Economics 440 Public Finance Winter 2022 Allen Head Answers to Assignment 2 1a. When firms are independent, the manufacturing firm chooses its output taking into account the cost of pollution only to itself. Thus its profit maximization problem is given by: 2 max Pm qm − 3qm − (x − 2)2 (1) qm ,x | {z } Cm (qm ,x) subject to: x = .5qm (2) 2 2 max Pm qm − 3qm − (.5qm − 2)2 = Pm qm − 3.25qm + 2qm − 4 (3) or qm The first-order condition is given by: Pm − 6.5qm + 2 = 0 (4) Letting Pm = 13.5 and solving we have: 6.5qm = 13.5 + 2 = 15.5 15.5 qm = ≈ 2.38 6.5 x = .5qm ≈ 1.19. 1b. The merged firm maximizes total profits from both enterprises. Its profit maximization problem is: 2 (5) max Pm qm + Pf qf − 3.25qm + 2qm − 4 − 3qf2 − 1.5qm qm ,qf {z } | | {z } Cm Cf Taking the first-order condition for choice of manufacturing output, qm , and solving for the case in which Pm = 13.5 we have: 13.5 + .5 − 6.5qm = 0 6.5qm = 14 qm ≈ 2.15 x = .5qm ≈ 1.075. 1 The level of pollution in this case differs from that in part a. because the merged firm takes into account the effect of pollution on the cost of fishing, whereas the independent manufacturer does not. The merged firm thus produces a lower level of manuafacturing output (qm = 2.15 rather than 2.38) and as a result pollutes less (x = 1.075 rather than 1.19). 1c. If the government imposes a tax on the independent manufacturing firm for each unit of output that it produces, then its profit maximization problem becomes: 2 max Pm qm − 3.25qm + 2qm − 4 − τ qm qm (6) Taking the first-order condition and solving under the assumption that Pm remains constant at 13.5, we have: 13.5 + 2 − τ − 6.5qm = 0 6.5qm = 15.5 − τ τ = 15.5 − 6.5qm We want to find the level of τ such that the independent manufacturer emits the same pollution as it would if it were merged with the fishing firm, i.e. x = 1.075. Since x = .5qm , it is sufficient to solve the equation above for the tax amount, τ that sets qm = 2.15, the quantity of manufacturing output produced by the merged firm: τ = 15.5 − (6.5 × 2.15) τ = 1.525. 2. To derive the Samuelson condition consider the social welfare maximization problem, where each household gets equal weight: max αx1 + (1 − α) ln y + αx2 + (1 − α) ln y x1 ,x2 ,y (7) subject to: h y i γ1 = ω1 + ω2 = Ω (8) δ Form the Lagrangian in the usual way, and derive the first-order conditions by differentiating with respect to x1 , x2 , y, and the Lagrange multiplier, λ, respectively: x1 + x2 + α = λ α = λ 1−α 1−α 1 h y i γ1 −1 + = λ y y γδ δ h y i γ1 x1 + x2 + = Ω δ 2 Now, divide the third first-order condition by λ and substitute using the first two: 1−α 1−α 1 + = αy αy γδ | {z } | {z } | 1 M RSyx h y i γ1 −1 δ {z M RT yx 2 M RSyx . (9) } Notice that this is the Samuelson condition. In this case, solving for y we can write γ 2δ(1 − α) y=γ . (10) α Thus, the optimal level of public good provision is determined by the parameters α, δ, and γ, and does not depend on how the private good is distributed across households. Notice also that when the planner weighs the two households the same, it does not matter how much each consumes of the private good. If the planner assigns higher weight to either of the households, then that household will get all of the private good; and the other will get utility only from the public good. 3a. To show that drivers have incentive to drive an inefficient speed, we compare the first order conditions from the individual drivers’ utility maximization problems to those from a social welfare maximization problem. Driver h chooses his/her speed to solve: max uh (sh ) − π(s1 , s2 )ch sh (11) The first-order condition for this problem is: ∂π ∂uh − h ch = 0 h ∂s ∂s h = 1, 2. (12) So from the two drivers’ problems with have two equations: ∂u1 ∂π − 1 c1 = 0 1 ∂s ∂s 2 ∂u ∂π 2 − c = 0 ∂s2 ∂s2 in the two unknown speeds, s1 and s2 . We compare these equations with the first order conditions from the following social welfare maximization problem: max ψ u1 (s1 ) − π(s1 , s2 )c1 + (1 − ψ) u2 (s2 ) − π(s1 , s2 )c2 (13) 1 2 s ,s Choosing (without loss of generality) ψ = max 1 2 s ,s 1 2 we have 1 1 1 u (s ) + u2 (s2 ) − π(s1 , s2 )(c1 + c2 ) 2 3 (14) The first order conditions for this problem are: ∂u1 ∂π − 1 c1 + c2 = 0 1 ∂s ∂s 2 ∂u ∂π 1 2 = 0 c + c − ∂s2 ∂s2 These two equations are clearly different from the two above derived from the individual drivers’ problems. In the social optimum, the planner recognizes that when an accident occurs, costs accrue to both drivers. Individual Driver h, however, considers only the cost to him or herself, ch , rather than the total cost, c1 + c2 . 