Tax Reforms and the Underground Economy: A Simulation-Based Analysis

Transcription

1 Tax Reforms and he Underground Economy: A Simulaion-Based Analysis Barbara Annicchiarico Claudio Cesaroni Sepember 2016 Absrac This paper sudies he effecs of several ax reforms in an economy where axes are parially evaded by means of undeclared work. To his purpose, we consider a wo-secor dynamic general equilibrium model calibraed o Ialy which explicily accouns for underground producion. We consruc various ax reform scenarios, such as ex-ane budge-neural ax shifs from direc o indirec axes, and ax cus on labor and business financed by decreases of governmen spending. Our resuls indicae ha neglecing he exisence of he underground secor may lead o severely miscalculae he macroeconomic effecs of ax reforms. Furher, he dimension of he underground secor is permanenly and considerably reduced by changes in he ax mix ha diminish he labor ax wedge. Reducions of he business ax prove o be highly expansionary in he presence of a sizable underground secor. Keywords: Dynamic General Equilibrium Model, Underground Economy, Tax Reforms, Ialy J.E.L. Codes: E62, O41, O52 We hank he edior and wo anonymous referees for heir excellen commens. This paper has also benefied from he commens and suggesions of Amedeo Argeniero, Andrea Cosa, Fabio Di Dio, Edgar L. Feige, Giammario Impullii, Elisabea Marzano, Libero Monefore, Alessandra Pelloni, Lorenza Rossi, seminar paricipans a he Universià degli Sudi di Roma Tor Vergaa and paricipans a he Shadow Conference 2015, Universiy of Exeer and a he 56h Annual Conference of he Ialian Economic Associaion, Universià Parhenope, Naples. Corresponding auhor: Diparimeno di Economia e Finanza, Universià degli Sudi di Roma Tor Vergaa, Via Columbia 2, Rome, Ialy; barbara.annicchiarico@uniroma2.i. Diparimeno di Economia e Finanza, Universià degli Sudi di Roma Tor Vergaa, Via Columbia 2, Rome, Ialy; claudio.cesaroni@uniroma2.i. 1

2 1 Inroducion This paper seeks o undersand he imporance of ax evasion in deermining he economic impac of ax reforms. Clearly, ax reforms should be carefully designed aking ino accoun he exising ax sysem of he economies involved along wih heir producion srucure. From his perspecive he presence of a non-negligible share of underground producion may subsanially influence he effecs of any reform aimed a reducing ax disorions. Indeed, ha beween a counry s ax sysem and he size of is shadow economy is a wo-way relaionship. On he one hand, here exiss a posiive relaionship beween he dimension of he ax burden on economic aciviy and he size of he informal economy. 1 See he evidence provided by Schneider (1997), Cebula (1997), Johnson e al. (1998), Giles e al. (2002), Dell Anno e al. (2007), Buehn and Schneider (2012), Giles and Caragaa (2001), Tanzi (1999), Schneider and Ense (2000), and Feld and Schneider (2010). On he oher hand, jus as he ax burden conribues o deermine he size of he shadow economy, he shadow economy iself affecs he oucome of fiscal reforms. In fac, srong reallocaion effecs beween he underground and he formal secors may be riggered by changes in he ax srucure, shaping he overall macroeconomic performance of ax reforms and he effecs on fiscal variables. In his regard, Busao e al. (2012) show how ax evasion and he exisence of underground economy miigae ax disorions and aenuae he response of oupu o resricive or expansionary ax changes; Fugazza and Jacques (2004) inroduce an underground secor ino a coninuous-ime search and maching model and show how he presence of an informal secor is likely o affec he impac of policy inervenions, boh on he labor force and on he governmen budge; in a wo-secor general equilibrium model Turnovsky and Basher (2009) examine he role of he informal economy in limiing he abiliy of he governmen o raise ax revenues. Saring from he work by Busao e al. (2012), he aim of his paper is o explore he effecs of a richer se of ax reforms aimed a promoing growh, while preserving fiscal soundness, in an economy characerized by underground producion, undeclared work and imperfec compeiion. 2 1 Of course here are oher facors playing an imporan role, such as he possibiliy of ax avoidance, which affecs he exen of ax evasion and, herefore, he size of he underground economy. See, e.g., Neck e al. (2012). 2 Along his line of research, see Felensein and Shamloo (2013). 2

3 Despie mos of he EU Member Saes in economic and fiscal disress, such as Greece, Ialy, Porugal and Spain, have high shares of shadow economies, as documened by Schneider (2012), his essenial aspec of heir producion srucure is overlooked in assessing he macroeconomic impac of changes in he ax and spending mix. 3 In his paper we seek o fill he gap by means of a wo-secor dynamic general equilibrium model calibraed o Ialy which explicily accouns for underground producion and embodies a rich se of axes. Specifically, we conduc our analysis in an arificial economy where he underground secor includes all marke-based legal producion of goods and services ha are deliberaely concealed from public auhoriies o avoid paymen of axes and social securiy conribuions or o escape cerain legal marke sandards and adminisraive obligaions (Schneider, 2012, p. 6). In paricular, we sar from he dynamic sochasic general equilibrium model of Orsi e al. (2014), which, in urn, represens an adapaion of he model of Busao and Chiarini (2004), and incorporae inernaional rade, nominal rigidiies and imperfec compeiion, along wih real rigidiies, such as invesmen adjusmen coss and habi persisence. We also inroduce axes on consumpion, labor income and business as well as social securiy conribuions (SSC) borne by employers and workers. We are so able o consruc various ax reform scenarios, recenly advocaed in economic and policy circles as a means o promoe growh, such as ex-ane budge-neural ax shifs involving reducions of disorionary axes on labor and business compensaed by an increase in he consumpion ax or counerbalanced by decreases of governmen spending. 4 We will see ha all hese fiscal reforms give rise o a resource reallocaion from underground o official producion or vice versa and have raher differen implicaions on he level of economic aciviy. 5 The ineres of he presen analysis on he Ialian economy is moivaed by wo main reasons. 3 A laudable excepion is given by he recen conribuion of Pappa e al. (2015) who sudy he effecs of fiscal consolidaion accouning for he presence of ax evasion and corrupion. In heir analysis conduced for Greece, Ialy, Porugal and Spain hey unambiguously show ha he recen consolidaion plans inroduced in hese economies end o increase ax evasion. 4 In doing so we enrich he analysis of Busao e al. (2012) who focus on he effecs of ax reducions on labor and income, wih public spending being deermined endogenously so as o preserve he balance of he public budge consrain a all imes. 5 In his respec, hrough he use of a macroeconomic model embodying an underground secor, his paper conribues o he vas lieraure sudying he effecs of ax disorions and measuring he impac of ax reforms involving ax shifs or changes in public spending me by adjusmens in axes (see e.g. Baxer and King 1993, Mendoza and Tesar 1998, Leeper and Yang 2008). 3

