Tax Compliance, Income Distribution and Social Norms

Transcription

1 Theoretcal Economcs etters, 017, 7, ISSN Onlne: ISSN Prnt: Tax Complance, Income Dstrbuton and Socal Norms Debora D Goacchno, Fabrzo Patrarca Sapenza Unversty of Rome, Rome, Italy ow to cte ths paper: D Goacchno, D. and Patrarca, F. (017 Tax Complance, Income Dstrbuton and Socal Norms. Theoretcal Economcs etters, 7, Receved: March 1, 017 Accepted: Aprl 1, 017 Publshed: Aprl 4, 017 Copyrght 017 by authors and Scentfc Research Publshng Inc. Ths work s lcensed under the Creatve Commons Attrbuton Internatonal cense (CC BY Open Access Abstract Ths paper studes the effect of ncome nequalty on tax evason. To dscuss the topc, we present a smple model, based on Benabouand Trole [1], that ncorporates ncentves for tax complance such as punshment and fnes, ntrnsc motvaton and socal norms. Snce we consder a regressve system of ncentves to comply, ncome nequalty ncreases the value of tax evason although overall propensty to comply s unaffected. In ths framework, we consder the hypothess that socal norms are group specfc as n the case of socal segregaton or status related networks. We show that all the negatve effects of nequaltes are amplfed: the dfference between the tax complance of the ncome groups and the value of tax evason ncrease. Keywords Tax Evason, Socal Norms, Income Dstrbuton 1. Introducton The relatonshp between ncome nequalty and tax evason s a complexone. Bloomqust [] fnds a postve correlaton between a measure of ncome n equalty (Gn coeffcent and the under reportng rate for wage and salary ncome. On one hand, f ndvduals ncentve for tax evason s ncome related, then the share of tax evaders and the total amount of tax evaded depend on the dstrbuton of ncome. On the other hand, the specfc desgn ofthe tax system, together wth socal norms and laws, affect the after-tax dstrbuton of ncome (see [3] for a dscusson. ere we focus on the frst ssue: the effect of ncome nequalty on taxevason. The theoretcal lterature that has tred to explan the reasons for tax complance has dentfed three motvatons: extrnsc, ntrnsc and reputatonal. In ther semnal paper, Allngham and Sandmo [4] consder DOI: /tel Aprl 4, 017

2 the decson on how much ncome an ndvdual should report and how much should he evade as dependng on hs ncome, the fne to be pad n case of detecton and the (subjectve probablty of detecton. They show that f the expected penalty s less than the tax rate then some tax evason s optmal from the taxpayers pont of vew. Moreover, f n dvduals are rsk-averse and absolute rsk-averson s decreasng, then rcher n dvduals wll evade more. Thus, n the standard models (see [4] [5] evason makes the tax system regressve. Along the lnes of Gordon [6], several papers have ncorporated tax morale nto the standard models of tax evason, recognsng that people abstan from tax evason not only for fear of beng caught but also for moral and socal consderatons: agents have ntrnsc motvaton to behave prosocal andthey also lke to appear.e. have a reputaton for beng prosocal. Taxmorale s nterpreted as a socal norm for tax complance and the strength of the norm s shaped endogenously, dependng on the populaton share of honest ctzens. If taxpayers care about socalstgma, the proporton of ndvduals who are expected to evadetax plays an mportant role n ndvdual evason decsons. The nterdependence of ndvduals evason decsons typcally results n amultplcty of equlbra. An economy wth a gven ncome dstrbuton and tax system can ether end up n a state wth a strong socalnorm, where most taxpayers pay all ther taxes, or a state wtha weak socal norm, where evason s more wdes pread. Km [7] nvestgates the role of the ncome dstrbuton on the exstence of multple equlbra, and, usng numercal smulatons, argues that multple equlbra arse f taxpayer ncomes are suffcently homogeneous so that the socal coordnaton effect domnates. Allowng for heterogenety n ndvduals concern for socal norms, Taxler [8] characterzes, analytcally, the multplcty of (socal equlbra n the share of evaders. e also consders tax complance n a populaton consstng of subgroups wth nter-group spll overs emphaszng the potental role of norm complance n reference groups for taxmorale n other groups. Buldng on Benabou and Trole [1], Besley et al. [9] use a smple model that ncludes ntrnsc motvaton, extrnsc ncentves and socal concerns to study the dynamcs of taxevason. They apply ths framework to estmate the dynamcs of tax evason n the UK after the ntroducton of a poll tax n1990. In dealng wth ndvduals ncentve for tax complance, our startng pont s the model n [1] as appled to tax evason n [9]. Ther model ncludes n a smple way all the three motvatons for tax complancementoned above. To study the effect of nequalty on tax evason, we modfy [9] by assumng two groups of agents dfferng n ther ncome, but dfferently from them, we do not consder the dynamcs of taxevason. In such a framework, ncome nequalty can affect tax complance through all ts determnants: the ntrnsc motvaton, whenever ts dstrbuton among agents changes wth ncome levels; the tax system, ncludng the punshment and fnes structure, as long as ts not proportonal; and fnally extrnsc motvatons and socal norm when they are determned accordng to ncome group specfc features. In our case we do not consder the frst motvaton, thus the dstr- 590

