Abstract
This paper examines the socially optimal (and also equilibrium) locations of two stores or libraries on a linear market of unit length. If each consumer has probability \(w\) of finding a desired product at each store, then the socially optimal locations are never completely centralized for full information, but are completely centralized for \(w \le 0.5\) when costly visit search is necessary. The Nash equilibrium locations of two stores, and various alternative models for the socially optimal locations of two stores, are also examined.
Similar content being viewed by others
Notes
A few of these consider more than two firms, but they are applicable to two firms.
Simply set the partial derivatives of (3) with respect to \(x_1\) and \( x_2\) equal to zero and then solve for \( x_1\) and \( x_2\). Calculating the second derivative matrix of (3) shows that AETD is a strictly convex function of \( x_1\) and \( x_2\) over the feasible solution set (see Definition 1), so we have a local and global minimum of AETD.
If there were three stores (or libraries), the consumer would not necessarily visit the closest store first (see footnote 2), even if (as in this paper) the prices of the stores are assumed to be fixed and equal (see paragraph 3 of Sect. 10).
If \(0 \le x \le x_1\), and store 1 does not have the product, then the consumer continues directly on to store 2, and then returns home. The right- hand side of (4) could also be written as \(2w( x_1 - x) + 2(1 - w)(x_2 - x)\). If \( x_1\le x \le (x_1 + x_2)/2\), and store 1 does not have the product, then the consumer passes by his home on the way to store 2, and then returns home for good.
Note also that the level curve that is tangent to the third boundary locus is strictly concave (when expressed in terms of \(x_2\) as a function of \(x_1\)) in a neighborhood of (0.5, 0.5). The level curve that forms the boundary of the less-than-or-equal set might curl around and be a closed curve that is interior to the loci \(x_2 = x_1, x_2 = 0, \hbox {and}\; x_1 = 1\) (this is not the feasible solution set) except at (0.5, 0.5). It is not necessary to determine its exact shape, since all that matters is that its only intersection with the feasible set is at (0.5, 0.5).
Proposition 4 is the same as Proposition 1 of Braid (2008), and Proposition 1 in Sect. 3 of this paper can be derived from it if we set \(b = {w^{2}}\) and \(c = w(1-w)\). Also, a variation of Proposition 4 with quadratic transportation costs is stated in a preliminary section of Braid (2011). However, the socially optimal location results for costly visit search (see Propositions 2 and 5), which are the main results of this paper, are completely new to the literature.
Variable prices might very well lead to different results. If \(w = 1\), there is a central location tendency, but a price equilibrium does not exist if the firms are inside the quartiles (see d’Aspremont et al. 1979). This can be avoided with mixed-price strategies (see Osborne and Pitchik 1987), or with travel costs that are proportional to distance squared (see d’Aspremont et al. 1979), in which case the sellers move apart to the locations (0, 1). Also, if \(w < 1\), the equilibrium prices might be infinite for any set locations, since each seller has a monopoly over consumers for whom it has the desired good and the other seller does not. To avoid this, it might be possible to introduce a reservation price of each consumer (as in Economides 1984).
If the length of the jurisdiction is \(L\), AETD would also be multiplied by \(L^{2}\).
I was assuming that each consumer has the same completely inelastic demand. This assumption is often, though far from always, used in spatial models. Mulligan (2012) generalizes this in several ways. In this section, I generalize it in the simplest possible way, by assuming that each consumer has the same (finite) reservation price, as in Beckmann (1972), Economides (1984) and Salop (1979).
As one of many possible alternatives, it might be possible to put one store at location 3/10, and consumers in the interval [1/10, 1/2] would visit it (since the equation \(wv = 2td\) implies that \(d =\) (1/2)(1/6)(12/5) \(=\) 1/5), and the other store at location 7/10, and consumers in the interval [1/2, 9/10] would visit it. However, even though a (slightly) greater number of consumers would visit a store, fewer would find what they want (each consumer who travels is visiting one store rather than two), and the overall consumer surplus would be less (and the overall producer surplus would also be lower, if the model is interpreted as stores, and if the price, assumed to be fixed, exceeds marginal costs, assumed to be constant).
