Labour and Product Market Reforms and Fiscal Deficit

Transcription

1 Labour and Produc Marke Reforms and Fiscal Defici Juha Kilponen Bank of Finland Ani Ripai Bank of Finland Revised: June 16, 2008 Absrac Using a DSGE model of a small open economy we sudy he response of he economy and income ax rae, in paricular, o he reforms in he labour and produc markes. The model is non-ricardian due o he disorionary axaion and buil-in life-cycle feaures. We assume ha he wage markup and he price margins are reduced by one per cen each. Boh consumpion and employmen increase permanenly. The public secor balances improve, allowing for roughly 1 percenage poin cu in labour income axes. Produc marke reform leads o a shor-run reducion in consumpion, leading o an ineremporal radeoff in reform seing. More acivis fiscal policy can dampen his radeoff. Keywords: compeiion, dynamic general equilibrium, public finance JEL classificaion numbers: E60, C68 This is a subsanially revised version of he Bank of Finland s Discussion Paper 5/2006. The model presened in he paper relies on he Bank of Finland s Aino model, which is an oucome of a join projec by Juha Kilponen, Mika Kuismanen (Minisry of Finance), Ani Ripai and Jouko Vilmunen. We are also indebed o Juha Tarkka for inensive discussions and guidance during he various sages of model developmen. We are also indebed o paricipans of he CEPR/ESI 2005 conference and HECER Economics Research Seminar for many useful suggesions. The usual disclaimer applies. Correspondence o firsname.lasname@bof.fi or Bank of Finland, PO Box 160, FIN Helsinki, FINLAND. The views expressed are hose of he auhors and do no necessarily reflec he views of he Bank of Finland. 1

2 1 Inroducion The fuure challenges in all developed economies are closely conneced wih fiscal policy and populaion ageing, which pu pressure on boh he produc and labour markes. This paper sudies he response of he Finnish economy, and public finances in paricular, o he permanen produc and labour marke reforms. To sudy he quesion we build on he ideas of Blanchard and Giavazzi (2003), where produc and labour marke (de)regulaion are modelled as changes in he elasiciy of subsiuion of differen ypes of goods and labour. In preparing quaniaive resuls, we use he DSGE model of he Finnish economy. A similar exercise for aggregaed Europe has been done by Bayoumi e al. (2004a) and applied o he Danish economy by he IMF (2004) using a model varian of Global Economic Model (GEM), build a he Inernaional Moneary Fund by Bayoumi e al. (2004b). Jonsson (2007) sudies he welfare losses of imperfec compeiion and is relaionship o disorionary axaion. Their resuls are qualiaively similar o ours. However, significan differences arise from he fac ha, unlike he GEM model, our model includes disorionary axes and households ha are essenially non-ricardian. Moreover, our model depics he Finnish economy as a small open economy and reas he res of he world as given, while he GEM can be exended o a muli-counry seup, wih explici rade relaionships beween srucurally similar economies. Finally, we inerpre he reforms as specific o Finland, so ha EMUwide moneary policy does no respond o hese wo reforms. Our model has a non-sochasic balanced growh pah a which he economic growh is deermined by exogenously given growh of labour saving echnology and populaion. Accumulaion of financial asses and physical capial reflec opimal ineremporal decisions of households and firms. Moneary policy follows a fixed-exchange-rae regime, reflecing Finland s small size wihin he Eurosysem. The model conains exogenously deermined markups in he domesic goods markes and labour markes. The firs markup, given by he ime-varying elasiciy of subsiuion beween differen produc brands, can vary emporarily due boh exogenously given variaion and Calvo-ype price rigidiies. A similar srucure holds for he labour markes. A household s lifeime consiss of wo disinc periods. We label households living in hese wo differen periods as workers and reirees, as in Gerler (1999). Workers and reirees differ in erms of heir effecive planning horisons, marginal propensiy o consume, as well as he efficiency of heir labour effor. In he model, he public accouns are closed 2

3 by a labour income ax rule, which responds o he public defici as well as he deviaion of public deb from he exogenously given arge. The model is non-ricardian in hree ways: Firs, he model has disorionary axaion in he form of a labour income ax, indirec axes (VAT) and corporae (profis) ax. Second, households have finie life. Consequenly, hey discoun governmen deb a a higher-han-marke rae. Finally, in he second sage of life (reiremen) households suffer from a reduced abiliy o work. The age-profile of heir produciviy decreases sepwise, which affecs discouning of labour income. Reforms ha increase compeiion in boh he produc and labour markes are welfare enhancing in he long-run. Increasing compeiion leads o increased consumpion, invesmens, employmen and producion poenial of he economy. However, increasing compeiion is associaed wih an iniial decline in privae consumpion and a sligh drop in labour effor. This is due o he wealh effec caused by he reducion in profis, as well as a emporary increase in he real ineres rae caused by he slowdown of expeced domesic inflaion. A similar phenomenon is called he ineremporal radeoff by Blanchard and Giavazzi (2003). This ineremporal radeoff provides a poenial poliical-economy explanaion of why he produc marke reforms seem o be difficul o implemen. Public secor finances are improved, in he sense ha he iniial public secor deb-o- GDP raio can be mainained a a lower labour income ax rae. Our sandard simulaion suggess ha a one per cen reducion in he price and wage markups respecively allows roughly a 1.1 percenage poin reducion in he wage income ax rae. The sensiiviy analysis shows ha he resuls appear mos sensiive o changes he labour supply parameer. I also demosraes ha he ineremporal radeoff may be dampened when combined wih acivis fiscal policy. The res of he paper is organised as follows. Secion 2 ses ou he model. Secion 3 presens he simulaion experimen in deail and discusses he resuls. The final secion concludes. 2 Life-cycle model Following Gerler (1999) a household s finie life-cycle consiss of wo disinc sages, acive working-age and reiremen period. Consumpion and labour supply decisions are affeced by he fuure prospec of reiremen, as well as he fac ha labour efficiency is assumed o decline in he reiremen sage. In paricular, he likelihood ha he worker will lose 3

