PhD THESIS KAPOSVÁR UNIVERSITY FACULTY OF ECONOMIC SCIENCE. Department of Finance and Economics. Head of PhD School: DR. SÁNDOR KEREKES Professor

Transcription

1 PhD THESIS KAPOSVÁR UNIVERSITY FACULTY OF ECONOMIC SCIENCE Deprtment of Finnce nd Economics Hed of PhD School: DR. SÁNDOR KEREKES Professor Thesis topic supervisor: DR. TAMÁS BÁNFI Professor Co-thesis topic supervisor: DR. ANETT PARÁDI-DOLGOS Associte professor APPLICATION POSSIBILITIES OF PROGRESSIVE CONSUMPTION TAX Prepred by: ERZSÉBET TERÉZ VARGA KAPOSVÁR 2014

2 1 TABLE OF CONTENS 2 RESEARCH BACKGROUND, AIMS MATERIAL AND METHODOLOGY RESULTS Results of efficiency comprisons Results of equity comprisons Results of neutrlity comprisons CONCLUSIONS NEW RESEARCH RESULTS RECOMENDATIONS PUBLICATIONS IN THE FIELD OF THE DISSERTATION

3 2 RESEARCH BACKGROUND, AIMS There is no consensus in economic nd more specificlly public finnce thinking regrding the priority of consumption nd income tx. I im to demonstrte in my thesis tht insmuch progressive consumption txtion is estblished, it is more suitble thn progressive income tx from the spects of most criteri, especilly tht of equity nd efficiency. The possibility hs been given since Fisher (1942) nd Káldor (1955) but in my thesis I will lso consider the questions of fesibility nd dministrtive burdens involved Although it is n old question my thesis looks t, it is nevertheless burning one currently. It is sign of the unhelthy stte of the economy tht in the period prior to the current crises the popultion got into net creditor position for which eqution two (prllel) fctors re responsible: the low level of svings nd credit covered consumption. Expenditure tx would be beneficil for both phenomen which inspire us to compre it to the income tx reigning currently. In Fisher s time excessive sving ws chrcteristic so the discourging of consumption wsn t desirble, which is one reson why his ide wsn t received with gret enthusism. The current er of crises mngement is subject to the sme evlution so this is not necessrily the time when consumption bsed tx should be introduced even if it might be beneficil in the long run. The view, tht liner or flt txes s they re more populrly known re not so beneficil s they were expected to be, is more nd more justified nowdys. It ws the simultneous presence of other fctors tht my be credited for the economic growth tht ws felt t the plce of the introduction of flt rte txtion. (see for exmple Erdös, 2012) The presumbly short lived liner txes will most likely be overtken by the ppliction of progressive txtion This is nother reson why it is especilly importnt to introduce nd describe the systems of progressive txtion nd the expnsion nd ppliction of the terms 2

4 used to describe liner txes by tx studies. We cn hrdly find scientific publictions which ssume progressive rtes of txtion. In my study I m compring nd contrsting progressive consumption tx with progressive income tx long importnt fetures which hve formed the bses of the expecttions of the given system of txtion for centuries. The weight of the criteri vry from er to er nd from writer to writer but non of them my be neglected. Efficiency, equity nd neutrlity re mong the spects to be exmined. (1) According to my first hypothesis, in one-period model the effect on efficiency of equivlent progressive consumption nd income txes cn be expected to be the sme. (2) Contrry to this, I expect bsed on my second hypothesis tht considering more periods progressive consumption tx doesn t distort intertemporl consumption decisions s much s progressive income tx s the ltter punishes fluctuting income flow. (3) In model encompssing more time periods we re ble to discuss svings too so ccording to my third hypothesis income tx levied on cpitl income distorts intertemporl decisions to greter extent thn consumption tx. (4) In my fourth hypothesis wht I expect is tht progressive consumption tx better stisfies the criterion of equity thn progressive income tx. (5) According to my fifth hypothesis progressive consumption tx stisfies the criteri of neutrlity while income tx hurts this criterion in some cses. 3

5 3 MATERIAL AND METHODOLOGY I m bsiclly mking use of the normtive pproch to public finnce in my study while trying to find the nswer to the question of which system of txtion would be better from society s point of view. In this process I cn t void the positive questions of public choice: to wht extent cn either system sfegurd ginst cheting nd wht re their respective dministrtionl burdens. I will del with these topics in the literture review section of my study. I will nlyse the prticulr tx system criteri in prtil model, since the tx theory literture often mkes do with thinking in single decision spce even in the compertive nlyses of liner txes, thereby limiting the number of endogen vribles. Nevertheless I will lwys show the vrious income pths of vrious individuls in order to study in ech period the decision mking position of people with high, low nd fluctuting incomes. Tx theory reserch mesure the efficiency of txes by the ded weight loss cused by them. Ded weight loss cn be clculted by consumer nd producer surpluses s well s from the compenstionl or equivlent vribles. However, these re bsed on uncertin estimtion processes nd subjective fctors. In itself, in order to demonstrte the presence of ded weight loss, it is sufficient to exmine the price rte of budget constrint relting to the tx. Price is n objective, evident dt so this is wht I concentrte on in my nlyses I will only concentrte on the question of whether in the given country the price rtio of the goods chnges s result of the txes nd whether or not ded weight loss is creted. During my equity comprison the nswer I ws looking for is how the choice of tx bse influences inequlity. Tht is, does progressive consumption tx or progressive income tx decreses the inequlities to greter extent present in society? In my disserttion I will clculte in two time period model 4