4 In he firs place, he Ialian labor marke is characerized by high ax wedges. In 2014 Ialy recorded a oal ax wedge as a percenage of labor coss of 48%, wih he share of firms social securiy conribuions accouning for more han 50% of his wedge. The average ax wedge for he OECD counries is abou 36%. 6 A he momen several ax reforms are underway in Ialy, aimed a reducing he ax wedge on labor and a simulaing economic aciviy, while preserving fiscal soundness. In his respec, our analysis aemps o quanify he poenial gains for he economy from ax reforms acing in his direcion. Furhermore, Ialy is characerized by he exisence of a sizable underground secor. Mos recen available esimaes of he Ialian Naional Insiue for Saisics (ISTAT) poin o he shadow economy being 20.6% of GDP, 7 Orsi e al. (2014) find ha he share of he underground secor is abou 23%, he esimaes of Schneider (2013) indicae ha he size of he shadow economy in Ialy is 21.1% of GDP in 2013, while Ardizzi e al. (2014) esimae a quie lower, bu sill relevan, value (16.5% of GDP) using he currency demand approach. 8 This aspec is paricularly significan when sudying he effecs of ax reforms able o change he incenives o supply and/or employ informal labor inpus. Our main resuls can be summarized as follows. Firs, consisenly wih previous findings, we show ha neglecing he exisence of he underground secor may lead o severely miscalculae he macroeconomic impac of ax reforms. This resul arises since a share of producion or worked hours, depending on he specific policy pu in place, being kep secre o avoid axaion, is no direcly affeced by he ax reform. In addiion, he exisence of he underground secor offers households and firms he possibiliy of shifing from one secor o anoher in response o changes in he ax mix. Second, wih imperfecly compeiive markes, he dimension of he underground secor crucially affecs he size of he price markup and is response o ax reforms, shaping he response of he economy o policy shifs, especially in he shor run. Third, he dimension of he 6 See he OECD Tax Daabase available online a hps://daa.oecd.org/ax/ax-wedge.hm. The ax wedge is measured as he raio beween he amoun of axes paid by a single worker a 100% of average earnings wihou children and he corresponding oal labor cos borne by he employer. 7 This value arises when considering only he marke economy, ha is excluding he non-marke services provided by Public Adminisraions. When he laer are included, he esimae of he underground economy reduces o 17.5 percen of GDP. See ISTAT (2010). 8 This discrepancy of esimaes can be parly explained by he fac ha differen definiions of he shadow economy enail differen measuremens. For insance, Buehn and Schneider (2016) claim ha explici shadow aciviies - ha is, shadow aciviies from black hours worked - only accoun for abou one hird of he shadow economy. See also Zizza (2002) for furher deails. 4

5 underground secor is subsanially decreased by fiscal inervenions envisaging subsanial labor ax wedge reducions, while he sronger effecs on oupu are observed when he business ax is reduced. The remainder of he paper is organized as follows. Secions 2 presens he model wih underground producion. Secion 3 provides deails on model paramerizaion and soluion sraegy. Secion 4 describes he alernaive fiscal reform scenarios and presens he resuls of he policy simulaion analysis. Secion 5 provides sensiiviy analysis. Finally, Secion 6 presens our conclusions and presens possible exensions in his line of research. 2 The Model We build a wo-secor dynamic general equilibrium model wih underground economy, a rich se of disorionary axes, and sicky prices. The Ialian economy is open o inernaional rade and is modeled as a small open economy. Firms are divided ino wo ypes: (i) inermediae goods producers ha engage in monopolisic compeiion, produce radable goods eiher in he formal or in he underground secor, and face price rigidiies; and (ii) compeiive final good producers ha aggregae he domesically produced inermediae goods wih impored inermediae goods ino a final non-radable good used for privae and public consumpion, and invesmens. The possibiliy ha inermediae good firms produce in he underground secor is moivaed by he exisence of ax disorions. In fac, informal producion enables firms o keep ransacions (and labor inpus) secre from fiscal auhoriies, hus eluding ax paymens and raising profis. 9 However, in each period inermediae good producers face a posiive probabiliy, p (0, 1), of being inspeced and conviced o pay back he oal amoun evaded, increased by a surcharge facor, s > 1. In he household secor, agens can allocae heir working ime beween formal and informal jobs, where he income derived from he laer remains unrepored. Since underground 9 On he conrary he final good secor can only operae formally. However, since producion of he final good secor uilizes domesic and foreign inermediae inpus a value added ax scheme could be inroduced ino he model, so allowing for ax evasion also in his secor. Noably, being based on a credi-and-refund scheme he VAT is srongly vulnerable o evasion, fraud and ax avoidance. For insance, over he period Ialy repored an average VAT gap of 33%. See European Commission (2015a). As we discuss in he Conclusions, we leave his imporan aspec for fuure research. 5

