Efficient Public Good Contributions by Impurely Altruistic Individuals By Emilson C. D. Silva* School of Business, University of Alberta, Edmonton, Alberta, Canada E-mail: [email protected] Xie Zhu Department of Economics, Oakland University, Rochester, MI 48309-4401 E-mail: [email protected] April, 2017 Abstract: We show that Becker’s Rotten Kid Theorem applies to situations involving impurely altruistic individuals who voluntarily contribute to a public good-individuals who we call “reluctant contributors”. Reluctant contributors care about prestige levels derived from their contributions. Reluctant contributors behave as “Good Samaritans” whenever they make contributions that add up to the socially optimal level of the public good and marginal rates of substitution between prestige and the numeraire good are equalized across contributors. There are two recipes for turnovers. In the first, the government redistributes income and prestige levels across contributors after contributions are made. In the second, the government redistributes income after contributions and exchanges of prestige units take place. Prestige units can be exchanged via brokerage services provided by a charity. We demonstrate that brokerage services in the market for prestige provide a rationale for the existence of charities. This paper also shows that a policy of subsidizing contributions is neutral in the presence of income redistribution. * Corresponding author. Efficient Public Good Contributions by Impurely Altruistic Individuals 1. Introduction Becker’s infamous Rotten Kid Theorem is as follows: “Each beneficiary, no matter how selfish, maximizes the family income of his benefactor and thereby internalizes all the effects of his actions on other beneficiaries” [Becker (1991), pp. 288]. Bergstrom (1989) formally demonstrates that the Rotten Kid Theorem holds as long as a benevolent parent considers his children - the rotten kids - as normal goods and the rotten kids’ utilities are transferable.1 More recently, Cornes and Silva (1999) extended Bergstrom’s analysis by showing that rotten kids will behave as Becker proposed whenever they care about the sum of their individual contributions to a pure public good (bad) and each rotten kid makes a strictly positive (negative) contribution.2 In this paper, we examine the circumstances under which voluntary contributions to a public good by individuals who care not only about the sum of their individual contributions, but also about their contributions per se - individuals who we shall call “reluctant contributors” - yield a socially optimal level of the public good. In other words, we investigate the circumstances under which reluctant contributors behave as “Good Samaritans”.3 1 See also Bernheim, Schleifer and Summers (1985) for circumstances where the Rotten Kid Theorem does not apply. Chiappori and Werning (1999) identify circumstances leading to interior and non-interior solutions. 3 One should not confuse our terminology - “Good Samaritans” - with the well known “Samaritan’s Dilemma” of Buchanan (1975). This dilemma says that an individual may become impoverished if he anticipates that his benevolent government is willing to help him out upon hard times. Bruce and Waldman (1990) show that in a two-period model parental altruism can result in the Samaritan’s Dilemma. This possibility is not present in our analysis because we consider a one-period model. In future work, we plan to extend our model to two periods and examine whether or not the “Rotten Kid Theorem Meets the Samaritan’s Dilemma”, as in Bruce and Waldman. 2 1 By postulating that contributors care about both the public good and the act of contributing per se, our approach builds on the impure public good model developed by Cornes and Sandler (1984) 4 and is closely related to the impure altruism model of Andreoni (1989, 1990)5. As Harbaugh (1998a) points out, contributors may derive positive utility from the act of contributing per se because publicly announced contributions endow contributors with “prestige”. 6 Indeed, Harbaugh (1998b) finds empirical support while Andreoni and Petrie (2004) providing experimental evidence for the hypothesis that donors derive utility from prestige.7 As in Harbaugh (1998a, 1998b), we shall adopt the convention that contributors obtain prestige from having their contributions publicly known. Society or private charitable organizations announce the levels of contributions after they are made. To simplify exposition, we shall refer to a contributor’s private utility from making a contribution as utility from prestige. This paper makes three main contributions to the literature, two of which are positive findings and one is negative. Starting with the good news, we show that reluctant contributors behave as “Good Samaritans” under two situations. In the first situation, reluctant contributors make strictly positive contributions to the public good in anticipation that their benevolent and egalitarian government will redistribute income and levels of prestige. The government redistributes levels of prestige by matching contributions and endowing contributors with prestige levels 4 The joint product impure public good model established by Cornes and Sandler (1984) has been developed to analyze a variety of impure public good provision problems ranging from climate protection to fighting terrorism. In the case of climate protection, reduction of greenhouse gas emissions generates not only global public benefits of mitigating global warming, but also local ancillary benefits such as improved health from reduction in conventional air pollutants - particulates, nitrogen dioxides and sulfur dioxides. See Caplan and Silva (2005), Loschel and Rübbelke (2005), Rübbelke (2005a), and Silva and Zhu (2005, 2011, 2015) for climate policy analyses in an impure public good framework. In fields other than climate protection, recent applications of the impure public good model include Kotchen (2005), who studies environmental friendly consumption; Mazzanti, Cainelli, and Mancinelli(2005), who study firms’ R&D decisions and social capital; Rübbelke (2005b), who studies terrorism. For more examples, see Sandler (1997) and Sander and Hartley (2001). For provision of impure public goods in monopolistic competitive markets, see Cavalletti, and Levaggi (2002). 