Economic Effects of the Corporate Income Tax Reforms: A Computable General Equilibrium Approach

Economc Effects of the Corporate Income Tax Reforms: A Computable General Equlbrum Approach Sung Ta Km Insll Y Chong-Bum An Sang-Don Lee Recently reform of the corporate ncome tax (CIT hereafter) s becomng one of the key subjects to rase the natonal compettveness n the nternatonal economy. In Korea there have been tax reform efforts relatng to CIT to meet the nternatonal trends several tmes. The statutory CIT rates have been lowered n many countres. The varous reducton and exemptons relatng to CIT have been consdered. However, Korean CIT system has been evaluated as outdated, n comparson wth fast movng nternatonal economc envronment. The motvaton of ths paper s to fgure out the rght drecton of the Korean CIT reform. The dscussons on Korean CIT reform have been lop-sded to the qualtatve analyss of tax reform approprateness. There are not many studes explanng how CIT affects the ncome dstrbuton by classes and producton n each ndustry. Usng the Korea Computable General Equlbrum (KOCGE) model, we try to analyze quanttatvely the effects of varous corporate ncome tax reducton plans on the GDP, the producton n 25 ndustres respectvely as well as on the welfare of the economy. We nvestgate fve measures for CIT reform n Korea. Frst, we examne the effects of scenaro B and C under whch the government lowers the CIT rate and ncrease other tax rates such as the ncome tax rate for scenaro B and the value-added tax rate for scenaro C respectvely satsfyng the government tax revenue constrant. Second, we examne the effects of scenaro D and E under whch the government lowers the CIT rate and s the government expendtures for scenaro D or s government lump-sum transfers for scenaro E satsfyng the government budget constrant. Thrd, we examne the scenaro A under whch the CIT and the ncome tax are fully ntegrated. Accordng to the smulaton results based on the KOCGE model, the scenaro E s most desrable tax reform measure for Korean CIT. Under scenaro E the government lowers 0% of the CIT rate and government lump-sum transfers, we can mprove both effcency and equty aspect of the economy by ncreasng the welfare level of the economy as well as mprovng ncome dstrbuton by classes. In addton, we found that the full ntegraton of the ncome tax and the CIT wll reduce both labor supply and savngs from households resultng lower welfare and aggravate the ncome dstrbuton. Keywords: Tax Reform, Corporate Income Tax, Korea CGE Model, Income Dstrbuton JEL Classfcaton: H2 Cheongju Unversty, Department of Economcs, Professor, e-mal: stkm@cju.ac.kr Natonal Assembly Budget Offce, Department Manager, e-mal: nsll@nabo.go.kr Sungkyunkwan Unversty, Department of Economcs, Professor, e-mal: cban@skku.ac.kr Korea Research Insttute for Vocatonal Educaton and Tranng, Research Fellow, e-mal: sdlee@krvet.re.kr

I. INTRODUCTION Recently due to rapd globalzaton openng of captal market has nduced nternatonal tax competton wth respect to n captal ncome tax rate. Countres jonng nternatonal tax competton try to ncrease captal nflows from foregn countres by reducng the tax rate on captal returns. Especally as the number of multnatonal frms has ncreased, the reform of corporate ncome tax (hereafter CIT) n Korea looks nevtable to secure Korean frms nternatonal compettveness. The purpose of ths paper s to examne varous CIT reform proposals to fnd the optmal scenaro for Korea. The problem of CIT reform s evdently to fnd an answer to a queston of whch scenaros among varous reform measures for CIT s optmal for Korea. We should have the norms by whch we evaluate each CIT reform measure and rank them n order to fnd the best scenaro for CIT reform. The norms are as usual to have effcency and equty aspects. The effcency crteron les n measurng the excess burden of the CIT that s socal welfare cost of the CIT, whle the equty crteron s to analyze the effect of CIT reform on ncome dstrbuton of households n the economy. We have to go through an ncdence analyss of the CIT for ncome dstrbuton analyss. Incdence analyses of the CIT are very mportant snce varous CIT reform measures affect ncome dstrbuton as well as socal welfare through changes of prce mechansm prces of captal, labor and fnal goods. The ncdence analyss of CIT may be classfed nto two sorts, partal equlbrum analyss and general equlbrum analyss. The former focuses on analyzng how taxes on corporate ncome affect the demands for factors. Ths approach has a couple of advantages. It s easy to do and also smple to understand the whole process. However, ths approach also has shortcomngs of beng lack of mplcatons for the real world economy snce the model s too smple. On the other hand the general equlbrum approach s to analyze how taxes on corporate ncome affect the economy as a whole. Ths approach has advantages of beng very realstc and gvng us a varety of mplcatons. However, ths approach has dsadvantages of beng too complex and of beng hard to nterpret the results of analyss. Recently the general equlbrum approach has been used more often than before due to very rapd development of computaton devces and soft wares. Harberger (962) wrote a semnal paper on ncdence of the CIT usng a two-sector general equlbrum model. Under plausble parameter values depctng the U.S. economy Harberger drew a concluson that the CIT burden mostly fell on captal. The two-factor, two-sector model postulated by Harberger assumes that economy-wde supples of the two factors captal and labor are perfectly nelastc, but that captal 2