4 par of his labour income due o reiremen, induces her o discoun he fuure income sream a a higher-han-marke rae. This reduces consumpion and increases saving. In his sense, acive working-age populaion saves for reiremen. Reirees discoun he fuure more heavily han acive working-age individuals, due o he consan periodic probabiliy of deah. Therefore, reirees propensiy o consume ou of wealh is greaer han ha of workers. In addiion o hese basic feaures from Gerler (1999), a few exensions have been made. Firs, we allow for a se of disorionary axes: labour income axes, sauory pension conribuions, indirec axes and corporae axes. Second, he labour markes are monopolisic, wih wage rigidiies ha arise from Calvo ype wage conracs. Third, individuals receive ransfers from boh he public secor (he sae) as well as from (sauory) pension funds. In modelling ransfers, we have followed he general feaures of he ransfer sysem described in he Finnish naional accouns. Finally, he model s supply side is based on consan elasiciy of subsiuion (CES) producion echnology wih facor augmenaion in he underlying echnological progress and monopolisic compeiion in he goods marke. This conrass wih Gerler (1999), which assumes compeiive markes and Cobb-Douglas echnology. 2.1 Households Consumers are assumed o be born as acive working-age individuals. Condiional on being an acive worker in he curren period, he probabiliy of remaining one in he nex period is ω, while he probabiliy of reiring is 1 ω. The ransiion probabiliy is independen of one s employmen enure, so ha he average enure of acive working age is 1/(1 ω). Once an individual has reired she faces a consan periodic probabiliy of deah, 1 γ. Also, given ha he survival probabiliy γ is assumed o be independen of reiremen enure, he average reiremen period is 1/(1 γ). Regarding populaion dynamics, in each period ( ˆN ω)n new acive working-age individuals are born, so ha he working-age populaion grows a he gross growh rae of ˆN N /N 1. Given consan probabiliies of reiremen and deah and ha cohors are large, he reiree populaion (N r ) evolves according o he equaion N r +1 = (1 ω)n + γn r. 4

5 The raio of reirees o he whole populaion, he old-age dependency raio, evolves according o ϕ N r N = 1 ω ˆN + γ ϕ 1 ˆN. (1) In he seady sae, where he populaion growh is a consan ˆN = ˆN, he old-age dependency raio is ϕ = (1 ω)/( ˆN γ). In consrucing he ineremporal decision problem, Gerler (1999) follows he recursive preferences inroduced by Epsein and Zin (1989) and modified by Farmer (1990). The preferences separae ineremporal elasiciy of subsiuion and relaive risk aversion. Assuming risk-neuraliy, he model soluion enails linear decision rules even wih idiosyncraic risk o income, asse reurn and lengh of life. This class of preferences is applied in he following equaion: {[ (C V i ) i ν = (1 l i ) 1 v] ρ c } 1 + β i [E (V +1 i)] ρc ρc, i = {w, r} (2) where E (V +1 w) = ωv w +1 + (1 ω)v r +1, β w = β (3) E (V +1 r) = V r +1, β r = βγ. V i denoes an individual s value funcion, and i = {w, r} indicaes wheher he individual is a acive working-age w or reired r. C i is consumpion and 1 l i denoes leisure. Thus, l i denoes he fracion of ime allocaed o work and parameer v is he elasiciy of he period uiliy funcion wih respec o consumpion. ρ c is he curvaure parameer, which implies a consan ineremporal elasiciy of subsiuion, σ = 1/(1 ρ c ). The reirees effecive discoun facor β r is adjused o ake ino accoun he probabiliy of deah, as finie life effecively implies a shorer planning horizon. A perfec annuiies marke is assumed in order o eliminae he impac of uncerainy abou ime of deah. An acive working age individual, in urn, faces a poenial risk of a decline in wage income due o reiremen. However, given he individual s preferences, only expeced fuure labour income affecs consumpion. A worker hus forms a cerainy equivalen of his random uiliy as shown in equaion (3). There are differen financial asses available for consumers: a one-period governmen 5

6 bond A S, one-period pension fund bond A P, one-period foreign bond A W and shares issued by domesic firms A F. A domesic one-period bond pays a nominal gross reurn of R = 1 + r, while he gross reurn on socks is R D 1 + r D. Similarly, a foreign bond pays a gross reurn of R W 1+r W. Thus, he gross nominal reurn on asse holdings of populaion group i = {w, r}, is R A i (1 + r S )(A is + A ip ) + (1 + r D )A if + (1 + r W )S A iw, i = {w, r} (4) where S ishe nominal exchange rae, r S denoes he one-period bond rae, and r W denoes he corresponding foreign one-period bond rae. Given profis Π D, he gross reurn is 1 + r D = [A F +1 + (1 K )Π D ]/A F, where K denoes corporae ax rae and profis 1 Π D are given by Π D = P Y W F L R K (inermediae goods producers) + R K P I I (capial renal firm) Reirees A periodic budge consrain of a reiree born a ime j,reiring a ime k, and surviving a leas unil + 1 is given by A rj +1 = R γ Arj + W (1 RS )ξl rj + T rjk P C C rj, (5) where A is he se of asses wih corresponding gross-reurn R from (4), W denoes nominal wage rae, RS reirees labour income ax rae, ξ he relaive efficiency of reirees labour effor (in workers labour unis), and T rjk denoes nominal ransfers (such as pensions) o an individual born a ime j and reiring in period k. The price level of consumpion goods are given by P C and he corresponding consumpion aggregae is C. A reiree chooses consumpion and asse accumulaion by maximising (2) subjec o (5). Since ime and risk aggregaors are linear homogeneous, reirees maximisaion problem can be urned ino a dynamic programming problem, where he consumpion-saving decision is separable from he porfolio opimisaion. Furhermore, aggregaion can be done by simply 1 See secion 2.4 for deails. 6

