3 Equilibrium and Efficiency in an Exchange Economy 341 The 2 x 2 Exchange Economy Key ideas: Pareto efficiency, Edgeworth Box diagram, first and second welfare theorems In this chapter the focus shifts from the individual agent to the market. Specifically we examine the allocation that results if all economic agents are price takers and prices adjust until markets clear. Rather than attempt to bring firms and consumers into the analysis all at once, we focus here on equilibrium in which there is no production, Consumers have endowments of commodities that they may exchange. As we see later, the ideas developed here extend very directly to economies with production. Even though this chapter focuses on equilibrium outcomes, itis helpful to keep in mind a possible adjustment process that might lead to equilibrium. Suppose that there is an auctioneer who calls out prices for each of the com- modities. Consumers and firms respond with the demands that they would make at these prices, The auctioneer lowers prices in markets where there is, excess supply and raises them in markets where there is excess demand. At a price-taking (Walrasian) equilibrium, all markets clear. The central results are the two welfare theorems. The First welfare the- ‘orem formalizes Adam Smith's argument about the “invisible hand” of the market place. Under extremely weak assumptions we establish that a Wal- rasian equilibrium (WE) is efficient. The Second welfare theorem estab- lishes that in a convex economy, any efficient allocation can be supported as a WE with appropriate lump-sum transfers. In this first section we focus on a simple exchange economy in which there are two consumers, A (Alex) and B (Bev). The set H = (A, B} is then the set of consumers. Each consumer has an endowment of two commodities. ‘Commodities are private. That is, each consumer cares only about his or her86 Equilibrium and Efficiency in an Exchange Economy vu 45 line UtsUt ak uA Figure 3.1-1. Utility possibility set. ‘own consumption. Consumer h, h € H has an endowment a, a consumption set X* = R} and a utility function U(x") that is strictly monotonic. Pareto Efficiency With more than one consumer, the social ranking of allocations requires weighing the utility of one individual against that of another. Suppose that the set of possible utility pairs (the “utility possibility set”) associated with all possible allocations of the two commodities is the shaded area depicted in Figure 3.1-1. Setting aside the question of measuring utility, one philosophical approach to social choice places each individual behind a “veil of igno- rance.” Not knowing which consumer you are going to be, it is natural to assign a probability of } to each possibility. If individuals are neutral toward tisk while behind the veil of ignorance, they will prefer allocations with a higher expected utility suas 1ue, This is equivalent to maximizing the sum of utilities, a proposal first put forth by Jeremy Bentham. In Figure 3.1-1, the Benthamite criterion picks the point B. A more recent philosophical argument developed by John Rawls argues that behind the veil of ignorance consumers will be infinitely averse to risk and thus will place all weight on the worst possible outcome. The social cri- terion then become the Max Min criterion. Maximizing the minimum utility is achieved by moving out along the 45° line to the utility possibility frontier: the point W in the figure.3.1 The 2 x 2 Exchange Economy 87 Economists tend to be agnostic when it comes to theorizing about social choice rankings. Instead they focus on minimizing unnecessary waste. The utility allocation A in Figure 3.1-1 is wasteful or “inefficient” because there are alternative allocations of goods that would make both individuals better off. Both the Benthamite and the Rawlsian allocations are said to be Pareto efficient (or simply “efficient”) because the only way to raise the utility of one individual is by reducing the utility of the other. Generalizing to more than two individuals we have the following definition. Pareto Efficient Allocation A feasible allocation of