5 Andreoni (1989, 1990)’s warm glow model made significant progress to the charitable giving literature by treating individual donations as contributing to an impure public good - individuals are motivated to give by an intrinsic feeling of warm glow from their individual gifts as well as the sum of their donations. See Andreoni (2004) for a review on the theoretical foundations, as well as the empirical and policy studies on charitable giving over the past 25 years. Some important questions investigated by the literature as reviewed in Andreoni (2004) are, e.g., optimal subsidy to giving, charitable contributions of time and money, fund-raising activities of charities as strategic players, measuring price and income elasticity of giving, etc. 6 Similarly, donors may want others to know about their donation due to the desire to demonstrate wealth and status (Glazer and Konrad (1996)), or to receive social approval (van de Van (2002)). Romano and Yildirim (2001) formulate a general utility function which could be used to model warm-glow or prestige and reputation motives for donation. 7 See, e.g., Rege and Telle (2004), Masclet et al. (2003), among others, for how public recognition of contributions and social approval can affect voluntary private contributions to public goods in experimental settings. 2 associated with their overall contributions to the public good. The matching grant changes the amount of prestige received by the grantee by an amount exactly equal to the grant amount. Therefore, such grants, which can be interpreted as grants in kind, have the key property of endowing grantees with levels of prestige that surpass (or fall short) of levels of prestige that they can obtain on their own through their original contributions to the public good. In the second situation, a market for prestige is created. Reluctant contributors make strictly positive contributions to the public good and trade prestige levels in anticipation that their benevolent and egalitarian government will redistribute income. The market for prestige is created by a non-profitable organization - a charity - whose responsibilities are to provide the public good and act as a “broker” in the market for prestige. Reluctant contributors utilize the charity to purchase and sell units of prestige, in excess of their direct contributions to the public good. Since we show that the resulting allocation of resources in this setting is socially optimal, we obtain a true Rotten Kid Theorem for the impure altruism model. Our bad news is the neutrality of a governmental policy of subsidizing contributions to the public good. Such a policy is shown to have no real effect on the allocation of resources whenever reluctant contributors anticipate that, after they contribute to the public good, the government will redistribute income. This result is novel for two main reasons. First, the policy neutrality results available in the literature are derived with the assumption that the center or the altruistic head of a household - i.e., a benevolent or altruistic agent - is able to precommit to lump-sum or distortionary income redistribution. This is a key assumption underlying the Ricardian Equivalence Theorem (see, e.g., Barro (1974) and Bernheim and Bagwell (1988)). It is also an important assumption - even if not explicitly stated - underlying the neutrality of income redistribution in economies with pure public goods (see, e.g., Warr (1982), Andreoni (1989), Bernheim (1986) and Boadway, Pestieau and Wildasin (1989)). In our model, the government plays a three-stage game with the reluctant contributors. This game is built after a three-stage game developed by Andreoni and Bergstrom (1996). The government announces subsidies and tax rates in the first stage, the reluctant contributors make their contributions to the public good in the second stage, and the government redistributes income amongst the reluctant contributors in the third stage. Therefore, the timing of the game played between the government and the reluctant contributors in our analysis is essentially different from the games examined by the policy neutrality literature. Indeed, our neutrality result comes from the fact that the government’s income-redistribution policy takes effect after subsidies, tax rates and 3 contributions to the public good have been made. The second reason for the novelty of our neutrality result is that it is derived in a setting where the public good is impure. Warr’s well known neutrality result, for example, depends crucially on the assumption that the privately provided public good is pure. As Bergstrom (1989) alerted us, the income-redistribution policy of a benevolent and egalitarian government, which is unable to precommit to incentive schemes, will not generally yield the Rotten Kid Theorem. But, as we shall demonstrate, this policy - being generally non-neutral - is an essential ingredient of a successful recipe for turnovers. Income redistribution when mixed with either prestige redistribution or prestige exchange is capable of transforming reluctant contributors into Good Samaritans! We organize the paper as follows. Section 2 presents the socially optimal allocation. Section 3 shows us the conditions that characterize the decentralized allocation, where the government does not intervene. In section 4, the government is endowed with a policy instrument to redistribute income. Section 5 examines the effectiveness of a governmental policy that subsidizes contributions to the public good in the presence of income redistribution. In section 6, the government is endowed with policy instruments to redistribute income and prestige levels. Section 7 introduces a market for prestige and analyzes the allocation of resources where contributors trade prestige levels and provide contributions to a charity in anticipation of the income-redistribution policy of the government. Section 8 concludes the paper. 2. Social optimum: the government’s most preferred allocation Consider an economy with n individuals, indexed by j, j = 1,..., n and a benevolent and egalitarian government. There are two goods. A private good can be used for either consumption or as an input in the production of a public good. For simplicity, we assume that it takes one unit of the private good to produce one unit of the public good. The public good level, G, equals ∑jgj. Contributor j is assumed to derive the following utility from contributing gj units of the private good to the public good and consuming xj units of the private good and G units of the public good: u j x j , pj , G u x , p g , G , j j j j = 1 , ... , n , where, for each j, u j is increasing in all arguments and strictly quasiconcave. The function p transforms contributor j’s contribution to the public good into units of prestige, pj. For simplicity, we assume that 4 pj p g j = g j , j = 1 , ... , n . Substituting this into contributor j’s utility function, enables us to write his utility function as follows: u j x , g , G , j j j = 1 , ... , n . The government’s objective function is a strictly quasiconcave transformation W of individual utilities: W u 1 n x1 , g1 , G , ... , u xn , gn , G , (1) where Wj=∂W/∂u j > 0 for all j. In words, the government cares about the welfare of each individual in society. Let us suppose that the government is persuasive enough so that it is able to convince all contributors how they should allocate their resources. The government determines its most preferred allocation - the social optimum - and instructs all contributors to make their choices so that the socially optimal conditions are satisfied. The contributors implement the socially optimal allocation by religiously following the instructions dictated by the government. All contributors will obtain prestige from their contributions because they will contribute to the public good on their own. The social optimum can be obtained by choosing {xj, gj}j = 1,...,n to maximize (1) subject to the resource constraint: j x j + j g j = j Ij , or j x j + G = I , (2) where Ij > 0 denotes individual j’s income level and I =∑j Ij represents the aggregate income level. Let λ be the Lagrangian multiplier associated with (2). Assuming the solution is interior, the first order conditions yield (2) and the following equations: j j = 1 , .... , n , W j ux = , j k n j = 1 , ... , n . Wj up + k = 1 Wk uG = , (3) (4) Equations (3) and (4) yield j k up uG + nk = 1 =1, j ux ux supk j = 1 , ... , n . (5) Equation (3) tells us that the marginal social utilities of income should be equalized. Income redistribution is socially desirable because social preferences are strictly convex. Equations (5) show that the typical Samuelson condition for optimal provision of a pure public good should be modified to account for the private benefits associated with 5 contributions to the public good. For each j, the first term of the left-hand side of (5) is individual j’s marginal rate of substitution between prestige and the private good. Hence, the marginal rates of substitution between prestige and the private good should be equalized across contributors: j i up up = i , j ux ux 3. i , j = 1 , ... , n . (6) Reluctant contributors on their own Let us now examine a situation where the contributors voluntarily decide how much to contribute to the public good without governmental intervention. Our objective in this section is to illustrate the shortcomings generated with decentralized behavior. We wish to clearly identify which incentives governmental policy should promote in order to induce reluctant contributors to voluntarily contribute to the public good at the socially optimal level. All contributors play a simultaneous Nash game, whereby contributor j chooses { xj, gj } to maximize u subject to: j x,g,g+G j x j + g j = Ij , j -j j xj 0 , gj 0 , (7) taking the choices of all other contributors as given. The contributions of all contributors other than j sum up to yield G-j. Assuming interior solutions,8 the Nash equilibrium is characterized by (7), for all j, and the following equations: j j u p uG + j =1, j ux ux j = 1 , ... , n . (8) Equations (8) show us that each contributor neglects the positive effect of his contribution on each other. The decentralized allocation deviates from the socially optimal allocation in three ways: (1) it violates the modified Samuelson condition (5); (2) it does not yield equalization of marginal social utilities of income; and (3) it does not imply equalization of marginal rates of substitution between prestige and the private good. Governmental policy should, therefore, induce reluctant contributors to voluntarily choose levels of contributions that satisfy (5) while equalizing marginal social utilities of income and marginal rates of substitution between prestige and the private good. 8 As in Andreoni (1990), we shall only consider equilibria with non-zero contributions. This restriction is not without loss of generality, since the efficiency properties of the equilibria studied are sensitive to whether or not the equilibria are characterized by corner solutions. By restricting our attention to equilibria with positive contributions, we wish to illustrate the set of circumstances under which voluntary contributions and governmental policy implement the socially optimal allocation described by (2) - (5). 6 4. Reluctant contributors anticipate income redistribution As discussed in the introduction, Cornes and Silva (1999) show that voluntary contributions to a pure public good yield an efficient aggregate level of provision provided that each contributor makes a positive contribution to the pure public good and a benevolent agent, who cares about all contributors, redistributes the incomes of contributors after contributions are made. Such a finding motivates us to study the implications of income redistribution in the current setting, with an impure public good. As we shall see, the fact that all contributors are impurely altruistic implies that income redistribution will generally fall short of providing contributors with the right incentives to voluntarily provide a socially optimal level of the public good. However, as we shall also demonstrate, income redistribution will be an essential policy tool in the arsenal of policy instruments that the government should use to induce reluctant contributors to behave as Good Samaritans. Assume that the government’s sole economic role is to redistribute income across all individuals. Income redistribution takes place after the reluctant contributors choose their contributions to the public good. When contributors make their choices, they know that the government will subsequently redistribute income in order to maximize its egalitarian objective function. That is, all contributors correctly anticipate that income transfers will be effected to equalize marginal social utilities of income, as in (3). In game theoretical terms, the contributors and the government play a two-stage game, whereby the contributors choose their contributions to the public good in the first stage and the government chooses income transfers in the second stage, after having observed the levels of contributions. The equilibrium concept utilized is subgame perfect equilibrium. Let τj denote the income transfer received (if positive) or paid (if negative) by contributor j. This individual’s budget constraint is x j + g j = I j + j , j = 1 , ... , n . (9) Since ∑jτj = 0, we obtain the economy’s resource constraint (2) by summing up equations (9) over all j. Let g-j denote the (n-1) - tuple of contributions which does not include gj. We are now able to formally describe the two-stage game C which is called Game 1 C as follows: Game 1 7 