis in some large corporaions ha are naurally he main conribuors o his ax collecion. In addiion, he fac ha he variable profi is no one of he aggregaes compiled in he elaboraion of he Naional Accouns and he mehod iself for he calculaion of his ax receips, in paricular he possibiliy of deducing losses from previous years, make i even more difficul o esablish a relaionship wih he macroeconomic scenario. Neves and Sarmeno (2001) used gross operaing surplus as a proxy o corporae profis and, given he proporionaliy of he ax, assumed an elasiciy of revenue wih respec o is macroeconomic base equal o one. Char 2 shows he growh raes of revenue from axes on corporae income in he period, adjused for he effecs of he main emporary and permanen ax policy measures. I could be referred, as examples, he IRC rae cus from 36 o 34 per cen (1997), from 34 o 32 per cen (2000), from 32 o 30 (2002) and from 30 o 25 per cen (2004), he change in he percenage used o calculae pre-paymens (2000) and a number of oher changes relaed o special pre-paymens. In he years when gross operaing surplus growh raes are available, he resuling developmens seem o sugges ha privae GDP is a more suiable variable o capure he dynamics of revenue from axes on corporae income. Afer he choice of he new macroeconomic base, defined in real erms, he revenue elasiciy was obained from he esimaion of equaion (5) by ordinary leas squares, using daa from 1995 o 2005, and is value corresponded o Taxes corp. inc where: ln cor RTaxes corp. inc. lngdp deflaor priv. Taxes real corp. inc. ln GDPpriv. cor R Taxes corp. inc. Revenue from axes on corporae income adjused for he effecs of ax policy measures; deflaor GDP priv. Privae GDP deflaor; (5) 80 Banco de Porugal Economic Bullein

7 Aricles Winer 2006 real GDP priv. Real privae GDP. Three aspecs should be highlighed in his esimaion. Firsly, he specificaion of equaion (5) assumes a uni elasiciy wih respec o he privae GDP deflaor, which capures he fac ha revenue from hese axes acually depends on a nominal base. Secondly, he period used in he analysis includes he year 2005, alhough his one was much affeced by effeciveness gains in ax collecion of a non-cyclical naure and difficul o quanify. However, he inclusion of his observaion does no change significanly he value obained for he elasiciy. Finally, a resul for he elasiciy higher han one does no reflec any progressiviy in his ax, bu only he use of a macroeconomic variable o capure he cyclical effec wih an average growh clearly below he one of ax receips. I should also be noed ha he series of axes on corporae income receips based on he esimaed elasiciy presens less volailiy han he acual series, as can be confirmed in Char 2. Char 2 GROWTH RATES OF TAXES ON CORPORATE INCOME Per cen "Adjused" acual series* "Consruced" series elasiciy-based Previous "consruced" series Noe: The previous consruced series is based on he nominal growh raes of gross operaing surplus (from1995 o 2000 in he previous Naional Accouns basis and from 2001 o 2002 in he Naional Accouns basis 2000) and a uni elasiciy Taxes on producion and impors Neves and Sarmeno (2001) adoped a procedure based on he daa from he Survey of Household Budges for 25 caegories of consumer goods and he respecive ax raes, which allowed he esimaion of a value of 1.1 for he elasiciy of axes on producion and impors, as a whole, wih respec o privae consumpion. This aricle esimaes individual elasiciies for he main axes on producion and impors, i.e., he Value Added Tax (VAT), he Tax on Oil Producs and he Car Tax, while oher axes are considered ogeher. As regards VAT, he esimaion of he elasiciy wih respec o real privae consumpion was based on equaion (6), using he daa for he revenue of his ax from 1995 o 2005, adjused for he impac of he main discreionary changes implemened in his period. Among hese, i should be highlighed he inroducion of an inermediae rae of 12 per cen in 1996 and he rise in he sandard rae from 17 o 19 per cen in 2002 and o 21 per cen in The growh raes adjused for hese effecs are presened in Char 3. Given ha he VAT is a ax on he sales of goods and services, i depends no only on he quaniy of goods and services raded, bu also on heir price. Therefore, equaion (6) assumes ha a 1 Economic Bullein Banco de Porugal 81