commodities is Pareto efficient (PE) if there is no other feasible allocation that is strictly preferred by at least one consumer and is weakly preferred by all consumers. For the special 2 x 2 case, Alex and Bev must share the aggregate endow- ment « = (oi, 2). Let £° be the allocation to Bev and let B be the set of allocations that Bev prefers over £%. This is depicted in Figure 3.1-2. For any x ¢ B, the allocation to Alex is x4 = @ — x4. Thus the best pos- sible allocation to Alex that leaves Bev no worse off is Alex's utility maxi- mizing allocation in B. Figure 3.1-3 shows Figure 3.1-2 rotated 180°, This depiction of the pref- erences of both consumers in a rectangle is called an Edgeworth box dia- ‘gram, Note that at the bottom left corner Bev consumes the entire endow- ment. Therefore this is the zero consumption point for Alex. Also shown is the indifference curve for Alex through ¢“ = » — #®, Any point in the > of Figure 3.1-2. Bev’s upper contour set.88 Equilibrium and Efficiency in an Exchange Economy Ur) =U4G) of Figure 3.1-3. Edgeworth box diagram, intersection of B and Alex's upper contour set (x|U4(x4) > U4(2e4)) is strictly preferred by both Alex and Bev. Then the allocation depicted (84, 2" = w — £4) is not PE. For efficiency, there can be no such mutually preferred alternative. One such allocation is depicted in Figure 3.1-4. As long as an allocation (£4, ¢” @ — £4} isin the interior of the Edgeworth box, a necessary condition for the allocation to be PE is that the slopes of the two indifference curves must be Figure 3.1-4. PE allocations with identical CES preferences.3.1 The 2 x 2 Exchange Economy 80 equal. Thus the graph of the PE allocations is the set of allocations to Alex (and hence Bev) satisfying a8 (gBy a 5 A = Gem Where = W984, a) Mathematically, an allocation is Pareto efficient if we © argMax {U(x4)|U%(x") = U%(2%), x4 +2" <0}. Example: Identical Constant Elasticity of Substitution (CES) Preferences If preferences are CES with elasticity of substitution o, both consumers have a marginal rate of substitution, MRS"(x") = k(3)"°. If a PE alloca- tion is in the interior of the Edgeworth box, the indifference curves of the two consumers must have the same slope; that is, af) ap ato =) =(5 hence 2, = 3, Appealing to the Ratio Rule! and then setting demand equal to supply, af xe xf taP xf xP aaa a ‘Thus, in a PE allocation each consumer is allocated a fraction of the aggre- gate endowment. It follows that for each consumer the marginal rate of substitution is ve wnstat) = (2) " G14) ‘The PE allocations are depicted in Figure 3.1-4. Walrasian Equilibrium for an Exchange Economy In a WE each consumer is a price taker. We write the set of consumers as H so in the two-person economy H={A, B). Consumer h ¢ H, with endowment « has preferences represented by the utility function U(x") where x" is the private consumption of consumer h. The value of the con- sumer’s endowment is p - wand so the consumer chooses a bundle of goods, x*(p, o'), that solves Max(U"(x)|p-x < p-a) them ay = Kas and by = kbaand soa, + by = kg +b). Hence 221 =90 Equilibrium and Efficiency in an Exchange Economy Let p20 be a price vector of this exchange economy. Define @ = Dhev" to be the vector of total endowments in the economy and X(p) = Cpe X"(p. 0%) to be total (or “market”) demand. Then the vector of excess demands is (p) =x(p) — @. ‘A market clears if either excess demand is zero or itis negative and the price of the commodity is zero. Definition: Market-Clearing Prices Let z;(p) be the excess demand for commodity j at the price vector p > 0. The market for commodity j clears if z;(p) <0 and pjzj(p) = 0. Walras’ Law We assume that the preferences of each consumer satisfy the local non- satiation axiom (discussed in Chapter 2). Given this axiom, each consumer must spend all of his income. To see this, we suppose instead that consumer / spends less than the value of his endowment and seek a contradiction. If the consumer’s optimal consumption bundle x" satisfies p «x < p- a", then for 4 > 0 and sufficiently small, the 5— neighborhood N(x", 8) lies in the budget