POLITICAL LOBBYING AND ASYMMETRY OF PIGOVIAN TAXES AND SUBSIDIES. Israel Finkelshtain and Yoav Kislev

Transcription

1 September. 11, 2000 POLITICAL LOBBYING AND ASYMMETRY OF PIGOVIAN TAXES AND SUBSIDIES by Israel Finkelshtain and Yoav Kislev Department of Agricultural Economics and Management, Hebrew University, P.O. Box 12, Rehovot, Israel* Abstract The effects of Pigovian taxes and subsidies on pollution control are compared for an economy with politically powerful producer and consumers. The analysis reveals sharp asymmetries between the two instruments of control. Where the polluting producers hold the political hegemony, taxes lead to over-production and excess pollution; while with subsidies, production and pollution levels are too small. Alternatively, if organized consumers dominant the political arena, the effects are reversed with taxes, production is sub-optimal; while with subsides, too much is produced. With political pressure, both control instruments may be worse than a hands off, non-intervention policy. The paper further shows that the preferred policy instrument cannot be determined on theoretical grounds; detailed knowledge of technical and behavioral parameters is required to find the best policy tool. * Please send correspondence to Israel Finkelshtain at the above address, or finkelsh@agri.huji.ac.il

2 I. Introduction and Summary A familiar finding of marginal economic analysis is that a socially optimal level of pollution can be achieved by either a tax per unit of discharge or a subsidy per unit of reduced emission. The symmetry of the two control instruments was however criticized on several grounds, the most common being the marked difference in their long-run effects. The general conclusion that emerged was that a tax regime is more efficient than a subsidy, since it yields fewer active firms, less pollution, and lower production costs. It was also shown that, with subsidies and in the long-run, pollution may be greater than its free market, non-intervention level (for a survey and references, see Cropper and Oates [5]). The existing literature on environmental economics mostly views the government as enacting policies to enhance social welfare. Intervention, however, more often than not, induces lobbying and political pressure. Interest groups organize to modify policies: either to fend-off threats or to exploit opportunities. In this paper we offer an examination of regulations to reduce pollution in a political environment where both producers in the polluting industries and non-producing citizens (regarded here as consumers) may form lobbies to promote their interests. The tool of the analysis is a model of the kind used by Grossman and Helpman [8] in their study of trade policies, modified to serve the needs of the present comparison between Pigovian taxes and subsidies. The most closely related study to ours is Fredriksson [7] who also used the Grossman and Helpman model to analyze Pigovian taxes and subsidies in a political environment. Our analysis differs from Fredriksson in several aspects. First, in Fredriksson mode, the Pigovian tax is levied on the production of the polluting industry, while a the subsidy is granted for investment in less polluting technology. Therefore the two are considered as complementary components of a policy, designed to encourage pollution reduction.. In our analysis, the two policy regime are substitutes, and industries are either taxed or subsidized to encourage reduction of pollution. Interest groups offer political contributions, rewards to politicians for policy modifications. The emerging political equilibrium depends 2

3 on the control regime and the relative strength of the lobbies. Under taxes, if producers are better organized, tax rates will be lower than needed to achieve first best social optimum and, if consumers are politically stronger, taxes will be too high and pollution too low. It goes the other way around with subsidies; now, when producers are better organized, subsidies will be too high and pollution less than optimal. Again, the opposite will be true where consumers have the upper hand. The asymmetric effects of the political influence on taxes and subsidies stem form the opposing goals of producers and consumers: the former attempt to reduce taxes and increase subsidies, the latter press for higher taxes and lower subsidies. A second difference between Fredriksson analysis and the current one concerns the reliance of Fredriksson results on WW menu auction game and solution concept. In our analysis, the asymmetry result and other results concerning the efficiency of the two policy regimes relay on considerably less restrictive assumption that the contract between the politicians and the lobbies if efficient. This property is common to both cooperative and noncooperative theories of political pressure. Finally, while Fredriksson focuses on the influence of exogenous factors on the efficiency of the tax regime in the presence of political pressure, we show that a tax regime may be inferior both to a non-intervention, free market equilibrium and to a subsidy regime. 3