Stage 1: Taking g-j as given, contributor j chooses gj > 0 to maximize u j x g j j , g- j , g j + G- j , I , g j , g j + G- j , j = 1 , ... , n . (10) Stage 2: The government chooses {xj }j = 1,...,n to maximize (1) subject to (2) after observing {gj, g-j, G}. It is easy to explain the crucial features of this game. The government, in the second stage, knows both the prestige levels reached by contributors, {pj }j = 1,...,n, and the aggregate level of the public good provided, G. Since pj = gj for all j, the government, therefore, knows {gj, g-j, G}. As we mentioned above, we obtain the economy’s resource constraint (2) by summing up equations (9) over all j. The government chooses nonnegative {xj }j = 1,...,n to maximize (1) subject to (2). Let λ be the Lagrangian multiplier associated with constraint (2). The first order conditions for this problem yield the constraint (2) and equations (3); that is, a system of n + 1 equations, with n + 1 endogenous variables {λ , xj }j = 1,...,n and n + 2 exogenous variables {G, I, gj, g-j }j = 1,...,n. Since the objective function is strictly quasiconcave and constraint (2) is linear, the solution to the government’s problem - assumed to be interior - is a unique constrained global maximum. Hence, we may apply the Implicit Function Theorem to obtain the n + 1 implicit functions λ (gj, g-j, G, I), xj(gj, g-j, G, I ), j = 1,...,n. The functions xj(gj, g-j, G , I) tell us how the government’s transfers respond to changes in the levels of prestige, aggregate public good and aggregate income. Substituting contributor j’s transfer function into his utility function yields his objective function (10). The problem this contributor faces in the first stage is now clear and well defined. He chooses his most desired contribution level anticipating that there will be transfers of the numéraire good amongst contributors in order to satisfy the government’s income redistribution objective. The following result is immediate. Proposition 1. The subgame perfect equilibrium for Game 1 is given by ∑jτj = 0, (2), (3) and the following conditions: j i k xk up up u + nk = 1 Gk = 1 - in_ j i + nk = 1 , j psubk ux ux ux j = 1 , ... , n . (11) Proof. Consider the second stage. Maximization of (1) subject to (2) yields the constraint and equations (3) as the set of first order conditions. As we explained above, this set of first order conditions enable us to derive the implicit functions xj(gj, g-j, G , I). Substituting these functions into constraint (2), we obtain: j x j g j , g- j , G , I Differentiating (12) with respect to G yields: 8 + G I. (12) j x j g j , g- j , G , I G = -1. (13) Consider the first stage. Assuming interior solutions, the first order conditions can be written as follows: j j xj xj u p uG + = - + , j j ux ux pj G j = 1 , ... , n . (14) Summing up equations (14) over all j, acknowledging equation (13) and rearranging, we obtain equations (11). The subgame perfect equilibrium for Game 1 deviates from the socially optimal allocation in that the modified Samuelson condition (5) is not satisfied and marginal rates of substitution between prestige and the numéraire good are not equalized. As we shall demonstrate below, equalization of marginal rates of substitution between prestige and the numeraire good is a necessary condition for the efficiency of an equilibrium of any game where reluctant contributors voluntarily decide how much to contribute to an impure public good. If, as in the setting considered by Cornes and Silva (1999), contributors did not care about prestige at all, the income redistribution policy of the government would suffice for contributors to contribute the socially optimal amount of the public good. The question we investigate in the next section is whether the government may induce reluctant contributors to behave as Good Samaritans with a policy of subsidizing contributions, in addition to its income redistribution policy. 5. Policy neutrality: subsidizing contributions in the presence of income redistribution Andreoni (1990) finds that contributions of impurely altruistic individuals - i.e., reluctant contributors - increases with the rate at which contributions are subsidized by the government. This result is derived from straightforward comparative statics applied to the equilibrium conditions for a simultaneous Nash noncooperative game played by contributors. Because the government is not a strategic player in his analysis, taxes and subsidies are exogenously given and income redistribution is not considered. We extend his analysis in two significant ways. First, the government is given authority to endogenously determine: (i) the subsidy rate applied to contributions, (ii) the taxes levied to finance the subsidy scheme and (iii) the income transfers needed to promote income redistribution. Second, we assume that the government is able to pre-commit to the subsidy scheme and taxes. That is, the government announces both subsidies and taxes before contributors decide how much to contribute to the public good. To be precise, we consider a three-stage game as follows. In the first stage, the government announces subsidies and taxes. All contributors observe these quantities and, in the second stage, choose 9 their contributions. In the third stage, the government chooses income transfers. Let s [0, 1] denote the standard subsidy rate and tj denote contributor j’s tax rate, j = 1,...,n. This contributor’s budget constraint is as follows: x j + (1 - s) g j = I j + j - t j s G , j = 1 , ... , n . (15) Contributor j is taxed for a share tj of the cost of the subsidy scheme. Hence, tj> 0 and ∑j tj = 1. Since ∑jτj = 0, we obtain (2) by summing up equations (15) over all j. We are now ready to present the formal description of the three-stage game. Game 2 Stage 1: The government chooses {s, t }, where t = (t1,...,tn ), to maximize W u1 x1 g1 s, t , g-1 s, t , G s, t , I , g1 s, t , G s, t ,.. , un xn gn s, t , g-n s, t , G s, t , I , gn s, t , G s, t (16) subject to: 0 s 1 , j t j = 1 , t j 0 , j = 1 , ... , n . Stage 2: Having observed {s, t } and taking g-j as given, contributor j chooses nonnegative gj to maximize (10). Stage 3: The government chooses {xj }j = 1,...,n to maximize (1) subject to (2), knowing {s, t, gj, g-j, G}. Stages 2 and 3 of this game are identical to stages 1 and 2 of Game 1, respectively. The additional set of exogenous variables in this game - i.e., the n+1 exogenous variables, (s, t) - are only relevant in the first stage of the game. This follows from the fact that the government is able to redistribute the numéraire good across individuals in the third stage of the game, irrespective of subsidies and tax rates. Applying the same logic of the previous section, the first order conditions for the third stage will enable us to define the functions xj (gj, g-j, G, I), j = 1,...,n. Inserting these functions into the contributors’ utility functions yields their objective functions (10) in the second stage of the game. Assuming interior solutions to the maximization problems of the second stage, as in the previous section, we obtain equations (14). Provided the sufficient second order conditions are satisfied and the Nash equilibrium of the second stage is (locally) unique, we will again be able to apply the Implicit Function Theorem to define the functions gj (s, t). Hence, the function G(s, t)=∑j gj (s, t) is well defined. Substituting these functions into the contributors’ utility functions and the implied utility functions into the social welfare function (1), we obtain (16). The problem facing the government in the first stage is now also well defined. 