and labor are perfectly moble between sectors n the closed economy. Theoretcal tax ncdence formulae for example derved by Harberger (962) are approprate only for small changes around ntal no-tax equlbrum. To examne the ncdence of large tax changes as well as to consder many more ndustry sectors and consumers, Shoven and Whalley (972) suggest a computable general equlbrum model. Ther method of computng an equlbrum s based on Scarf s (967, 973) algorthm and some other technques. To analyze large perturbatons of the general equlbrum we need to assume specfc functonal form for producton functons of ndustres and for utlty functons of households. The parameter values of those functons are selected such that ntal equlbrum values of the model are exactly reproducton of the base year data set. There are several features of the computable general equlbrum models that contrbute to the tax ncdence studes. Above all, the appled general equlbrum taxaton models may use dsaggregate data on both producton and consumpton. We can delneate two levels of dsaggregaton of producton sectors: medum and hgh. Wth ths specfcaton we may examne the effects of the CIT on each ndustry outputs drectly. Dsaggregaton of households by ncome classes n the appled general equlbrum model makes t possble to examne the effect of the CIT on ncome dstrbuton. The computable general equlbrum model s very powerful n examnng both effcency and equty effects of varous tax reform proposals. Thus, n ths paper we use the Korean Computable General Equlbrum (KOCGE) model to examne varous CIT reform measures on the welfare as well as ncome dstrbuton of the Korean economy. Up to now many of the CIT related studes n Korea, for nstance Roh and Km (996), Won (996), Yoo(993) and Cho (996), have pad attenton to estmaton of the effectve tax rate for the CIT. Others ncludng Kang (995), Lee (999), and Yoo (997) analyze the CIT reform measures n Korea. An (996) examned the effect of the CIT rate on nvestment and savngs usng the open macro economy model and Km (996) estmated the excess burden of the CIT for Korea. Lee (997) and Lee and Km (998) studed the ncdence of the CIT n Korea. Kwack (994) n partcular estmated the complance cost of the CIT n Korea. Park and Lee (2000) set up the CIT revenue forecastng model for Korea. Of these studes related to the CIT n Korea, few have examned the effects of the CIT on ncome dstrbuton, the outputs by ndustres at hgher level of dsaggregaton of producton sector, as well as backward shft of the CIT burden through captal and labor prce changes. Therefore n ths paper we nvestgate those effects of varous CIT reform proposals usng the KOCGE model. The paper s organzed as follows. Followng Secton I we brefly ntroduce the KOCGE model focusng on modfcaton 3

of the corporate sector n secton II. In secton III we explan fve reform scenaros for the CIT n Korea. In secton IV, we analyze the effects of fve CIT reform scenaros on the welfare cost of the economy, ncome dstrbuton, labor supply, savngs as well as ndustry outputs. In secton V we explan the polcy mplcatons of analyses. II. KOCGE MODEL. Overvew of KOCGE Model The KOCGE (Korean Computable General Equlbrum) model used n ths paper bascally follow Km et al. (2003, 2005). The KOCGE model s based on the Ballard- Fullerton-Shoven-Whalley (985) type model and revsed to accommodate Korean stuaton. Four mportant factors ncorporated n the KOCGE model are resource endowments of the consumers, demand functon, producton technology, and the equlbrum condton. The model n ths paper has three sectors, consumpton, producton, and government. Households n the consumpton sector are dvded nto 0 ncome groups wth each household category maxmzng ts utlty functon subject to a gven budget constrant. Busness frms n the producton sector maxmze ther profts producng an overall of 26 dfferent manufacturng goods and 0 consumpton goods under a CES producton functon, usng labor and captal. The government collects taxes, such as personal ncome tax, specal consumpton tax, corporate proft tax, value-added tax and spends wthn the budget so as to keep the budget balanced. () Producton In the model labor L and captal K are the two basc factors of producton employed to produce goods and servces. It s assumed that both labor and captal are homogeneous and easly moble among the sectors. Captal s owned by the ten consumer groups and by government, and we denote endowments by K j (j =,..., 0) and K G, respectvely. Captal can be used n any of the 26 producer ndustres or n the general government sector. These uses of captal are denoted by K ( =,..., 26). Only consumers have endowments Ej (j =,..., 0) of labor, but because they also consume lesure, ther actual supples are Lj (j =,..., 0) wth lesure denoted Ij (j =,..., 0). Labor can be used n any of the twenty-sx sectors and for each consumer, then, we have Ej = Lj + Ij. In total, we have, 4

0 0 0 26 0 j = j j j j= j= j= = j= E = E L + I = L+ I = L + I () Each of these factors s defned n servce unts per perod. When a unt of captal servces s rented out for one perod, the owner receves a prce, P K, whch s net of factor taxes and deprecaton. In addton to the rental prce, P L and P K, whch are pad to factor owners, producers are requred to pay factor cost ncludng factor ncome taxes. These taxes dffer by sector. Therefore the th ndustry producer faces gross of labor ncome tax P L as follows, P = P ( + τ ) (2a) L L L When t comes to captal ncome producer pays the corporate ncome tax at rate of τ C only for corporate sector. The corporate ncome tax rate τ C s the rato of the corporate ncome to total taxable captal ncome. Snce the retaned earnngs are not taxable the after tax captal prce becomes P = ( + τ KENC ( re )) P C K K τ C (2b) where KENC = rato of corporate sector captal to total th ndustry captal, re = rato of retaned earnngs to total captal ncome of corporate sector n th ndustry. Captal and labor appear n a constant elastcty substtuton (CES) value-added functon of the form: σ σ σ σ σ σ [ δ ( δ) ] VA =Φ L + K =,..., 26 (3) where Φ and δ are producton parameters, and σ s the elastcty of substtuton. For expostonal smplcty, we have suppressed the subscrpts of all varables and parameters. Ths model uses a 26 26 fxed coeffcent nput-output matrx, denoted by A, wth columns gvng the ntermedate nput requrement per unt of output. The ndustry outputs are represented as Q ( =,..., 26). A sngle output s characterzed by cost mnmzaton for each unt of output. Mnmzaton of factor costs ( PL L + PK K ) subject to the constrant that VA = n equaton (4) yelds the labor and captal demand requrement per unt of value-added as shown n equaton (5) and (6). 5

mn LK, L K σ σ σ σ δ σ σ s.t. VA = Φ [ δl + ( ) K ] = P L + P K (4) σ ( σ ) σ PK R δ L =Φ ( δ) δ =,, 26 + (5) ( δ P ) L σ ( σ ) σ ( P ) L R δ K =Φ δ ( δ ) =,, 26 + (6) δ PK To fnd out the demand for labor and captal, we denote the value-added contrbuted by ndustry by VA and then the contrbuton rato to output, var, s defned n equaton (7) and determned exogenously. VA var = (7) Q Once var s determned, the demand for labor and captal follows σ ( σ ) σ PK L VAR L VA δ = = Φ ( δ) δ + ( δ P ) L (8) σ ( σ ) σ ( P ) L K VAR K VA δ = = Φ δ ( δ) + δ PK (9) Gven parameters δ, Φ and σ for each ndustry, we use the net of tax factor prces together wth the tax rate to calculate each producer s gross-of-tax prce for each factor. Thus the tax system dstorts factor nput decsons. We assume perfect competton n both the factor and good markets. Therefore there s no economc proft left after the producer pays for factor costs and taxes. The prce of a producer s 6