7 summing over he reirees, since he decision rules are linear and here is no heerogeneiy among he reirees. I can be shown ha he consumpion funcion for reirees is given by P c C r = ɛ π [R A r + H r + S r ] (6) where H r and S r denoe discouned afer-ax values of labour income and ransfers. Their recursive represenaion is given by H r = (1 RS )W ξl r + Hr +1 ˆNR +1 /γ, Sr = T r + T +1 r ˆNR +1 /γ. Since he populaion of pensioners grows a he gross rae ˆN and oal social securiy paymens are disribued equally among hem, ˆN eners he discoun facor for fuure social securiy ransfers. Fuure aggregae labour income is similarly discouned. The reirees marginal propensiy o consume ou of wealh, ɛ π, evolves according o following non-linear difference equaion: ( W /P c ɛ π = 1 W +1 /P+1 c (1 RS ) (1 RS +1 ) ) (1 v)ρc 1 ρc β 1 1 ρc ( R +1 ˆP c +1 ) ρc 1 ρc ɛ π ɛ +1 π +1 γ (7) where ˆP +1 c P+1/P c c. The reirees marginal propensiy o consume varies wih expeced real ineres rae R +1 / ˆP +1 c as well as wih he expeced changes in real ne wage income. As in he sandard Yaari (1965) and Blanchard (1985) models, he likelihood of deah, (1 γ) in (7), raises he reirees marginal propensiy o consume. This can be easily seen by considering a case of logarihmic preferences, where σ 1 (i.e. ρ C 0). In his case ɛπ = 1 βγ Workers As regards o workers, he budge consrain of a household of acive working-age (denoed by h) and born a ime s is given by A whs +1 = R A whs + (1 W S W P )W L ws + T ws P C C ws (8) and where T ws denoes ne fiscal ransfers o workers. As in he case of reirees, we assume ha hese ransfers are independen of workers age. A worker chooses consumpion, labour supply and asse accumulaion by maximising (2) subjec o (8) and o he consrains ha 7

8 become operaive once she reires. Ineremporal maximisaion gives rise o a raher complicaed Euler equaion, bu i can be shown ha he consumpion plan of acive working-age individuals can be aggregaed as P C C w = π (R A w + H w + S w ), (9) where π is he marginal propensiy o consume of an acive working-age individual and H w and S w denoe discouned human and social securiy wealh. Marginal propensiy o consume ou of wealh saisfies he non-linear firs order difference equaion (( ) 1 W S W P W /P C π = 1 W +1 /P+1 C ) (1 ν)ρc 1 ρc β 1 1 ρ ( Ω +1 R +1 ˆP C +1 ) ρc 1 ρc π π +1 where W S is he labour income ax rae and W P he pension conribuion rae. The facor Ω +1 ha weighs he gross real reurn R +1 / ˆP C +1 evolves according o ( Ω +1 = ω 1 1 W S +1 W P +1 ) 1 υ + (1 ω )ɛ 1 ρc ρc +1 ( 1 ξ ( ) 1 RS +1 ) 1 υ where RS +1 is reirees labour income ax rae and ɛ +1 > 1 is he raio of reirees marginal propensiy o consume o ha of he acive working-age individuals. H w in (9) is he discouned sum of he wage bill of acive working-age individuals and S w is he sum across workers alive a of he capialized value of social securiy. Boh of hese measures ake ino accoun corresponding discouned values a he ime of reiremen. Formally, H w = ( 1 W S S w = T w W P ) W L w + ω ( (ɛ +1 ) 1 1 ρc ( 1 ξ(1 RS +1) + (1 ω) R +1 Ω +1 ˆN ( ) 1 υ 1 (1 S w + ω W +1 S W +1 P ) +1 R +1 Ω +1 ˆN ( (ɛ +1 ) 1 1 ρc + (1 ω) 1 ξ(1 RS +1) R +1 Ω +1 ˆN 1 (1 W S +1 W P +1 ) ) 1 υ H w +1 R +1 Ω +1 ˆN ) 1 υ ϕ 1 H r(+1) +1 ) 1 υ ϕ 1 S r(+1) +1, (10) (11) 8

9 where H r(+1) +1 measures he aggregae value of human wealh for a working reiree who reired a ime + 1 bu was sill working a ime. Similarly, S r(+1) +1 measures he value of oal social securiy for a reiree who reired a ime + 1 bu was sill working a ime. The presence of Ω +1 > 1 in he denominaor of (10) (11) shows how workers discoun fuure income sreams a a higher rae han ha a which he governmen can borrow, R. This in urn has a endency o reduce a working-age individual s consumpion and increase saving. Ω +1 varies posiively wih reirees marginal propensiy o consume relaive o ha of acive working-age individuals. I also depends posiively on reiremen probabiliy and ax raes. This can be seen mos easily by looking a he seady sae value of Ω in he special case where reirees and acive working-age inviduals face he same labour income ax rae : ( Ω = 1 ξ(1 ) ) 1 υ [ ] ω + (1 ω)ɛ 1 1 σ. Moreover, noice ha in he special case of logarihmic preferences (σ 1) marginal propensiy o consume is consan, and i depends only on discoun rae β (π = 1 β) Disribuion of wealh and aggregae consumpion working-age persons and reirees differing marginal propensiies o consume are refleced in he rae a which he wo groups accumulae financial asses. As hese asses are accumulaed a differen raes, aggregae consumpion depends on how financial asses are disribued among he wo groups. In oher words, we need a sae equaion for he disribuion of financial wealh among he wo groups. Le λ f +1 Ar +1/A +1 be he share of financial asses held by reirees and 1 λ f +1 Aw +1/A +1 he share of financial asses held by working-age individuals. I can be shown ha reirees share of financial wealh evolves according o λ f +1 /ω = λf (1 ɛπ ν )R A A +1 ω + RS (1 τ )ξw N r + T r ɛπ ν (Sr + H r ) + A +1 (1 ω). ω 9

10 Finally, aggregae consumpion is obained simply by summing up (6) and (9) and using λ f +1 : C = π {[(1 λ f ) R A + H w + S w ] [ ]} + ɛ λ f R A + H r + S r The equaion for aggregae consumpions shows ha ransfers markedly influence he evoluion of he disribuion of wealh, which in urn affecs aggregae consumpion. Labour income axes affec consumpion direcly via he discouned lifeime human wealh and income ransfers, bu also indirecly hrough heir effec on labour supply and on he disribuion of asses beween reirees and acive working-age populaion. Given ha he working-age populaion discouns fuure income sreams a a higher rae han ha a which governmen can borrow, fiscal policy ha pospones ax increases ino he fuure booss consumpion in he shor-run. 2.2 Aggregae labour markes In he model, labour supply is deermined endogenously via households opimal decisions on consumpion and labour supply. Each individual has one uni of ime which he may use o work or o enjoy leisure. Reirees as well as hose of acive working-age may paricipae in he labour markes, bu reirees are less producive han acive working-age individuals. In addiion, he labour marke is imperfecly compeiive due o he wage seing power of workers. Moreover, in he shor erm, real wages can depar from he desired level, due o he slow adjusmen of nominal wages, reflecing he long duraion and overlapping naure of wage conracs. Following Erceg e al. (2000), we assume ha here exiss a labour-aggregaor ha uses he following consan-elasiciy-of-subsiuion (CES) producion funcion o aggregae differen working-age labour ypes (j): [ 1 L w = 0 L w (j) ρl dj ] 1 ρ L, where L w (j) denoes he demand for ype j workers. Cos minimisaion of he labouraggregaor implies ha he demand for ype j workers depends upon he relaive wage and 10

