Tax Ratio Threshold and Economic Growth in Iran

Transcription

1 Ausralian Journal of Basic and Applied Sciences, 4(8): , 2010 SSN ax Raio hreshold and Economic Growh in ran 1 Ahmad Jafari Samimi, 2 ounes Nademi, 3 Hoda Zobeiri 1 Professor of Economics & Vice presiden in Research Affairs Deparmen of Economics, Universiy of Mazandaran, Babolsar-ran 2,3 Ph.D. Suden Deparmen of Economics, Universiy of Mazandaran, Babolsar-ran Absrac: he purpose of his aricle is o esimae he rae and effec of hreshold ax on economic growh in ran during o do so, firs we divide ran s economy ino wo secors namely, public and non-public secor based on Ram (1986) model. hen we esimae a linear model.due o poor resuls of he esimaed linear model we specified a nonlinear model on explain he role of ax on economic growh in ran.. he resuls indicae ha here exiss a hreshold ax rae of 22%. n oher words, when ax rae is less han 22%. hen as a resul of an increase in ax revenue rae of economic growh will increase and beyond a ax rae of 22%. will have a derimenal effec on economic growh of he counry. Key words: ax Rae; Economic Growh; hreshold Model; Nonlinear Model. NRODUCON Economic growh and is deerminans have been a criical issue especially for developing counries afer he world war. ax rae is one of he variables which may have boh posiive and negaive effec on economic growh. he negaive effec is due o behavior and choice of individuals and firms. n fac when ax rae increases, invesing reurns (boh physical and human) decreases which in urn leads o decrease invesing inensives and reducion of economic growh. he posiive effec is due o increase governmen revenue. When ax rae increases he governmen abiliy o spend on healh, educaion, infrasrucures and so on increase and as a resul boos he rae of economic growh. he raio of ax revenue o gross domesic producion has increased remarkably in many counries hrough wenieh cenury. his raio was 5 o 10 percen in beginning of wenieh cenury bu i reach o 30 o 40 percen in he end of wenieh cenury. According o managing and planinig insiusion repors, in 2006 ran ax reavenue become 20 imes as much o Now he quesion is he effec of his excessive ax revenue on economic growh during las 40 years. Afer he inroducion, secion 2 is devoed o lieraure reviews. Secion 3 inroduces he model specificaion. hen Secions 4 provide empirical resuls and finally secion 5 presens a summary of main conclusion. Review of ieraure: ax rae affec economic growh hrough individuals and firms decisions. Several models have ried o idenify his relaionship beween ax rae and economic growh. Apar from deails, all of hem invesigae his process wih wo differen pars (Myles 2009).his wo pars are as follows: g y = g y ( a 1 ( 1, 2 ), a 2 ( 1, 2 )) (1) n which a 1 and a 2 represens wo sor economic aciviies such as educaion and research. 1 and 2 are ax raes. g y (.) is equilibrium growh rae as a funcion of economic aciviy. Using (1) he effec of ax rae change on economic growh is equal o: g y / d i = (dg y / da i )( da i / d i ) (2) Corresponding Auhor: Ahmad Jafari Samimi, Professor of Economics & Vice presiden in Research Affairs Deparmen of Economics, Universiy of Mazandaran, Babolsar-ran jafarisa@umz.ac.ir 2242