6 producion requires informal labor, we assume ha firm inspecions allow he fiscal auhoriies o simulaneously deec boh irregularly produced oupu and irregularly employed workers. Thus, households and inermediae firms face he same probabiliy of being deeced. We also assume ha he penaly o be paid if deeced is he same. 2.1 Final Good Secor The represenaive firm producing he final non-radable good Y F produced inermediae goods Y H a consan elasiciy of subsiuion (CES) echnology: Y F = [ combines a bundle of domesically wih a bundle of impored inermediae goods M, according o (1 α M ) 1 σ M ( ) Y H σm 1 1 σ M + αm σ M 1 σ σm M M ] σ M σ M 1, (1) where σ M > 0 is he elasiciy of subsiuion beween domesically produced goods and inernaionally produced goods, 0 < α M < 1 represens he share of foreign inermediae goods used in he [ ] ( ) θy [ ] θ θ producion of he final goods, while Y H 1 = 0 Yi, H Y 1 M θ Y 1 θ Y θm 1 θ M 1 1 di and M = 0 M θ M j, dj, where θ Y, θ M > 1 denoe he elasiciies of subsiuion beween he differeniaed inermediae goods produced a home and abroad. Le P i, denoe he price of he differeniaed produc i, hen, a he opimum, he demand for Y H i, is Y H i, = ( Pi, P ) θy Y H, (2) where P = ( ) 1 0 P 1 θ 1 Y 1 θ i, di Y demand for he generic j impored good is found o be is he price index of domesically produced goods. Similarly, he M j, = ( ) P M θm j, M, (3) P M where P M j, is he price of good j and P M is he price index of impored goods, P M = [ ( ) ] Pj, M 1 θm 1 θ M dj. Given echnology (1), cos minimizaion delivers he overall demand for domesic and impored 6

7 goods: ( ) σm Y H P = (1 α M ) P F Y F, (4) M = α M ( P M P F ) σm Y F, (5) where P F [(1 α M ) (P ) 1 σ M + α M ( P M ) 1 σm ] 1 1 σ M is he final good price index. 2.2 Inermediae Good Secor wih Underground Producion The marke is populaed by a coninuum of firms acing as monopolisic compeiors which can avoid paymen of axes and social securiy conribuions by employing illegal labor. 10 We furher assume imperfec price adjusmen according o a iming-dependen pricing scheme à la Calvo, where he probabiliy of changing price is consan and does no depend on he sae of he economy. 11 Le Y i, denoe he producion of he differeniaed inermediae good i, which is allocaed beween domesic producion of he final good and exporaion, herefore we have Y i, = Y H i, + X i,, where X i, denoes he quaniy expored. We furher assume ha expor demand of variey i is as in (2), herefore he overall demand for inermediae inpu i is Y i, = ( Pi, P ) θ Y Y. (6) As proposed by Busao and Chiarini (2004), and more recenly by Orsi e al. (2014) and Pappa e al. (2015), he ypical firm i can produce is specific inermediae good wih wo differen echnologies: Yi, m = A m ( Hi, m ) α K 1 α i,, (7) Yi, u = A u ( Hi, u ) αu K 1 αu i,, (8) where α, α u (0, 1), A m and A u denoe oal facor produciviy (TFP) in he regular and in he 10 Noably, perfec compeiion can be rejeced for almos all secors in all counries and Ialy is no an excepion. See Chrisopoulou and Vermeulen (2012). 11 The exisence of price sickiness is fully suppored by empirical evidence. See, e.g, Alvarez e al. (2006) and Dhyne e al. (2006). In general, a marked heerogeneiy in he frequency of price adjusmen is found across secors. Furher, prices seem o change less frequenly in Europe han in he Unied Saes. 7

8 irregular secors. In order o reflec he fac ha less skilled and educaed people are more likely o work in he informal secor compared o hose who aained higher educaion levels, and ha he informal echnology is labor inensive compared o he regular one, 12 we assume secor-specific TFPs and ha α α u - precisely, α u > α and A m > A u. According o he above producion funcions underground labor inpus use he same capial sock ha is used by regular labor. In his sense, similarly o Busao e al. (2011) and Orsi e al. (2014) we assume a moonlighing producion scheme where firms rely on he regular echnology by day and on he informal one by nigh. 13 Goods produced in he wo secors are homogeneous and indisinguishable. Denoing by P m i, and Pi, u he price of good i produced, respecively, in he regular and informal marke, his assumpion implies ha P m i, = P u i, = P i,, where P i, is he price ha he i-h inermediae producer charges o final good producers. Therefore, oal oupu produced by he inermediae firm i is simply given by Y i, = Y m i, + Y u i,. (9) Equaion (9) shows ha Y m i, and Y u i, are perfec subsiues in producion.14 Moreover, (9) implies ha firm i could cease using he informal echnology wihou affecing is producion capaciy. In his perspecive, he informal echnology merely offers an addiional channel along which firms maximize heir profis by reducing he disorions generaed by axaion, as i will become clear in a 12 For insance, Gallaway and Bernasek (2002) and Marcelli e al. (1999) show ha, a leas in urban seings, low skilled workers are more likely o work in he irregular secor compared o hose workers who aained a higher level of educaion. Amaral and Quinin (2006) link he presence of low-skill workers in he underground secor o he fac ha informal managers have access o less ouside finance han heir formal counerpar, and hey subsiue par of physical capial wih low-skill workers. 13 On he noion of moonlighing producion, see e.g. Bajada (1999) and Cowell e al. (1990). This scheme allows us o capure he fac ha firms operae simulaneously in he official and in he unofficial secors. Several causes have been lised in he lieraure o explain he coexisence of formal and informal producion wihin he same firm in conras o a ghos ype producion scheme, where firms op o operae only in he informal secor. For insance, paricipaion in he official marke faciliaes access o credi and public secor services and grans propery righ proecion. Official producion may also represen a valuable shield agains fiscal audis for firms engaging in underground producion. Moreover, openness o rade implies ha expor producs should mee he requiremens of he desinaion counry, which looks in principle less likely for ghos, unregisered, firms. Our choice is also suppored by he empirical evidence showing he predominance of his scheme of informaliy for Ialian firms (see Manegazza e al. 2014). 14 This propery, ogeher wih homogeneiy and he fac ha boh Yi, m and Yi, u can be eiher used in he final good producion funcion wih no addiional coss, suggess ha his model wih underground economy could be more appropriaely defined a wo-echnology model raher han a wo-secor model. 8