11 aggregae labour demand index: ( L w W (j) (j) = W ) 1 1+ρ L L w (12) where 1/(1+ρ L ) is he elasiciy of subsiuion beween differeniaed labour inpus. W (j) denoes he wage rae of a ype j worker and he aggregae wage, W, is defined as [ 1 W = W (j) 0 ρ L ρ L +1 dj ] 1+ρL ρ L. Following he now sandard approach in he lieraure, i is assumed ha only a fracion of ζ W randomly chosen workers can re-se heir wages in each period. For hose no able o opimise in period, he wage is mechanically adjused using he previous period s growh rae of wages (dynamic indexing). The behaviour of aggregae nominal wages is characerised by he following wo equaions: 2 W = 1 ρ L (1 v) P c C w / ( ) 1 W S W P v[(1 ϕ)n L w ] 2 log(w ) = β E 2 log(w +1 ) + (1 ζw )(1 ζ W β) ζ W [log(w ) log(w )] (14) (13) where ρ L is he inverse of he wage markup and N is populaion. The opimal wage rae W is direcly derived from he aggregae version of an acive working-age individual s labour supply decision, and aking ino accoun he individual s labour demand consrain (12). Reirees labour supply is deermined by he corresponding firs order condiion: (1 v) L r v P c C r = ϕn ( ). ξw 1 RS 2.3 Public secor, pension fund and fiscal rules The general governmen (public secor) is divided ino wo secors: sae (cenral governmen) and (sauory) pension funds. The sae axes as follows: labour income of workers a he rae W S reirees income a RS, firms social securiy conribuions a F S, corporae (profis) a K, and indirec axes a he rae C. Sae consumpion C S consiss of marke goods C SF, which are provided by he consumpion goods reailer (privae secor), and 2 I is worh noing ha we assume ha here exis sae coningen securiies ha allow equilibrium consumpion and asse holdings o be equal for all workers, despie heerogeneous wages and labour supply. Noe also ha he second equaion is obained afer log-linearizaion. 11

12 non-marke goods Y S produced by he public secor iself, using a simple linear producion echnology Y S = Λ S ξ G L S, where Λ S is a echnology facor. Public secor employees have differen produciviy level ξ G which can differ from ha of privae secor employees. Sae The sae also pays axable and non-axable income ransfers o boh working-age and reired individuals. In addiion, i issues one-period governmen bond A S ha pay a nominal reurn r. Each period, he following budge consrain holds (A S A S 1) (ne lending) = W S W L w + RS ξw L r (income ax revenues) + K Π (corporae income ax revenues) (15) + C P C C F (indirec axes) + F S W L (firms social securiy conribuions) P C C SF P O Y S (governmen consumpion) P I I S (governmen invesmen) T w T r (oal ne ransfers) r A S 1 (ineres paymens). Public secor revenues mus be in harmony wih expendiures on public consumpion and invesmen, income ransfers and ineres expendiure on public deb. This is ensured in he model by he fiscal policy rule W S = κ[(a S A S 1)/Y Ās (1 1/Ŷ)], (16) where Ŷ Y /Y 1 denoes he gross growh rae of privae producion, (A S A S 1)/Y is he fiscal defici expressed as a share of privae producion, Ā s is an exogenous arge for he cenral governmen deb raio, and κ is he fiscal rule adjusmen parameer which conrols he speed of adjusmen of he labour income ax rae o deviaions of public deb from is long-erm arge. In principle, he higher he value of κ, he more concerned is he sae abou balancing is budge. 3 3 See for insance Railavo (2004) for a discussion of alernaive fiscal policy rules and heir sabiliy properies. 12

13 Sauory pension funds The pension scheme in Finland is a defined benefi scheme in he sense ha pension benefis are no direcly dependen on workers conribuions o employmen pension schemes or on he yields of pension funds. The conribuion raes are adjused in for possible shorfalls in fund balances. Nearly all old age pensions are provided by employmen pension insiuions or naional pension insiuions closely conrolled by he sae. The Finnish sauory pension sysem is approximaely 20 per cen funded. Oherwise, i funcions as a decenralised pay-as-you-go (PAYG) sysem 4. The fac ha he pension scheme is defined benefi and parly funded moivaes reamen of pension funds as separae from he cenral governmen and modelling hem as having heir own flow budge consrains and budge balancing rules. Furhermore, we can consider he funded par of he pension sysem as conracual savings and he PAYG par as a ransfer from workers o pensioners. Accordingly, we assume ha he fund collecs pension conribuions and disribues pensions o reirees T P R. Pension fund also accumulaes financial asses A P according o he flow budge consrain (A P A P 1) (ne lending) = F P W [L F + ξ G L G ] (social securiy conribuions of employer) (17) + W P [W L w ] (social securiy conribuions of workers) T P R (oal ransfers paid o reirees) r A P 1 (ineres paymens) where F P is he employer s pension conribuion rae and T SW sae o workers ha are reaed as labour income, and finally T P R oher ransfers from pension funds o reirees. denoes ransfers from he denoes pensions and Furhermore, i is assumed ha pension funds have a long-run funding arge, expressed as a raio of financial asses (A P ) o oupu. This arge is evenually achieved by appropriaely adjusing he pension conribuion raes. This pension conribuion rule is similar o he fiscal rule (16). Moneary policy Moneary policy reflecs Finland s small share (1.6 per cen) in he euro area. Consequenly, he feedback from Finnish economy o euro area level is very modes. 4 There is also a naional pension scheme covering all ciizens, bu is role is diminishing. A he same ime, volunary pension schemes and parly ax deducible are becoming increasingly popular. 13