4 II. The Economy and the Environment There are n + 1 industries in the economy, producing n + 1 consumer goods, y 0, y 1,..., y n. The production of goods 1,..., n pollutes the environment. The pollution technology is simple; total pollution, E, is given by E = n e i (y i ), (1) i=1 where y i is the total domestic production of good i and e i is increasing. That is, pollution increases with production, and total contamination is the sum of pollutants emitted by all industries. The economy is populated by N individuals, who derive utility from the consumption of the n + 1 goods and disutility from pollution. Preferences are identical, but individuals may own varying amounts of factors of production. The utility function of a typical individual, j, is given by u j = x j 0 + n i=1 u i (x j i ) u(e), (2) where all the sub-utilities are concave, second-order differentiable, and increasing, and x j i is the consumption of good i by individual j. Total domestic consumption of good i is x i and it is identical in a closed economy to y i. In an open economy, domestic production need not be the same as consumption. Good 0, with a price p 0 = 1, serves as numeraire and defines the dimension of utility. The competitive market prices of the other goods are {p c 1,..., p c n}. Producer prices are {p p 1,..., pp n} and they may be lower or higher than market prices if taxes are imposed or subsidies granted. (Where appropriate, a superscript t or s replaces the superscript p to indicate the regime, taxes or subsidies.) The preferences in Eq. (2) yield n + 1 demand functions: x j i = d i(p c i), i = 1,..., n and x j 0 = Ij 4 n p c id i (p c i), (3) i=1

5 where I j is income of individual j and, by assumption, some quantity of the numeraire good x 0 is always demanded. With this structure, cross price elasticities of the demand functions for goods 1,..., n as well as their income elasticities are all zero. The consumer surplus of individual j is a function of the vector of prices of all goods p c : n n cs j (p c ) = u i [d i (p c i)] p c id i (p c i), (4) i=1 i=1 and, as utility functions are identical, the aggregate, economy-wide, consumer surplus is N CS(p c ) = cs j (p c ) = Ncs j (p c ). (5) j=1 The technology of the domestic production of good 0, y 0, is linear with labor the only input; that is, y 0 = l 0, (6) where l 0 denotes labor force in the 0 industry. The labor market is competitive and, accordingly, the wage rate is 1. All other goods are produced by combining labor and an industry specific fixed factor, a i, i = {1,..., n}, owned by a fraction of the population in the economy. In the absence of taxes or subsidies, profits of producer j in industry i are Π j i (aj i, pp i ) = pp i yj i c(aj i, yj i ), (7) where c(a j i, yj i ) is a cost function; it decreases with aj i and increases and is convex in yj i. Note that profits depend only on the specific factor endowment and the producer s price of the manufactured good. The profit function, Π j i (aj i, pp i ), increases in both its arguments. 5

6 III The Political Economy The political process solves two allocation problems: (a) resource allocation and the composition of production, determined by the choice of policy instruments and their levels; and (b) the division of the political spoils between the parties to the process. In our analysis, the division of the political surplus is set by the rewards contributed by the lobbies, the higher the contributions, the larger the share of the politicians and the smaller the share of the interest groups. In principle, political outcomes depend on the nature of the process of policy formation. Two groups of models have been applied to the study of policy formation in the presence of organized interest groups and political activity. In the first, the government is viewed as setting policy parameters aimed at maximizing a political support function, trading welfare of voters with diverse interests. This approach was taken by, among others, Peltzman [11] and Hillman [9]. The second group of models employs explicit game formulation; examples are Zusman [13] using a Nash [9] cooperative bargaining game, and the Grossman and Helpman [8] work we follow, where the political process is modelled as a non-cooperative auction game. The models differ, but they share a common property: the equilibrium reached is politically efficient, it is located on the polity s contract curve. Moreover, as we show later, with linear political rewards (defined below), the allocation of the controlled factor is independent of the magnitude of the political contributions and all the above models predict identical allocations (Hillman does not specify rewards explicitly). In other words, with linear political rewards, all these models suggest identical solutions to the first allocation problem, and the solution to the second problem the division of the spoils is independent of the solution to the first. We are making use of this independence property in the analysis to follow. 1 A. Policy Instruments 1 We have also used the same property in Finkelshtain and Kislev [6]. 6