2 Aus. J. Basic & Appl. Sci., 4(8): , 2010 n which da i / d is represen effec of ax rae on economic aciviy. dg y / da i is represens effec of economic aciviy on economic growh. he effec of ax rae on economic growh is equal o effec of ax rae on economic aciviy plus effec of economic aciviy on economic growh. Now if he da i / d i is large, bu dg y / da i is lile, he resul is ax rae is no influence economic growh effecively. n conrary if da i / d is small bu dg y / da i is large, hen ax rae is imporan facor of economic growh. herefore he effec of ax on economic growh is no clear in heoric analysis. Several sudies have ried o invesigae he effec of ax rae on economic growh which mos of hem ( no all of hem) conclud ha increase of ax rae have negaive effec on economic growh. Some of hese sudies are poined ou as follows: Mofidi and Sone (1990) invesigaed he effec of ax on economic growh of 50 developing counries during heir purpse was analyse he effec of ax on microeconomic level in order o remove he business cycles. hey resuls indicaed ha ax increase had negaive effec on privae invesmen and economic growh. Hanson (2002) examined he effec of ax on wealh and economic growh in 20 of OECD counries and have made same resuls. Holcombe and acomble (2004) sudied he effec of marginal ax rae change on per capia income in 20 developing counries hrough secional analysis. hey compared counries which have increased ax rae wih oher counries which had lower ax rae during hey concluded ha he counries wih larger ax rae have had lower capia income relaively. Furceri and Karras (2008) invesigaed he effec of ax rae changes on economic growh of 26 OECD counries empirically. heir resuls indicaed ha 1 percen increase in ax rae lead o 0.5 o 1 percen decrease in capia GDP. hey also come o he conclusion ha increase goods and services ax rae have more negaive effec as compared wih increase income ax rae. Jafari Samimi and Hasanzadeh (2001) argued he effec of ax on economic growh wih regard o endogenous growh models. heir findings indicae ha he effec of ax on economic growh is no clear. Furhermore he resuls are senseives o parameers and ype of daa ( ime series or secional) ha used in he model. n sum hey conclude ha he effec of ax on economic growh is weak. Samadi, Zahedfar and Faramarzi (2008) showed ha if income disribuion be righ which means capured ax from wealh people and paid o poor people, hen increas ax revenue lead o increase economic growh. ncrease ax revenue in ran based on empirical resuls during , cause o decrease economic growh. Because in ran mos of axes gaherd from low income people and mos expendiures do for high income people. Model Specificaion: We have used he Ram (1986) model as following: g g G e ( ) ( ) (1) G Regression (1) shows ha he variables which affec economic growh ( 1 ) include he invesmen rae ( ), growh of labor force ( gg ), and he muliplicaion effecs of governmen expendiure growh ( gg ) imes governmen size (G/). Now, we assume ha budge is balance. herefore G=, G and G. So we modify Eq.(1) as following: g g e ( ) ( ) n Eq.(2), we idenify he muliplicaion effecs hrough he sign of 3. his indicaes ha he governmen secor has a reciprocal effec on economic growh hrough wo ways: one is he direc conribuion of he governmen secor and he oher is he indirec effec hrough he non-governmen secor (exernaliy effec). Regression (2) is a radiional linear economic growh model, bu we aler he linear model ino he wo regime AR model of Hansen (1996, 2000). he model can be shown as follows: g 13 g e q g 23g e q ( ) ( ) if ( ) ( ) if 2243 (2) (3)

3 Or as one nonlinear regression such as: Aus. J. Basic & Appl. Sci., 4(8): , 2010 ( ) ( ) g 13 g q ( ) ( ) g g q e he hreshold value can be found by esimaing he regression (4) hrough finding he minimum Error Sum of Squared in a re-order hreshold variable. he hreshold variable can be se by he exogenous variables ou of he heoreical model. For example, in his paper we se ( ) as he hreshold variable. We can also apply he saisic coming from he hreshold variable. For insance, we adop he heeroskedasiciy-consisen agrange muliplier (M) of Hansen (1996) o es he null hypohesis of he linear assumpion. Once he esimaor can be found, we hen sar wih he saisical es, bu he es procedure of Eq. (4) is differen from he radiional es. Under he null hypohesis of no hreshold effec, he hreshold parameers will be unidenified. his will cause he radiional es saisic in a large sample disribuion o no belong o he 2 disribuion, bu raher o a non-sandard and non-similar disribuion which is affeced by nuisance parameers. his will cause he criical value of he disribuion o no be esimaed hrough simulaion. n order o overcome he difficuly, Hansen (1996) uses a saisic of his own large sample disribuion funcion o ransfer and calculae he asympoic p-value of a large sample. Under he null hypohesis, he disribuion of he p-value saisic is uniform, and his kind of ransformaion can be calculaed hrough boosrap. he null hypohesis o es Eq. (4) is as follows: H i 0 1i 2i : ; 1,2,3,... f H 0 is no rejeced hen he relaionships beween economic growh and he ax raio ( ) (4) (5) would be he linear regression as he regression (2). his means here exiss no hreshold effec. Oherwise, if H 0 hypohesis is rejeced, i means ha here exis differen effecs beween he wo regimes of 1i and 2i. he F-es saisics is as follows: F 1 RSS RSS ˆ n which RSS 0 and RSS 1 are he residual sum of squares under he null hypohesis and he alernaive, respecively. Empirical Resuls: his paper uses Hansen (1996, 2000) hreshold regression model o sudy wheher a non-linear relaionship beween ax raio and economic growh exiss in ran. As able 1 shows, we adop Hansen (1996, 2000) advice o use he boosrapping model. While he hreshold variable is ax revenue divided by GDP, we find ha F-saisic is (-4.62), which is significan a 1% level. he hreshold value is 22%, and his means ha one hreshold exiss. As able 1 show, only he invesmen rae ( /) has a significan and posiive effec on economic growh in he linear model, bu he oher variables haven a significan effec on economic growh in he linear model. n he nonlinear model, while ax revenue divided by GDP is he hreshold variable, since he ax raio is small (he hreshold value is less han 0.22), he invesmen rae and he muliplicaion effecs of ax revenue g growh ( ) imes ax raio (/) have a significanly posiive effec on economic growh in ran. Bu when he ax raio is large (he hreshold value is larger han 0.22), he muliplicaion effecs of ax revenue growh ( g ) imes ax raio (/) has a significanly negaive effec on economic growh, and he oher variables (6) 2244