3 buton of the ntrnsc motvaton s not correlated toncome. We wll consder a tax system yeldng a regressve system of ncentves to tax complance, ncludng progressve taxaton and nonproportonal fnes. Fnally, we wll us ths settng to analyze the case n whch socal norms are group specfc, whch corresponds to the dfferent cases of socal segregaton or network effects when networks have a socal structure. We wll show that ncome specfcsocal norms ncrease the negatve effect of nequaltes on taxevason, havng a further regressve effect on dsposable ncomes. At the same tme, the average socal norms wll not change snce the hgher number of payers between the agents on the bottom of the dstrbuton wll ncrease exactly as the number of rcher agents evadng.. Baselne Framework We bult up our analyss on the general model n [9]. There s a contnuum of agents of unt mass maxmzng: max y x( e ( m+ v e+ E( ve ; (1 c where y s ncome, x the tax, e ( 0;1 the evason decson (yes = 1; no = 0, m the expected economc cost of beng caught, v the subjectve cost of evadng havng a densty dstrbuton g(v, μ the senstvty to socal norms, E( ve the socal norm effect gven the choce e. There s a cutoff value v such that: m x+ v = E( ve= 1 E( ve= 0. ( Gven the densty functon of v and ts cumulate G(v, tax wll be evaded by N T = xg v. = G(v agents for a total tax evason of N ( If v ~ U( 0, φ, then 1 : ( 1 ( 0 E ve= E ve= = v φ ; (3 and thus usng Equaton (: φ + x m v =. (4 The share of tax evaders wll be: N v φ + x m ; (5 φ φ 1 = = ( for a total evason T N equal to: x m TN xn x φ + = =. (6 φ 1 ( 3. Income Inequalty and Regressve Incentves to Comply avng defned the benchmark model, we can now ntroduce the hypotheses of a constant fne systems and ncome nequalty. We consder two types = (;, v φ v E v e= 1 E v e= 0 = E v< v E v> v =

4 havng ncome y y wth α representng the share of agents of type. To use the prevous secton as benchmark we take: ( 1 α y = αy + y = y. (7 If the dstrbuton of v s the same for each type, the number of evaders and the tax evaded wll not change. et s consder nstead the case of proportonal tax x = ty and aconstant m. All the followng results same propostons hold when m= m0 + my 1 as long as m1 < t once we substtute t wth t m1 and m wth m 0. In the same way, they hold when taxaton s not progressve usng the opposte transformaton. The maxmzaton problem becomes: max y x e m+ v e+ E ve (8 c For each type, there s a cutoff value v such that: where: m ty + v = E ve= 1 E ve= 0 ( = 1 ( = 0 = ( α ( < + α ( < ( α E( v > v + αe( v > v E v e E v e E v v E v v If v ~ U( 0, φ, then we wll have the equvalent of Equaton (3: ( 1 ( 0 (9. (10 E ve= E ve= = v φ ; (11 where: usng 9 3 : then 4 : Furthermore we have: ( 1 α v = v + αv ; (1 v v = t y y ; (13 v = v αt y y (14 1 ( α v = v + t y y. (15 φ + ty m v = = v ; (16 and thus: ( α E( v< v + αe( v< v ( 1 α E( ( v> v αe( v> v v v φ v φ v = ( α + α ( α α = α v α + v φ ( 1 3 Indeed we have: ( φ m ty + v = v =., 4 m ty + v = v φ α ( v v = v φ αt ( y y v φ ty m α t ( y y + + = 1.. ; 59