7 The evolution of the political process is viewed as proceeding in two stages. In the first, the government, aiming at pollution control, chooses for each industry a regulation instrument. We denote by T the set of taxed industries and by S the set of industries, which are offered subsidies to encourage reduction in the production of the polluting good. The sum of industries in the two sets is smaller or equal to n. In the second stage, once an instrument was chosen, the producers in the industry and consumers may lobby to affect the adopted policy, but, by assumption, the choice itself is not subject to political debate and influence. 2 When a per-unit, industry specific tax, t i, is imposed on production, producers face the price p t i = pc i t i and a firm producing y j i, pays taxes to the amount t iy j i. The profit of firm j in industry i is then π j i (t i) = Π j i t iy j i (pt i) j {1,..., N}, i T. (8) A subsidy is granted for the reduction of production below some predetermined, firm specific level, ȳ j i. Firm j is paid s i(ȳ j i yj i ) and its profits are π j i (s i) = Π j i + s i[ȳ i y j i (ps i )] j {1,..., N}, i S. (8 ) Note that, given ȳ j i, the subsidy regime can be viewed as made of a fixed transfer of s iȳ j i and a per unit tax, making the producer s marginal price p s i = pc i s i. The per capita tax revenue is (p c i p t i) y i(p t i ) N i T Per capita subsidies amount to = i T t i y i N = Qt. (9) (p c i p s i ) [ȳ i y i (p s i )] N i S = i S s i [ȳ i y i (p s i )] N = Q s, (9 ) 2 Several authors analyzed choice of a control regime. For example, Buchanan and Tullock [3] concluded that politicians will, generally, prefer quantitative controls over price policies. 7

8 Net government revenue, expressed in per-capita values, is Q = Q t Q s, (10) and it is distributed to the entire population in equal lump-sum transfers; that is, the public at large is the beneficiary of the taxes collected and it also carries the burden of the subsidies paid to producers in industries under a subsidy regime. 3 Income of an individual in this economy is the sum of earnings from labor, profits, and a lump-sum transfer, Q: I j = l j + n i=1 Total economic surplus in the economy, V, is given by V = N [l j + j=1 π j i (aj i, pp i ) + Q. (11) n π j i + Q + csj u(e)]. (12) i=1 B. Interest Groups Political lobbies represent diverse interests. Producers may join forces in an effort to reduce taxes or increase subsides; consumers will strive to maximize consumer surplus plus transfer payments from the government minus pollution ( greens are included among the consumers). Lobbies are seen as contributing monetary rewards to politicians they try to influence: industry i s lobby is contributing the sum R i (dollars per year) as political rewards. These may take the form of aid in campaigns, demonstrations, letter writing, or even outright bribes. We analyze the effect of collaboration in the influence groups, but do not discuss the structure of the lobbies and modes of collaboration. Also, by assumption, the individual political contribution is not determined in the political equilibrium; it is left to the lobby to charge its members. 3 This is a simplified representation of taxes being used to provide public services and subsidies reducing the funds available for public goods. 8

9 We consider two types of political lobbies: (a) L producer lobbies, representing the interests of owners of specific factors of production in both taxed and subsidized industries; and (b) a consumer lobby, representing the interests of the consumers. We begin with the former. The welfare function of producer lobby i is: W i = ρ i π i (p p i ) R i i {1,..., L}, (13) where π i (p p i ) is total industry profits, ρ i is the share of the profits received by producers who are lobby members, R i is the lobby s contribution to the politicians. There may be fewer lobbies than industries, L n, not all producers are necessarily organized. 4 Like producers, consumers may also be organized in a lobby. Producers may participate in both their industry s lobby and in the political organization of the consumers. 5 The welfare of this last lobby, indexed h, is made of consumer surplus of its members plus government transfers and labor income minus pollution damage and political contributions; namely, W h = l h + νn[cs(p c )/N + Q u(e)] R h. (14) where ν stands for the share of the members of the consumers lobby in the population. C. The Government The politicians, accepting the rewards, are willing to modify regulation policies. It is assumed that R i i {1,..., L, h} are direct transfers from the interest groups and that the receiving politicians are maximizing the weighted sum W g in W g = αv + R, (15) 4 Formally, the model could accommodate non-polluting industries and polluting industries subsidized to increase production (agriculture in some countries), but such extensions stretch the analysis beyond the focus of our discussion. 5 People may often be found in opposing or competing lobbies. Farmers, as an example, may lobby for higher subsidies to agriculture and, at the same time, participate in political activity promoting public expenditures on schooling. 9