10 Game 2 is purposely built very similar to a three-stage game considered by Andreoni and Bergstrom (1996). Our construction of the subsidy scheme and tax rates mirrors theirs. Our precommitment and timing assumptions for the first and second stages of the game are also identical to their own. Unlike us, they assume that contributors are purely altruistic and do not consider income redistribution. The third stage of their game involves no strategic choice, but simply the collection of taxes and subsequent payment of subsidies. As they acknowledge, the third stage of their game can be collapsed into the second stage without altering the Nash equilibrium for the second stage. Their analysis focuses on the Nash equilibrium for the second stage. Before we characterize the subgame perfect equilibria for Game 2 - as we will see there is a continuum of equilibria - as well as some of their key features, it is important to note that, because the government makes interpersonal income transfers after all the other strategic variables have already been determined, there is no loss in generality in letting t1 = 1, ti = 0 for I = 2,...,n, and considering the equilibria for Game 2 given these tax rates. All contributors know that the government’s income redistribution policy will equalize marginal social utilities of income in spite of the values set for the tax rates. We are now ready to state our negative result: Proposition 2. (Policy Neutrality) A policy of subsidizing contributions is completely neutral in the presence of income redistribution. Proof. Consider the third stage of the game. As we mentioned above, the conditions that characterize the equilibrium in this stage are exactly the same as the conditions that characterized the equilibrium for stage 2 of Game 1. It is important to note, however, that the equilibrium in stage 3 also satisfies the budget constraints (15). Since we assume that t1 = 1, ti = 0 for i = 2,...,n, we can write: 1 s , g1 , g- 1 , G , I , I1 i s , gi , g- i , G , I , Ii x1 g1 , g- 1 , G , I xi gi , g- i , G , I Note that: 11 - - I1 + s G + (1 - s) g1 , (17) Ii + (1 - s) gsubi , i = 2 , ... , n . 1 1 x1 x1 + = + +1, p1 G p1 G (18) i i xi xi + = + +1- s , pi partialG pi G i = 2 , ... , n . Differentiating equations (17) with respect to s yields: 1 = G - g1 G- 1 , s (19) i = - gsubi , i = 2 , ... , n . s Consider the second stage. Assume that the Nash equilibrium for the second stage is (locally) unique. Equations (14) can be utilized to implicitly define the functions gj(s). Hence, the function G(s) =∑j gj (s, t) is also well defined. Inserting these functions into equations (17) yields: 1 s , g1 (s) , g- 1 (s) , G(s) , I , I1 i s , gi (s) , gsub- i(s) , G(s) , I , Ii x1 g1 (s) , gsub- 1(s) , G(s) , I xi gi (s) , g- i (s) , G(s) , I - I1 + s G(s) + (1 - s) g1 (s) , - Isubi + (1 - s) gi (s) , (20) i = 2 , ... , n . Differentiating equations (20) with respect to s and utilizing (18) and (19), we obtain: x1 x1 1 1 n + + s i = 2 G pi G pi d gi 1 - G- 1 = 0 , = s ds (21) xi xsubi i i + k _ i G G p p k k d gsubk i + gi = 0 , = ds s i = 2 , ... , n . It should be clear that the solution to the system of equations (21) is given by: d gj ds = 0, j = 1 , ... , n . (22) Hence, d gj dG = nj = 1 = 0. ds ds (23) In the first stage of the game, we have the following implied social welfare function: W u1 x1 g1 ( s ), g-1 ( s ), G ( s ) , I , g1 ( s ), G ( s ) ,.. , un xn gn ( s ), g-n ( s ), G ( s ) , I , gn ( s ), G ( s ) The Lagrangian for the maximization problem in the first stage is as follows: 12 (24) L = W u1 x1 g1 ( s ), g-1 ( s ), G ( s ) , I , g1 ( s ), G ( s ) ,.. , un xn gn ( s ), g-n ( s ), G ( s ) , I , gn ( s ), G ( s ) + s + (1 - s) The first order conditions are: (25) -=0, s = 0 , 0 , s 0 , (26) (27) (1 - s) = 0 , 0, 1- s 0 . In writing (25), we made use of equations (22) and (23). Conditions (25) - (27) imply that any s [0, 1] solves the maximization problem. Proposition 2 tells us that subsidies have no real effect on the allocation of resources. It follows that the level of public good in any equilibrium with a positive subsidy rate will correspond to the level of public good in the equilibrium where this rate is zero. Since the latter equilibrium allocation is equivalent to the equilibrium allocation described by Proposition 1, the ability of the government of pre-committing to the subsidy scheme is of no real value. The equilibria for Game 2 are as inefficient as the equilibrium for Game 1. This begs the question: Is there some governmental policy other than subsidizing contributions which, when implemented together with the income redistribution policy, induces reluctant contributors to voluntarily provide the socially optimal level of the public good? As we shall demonstrate in the following section, the answer is an affirmative one. We show that a governmental policy of actively providing contributors with prestige C a commodity whose market is missing C promotes the right incentives for reluctant contributors to behave as Good Samaritans. 