6 whether it is progressive income tx or progressive consumption tx decreses the Gini coefficient (used to indicte inequlity in socil income distribution) I used the tx pyer s life time income to describe the shping of neutrlity nd I presented n exmple when progressive income tx chnged the rnk order of individuls. Besides the consumption tx I demonstrte tht n individul s txed life time income cn only be higher thn n other individul s if the sme reltion ws true for the before tx income. I need to determine the consumption creer for both the efficiency nd the equity comprisons besides the income creer. Insted of the minstrem utility mximlising individul I ssumed consumption smoothing individuls, which is generlly ccepted, less restricting method for ssessment. Consumption smoothing is the observble phenomen tht throughout their life the consumption pttern of individuls follows more or less stedy pth.it is much less voltile thn their income, which cn hve fluctuting vlue throughout their life. The consumption functions used in mcroeconomic models grb this behviour in mny different wys. In my disserttion I will demonstrte the concepts of Keynes, Modiglini, Friedmn nd lso the bufferstocks nd hbit-persistence theories. The newer nd newer models del with the shortcomings of the previous models but in the end they explin the sme observtions: wht consumer decision precedes the phenomen of consumption smoothing. In the multitemporl model of efficiency, equity nd neutrlity I consider tx pying obligtion for both consumption nd income tx only in the period in the budget constrints in which they cn be defined s income or consumption. This is evident in the cse of income tx s the employer continuously deducts the tx from the wges of the employees who hence gets the net, fter tx ernings. In cse of consumption tx the mgnitude of the tx cn only be determined lter s it not only depends on the income erned in the period but lso on the end of your bnk ccount blnce mong others. So, in 5

7 order to ensure continuous revenues for the stte, I think it is necessry to py tx dvnce in the bse period, the bse of which my be the previous period. If we ccept the ssumption of consumption smoothing, we won t (significntly) devite from the tx burden which is determined lter. I must emphsize here the strictness of the models nd formultions which obviously limit the gined results. One such importnt limittion is tht the consumers do not leve n inheritnce, nd do not inherit nything themselves nd hve no welth t the strt of their lives. These re very strong ssumptions, even if mny economic models re bsed on them. 6

8 4 RESULTS 4.1 Results of efficiency comprisons The comprison of progressive consumption tx nd lbour income tx during the intrtemporl choices mde on the consumption of vrious goods. Let us exmine how progressive consumption tx nd lbour income tx influence the decision set of individuls, ssuming one period. Let s strt from the position tht the consumer s (exogene for now) income must cover his or her expenditures, which my be consumption driven or public duties to the stte. 1 It is esy to see tht given the correltions below, within the sme period the decision spce of prticulr types of tx will be the sme nd there won t be ny distorting effect. t j j, j (,f) (4.1) 1 j C Y( 1 ), (4.2) Where t: tx rte of expenditure tx nd τ: tx rte of income tx. Prcticlly, in single time period model neither the generl progressive consumption tx (covering ll products) nor progressive income tx modifies the reltive prices of given goods so neither of the txes distort the consumer decision bout which good to choose. As soon s we mke the individuls income endogen, through choosing n income the individul will be ble to influence his or her tx duties nd his or her decision will be distorted. 1 As we re tlking bout single period, we don t hve to consider svings 7

9 4.1.2 The comprison of progressive consumption tx nd lbour income tx in light of the intrtemporl choices of work nd leisure time. Let s consider the cse when the income of the individul is bsed on his or her decision, which is the result of the choice between time llocted to work nd leisure. I strt from the fct tht the time people hve is limited (for exmple 24 hours dy or 168 hours week.) If the individul spends time on work, he or she will receive given mount of externlly determined remunertion which cn be spent on consumption but fter certin level of consumption, spending time on leisure might be worth more to the individul thn on work. Let us further suppose tht the decision mker lives only certin mounts of time, nd cn consume single product, which product should be numerire good by nture, so its price should be unit. By choosing the mount of time spent on work, the individul cn optimise the wellbeing chievble by work nd leisure. We cn write the budget constrint (without txes s yet) of the decisions ccordingly: C Y w (L L) (4.3) Where: C: mount of consumption, Y: erned (work) income, w :rel hourly wge (expressed in numerire good), : ll the time t the individuls disposl, L: mount of leisure. With progressive consumption txes, the written budget constrint in (4.4) nd (4.5) chnges thus. Y w L w L C, if Y E (4.4) 1 t 1 t 8