3b. If Driver h pays fine th in the event of an accident, then his/her maximization problem becomes: h = 1, 2. (15) max uh (sh ) − π(s1 , s2 ) ch + th sh The first-order condition for this problem is: ∂uh ∂π h h − c + t =0 ∂sh ∂sh h = 1, 2. (16) Now, if t1 = c2 and t2 = c1 , then the first-order conditions from the individual households’ problems are identical to those for the social welfare maximization problem. In this case, the efficient fine involves each driver paying the other driver’s accident cost as well as incurring their own. 3c. No, it is not. If drivers are fined in the event of an accident, then the cost to each driver is effectively paid twice. First, Driver h incurs cost ch , because of the accident. Then, he/she has to pay the other driver’s cost as a fine. In the end the government ends up with t1 + t2 = c2 + c1 in revenue from fines, which has reduced drivers’ utility. Now we might think that the government could spend this revenue in such a way that it would compensate the drivers for the lost utility due to fines. If people know they are going to get their fines back, however, then the fine will not induce them to drive any slower! This policy does induce drivers to drive the correct speeds, but it lowers their expected utility through fines. 4a. If households take G as given, then they will not voluntarily contribute anything for provision of public good. Each household thinks that their individual contribution has no effect on the stock of the public good. In particular, they all think that if they contribute nothing, this will have no effect on the amount of public good available for their consumption. So, each household maximizes utility by spending all of their income on private consumption, and contributing nothing to public good provision: xy = Y gh = 0 4 h = 1, . . . , H (17) Now in this case, feasibility dictates that G= M X g h = 0. (18) h=1 Since no public good is provided in equilibrium, each household’s gets utility is U (xh , G) = αY × 0 = 0. (19) This equilibrium allocation is inefficient. There are many ways to see this. Recall that U (xh , G) = 0 for all households. Now consider the following feasible allocation that is Pareto superior to the equilibrium. Suppose we take a small amount, , of private good away from each household and use it to make public good. Now the total amount of public good produced is H X G= = H > 0. (20) h=1 Utility in this case is U (xh , G) = α(Y − )H > 0. (21) Since all households get higher utility in this feasible allocation than in the equilibrium allocation, the equilibrium allocation is not Pareto efficient. Another way to see that the equilibrium is not Pareto efficient is to derive the Samuelson condition and show that it does not satisfy it. Note that with the given utility function, h M RSGx xh = G h = 1, . . . , H (22) and M RTGx = 1. So, the Samuelson condition can be written: H X xh h=1 G =1 or G= H X xh . (23) h=1 Clearly this is not satisfied if G = 0 and xh = Y for all h. 4b. If the government levies lump-sum tax, τ , on each household, then the households all have maximization problem: max αxh G (24) xh ,G subject to: xh + g h = Y − τ (25) Households still take the stock of the public good as given, and so they continue to consume their (now after-tax) income, and contribute nothing voluntarily to the production of public good. So the equilibrium allocation is: G = Hτ xh = Y − τ 5 h = 1, . . . , H. (26) Now we want the τ so that Hτ is the efficient level of public good. Recall the Samuelson condition: H X G= xh . (27) h=1 In the equilibrium here, this becomes: Hτ = H(Y − τ ) solving, τ= Y . 2 (28) With these taxes we have xh = Y 2 G= HY . 2 (29) Clearly, this allocation (the equilibrium with taxes) satisfies the Samuelson condition. 4c. In this case, households realize that when they change their level of contribution to provision of the public good, g h , the stock, G, is affected. In this case, they solve: " # H X (30) max αxh g h + gj xh ,g h j6=h {z | } G subject to: xj + g j = Y (31) Their Lagrangian is, " Lh = αxh g h + H X # g j − λ h xh + g h − Y . (32) j6=h The first order conditions are: α H X gj = λ j=1 αxh = λ xh + g h = Y Now, assume that all households behave symmetrically, that is g h = g, xh = x, etc.. Using the first two first-order conditions to eliminate the third we have: H X gj + gh = Y j=1 6 → Hg + g = Y (33) Solving, we get Y H +1 The stock of the public good satisfies, g= x= G = Hg = HY H +1 HY = x. H +1 (34) (35) Note: G 6= 0. With the number of households small enough so that they recognize the influence that they have on the stock of the public good, they will realize that it is not in their interest to contribute zero. Still, this is not the efficient outcome, as it does not satisfy the Samuelson condition (assuming H > 1!). Households consider only the benefits of their contribution to public goods provision that accrue to them directly, not the total benefits to society. 7