9 momen. Thus, he underground economy exiss as i is opimal for firms o use boh echnologies. Given he exisence of wo echnologies and axaion, firm i oal expeced revenues, Λ i,, a ime are E{Λ i, } = P i, [ (1 τ y )Y m i, + (1 psτ y )Y u i,], (10) where τ y is a business ax on firms revenues. Equaion (10) implies ha a necessary condiion for he exisence of underground producion is ha 1 psτ y > 0, as firms have an incenive o use he informal echnology as long as revenues from his aciviy are expeced o be posiive. To produce Y i,, firm i hires (regular and informal) workers and rens capial in compeiive markes. Expeced oal coss, Φ i,, a ime are hen given by E{Φ i, } = (1 + τ f,s )W m H m i, + (1 + psτ f,s )W u H u i, + R k K i,, (11) where τ f,s denoes social securiy conribuion rae paid by employers. In order o derive he opimal demand for regular labor, informal labor, and capial, firm i chooses coningen plans {H m i,, Hu i,, K i,} =0 ha maximize he presen discouned value of is fuure sream of profis subjec o he demand for inermediae goods (6), echnology consrains (7) and (8), he ideniy (9), and given he vecor of prices {P, W u, W m, R k } =0. The soluion o his problem is characerized by he following se of firs order condiions: (1 + τ f,s )W m = αmcn i, Y i, m, (12) H m i, (1 + psτ f,s )W u = αumcn i, Y i, u H u, if 1 psτ y > 0 i, (13) Hi, u = 0, oherwise, R k = MC N i, [(1 α) Y i, m + (1 α u ) Y i, u K i, K i, ], (14) where MC N i, represens firm i nominal marginal cos, ha is he Lagrange muliplier aached o he consrain (9). Equaions (12) and (13), respecively, describe he demand for regular and informal labor. Wih imperfec compeiion he marginal cos drives a wedge beween he wage 9

10 and he marginal value of labor. Equaion (14) is he demand for capial, where again he presence of imperfec compeiion implies he exisence of a wedge beween he renal rae of capial and is marginal value. Because of heir marke power, inermediae firms opimally choose he price of heir oupu. However, during each period = 0, 1, 2... only a fracion (1 ψ) of inermediae producers se a new price, while he remaining fracion ψ charge he previous period price. The probabiliy of changing price is consan over ime, regardless of how long i has been since he las price change, and is independen of economic condiions. Thus, he ypical firm able o rese is price in period will choose a price P i, ha maximizes he presen discouned value of expeced profis unil he seleced price remains effecive, namely: max P i, E j=0 ψ j ( Q,+j Λi, MCi,Y N ) i,, (15) subjec o he sequence of consrains (6) and (9). In (15) he erm Q,+j denoes he nominal sochasic discoun facor used a ime by shareholders o value dae + j real profis and is relaed o he household s marginal uiliy of real wealh λ, i.e. Q,+j = β λ +j λ The firs-order condiion for his problem delivers he soluion P i, = θ Y P θ Y 1 E { j=0 ψj Q R,+j E { j=0 ψj Q R,+j MC i,+j [1 τ y +j + τ y +j ( P+j P ) θy Y+j } P P +j. (1 ps) Y u i, Y i, ] ( P+j P ) θy 1 Y+j }, (16) where Q R,+j = P +j P Q,+j is he sochasic discoun facor for real payoffs and MC i,+j = MCN i,+j ( 1 denoes he real marginal cos. Given price saggering, he aggregae price level P = 0 P 1 θ Y i, di [ evolves according o P = ψp 1 θ Y 1 + (1 ψ) P ] 1/(1 θy ), ha is o say ha he price level is jus P +j ) 1 1 θ Y a weighed average of he las period price level and he price se by firms adjusing in he curren period. This equaion can be rewrien as follows: 1 = ψπ θ Y 1 + (1 ψ) p 1 θ Y, (17) 10

11 where Π = P /P 1 (i.e. PPI inflaion) and p = P P. Noice ha in he deerminisic seady sae and in he absence of rend inflaion (i.e. Π = 1), he opimal condiion (16) collapses o: P = P = θ Y MC N θ Y 1 1 τ y + τ y (1 ps) Y. (18) u Y Clearly, in he absence of he business ax, he seady-sae price markup will only depend on he elasiciy of subsiuion beween inermediae goods. 15 In his case, insead, he price markup is shown o be increasing in he business ax rae, while he possibiliy of evading his ax hrough undeclared producion is found o miigae his effec. In his sense, marke power gives producers he possibiliy of shifing he burden of he business ax oward consumers. In wha follows we will show ha his feaure of he model is no innocuous for our resuls. Tax reforms reducing he share of he underground secor, Y u Y, will end o increase he markup in he long run (imperfecly compeiive producers will be able o shif a par of he burden of he business ax o consumers), while severely condiioning he response of he economy in he shor run, due o he imperfec price adjusmen involved in he Calvo s pricing scheme. In fac, firms able o rese heir prices will end o decrease sharply heir markups in he shor run in response o he fiscal reforms reducing he disorions of he economy in he aemp o capure major marke shares, or o increase hem in response o a srong deerioraion in he erms of rade which, in urn, reduces sharply he compeiive pressure from he foreign marke. 2.3 Aggregaion As shown in Appendix A, he sudy of inermediae firms marginal producion coss poins ou ha hese are he same for all firms, meaning ha all producers which are able o se heir price opimally make he same price decision, ha is, ˆP i, = ˆP. In equilibrium facor and good markes clear, hence he following condiions are saisfied for all : H m = Hi, mdi, Hu = Hi, u di, K = K i, di and Noably, monopolisic compeiion in he inermediae goods secor generaes an average markup, which lowers oupu wih respec o he compeiive economy. In his model, he effecs of he lack of compeiion are furher complicaed by he exisence of he business ax and of he underground secor. 11