14 A reasonable approximaion is ha he euro area policy rae and he foreign exchange rae are exogenous for he Finnish economy. We also assume ha he exchange rae is fixed, i.e. S = S Firms and echnologies The supply side of he model is essenially based on a single good. This is an inermediae good ha is a consan-elasiciy-of-subsiuion (CES) aggregae of a coninuum of brands. The domesic inermediae good is combined wih he impored good o obain hree differen final goods: consumpion goods, capial goods and expored goods. These final goods differ wih respec o ype of impored facor and elasiciy of subsiuion beween domesic and impored facors. Domesic producers of inermediae producs purchase heir capial inpus (capial services) in a compeiive capial marke (from companies providing capial services) in which capial is freely for sale and ransferable for use by oher companies. The domesic inermediae goods producers operae in monopolisic produc markes due o he imperfec subsiuabiliy of heir producs, and prices are sicky Domesic inermediae goods producers The domesic composie inermediae good, Y, is produced according o he following consan elasiciy of subsiuion (CES) producion funcion, which combines individual goods Y (j) (Dixi and Sigliz (1977)): [ 1 Y = 0 Y (j) ρz d j ] 1 ρ z The parameer ρ z [ 1, ) deermines he elasiciy of subsiuion, 1/(1 + ρ z ). For nonposiive values of ρ z he inermediae goods are gross subsiues. Perfec subsiuabiliy, and hence perfec compeiion, is obained by leing ρ z approach 1, so ha in his case he elasiciy of subsiuion approaches infiniy. We allow for ime variaion in he elasiciy of subsiuion. Cos minimizaion implies he condiional demand funcion for he individual good j (j [0, 1]) ( P (j) Y (j) = P ) 1 1+ρ z Y (18) 14

15 and he price index for he composie domesic inermediae good [ 1 ρ z 1+ρ P = P (j) z d j 0 ] 1+ρz ρ z (19) Domesic inermediae goods, Y (j), are produced by producers who face monopolisic compeiion. They ake he producion echnology and facor augmening echnical rends as exogenously given. The producion funcion is of he CES ype and akes he specific form of consan-reurns-o-scale: 5 Y (j) = [ δ ( Λ K ) ρ ( K + (1 δ) Λ L L F ) ρ ] 1/ρ (20) The facors of producion include homogeneous capial services 6, K, and homogeneous labour, L F. Λ K and Λ L denoe ime-varying 7 capial and labour-augmening echnical progress, which are unobservable o he economerician and are common o all firms. The elasiciy of echnical subsiuion is given by 1/(1+ρ), where ρ is he subsiuion parameer in he producion funcion. δ is he share parameer. Cos minimizaion implies he following real marginal coss: MC (j) P (j) = [ ( ) ρ ( ) ] δ 1 R 1+ρ 1 1+ρ + (1 δ) Λ K 1+ρ W F ρ 1+ρ ρ 1+ρ P (j) Λ L, (21) P (j) where R denoes he nominal renal price of capial services and W F = (1 + F P + F S )W represens nominal labour coss. Firms pension and social securiy conribuions are denoed by F P and F S. Price level, P (j) = Υ MC (j), is deermined by he markup Υ 1/ρ z over marginal coss. Noe, ha he markup is no uniy in he seady-sae case since he seady-sae elasiciy of subsiuion 8 beween inermediae goods is generally finie. The firs order condiions wih respec o labour and capial services are given by he log-linear equaions log δ υ ρλ K + (1 + ρ)(y k ) = r p (22) log(1 δ) υ ρλ L + (1 + ρ)(y l ) = w F p (23) 5 According o Ripai and Vilmunen (2001) and Jalava e al. (2006) his seems o be a reasonable assumpion for he Finnish daa. 6 Capial is rened from capial renal firms, leasing firms. 7 We do no specify heir sochasic properies a his sage. See Ripai and Vilmunen (2001) for furher discussion of heir properies and esimaes using aggregae Finnish daa. 8 The elasiciy of subsiuion is Υ/(Υ 1) in erms of Υ. 15

16 where r is he log of he nominal renal rae of capial services, w F = W /(1 F S F P ) he log of nominal wages (compensaion per employee), and p he log of he oupu price. Due o he monopolisic compeiion in he marke for oupu, he ime-varying slope of he demand curve, υ log(υ ), eners o he firs order condiions. Noe also ha λ log Λ for each ype of λs in he model. The dynamics of he price level P (j) of producer j arises from he assumpion ha a firm changes is price level when i receives a random price-change signal Calvo (1983). The probabiliy of receiving a price change signal is given by 1 ζ (ζ [0, 1]). Since here is coninuum of inermediae producers, 1 ζ also represens he share of producers ha have received such a signal and hus had an opporuniy o change heir prices. The average ime beween price changes is given by 1/(1 ζ). For hose firms ha do no receive a price-change signal in period, he price level is mechanically adjused using he period 1 inflaion rae (Alig e al., 2005). Assuming symmery of he firms, we obain he following log-linearized, aggregae pricing equaion for he inermediae goods producers: p = M E p +1 + (1 ζ)(1 ζm) ζ [υ + mc p ]. (24) Producer price inflaion is hus deermined by expeced producer price inflaion and changes in he slope of he demand curve (ime-varying markup) and real marginal coss Capial renal firms Capial is a homogeneous facor of producion ha is owned by a firm ha rens capial o producers of domesic inermediae goods. I operaes under perfec compeiion. The capial renal firm may choose beween physical capial accumulaion K p or a higher uilizaion rae U, wih K = U K p 1 (U [0, 1]). Physical accumulaion generaes real adjusmen coss in he form of los capial sock, whereas he capial uilizaion rae affecs he depreciaion of he capial sock. Capial accumulaion is given by K p + γ 1 2 ( K p γ 2 K p ) 2 1 K p 1 } {{ } a K (K p,kp 1,Kp 2 ) ( = K p 1 1 δ 0 + δ ) 1 U 1+δ2 1 + δ 2 + I (25) }{{} D(U ) 16