good s P ( =,..., 26) must be set such that zero economc proft condton s satsfed. The before tax prce of one unt of the th good s the cost-coverng prce of producer goods, that s, the prce beng pad to value-added (V ) n equaton (0) and ntermedate good (a j ) whch s nputted n one unt of the producer s good. Ths s shown by equaton (). V = P R + P R =,, 26 (0) L L K K T P= ( I A ) V P V.. P (26 ) = V (26 ) =.. P V 26 26 () where A T s the transpose matrx of the nput-output matrx A. We mpose dfferent value added tax rates on each ndustry s ntermedate purchases from each other ndustry. Each ndustry s supposed to share ts tax burden at the same level, the effectve value added tax rate of each ndustry s dfferent from each other, due to the economc polces of the government. We mpose dfferent tax rate on each ndustry. To analyze the performances of tax admnstraton, we add tax avodance rate n ths model, as follows. P = P[ + τ {( a ) enc + ec}] (2) v v Here τ v s the value added tax rate, a v s the avodance rate on value added tax. ec and enc are respectvely the rato of corporate sector to whole ndustry and that of non-corporate sector wthn each ndustry, enc + ec =. The tax avodance rate s defned as follows. a v total amount of value added tax n economy value added tax actually pad = total amont of value added tax n economy (3) =,, 26 Consumer goods X m (m =,..., 0) are produced as producer goods Q through a fxed coeffcent Z matrx as shown n (4). Each of the coeffcent Z m n the Z matrx 7

gves the amount of producer goods, needed to produce one unt of consumer good m. Snce perfect competton s assumed, producers make zero profts after payments for factors, factor taxes, and output taxes. The zero proft condton also apples to the producton of consumer goods. The costs covered n consumer good prces are gven by P m n equaton (4). 26 m m = P = z P m =,, 9 (4) When consumers purchase good X m, they must pay addtonal value added taxes. We model sales taxes on the purchase of each good at rates τ m (m =,..., 9). Grossof-tax prces pad by consumers are shown n (5). Sales tax τ m nclude all the taxes that the consumers face, for example, specal consumpton tax, telephone tax, lquor tax, stamp tax, securty tradng tax. P = P ( + τ ) (5) m m m (2) Households We assume that the households n ths model own all the goods and factor of payments. The market demand functons for goods are non-negatve and contnuous n prces. Indvduals make savng decsons based on expectatons about the resultng ncrement n future consumpton. We assume that expectatons are myopc n the sense that ndvduals expect all current prces, ncludng the return to captal, to reman constant through all future perods. The household chooses the demand for consumer goods based on three-stage maxmzaton of the nested utlty functon. In the frst stage, consumers choose present consumpton H and future consumpton C F, to maxmze ther CES utlty functon. Equaton (6) shows the frst stage consumers maxmzaton problem. Each consumer group has ts own set of parameters and values, whch s a CES utlty functon; we suppress ndexes for expostonal smplcty. H, CF σ2 σ2 σ2 2 2 2 2 2 α F max U ( H, C ) = [ α H + ( ) C ] F s.t. σ σ σ σ σ PP s I = PHH + CF PK γ Here I s current expanded ncome after taxes and transfers, α s a weghtng parameter and σ 2 s the elastcty of substtuton between H and C F. γ s the (6) γ 8

physcal servce flow per unt of captal goods purchased, whch s assumed exogenous. In the second stage, consumers dstrbute ther present consumpton H, after savng. They choose ther composte good X and lesure lmaxmzng a CES Utlty functon U( X,l). The 2 nd stage maxmzaton problem s wrtten n equaton (7). X, l σ σ σ σ max H = [( β) X + β l ] s.t. σ σ σ σ I P S = P X + Pl s l (7) where β s a weghtng parameter, and σ s the elastcty of substtuton between X and l. The prce of lesure, P l s taken to be the after-tax return to labor of each group. Snce a unt of labor earns P L after factor taxes, P l = P L (- τ j ), where τ j d the jth consumer s personal margnal tax rates. After spendng P l lon lesure, consumer has I Ps S = P X + Pl l avalable to spend on the consumpton components on X, n the thrd stage. They choose X m (m=,,, 9) to maxmze a Cobb-Douglas form of the subutlty functon n (8), s.t. max X = X 9 m= m m 9 s l m= m m I PS Pl = X P I λ (8) The λ m weghtng parameters are the Cobb-Douglas expendture shares. Constraned maxmzaton of the subutlty functon, X, provdes the demand functons as follows mj ( I j PS Sj Pl l ) j j Pm λ X = m =,, 9, j =,, 0 (9) mj Wth the mportant property of the nested Cobb-Douglas and CES utlty functon, we derve the followng prce functons n (20), (2), and (22) 5 P ( m λ P = ) m (20) m= λ m 9

( ) ( ) l P = [( β) P( σ ) + βp σ ] σ (2) H P U ( σ2) PP S ( σ2) ( σ2) H α PK γ = [ αp + ( )[ ] ] (22) (3) Government The government collects personal ncome tax, value-added tax, sales tax, and factor payment tax as tax revenues and mantans a balanced budget. The personal ncome tax follows margnal rates that dffer among ncome groups. It also ncludes specal features that dscrmnate by ndustry. For example, ndustry wth more corporate sector wll have a hgher tax burden. Wthn the corporate sector, the ndustry wth hgher resdual earnngs wll pay more taxes than the ndustry wth hgher dvdends and nterest payments. We assume that there s a tax evason rate a v assgned to each ndustry n the non-corporate sector, whch ncludes mostly selfowned frms. Ths s shown n equaton (2). The sum of each ndustry s captal ncome net of corporate ncome s the same as the captal ncome receved by the ten consumer classes. The rght hand sde of equaton (23) s the sum of captal ncome receved by the ten consumer groups and the left hand sde s the sum of captal expendture pad by twenty-sx ndustres. 26 0 CAI = CAI (23) = j= j Each of the 0 consumer groups has a margnal tax rate on all captal and labor ncome, denoted by τ j (j =,..,0). Many transfer payments are not subject to ncome tax. In our model we assume that all transfers are tax-exempt, whle labor ncome s fully taxable. Ths s expressed by the followng formula for ncome taxes pad by group j. I T = B + τ P L + τ ( a ) P K (24) j j j L j j j K j The ntercept of each lnear tax functon, B j may be nterpreted as a knd of socal securty tax: t s negatve reflectng the fact that margnal tax rates exceed average tax rates. Whle margnal changes n ncome are taxed at the approprate margnal rate for each group, ths margnal rate does not change as ncome changes. Expanded ncome, I j, equals transfers plus labor and captal ncome, plus the value of lesure, 0