set. Yet by local non-satiation, there must be some consumption bundle in M(x", 6) that is strictly preferred to x", However, if this is true, x” cannot be ‘optimal after all so we have a contradiction. We now show that for any price vector p the market value of excess demands must be zero. First note that p (Ee'-e)-De = pooh) Because all consumers spend their entire wealth the right-hand expression is zero. Hence P+ 2p) = pi (x—a) Pizi(p) + prza(p) = 0. ‘This is known as Walras’ Law. Suppose that the first market clears, that is pizi(p) = 0. Then p»2z2(p) = 0 so market 2 must clear as well? Note also that if one market does not clear, the other cannot clear either. Definition: Walrasian Equilibrium The price vectorp > 0 is a WE price vec- tor if all markets clear. For our two-commodity model, it follows from Walras’ Law that we need to consider market clearing in only one market. Take any price vector 2 With H consumers and n commodities an identical argument establishes that p-2(p) = O.Thusifn ~1 markets clear then the remaining market must also clear3.1 The 2 x 2 Exchange Economy 91 on Figure 3.1-5. Excess supply of commodity 1, P=(p1, P2). Given an endowment allocation (w*,*}, each consumer chooses his utility-maximizing consumption bundle. Itis helpful to depict trades in the Edgeworth box diagram. However there is one important caveat. From the viewpoint of each consumer, the budget set is the set of non-negative consumption bundles between the origin and the budget line. Consumers know nothing about the size of the Edgeworth box. In equilibrium all markets must clear. However, the demands of an individual consumer do not take into account aggregate supply constraints. For this reason, each side of the Edgeworth box is depicted as an axis for either Alex or Bev and indifference curves are depicted extending outside the box. As depicted, in Figure 3.1-5, Alex wants to trade from the endowment point N to his most preferred desired consumption ¥“, whereas Bev wishes to trade from N to ¥®. Thus, there is excess supply of commodity 1. By lowering the price of commodity 1 (relative to commodity 2) the bud- get line becomes less steep until eventually supply equals demand. The Wal- rasian equilibrium E is depicted in Figure 3.1-6. Equilibrium and Efficiency In Figure 3.1-6 the Walrasian equilibrium (WE) allocation is in the interior of the Edgeworth box. Thus the marginal rates of substitution must both be equal to the price ratio: MRS4(z4) MRS4(z"), where2 Equilibrium and Efficiency in an Exchange Economy o lve=ure"y ot x Figure 3.146, Walrasian equilibrium. Comparing this condition with the necessary condition for an allocation to be PE, it follows that the WE allocation must be PE. To prove that this result holds very generally, we appeal to the Duality Lemma. In Section 2.2 we argued that if the local non-satiation property holds, then the utitity-maximizing bundle is cost minimizing among all pre- ferred consumption bundles. Formally, if a = argMax{ U"(x")|p-x" <= p- a"), then pe" = Min{ p- x*|U"(x") > U"@)), With this observation, the proof that a WE is Pareto efficient is short and simple. Proposition 3.1-2: First Welfare Theorem for an Exchange Economy If preferences satisfy local non-satiation, a WE allocation in an exchange economy is PE. Proof: Let {¥"},,., be a WE allocation for the exchange economy with endowments {o/"},,..,. Let p > 0 be the WE price vector. Because ¥* max- imizes consumer hs budget constrained utility, any strictly preferred bundle x must cost strictly more, that is px" > p-¥*. Moreover, by the Duality ‘Lemma (Lemma 2.2-3) any weakly preferred bundle x" must cost at least as much as ¥*; that is p- x" > p-¥".3.1 The 2 x 2 Exchange Economy B Consider any allocation {x"),.., that is Pareto-preferred to ("}jcy. Because none of the consumers can be worse off in the Pareto-preferred allocation, it follows that poxt— p20, hel Moreover at least one consumer must be strictly better off. Thus p-x"—p-%">0_ for someh. Summing over consumers, (gr-EH)>0 heh het Also all markets clear in a Walrasian equilibrium. Therefore o(E het ‘Combining these