17 where a K ( ) denoes he adjusmen coss of he physical capial sock. The depreciaion facor D(U ) (D( ) [0, 1]) is an increasing funcion of he capial uilizaion rae, D (U ) > 0 (see Baxer and Farr, 2001). The capial renal firm maximizes is expeced discouned profis max E {U }{I } s=0 M,+s Π K +s subjec o he capial accumulaion equaion (25) and he definiion of capial services. The momenary profis are given by Π K =R K P I I =R U K p 1 P I {K p + a K (K p, K p 1, Kp 2 ) Kp 1 [1 D(U )]} The price index of invesmen goods is he price index of he domesic invesmen good reailer, P I. Since he firm is owned by households, he fuure profis are discouned using he nominal sochasic discoun facor (pricing kernel) M,+s. The opimal level of capaciy uilizaion is given by he following firs order condiion w.r.. U : R P I = δ 1 U δ2 (26) which relaes renal price o marginal depreciaion of capial sock. The firs order condiion w.r.. he capial sock K p is given as [ P I K p γ 2 K p ] 1 E 1 + γ 1 K p 1 + E M,+1 {R +1 U +1 P I +1 [ γ 1 (1 + γ 2 ) Kp +1 γ 2 K p ( K p +1 γ γ ( )]} 2 K p ) (K p ) 2 1 δ 0 δ 1 U 1+δ δ 2 K p E M,+2 P I +2γ 1 γ 2 K p +2 γ 2 K p +1 K p +1 = 0. (27) Due o he end-of-period iming of he physical capial sock, accumulaed physical capial is in use in he following period. Hence, i is he expeced following period s renal rae ha governs he curren period invesmen decision. 17

18 2.4.3 Domesic reailers The economy is inhabied by hree reailers, i.e. final goods producers. The firs one specializes in consumer goods (j = C), he second one in capial goods (j = I), and he hird one in expored goods (j = X). They combine domesic inermediae inpu produced by he inermediae goods producers and secor specific impored goods and services and operae under perfec compeiion. This means ha hey do no produce any value-added and can be considered as aggregaors ha reflec he way consumers or he capial renal firm (and public secor) subsiue beween domesic and foreign inermediae goods and services. The producion srucure of reailers is similar in all hree cases. Wha varies is he elasiciy of echnical subsiuion and he ime-pah of facor-augmening echnical changes. The echnical changes are assumed o be consan in he balanced-growh-pah, bu here may be shifs during he ransiion. Their developmens affec he relaive prices of final goods. The CES aggregaors are [ j = δ j ( Λ jy ) ρ j ( Y j + (1 δ j ) Λ jm M j ) ρ j ] 1/ρ j, j = {C, I, X}. Condiional facor demands are of he form Y j = ( δ j) 1 1+ρ (Λ j jy M j = ( 1 δ j) 1 1+ρ (Λ j jm and he price index is given by P j = ( δ j) 1 ( P 1+ρ j Λ jy ) ρ j 1+ρ j ( P P j ) ρ j 1+ρ j ( ) 1 1+ρ j j ) ρ j 1+ρ j + ( ρj 1 δ j) ) P Mj 1 1+ρ j P j j j = {C, I, X} 1+ρ j ( P Mj Λ jm ) ρ j 1+ρ j ρ j +1 ρ j. j = {C, I, X} According o our esimaion, i urns ou ha in he producion of consumpion goods and capial goods, he inpus are gross subsiues and in he producion of expors, he domesic and foreign inpus are gross complemens (he elasiciy of subsiuion is less han uniy). Noe also ha consumpion goods form he ax base for indirec axaion. In his case he price index P C in he above formulas should be replaced by he indirec ax rae facor (1 C )P C. 18

19 2.4.4 Imporing firms In deriving a model for impor prices, we follow he approach derived by Bes and Devereux (1996 and 2000) and applied o Finnish aggregae impor daa by Freysäer (2003). We assume ha a fracion of imporers price heir produc in local (Finnish) currency and he res in producer (heir own) currency. Their pricing conains idenical fricions in he form of Calvo (1983), i.e. hey may change heir price only in he even of a random price-change signal. And heir marginal coss are idenical. Aggregaion of prices over hese wo ypes of imporers yields an impor-price Euler equaion where impor prices depend on expeced fuure impor price inflaion and curren and expeced fuure changes in foreign exchange raes, and on he real marginal coss of imporers. To model a delayed pass-hrough of foreign exchange raes, we assume ha a fracion ω j of firms in each final goods secor, j {C, I, X}, price heir producs in he Finnish currency, he res price in heir own producer currency. Each of hese wo ypes of firms have Calvo price rigidiies (price change probabiliy 1 ζ j, j {C, I, X}) wih dynamic indexing (see, for example, Alig e al., 2005). The aggregaion of he wo ype of firms leads o he following pricing scheme 2 p Mj =R E 2 p Mj +1 + (1 ζj )(1 ζ j R ) ζ j (s + mc j p Mj ) + (1 ω j )( 2 s R E 2 s +1 ), j = {C, I, X} where s is he nominal foreign exchange rae in domesic currency (e.g. EUR/USD), mc j he marginal cos for imporing firms, p Mj he impor price for secor j. The parameer R denoes he foreign seady sae discoun facor, and 1 ζ j he probabiliy of reopimising prices. 2.5 Marke equilibrium All markes are in equilibrium a every poin of ime. The capial goods marke is in he equilibrium if he supply of capial services by he capial renal firm equals he demand for capial services by inermediae goods producers. Similarly he demand of labour equals is supply. There is an exra complicaion because labour supply and labour demand are measured in differen efficiency unis, he former (L LW ) in erms of working-age populaion 19

20 and he laer (L LF ) in erms of privae secor employmen: L LW = L W + ξl R (28) L LF = L F + ξ G L G. (29) In erms of number of employed, labour supply equals labour demand, i.e. L S = L D as follows: L S = L W + L R = L LW + (1 ξ)l R (30) L D = L F + L G = L LF + (1 ξ G )L G. (31) In he inermediae goods secor, he demand for inermediae goods by reailers and exporers ogeher equals oal supply: Y C + Y I + Y X = Y. (32) Sock markes clear when he supply of shares equals he demand for shares. Markes for consumpion goods clear when governmen purchases C SF plus privae consumpion C equals he supply of consumpion goods C T C SF + C = C T (33) and he markes for capial goods clear when public I S plus privae I invesmens equals o he supply of capial goods I T : I S + I = I T. (34) Expor demand and supply are given by ( P X S P W ) ρ W M = X, (35) where he lef-hand-side of equaion (35) represens world demand for expors and M and P W are aggregae res-of-he-world impors and he uni price in erms of foreign currency. When marke clearing condiions (28) (35) hold workers and pensioners budge 20