mnus ncome taxes. Therefore, for each group j (j =,, 0 ), we have I = ( R G) d B + E P ( τ ) + K P [ τ ( a )] (25) j j j j L j j K j j Here, G s government expendtures, whch s the sum of government s fxed captal formaton (Q FIG ) and government s consumpton expendtures (Q FG ), that s G = Q FIG +Q FG, and s exogenously determned. We dvde government expendtures nto two categores. Some publcly suppled goods and servces are offered free of charge, whle other expendture for goods and servces and nvestment are subject to a user s charge. The government dstrbutes the resdual after payng expendtures (G) from tax revenues (R) to each group of consumers as a transfer. The dstrbuton for j th consumer group s d j (R-G) whch can be nterpreted as a knd of socal securtes. Snce the government transfers to the consumers all resduals, the government budget s always balanced as shown n (26). 0 9 0 [ B + τ P L + τ ( a ) P K ] + ( τ P X ) j j L j j j K j m m mj j= m= j= 9 26 0 26 ZmPτv Xmj τlpll τkpkk m= = j= = + [( )( )] + ( + ) (26) 0 ( G+ d ( R G)) = 0 j= j 2. Equlbrum of Model The equlbrum of the model s defned under gven governmental polcy varables {G, d j } and parameter values (enc, ec, a v, a j), dstrbuton varables {H j, C Fj, S j, X, l, X mj, R L, R K, Q, L, K } and prce varables {P H, P S, P L, P K, P, P m, P, P l } (=,.. 26, j=,...0) satsfyng the followng condtons:

A. Optmzaton condton A-. Producer s optmzaton condton: equaton (4) A-2 consumer's optmzaton condton A--. st step optmzaton condton: equaton (6) A--2. 2 nd step optmzaton condton: equaton (7) A--3. 3 rd step optmzaton condton: equaton (8) B. Government s budget balanced condton: equaton (26) C. Feasblty Condton D. Market equlbrum condton D-. Goods market equlbrum condton D-2. Factor market equlbrum condton () Goods Market Equlbrum In the goods market, the demands for fnal producer goods are dvded nto consumpton demand (Q FC ), nvestment demand (Q FI ), government consumpton demand (Q FG ), and foregn demand (Q FX ). We assume that foregn demand s balanced. The equlbrum condton for the goods market s determned where total demands for fnal producer goods equal total supples n producton sector. Q I A Q Q Q Q FC FI FG FX = ( ) ( + + + ) (27) Here the consumpton demand for fnal producer good, follows Q FC, s defned as 9 FC ( m m) m= Q = Z X P =,, 26 (28) Q FC 26 = Q (29) = FC In ths model, the nvestment demand for fnal producer goods, Q FI, s the sum of the household sector s savngs (Q FIH ), the corporate sector s savngs (Q FIC ), the government s savngs (Q FIG ) and deprecaton (D), and changes n nventores (IV) as shown n (30). We assume that government and corporate sector savngs and nventory changes are gven exogenously. 2

FI FIH FIC FIG Q = Q + Q + Q + D+ IV (30) Here, household s savngs equals nvestment demand for fnal producer goods, as shown n (3). Q = Z S, =,, 26 (3) FIH k 0 H (2) Factor Market Equlbrum Captal market equlbrum s as follows 26 26 0 K FIC D Q K j = = j= ( ) = 0 (32) Labor market equlbrum s as follows 26 0 L L = 0 (33) = j= j 3. Dervaton of Model Equlbrum The equlbrum n ths model s summarzed by 565 equatons wth 565 varables. To solve ths system of equatons, we replace the varables n equatons reducng ths system of equaton nto 3 equatons wth 3 varables, P L, P K, R, as follows: 26 0 F L Lj = j= λ ( I P S Pl ) = 26 0 0 0 0 mj j S j l j [ Tj ( Zm ( ))] R L Lj k= j= m= j= P j= m = 0 (33) 3

26 0 2 F K K j = j= λ ( I P S Pl ) = 26 0 0 0 0 mj j S j l j [ Tj ( Zm ( ))] R K K j k= j= m= j= P j= m = 0 (34) 0 9 0 3 F Bj + τ jpllj + τ j aj PKK j + τmpmxmj j= m= j= [ ( ) ] ( ) 9 26 0 26 m τv mj τl L τk K m= = j= = = 0 [( Z P )( X )] + ( P L + P K ) R (35) III. SIMULATION FOR THE CORPORATE INCOME TAX REFORM PROPOSALS IN KOREA. Corporate Income Tax Reform Proposals () Full Integraton of Indvdual Income Tax and Corporate Income tax: Scenaro A Ths scenaro s to abolsh the CIT and ncrease the ndvdual ncome tax rate for satsfyng the real tax revenue constrant. Thus the followng condton must be satsfed. R() = R(0) Q where Q = L L 9 Pm m= 9 Pm m D () x (0) m D (0) x (0) m (36) Here number 0 n the bracket represents the benchmark equlbrum, whle number 0 represents the counter-factual equlbrum. Hence R (0) represents the tax revenue n benchmark equlbrum whle R () represents the tax revenue n counterfactual equlbrum. Q L s the Laspeyres prce ndex, and P m () and P m (0) represent the D new equlbrum and the benchmark equlbrum prce vector respectvely. x m (0) denotes the consumpton level of good m n benchmark equlbrum. (2) The CIT rate cut wth other Tax rate Increase : Scenaro B and C 4