10 where V is defined in (12) and R = R i is the sum of the political contributions from all lobbies. The parameter α 0 represents the preference of the politicians for society s welfare relative to political bribes. The lower α, the lower the ethical norms of the politicians. We use the term linear reward system to characterize the way the rewards enter additively, in (13) (15), the welfare functions of the lobbies and the politicians. As indicated, this linearity will be shown to simplify the analysis considerably by separating the determination of factor allocation, on the one hand, and rewards, on the other. The politicians (the government), striving to maximize their welfare, W g, are seen as striking a set of contracts, deals with the lobbies, trading regulations for political rewards. For most of our analysis we do not need to specify the details of the deals. It is only assumed that they are politically efficient. D. The Polity s Contract Curve The political contract is Pareto efficient: no party to the contract can improve its position without reducing the welfare of any of the other parties. Efficiency is sufficient for qualitative and quantitative characterization of the political equilibrium and its properties. An efficient agreement between the government and the interest groups is defined by the solution to r, p = arg max r,p W g(r, p) S.T. W i (R i, p) W i i {1,..., L, h}, (16) where r = R 1,..., R L, R h is the vector of contributions of the industrial and consumer lobbies, and W i > 0 marks lobby s i, reserved welfare. 6 An examination of (16) will further clarify the nature of the political process. 6 Since both producer prices and prices faced by consumers are affected by the solution to (16), we economize and omit the superscripts p or c on the vector of prices, p. Also, taxes, t, and subsidies, s, are the differences between producer and consumer prices and are not written explicitly in (16). 10

11 Because of the simplifying assumptions underlying the preference function (2), a control instrument in industry i affects only that industry and the consumers (and, indirectly, industry 0). Consider a tax, it modifies producer profits, it reduces pollution and, in addition, all consumers share in the tax revenue as recipients of the lump-sum transfers. Consequently, the first order conditions of the maximization in (16) can be written as separate though interconnected n industry sets. For industry i, by maximizing first with respect to t i and then with respect to R i, one gets, W g t i = λ i W i t i + λ h W h t i (17) and λ i = λ h = 1 (18) where the λ parameters are Lagrange multipliers. The equality in (18) is due to the linearity of the rewards, that they are added to and subtracted from the welfare of the politicians and the lobby members, respectively. Examining (13) (15) we can see that the term R i or R h do not appear in any of the partial derivative in (17). As indicated, the control variable t i is determined independently from the political reward. Combining (17) and (18), one realizes that (16) would have had an identical solution if W g was maximized subject to a given sum of W i + W h. That is, the two separate political contracts between the politicians and each lobby affected by the control instrument in industry i are equivalent to a single contract between the politicians and the interested public; where the welfare of the interested public is the sum of the welfare of the members of the industrial lobby and the organized consumers. This equivalence is also a reflection of both groups attempting to influence the political process by money-like donations while, for the politicians, dollars have the same value whether contributed by producers or consumers. In addition, it follows from (17) and (18) that the set r, p, solving (16), also maximizes W w = αv + W i + W h ; (16 ) 11

12 that is, the political process can be viewed as maximizing a weighted sum, W w, with the weight 1 + α given to the welfare of members of lobbies producers and consumers and the weight α attached to the welfare of non-members. 7 IV. Political Equilibrium in a Closed Economy We begin with a description of optimal regulations with taxes or subsidies and a brief discussion of the free market, non-intervention equilibrium. Political pressure is then introduced and the asymmetric effects of the control measures is demonstrated. The implications of the political equilibrium are discussed. Political equilibrium in a small open economy is analyzed in the next section. A. Welfare Maximization First best, welfare maximizing output levels are the solution to N n [ max Π j y1 1,...,yn N 0 i j=1 i=1 u(e)]. (19) Accordingly, the first order conditions, with respect to each firm production level, are (primes indicate derivatives): p c i = c j i (y j i ) + Nu e i i {0,..., n} j {0,..., N}. (20) In the absence of political pressure, the government, aiming to achieve social optimum, sets taxes or subsidies as t i = Nu e i or s i = Nu e i and the private first order conditions become p t i = c j i (y j i ) i T (21) and p s i = c j i (y j i ) i S, (21 ) 7 This observation is due to Zusman[13] and Grossman and Helpman [8]. 12