12 Y m = 1 0 Y m i, di, Y u = Yi, u di, where = 0 ( Pi, P ) θy di is a measure of price dispersion. I is easy o see ha evolves according o a non-linear firs-order difference equaion of he form: = (1 ψ) ˆp θ Y + ψπ θ Y 1. (19) Aggregae producion Y = Y m + Y u is hen found o be Y = 1 [ A m (H m ) α K 1 α + A u (H u ] ) αu K 1 αu, (20) where, evidenly, price dispersion creaes a wedge beween aggregae oupu and producion inpus, hus generaing oupu dispersion ha is increasing in he degree of price rigidiy, ψ. 2.4 Households The economy is populaed by a large number of idenical, infiniely living, agens who seek o maximize he following objecive funcion, which is separable in consumpion and labor: U 0 = E 0 =0 β {( C η C 1 ) 1 σ 1 1 σ Γ 0 (H m + H u ) 1+ξ 1 + ξ } (H u ) 1+ϕ Γ 1. (21) 1 + ϕ where β (0, 1) is he subjecive facor, C denoes consumpion, σ > 0 is he inverse of he ineremporal elasiciy of subsiuion and η (0, 1) regulaes he inensiy of exernal habi relaive o he lagged aggregae consumpion C 1 and inroduces non-separabiliy of preferences over ime. Wih exernal habi a rise in curren aggregae consumpion increases he marginal uiliy of individual consumpion in he nex period. This specificaion of preferences capures he excess of smoohness of consumpion observed in he daa and induces a slower adjusmen o fiscal reforms. 16 The firs negaive erm in (21) is he sandard disuiliy of working aciviies, which is regulaed by he scale parameer Γ 0 0, and where he sum beween hours spen working in he formal secor, H m, and underground secor, H u, represens he oal amoun of hours worked in he economy; 16 For evidence of habi formaion in consumpion of Ialian households, see Rossi (2005). 12

13 he second negaive erm represens he idiosyncraic cos of working in he underground secor, which is regulaed by he scale parameer Γ 1 0. Finally, ξ > 0 and ϕ > 0 are posiive parameers affecing he elasiciy of labor supply. As in Busao and Chiarini (2004), Busao e al. (2012) and Orsi e al. (2014), we assume ha individuals face an idiosyncraic disuiliy of working in he underground secor which can be jusified by he lack of any social and healh insurance along wih job insecuriy. 17 Furher, he addiional disuiliy aached o he supply of underground labor is mean o capure he subjecive cos due o he social sigma associaed wih paricipaion in he informal secor. 18 Pu i differenly, he second negaive erm in (21) capures he exensive marginal dimension of informal labor supply. Households consume, supply labor per uni of ime, own physical capial which hey ren ou o producing firms, pay axes on consumpion and labor income, and pay social securiy conribuions. The sock of capial held by households, denoed by K, evolves according o he following law of moion: [ ( )] I K +1 = (1 δ)k + 1 S I, (22) I 1 where δ (0, 1) is he consan depreciaion rae of capial, I denoes invesmen, and ( ) I S = φ ( ) 2 I I 1, (23) I 1 2 I 1 are convex invesmen adjusmen coss he inensiy of which is regulaed by he parameer φ I > 0. By penalizing fas invesmen change, convex adjusmen coss represen he coss relaed o insallaion, disrupion of producive aciviies, he need o rerain workers, he need o change he producion process ec... The presence of adjusmen coss prevens invesmens from over-reacing o changes in he economic environmen. 17 Despie he absence of adjusmen coss while shifing labor supply or demand from a marke o anoher, each labor marke has is own specific characerisics ha his exra erm ries o represen. 18 On hese behavioral aspecs of households regarding paricipaion o shadow economy and ax moraliy, see e.g. Gordon (1989) and Dhami and Al-Nowaihi (2007). 13

14 The represenaive household s period-by-period budge consrain is given by (1 + τ c ) P F C + P F I + +Q B +1 + S F +1 = B + R S F + R k K + (1 τ h τ h,s )W m H m + (24) + [1 ps(τ h + τ h,s )]W u H u + D + T ls, where τ c represens he ax rae on consumpion, B +1 is he quaniy of one-period, nominally risk-free bonds purchased in period and mauring in period +1, while B is he quaniy of bonds carried over from period 1. Each bond pays one uni of currency a mauriy and is price is Q. 19 F denoes he quaniy of one-period foreign asses, R is he risk-adjused (gross) ineres rae, S is he nominal exchange rae, defined as he home currency per uni of foreign currency and R k is he nominal rae of reurn on capial. Households earn a ne salary (1 τ h τ h,s )W m for working one uni of ime in he regular secor, where τ h he workers SSC rae, and W m is he labor income ax rae, τ h,s represens is he nominal wage of formal jobs, and obain he nominal wage W u W u for each hour hey work in he underground secor. As previously discussed, he informal wage is subjec o he income and social conribuions ax raes - increased by a penaly - only in probabiliy, ha is when he fiscal auhoriies deec he informal employmen relaionship beween workers and firms. Finally, D denoes nominal profis received by he represenaive household for he ownership of inermediae-good producing firms, and T ls from households o he governmen secor. are nominal lump-sum axes flowing The problem of he represenaive household consiss in choosing opimal coningen plans {C, H m, H u, B +1, F +1, K +1, I } =0 so as o maximize lifeime uiliy (21), subjec o he sequence of consrains (22) and (24). We focus our aenion on he opimal labor supply condiions, while we refer o Appendix A for furher deails. Denoing by λ he Lagrange muliplier associaed wih he household ineremporal budge consrain (24) expressed in real erms - ha is, he marginal uiliy of real wealh - from he 19 Clearly, he price a ime of a risk-free asse ha pays one uni of currency in period + 1 has o be equal o he inverse of he risk-free nominal (gross) ineres rae R, ha is Q = R 1. 14