These proposals may be consdered snce the other tax rate must be ncreased to keep real tax revenue constant when the CIT rate s cut. Here we consder two scenaros. Scenaro B s to ncrease the ndvdual ncome tax rate when the CIT rate s cut. Scenaro C s to ncrease the value-added tax rate when the CIT rate s cut. For both tax reform proposals the magntude of the CIT rate s 0%, 20%, 30%, 40%, and 50% of current rate for smulaton. (3) The CIT rate cut wth Government Expendture Decrease : Scenaro D and E If the government ncreases other taxes to keep ts revenue constant when t s the CIT rate, t should meet the resstance from tax payers who wll bear more tax burdens after tax reform. Thus, more realstc tax reform proposal for the CIT s to government expendtures n accord wth the CIT revenue reducton. There are two sorts of government expendtures government consumpton expendtures and transfer payments. Therefore, there are two knds of tax reform proposals. Scenaro D s to cut the CIT rate and government consumpton expendtures accordngly, Scenaro E s to cut the CIT rate and government transfer payments accordngly. These two scenaros must satsfy the government balanced budget constrant, whle scenaro B and C must satsfy the equal tax revenue constrant. The magntudes of the CIT rate are 0%, 20%, 30%, 40%, and 50% for scenaro D and E 2. Economc Effects of Corporate Income Tax Reform Measures () Welfare Effects The mpact of specfc tax reform on the effcency of the economy can be calculated by ether by the Equvalence Varaton (EV) or the Compensatng Varaton (CV) measures. In ths paper, we use the Hcks Compensatng Varaton depcted as follows CV E u p E u p N N 0 0 = (, ) (, ) (37) Here, the functon E( ) represents the expendture functon. u 0 and p 0 are the utlty level and prce vector respectvely under the ntal benchmark equlbrum. u N and p N 5

are the utlty level and prce vector respectvely under a new counter-factual equlbrum after changng the polcy varables. E(u N, p N ) s the expendture level needed to keep the utlty level u N under the new prce vector p N. The Compensatng varaton represents the ncome compensaton needed to keep the consumer at the ntal utlty level as the prce vector changes. When consumer s preference are assumed homothetc, Compensatng Varaton CV can be expressed as follows CV u u u N 0 N = I (38) N Here I N s the ncome level n the counter-factual equlbrum. The total Compensatng Varaton s the sum of Compensatng Varaton of each ncome class. That s H CV = CV j= j where CV j s Compensatng Varaton for ncome class j, and H s total number of the class n ths model and H=0. (2) Income Dstrbuton Effects The mpact of specfc corporate ncome tax reform on the equty of the economy can be estmated ether by the Gn Index or the Decles Dstrbuton Rato. The latter s measured by the rato of total ncome of the lower 20% ncome class to total ncome of upper 20% class. The hgher the Decles Dstrbuton Rato s, the farer the ncome dstrbuton. The lower the Gn Index the farer the dstrbuton of ncome. (3) Effects on Labor Supply and Savng Dependng on whether corporate tax reform ether encourages or dscourages each ncome group s wll to work and wll to save, ncome dstrbuton and welfare of the economy change. So, t s mportant to measure how corporate tax reform affects labor supply of each ncome group (L j, j=,, 0) and total labor supply( 0 Lj = L ) j= n the economy. Furthermore, t s also mportant to measure how corporate tax reform affects savngs of each ncome group (S j, j=,, 0) and total savngs( 0 Sj = S ). j= 6

(4) Effects on ndustral producton and factor ncome To analyze the corporate tax ncdence, the effect of corporate tax reform on producton by each ndustry (Q, =,,26) and on factor ncome need to be analyzed. By comparng ntal factor prces ( P (0), P (0) ) wth factor prces after tax reforms L ( P (), P () ), we analyze the splt of corporate tax burden ether to the labor ncome L K or to the captal ncome. The corporate tax reform effect on tax revenue can also be measured. K IV. SIMULATION RESULTS Smulaton results of fve Korean corporate reform scenaros are summarzed as <Table > Effcency CV Income dstrbuton A Integrate CIT (a) &PIT <Table > Smulaton Results B CIT (a) & PIT C CIT (a) & VAT 0% All CIT 20~50% proportonally D CIT (a) & Govt Exp All CIT 40% max E CIT (a) & Tansfers 0% 20~50% Improved Improved Worsened Worsened Improved Revenue Unchanged Unchanged Unchanged 0%.7% 50% 7.8% Producton (c) (c) (c) (c) Tax Incdence (b) Labor Supply PL (.046%) PK (7.977%) P L (0.570%) PK (4.494%) P L (0.502%) PK (4.500%) PL (.346%) PK (2.50%) P L (0.84%) PK (5.327%) (c) (c) (c) (c) Savngs (c) (c) (c) (c) Note: (a) Corporate Income Tax rates are reduced by 0%, 20%, 30%, 40%, 50% for each scenaro. (b) Changes n PL and P K are for the case of 50% CIT rate reducton. (c) Sze of (ncrease) n ndustral producton s proportonal to sze of CIT rate reducton 7

. Welfare Effects For fve dfferent corporate ncome tax reform scenaros, the welfare effects are analyzed. The results are summarzed as <Table 2>. Frst, n cases (scenaro B and C) of reducng corporate ncome tax rate by 0%, 20%, 30%, 40%, 50% respectvely and ncreasng other tax rates for the equal revenue constrant, total welfare of the economy s d. Meanwhle, n case (scenaro D) of reducng the government consumpton expendtures to compensate tax revenue reducton, total welfare of the economy s ncreased generally. Second, the case (scenaro D) of reducng the government consumpton expendtures ncreases the welfare of the economy more than case (scenaro E) of reducng the transfers. Increasng the economc welfare s measured as about 5.4 trllon Won (2.02% of GDP), n case (scenaro D) of reducng the government consumpton expendtures to compensate 40% of CIT tax rate reducton. Thrd, Integraton of corporate and personal ncome tax (scenaro A) s total welfare of the economy by about.3 trllon Won (0.47% of GDP). Fourth, only the case (scenaro B) of reducng corporate ncome tax rates and ncreasng personal ncome tax rates by 0% ncreases total welfare of the economy. In all other case of ncreasng tax rate, especally ncreasng value-added tax rate, total welfare s. So, t should need to be careful to the CIT rate and to ncrease the other tax rates, consderng welfare reducng effect as well as the tax resstance effect. <Table 2> Welfare Effects of CIT Reform Scenaros by Measurng Changes n CV (Unt: Bllon Won) A B C D E CIT Rate Integraton of CIT & CIT & CIT & CIT & Reducton CIT & PIT PIT VAT Govt Exp Tansfers 0% 0% 08.2-2.4,08.9 20~50% -,258.6 20% -75.2-705.4,687.2-53.0 30% -249.9 -,83.5 2,23.9-27.2 40% -46.3 -,646.4 5,374. -373.2 50% -574.6-2.095.0 3,238.9-52.4 8