results yields °( Because p> 0, it follows that there must be some commodity j such that EpeyX) Engie] > 0. Thus all Pareto-preferted allocations are infeasible? a Clearly consumers’ WE allocations depend on their initial endowments. ‘Thus the outcome will change if a government intervenes and redistributes income. We now argue that, as long as preferences are convex, any PE allocation is also a WE allocation with the appropriate redistribution of resources. We first sketch the argument for the two commodity case. Consider the PE allocation £4 — $4 in Figure 3.1-7. The indifference curves are tangential at ¢4, that is, ae) Hea ‘Therefore, for some 6. ae eM) > Note that the proof does not appeal othe assumption that these ofconsumersisH = [A,B Itholds for any numberof consumer94 Equilibrium and Efficiency in an Exchange Economy \ o ur =Ue") % Figure 3.1.7. PE allocation (24,8" = — Since the upper contour set X4 = {x4|U4(x4) = U4(24)} is convex, it fol- lows from Lemma 1.1-2 that Ue) = UAC) > Ge) (e428) 20, Making the same argument for Bev, ue UP (x?) = URRB) = (88) (xP — 29) 2 0. Choose p = *2*(2"). Then re =6pand so UA(xA) = UA(RA) = p-x4 > p84 and UF (x8) > U8(89) = p-x® = p- 28. Suppose that UA(x4) > U4(%4) and p-x4 = p- £4, Then, since U is strictly increasing and continuous, there exists 5 >> 0 and sufficiently small such that U4(x4 — 8) > UA(#4) and p-(x4 —8) < p- £4, However, we have just argued that U4(x4) > U4(e4) > p-x4 > p- 24,0 thisis impos- sible. Therefore UA(x4) — UA(R4) > O => pox > p24 Hence no allocation that Alex strictly prefers is in his budget set (x4|p Ac pth x4 < p34), ‘An identical argument holds for Bev. Because demand equals supply for each individual, all markets clear. Therefore, the vector p is a WE price vector. Define the transfer payment 7" = p- (" — a"), h € H.3.1 The 2 x 2 Exchange Economy 95 Because Dycyt" = Cpcy 0" the sum of these transfers is zero so this is a feasible redistribution of wealth. The budget constraint p- x" < p- £" can be rewritten as follows pists poh +T ‘Then given transfers T*, h € H, the price vector p is a WE price vector. ‘This argument holds for PE allocations in which consumer h € H has an allocation £" >> 0. We now prove a more general result. Proposition 3.1-3: Second Welfare Theorem for an Exchange Economy Tn an exchange economy with endowments {o"} ,> 0, he H. Then any PE allocation {2"}),-4 where £" # 0, h € H, can be supported by aprice vector p > 0. Proof: We prove this for the two-person exchange economy. If £4,£° = o4 +08 — £4 isa PE allocation then £4 € argMax {U4(x4)x4 4.2% < 04 +0, U(x) > UR(R®)} 3.2) Let the aggregate endowment be w. We begin by examining the feasible set of this optimization problem. Consider ¥ = 6 > Oand¥" = w — 6, Because U®(.) is strictly increasing and 0 < $" 0 it follows from (3.1-3) that v > 0. From (3.1-5) it then follows that aP=0. ot to* (3.1-6) Because £” > 0 and 347 >> O it follows from (3.1-4) that . > 0. Now consider an economy with endowments é/" = £", h € H and consider the price vector p = v. Consumer h chooses w= argMax(UM(a")Iv -x* < v2) ‘The FOC for this optimization problem a aut we) =v <0, where " (ao - wy) =0. Moreover, because U*(.) is quasi-concave the FOC is also sufficient. Choose a4 = Land 4 = 1/y. Then, appealing to (3.1-3) and (3.1-4), the FOC hold atv =<'heH, Thus at the price p = v no consumer wishes to trade, Therefore supply ‘equals demand and so the price vector is a WE price vector. Finally define transfers 7 = v - (#" — w). Appealing to (3.1-6), the sum_ of these transfers is zero. Consumer h’s budget constraint with these trans- fersis viat On! Om = MRSp(o1, 0). (3.1-7) ‘This is depicted in Figure 3.1-9. We now explore the implications of this assumption on the Pareto efficient allocations. Consider the Edgeworth box diagram shown in Figure 3.1-10.98, Equilibrium and Efficiency in an Exchange Economy | Figure 3.1-9, Alex has a stronger preference for commodity 1 First note that along the dotted diagonal line (x}, x4) = ¢w. By hypoth- esis, preferences are homothetic. Because Alex places a higher value on commodity 