13 where p t i = pc i t i and p s i = pc i s i for taxes and subsidies, respectively. 8 B. Asymmetry of Political Effects The political equilibrium is characterized by tangency conditions defined by the solutions to (16). They are developed now in two stages; in the first, we find the marginal effects of the policy instruments, taxes or subsidies, on the welfare of each of the lobbies. These are combined in the second stage into a pair of equilibrium conditions which forms the first proposition of the paper. We begin with a lobby of producers owners of the specific factors of production in industry i. Using (13) one gets, 9 W t i t i = ρ i y i c u c, (22) and W s i s i = ρ i [ȳ i + y ic u c ], (22 ) For the consumers, receivers of the net tax revenue and affected by pollution, it follows form (14) that W t h t i = ν[ t i Nu e i y ic ], (23) u c and W s h s i = ν[ s i Nu e i y i c ȳ i (u c ) ]. (23 ) u c Recall that ν is the share of organized consumers in the economy s population. Next we notice that for the politicians W t g t i = α( t i Nu e i ), (24) u c 8 The same rates of taxes and subsidies also maximize the social welfare function in (12). 9 The following equalities were utilized in deriving (22) (26 ): π/ p p = y (Hotelling s Lemma), p c = u, p c / y = u, p p = c where c is the supply function, and p p / y = c. Also, from t = p c p t, t/ y = u c, and y/ t = 1/(u c ). 13

14 and W s g s i = α( s i Nu e i u c ), (24 ) It follows from (17) and (18) that the tangency condition, the first order condition solving (16), is W t i t i + W t h t i = W t g t i, (25) and similarly for subsidies. This leads to Proposition 1 characterizing the political equilibrium. Proposition 1: Consider the regulation of polluting industries by Pigovian taxes or subsidies in a closed economy with political lobbying by groups of producers and consumers. Then, the political equilibrium levels of taxes and subsidies satisfy the following conditions: Nu e i t i p t i = ρ i ν (α + ν)ɛ i i T, (26) and Nu e i s i p c i = (ν ρ i)[σ i (ɛ i β i η i ) + ɛ i ] (α + ν)ɛ i η i i S. (26 ) where ɛ i is the industry aggregate supply elasticity, η i is the elasticity of the demand, σ i = (ȳ i y i )/y i, and β i = p p i /pc i. Note that the denominator in the right hand side (rhs) of (26) is positive and that, in the rhs of (26 ), the term in the square brackets in the numerator is positive and the denominator is negative. The signs of both equations depend on the difference ρ i ν. Proof: Write (25) explicitly in terms of (22), (23), and (24) or (22 ), (23 ), and (24 ) and rearrange terms to get (26) and (26 ). By (26) and (26 ), (i) If the share of profits of the politically organized producers in industry i exceeds the fraction in the population of the politically active consumers (ρ i > ν), then the tax (subsidy) level is smaller than (exceeds) the socially optimal level and the equilibrium production 14