15 firs-order condiions wih respec o H m and H u we obain H m + H u 1 ξ = λ [ (1 τ h τ h,s ) W m P Γ 0 ] 1 ξ, (25) ( [ ] 1 1 ps(τ h h,s W u ) +τ ) (1 τ P h h,s λ ϕ τ ) W m 1 ϕ [ ] P H u Γ 1, if 1 ps(τ h + τ h,s W u ) P = 0, oherwise, (1 τ h τ h,s ) W m P, (26) where equaion (25) deermines oal labor supply, whereas equaion (26) describes how households opimally allocae ime o underground aciviies. Because of he addiional disuiliy semming from working in he underground secor, households supply informal labor as long as he afer-ax wage differenial beween informal and formal wages is non-negaive. From hese wo firs-order condiions we also noice he role played by parameers ξ and ϕ in his model, wih he former deermining he responsiveness of oal labor supply (legal and illegal) o changes in he ne wage paid in he legal secor, and he laer deermining he responsiveness of illegal labor supply o he ne wage differenials beween he wo secors. 2.5 Public Secor, Marke Clearing and Curren Accoun Each period he consolidaed governmen faces an exogenous sream of expendiures, G. In order o finance hese expendiures, he governmen issues one period risk-free bonds wih nominal value B, and collecs axes in he amoun T. Hence, he governmen budge consrain in period can be wrien as R 1 B +1 = B + P F G T, (27) where R is he risk-free nominal (gross) ineres rae. Because of he rich srucure of axes in his economy and he exisence of underground producion, ax revenues, T, are given by T = (τ h +τ h,s +τ f,s )W m H m +ps(τ h +τ h,s +τ f,s )W u H u +τ c P F C +τ y P (Y m + psy u )+T ls. (28) 15

16 In wha follows we assume ha he governmen adops a balanced budge rule a all imes, herefore he lump-sum componen of axaion, T ls, is adjused ex pos so as so clear any possible imbalances arising from changes in ax revenues and/or public consumpion. As a consequence he real sock of public deb is kep consan a is iniial baseline level. The ne foreign asse posiion of his economy, expressed in erms of domesic currency, follows he process S F +1 = R S F + P X P M M, (29) where he risk-adjused ineres rae on foreign asses, R, depends on he risk premium. In paricular, as in Schmi-Grohé and Uribe (2003) we use he following funcional form for he risk premium: ρ F = ϕ F (e (f f) 1), where ϕ F is a posiive parameer, f = S F /P and f is he seady sae level of ne foreign asses in real erms. 20 Le R be he unadjused ineres rae facor, hen R = R + ρ F. To close he model we need o specify expor demand, X. We assume his funcional form: X = α X ( P S P F ) σx Y F, (30) where α X is a posiive coefficien, σ X denoes he elasiciy of expor demand, P F is he price of he final good prevailing abroad, while Y F denoes he overall foreign final producion. The erms of rade is defined as, T ot : T ot = P P M = P S P, (31) where P is he price of radable goods produced abroad expressed in foreign currency. T ot can hen be inerpreed as he amoun of impor goods he economy can purchase per uni of expor goods. In his sense a deerioraion in he erms of rade implies ha a greaer volume of expors is needed o finance a given amoun of impored goods. In addiion, when he erms of rade worsens households face a negaive income effec and end o subsiue impored goods wih domesic 20 In oher words, here is a premium on foreign bond holdings which is a funcion of he aggregae ne foreign asse posiion of he domesic households. The exisence of his deb elasic risk premium is necessary o induce saionariy of he equilibrium dynamics. For deails, see Schmi-Grohé and Uribe (2003). 16

17 goods. Turning o moneary policy, we rea he nominal exchange rae as a consan (S = 1) and conduc all our analysis under his assumpion. 21 According o ISTAT bilaeral rade saisics in 2015, in fac, Ialian expors represen abou 11% of euro area world expors. This is why changes in he rade balance posiion of Ialy compared o rading parners ouside he moneary union are expeced o produce only minor changes in he nominal exchange rae beween he euro and he oher counries. For his reason we can safely ake he nominal exchange rae as given. 22 Finally, he resource consrain of he economy immediaely follows: Y = P F P C + P F P G + P F P I + X S P P M. (32) All he equilibrium condiions describing he baseline model are summarized in Appendix A. 3 Calibraion The model is calibraed a quarerly frequency o mach seady-sae grea raios for he Ialian economy and consisenly wih he exising DSGE modeling lieraure. Given he lack of daa for non-observable variables in he shadow economy, some parameers and variables are se according o he Bayesian esimaes provided by Orsi e al. (2014). Precisely, we draw from heir work he seady sae share of irregular producion o oal oupu, Y u /Y i.e., he size of he underground economy which is equal o 0.228, he inverse of he elasiciies of oal labor supply, ξ, and of informal labor supply, ϕ, equal o and , respecively, and he oupu elasiciies of labor in he regular echnology, α, and in he underground echnology, α u, respecively equal o and According o he esimaes by Orsi e al. (2014) we also se he ineremporal elasiciy of subsiuion σ o The discoun facor β is se o 0.99, which implies an annual real (and nominal, given a zero seady sae level of inflaion, Π = 1) ineres rae of 4%. The capial depreciaion rae δ is se o 0.025, while he invesmen adjusmen cos parameer, φ I, is se o 2.5, 21 This assumpion implies ha all he observed changes in he erms of rade are o be ascribed o changes in he price level P, since P is also kep consan. See equaion (31). 22 This assumpion allows us o swich off any possible feedback channels coming from moneary policy. 17