21 consrains, (8) and (5), he general governmen budge consrain (15), and pension fund s budge consrain (17) imply he following accumulaion of (ne) foreign asses: S A W = (1 + r F )S A W 1 + P X X P MR M R P MC M C P MI M I }{{} rade balance (36) where he lower line defines he rade balance. To close he model we specify sochasic processes for he exogenously given variables: ˆΛ L, ˆN, Υ, Λ K, Λ CY, Λ CM, Λ IY, Λ IM, Λ XY, Λ XM, Λ G, I S, C SF, T w, T r, Y S, K, F P, F S, C, T P RT, r F, S, Pˆ W, M. For each variable, we provide a univariae firs order auoregressive equaion of he form x = (1 ρ x ) x + ρ x x 1 + ε x, ε x IID(0, σ 2 x) (37) where x is one of he variables above. An imporan par of he calibraion exercise is calibraion of he seady-sae values of he exogenous variables. 2.6 Model calibraion The parameers affecing he demographics is calibraed o approximaely fi he demographic srucure in he near fuure, where he reirees share of he whole populaion, here defined as individuals of age years, is roughly 25 %. Table 1 gives he implied probabiliy of reiremen and deah. Corresponding reiremen and acive working-age periods are hen roughly 13 and 48 years. Annual ne growh rae of populaion is se o 0.16 percen. These demographic assumpions reflec Finland s siuaion over he nex decade, according o official demographic projecions. In order o fi he paricipaion raes close o averages, we se he relaive efficiency, ξ, of reirees a 32% of ha of acive working-age. The elasiciy of periodic uiliy wih respec o consumpion, ν, is se a 0.844, and he ineremporal elasiciy of subsiuion, 0.4, is based on he esimae by Viianen (2002). These parameer values are summarized in able 1. We calibrae he seady-sae produc marke markup o This is low compared o he esimae of Marins e al. (1996), which finds a manufacuring secor markup of

22 Table 1: Parameer values Parameer Explanaion Value v Consumpion share in uiliy σ Iner-emporal subsiuion 0.4 γ Probabiliy of surviving (12.5 years) ω Probabiliy of remaining worker (48 years) ξ Relaive labour efficiency of reirees 0.32 ˆN Populaion growh rae, p.a. 0.16% M = R Discoun rae /4 γ 1 Capial adj. cos (level) 300 γ 2 Capial adj. cos (lag) 0.95 δ 0 Rus and dus in depreciaion δ 1 Wear and ear in depreciaion 0.03 δ 2 Wear and ear curvaure 4.5 1/(1 + ρ) Subs. in inermediae goods 0.58 δ corresponding share parameer 0.1 ζ Calvo parameer in goods marke 1 1/6 ζ W Calvo parameer in labour marke 1 1/8 Υ Seady-sae produc markup /ρ L Seady-sae wage markup /(1 + ρ C ) Subs. in consumpion reailer 2.5 δ C corresponding share parameer /(1 + ρ I ) Subs. in capial reailer 2.2 δ I corresponding share parameer /(1 + ρ X ) Subs. in exporer 0.45 δ X corresponding share parameer 0.51 ζ C Calvo, impored consumpion 0.88 ω C Share of LCP firms in above 0.6 ζ I Calvo, impored capial goods 0.95 ω I Share of LCP firms in above 0.3 ζ X Calvo, impored maerials 0.6 ω X Share of LCP firms in above 0.9 ρ W Expor price elasiciy 1.24 Their mehod, however, differs markedly from ours. 9 Kilponen and Sanavira (2004) find ha he average price-cos margin in Finnish indusry is roughly 8 per cen, based on microdaa from annual Indusrial Saisics surveys ha cover essenially all Finnish manufacuring plans employing a leas 20 persons. The labour marke markup is calibraed o 1.30, o mach he labour share and employmen rae in recen Finnish daa. This wage markup is roughly he same as ha observed in Europe on average. The elasiciy of echnical subsiuion beween capial and labour follows he esimae by Jalava e al. (2006), The elasiciy of subsiuion beween domesic inermediae 9 Anoher reason for a relaively small esimae of he seady-sae markup resuls from he model consrucion: The profi funcion of he capial renal firm is nonlinear, and i yields posiive seady-sae profis even in he case of perfec compeiion. Hence a significan par of he privae secor profis are generaed by ha firm insead of by inermediae goods producers. Therefore our esimae is enaive bu consisen wih he daa and he parameer values in he model. 22

23 goods and impored consumpion goods is calibraed o 2.5 and is marginally higher han ha of he capial goods reailer, 2.2. For expors, he elasiciy of subsiuion is esimaed a In order o illusrae how well he model accords wih he recen daa, we use he daa from and calculae averages of several macro economic variables. The reason for no using a longer ime span is ha Finland experienced major srucural changes during he 1990s recession whereas we wan o fi he balanced growh pah o he more recen economic environmen. Table 2 gives summary saisics of hose macro economic variables. The model s implied seady sae porrays he expeced demographic change in he forhcoming decades and herefore differs somewha he mos recen daa. In paricular, here is a higher dependency raio, implying a bigger privae consumpion share, refleced in a larger impor share. Table 2: Seady sae shares and he daa Variable The daa Seady- (% of privae producion) ( ) sae Impors Expors Consumpion Privae Public Invesmens Privae Public Employmen rae, % Labour share, % I/K, % Dependency raio, % Income ax rae, % Pension conribuions, % Produc and labour marke reforms Gersbach (2000) 10 classifies possible mechanisms of labour marke reforms ino hree caegories: direc effec of lower markup, simulaion of echnical change, and expanded se of produc varieies. The seminal aricle by Blanchard and Giavazzi (2003) sudies he poliical economy of produc and labour marke reforms in a model ha combines he firs and he hird of he above mechanisms. A novel feaure of heir model is ha he produc marke markup, which is relaed o he number of firms, is endogenously deermined by an enry 10 See also Høj and Wise (2004). 23