10 C Y E(t f f, if Y> E. (4.5) 1 t f t ) w L E(t 1 t t ) w L f In the consumption - leisure spce, the consumption constrint breks t the L' L (1 t ) E / w leisure mount s cn be seen in figure 4.1. In cse of consumption less or more thn L, scrificing one unit of leisure will result in w/( 1 t f ) nd w/( 1 t ) mke up consumption. Besides, the budget line does not only brek but lso chnges from the originl. Its steepness chnges on both sections s the originl exchnge reltionship ws the w rel wge so the choice is distorted. There is lso substitution effect besides the income effect Grph: Budget constrint of leisure - consumption (originl work) Let s look t the decision sitution bove with progressive lbour income tx. The budget constrint chnges ccording to eqution (4.6) nd (4.7). C Y(1 ) w (L L)(1 ), if Y Y (4.6) C w(l L)(1 ) Y( ), if Y> Y. (4.7) f f The constrint lso brkes in this cse, notbly t the L '' L Y / w point. In cse of choosing more or less free time thn L - the unit leisure scrifice w 1 ) or w 1 ) leds to dditionl consumption. So the originl ( f ( exchnge position (-w) hs chnged in this sitution, too. The tx distorts the decision. Since both txes distort the choice, the question is which one does more so. It is esy to see tht in the cse of equivlent rtes nd bnds the effects 9

11 of the txes re exctly the sme s the breking points re mthemticlly the sme (L = L ) nd the steepness ( w/( 1 t f ) = w 1 ) nd w/( 1 t ) = ( f w 1 ) ). Tht is, for both the consumer tx nd the lbour income tx there is ( substitution effect in the choice between work nd free time thereby distorting the economic decisions. Bsed on subsections nd 4.1.2, in single time period model there is no efficiency difference between progressive consumption tx nd progressive income tx, their distorting effect in the sme decision sitution is the sme mgnitude Compring nd contrsting progressive consumption nd lbour income tx during consumption choices mde in different periods. Let us exmine the intertemporl decision fced by the consumer hving exogen income. Let us strt gin by nlysing the behviour of progressive consumption tx. Looking t two time periods, the budget constrint will turn out ccording to Figure 4.2. where the totl consumption in given periods is mesured on the different xes. The dotted line denotes the bnd borders of consumption tx (C ) t which the budget constrint is broken due to the progressive tx rte just s it hppened in cse of the work In the vrious periods the rte bnds divide the goods spce into four prts nd the budget constrint is differently modified in the different prts. In section 2 nd 3 the originl nd the fter tx budget constrints re prllel s their steepness is the sme, their vlue is: -(1+r). The result of this prllelism is tht there is no substitution effect in either sections 2 or 3 only income effects. The consumer gets into these situtions if he or she wnts to smooth consumption. Assuming consumption smoothing, the consumption creer is typiclly rther stright throughout one s life so from here on our point 10

12 of deprture will be tht the tx pyer devites from the bnd border in both periods in the sme direction, mening tht he or she will choose good depending on income in section 2 or 3. In wht follows section 2 nd 3 will be regrded s the relevnt decision plne. Figure 4.2. Budget constrint in cse of progressive consumption tx. 4.1 Tble Income flow sub cses ( Own design) Income flow: Y1 Y Y2 Y Low Y2 Y Incresing Y1 Y Decresing High We must differentite four types of income flow in order to nlyse the effect of income tx. These re summrised in tble 4. The so clled high nd low income flow denote those cses where someone hs lbour income bove nd below the income tx bnd border ( ) in both periods.the income flow is incresing or decresing for the person who hd lbour income bove or below the tx bnd border in the first period nd lbour income below or bove in the second. Bsed on the clcultions, we found tht the lst two income flows behves the sme wy with respect to txtion so the number of cses to 11

13 HIGH income flow UNSTEADY income flow LOW income flow be exmined my be reduced to three. ( low, unstedy nd high income pths) Low consumption pth Borderline consumption pth High consumption pth Figure 4.3.Budget constrints fter the introduction of progressive lbour income tx nd consumption tx in cse of different consumption nd income pths (own design bsed on own clcultions ) The consumption pth cn turn out in three different wys depending on the life pth income (ssuming consumption smoothing) hppening in both periods on, below or bove the rte border. Similir to the income flow, we cn cll them high, low or borderline consumption pth. In totl then, we cn imgine 12

14 three- three different income nd consumption pths which in theory cn be pired, so in totl we must exmine in 3x3 how progressive consumption nd income bsed txtion cn shpe up in intertemporl decisions. Figure 4.3. sums up the chieved results with grey denoting the budget constrints of income tx, blck is the budget constrint of consumption tx nd dotted line is border of consumer tx rte. Let us consider the cses in the relvent decision spces. (1) Low consumption pth ( C1, C2 E ) low income flow ( Y1, Y2 Y ) After levying consumption nd income tx, the budget constrins coincide.there is only income effect in both cses nd there is no substitution effect. (2) Low consumption pth ( C1, C2 E ) unstedy income flow (2. decresing Y1 Y Y2 or 2b incresing Y1 Y Y2 ) After levying consumption nd income tx the budget constrints do not coincide.although there is no substitutiton effect, the income effects differ. (2) In cse of decresing income flow the income tx Y Y)( ) sum, (2b) In ( 1 f cse of incresing income flow ( Y 2 Y)( ) /(1 r) sum extr is deduced from the life pth income discounted for the first period. (3) Low consumption pth ( C1, C2 E ) high income flow ( Y1, Y2 Y ) This is only theoreticl possibility s the person with high income flow is bound to choose high consumption pth due to utility mximlistion s he or she is covering expenditure. (4) Borderline consumption pth ( C1, C2 E ) low income flow ( Y1, Y2 Y ) This is only possible in cse of Y 1 Y 2 Y so we re only nlysing single point.so the only question tht mkes sense is looking t wht hppens in the cse of given txes if one wnts to move from consumption point ( E, E). In f 13