15 and pollution exceed (are smaller than) optimum levels. Production and pollution fall short of (exceed) the free market, non-intervention levels. (ii) If the share in the profits of the organized producers falls short of the politically active fraction of the population of consumers ρ i < ν, then the relative magnitudes in (i) are reversed. (iii) If the political pressures are balanced ρ i = ν, then the first best tax (subsidy) level is achieved. (vi) In all cases, the larger the supply elasticity and the politicians ethical norms, the smaller the absolute value of the deviation of taxes and subsidies from the optimal level of the instruments. C. Interpretations and Implications We turn now to further interpretation of Proposition 1 and the implied properties. By the proposition, if producers in the polluting industry are better organized than the consumers in the economy, then under a tax control, the political equilibrium is a compromise, with production between free market and socially desired levels. A subsidy regime, again with producers comparatively more powerful, induces too little production and too little pollution. The intuitive interpretation is simple. Under taxes, the producers press to reduce the tax; while with subsidies, they fight to increase the subsidy, up to and above the social optimum. The situation is reversed where the consumers dominant the political arena. The consumers are the beneficiaries of the tax revenue and pay a share of the subsidy bill. As a result, they press to raise taxes and to reduce subsidies; if they have the upper hand, taxes will be above the optimum and subsidies too low. The asymmetry between the two regimes is independent of the political power structure, it arises whether the producers are better organized or the consumers are politically more active. Symmetry is maintained only when the control instruments are set at their 15

16 optimal levels, t i = Nu e i or s i = Nu e i, and this is achieved either when the political power is balanced (ρ i = ν) or supplies are infinitely elastic. Only under these conditions the familiar finding is preserved: a socially optimal pollution level is achieved by either a tax or a subsidy. The welfare implications are illustrated in Figure 1 for a political equilibrium dominated by the producers. Social optimum is given by the pair p w, y w. In the tax panel the non-intervention price is p n leading to industrial output level of y n. The welfare loss due to the externality is measured by the area of the triangle acg. With a tax t, the price and the quantity are p t, y t. The welfare loss in this political equilibrium (ν < ρ i ) is measured by the smaller triangle abd; indicating that intervention by means of a Pigovian tax, even if modified by politically powerful producers, is welfare enhancing. Under a subsidy, on the other hand, the welfare loss in a free market is bgc and in the political equilibrium depicted in the figure it is abd. The second triangle may be larger than the first: in a political equilibrium dominated by polluters, a subsidy may reduce welfare relative to non-intervention conditions. Where, alternatively, consumers dominate the political stage, taxes will be too high and subsidies too low for social optimum. One concludes that, in the presence of political pressure, government intervention may worsen resource allocation. Depending on the structure of power, worsening may occur both under taxes and under subsidies. Moreover, since the alternative instruments operate in opposite directions and neither is consistently welfare maximizing, the preferred control cannot be determined a-priori. Contrary to the conventional conclusion, taxes are not the safely preferred instrument the welfare loss tax triangle abd may be larger than the corresponding subsidy triangle. The role of supply elasticity in a tax regime is illustrated in Figure 2. As drawn, socially optimal production is y w, attained when the pollution tax, t i, is equal to the distance ab in the diagram. Political equilibria, either on the supply function S 0 or on S 1, are y 3, y 4 when the producers are stronger and ρ i > ν, and they are y 1 and y 2 when 16

17 consumers have the upper hand. On the comparatively more elastic supply, S 0, equilibrium production is closer to the optimum, y w. This can be verified by examining the sign of y ɛ = y t t ɛ = 1 p t i (ρ i ν) u c (α + ν)ɛ 2, (27) where t/ ɛ is calculated form (26). Also in the diagram, the tax rates (the distances between the demand and the supply curves) deviates less from ab for the more elastic supply S 0. To provide the explanation for the effect of the supply elasticity under taxes, we turn to examine the political implications of its magnitude. We commence with the produces and show that the larger the supply elasticity the less profitable the political lobbying. Consider an equilibrium under a tax regime and, for simplicity, assume a single firm industry. Suppose now that the government attempts to decrease production (and, consequently, pollution) by a given amount y i. The government objective can be accomplished by increasing the tax rate such that producer price will change by p t i = y i c. Profits will then change by π = y i c y i = p i y i ɛ i. (28) The less elastic the supply, the more profits decrease for a given change in producer price. Hence producers in industries with comparatively less elastic supply will struggle harder to modify pollution reducing policies. ( Struggling harder in our analysis is reflected in the slopes of the indifference curves in Figure 3 below and, as a result, in the political contract.) We turn to the government to show that the larger the supply elasticity the larger the loss the politicians take from yielding to the political pressure. Suppose now that the producers attempt to decrease the tax rate by t i. If successful, social welfare will decrease by the amount V = t(u c Nu e ) y t = 17