1, MRS4(94@) > MRSg(04@) = MRSp(0%0). It follows that the Pareto efficient allocations must lie below the diagonal. Let C be an efficient allocation and C’ be a second such allocation pre- ferred by Alex. In Figure 3.1-10, C’ must lie to the northeast of C. Intuitively, because Alex’s consumption is higher and Bev’s is lower at C’, the marginal rate of substitution at C’ will be higher, reflecting the greater influence of ‘Alex's stronger preference for commodity 1 To confirm this, let the MRS at C be m. Given homothetic preferences, for any point above the line 04D, MRS4(xj, x2) > m. Also, for any point above the line OF, MRS4(x,.x2) < m. Hence in the upper shaded region oO ‘Commodity 2 Commodity 1 Figure 3.1-10, Pareto efficient.3.1 The 2 x 2 Exchange Economy 99 Alex has a MRS exceeding m, while Bev has a MRS less than m. It follows that no such point can be Pareto efficient. A symmetric argument establishes that allocations in the lower shaded region are also not Pareto efficient. Consider the efficient allocation C’ to the northeast of C. Because this point lies below O4D and above OpF, it follows that af xt] | <<] > neH, .1-8) Ae “tle and MRS"(C’) > m=MRS"(C), hE H. (3.1-9) It follows that, on the map of Pareto efficient allocations between C to C’, the marginal rate of substitution increases. Hence the supporting prices must have the property that the relative price of commodity I rises. Thus the greater the wealth of Alex relative to Bev, the higher will be the relative price of commodity 1. Because we refer to these results later, we summarize them below. Proposition 3.1-4: Pareto Efficient Allocations with Homothetic Preferences In the 2x2 exchange economy, suppose both consumer A and B have homothetic preferences. Suppose also that at the aggregate endowment, consumer A has a stronger preference for commodity 1. Then at any inte- rior efficient allocation, xf xP x w xf Moreover, along the locus of efficient allocations, as consumer A’s utility rises, the consumption ratio x# /x/' and marginal rate of substitution of x; for 29 of both consumers rise, Exercises Exercise 3.1-1: Prices with Quasi-Linear Preferences Consider a two- person economy in which the aggregate endowment is (01, 2) = (100, 200) and both have the same quasi-linear utility function U(x") = x! +. (a) Solve for the Walrasian equilibrium price ratio under the assumption that the ‘equilibrium consumption of commodity 1 is positive for both individuals, (b) What is the range of possible equilibrium price ratios in this economy? Exercise 3.1-2: Pareto Efficient Allocations (a) If U4 and U® are strictly increasing, explain why the allocation (24, 28) (o +08, 0) isa PE and WE allocation,100 Equilibrium and Efficiency in an Exchange Economy Suppose that U4 = xf! + 10Inxf and U® = Inx? + x#. The aggregate endowment is @ = (20, 10). (b) Show that the PE allocations in the interior of the Edgeworth box can be ‘expressed in the form #4 (©) Suppose that «3! = f(«!). How does the equilibrium price ratio change as « increases along this curve? (d)_ Which allocations on the boundary of the Edgeworth box are PE allocations? Exercise 3.1-3: Walrasian equilibrium Suppose half the population (the Biggs) each have an endowment (24, 8) and the other half (the Littles) each have an endowment (20, 10). Each Mr. Bigg has a utility function U® = Inxf + Inx?. fh Ms. Little has a utility function UF =In(4+x}) +In(6+x4) (a) By solving for and then aggregating individual demands, solve for the equilib- rium price ratio. (b) Solve also for the contract curve and depict it in a neat Edgeworth box diagram showing the bottom left-hand corner as the zero consumption point for a rep- resentative Ms. Little and the top right-hand corner as the zero consumption point for a representative Mr. Bige. (©) Explain carefully why the equilibrium price will not change if endowments are reallocated in favor of the Biggs. (d) What will be the equilibrium price ratio if the Littles have an endowment of (8,0) while the Biggs have an endowment of (36, 18)? (©) What if the Littles have an endowment of (2, 0) and the Biggs have an endow- ment of (42, 18)? Exercise 3.1-4: Linear Preferences Consider a 2 person economy in which Alex’s preferences are represented by the utility function U(x) = 2x, + x9, while Bev’s preferences are represented by the utility function U(x) = xy + 2x. The total endowment is (30, 20). (a) Characterize the PE allocations and dey them in an Edgeworth box. Show that if Alex has a sufficiently large fraction of the total endowment the equilibrium price ratiop1/p2 = 2. What if Bev has a large fraction of the total endowment? (b) For what endowments will the price ratio ie between these two extremes? Char- acterize the Walrasian equilibrium, (©) Show that for some endowments a transfer of wealth from Alex to Bev has no effect on prices. Also show that for other endowments there is no effect on the ‘Walrasian equilibrium allocation.3.2 The Fundamental Welfare Theorems 101 Exercise 3.1-5: More on the Biggs and Littles Suppose that the Biggs only like commodity 1, whereas the Littles only like commodity 2. Depict the PE allocations in a neat Edgeworth box diagram. Is there a Walrasian equi- librium? If so depict it. Exercise 3.1-6: Market Excess Demand Consider the following two-person exchange economy. Alex and Bev each has a consumption set X= {x|x > (2,2)}. Alex has a consumption utility function U4 = (xf! —2)5(x#! — 2) and endowment 4 = (7 +a, 1 ~ a). Bev has a utility function U = (x? — 2)(xP = 2)° and endowment w” = (1 —a,7 +a). The parameter a € [0, 1]. (a) Show that for both consumers to be able to purchase a bundle in their consump- tion sets the price ratio must satisfy Stal pr lta Ita” p (b) Solve for each consumer's demand for commodity 1 (c) Hence show that the market excess demand function for commodity 1 is, cin 48) (@) For what values of a does the excess demand for commodity 1 increase as the price of commodity 1 increases? Provide some intuition for this paradoxical result, HINT: Consider the income effects of an increase in p}, (©) Ia = O characterize the Walrasian equilibrium prices and allocations. 3.2 The Fundamental Welfare Theorems Key ideas: Walrasian equilibrium, First and Second welfare theorems We now consider an exchange economy with an arbitrary number of com- ‘modities and consumers. As we see, the insights gleaned from the two- person two-commodity economy generalize. Commodities are private, that is, consumer h € H = (1,..., H) has preferences over his own consump- tion vector x" = (x},...,x!) and not over those of other consumers. Let X" CR" be the consumption set of consumer h. That is, preferences are defined over X". We assume that consumer h has an endowment vector a! e X*. A consumption allocation in this economy (x"}, 0 be the price vector. Consumers are price takers. Consumer h has an endowment o/ She chooses a consumption bundle ¥* in her budget set (x! € X*|p.x" < i pol Walrasian Equilibrium Consumer h chooses a most preferred consumption plan ¥* in her budget set. That is, u*@) = UNC), for all x! such that px" < p- a! ‘Let ¥ = }¥" be the total consumption of the consumers. Excess demand is then Definition: Walrasian Equilibrium Prices The price vector p>0 is a Walrasian equilibrium price vector if there is no market in excess demand (@ <0) and pj = 0 for any market in excess supply (Z) < 0). Welfare Theorems ‘The proof of the First welfare theorem is exactly the same as for the two- commodity case examined in Section 3.1. In the proof of the Second welfare theorem for the two-person exchange economy we assumed quasi-concavity and differentiability of utility functions and appealed to the Kuhn-Tucker3.2 The Fundamental Welfare Theorems 103 conditions. Here we drop the differentiability assumption and appeal directly to the Supporting Hyperplane Theorem.> ‘The allocation (£"}jcH is PE if there is no alternative allocation that increases the utility of consumer 1 without lowering the utility of at least one other agent. Thus {2"},<1 must solve the following optimization problem: Max Ul(e)UN) = UNE") A= 2,... HO (o" = x") > 0,2" RY on he Define V(x) to be the maximum utility for consumer 1 given an aggregate supply of x, that is, VG) = Mas | UiGeh|wAey = ume a h=2, storm Sota (32-1) = Note that ("<1 solves this optimization problem if x . Lemma 3.2-1: Quasi-Concavity of V'(.)