15 cse of income tx there is still no substitution effect. There is in cse of consumption tx but the decision mker must hve rther unrelistic preferences to be even worth mking move. Hence in cse of consumption smoothing there is prcticlly no difference between consumption nd income tx. (5) Borderline consumption pth ( C1, C2 E ) unstedy income flow (5. decresing Y1 Y Y2 or 5b. incresing Y1 Y Y2 ) If the decision mker hs unstedy income flow the borderline consumption line cn be chieved by consumption tx but t the sme level of life time income this cn t be chosen with income tx becuse with income tx the effect of income is greter (5) with decresing income flow Y Y)( ) sum, (5b) ( 1 f in cse of incresing income ( Y 2 Y)( ) /(1 r) sum (gin discounted for f the first period). So in the relevnt decision spce in cse of mrginl chnge lthough the consumption tx distorts, there is substitution effect but even like this it secures bigger choice thn income tx since this substrcts more. (6) Borderline consumption pth ( C1, C2 E ) high income flow ( Y1, Y2 Y ) This is lso just theoreticl possibility s the person with high income flow covers his or her expenses so high consumption pth will be chosen for utility mximlistion.. (7) High consumption pth ( C1, C2 E ) low income flow ( Y1, Y2 Y ) This is not even theoreticl possibility becuse high consumption pth cnnot be ensured with low income flow. (8) High consumption pth ( C1, C2 E ) unstedy income flow (8. decresing Y1 Y Y2 8b. incresing Y1 Y Y2 ) After levying income tx nd consumption tx, the budget constrints do not coincide, lthough there is no substitution effect nywhere but the income 14

16 effects differ. (8) In cse of decresing income flow the income tx ( Y Y2 )( ) /(1 r) sum, (8b) in cse of incresing income flow ( 1 f f Y Y )( ) sum extr is deduced from the life creer income discounted for the first period. (9) High consumption pth ( C1, C2 E ) high income ( Y1, Y2 Y ) The budget constrints coincide fter levying income nd consumption tx, There is only income effect in both cses nd neither hs substitution effect. Altogether the txpyer is worse off with progressive income tx in cses (2), (5) nd (8), tht is when the income flow is unstedy between the time periods. In ny other sitution involving decisions within single time period or between periods, income tx nd consumption tx hve the sme effect on it in ll the relvent decision spheres. The only exception is the point when in both periods Y income is mde from which E is consumed becuse here the consumer tx ensures worse exchnge possibility if he or she wnts to devite from perfect consumption smoothing. We cn prcticlly sy nevertheless tht progressive income tx results in t lest s much but in cses there is even more (with unstedy income ) excess withdrwl thn consumption tx. Exmining the question from socil point of view, we do not find such serious diversion. The excess withdrwl ppering with progressive income tx (stemming from unstedy income flow) is relised for the stte. It is only the dedweight loss tht ruins the cumultive blnce of tx levier nd tx pyer for which the substitution effect is responsible. However this effect is the sme in lmost every decision mking sitution for the two tx types. There is only vrince in the ( C, C) consumption pth for the disdvntge of the consumer tx. Wht this mens in prctice is tht there is no difference between the two rteble vlues from the perspective of efficiency. They cuse the sme level of dedweight loss. Due to the differing level of tx withdrwl there is only the question of justice, to be discussed in nother chpter. 15

17 4.1.4 The expnsion of the comprison of progressive consumption tx nd lbour income tx to cpitl income tx in mking consumption choices in different periods It is not worth looking t cpitl income tx in single time period decision mking (since people mke no svings) so I did not differentite income bsed on its source. In multi period model, svings generte n interest in the next period nd they mybe levied with positive level tx. We sw in sub-chpter tht income tx does not distort intertemporl consumption nd hence sving decisions if it only burdens income tx nd not cpitl income tx. In relity it is rther rre tht government doesn t use the revenue source provided by cpitl income tx. In this chpter of my disserttion I will exmine how the intertemporl budget constrint nd hence consumer decisions re influenced by liner interest tx levied beside progressive lbour income tx. Using the previous symbols, the new budget constrint is given by equtions (4.8) nd (4.9)., if (4.8), if (4.9) Where stnds for the work income in period i, nd t k is the (liner) rte of cpitl income tx. If the consumer hs svings, tht is he or she consumes less thn his or her totl lbour income in this period, thn interest tx must be pid ccording to (4.8). The interest tx distorts the intertemporl consumption decision s it my chnge the reltive price of consumption in the vrious periods nd hs substitution effect with the consumer. Becuse if he or she would like to increse the consumption of the second period by unit, he or she must decrese tht of the first period by 1/(1+r(1-t k )) unit while before the introduction of interest tx he or she only hd to give up 1/(1+r). Consumption in the first period becme reltively cheper with the tx so it is worth substituting tht of 16