18 = t p y(u c Nu e )ɛ. Hence the government will yield easier to the pressure of industries with comparatively less elastic supply. In the extreme case, when supply is infinitely elastic, both producers and consumers are indifferent to the tax. Producers are not affected by it and consumers receive the tax they pay. Social optimum prevails. These findings may seem to contradict the established Ramsey Boiteux tradition (Atkinson and Stiglitz [1]) of optimal taxation by which the more elastic the supply the more socially harmful is an intervention in prices. The apparent contradiction is resolved by recognizing that when taxes are levied to raise revenue, optimal rates minimize their effect on resource allocation, while with pollution the purpose of the taxes is to modify inefficient use of resources. In some respects, Proposition 1 modifies, for a political economy, the Coase [4] and Weitzman [12] conclusion that property rights affect the distribution of the cost of pollution abatement but they do not modify the nature of the solution to an externality problem. Accordingly, if the producers own the right to pollute the environment, a Pigovian subsidy secures maximum social welfare; if, alternatively, the government owns these rights the polluters should be taxed. Resource allocations will be the same. With political pressure, unless balanced, resource allocations differ. D. Independence of Policies and Rewards The magnitude of the political rewards, R i, R h, does not appear in the equations characterizing the political equilibrium, (26) and (26 ). This confirms our earlier assertion that, with a linear reward system, the equilibrium may be computed recursively. Figure 3 clarifies this point. The figure is drawn for a tax regime and for the government and producer lobby i, and, for simplicity, with ν = 0, that is consumers are not organized in the political areana. 18

19 The horizontal axis in the diagram is the tax imposed, the vertical axis the reward. The curves W i 0, W i 1, are the iso-welfare (indifference) curves of the lobby, where a higher curve represents lower welfare because of a larger political reward. Similarly, the curves W g 0 and W g 1 are the government s iso-welfare curves with welfare increasing the higher the curve. The slopes of the iso-welfare curves are W i t i / W i R for the government. Because of the linearity of the rewards, W i R for the producer lobby and W g t i / W g R = 1 and W g R = 1 [these are the λ values in (17) and (18)]. Thus the iso-welfare curves, both the lobby s and the government s, are parallel along vertical lines. The social optimum tax level is marked in Figure 3 as t w and the political equilibrium is marked t t. With parallel iso-welfare curves, tangency is along a vertical contract curve. The section ab is the core of the contract between the parties: neither the government nor the lobby will accept a political agreement that lowers its own welfare compared to a no-reward social optimum. The reward R i will be determined in the political process. One such process is outlined in the last section of the paper; here we just note that with linear rewards with vertical contract curves its exact magnitude will not affect the location of t t. Any set of rewards (point c is one possibility) will lead to the same set of taxes or subsidies, provided it is in the core and satisfies the participation constraint of all contracting parties If either the cost of raising R i was changing with the size of the contribution or the effect of the rewards on the politicians was not constant, the magnitude of the λ parameters indicating tradeoff between cost and effect would have measured industry specific power to influence policies. The contract curve in such a situation would not have been a vertical line. See Zusman [13] for a generalization in this direction. 19

20 V. A Small Open Economy In an open economy, the supply of the goods x i i {0,..., n} is perfectly elastic at the international prices p c i i {0,..., n}. Consumers welfare depends, therefore, on the domestic production levels y i i {0,..., n} only through their effect on pollution. As demand parameters do not appear in condition (26), characterizing political equilibrium for a taxed industry, the condition applies to an open economy, as well as to a closed economy. The equilibrium condition for a subsidized industry is made simpler by moving from a closed to an open small economy. We start with the effect of subsidies on the welfare of the three parties to the political contract and continue to the second proposition of the paper, W i s i = ρ i (ȳ i y i ) (29) W h s i = ν[ Nu e i s i c (ȳ i y i )] (30) W i s i = α( Nu e i s i ) (31) c Proposition 2: Consider the regulation of polluting industries by Pigovian taxes or subsidies in a small open economy with political lobbying by groups of producers and consumers. Then, the political equilibrium levels of the tax and the subsidy are given by the following equations: and Nu e i t i p t i s i Nu e i p s i = ρ i ν (α + ν)ɛ i, (26 repaeted) = σ i(ρ i ν) (α + ν)ɛ i (32) Proof: Eq. (26) is repeated. For (32), write (25) for subsidies in terms of (29), (30) and (31) and rearrange terms. Remarks (i) (iv) of Proposition 1 apply also to Proposition 2. 20