Since some consumers do not make a trip with any of these socially optimal locations, it is natural to imagine that it might be socially optimal to have more than two stores (or libraries). In fact, in terms of libraries, the significant cost of getting downtown for many consumers is a reason for a large city to have a central (main) library and a significant number of branch libraries (paragraphs 2 and 3 of Sect. 10 are somewhat related to this).
Visit search is also assumed in much of Braid (2014), which has variable prices but fixed store locations.
References
Anderson SP, de Palma A (1988) Spatial price discrimination with heterogeneous products. Rev Econ Stud 55:573–592
Anderson SP, Neven D (1991) Cournot competition yields spatial agglomeration. Int Econ Rev 32:793–808
Beckmann MJ (1972) Spatial Cournot oligopoly. Pap Reg Sci Assoc 28:37–47
Berry BJL, Garrison WL (1958a) The functional bases of the central place hierarchy. Econ Geogr 34:145–154
Berry BJL, Garrison WL (1958b) A note on central place theory and the range of a good. Econ Geogr 34:304–311
Bigman D, ReVelle C (1978) The theory of welfare considerations in public facility location problems. Geogr Anal 10:229–240
Biscaia R, Mota I (2013) Models of spatial competition: a critical review. Pap Reg Sci 92:851–871
Braid RM (1996) The optimal locations of branch facilities and main facilities with consumer search. J Reg Sci 36:217–234
Braid RM (1999) The socially optimal locations of three stores with stockouts or limited product selections. Econ Lett 64:363–368
Braid RM (2008) Spatial price discrimination and the locations of firms with different product selections or product varieties. Econ Lett 98:342–347
Braid RM (2011) Bertrand-Nash mill pricing and the locations of two firms with partially overlapping product selections. Pap Reg Sci 90:197–211
Braid RM (2014) Search costs decrease prices in a model of spatial competition. Pap Reg Sci 93:125–139
d’Aspremont C, Gabszewicz JJ, Thisse J-F (1979) On Hotelling’s ‘Stability in competition’. Econometrica 47:1145–1150
de Fraja G, Norman G (1993) Product differentiation, pricing policy and equilibrium. J Reg Sci 33:343–363
de Palma A, Ginsburgh V, Papageorgiou Y, Thisse J-F (1985) The principle of minimum differentiation holds under sufficient heterogeneity. Econometrica 53:767–781
Eaton BC, Lipsey RG (1982) An economic theory of central places. Econ J 92:56–72
Economides N (1984) The principle of minimum differentiation revisited. Eur Econ Rev 24:345–368
Economides N (1986) Minimal and maximal product differentiation in Hotelling’s duopoly. Econ Lett 21:67–71
Fischer JH, Harrington JE (1996) Product variety and firm agglomeration. Rand J Econ 27:281–309
Fujita M, Thisse J-F (2002) Economics of agglomeration: cities, industrial location, and regional growth. Cambridge University Press, Cambridge
Gupta B (1992) Sequential entry and deterrence with competitive spatial price discrimination. Econ Lett 38:487–490
Hamilton JH, Thisse J-F, Weskamp A (1989) Spatial discrimination: Bertrand versus Cournot in a model of location choice. Reg Sci Urban Econ 19:87–102
Harwitz M, Lentnek B, Rogerson P, Smith TE (1998) Optimal search on spatial paths with recall: theoretical foundations. Pap Reg Sci 77:301–327
Harwitz M, Lentnek B, Rogerson P, Smith TE (2000) Optimal search on spatial paths with recall, part II: computational procedures and examples. Pap Reg Sci 79:293–305
Hsu SK (1983) Pricing in an urban spatial monopoly: a general analysis. J Reg Sci 23:165–175
Massam BH (1974) A note on expected distances. Geogr Anal 4:377–381
Meza S, Tombak M (2009) Endogenous location leadership. Int J Ind Organ 27:687–707
Mulligan G (2012) Stochastic shopping and consumer heterogeneity. Stud Reg Sci 42:93–107