6. A winner recipe’s key ingredients: income redistribution and prestige grants Consider now a setting where the government takes two actions after contributions are made. First, as before, it redistributes income amongst the reluctant contributors. Second, it allocates additional prestige units to contributors with matching grants. The sole purpose of the “prestige policy” is to endow contributors with more (or less) prestige than the prestige levels they can achieve on their own through their contributions to the public good. Let mj denote the matching grant amount received by contributor j, j = 1, ..., n. Contributor j’s utility function is now written as follows: u j x , p , G j j = uj xj , p g , m , G , j j j = 1 , ... , n , where, for simplicity, we assume that each contributor j derives prestige from his actual contribution to the public good, defined as the sum of his own contribution and the matching grant received from the government: 13 pj = p g ,m = g j j j + mj , j = 1 , ... , n . Substituting these expressions into the contributors= utility functions, enables us to write the utility functions as follows: u j x , p , G j j = u j x j , g j + mj , G , j = 1 , ... , n . Contributor j=s budget constraint is x j + g j = I j + j + m j , j = 1 , ... , n . (28) Since the objective of the prestige policy is not to supplement private provision of the public good but to alter prestige levels that can be achieved by contributors on their own, the prestige policy must satisfy the following constraint: n j = 1 mj = 0 . (29) The matching grant policy of the government considered here implies that some contributors will receive negative grants. This fully captures the notion that prestige redistribution is a zero-sum game - i.e., for some contributors to receive additional prestige from the government, some others have to give up an equal amount of their initial endowment of prestige. It may help the reader to think of the prestige redistribution policy of the government as policy of granting medals to heroes - i.e., of granting medals to Good Samaritans! The analogy is a good one, since everyone should agree that a policy of granting medals to heroes is a zero-sum game.9 Given equation (29) and ∑jτj = 0, we obtain the economy resource constraint (2) by adding up the budget constraints (28) over all j. Now, note that when the government chooses mj, for all j, each individual’s contribution to the pure public good is already predetermined. Since the government takes gj, for all j, as given and chooses mj, for all j, it, in fact, chooses pj, for all j. Hence, the problem faced by the government in the second stage of the game it plays with the reluctant contributors C formally described below C is to choose {xj , pj }j = 1,...,n to maximize: W subject to: u 1 n x1 , p1 , G , ... , u j x j + G = I , j pj = G , 9 xn , pn , G . (30) (2) (31) Our results do not depend on the restriction imposed by equation (29). The results go through even if the prestige redistribution policy of the government granted additional prestige to everyone. This would require positive public provision of the pure public good. It can be shown that the results remain the same as long as * n j = 1 mj G , where G* denotes the socially optimal amount of the pure public good. 14 where, given our definition of pj, we obtain constraint (31) by summing up all pj’s and noting equation (29). It is important to note that the exogenous variables in the government’s maximization problem are G and I. Since W is strictly quasiconcave and the constraint (2) is linear, the solution to the government’s problem - which will be assumed to be interior - will be a unique and global constrained maximum. By applying the Implicit Function Theorem, we will be able to implicitly define the following important functions: xj = xj ( G , I ) , pj = pj ( G , I ) , (32a) j = 1 , ... , n , j = 1 , ... , n . (32b) Functions (32) tell us how the government responds to both the aggregate private provision of the pure public good and the aggregate income level. As we shall demonstrate below, the key difference between this situation, where the government is able to redistribute prestige levels, relative to the situation where the government is not endowed with instruments to redistribute prestige is that now information about the aggregate quantity of the pure public good suffices for the implementation of a socially optimal allocation of resources. For concreteness, we formally describe the two-stage game played by contributors and the government - called Game 3. The game is as follows: Game 3 Stage 1: Taking g-j as given, contributor j chooses gj > 0 to maximize u j x j ( g j + G- j , I ) , p j ( g j + G- j , I ) , g j + G- j , j = 1 , ... , n . (33) Stage 2: The government chooses {xj , pj }j = 1,...,n to maximize (30) subject to (2) and (31), after observing {G}. We are now ready to present the first of our two main results: Proposition 3. The subgame perfect equilibrium for Game 3 is socially optimal. Proof. Consider the maximization problem in the second stage. Let λ and ψ be the Lagrangian multipliers associated with (2) and (31), respectively. Assuming the solution is interior, the first-order conditions yield (2), (31) and the following equations: j W j ux = , j W j ux = , j = 1 , .... , n , j = 1 , .... , n . 15 (3) (34) Note that equations (3) and (34) imply equalization of marginal rates of substitution between prestige and the numéraire good; that is, they imply equations (6). Equations (2), (3), (31) and (34) implicitly define functions (32) as well as functions λ = λ (G, I ) and ψ = ψ (G, I ). Substituting functions (32) into equations (2) and (31) yields: j x j ( G , I ) + G = I , j pj ( G , I ) = G . (35a) (35b) Differentiating equations (35) with respect to G, we obtain: xj ( G , I ) = -1, G pj ( G , I ) =1. j G j (36a) (36b) Now consider the first stage of the game. Substituting equations (32) into contributor j’s utility function, the problem faced by contributor j is to choose gj > 0 to maximize (33), taking g-j as given. Assuming interior solutions, the first order conditions for the contributors’ problems are: or pj ( G , I ) xj ( G , I ) j + uGj = 0 , + upj ux G G j j xj ( G , I ) up p j ( G , I ) uG + j + = 0, j G G ux ux j = 1 , ... , n , j = 1 , ... , n . (37) Summing up (37) over all j and noting (36a) enables us to write: j j up j ux pj ( G , I ) j + j uGsupj = 1 . G ux (38) Since the first order conditions of the second stage maximization problem imply that j i up up = i , j ux ux (6) i , j = 1 , ... , n , equation (38) can be rewritten as follows: i pj ( G , I ) u subGj up + =1, j j i j G ux ux i = 1 , ... , n . Utilizing equation (36b), we are able to rewrite (39) as follows: i j up u + j Gj = 1 , i ux ux 16 i = 1 , ... , n , (39) which is the modified Samuelson condition (5). Proposition 3 is a “sophisticated” Rotten Kid Theorem. Besides redistributing income, a benevolent agent - the government - who cares about all his children - the reluctant contributors - promotes the right prestige incentives so that the reluctant contributors, acting selfishly, internalize all externalities. Income redistribution is necessary but not sufficient for our Rotten Kid Theorem because, in our setup, there is no market for prestige, where reluctant contributors can trade prestige for the private good. The governmental prestige policy