8 Winer 2006 Aricles per cen increase in he privae consumpion deflaor is proporionally refleced in VAT revenue. The value obained for he elasiciy, VAT =1.69, poins o a more-han-proporional effec on revenue. This may imply ha in periods of economic expansion (recession) he households end o change heir consumpion paern owards more (less) goods axed a he sandard rae and less (more) subjec o he reduced rae. Moreover, in he period under review, oher srucural facors have occurred, such as he modernisaion of he disribuion circuis, whose posiive effec canno be quanified, and, as such, may lead o some overesimaion of he elasiciy wih respec o he economic cycle. Neverheless, acual and elasiciy-based growh raes are relaively close, even in mos recen years, as can be seen in Char 3. cor deflaor real RVAT C VAT C ln ln ln (6) where: cor R VAT VAT revenue, including European Union own receips, adjused for he effecs of ax policy measures; C deflaor Privae consumpion deflaor; C real Real consumpion deflaor. As regards he Tax on Oil Producs, changes in ax raes were frequen beween 1995 and However, aking ino accoun he quaniies of fuel consumed, he eliminaion of is effecs on he series used in he esimaion of he elasiciy is quie sraighforward (Char 4). However, wo aspecs should be highlighed. On he one hand, privae consumpion of services and non-durable goods measured in real erms is deemed o be an appropriae, bu no perfec, macroeconomic base for he quaniies of fuel consumed. On he oher hand, ax revenue does no depend on acual prices, since is defined in moneary unis per lire of fuel consumed. The elasiciy of revenue from he Tax on Oil Producs wih respec o real privae consumpion of services and non-durable goods TOP esimaed for he period is In mos recen years, in paricular from 2003 o 2005, he series consruced on he basis of he esimaed elasiciy seems o overesimae he ax revenue acual growh. This resul should be mainly explained by he fac ha he volume of privae consumpion of services and non-durable goods does no accuraely reflec he quaniies of fuel sold. To a lesser exen, i may also Char 3 Char 4 VAT GROWTH RATES GROWTH RATES OF THE TAX ON OIL PRODUCTS 2 "Adjused" acual series "Consruced" series elasiciy-based 1 "Adjused" acual series "Consruced" series elasiciy-based 15.0 Per cen 1 Per cen Banco de Porugal Economic Bullein

9 Aricles Winer 2006 be influenced by changes in he composiion of vehicles in circulaion, wih a growing share of diesel vehicles, as his fuel is axed a lower raes. ln cor real RTOP TOP lncnon durables (7) where: cor R TOP Revenue from he Tax on Oil Producs adjused for he effecs of ax policy measures; real C non durables Privae consumpion of services and non-durable goods measured in real erms. The Car Tax, on is urn, depends on cerain characerisics of he vehicles, such as he cylinder capaciy in cubic cenimeres. I is esablished in ables ha are updaed every year in he Sae Budge Law, usually in line wih expeced inflaion. Therefore, revenue from his ax derives chiefly from he number of vehicles sold and heir qualiy, measured by he privae consumpion of durable goods in real erms for he purpose of calculaing he elasiciy, and from expeced inflaion considered in he differen Sae Budges, which jusifies he specificaion of equaion (8). In addiion, changes in he ax srucure, similarly o changes inroduced in oher axes, may jusify he correcion of he observed series via he impac of such measures. However, in he period analysed, hese were no significan, and he mos relevan occurred in 2001 wih an increase in he axaion on off-road vehicles. From he esimaion of equaion (8) for he period, based on ordinary leas squares, resuled an elasiciy of CT =1.33, which brings he growh of he esimaed series based on he elasiciy reasonably closer o he one of acual revenue, as can be observed in Char 5. budg. real RCT Inf CT Cdurables ln ln ln (8) where: R CT Acual revenue from he Car Tax; Inf budg. Price index consruced on he basis of he expeced inflaion included in he Sae Budge; real C durables Privae consumpion of durable goods measured in real erms. Finally, as regards oher axes on producion and impors, a uni elasiciy was assumed wih respec o real privae consumpion. In effec, his aggregae includes a number of axes, much differen in heir incidence and calculaion. Sress should be laid, for heir relevance, on he Tax on Tobacco, he Samp Tax, he Municipal Tax on Real Esae and he Municipal Tax on Real Esae Transacions. 