5 Fnally 5 : v v v N = α + ( α = = N. (17 ( α α 1 N T = T + t y y (18 As expected, the type evade more than the type, the two thresholds v and v are respectvely ncreasng and decreasng n nequalty (Equaton (15, Equaton refdue, and the value of evason s ncreasng n nequalty (Equaton (18. owever, the number of evaders and the average socal norms vare unaffected by nequalty (Equaton (16: the ncrease n nequalty brngs about an ncrease of tax payers of group equal to the ncrease n evaders of group. 4. Group Specfc Socal Norms If the two types are not n contact we wll have 6 : ( y y α v = v t > v ˆ ( α ( y y vˆ = v + t > v ˆv (19 (0 = v (1 Evason n group s lower and evason n group s hgher. Thus, the dfferences between the socal norms of the two groups wden. Coherently wth Equaton (1 we have: ˆ φ ty m φ ty m N = α + + ( α + = N 1 1 ( ( Average tax complance and the number of evaders doesn t change. Fnally 7 : ( α ( ( ˆ α 1 T = T + t y y > T. (3 N N 1 The value of evason ncreases. Takng together all these results we can conclude that when socalnorms are group specfc and economc ncentves to comply are regressve, the effects of 5 v v ( 1 α ( 1 α α α ( 1 α v + v v v α v v T = t y + t y = t y + t y ; N usng Equaton (13: v + ( α t( y y v αt ( y y ( 1 v α α T = αt y + ( α N t y = t y + t ( y y ; usng eq.6 we obtan Equaton (18. 6 φ + ty m y y v = = v + t. ˆ ˆ φ + ty m φ + ty m t( y y t ( y y 1 1 T = ty α + ty α = T + ty α + ty α N N 1 ( 1 ( 1 ( 1 ( 7 α ( α( y y y + α( α ( y y y = T + t N 1 (. 593

6 nequaltes are amplfed and there s a fnal further regressve effect snce, as n the case of hgher nequalty, segregaton ncrease the number of tax payers of group to the same extent of the ncrease n evaders of group one. 5. Conclusons In ths artcle, we have ntroduced a framework to assess the role of ncome nequaltes n a standard model of tax complance wth socal norms. We have consdered economc ncentves to comply decreasng n ncome as a conse- quence of progressve taxaton or regressve fnes. Thus, hgher ncome nequa- lty ncreases the value of evason whle overall socal norms are unaffected snce the ncrease of evaders nthe rcher class s equal to the decrease of tax payers n the lower ncome class. The regressve nature of the ncentves to comply also determnes a further regressve effect on the dstrbuton of dsposable ncome. Once ntroduced the benchmark framework, we have consdered the hypo- thess that socal norms are formed separately by ncome groups as n the case of socal segregaton or network effects when networks have a socal structure. Group specfc socal norms amplfy all the negatve effects of nequaltes: the dfference between the tax complance of the two groups and the value of tax evason ncrease whle overall socal norms are constant. Ths model provdes a benchmark for the analyss of tax complance when agents are heterogeneous and socal norm are group specfc. Polces amed at ncreasng complance can rely on the socal norm effects, nstead of hgher controls or hgher fnes and punshments. From these perspectve, snce the dsclosure of nformaton about evaders s not always an opton, analternatve channel s to reduce segregaton effects n the process of formaton of socal norms. Acknowledgements We thank for the support: Progetto d Rcercad Ateneo 105-Sapenza Unversty of Rome. References [1] Trole, J. and Bnabou, R. (011 aws and Norms. NBER Workng Paper, No [] Bloomqust, K.M. (003 Income Tax Evason: A Theoretcal Analyss. OECD Papers, 3, 1-7. [3] ee, K. (016 Norms and Monetary Fnes as Deterrents, Anddstrbutve Effects. Journal of Economcs, 190, 1-7. [4] Sandmo, A. and Allngham, M.G. (197 Income Tax Evason: Atheoretcal Analyss. Journal of Publc Economcs, 104, [5] Ytzhak, S. (1974 A Note on Income Tax Evason: A Theoretcal Analyss. Journal of Publc Economcs, 3, [6] Gordon, J.P.F. (1989 Indvdual Moralty and Reputaton Costs as Deterrents to Tax Evason. European Economc Revew, 33, [7] Km, Y. (003 Income Dstrbuton and Equlbrum Multplcty n a Stgma-Based Model of Tax Evason. Journal of Publc Economcs, 87,

7 [8] Traxler, C. (017 Socal Norms and Condtonal Cooperatve Taxpayers. European Journal of Poltcal Economy, 6, [9] Persson, T., Besley, T. and Jensen, A. (015 Norms, Enforcement and Taxevason. CEPR Dscusson Paper, 104, Submt or recommend next manuscrpt to SCIRP and we wll provde best servce for you: Acceptng pre-submsson nqures through Emal, Facebook, nkedin, Twtter, etc. A wde selecton of journals (nclusve of 9 subjects, more than 00 journals Provdng 4-hour hgh-qualty servce User-frendly onlne submsson system Far and swft peer-revew system Effcent typesettng and proofreadng procedure Dsplay of the result of downloads and vsts, as well as the number of cted artcles Maxmum dssemnaton of your research work Submt your manuscrpt at: Or contact tel@scrp.org 595