_ If U", h ¢ H is quasi-concave then so is the indirect utility function V1(-). Prooft Consider the aggregate endowments a and 6. We must show that for any convex combination ¢ = (1 —A)a +b, V'(c) > Min{Vi(a), V1(b)). Suppose that the allocation{a"},,.., solves the optimization problem (3.2-1), with aggregate endowment a and that (5"), . solves the optimization prob- lem when the aggregate endowment is b. Then V'(a) = U'(a") and V(b!) = U'(b!), We must then show that V'(c) = Min{V#(a), V4(b)}. Because {a"},,.., is feasible with a and {b"}/!_, is feasible with b, the convex combination {c"},.4, is feasible with aggregate endowment c = (1 —A)a + Ab. Then V'(c) > U'(c!). Appealing to quasi-coneavity, U'(c!) > Min{U(a"), U'(b")}. Combining the last two inequalities, V'(c) > Min{U"(a'), U'(b")} = Min(V1(a), V"()} i Proposition 3.2-2: Second Welfare Theorem for an Exchange Economy Consumer h ¢ H has an endowment o! ¢ R\. The consumption set for each individual X" is the positive orthant R. Suppose also that utility functions U"(), h € Hare continuous, quasi-concave and strictly increasing. If ("}jey 5 Aswe seein Chapter 5, this proot can be easily modified to include proton.104 Equilibrium and Efficiency in an Exchange Economy FOO Figure 3.2-1. Supporting hyperplane. where 2" # 0,6 h € His a PE allocation, then there exists a price vector p > 0 such that Ut(xh) > UN") > pox > ps", hed, Proof: Define t Via) Max UN(x!)|UN(") = UNA), = 2,2. Hx Doxt of. ih Appealing to Lemma 3.2-1, V¥(.) is quasi-concave. Also V4(.) is strictly increasing because U'(.)is strictly increasing and any increment in the agere- gate supply can be allocated to the first consumer. We have already noted that {@"},<4 is the solution of this optimization problem ifx =o. Moreover, because U"(.) is strictly increasing, (322) fat Because « is on the boundary of the set {x|V!(x) > V'(w)}, it follows from the Supporting Hyperplane Theorem that there isa vector p #0, such that V(x) > Via) > pox > p-o and V(x) > Vw) > p-x2 p-o. (23) ‘The supporting line through the aggregate endowment vector is depicted in Figure 3.2-1 6 irthere are M consumers wit a zero allocation, west these aside an appeal tothe theorem for he TEM consumers with non-zero allocations. Because ll feasible allocations are strilly preferred tothe zero allocation, the theorem then extends immediately.References and Selected Historical Reading 105 We now argue that the vector p must be positive. If not, define 8 = (61, ---,4x) > 0 such that 5; > 0 if and only if pj <0. Then V'(w+8) > V1(w) and p- (w+) < po. But this contradicts (3.2-3) so p must be posi- tive after all. From (3.2-3) and the definition of the indirect utility function, if H= p-x=p-ox"> p-w. (32-4) fet ut(x') > UN e), = 1 Substituting for « from (3.2-2) it follows that t t UN xh) > UN 1...,H=> p-Yoxt= poe. = mi Setting x* = 2", k # h, we may then conclude that for consumer h, UNG") > UNG) => pox = pa 625) It remains to show that any strictly preferred bundle costs strictly more. ‘Suppose instead that U'(x") > U(z") and p- x" = p- 2%. Then for all 4 € (0,1), p- 2x" < p-2*. Also because U*(.) is continuous, for alld sufficiently close to 1, UN(.x") > UM(t"). But these two inequalities contradict (3.2-5). Hence UM(x") > UR") = px! > p- 2". o Exercise 3.2-1: Walrasian Equilibrium with Identical Homothetic Prefer- ences Suppose each consumer has a consumption set and the same strictly increasing, quasi-concave, homothetic utility function U ¢ C!. Char- acterize Walrasian equilibrium prices. References and Selected Historical Reading Edgeworth, Francis Y. (1881). Mathematical Psychics: An Essay on the Application of Mathematics to the Moral Sciences. London: C. K. Paul. Pareto, Villredo. (1909). Manuele di economia politica. Trans. Augustus M. Kelley, 17. Rawls, John, (1971). A Theory of Justice. Cambridge, MA: Belknap Press. Walras, Leon, (1874). Eléments d’économie politique pure, ou théorie de la richesse sociale. Lausanne: F. Rouge Editeur.