3 In addiion, in he period under analysis, he srucure and raes of hese axes were subjec o some changes, whose individual effecs are difficul o esimae. As a resul, in parallel wih he impac of he price level on he revenue of hese axes, he growh of his aggregae usually exceeds he one esimaed on he basis of a uni elasiciy, as illusraed in Char 6. In mos recen years, he difference beween boh series is parly explained by he increase in he Tax on Tobacco included in he differen Sae Budges, as well as by he reform of axaion on real esae. The laer, on accoun of he updaes of he values of he properies o be used for ax purposes, has been behind he srong rise in revenue from he Municipal Tax on Real Esae. Table 3 summarises he values esimaed and assumed for he elasiciies of he differen axes on producion and impors wih respec o heir bases and real privae consumpion, as well as he calculaion (3) The Samp Tax receip relaed wih banking operaions is reclassified as ax on income and propery in Naional Accouns. I should also be highlighed ha revenue from axes on propery is recorded as ax on producion and impors in Naional Accouns, bu i is classified under direc axes in Public Accouns. Economic Bullein Banco de Porugal 83

10 Winer 2006 Aricles Char 5 Char 6 GROWTH RATES OF THE CAR TAX GROWTH RATES OF OTHER TAXES ON PRODUCTION AND IMPORTS Acual series "Consruced" series elasiciy-based 1 Per cen Per cen Acual series "Consruced" series elasiciy-based Table 3 THE ELASTICITY OF TAXES ON PRODUCTION AND IMPORTS Taxes on producion Elasiciy wih Elasiciy wih Average weigh on oal Toal elasiciy of axes and impors respec o he base respec o privae axes on producion and on producion and consumpion (a) impors (a) impors wih respec o privae consumpion VAT Tax on oil producs Car ax Oher Noe: (a) Based on average weighs from 1995 o of an aggregae elasiciy for hese axes as a whole. This aggregae elasiciy sands a 1.3, above he 1.1 value considered in Neves and Sarmeno (2001) Social conribuions As presened in Secion 2, when applying he ESCB cyclical adjusmen mehodology, only social conribuions of he privae secor sysem are deemed o be affeced by economic aciviy, so ha he reamen of public revenue is consisen wih ha of expendiure. As far as he elasiciy is concerned, given he exisence of single raes for he share of employers (23.75 per cen) and of employees (11 per cen), Neves and Sarmeno (2001) assumed ha he behaviour of revenue from hese conribuions wih respec o he privae secor real wage bill was proporional. As illusraed in Char 7, his assumpion seems o replicae raher closely he acual growh in privae secor social conribuions in he period analysed and, as such, i will be mainained. Wih regard o ax policy measures or oher facors which may disor he relaionship wih he macroeconomic base, he series in he char was only adjused for he exraordinary effec of a Sae s paymen o Social Securiy in 2004, regarding social conribuions he Sae commied o pay on behalf of small farmers beween 2001 and Banco de Porugal Economic Bullein

11 Aricles Winer 2006 Char 7 GROWTH RATES OF SOCIAL CONTRIBUTIONS OF THE PRIVATE SECTOR SYSTEM 15.0 "Adjused" acual series "Consruced" series elasiciy-based 1 Per cen Unemploymen benefis In mos EU Member Saes he cyclical componen of unemploymen, wihin he framework of he ESCB mehodology, is obained wih he Hodrick-Presco filer. However, in he Poruguese case, he unemploymen gap was esimaed on he basis of he difference beween acual unemploymen and naural unemploymen, since conclusions from a number of empirical sudies for Porugal poin o reasonably sable values for he naural rae of unemploymen since he beginning of he 1980s (see equaion (9)). In his conex, Neves and Sarmeno (2001) assumed a naural