18 the second period. This leds to drop in svings nd encourges to consumer to lrger thn optiml credit uptke. There is no need to exmine the effect of liner income tx s cpitl income is lredy txed by progressive consumption tx in such wy tht it does not encourge credit uptke (t lest not lrger thn optiml) nd does not discourge svings (in cse of reltively smooth consumption, tht is, when in both periods the consumer chooses between the sme two bnd borders) Figure 4.4. Intertemporl budget constrints with cpitl income tx. (own work) 4.2 Results of Equity Comprison The construction of the Model Stylized fcts First period Tble 4.2: Pre-tx life-pth incomes Second period (w, p ) (w m, p m ) (w, p ) Y = w + w /(1+r) Y m = w + w m /(1+r) (w m, p m ) Y m = w m + w /(1+r) Y mm = w m + w m /(1+r) Supposing tht the economic individuls live for two periods, in which prt of the society erns w, while the others ( prt of society) ern w m work-relted income, where. Let us 17

19 suppose tht no-one hs ny initil welth, nd let us leve out of considertion ny possible inheritnce, tht is to sy, everyone exhusts their reserves by the end of the second period. The income flow of the two periods re independent of ech other, creting four types of income flow, s Tble 4.2 illustrtes. (henceforth denoted ) prt of the people hve life-ernings for the first period discounted s:, where i,j {,m}, r denotes the interest rte for credit lon nd the yield on svings is the sme Defining the Gini coefficient The verge bsolute devitons rte is clculted by eqution (4.10). (4.10) After levying the income tx, the bove formul is modified by clculting creer ernings with ernings fter tx, so the Gini co-efficient fter txtion is to be mrked L τ. For this clcultion, I m ssuming the income tx brckets to fll between low nd high slries, formlly to be denoted:. For the clcultion of the Gini co-efficient in reltion to consumption tx (henceforth mrked L t ), I m ssuming perfect consumption smoothing, nd hve thus determined consumption from creer-ernings (see formuls 4.11 nd 4.12 ), nd the extent of consumption tx. C Y 1 r)/[(2 r)(1 t )], ify ( 1 r)/[(2 r)(1 t )] C (4.11) ij ij ij( C Y C(2 r)( t t )/(1 r)](1 r)/[(2 r)(1 t )], [ ij f f ij ify ( 1 r)/[(2 r)(1 t )] C (4.12) ij where is the periodicl consumption of the individul subject whose life pth income is, nd the tx bnd boundries nd tx rtes develop ccording to the estblished equivlency-criterion. L t is the life-pth income decresed by consumption tx, s illustrted by formul

20 4.2.3 Clibrtion of the model In numericlly denoting the model, I m working on the ssumption tht the income tx bnd is the simple verge of the low nd high incomes.. (4.13) The low income is understood to be one unit nd the results hve been exmined with two levels of high income; in the cse of (I) s twice the low income nd in (II) s ten times the low income. (I) (4.14) (II) (4.15) The low nd high income tx hs been djusted to current income txtion rtes, nd determined s 20 nd 40%, of which consumption tx rtes re set s 25 nd 67% Results The resulting differences in the Gini co-efficient hve been formulted s follows: (4.16) Next to the consumption nd income txes, the difference of the Gini coefficient ( L) shows which tx resulted in greter reduction of socil inequlity. When the consumption tx is more productive in this respect, the L vlue is positive, wheres when the income tx is more successful in reducing inequlity, the vlue of L is negtive. 19

21 0,025 0,02 0,015 0,01 0,005-2E-16-0,005 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9-0,01-0,015-0,02-0,025 deltl (wm=10*w) deltl (wm=2*w) Figure 4.5: Differentition of the Gini coefficient with respect to low income (my own design) The results re gretly dependent upon the composition of socil levels; figure 4.5 shows the vlue of L with respect to the rtio of low incomes (p ). The following results cn be clerly determined from the figure: 1. The greter the difference between high nd low incomes, the longer the intervl nd the greter the extent to which progressive consumer txes, long with the Gini coefficient, reduce socil inequlity more thn progressive income txes. 2. Both consumer txes nd income txes (depending on socil composition) cn be both more nd less successful thn the other in decresing inequlity. Consumer tx is more successful t longer intervls; however, in cses when the rtio of low incomes is 0.55 (twice difference in slries) or 0.7 (ten times difference in slries), while income tx is more successful when the rtio of the poor in the populce is higher. I hve mde the clcultions bove ccording to n lterntive criterion of equivlence; nmely, keeping the terms relting to the tx rte (s 4.1), yet determining the consumer tx-bnd not s in 4.2 but, insted, so s to ensure 20

22 stte revenues on level with income tx. Figure 4.6 illustrtes the differences of the Gini coefficient long this line of clcultion: 0,025 0,02 0,015 0,01 0, ,005 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9-0,01-0,015-0,02-0,025 deltl (wm=10*w) deltl (wm=2*w) Figure 4.6: Difference between Gini coefficients reltive to low incomes with view to lterntive criteri of equivlency (my own design) With the following results: 1. Similr results s before: the lrger the difference between high nd low incomes, the greter the extent to which progressive consumer txes reduces socil inequlity - s mesured by the Gini coefficient - more thn progressive income. 2. Given the criterion of identicl stte revenue, progressive consumer txtion is more efficient thn progressive income txtion t reducing socil inequlity, no mtter wht the composition of society is like. 4.3 Results of neutrlity comprison According to the tx system s condition of neutrlity, the (finncil) sttus of the mrket plyers in comprison to ech other cnnot chnge (Blogh, 2003). It is this expecttion tht progressive income tx clerly does not meet, s it punishes fluctuting income-flow. Nmely, if person A nd B both hve the 21