4 Aus. J. Basic & Appl. Sci., 4(8): , 2010 haven a significanly effec on economic growh in ran. hus, we can make sure ha he non-linear relaionship beween ax raio and economic growh as he shape of inversion U exiss in ran when ax revenue divided by GDP is he hreshold variable. Moreover, he labor force growh has no a significanly impac on economic growh in boh of wo regimes axes able 1: Economic Growh and ax Raio Variables inear Model ax Raio hreshold value (%) Coefficien -saisics saisics >0.22 -saisics nercepion -0.09*** ** / 0.98*** *** g ( g ) ( ) ** ** R Ramsey rese (F) Jarque-Bera (Normaliy) 14.49*** 1.71 HE(Breusch-Pagan-Godfrey) F value of hreshold es -4.62**** ndicaes significance a 10% level. ** ndicaes significance a 5% level. ***ndicaes significance a 1% level. Conclusion: n his paper, we have modified he Ram (1986) wo-secor producion model ino a hreshold regression model and apply Hansen (1996, 2000) mehod o es he hreshold effec. he empirical resuls indicae ha hreshold effec exis beween ax raio and economic growh in ran. While ax revenue divided by GDP as he hreshold variable, he hreshold regime is 22%. his indicaes ha when he ax raio is smaller han he regime, economic growh is promoed under expanding governmen expendiure, bu if he ax raio is larger han he regime, hen he economic growh decreases. REFERENCES Armey, R., "he Freedom Revoluion", Washingon DC: Rognery Publishing Co. Dar, A. and S. khalkhali, "Governmen Size, Facor Accumulaion, and Economic Growh: Evidence from OECD Counries, Journal of Policy Modeling, 24: Engen, E. and J. Skinner, "Fiscal Policy and Economic Growh", n paper presened a NBER conference on axaion. Folser, S. and M. Henrekson, "Growh Effecs of Governmen Expendiure and axaion in Rich Counries", European Economic Review, 45(8): Furceri, D. and G. Karras, " ax Changes and Economic Growh: Empirical Evidence for a Panel of OECD Counries", OECD Economic Deparmen, Working Paper, pp: 641. Gwarney, J., R. awson and R. Holcombe, 1998 "he Size and Funcions of Governmen and Economic Growh", Join Economic Commiee. Hanson, A., 2002." he Wealh ax and Economic Growh", Deparmen of Economics, und Universiy. Holcombe, R.G. and D.J. acomble, " he Effec of Sae ncome axaion on Per Capia ncome Growh", Public Finance Review, 32: Hsieh, E. and K. ai, "Governmen Spending and Economic Growh; he G-7 experience" Applied Economics, 26: Jafari Samimi, A. and A. Hasanzadeh, " nvesigaion he Effec of ax on Economic Growh; heoreical and Empirical Analysis", Economic researches, 1(2): Kormendi, R.C. and P. Meguire, "Governmen Dep, Governmen Spending, and Privae Secor Behavior", American Economic Review, 76(1): andau, D., Governmen Expendiure and Economic Growh; A cross-counry sudy, Souhern Economic Journal, 49(3): in, S., "Governmen Spending and Economic Growh", Applied Economics, 26(1); Mofidi, A. and J. Sone, " Do Sae and ocal axes Affec Economic Growh", Review of Economics and Saisics, 72(4):

5 Aus. J. Basic & Appl. Sci., 4(8): , 2010 Myles, D.G., 2009." Economic Growh and he Role of axaion- heory", OECD Economic Deparmen, Working Paper, pp; 713. Ram, R., "Governmen Size and Economic Growh; A New Framework and Some Eevidence from Cross Secion and ime-series Daa", American Economic Review, 76(1): Samadi, S., A. Zahedmehr and A. Faramarzi, "he Effec of Fiscal Policy on ncome Disribuion and Economic Growh of ran", Business Research, 49: Sheehey, E., "he Effec of Governmen Size on Economic Growh, Easern Economic Journal, 19 (3): Sheng ung C. and. Chien Chiang, "Governmen Size and Economic Growh in aiwan: A hreshold Regression Approach, Journal of Policy Modeling, 27: Vedder, R.K. and.e. Gallaway, "Governmen Size and Economic Growh" paper prepared for he Join Economic Com. 2246