rae of unemploymen of 5.0 per cen, as well as he proporionaliy beween expendiure on unemploymen benefis and he number of unemployed. no. unemployed naural rae of unemploymen labour force Unemployed gap naural rae of unemploymen labour force (9) In Porugal, daa on he number of unemployed regisered in he Employmen and Professional Training Insiue (Poruguese acronym: IEFP) are available on a regular basis. This series, which in he pas showed a raher volaile difference vis-à-vis he number of unemployed in he economy obained on he basis of he Labour Force Survey, is currenly more accurae o explain he developmens in expendiure on unemploymen benefis. In fac, in recen years, changes in he rules for graning hese benefis and in he monioring by employmen offices of job seeking effors by he unemployed rendered his variable a good indicaor for developmens in expendiure on unemploymen benefis. Char 8 shows ha in mos years under analysis, as anicipaed, raes of change in unemploymen benefis sand above hose in he number of regisered unemployed, which illusraes a posiive growh in he average unemploymen benefi, in line wih wage developmens. Neverheless, in addiion o he effec of he change in he srucure of unemploymen hroughou he period and in he relaed benefis, he difference beween he wo series is also affeced by changes inroduced in he meanime in he rules for graning unemploymen benefis. The use of he number of regisered unemployed requires changes in he rend number of unemployed aken ino accoun in he cyclical adjusmen. In fac, as illusraed in Char 9, he raio of he number of regisered unemployed (calculaed on he basis of IEFP daa) o he labour force is clearly higher han Economic Bullein Banco de Porugal 85

12 Winer 2006 Aricles Char 8 Char 9 GROWTH RATES OF UNEMPLOYMENT BENEFITS AND THE NUMBER OF UNEMPLOYED RATIOS OF THE NUMBER OF UNEMPLOYED TO THE LABOUR FORCE Unemploymen benefis Number of (regisered) unemployed 9.0 Raio of he number of regisered unemployed o he labour force Raio of he number of unemployed o he labour force Per cen Per cen he one based on he number of unemployed in he economy (calculaed according o he Labour Force Survey resuls), and poins o an average value of 7.2 per cen beween 1995 and 2005, compared wih 5.6 per cen in he case of unemployed in he economy. This laer value is higher han ha considered by Neves and Sarmeno (2001) for he naural unemploymen rae, alhough i is quie in line wih he reassessmen made by Dias e al. (2004). These auhors, by using quarerly daa from 1983 o 2003 and alernaive approaches, obained average esimaes for he NAIRU 4 a around 5.5 per cen. Hence, in he pracical implemenaion of he cyclically adjused balances mehodology, he gap in he number of unemployed will be calculaed on he basis of he difference beween he number of regisered unemployed and he average number of unemployed equal o 7.2. per cen of he labour force Semi-elasiciy of he budge balance The change in some elasiciies of fiscal variables vis-à-vis heir macroeconomic bases, elaboraed hroughou his Secion, allows he updae of he budge balance s semi-elasiciy wih respec o GDP. In his conex, i is also convenien o revise he elasiciies of macroeconomic bases wih respec o GDP, based on he currenly available Naional Accouns daa and drawing on he shock scenario specified by Bouhevillain e al. (2001). As menioned in Secion 2, his exercise is based on he relaionship among he differen macroeconomic bases via Naional Accouns ideniies and, as such, is calculaion does no require a simulaion using a macroeconomic model. Table 4 shows he resuls obained and poins o a semi-elasiciy of This increase from he previous esimae (0.43) reflecs he increase in he elasiciies of budge variables vis-à-vis he respecive macroeconomic bases, since he updaing of he sensiiviies of he differen macroeconomic bases wih respec o GDP would by iself lead o a decline in he semi-elasiciy. In addiion, i should be noed ha he curren esimae for he semi-elasiciy is very close o he values currenly used by he European Commission (0.45) and he OECD (0.46). (4) Non-acceleraing inflaion rae of unemploymen, i.e. an unemploymen rae ha is compaible wih a consan inflaion rae. 86 Banco de Porugal Economic Bullein