23 sme life-pth income, but person A hs cquired it unevenly, he will hve greter tx burden thn person B, with n even nd stedy income flow. In contrst, due to the effect of consumption smoothing, consumption pth is selected which is identicl to the life-pth income, resulting in n identicl consumption tx burden, nd their (finncil) sttus in reltion to ech other remins unltered. Equtions 4.17 nd 4.18 show, using the sme denottions s erlier, the life-pth income fter the levying of consumption txes. Yij ( 1 r) /[(2 r)(1 t )], ify ( 1 r)/[(2 r)(1 t )] C (4.17) ij [ Y C(2 r)( t t ) /(1 r)](1 r) /[(2 r)( 1 t ij f f ify ij ( 1 r) /[(2 r)(1 t )] C. (4.18) Consumer txes cn be considered neutrl when life-pth incomes reltion to one nother remin the sme s before txtion; tht is to sy, the following is true: (4.19) Three different cses must be exmined to gin confirmtion of this: (1) Cse 1: Both life-pth incomes only llow for consumption below the consumer tx bnd-border, in the following reltion: 1 2 Y (1 r) /[(2 r)(1 t )] Y (1 r) /[(2 r)(1 t )] C ij ij 22

24 This is evident, s 4.19 follows, from seeing tht the fctor, defined in 4.20 nd derived in 4.17, is positive, resulting from the nture of the low rte nd interestrte. 0 (2 r t ) /[(2 r)(1 t )] (4.20) (2) Cse 2: Both life-pth incomes llows for consumption over the consumer tx bnd border, in the following reltion: C Y (1 r) /[(2 r)(1 t 1 ij )] Y 2 ij (1 r) /[(2 r)(1 t In this cse, we need discern the following inequlity (4.21) tht is reduced to (4.22) fter the deduction of the constnt, nd results, s in the previous cse, from the nture of the high rtes nd interest-rte. )] 0 f f (4.21) (2 r t ) /[(2 r)(1 t )] (4.22) (3) Cse 3: One of the life-pth incomes is below the consumer tx bnd border, while the other llows for consumption over it, in the following reltion: Y (1 r) /[(2 r)(1 t 1 ij )] C Y 2 ij (1 r) /[(2 r)(1 t We must prove the disprity of (4.23) in order to fulfil the condition of neutrlity. )] (4.23) On the right-hnd side of the inequlity, cn be replced with the smller prt Y (1 r) /[(2 r)(1 )], nd fter redistribution we get the inequlity of (4.23), 1 ij t tht is, tht the reltion is required for the condition of neutrlity which ws our primry supposition nd strting point. 23

25 Thus, we hve seen tht progressive consumption tx does not violte neutrlity, in the sense tht it does not chnge the rnk order of n individul life-pth income. In the cse of income tx, we cn esily find counter-exmple which violtes the criterion of neutrlity. Let individul A denote the pre-tx income of 400 units in the first yer, nd 200 units in yer two. Menwhile, let B denote n individul with n income of 300 units during both periods. Clculting with 5% rel interest-rte, the life-pth income, discounted for the first period (rounded off to 590.5) is higher thn tht received by B (585.7). However, B is in significntly better position fter-tx, s units remin in this cse, wheres A is left with only The bove counter-exmple proves tht progressive work-relted income cn violte requirements of neutrlity in some (not ll) cses, s when it chnges the rnk order of individuls life-pth income. 24

26 5 CONCLUSIONS Tking into considertion the stringent limittions of the model (prtil pproch, no inherited or prior ssets, consumption smoothing by decisionmkers), we cn drw the following conclusions: On the bsis of effeciency studies - in periodicl model - there is no difference in efficiency between progressive consumption nd income tx, nd both distort decision-mking in the sme situtions nd to the sme extent; specificlly, in the choice between work nd leisure ctivities. In the model encompssing severl time periods, there is hrdly ny discernble difference in efficiency between the two tx rteble vlues nd, contrry to expecttions, they both result in the sme dedweight loss. Due to the vrying tx burdens on the tx pyer, only problems with equlity rise since, despite equivlency criteri, progressive consumption tx deducts less from individuls with identicl incomes thn does progressive income txtion nd, consequently, stte revenues will be less, s well. However, severl periodicl models meet with our previous expecttion by showing tht when income txtion is extended to cpitol income - progressive consumption txtion cuses less distortion thn income txtion. It would therefore pper tht consumption txtion is better, from the point of view of efficiency, insofr s it is contrsted to income tx; however, in the cse of lbour income tx, there is no difference in distortion. As fr s equity is concerned, it hs lredy been shown tht it cn perform better lso when fced with n income tx only levied ginst lbour income txes. Under the originl equivlent systems of txtion, both consumption nd income txes (depending on the structure of society) cn be both more nd less successful thn the other in levelling socil inequlity. The greter the difference between high nd low incomes, the lrger the intervl, nd the greter the extent to which progressive consumer txes reduce socil inequlity - s mesured by 25