fulfills this market vacuum by equalizing marginal rates of substitution between prestige and the numéraire good. 7. Another winner recipe’s key ingredients: income redistribution and a market for prestige 17 As we observed in section 6, there are efficiency gains to be had from properly altering the levels of prestige that contributors can obtain on their own through their contributions to the public good. We also noticed that, due to the absence of a market for allocating prestige between contributors, governmental action is needed to induce rotten contributors to behave as Good Samaritans. In this section, we investigate the efficiency properties associated with the creation of a market for prestige. To simplify the exposition, we will assume that n = 2 and that the social welfare function is utilitarian. The government’s sole policy instrument is the interpersonal income transfer used to redistribute income between the two reluctant contributors. Suppose that an entrepreneur who cares about the public good - say, contributor 1 - forms a non-profitable organization - say, a charity - with the basic purpose of providing the public good. Contributor 1, however, knows that both he and contributor 2 care not only about the public good, but also about the level of prestige that originates with the act of contributing to the public good. Contributor 1 also knows that there are efficiency gains to be had from creating a market mechanism whereby he and contributor 2 trade units of prestige for units of the private good, because their marginal rates of substitution between these two commodities will generally be different in the absence of trade. Contributor 1, therefore, decides that his organization will not only be a vehicle for the provision of the public good, but will also function as a “broker” in the market for prestige to be created. The charity works as follows. It sells gj+ zj-zk units of prestige to contributor j, j, k = 1, 2, j … k. The total revenue collected by the charity is, thus, g1+ g2. This quantity is used to finance the provision of G units of the public good. The brokerage function performed by the charity amounts to collecting zj units of the private good from contributor j in order to purchase zj units of prestige from contributor k, j, k = 1, 2, j … k. Hence, zk, k = 1, 2, k … j, denotes either the quantity of prestige sold by contributor j to the charity or the amount of private good offered, in excess of gk, by contributor k. Contributor 1 and the charity are assumed to be one and the same agent. The charity’s budget constraint is x1 + g1 + G + z1 - z2 = w1 + G + 1 , which implies x1 = I1 + 1 - g1 - z1 + z2 . (40a) Since τ1 + τ2 = 0, it follows that τ2 = - τ1. Hence, contributor 2's budget constraint yields x2 = I2 - 1 - g2 - z2 + z1 . 18 (40b) The charity, contributor 2 and the government play a two-stage game - called Game 4 - as follows. Game 4 Stage 1: Taking {g2, z2} as given, the charity chooses {g1, z1} to maximize 1 u I1 + 1 g1 , g2 , z1 , z2 - g1 - z1 + z2 , g1 + z1 - z2 , g1 + g2 . (41a) . (41b) Taking {g1, z1} as given, contributor 2 chooses {g2, z2} to maximize u 2 I2 - 1 g1 , g2 , z1 , z2 - g2 - z2 + z1 , g2 + z2 - z1 , g1 + g2 Stage 2: Having observed (g1, g2, z1, z2), the government chooses τ1 to maximize 1 u I1 + 1 - g1 - z1 + z2 , g1 + z1 - z2 , g1 + g2 + u 2 I2 - 1 - g2 - z2 + z1 , g2 + z2 - z1 , g1 + gsub2 . (42) Since there are only two contributors, we find it easier to utilize the budget constraints to solve for the xj’s and then substitute these quantities in the utility functions. Because in this section we assume that the social welfare function is utilitarian, the government’s objective function in the second stage of the game is given by (42). Note that in the second stage there is only one choice variable, τ1, but there are four exogenous variables, (g1, g2, z1, z2). The solution to the government’s maximization problem - assumed to be interior - is a unique and global constrained maximum. The first order condition that characterizes this solution enables us to implicitly define the function τ1(g1, g2, z1, z2). This function gives the government’s responses to changes in the levels of contributions and prestige. Inserting this function into equations (40) and the implied xj’s into the utility functions, we obtain the utility functions (41). The maximization problems in the first stage are now well defined. Our second main result is presented in Proposition 4. Proposition 4: The subgame perfect equilibrium for Game 4 is socially optimal. Proof. The first order condition for the second stage can be written as follows: 1 u x I1 + 1 g1 , g2 , z1 , z2 - g1 - z1 + z2 , g1 + z1 - z2 , g1 + g2 - u 2 x I 2 - 1 g1 , gsub2, z1 , z2 - g2 - z2 + z1 , g2 + z2 - z1 , g1 + g2 = (43) 0. Equation (43) yields the following response functions: 1 2 1 1 1 ux x + ux G - ux G - ux p = , 1 2 g1 ux x + ux x 2 2 1 2 1 ux p + ux G - ux G - ux x = , 1 2 g2 ux x + ux x 19 (44a) (44b) 1 2 1 ux p + ux p = 1- 1 , 2 z1 ux x + ux x 1 2 1 ux p + ux p = 1 -1. 2 z2 ux x + ux x Note that: 1 1 = , z1 z2 1 1 1 = . z1 partialg1 g2 (44c) (44d) (45) Assuming interior solutions, the Nash equilibrium in the first stage is characterized by the following equations: 1 1 1 uG u p + 1 = 1, 1 g1 ux ux 2 2 1 uG u p + = 1+ , 2 2 g2 ux ux 1 up = 11 ux 2 up = 1+ 2 ux 1 , z1 1 . z2 (46a) (46b) (46c) (46d) Then, the first equation in (45), (46c) and (46d) imply that 1 2 up up = 2 . 1 ux ux (47) 1 1 2 1 1 uG uG up usubp + 2+ 1+ = 2+ . 1 2 g2 g1 ux ux ux ux (48) Adding up (46a) and (46b) yields 2 Substituting (46d) into (48) and rearranging, we obtain 1 1 2 1 1 1 uG uG u p + + = 1+ . 1 2 1 g2 g1 z2 ux ux ux (49) Since the two equations in (45) imply that 1 1 1 = 0, g2 g1 z2 it follows that equation (49) corresponds to the modified Samuelson condition (5) for n = 2. € Proposition 4 is a “true” Rotten Kid Theorem. With the creation of a market for prestige, reluctant contributors behave as Good Samaritans if they anticipate that their benevolent government will redistribute income after levels of contributions and prestige have been allocated. 