13 Aricles Winer 2006 Table 4 THE UPDATED SEMI-ELASTICITY OF THE BUDGET BALANCE WITH RESPECT TO GDP IN THE ESCB METHODOLOGY Elasiciy of Elasiciy of he fiscal he variable wih macroeconomic respec o he base wih macroeconomic respec base o GDP Elasiciy of he fiscal variable wih respec o GDP Weigh of he fiscal variable on Conribuion o he semi-elasiciy of he budge balance wih respec o GDP GDP (a) (1) (2) (3)=(1)*(2) (4) (5)=(3)*(4) Taxes on household income Macroeconomic base: privae secor wages Macroeconomic base: privae employmen Taxes on corporae income Taxes on producion and impors Social conribuions Macroeconomic base: privae secor wages Macroeconomic base: privae employmen Unemploymen benefis Toal: 0.50 Noe: (a) The raios of he fiscal variables o GDP are based on Naional Accouns and correspond o averages for he period. In order o obain raios ha are more consisen wih he mehodology o calculae cyclically adjused balances adoped by he ESCB, IRS revenue from final wihholding schemes and public secor labour income was excluded in he case of axes on household income, acual conribuions o he civil servans pension sysem and impued conribuions were deduced from oal social conribuions, while he European Union own receips were added o axes on producion and impors. 4. CYCLICALLY ADJUSTED BALANCES On he basis of he mehodology implemened by Neves and Sarmeno (2001), Banco de Porugal had already changed he macroeconomic bases of corporae income axes and unemploymen benefis o, respecively, privae GDP and he number of regisered unemployed. Vis-à-vis he figures now presened, he elasiciy of corporae income axes was higher and he average share of regisered unemployed wih respec o he labour force was lower. Therefore, he differences beween he laes version of cyclically adjused balances made public by Banco de Porugal (in he Auumn 2006 issue of he Economic Bullein) and ha which would resul from he updae of parameers carried ou in his aricle are no very si as can be seen in Table 5. The revision of he oal balance adjused for cyclical effecs and emporary measures reaches a he mos an absolue value of 0.3 percenage poins of GDP in 2005, sanding below ha figure in all oher years considered in his analysis. The differences regarding he change in he primary balance adjused for cyclical effecs and emporary measures, an indicaor ha is commonly used o assess he fiscal policy sance, are negligible. Overall, changes inroduced in his aricle mainly enail a reducion of he cyclical componen, which is less posiive (negaive) during srong (weak) economic growh periods (Char 10). This resuls from he lower elasiciy of axes on corporae income, which more han offses he higher elasiciy of axes on producion and impors as whole wih regard o privae consumpion and he increase in he average share of regisered unemployed wih respec o he labour force (Char 11). Char 12 illusraes he effec of he composiion of economic growh on he cyclical componen of he budge balance, defined as he difference beween he cyclical componen calculaed using he ESCB mehodology, in line wih he updae described previously, and ha obained hrough he applicaion of he 0.5 semi-elasiciy of he budge balance wih respec o GDP o he oupu gap. I should be noed ha a posiive value for his effec means ha he behaviour of he differen macroeconomic bases as a whole, given heir fiscal elasiciies, was more favourable o he balance han he one which would re- Economic Bullein Banco de Porugal 87

14 Winer 2006 Aricles Table 5 CYCLICALLY ADJUSTED BALANCES: COMPARISON WITH THE LATEST FIGURES MADE PUBLIC As a percenage of GDP Cyclically adjused oal balance, excluding emporary measures Previous Revised Difference: revised-previous Change in he cyclically adjused primary balance, excluding emporary measures Previous Revised Difference: revised-previous Char 10 Char 11 OUTPUT GAP AND THE REVISION OF THE CYCLICAL COMPONENT OF THE BUDGET BALANCE REVISED CYCLICAL COMPONENTS Per cen Oupu gap Revision of he cyclical componen of he budge balance (RHS) Percenage poins of GDP As a percenage of GDP Corporae axes: revised Corporae axes: previous Taxes on producion and impors: revised Taxes on producion and impors: previous Unemploymen benefis: revised Unemploymen benefis: previous sul from a macroeconomic scenario feauring a balanced growh of he various GDP componens. In he period under review, his effec reached a maximum absolue value of around 0.3 per cen of GDP, i.e. of a fairly small magniude. Moreover, Char 12 also illusraes he semi-elasiciy of he budge balance wih respec o GDP implici in he cyclical componen of he balance obained by applying he ESCB mehodology. Given ha he composiion effec is calculaed as previously described, he fac of i being posiive (negaive) only ranslaes ino a semi-elasiciy above (below) 0.5 if he oupu gap is posiive, as explained in equaions (10) and (11). On average, in he period analysed, he implici semi-elasiciy is in fac very close o he 0.5 value esimaed in subsecion 3.6, in spie of he non-zero cumulaive value of composiion effecs. 88 Banco de Porugal Economic Bullein

15 Aricles Winer 2006 Char 12 COMPOSITION EFFECT AND THE IMPLICIT SEMI-ELASTICITY Composiion effec (% of GDP) Implici semi-elasiciy Composiion effec Cyclical componen 05. oupu gap (10) ESCB Cyclical componenescb Implici semi elasiciy oupu gap (11) 5. THE APPLICATION OF THE DISAGGREGATED FRAMEWORK The updae of some of he parameers considered in he cyclical adjusmen mehodology used by ESCB allows for an improvemen of he disaggregaed framework for assessing public finances inroduced by Kremer e al. (2006). This approach was recenly applied o he analysis of he 2005 fiscal developmens in he 2005 Annual Repor of Banco de Porugal. 5 In analyical erms, his approach is based on he analysis of he change in he raio of he main revenue and expendiure iems o nominal rend GDP, excluding he effecs of he economic cycle and emporary measures. Wih regard o axes and social conribuions, he change in heir srucural raios is broken down, for every iem, ino four explanaory facors: (i) he fiscal drag, which given nominal growh of he rend macroeconomic base, corresponds o he auomaic effec on ax revenue resuling from an elasiciy of he fiscal iem vis-à-vis he respecive base no equal o one; (ii) he decoupling of he macroeconomic base from GDP ha accouns for he effec of he growh differenial beween he rend macroeconomic base and nominal rend GDP on ax revenue; (iii) legislaion changes, whose fiscal impac corresponds o official esimaes or, in heir absence, is calculaed on he basis of daa available; and (iv) he residual, which allows o quanify he par of he developmens in srucural raios of axes and social conribuions ha is no explained by he previous hree facors. The residual componen is an imporan elemen of his framework and i may conribue o a beer undersanding of pas developmens, by indicaing he quaniaive imporance of paricular unsysemaic evens. I may also show favourable or unfavourable rends in specific fiscal iems. Moreover, i may reveal he need o reassess he impac of legislaion changes or fiscal elasiciies. (5) See Box 6.1 A disaggregaed framework for he analysis of public finances, 2005 Annual Repor, Banco de Porugal Economic Bullein Banco de Porugal 89

16 Winer 2006 Aricles Beyond he adjusmens resuling from he new cyclically adjused figures for he differen fiscal iems, he breakdown of axes on producion and impors ino he caegories defined in his aricle may also be included. Table 6 shows he srucure used in he disaggregaed framework, incorporaing all updaes and improvemens. 6 Three facors are paricularly relevan regarding his srucure. Firsly, he effecs of legislaion changes considered correspond o hose assumed in Secion 3 for he esimaion of elasiciies and he graphical display of resuls. Secondly, he impac of he fiscal elasiciy in he case of VAT, he Tax on Oil Producs and he Car Tax is nil, alhough he elasiciies used in he calculaion of he cyclical componen of he budge balance are above one. Indeed, he small number of observaions considered, as well as he insufficien breakdown of he differen cyclical and non-cyclical facors ha influence revenue from hese axes, do no allow, in srucural erms, o assume ha revenue reacs more han proporionally o