27 the Gini coefficient - more thn progressive lbour-income tx. However, given the criterion of identicl stte revenues, progressive consumption txtion is more effective in reducing socil inequlity thn progressive lbour-income txtion, irrespective of different levels nd lyers of society. Therefore, the hypothesis relting to equity ws, in prt, proven to be correct. Consumption tx clerly met the requirements of neutrlity, while progressive income tx cn esily be t odds with it. The chpter reviewing some relevnt literture shows tht simplicity fvours the current income txtion lthough, grnted some ltertions, progressive consumption tx cn be similrly simplified (insofr s the current income tx cn be sid to be simple). Ultimtely, consumption tx better serves the function of the stte to ct s stbilising influence. On the one hnd, it cn prevent crisis similr to the current one; on the other hnd, it cn better serve recovery from such crisis, should one develop, ssuming tht consumption tx hd lredy been introduced beforehnd. 26

28 6 NEW RESEARCH RESULTS My reserch indictes tht the following new nd scientific conclusions cn be inferred in prtil model, insofr s prior inherited ssets, initil welth re excluded, nd consumption smoothing is ssumed: (1) Bsed on the literture, the most expedient demrction between direct nd indirect txtion is the seprtion of the personl nd the mteril. In this sense, progressive consumption tx cn be considered direct tx. Interpreting the concept of direct tx is not mong the primry ims of this thesis; nonetheless, it cn be viewed s n importnt reserch result. Hungrin nd interntionl literture on the subject of txtion does not define direct nd indirect tx in quite the sme wy. I encountered this problem of definition in reltion with the clssifiction of expenditure tx n issue on which I tried to define my position nd to present points of interest worth considering. The process of devolution is complex nd, in my opinion, does not lend itself to ccruement, while enumertion my only led to the exclusion of new txes. The mrked difference in the txtion of persons or mteril ssets is tht the former mkes it possible to tke individul chrcteristics into considertion; furthermore, it t lest llows for not only liner, but progressive txtion. Progressive txtion txes persons, progressively, nd cn therefore unequivoclly be clssified s direct tx. (2) Principles of equivlency introduced to progressive tx-rtes. According to the originl principle, two progressive system of txtion cn be deemed to be equivlent, if they possible n identicl consumption stemming from n identicl income; tht is, if the fter-tx budget restrints re identicl in given period. As n lterntive criterion, two txes must yield n identicl stte revenue in such wy tht tx-rtes ensure the sme chnge in reltive prices, ccording to the originl principle. 27

29 Literture on the subject typiclly only compres flt-tx rtes, however, the principles of equivlency cnnot be pplied to progressive rtes without modifictions. The introduction of new principles of equivlency ws necessry in order to verify the hypothesis of the thesis. Hving ccepted the equivlency criterion of Rosen Gyer (2010) (sme chnge in reltive prices), I hve formulted two different criteri of equivlency for cses of progressive txrtes. (3) In periodicl model, the effects of progressive consumption nd income tx on efficiency re the sme. Insofr s the decision-mker is in effect for only single period, budget restrictions re modified in the sme wy in the respective decision-mking res, during both the process of choosing between individul ssets nd choosing between work nd leisure. The correltion ws noted in reltion to the originl criterion of equivlency, lthough it clerly lso pplies to the lterntive principle of equivlency, s well, since the sme displcement of sme budget restrictions results in sme stte revenues. (4) Tking severl periods into considertion, the distortion effects of consumption tx nd progressive income tx re the sme with the exception of one prticulr cse. Although the movement of budget restrictions re not the sme, with the exception of the cse where the decision-mker is consumer on the bndboundry of consumption tx. There is no difference in efficiency in the cse in which prllelism is involved, the substitution effect is not pprent in either instnce. Nevertheless, effects of income do differ from one nother: progressive income tx operting long the originl principle of equivlency deducts more thn consumption tx from those who fll into different tx-rtes t different periods; however, this rises questions concerning equity. Chnging their prior decisions is worthwhile only for those on the boundry nd with specil preferences, but if we ccept consumption smoothing, then it is certinly 28

30 not worthwhile. In this cse there is therefore no sense in tlking bout the distorting effects on decision-mking of differences between consumption nd income tx. (5) Income tx levied on cpitl income distorts inter-temporl decisions to greter mgnitude thn consumption tx. The inter-temporl budget restriction breks in the cse of interest-tx: while the budget constrints of borrowers does not chnge, substitution effect, resulting in cn occur in the phse relted to those with svings. (6) In certin cses, progressive consumption tx is more in line with expecttions of equity thn progressive income tx. Where the originl criterion of equity is concerned, progressive consumption tx is only ble to reduce socil inequlity to greter degree thn income tx given certin sections of society. In contrst is the criterion of lterntive equivlency where, regrdless of ny poverty rtio, the Gini coefficient indicted greter mesure of equity for progressive consumption tx. (7) Progressive consumption tx stisfies the criterion of neutrlity while progressive income tx often violtes it. I exmined neutrlity by compring the reduced vlues of lifetime txtion on creer ernings. I demonstrted tht progressive consumption tx lwys remins neutrl, given n rbitrry income-flow, if we ccept consumption smoothing. On the other hnd, I cited only single exmple for income tx which, however, cnnot be considered exceptionlly rre. (8) I consider the most importnt scientific result of my thesis to be the presenttion of the effects of progressive txes. University text-books nd most rticles go no further thn nlysing liner tx. Constrints nd criteri relting to specific progressive txes re not prescribed. Furthermore, comprisons of income nd consumption tx found in relevnt literture tend to del with generl income tx; however, my thesis drws ttention to the wy in which there re distinct dvntges to txes levied only on lbour income tx. 29