18 which is an inermediae value beween he paramerizaion used by Forni e al. (2010b) and by Forni e al. (2010a) for he Ialian economy. The parameer regulaing he exernal consumpion habi componen is se o 0.7, in line wih he exising lieraure. 23 Toal hours worked in he economy, H = H m + H u, amoun o In order o assess he share of irregular o oal hours, H u /H, we rely on counry-specific daa provided he Euro Baromeer 2007 Survey of Undeclared Work in he European Union, as repored by Andrews e al. (2011). According o his informaion, he share H u /H is se o , which delivers H m = and H u = The probabiliy of a firm being inspeced, p, is se o 3%, which is he value esimaed by Busao and Chiarini (2004) using daa on firm inspecions released by he Ialian Minisry of Labor; he penaly s ha a ax evading firm is conviced o pay if deeced is se o 30% of he oal amoun evaded, which is consisen wih he curren Ialian Tax Law. 25 We se he elasiciy of subsiuion beween inermediae goods, θ Y a 6 so ha he corresponding price markup maches he esimaes by Chrisopoulou and Vermeulen (2012) who produce esimaes in he manufacuring secor for Ialy and he res of he Euro area. The probabiliy ha firms do no revise heir price, ψ ha is, he degree of price sickiness is se o The equilibrium level of oupu Y is normalized o 1, and he deb-o-gdp and consumpiono-gdp raios are se o 130% and 60%, in line wih he mos recen daa released by he Ialian Naional Insiue for Saisics. We se he level of expors X so as o mach he average expor- GDP share for he period consisenly wih he World Bank Indicaors, i.e Similarly, also he average share of impors for he same period amouns o 26% of GDP. Given hese daa we assume ha he ne asse posiion of Ialy agains he res of he Euro area is zero, (F = 0). The elasiciy of subsiuion beween domesic and impored radable goods, σ M, is se o 1.5 consisenly wih Forni e al. (2010b). Since each differen ype of ax accouns for a differen shares of he Ialian governmen ax revenues, we se he model ax raes accordingly. To his end we use OECD Tax Daabase for he year 2014, according o which axes on personal income represens 26.6% of oal ax revenues, 23 For insance, Boldrin e al. (2001) find a poin esimae of 0.7, while Chrisiano e al. (2005) esimae a value of For his share we follow Orsi e al. (2014). 25 See Busao and Chiarini (2004); Orsi e al. (2014). 18

19 axes on corporae income 7.2%, employers social securiy conribuion 20.1%, employees social securiy conribuion 5.6% and consumpion axes 23%. Given he model ax base under he baseline calibraion, he implici ax raes immediaely follow. The remaining share of axaion deermines T ls (i.e. he lump-sum componen) which represens a sor of cach-all componen of axaion, capuring all axes no covered by he model. 26 Finally, his calibraion delivers implici values for public expendiure, (G), he scale parameers regulaing labor disuiliies in he agens preferences, (Γ 0, Γ 1 ), oal facor produciviy in he wo producion echnologies (A m, A u ), and impors (M). Parameer values are repored in Table 1, while Table 2 summarizes he seady-sae values of he main macroeconomic variables of our model. Using his paramerizaion, he non-linear version of he model is solved using a Newon-ype algorihm. To conduc our simulaion exercise, we rea all he ax raes and governmen spending as exogenous variables and examine he deerminisic response of he economy o unexpeced permanen changes in hese variables occurring a he beginning of our simulaion ime horizon. 27 The analysis of he effecs of permanen shocks requires solving a wo-poin boundary-problem, specifying he iniial condiions for he predeermined variables and he erminal condiions for he forward looking variables. 4 Fiscal Reform Scenarios The simulaion exercise quanifies he poenial macroeconomic impac of a se of ax policy experimens parially inspired o he fiscal reform packages being underaken or presenly under consideraion in Ialy. 28 In paricular, over , Ialy reduced personal income ax by means of measures argeing low-income earners, such as he exension of ax credis as par of he 2014 Law 26 For insance we are no considering axes on capial rens. In addiion, for simpliciy excise axes and value added axes collapse in a unique consumpion ax. 27 Noably, deerminisic simulaions are carried ou when sudying he effecs of srucural and/or fiscal reforms involving permanen changes in some srucural parameers and/or ax raes. 28 Recenly many EU Member Saes have been aemping o srike he balance beween economic recovery and fiscal soundness by reducing he ax burden on labor and by narrowing he ax base of corporae income axaion, while increasing consumpion axes. Lower ax wedges on labor, in fac, ough o lead o a significan rise in labor uilizaion and, hereby, o a higher level of economic aciviy. Similarly, alleviaing he ax burden on business may simulae invesmen. However, given he risk for public finance, lower labor and business axes need o be compensaed by increases in oher sources of revenue, such as consumpion axes, or by public spending cus. For furher deails, see European Commission (2015b). 19

20 on Sabiliy, and he inroducion of a special personal income ax bonus for 2014 for low-income employees. 29 These kinds of inervenions are mapped ono he model by reducing he ax rae on labor income τ h. In he same period labor coss became deducible from he regional producion ax. The reducion of ax burden relaed o he regional producion ax is capured by a decrease of he ax rae on business τ y. Addiionally, employers SSC for new employees hired in 2015 on permanen conracs has been waived for hree years. This measure ranslaes ino a reducion of he SSC ax rae bearing on firm τ f,s. Turning o indirec axes, over he period he sandard VAT rae was increased by wo percenage poins. As common pracice in DSGE modeling, his change in VAT is capured by variaion in he consumpion ax rae τ c. Finally, according o he Ialian Minisry of Economics and Finance, he recen spending review plan resuled in savings of 18 billion euros in The exra measures presen in he Sabiliy Law for 2016, are expeced o deliver savings for 25 billion euros for In wha follows, he simulaion hypoheses concerning he size and he mix of he ax shifs, assumed in each case, are arbirary. In paricular, we design five scenarios in which a permanen reducion of one of he disorionary ax raes is me by a permanen change in anoher disorionary ax rae or by a conemporaneous governmen spending cu, so as o ensure ha he fiscal reform is designed o be ex-ane budge neural. For he sake of comparabiliy he size of each inervenion is se a 0.5% of oupu. Simulaions are carried ou under he assumpion ha all policy changes are unanicipaed by agens and ake place immediaely. Forward looking agens adjus heir behavior accordingly, anicipaing he long-run effecs of he ax reform pu in place. To make our analysis as much ransparen as possible and uncover he role played by he exisence of an informal secor on he performance of differen ax reforms, we carry ou our experimens also for an economy wih a small informal secor, where we calibrae he size of he underground economy a 5%, and for an economy wih an underground secor accouning for 40% of oal producion. Imporanly, he 29 Moreover, in he las years personal income ax has also been reduced for hose saring a new business and o he benefis of hose embarking on home renovaion works. Oher secors involved regards ar, spor, and ourism. 30 See he noe available from Ialian Minisry of Economics and Finance websie a hp:// hml and he informaion made available from he spending review working group a hp://revisionedellaspesa.gov.i/. 20