20 8. Concluding remarks We have, among other things, established a Rotten Kid Theorem in the impure public good model. An important implication of our analysis is that the creation of a market for prestige provides a rationale for the existence of charitable organizations.10 Charities should not only create the market for prestige, but also operate as brokers. Proper brokerage encourages reluctant contributors to agree with society about how much of the public good should be provided. To our knowledge, Harbaugh (1998a) is the only other study that formally examines prestige as a motivation for making charitable contributions.11 Harbaugh (1998b) provides empirical evidence that charities advertise contributions and allocate contributors into several categories according to the levels of their contributions. He also finds empirical support for the hypothesis that contributors possess a taste for prestige. These pieces of evidence appear to indicate that charities not only supply a market for prestige, but also operate as brokers in such a market. It is, therefore, quite possible that the levels of public goods provided by charitable organizations may indeed be socially optimal. 9. References Andreoni, J., 1989, Giving with impure altruism: applications to charity and Ricardian equivalence, Journal of Political Economy 97, 1447-1458. Andreoni, J., 1990, Impure altruism and donations to public goods: a theory of warm-glow giving, Economic Journal 100, 464-477. Andreoni, J., 1998, Toward a theory of charitable fundraising, Journal of Political Economy 106, 1186-1213. Andreoni, J., 2004, Philanthropy, Forthcoming, the Handbook on the Economics of Giving,Reciprocity and Altruism. L.-A. Gerard-Varet, S-C. Kolm and J. Mercier Ythier, eds., in the series Handbooks in Economics, K. Arrow and M.D. Intriligator, General Editors. Andreoni, J. and Bergstrom, T. C., 1996, Do government subsidies increase the private supply of public goods?, Public Choice 88, 295-308. Andreoni, J., and Petrie, R., 2004, Public goods experiments without confidentiality: a glimpse into fund-raising. Journal of Public Economics 88, 1605–1623. 10 In our simple model, we neglected the important fundraising function performed by charitable organizations. Andreoni (1998, 2004) and Rose-Ackerman (1982) are good examples of the literature on charitable fundraising. We believe that linking fundraising efforts with governmental matching grants may, in some circumstances, give rise to a Rotten Kid Theorem. This is an interesting agenda for future research. 21 Barro, R. J., 1974, Are government bonds net wealth?, Journal of Political Economy 82, 1195-1197. Becker, G. S., 1991, A treatise on the family, enlarged edition, Harvard University Press: Cambridge, Massachusetts. Bergstrom, T. C., 1989, A fresh look at the rotten kid theorem - and other household mysteries, Journal of Political Economy 97, 1138-1159. Bernheim, B. D., 1986, On the voluntary and involuntary provision of public goods, American Economic Review 76, 789-793. Bernheim, B. D. and K. Bagwell, 1988, Is everything neutral?, Journal of Political Economy 96, 308-338. Bernheim, B. D., A. Schleifer and L. H. Summers, 1985, The strategic bequest motive, Journal of Political Economy 93, 1045-1076. Boadway, R., P. Pestieau and D. Wildasin, 1989, Tax-transfer policies and the voluntary provision of public goods, Journal of Public Economics 39, 157-176. Bruce, N. and M. Waldman, 1990, The rotten-kid theorem meets the Samaritan’s dilemma, Quarterly Journal of Economics 105, 155-165. Buchanan, J. M.., 1975, The Samaritan’s dilemma, in Altruism, morality and economic theory, Edmund S. Phelps, Russell Sage Foundation: New York, pp. 71-85. Caplan, A. J. and Silva, E.C.D., 2005, An efficient mechanism to control correlated externality: redistributive Transfers and the coexistence of regional and global pollution permit markets, Journal of Environmental Economics and Management 49, 68-82. Cavalletti, B. and Levaggi, R., 2002, The provision of impure public goods in a non-competitive market Economic Modelling 19(1), 41-63. Cornes, R. C. and Sandler, T., 1984, Easy riders, joint production, and public goods, Economic Journal 94, 580-598. Cornes, R. C. and Silva, E. C. D., 1999, Rotten kids, purity and perfection, Journal of Political Economy 107, 10341040. Glazer, A. and Konrad, K. A., 1996, A signaling explanation for charity, American Economic Review 86, 1019-1028. Harbaugh, W. T., 1998a, What do donations buy? A model of philanthropy based on prestige and warm glow, Journal of 11 See footnote 6 for studies on similar motives such as social approval, reputation, etc. 22 Public Economics 67, 269-284. Harbaugh, W. T., 1998b, The prestige motive for making charitable transfers, American Economic Review (Papers and Proceedings) 88, 277-282. Kotchen M. J., 2005, Impure public goods and the comparative statics of environmentally friendly consumption, Journal of Environmental Economics and Management 49 (2), 281-300. Löschel, A. and Rübbelke, D., 2005, Impure Public Goods and Technological Interdependencies, FEEM Working Paper No. 60.05. Masclet, D., Noussair, C., Tucker, S. and Villeval, M., 2003, Monetary and non-monetary punishment in the voluntary contributions mechanism, American Economic Review 93 (1), 366-380. Mazzanti, M., Cainelli, G. and Mancinelli, S., 2005, Social capital, R&D and industrial districts, Fondazione Eni Enrico Mattei, Working Papers 84.05. Rege, M. and Telle, K., 2004, The impact of social approval and framing on cooperation in public good situations, Journal of Public Economics 88, 1625-1644. Romano, R. and Yildirim, H., 2001, Why charities announce donations: a positive perspective, Journal of Public Economics 81, 423–447. Rose-Ackerman, S., 1982, Charitable giving and Aexcessive fundraising, Quarterly Journal of Economics 97, 193-212. Rübbelke, D., 2005a, Foreign Aid and Global Public Goods: Impure Publicness, Cost Differentials and Negative Conjectures, International Environmental Agreements 5 (2), 151-173. Rübbelke, D., 2005b, Differing Motivations for Terrorism, Defence and Peace Economics 16(1), 19-27. Sandler, T., 1997, Global challenges: an approach to environmental, political and economic problems. Cambridge: Cambridge U. Press. Sandler T. and Hartley, K., 2001, Economics of alliances: the lessons for collective action, Journal of Economic Literature 39 (3), 869-896. Silva, E.C.D. and Zhu, X., 2005, Emissions trading of global and local pollutants, pollution havens and free riding, Journal of Environmental Economics and Management 58 (2), 169- 182. Silva, E.C.D. and Zhu, X., 2011, Efficient International Environmental Agreements for Correlated Transnational 23 Pollutants in the Presence of Free Trade of Goods and International Transfers, Strategic Behavior and the Environment 1 (2), 175 - 197. Silva, E.C.D. and Zhu, X., 2015, Overlapping International Environmental Agreements, Strategic Behavior and the Environment 5 (3-4), 255- 299. Van de Ven, J., 2002, The Demand for Social Approval and Status as a Motivation to Give, Journal of Institutional and Theoretical Economics 158, 464-482. Warr, P. G., 1982, Pareto optimal redistribution and private charity, Journal of Public Economics 19, 131-138. 24
© Copyright 2026 Paperzz