21 VI. The Division of the Political Surplus Recall that equilibrium policies are on the political contract curve. So long as they are positive and acceptable to both the politicians and the lobbies, the magnitude of the political contributions does not affect the political contracts. The contributions then divide the political surplus between the politicians and the interest groups. This division depends on the particulars of the negotiation procedure; several such procedures have been suggested and formulated in the economic literature. We conclude with a sketch of the determination of the political rewards in the non-cooperative game as suggested by Bernheim and Whinston [2] and adapted to models of political economies by Grossman and Helpman [8]. The logic of the solution to the game is simple: the political contribution of each lobby is the minimum sum acceptable to the politicians for that lobby to be included in the equilibrium. In other words, consider lobby i, if it does not participate and R i = 0, a political equilibrium will be reached as a contract between the politicians and all other lobbies. The welfare of lobby i in the equilibrium from which it is excluded is lobby i s reservation welfare. If it joins the political contract, lobby i will pay as a political reward the difference between the equivalent of its welfare when included in the contract and its reservation welfare. A few implications emerge. 1. A lobby will not participate in the political contract if it cannot secure its reservation welfare and, likewise, the politicians will not accept a lobby the reward of which will not cover the social damage that its inclusion may entail [see (16)]. For example, highly polluting industries may find it difficult to raise the funds needed to affect policies, even if they are well organized. 2. Consider an economy in which consumers are not organized and there is a single producer lobby in industry i. Then, the government s reservation welfare coincides with the first best social policy. To accept the contract with lobby i the politicians have to be compensated for the reduction in welfare. Equilibrium will be on the government s isowelfare curve passing, in Figure 3, through (t w, 0). The equilibrium will then be represented 21

22 by point b on the contract curve. In this case, the welfare the politicians enjoy in the political equilibrium is identical to their welfare in the first best social optimum. Lobby i will have to offer a higher political reward (for example, point c) if other interest groups, particularly consumers, lobby actively and their political contributions must be balanced for lobby i to be included in the political contract. 3. Another interesting situation emerges when all lobbies, including the consumers, are of equal relative size; that is ρ i = ν, i L. Then the lobbies balance each other and the equilibrium reached coincides with the first best policy. The interest groups do not modify the implemented policies, but they pay the politicians considerable amounts. Each lobby must protect its position; if it is not included in the political set of contracts, policies will be set against its interest. 22

23 References 1. A. B. Atkinson and J. E. Stiglitz, Lectures on Public Economics, McGraw-Hill, New York (1980). 2. B. D. Bernheim and M. D. Whinston, Menu auction, resource allocation, and economic influence, Quarterly Journal of Economics 101, 1-31 (1986). 3. J. M. Buchanan and G. Tullock, Polluters profits and political response: direct controls versus taxes, American Economic Review 65, (1975). 4. R. H. Coase, The problem of social cost, Journal of Law and Economics 31, 1-44 (1960). 5. M. L. Cropper and W. E. Oates, Environmental economics: a survey Journal of Economic Literature 30, (1992). 6. I. Finkelshtain and Y. Kislev, Prices vs. quantities: the political perspective. Journal of Political Economy 105, (1997). 7. P. G. Fredriksson, The Political Economy of Pollution Taxes in a Small Open Economy. Journal of Environmental Economics and Management 33, (1997). 8. G. M. Grossman and E. Helpman, Protection for sale, The American Economic Review 84, (1994). 9. A. Hillman, The Political Economy of Protection, Harwood, Chur, Switz. (1989). 10. J. Nash, Jr., The bargaining problem, Econometrica 18, (1950). 11. S. Peltzman, Toward a more general theory of regulation, Journal of Law and Economics 19, (1976). 12. L. M. Weitzman, Prices vs. quantities, Review of Economic Studies 41, (1974). 13. P. Zusman, The incorporation and measurement of social power in economic models, International Economic Review 17, (1976). 23

24