rend values of he respecive macroeconomic bases. This procedure corresponded o ha iniially adoped in he implemenaion of he disaggregaed framework for axes on producion and impors as a whole, which also presened a cyclical elasiciy higher han one. Finally, in he case of he Tax on Oil Producs, he effec of he decoupling of he macroeconomic base from GDP is likely o be sysemaically negaive, given ha, in addiion o he impac of changes in he average ax raes, fully regarded as a legislaion change, ax revenue seems o depend only on he consumpion of fuel, whose developmens are measured by he volume of he privae consumpion of services and non-durable goods. By conras, in he case of he Car Tax, i is assumed ha he annual updae of he ables of his ax, proxied by he privae consumpion deflaor of durable goods, is par of he ax base and, as such, is effec on he change in he srucural raio of he revenue is capured in he decoupling of he macroeconomic base from GDP. Wih regard o resuls, and focusing he analysis on he residuals of he breakdown of axes on producion and impors, menion should be made o: (i) he very favourable behaviour of VAT revenue, in paricular from 1996 o 2000, which may be parly due o he facors menioned in subsecion 3.3; (ii) he residuals of he Tax on Oil Producs and he Car Tax ha are virually nil for all years, indicaing ha legislaion changes wih idenified impac, macroeconomic bases chosen and esimaed elasiciies explain very saisfacorily developmens observed during his period; (iii) he posiive residual in almos every year of oher axes on producion and impors, as expeced, given ha legislaion changes in hese axes, which end o lead o an increase in axaion, were no explicily considered in his analysis. (6) I should be referred ha, vis-à-vis he figures shown in Box 6.1 of he 2005 Annual Repor of Banco de Porugal, Table 6 akes ino accoun he revised general governmen accouns, compiled wihin he framework of he excessive defici procedure noificaion of Sepember Banco de Porugal Economic Bullein

17 Aricles Winer 2006 Table 6 CHANGES IN GENERAL GOVERNMENT BALANCES, REVENUE AND EXPENDITURE, ADJUSTED FOR CYCLICAL EFFECTS AND TEMPORARY MEASURES As a percenage of rend nominal GDP Acual balance (improvemen +, deerioraion -) (a) Cyclical componen Temporary measures Srucural balance (improvemen +, deerioraion -) Ineres paymens due o changes in he average ineres rae on public deb due o changes in he deb sock Srucural primary balance (improvemen +, deerioraion -) Toal revenue Taxes on household income Fiscal drag Decoupling of he macroeconomic base from GDP Legislaion changes Residual Taxes on corporae income Fiscal drag Decoupling of he macroeconomic base from GDP Legislaion changes Residual VAT Fiscal drag Decoupling of he macroeconomic base from GDP Legislaion changes Residual Tax on oil producs Impac of budge elasiciy Decoupling of he macroeconomic base from GDP Legislaion changes Residual Car ax Fiscal drag Decoupling of he macroeconomic base from GDP Legislaion changes Residual Oher axes on producion and impors Fiscal drag Decoupling of he macroeconomic base from GDP Legislaion changes Residual Social conribuions Fiscal drag Decoupling of he macroeconomic base from GDP Legislaion changes Residual Memo iem: included in expendiure (b) Toal axes and social conribuions Fiscal drag Decoupling of he macroeconomic base from GDP Legislaion changes Residual Memo iem: included in expendiure (b) Oher revenue of which: from he EU (c) Primary expendiure Social paymens of which: Pension paymens Unemploymen benefis Social benefis in kind Compensaion of employees Inermediae consumpion Subsidies Invesmen Oher expendiure (d) Noes: (a) Change in he acual balance, he cyclical componen and he effecs of emporary measures as a percenage of nominal GDP. Due o differences in he denominaor, he change inhe srucural balance as a percenage of nominal rend GDPmaydiffer slighly fromhe change inhe raio of acual balance excluding cyclicaleffecs and emporary measures o nominal GDP. (b) Par of he residual of acual social conribuions relaing wih social conribuions of he civil servans sysem and impued social conribuions, boh recorded on he expendiure side under he iemcompensaion of employees. (c) Includes oher curren revenue, sales and capial revenue. (d) Includes oher curren primaryand capial expendiure. Economic Bullein Banco de Porugal 91