31 Consumption tx nd the tx exemption of svings is not the sme s the tx exemption of cpitl income nd this is the cse to such degree tht the former is demonstrtively better thn the ltter; primrily on the grounds of equity, neutrlity nd stbility. 30

32 7 RECOMMENDATIONS (Theoreticl nd Prcticl Usge) The following conclusion cn be drwn from working ssumptions in the thesis prtil model section: progressive tx cn better stisfy expecttions of tx systems given the condition tht pproprite simplifictions re mde when plced into prctice - nd lso tht certin terms re disregrded, s is the cse with the current income tx. When pplied in prctice, this mens tht the current income-tx system needs to be trnsformed in such wy tht it pproximtes or pproches expenditure tx. This endevour hs two stges: firstly, rteble-vlue needs to be extended to s mny income ctegories s possible; secondly, svings for the given period need to be under exemption. The more we re inclined towrds progressive consumption tx, the more pprent become those effects presented in this thesis which re principlly in line with requirements of equity, neutrlity nd stbility. For those who would not ccept wider ppliction of expenditure tx, the Advisory Commission on Intergovernmentl Reltions (1974) only recommends the introduction of progressive consumption tx for nrrow section, despite every positive nd simplifying effect it my present. For instnce, if consumer expenditure would exceed txble income by some previously stimulted miniml mount, then consumption would be rteble-vlue bsed, nd tx would be levied ccording to the income tx rte. A socilly wide-rnging discourse-presenting the dvntges of the system- is certinly necessry prior to the introduction of the system, due to high dministrtion expenditures involved; moreover, it is highly recommended tht the system be initilly introduced nd tried on smll bse s n instructive tril period (for instnce 100,000 thousnd erned bove $100,000 in 1971). 31

33 8 PUBLICATIONS ON THE TOPIC OF THE THESIS Peer-reviewed ppers in Hungrin Vrg, Erzsébet Teréz (2014): Lehet-e közvetlen fogysztási dó? Mgyr Tudomány, 175. évfolym, 11. szám, Budpest. pp Vrg, Erzsébet (2014): Ekvivlens dók htás társdlmi egyenlőtlenségre. Köz-Gzdság, 9. évfolym, 3. szám, Budpest. pp Vrg, Erzsébet (2013): Adózási foglmk újrgondolás és rendszerezése. In: Bánfi Tmás Kürthy Gábor (ed.): Pénz, világpénz, dó, befektetések. Budpest. pp Vrg, Erzsébet (2012): A jövedelem és fogysztás lpú sávosn progresszív dózttás összevetése. In: Fülöp Péter (ed): Tvszi Szél Konferencikötet, Doktornduszok Országos Szövetsége, Győr, pp Peer-Reviewed Ppers in Foreign Lnguge : Vrg, Erzsébet Teréz (2013): Comprison of the equity effect of Lecture Notes: progressive income nd consumption txtion. Közgzdász fórum, issue 6., Kolozsvár. pp Kürthy, Gábor Vrg, Erzsébet (2012): Pénzügytn gykorltok. Presenttions: Tnszék Kft., Budpest. Vrg, Erzsébet (2011): A végén csttn bnkdó. Budpesti Corvinus Egyetem, Közgzdságtudományi Kri Konferenci. Vrg, Erzsébet (2011): A progresszív fogysztási és jövedelmi dó htékonysági összevetése. Budpesti Corvinus Egyetem, Közgzdságtni Doktori Iskol VII. éves Konferenci. 32

34 Vrg, Erzsébet (2010): Kidási dó és htékonyságánk feltételei. Budpesti Corvinus Egyetem, Közgzdságtni Doktori Iskol VI. éves Konferenci. Vrg, Erzsébet (2007): Az dólp örök dilemmáj. I. Kposvári Gzdságtudományi Konferenci. Vrg, Erzsébet (2007): A Public Finnce és Finnzwissenschft hgyományink összevetése - A közjvk és közösségi igények vontkozásábn. Musgrve emlékkonferenci, Budpesti Műszki Egyetem. Vrg Erzsébet (2007): Tőkejövedelmek dóterheinek optimális szintje.. OTDK, Miskolc. Foreign Lnguge Presenttions: Vrg, Erzsébet (2010): Expenditure tx - Assumptions of the efficiency. Europen Doctorl Seminr, Srjevo. Vrg, Erzsébet (2009): Comprison of progressive consumption nd income txtion especilly in the light of efficiency. Europen Doctorl Seminr, Budpesti Műszki Egyetem. Trnsltion of Foreign Lnguge Article: Musgrve, R. A. (1996): A Public Finnce és Finnzwissenschft hgyományink összevetése. Prt I in: Köz-Gzdság issue 3. pp , prt II in: Köz-Gzdság yer issue 4. pp Hungrin Summry of Foreign Lnguge Article: Fogel, R. (2006): Szökés z éhségtől és kori hláltól Angus Deton nyomán. In: Köz-Gzdság, issue 2. pp