21 dimension of he informal secor deermines he size of he ax base involved by he ax change envisaged in each scenario. We consider hree scenarios feauring ax swaps amouning o 0.5% of oupu (iniial level), namely S 1 combines a reducion of he ax rae on business τ y wih an increase in he consumpion ax τ c, S 2 presens a ax shif from labor income o consumpion 31 and S 3 embodies a reducion of he employers SSC rae, τ f,s, compensaed by an increase in he consumpion ax rae. We hen consider wo addiional scenarios, S 4, S 5, in which he reducions of axes envisaged in S 1 and S 2 are financed by immediae public spending cus. Noice ha in all he experimens he implied changes in he ax raes on labor differ in he hree economies since he size of he ax base involved by he reform is differen. Table 3 summarizes he hypoheses made for each ax policy experimen in our benchmark economy (Y u /Y =0.228) and in he economy wih a small and wih a large illegal secor (Y u /Y =0.05, 0.4). In conducing hese experimens he model has been re-calibraed by changing proporionally he iniial share of worked hours in he underground secor, while keeping all he oher resricions and ax raes as in he benchmark economy. In wha follows we look a he long-run effecs on he main macroeconomic variables by comparing he afer-reform seady sae equilibria. We also look a he response of he economy afer 1, 5, and 10 years, so uncovering he role played by nominal and real fricions in shaping he iniial response of he economy. 32 For each ime horizon we repor a quaniaive evaluaion of he welfare gains or losses associaed wih each ax reform scenario. As firsly proposed by Lucas (1987), welfare gains or losses associaed wih a specific policy are compued in erms of consumpion equivalen unis. Tha is, as i is shown in Appendix A, we compue he permanen change in consumpion ha leaves households indifferen beween he uiliy derived by remaining in he pre-ax reform equilibrium and he uiliy obained when he ax reform is adoped. Finally, in order o capure he bang-for-he-buck aspecs of he reforms implemened we also look a he effecs of 31 This could be inerpreed as a reducion in he sum of he ax on labor income and workers SSC, reflecing he fac ha hese wo wedges isomorphically affec households decisions. 32 I should be noed ha he model is calibraed and solved on quarerly basis. We do no show he full ransiional dynamics, raher we repor he effecs of ax reforms which maerialize afer 4, 20 and 40 quarers, and hen heir long-run impac. 21

22 fiscal inervenions on ne ax revenues defined as oal axes minus he lump-sum componen of axaion as a share of of oupu, i.e. ( T T ls ) /Y. 4.1 Resuls In his secion we illusrae he effecs of all ax reforms showing he impac a differen ime horizons and considering hree differen hypoheses regarding he dimension of he underground economy, small Y u /Y =0.05, baseline Y u /Y =0.228 and large Y u /Y =0.4. As said above, fully unanicipaed fiscal reforms are implemened saring from he beginning of he simulaion. However, habi in consumpion, adjusmen coss on invesmen and nominal price sickiness preven variables from jumping immediaely o he new seady sae. The impac of fiscal reforms on he levels of economic aciviy are shown in Tables 4a and 4b, where oupu, consumpion, invesmen, hours, he price markup and he erms of rade are expressed in percenage deviaions from heir iniial value, while he share of underground economy, ne ax revenues (T T ls as % of oupu) and he rade balance (X 1 T ot M as % of oupu) are measured in percenage poin deviaions from heir iniial level. For each variable, scenario and dimension of he underground secor we repor a vecor of four elemens referring o he impac of he reform a differen ime horizon. The firs hree elemens refer o 1, 5 and 10 years, while he fourh elemen measures he impac of he reform on he long-run (seady-sae) levels of economic aciviy. For each scenario we firs discuss he general endencies of all variables and hen compare he resuls observed for differen shares of he underground secor. Furher, having normalized he pre-reform level of oupu o one, he inerpreaion of he resuls is raher sraighforward Shifing he Tax Burden from Business o Consumpion We sar by discussing he case in which we reduce he business ax by 0.5% of oupu and increase consumpion ax so as o leave he budge balance unalered, a leas ex ane (S 1 ). Oupu, invesmen and hours sar increasing immediaely, while consumpion declines as a resul of major axaion. The srong reacion of invesmens has wo effecs. Firs of all, i induces a emporary excess of demand yielding a posiive drif on PPI inflaion and herefore emporarily improving 22

23 he erms of rade of he economy. Second, he higher demand of impors for invesmen purposes induces a negaive rade balance which is absorbed already afer five years. A laer sages, he major increase in oupu delivers a fall of he erms of rade along wih a posiive impac on he rade balance. As a consequence of he lower consumpion, he marginal uiliy of wealh increases inducing households o supply more labor in boh secors. In paricular, we observe a major effec on worked hours in he irregular secor. Despie he fac ha on he demand side he new ax mix favors he employmen of workers in he formal secor, on he supply side households find i opimal o increase he supply of irregular labor in he aemp of overcome he negaive income effec bearing on hem. During he adjusmen process he laer effec dominaes he former, so explaining why he share of he underground economy is slighly higher afer a year from he implemenaion of he reform. We noice ha he average markup declines as a direc consequence of he drop in τ y, especially a earlier sages, when he effecs of nominal rigidiies are more pervasive. Ne ax revenues, insead, are shown o increase already afer one year of he reform, while welfare worsens as a resul of he lower level of consumpion and he higher labor effor. Furher, i should be noed ha on he one hand consumers suffer from he burden of a high ax on consumpion, on he oher hand, hey benefi form he reducion of he price markup. However, he laer effec does no offse he former. We now urn our aenion o he differences across economies displaying differen shares of he underground secor. Firs, we noice ha in he shor run he differences on oupu are negligible, while hey end o widen in he long run. The major effecs are deeced in he economy characerized by a small informal secor. In general, we observe ha he underesimaion of he size of he underground secor may lead o overesimae he posiive impac on all macroeconomic variables especially in he long run, while he miscalculaion is minor in he shor run, where openness o inernaional rade is he key facor deermining he smoohed difference. When he informal secor is small, in fac, srong re-allocaion effecs end o be operaive, inducing a sharper posiive reacion of oupu. This is eviden in he long run. However, in a open economy he posiive response of oupu induces a worsening of he erms of rade which revers back o he economy 23