18 Winer 2006 Aricles 6. CONCLUSIONS The analysis of Naional Accouns daa is very imporan for undersanding he relaionship beween fiscal variables affeced by he economic cycle in he ESCB mehodology and he respecive macroeconomic bases. In he case of axes on corporae income, his analysis le o he conclusion ha privae GDP, alhough shifing away from he acual base of hese axes, is able o capure revenue developmens more accuraely. Given ha his macroeconomic variable growh is, on average, below ha of revenue from axes on corporae income, he new esimae for he elasiciy of revenue, based on a series from 1995 o 2005 adjused for he effecs of discreionary measures, is higher han one. As regards axes on producion and impors, which up unil now had been deal wih as a whole in sudies on he cyclical adjusmen of fiscal variables, he presen aricle explains he developmens of he main axes VAT, Tax on Oil Producs, and Car Tax using he mos appropriae Naional Accouns aggregaes. Therefore, similarly o he descripion made above when commening axes on corporae income, elasiciies were esimaed for he hree axes wih respec o privae consumpion, privae consumpion of services and non-durable goods and privae consumpion of durable goods, respecively. The values obained were, in all cases, above one. As regards he remaining par, consising of very differen axes, a uni elasiciy wih respec o privae consumpion was assumed. As a whole, he new elasiciy of axes on producion and impors wih respec o privae consumpion reflecs a more significan response of his revenue o he economic cycle han ha iniially assumed in he implemenaion of his mehodology. Turning o unemploymen benefis, he changes inroduced over recen years made i possible ha he series of he number of unemployed regisered in he IEFP currenly describes more appropriaely is developmen han he number of unemployed obained from he Labour Force Survey. In he calculaion of he cyclical componen, i was also necessary o adjus accordingly he average value of he raio of he regisered number of unemployed o he labour force. In his aricle, he elasiciies of axes on household income and social conribuions of he privae secor sysem were kep unchanged vis-à-vis he assumpions considered in he iniial implemenaion of he ESCB cyclical adjusmen mehodology. Acually, i was confirmed ha hey explain he developmens of hese iems of revenue much saisfacorily, in paricular in he mos recen period. As a whole, he changes inroduced led o a negligible increase in he semi-elasiciy of he budge balance as a percenage of GDP wih respec o GDP, when compared o he value obained when repeaing he procedures adoped in Indeed, alhough he revision of he elasiciies indicae more sensiiviy regarding he respecive macroeconomic bases, an updae of he elasiciies of he macroeconomic bases wih respec o GDP poins in he opposie direcion and parly offses he iniial effec. Presenly, he value obained for he semi-elasiciy of he budge balance is 0.50, sanding relaively close o he ones used by he European Commission and he OECD. Banco de Porugal had already inroduced some of he above-menioned changes in he calculaion of he cyclically adjused balances. Therefore, in view of he laes figures made public, he revisions of he oal balance adjused for cyclical effecs and emporary measures, an indicaor usually used in he assessmen of he underlying budgeary posiion, are relaively negligible in he period under review. Differences in he change of he primary balance adjused for cyclical effecs and emporary measures are also insignifican. As regards he laes esimaions made public, he cyclical componen of he balance is now of a smaller magniude, i.e., less posiive (negaive) in periods of srong (weak) economic growh. Improvemens inroduced in he calculaion of he cyclically adjused balances also make i possible o furher develop he analysis based on he disaggregaed framework of public finances. This aricle il- 92 Banco de Porugal Economic Bullein