2. Income Dstrbuton Effects The smulaton results of corporate ncome tax reform on the Income dstrbuton of the economy are summarzed as <Table 3>. If Gn Index after tax reform s d compared to the benchmark equlbrum ndex level, 0.27858, then the dstrbuton of ncome s mproved. Corporate ncome tax burden splts to labor and captal ncome. In cases (scenaro B) of reducng CIT rate and ncreasng PIT, ncome dstrbuton s worsened, due to progressve Korean ncome tax scheme. Meanwhle, n case (scenaro C) of ncreasng VAT rate, ncome dstrbuton s mproved, due to regressvty of VAT tax scheme. What s nterestng s the case (scenaro D and E) of reducng the government expendtures. Reducng the government consumpton expendtures makes worse the ncome dstrbuton, meanwhle, reducng the government transfers mproves the ncome dstrbuton. The beneft of government expendtures s equal to all ncome groups, due to ts non-exclusveness. Ths makes ncome dstrbuton worse. Government transfer s proportonal to ncome level of each ncome group. So the burden of reducng transfers s progressve and the ncome dstrbuton s mproved. If transfers are dstrbuted dfferently, ncome dstrbuton results of transfers change as well. <Table 3> Income Dstrbuton Effects of CIT Reform Scenaros by Measurng Changes n Gn Index A Integraton of CIT & PIT CIT Rate Reducton B CIT & PIT C CIT & VAT D CIT & Govt Exp E CIT & Tansfers 0% 0.27795 0.27899 0.27885 0.27834 20% 0.27736 0.27938 0.2790 0.278 0.27340 30% 0.27679 0.27975 0.2796 0.27789 40% 0.27625 0.2800 0.2793 0.27768 50% 0.27573 0.28042 0.27944 0.27747 mproved Income Dstrbuton mproved worsened worsened mproved In KOCGE model, government transfer s defned as socal securty beneft from government. Accordng to Annual Report on the Famly Income and Expendture Survey, socal securty beneft for each ncome group s progressve. 9

Table 4a and Table 4b show the effect on labor supply and savngs of fve dfferent polcy changes for the corporate ncome tax. Let us begn to dscuss the effect of the corporate ncome tax change on labor supply. Labor supply appeared to n both the polcy scenaros A and B whereas scenaro A nvolves the ntegraton of corporate ncome tax nto the ndvdual ncome tax system and scenaro B ndcates an ncrease n ncome tax burden. Such the n the labor supply n overall economy resulted from the work ncentve s due to the after tax labor ncome caused by the rse of the ncome tax rate. The fundamental cause of the labor supply due to the ncome tax rate rse s that the wage elastcty of lesure demand s -0.2. <Table 4a> The Effect of the Corporate Income Tax change on Labor Supply (base equlbrum labor supply=50,075, ten bllon Wons) Scenaro A % of the ntegraton of reducton of corporate corporate ncome tax and ncome tax ndvdual rates ncome tax Scenaro B corporate Scenaro C corporate Scenaro D corporate Scenaro E corporate ncome tax rate ncome tax rate ncome tax rate ncome tax rate & ndvdual ncome tax ncrease & VAT ncrease & government expendture & transfer expendture 0% fall 49,954(-0.08)49,973(-0.068)49,94(-0.08) 50,227(0.0) 49,085 (-0.660) 20% fall 49,840(-0.57)49,878(-0.32)49,820(-0.70) 50,37(0.97) 30% fall 49,732(-0.229)49,788(-0.9)49,73(-0.228) 50,509(0.289) 40% fall 49,628(-0.298)49,705(-0.247)49,648(-0.285) 50,639(0.376) 50% fall 49,529(-0.304)49,625(-0.300)49,569(-0.337) 50,765(0.459) labor supply labor supply labor supply labor supply labor supply Note: ) Per unt labor supply s computed as Won. 2) The number n parenthess represents the percentage rate of change. The rate of change ndcates the rato of ncrease or to the base equlbrum. The scenaro C of the n the corporate ncome tax rate and the ncrease n the value added tax rate also turned out to labor supply. The n the labor supply shown n scenaro B s greater than that n scenaro C. Ths s because the rse of the value added tax rate ndrectly s the after tax labor ncome through the prce rse of consumpton goods followed by the n the real ncome, whle the rse of the ncome tax rate s the after tax labor ncome 20

drectly The polcy scenaro D representng the n both the corporate ncome tax and the government expendture s labor supply, whle the scenaro E decreasng the transfer ncome ncreases labor supply. The reason why the scenaro D s labor supply eventually s that dmnshng work ncentve s due to the fact that the n the government consumpton expendture from the corporate ncome tax nduces both the aggregate demand and wage to. Table 4b summarzes the effects of the corporate ncome tax change on the household savngs. Both A and B appear to savngs. The reason why the household savngs due to the ncome tax ncrease followed by the corporate ncome tax s that the elastcty of savngs to the nterest rate s 0.4%. Therefore the ncome tax ncrease lowers the after tax nterest rates followed by the reducton of the household savngs <Table 4b> The Effect of the Corporate Income Tax change on Savngs (base equlbrum labor supply=4,29, ten bllon Wons) Scenaro A % of the Scenaro B Scenaro C Scenaro D Scenaro E ntegraton of reducton of corporate corporate corporate corporate corporate corporate ncome tax rate ncome tax rate ncome tax rate ncome tax rate ncome tax and ncome tax & & & & ndvdual ncome tax rates ndvdual ncome tax ncrease VAT ncrease government expendture transfer expendture 0% fall 4,207(-0.09) 4,305(0.598) 4,37(0.682) 4,263(0.303) 20% fall 4,95(-0.73) 4,386(.67) 4,399(.26) 4,304(0.59) 4,6(-0.725) 30% fall 4,83(-0.256) 4,463(.70) 4,477(.33) 4,343(0.864) 40% fall 4,73(-0.33) 4,537(2.229) 4,553(2.34) 4,380(.25) 50% fall 4,62(-0.404) 4,607(2.728) 4,625(2.848) 4,45(.375) n n n n n savngs savngs savngs savngs savngs Note: The number n parenthess represents the percentage rate of change. Savngs are shown to ncrease n all the scenaros such as scenaro C (corporate ncome tax reducton and the VAT ncreases, scenaro D ( n the government consumpton expendture) and scenaro E (the n both corporate ncome tax and the transfer expendture). When we consder the effects on labor supply and savngs, the n both the 2

