This action might not be possible to undo. Are you sure you want to continue?
David Ong March 4, 2008
Monopoly (Heterogenous Products)
In Hotelling (Linear City or spatial) model, everyone is "a little di¤erent" so that each person has a di¤erent response to price changes. Hence, they don’ just all go to the cheaper …rm as in Bertrand Competition with t homogenous goods. Notation: F C =…xed cost of each of n outlets, N =number of consumers, a=reservation value of the all consumers, c=marginal cost. Here CS=consumer surplus. The tan colored region is revenue minus variable costs. Note that …xed costs are not drawn.
a1 a a2 a2 a3 a4 a5 a6 a7 a8 a9
Transport costs P=a-tx/2 π(N,n)+F c 0 VC
Here every consumer has the same valuation before travel costs are taken into account. t is the cost per unit of travel and x is the distance travelled. When travel cost tx is taken into account consumers become di¤erentiated in their willingness to buy. Now the maximal price that any consumer i is willing to pay is pi = a txi , which depends upon how far i has to travel. As is generally the case, a monopolist must charge all consumers the same price. We assume that the monopolist sells to everyone. If the whole market is x long, then the monopolist charges so that y is the one whose willingness to pay is equal to the sum of his travel costs and the price that the monopolist is charging: a = p + tx:
Pro…t Maximizing Variety
As always, we want to maximize pro…t. (Q) = (P c) Q FC
But, now we need to adapt the pro…t function to this problem.
the change in total surplus is only determined by the net change in transportation costs (purple areas lined with green). We multiply the average pro…t the number of N to get pro…t function. Combining price and quantity. 2 . which means that the highest price which can be charged is equal to the value of the consumer minus the highest travel cost=tx. n (n + 1) The pro…t maximizing number of varieties is the last n + 1 such that (n + 1) n 1.2 Socially Optimal Variety If you look at the change in surplus when we increased the number of outlets from 1 to 2. That’ because social s surplus is the sum of CS and PS.. p (n) = a tx 2n where x is the greatest distance. pro…t is maximized if we produce up to last unit that marginal pro…t is greater than zero. Since we sell to everyone and the number of varieties n a¤ects the price. pro…t maximization amounts to a condition on the number of outlets. (n) = a tx 2n c N nF C Now. you will see that though some consumer surplus CS became the producer surplus PS. we want to minimize social cost. If we were at Q then the pro…t of the next unit would be (Q + 1) : We would build that Q + 1 unit if1 : (Q + 1) (Q) 0 In the Hotelling model. we can …ll in the usual pro…t max equation. 2. with the di¤erence that: 1. From standard economics. we can only make discrete choices here. 1 We could have also used Q instead of Q + 1 but then we couldn’ use the "marginal" concept and we would have to delve t into negative numbers of the …rst unit.e. we have to take into account …xed costs also in this choice. travel cost and …xed costs. The only thing that changes social surplus is the net change in transportation costs and …xed costs.For simplicity. which I will normalize to 1. we are going to assume that all N consumers are served. all we have to do is decide how many stores to have. With n shops evenly spread apart. x becomes the fraction of consumers that are served. Note that when distance is normalized. To maximize social surplus. we build up to the ast n + 1 outlet that still gives an increase in pro…ts: (n + 1) This amounts to: (n + 1) = a tx 2 (n + 1) c N (n + 1) F C a tx N 2n tx 2n c N nF C = (n) (n) tx N + FC 2 (n + 1) FC txN 1 2n (n + 1) txN 2F C Nt 2F C . the farthest that any customer needs to travel is 2n . Minimizing travel costs thus maximizes the price the …rm can charge. Hence. in choosing n we are in e¤ect choosing a price that maximizing pro…ts.i. Travel costs are minimized if varieties are spread x evenly. Thus. social surplus is unchanged when some CS becomes PS or vis versa.
and 2. the lowest travel cost consumer at distance 0. Multiplied by the number of 8 consumers. the best place to put them is 1 x and 3 x which again. and 2. we look for the last n + 1 outlet that decreases costs. 1.n)+F Transport cost to PS c 0 VC 1 If everyone is served. So. the best place to put the outlet is in the middle. the highest travel cost consumer at distance x. C (n + 1) C (n + 1) = C (n) 0 txN + nF C = C (n) 4n txN + (n + 1) F C 4 (n + 1) 1.3 Costs Lets develop the calculation of costs through an example. who has travel costs is then 1. minimizes travel costs. the social optimal number of outlets is the cost minimizing number of outlets.3. tx is 2 4 = tx : The pattern is then 4n where n is the number of outlets.3. who has zero travel costs. who has zero travel costs. the middle person’ travel cost. the total travel cost is then: C (n) = txN + nF C 4n 3 . therefore. = tx 4 . where the marginal or next unit is still resulting in a cost decrease.2 0+ tx 2 2 tx 2 . the lowest travel cost consumer at distance 0. the highest travel cost consumer at distance x.1 Social Costs When There Is One Outlet If there were a single outlet.a1 a a2 a2 CS a3 a4 a5 a6 a7 CS a8 a9 tx/4 Transport cost CS to Transport cost CS to PS Transport cost P=a-tx/4 Transport cost to PS π(N. maximal consumer travel costs is tx : The average travel cost between 2 1. The 4 4 maximal travel costs is then tx : The average travel cost between 4 1. who has travel costs 0+ tx tx 4 . s Social Costs When There Are Two Outlet If there are 2 outlets.