Mulligan G, Partridge MD, Carruthers JI (2012) Central place theory and its reemergence in regional science. Ann Reg Sci 48:405–431
Narula SC, Lentnek B, Harwitz M (1987) A contextual analysis of the journey-to-shop with price uncertainty. J Reg Sci 27:403–418
Osborne MJ, Pitchik C (1987) Equilibrium in Hotelling’s model of spatial competition. Econometrica 55:911–922
Quinzii M, Thisse J-F (1990) On the optimality of central places. Econometrica 58:1101–1119
ReVelle C (1991) Siting ambulances and fire companies. J Am Plan Assoc 57:471–484
Salop SC (1979) Monopolistic competition with outside goods. Bell J Econ 10:141–156
Stahl K (1982a) Differentiated products, consumer search, and locational oligopoly. J Ind Econ 31:97–113
Stahl K (1982b) Location and spatial pricing theory with nonconvex transportation cost schedules. Bell J Econ 13:575–582
Tabuchi T, Thisse J-F (1995) Asymmetric equilibria in spatial competition. Int J Ind Organ 13:213–227
Wolinsky A (1983) Retail trade concentration due to consumers’ imperfect information. Bell J Econ 14:275–282
Acknowledgments
I would like to thank two anonymous referees for helpful comments and suggestions.
Author information
Authors and Affiliations
Corresponding author
Appendix
Appendix
This Appendix derives Eqs. (13) and (14), which are used in Propositions 7 and 8 of Sect. 7. In the full-information case of Sect. 3, if the two stores (libraries) are at the locations described in Proposition 1, it follows by substituting these locations into (3), multiplying by tD and adding 2\(f\), that aggregate social costs are
If, on the other hand, there is a single combined store (library) at location 0.5, the aggregate expected travel distance is the probability that a consumer makes a trip, which is \(2w - w^{2}\), multiplied by the average round-trip travel distance of consumers who do make the trip, which is 0.5. Thus, aggregate social costs are simply
Comparing (15) and (16), it follows that one single combined store in the middle is better than two separate stores if and only if
which is the same as (13) of the text.
In the costly visit-search case of Sect. 4, if \(0 < w \le 0.5\) and the two stores (libraries) are at the locations described by (12) in Proposition 2, it follows by substituting these locations into (8), multiplying by tD and adding 2\(f\), that aggregate social costs are
If there is a single combined store (library) at location 0.5, the aggregate expected travel distance is the probability that the consumer makes a trip, which is 1 (since consumers do not know ahead of time whether they will find what they want), multiplied by the average round-trip travel distance of consumers who do make the trip, which is 0.5. Aggregate social costs are thus
Comparing (18) and (19), the single combined store at location 0.5 is always better than two separate stores at the same location 0.5 as long as \(F < 2f\), which is an obvious result. Consider now the other case in Proposition 2. If \(0.5 \le w \le 1\), and if the two separate stores are at the locations described by (11) in Proposition 2, then substituting these locations into (8), multiplying by tD, and then adding 2\(f\), shows that aggregate social costs are
If there is a single combined store at location 0.5, the aggregate social costs are given by (19). Comparing (19) and (20), it is seen that one single store in the middle is better than two separate stores if and only if
which is the same as (14) of the text.
Rights and permissions
About this article
Cite this article
Braid, R.M. The socially optimal and equilibrium locations of two stores or libraries with consumer search. Ann Reg Sci 53, 123–136 (2014). https://doi.org/10.1007/s00168-014-0620-6
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00168-014-0620-6