corporate ncome tax and the government expendture s more effectve than the corporate ncome tax reducton together wth other tax ncreases. 3. The Effect on Producton Across Industres <Table 5> The Effect of the Corporate Income Tax Change on the Industry Producton Scenaro A ntegraton of corporate ncome tax and ndvdual ncome tax 49,085(-0.660) Total producton (Base Equlbrum Industry Producton = 580,432 unt Bllon Wons) % of the reducton Scenaro B corporate Scenaro C corporate Scenaro D corporate Scenaro E corporate of ncome tax rate ncome tax rate ncome tax rate ncome tax rate corporate & & & & ncome tax ndvdual VAT ncrease government transfer rates ncome tax ncrease expendture expendture 0% fall 49,954(-0.08)49,973(-0.068)49,94(-0.08) 50,227(0.0) 20% fall 49,840(-0.57)49,878(-0.32)49,820(-0.70) 50,37(0.97) 30% fall 49,732(-0.229)49,788(-0.9)49,73(-0.228) 50,509(0.289) 40% fall 49,628(-0.298)49,705(-0.247)49,648(-0.285) 50,639(0.376) 50% fall 49,529(-0.304)49,625(-0.300)49,569(-0.337) 50,765(0.459) Total producton Total producton Total producton Total producton Note: ) Per unt labor supply s computed as Won. 2) The number n parenthess represents the percentage rate of change. The rate of change ndcates the rato of ncrease or to the base equlbrum. In Table 5, the scenaros B and C both are shown to the total ndustral productons, ndcatng that the degree of n the total ndustral producton s proportonal to the corporate ncome tax rate. But producton levels appeared to ncrease n such ndustres as general machne, precson machne, constructon ndustry, real estate and busness servce. The scenaro D appeared to the total ndustral producton, whle the scenaro E d t. Ths also ndcates that the changes n the ndustral total producton become greater n the proporton of the level of fall n the corporate ncome tax rates. In the case of the scenaro D, producton levels are shown to ncrease n all the ndustres except the ndustres of ceramc, the frst metal, metal product, general machne, electrcty and electronc meter, precson meter and constructon. Those ndustres appear to obtan the benefts from the corporate ncome tax rate reducton 22

regardless of scenaros. In the case of scenaro E, on the other hand, producton levels are shown to ncrease n all the ndustres except the ndustres of textle and leather, wholesale and retal, telecommuncaton, socal and servce. 4. The Effect on Tax Revenue In prncple, tax revenue should be unchanged n the scenaro of the ncrease n other tax system along wth the abolshment of the corporate ncome tax. Thus, the changes n tax revenue are found only n the corporate ncome tax and the government expendture. When the corporate ncome tax rates fall by 0%, 20%, 30%, 40%, 50% and the government expendture, tax revenues appear to by 2.%, 3.9%, 5.7%, 7.3%, and 8.9% respectvely. The level of the tax revenue s shown to be smaller n scenaro E than n scenaro D. V. POLICY IMPLICATIONS. Polcy Implcaton from the Smulaton of the Corporate Income Tax Changes The smulaton results of the corporate ncome tax change usng our computatonal general equlbrum model gve rse to the followng polcy mplcatons. Frstly, n order to compensate for the n tax revenue due to the reducton of the corporate ncome tax rates, the reducton of the government expendture appeared to be superor, n terms of welfare aspect, to the ncrease n ncome tax or consumpton tax. Ths mples that the dstorton of resource allocaton from tax ncrease s greater than that from the n the government expendture. It also mples that the reducton of the corporate ncome tax rate n 200 can be justfed and result n a desrable outcome f t comes wth the n the expendture. Unfortunately, Korea experenced lack of efforts of reducng government expendture for the 2002 budget. The second mplcaton s that n the case of reducng corporate ncome tax rate, the n the government expendture ncreases natonal savngs, whle ncome tax ncrease ends up wth reducng natonal savngs. Ths also mples that the n the personal savngs resulted from the ncome tax ncrease outweghs the ncrease n the corporate savngs. As a result, we can learn that the dstorton effect of tax also appears to be large n Korea. The thrd mplcaton nvolves the comparatve advantage of the ncome tax rase 23

over the VAT rase n terms of equty n the case of tax rase to compensate for revenue due to the corporate ncome tax rate reducton. Ths s because both progressvty of ncome tax and the regressvty of VAT nduce the superorty of ncome tax rase. The fourth and fnal mplcaton s that the ntegraton of corporate ncome tax nto ncome tax (abolton of corporate ncome tax n other word) gves rse to negatve effects on savngs, labor supply and effcency all together wthn context of our model. Ths ndcates that abolshng corporate ncome tax on the purpose of effcency gan cannot show any ncrease n effcency n our smulaton. 2. Issues nvolved n changng the corporate ncome tax law Four polcy mplcatons derved from smulaton usng our KOCGE model can be dfferent f we apply some other assumptons and parameter values. However, our senstvty analyss results ndcate that major lessons from the conclusons and polcy mplcatons would be unchanged. Our conclusons and mplcatons can provde a momentum for the more detaled and rgorous dscusson about the tax polcy change. In partcular, further dscusson of the corporate ncome tax change requres a full consderaton of crcumstance and tax envronment faced by Korean companes. An mportant polcy suggeston based on the smulaton results gves rse to the efforts of reestablshment and reducton of varous tax reductons or exemptons and quas taxes for the revson of the corporate ncome tax system. Broadenng the tax base of the corporate ncome tax from the efforts enable the tax rate reducton to play a more mportant role n actvate the economy and enhance the nternatonal compettveness. More specfcally, as meetng the open economy and nternatonalzaton, such efforts can construct a base for ntroducng foregn captal and actvatng nvestment through the tax rate reducton. In addton, more efforts should be gven to the smplfcaton of procedures and forms for the tax exemptons used by the small and medum enterprses, through whch more effcent assstance for those enterprses can be guaranteed. REFERENCE Ahn, Jong-Seok, An Analyss on the Effect of Tax Polcy on Investment and Savngs n Open Economy, (n Korean) Revew of Fscal and Fnancal Studes, Vol. 3, No. 2., 996. 24