x = 1: (n) = a tx 2n c N nF C What is the pro…t maximizing number of outlets? outlets 1 2 3 Price 10 221 10 222 10 322 VC 9 1 9 1 9 1 Total FC 1 3 2 3 3 3 Total 69: 70:5 69 Another way to put the same thing using pro…t equation explained above: (9. Gresebomm’ ¯ s Fries is a gourmet french fry stand franchise. 2) = 10 (9. N = 9. 3) = 10 It would open 2 outlets. F C = 3. this means a = 10. Where would these be located? At equal distances to minimize travel costs. 1) = 10 (9. Everyone loves a serving of Gresebomm’ fries about as much as they love $10.Total costs starts high because of …xed costs and approach some minimum. c = 1.4 Example: There are only 9 equally distributed people in the mile long x=1 town of Couch Potato Texas. txN + (n + 1) F C 4 (n + 1) n (n + 1) txN + nF C 4n txN 4F C N xt 4F C : The optimal number of …rms from social perspective is the last n + 1 such that : (n + 1) n 1. From the above equations. t = 2. The last n + 1 where costs stops decreasing is the above expression. 2 21 2 22 2 32 1 9 1 9 1 9 1 3 = 69 2 3 = 70: 5 3 3 = 69 4 . but all su¤er a travel cost from s walking of t = $2 per mile. His costs consists M C = 1 for lard and F C = 3 for the stand.
the connection between hotelling and interdependent markets. 1.when there are equilibria. Salop. The …rm doesn’ care about the t consumer’ travel costs per see. The di¤erence between what is socially optimal and what is pro…t maximizing comes from the fact that the …rm cares about the travel costs of the consumer only in so far as the …rm can induce the consumer to travel and purchase. the socially optimal number of outlets is 2. there may be some wiggleroom for the best price.5 Finding the Pro…t Maximizing Price when the Monopolist Does not Sell to Everyone Since the choice of the price via the choice of n was descrete. interpretating Hotelling in terms of standard demand. so it won’ a¤ect the socially optimal number t of outlets.) From any …xed location. (Refer to graph from section.What price would it charge? P =a tx = 10 2n 2 = 9:5 3 2 What would be the socially optimal number of outlets? outlets 1 2 3 Total Trans Cost 2 221 9 2 222 9 2 322 9 VC 9 1 9 1 9 1 Total FC 1 3 2 3 3 3 Total 16: 5 17: 25 19: 5 Note that variable cost is constant for any number of outlets. Thus. the text is using some consumer between the location of the vendor and x (p. 2) = S (9. Society cares about total travel AND …xed costs. we can make it a function of p: p) t This is a sort of "quantity demanded" function. but only s s sets a bound on how much it can charge. to maximize it. Price is pro…t maximizing if the loss from either increasing or decreasing the price is outweighed by the loss. Pro…t for N customers from both directions is then: r= = max 2N (p p (V c) (V t p) There is no …xed costs. since that’ not a part of its costs. let r (be the fraction) of the total distance of the consumer served. because we take number of stores and location to be given. Travel costs are not borne by it. given n. Another way to put the same thing using cost equation explained below C (n) = txN nF C: 4n S (9. 3) = 2 221 2 222 2 322 1 9 1 9 1 9 1 3= 2 3= 3 3= 16: 5 17: 25 19: 5 Thus. their total 2 travel cost is: txN : 4n THE FOLLOWING IS OPTIONAL. competition. 1) = S (9. FOC: 2N which means that p = V +c 2 : (V t p) (p t c) = 2N ((V t 2p + c)) = 0 2N 2 (V c) t Next:demand functions in Hotelling from Carabas. Then p (n) = V tr: Since we have allowed r to be variable. Social surplus includes …rm and consumer surplus. The average x tx travel costs of someone travelling from 2n away is 1 0 + 2n : When there are N such people. 142). where instead tx of using the farthest consumer at x in p (n) = V 2n . Bertrand as the limiting case of Hotelling = 5 . we just want to minimize travel costs.
This action might not be possible to undo. Are you sure you want to continue?
We've moved you to where you read on your other device.
Get the full title to continue reading from where you left off, or restart the preview.