24 condiion. The very same feaure in heir model also creaes a link beween labour and produc markes. In heir model produc marke deregulaion leads o higher real wages, and o lower unemploymen in he long run. This resul has been challenged by Specor (2004) who, by imposing decreasing reurns o labour (producion funcion), shows ha real wages may fall even in he long-run. Building on a model wih labour marke maching model and efficiency wages, Amable and Gai (2004) demonsrae ha produc marke deregulaion may lead o employmen losses even in he long run. Ebell and Haefke (2003) also builds on a maching model and, due o hiring exernaliies, finds ha quaniaively he produc marke deregulaion has a small effec. The relaionship beween degree of produc marke compeiion and echnical change has been an imporan par of he research agenda of Aghion and Howi 11. This is quaniaively poenially he mos imporan channel. Their argumen ges some empirical suppor from Nicolei e al. (2001) and Kilponen e al. (2004). The heoreical linkages of labour and produc marke reforms, as se ou by Blanchard and Giavazzi (2003), also ge empirical suppor from Nicolei e al. (2001) and Jean and Nicolei (2002) 12. Our approach differs from he above in hree imporan dimensions: Firs, boh produc and labour marke markups are exogenously given. For his reason, here is no link from markup o echnical change. Second, in our model, agens opimize ineremporally and due, o nominal and real rigidiies, our model is able o analyze dynamics along he ransiion pah. Finally, as emphasized in he previous secion, he model is non-ricardian and so includes a role for fiscal policy 13. Here, produc marke reform refers o an increase in he elasiciy of subsiuion of differen produc brands. The degree of produc marke compeiion in he model is deermined by he ime-varying parameer Υ 1/ρ z in he pricing equaion (24). I is he inverse of he subsiuion parameer in he aggregaor of domesic inermediae goods (18). Less han perfec subsiuabiliy beween goods may capure effecs such as compeiion regulaion, horizonal collusion in produc markes, public ownership of domesic firms, differences in produc sandards, ec. The imperfec compeiion is limied o he inermediae goods secor, since final goods producers, including exporers, operae under perfec compeiion 11 See, for example, Aghion and Howi (1994) Aghion e al. (2005) Aghion and Howi (2006). 12 The measure of wage premia in he above-menioned sudies is based on he characerisics of workers, working condiions and firms. Therefore, i does no direcly correspond o our labour marke markup measure. 13 Bokan and Hughes Halle (2006) enrich he Blanchard and Giavazzi (2003) model by wih labour income and payroll axes. 24

25 in heir produc markes. 14 An example of his kind of reform was he change in Finnish legislaion in 2003 ha forced mobile phone operaors (carriers) o se up a sysem where he clien could keep he phone number afer swiching o anoher operaor. Mobile phone call prices declined subsanially afer he reform and conribued (ogeher wih a decline in handse prices) 0.3 percenage poin o he annual inflaion rae. 15 (Bank of Finland, 2005) The emergence of ne-sores provides anoher example ha can be inerpreed as an increase in he elasiciy of subsiuion. Similarly, we use labour marke reform o refer o an increase in he elasiciy of subsiuion of differen labour ypes. The degree of imperfecion in he labour markes is given by he process for ρ L in equaion (13). 1/ρ L deermines he premium over he marginal rae of subsiuion beween consumpion and leisure, which is he relevan measure of he marginal cos of changing labour by one uni. I capures facors like he bargaining power of labour unions, he unionisaion rae, minimum wage legislaion, unemploymen benefis, hiring and firing coss, immigraion policies ec. In his sense, he parameer can be hough of as capuring he essenial non-compeiive feaures of he wage seing process. In he following simulaion exercise we assume an unanicipaed insananeous decline of he markup process by one per cen. This resuls o a new seady-sae. The persisence of he shock is known by agens. The model is linearized around he erminal ( afer-shock ) seady-sae. 16 Due o his choice of shock srucure, he resuling dynamic pah porrays inrinsic dynamics of he model. Høj and Wise (2004) lis in deail he possible resricions o produc marke compeiion in Finland. They also provide esimaes of he macroeconomic effecs of increased compeiion. According heir esimae, based on empirical work by Nicolei e al. (2001), if Finland moved owards bes pracice for produc marke liberalisaion in he OECD, hen he employmen rae could increase by anoher 1/4 1/2 percenage poin (page 36). 3.1 Responses o produc and labour marke reforms. We sudy he responses of he economy o produc and labour marke markup shocks, boh separaely and combined. Table 3 repors he new seady-saes relaive o he iniial seady-sae repored in able 2, and figure 1 whows he dynamic (firs 20 years) responses. 14 The GEM model of Bayoumi e al. (2004b) allows for imperfec compeiion in various componens of final goods and is able o make his breakdown. 15 This is surprisingly large number given quaniaive inflaion responses of our model. 16 The simulaion is carried ou wih he forecas command of Dynare (Juillard, 2003) and crosschecked wih deerminisic simulaions (simul command) of Dynare 4. 25

26 Since he model is close o linear, he responses of combined shocks are close o he sums of individual shocks. In each case he size of a shock is one per cen from he iniial seady-sae value. This means ha he produc marke markup declines from o and ha of he labour marke from 1.30 o Consequenly, he labour markup shock is marginally larger, in percenage poins, han he produc markup shock. The reducion in produc marke markup resuls an increase in facor demands ha raises he real wages, employmen, and capial sock. Due o he decrease in domesic price level and he fixed exchange rae regime, he real exchange rae depreciaes. The reducion in price markup raises he marginal produc of capial and leads o a higher capial-oupu raio. Consequenly, he consumpion mus give room o invesmens o reach he new capial-oupu raio. Household financial wealh will decline due o he reducion in firms profis and hus in he value of equiy. Deceleraion of consumer price inflaion booss he real ineres rae, furher iling he consumpion profile. Increased real wages and employmen operae in opposie direcions and consumpion exceeds is original seady sae level afer 10 years. Fiscal policy plays an impor role in he ransmission of he price markup shock. Labour income ax revenues increase due o he increase in employmen and real wages. This is, however, no enough o balance he budge, due o he significan decline in indirec and corporae ax revenues. The budge is driven ino defici. The budge closure rule implies an adjusmen in he labour income ax ha increases ax disorions. Pension scheme financing relies solely on labour income axes. Therefore, he pension conribuion rae monoonically declines. Due o he ulimae reducion in disorions (markup), he permanen response of consumpion o produc marke deregulaion is posiive. Given his, here is a permanen improvemen in he fiscal balance. This gives room for lowering he labour income ax rae and pension conribuion rae by 0.5 and 0.4 percenage poins respecively. A one per cen shock o he labour marke markup resuls in responses ha are only 1/3 of he magniude of he produc markup shock. Responses also porray subsanial differences. As described in secion 2.2, working-age persons use heir monopoly power in he wage-seing. Their labour supply elasiciy is smaller han ha of reirees (see discussion in nex secion). A reducion in he labour marke markup shifs he supply of labour downwards, producing a rise in employmen and a reducion in real wages (see able 3). Conrary o he produc marke reform, he wealh effec is small due o he modes increase in he value of he capial sock. Consumpion increases. Improved public finances enable an easing of he ax burden by 0.17 and 0.05 percenage poins. 26