Auerbach, A. and Joel Slemrod, The Economc Effects of the tax Reform Act of 986, Journal of Economc Lterature, 997, pp. 589-632. Ballard, Charles, Don Fullerton. John Shoven, and John Whalley, A General Equlbrum Model for Tax Polcy Evaluaton, Chcago: Unversty of Chcago Press, 985. Bradford, Davd, Factor Prces May be Constant, but Factor Returns are Not, Economcs Letters, Vol., No. 3, 978, pp. 99-203. Cho, Kyekhen, Analyss of the Effectve Tax Rate n Korea: for the 9 Sectors n the Manufacturng Industry, (n Korean) The Korean Journal of Publc Fnance, Vol., No., 996, pp. 85-24. Fuchs, Vctor, Alan Krueger, and James Poterba, Why Do Economsts Dsagree about Polcy? The Role of Belefs About Parameters and Values, NBER Workng Paper 65, 997. Gravelle, Jane and Laurence Kotlkoff, The Incdence and Effcency Costs of Corporate Taxaton When Corporate and Noncorporate Frms Produce the Same Goods, Journal of Poltcal Economy, Vol. 97, 989, pp. 749-790. Gravelle, Jane and Laurence Kotlkoff, Corporate Taxaton and the Effcency Gans of the 986 Tax Reform Act, Economc Theory, Vol. 6, 995, pp. 5-8. Harberger, Arnold, The Incdence of the Corporaton Income Tax, Journal of Poltcal Economy, Vol. 70, 962, pp. 25-224. Harbereger, A. C., Effcency Effects of Income from Captal, Effects of Corporaton Income Tax, ed. M. Krzyzanak, Wayne State Unv. Press, 966. Harberger, Arnold, The ABC's of Corporate Tax Incdence: Insghts nto the Open Economy Case, Tax Polcy and Economc Growth, Washngton D.C.: Amercan Councl for Captal Formaton, 995. Kang, Dong-Ku, A Study on the Corporate Income Tax Reform Measures n Korea, (n Korean) Korea Insttute of Publc Fnance, 995. Km, Chongung, The Excess Burden of Corporate Income Tax and Indvdual Captal Income Tax, (n Korean) Kyungsan Unversty Paper Revew, Vol. 4, 996, pp. 7-33. Km, Sung Ta, I.S. Y, C. A, and S. Lee, Estmaton of Excess Burden from Income Tax, Value-added Tax, and Commodty Tax n Korea: A Computable General Equlbrum Analyss, (n Korean) presented at the Internatonal Insttute of Publc Fnance n Jeju Island, Korea, August 2005. Km, Sung Ta, I.S. Y, C. A, and S. Lee, Welfare Effects of Reducng the Tax Evason Rate: A Computable General Equlbrum Analyss for Korea, (n Korean) presented at the conference Korea and the World Economy II, Seattle, USA, August 2003. 25

Km, Yu-Chan, The Effect of the Corporate Tax Rate Reducton on Investment, (n Korean) Revew of Fscal and Fnancal Studes, Vol. 2, No., 995. Kotlkoff, Laurence and L. H. Summers, Tax Incdence, In Handbook of Publc Economcs, edted by Alan J. Auerbach and Martn S. Feldsten, Amsterdam: North-Holland, 987. Kwack, Tae-Won, A Study on the Estmaton of Tax Admnstraton Cost n Korea, (n Korean) Korea Insttute of Publc Fnance, 995. Lee, June-Q, Equty of Captal Taxaton and the Method of measurng Corporate Tax Burden, (n Korean) Socal Scence Bulletn, Kyunghee Unversty, 99, pp. 239-258. Lee, D.K. and Y.R. Km, An Emprcal Study on Imputed Tax on Tax Favored Frm and the Corporate Income Tax Burden, (n Korean) Industral Economcs Study, Vol. 2, No.4, 998, pp. 45-478. Lee, Myung-Hoon, Taxaton on Unexplaned Interest Payments for Corporate Income Tax, (n Korean) The Korean Journal of Publc Fnance, Vol. 8, 993, pp. 3-60. Mutt, John and Harry Grubert, The Taxaton of Captal Income n an open Economy: The Importance of Resdent-Nonresdent Tax Treatment, Journal of Publc Economcs, Vol. 27, 985, pp. 29-309. Park, Kback and Myungheon Lee, A Study on the Estmaton of Tax Revenues, (n Korean) Research Report of KIPF 00-03, 2000. Rhee, Woo-Tak, A Study on the Evaluaton of and Reform for Corporate Taxatons, (n Korean) Taxaton Studes, Vol. 3, 999, pp. 67-00. Roh, H.S. and T.S. Km, Relatonshp between Frm Sze and the effectve Corporaton Income Tax Rate, (n Korean) Korean Tax Journal, Vol., 996, pp. 57-88. Shoven, J. B. and J. Whalley, Applyng General Equlbrum, Cambrdge Surveys of Economc Lterature, 992. Sbert, Anne, Taxng Captal n a Large Open Economy, Journal of Publc Economcs, Vol. 4, 990, pp. 297-39. Slemrod, Joel, Taxaton Wth Internatonal Captal Moblty, In Uneasy Compromse: Problems of a Hybrd Income-Consumpton Tax, edted by Henry J. Aaron, Harvey Galper, and Joseph A. Pechman, Washngton D.C.: The Brookngs Insttuton, 988. Slemrod, Joel, Free Trade Taxaton and Protectonst Taxaton, Internatonal Tax and Publc Fnance, Vol. 2, 995a, pp. 47-489. Slemrod, Joel, Professonal Opnons About Tax Polcy, Natonal Tax Journal, Vol. 48, 995b, pp. 2-47. 26

Won, Yunh, An Analyss on Captal Income Tax through Estmatng the Effectve Tax Rate, (n Korean) Korean Tax Journal, Vol., 996, pp. -56. Yoo, B.H., An Analyss on the Effectve Corporate Income Tax Rate n Korea, (n Korean) Yungnam Commerce Bulletn, Vol., 993, pp. 245-264. Yoo, S.K. Reform of Frm Accountng Standard and Measures for the Corporate Income Tax, KIPF Polcy Report, 97-09, 997. 27