Abstract
This study incorporates the strategic behavior of outsourcing with a variant Hotelling model to explore the role of input outsourcing in determining equilibrium locations for firms under quadratic transportation. Given that strategic input outsourcing occurs, we show that the cost-efficient integrated firm will locate as far away from its rival as possible, so as to increase its rival’s input price when its own cost advantage is small; at the same time, the cost-inefficient downstream firm likes to locate closer to its rival to lower the input price. Hence, there is an interior locational equilibrium, and the principle of maximum differentiation does not hold. When an integrated firm’s cost advantage is large, the principle of maximum differentiation is valid. However, when the integrated firm’s own cost advantage is even larger, the integrated firm can become a monopolist through strategic input outsourcing. Under this case, the equilibrium location depends on the magnitude of the input homemade ratio when the input homemade ratio is small. Otherwise, the integrated firm would like to locate at the middle of the market.
Similar content being viewed by others
Notes
The previous version of this paper (Lin and Tu 2013) was presented at “International Conference on Applied Economics 2013.” The version here has a good number of modifications with location equilibrium under a monopoly market structure and offers some different interesting results.
An other example of our theoretical framework goes as follows. Samsung, a leading mobile phone maker, is a vertically integrated firm, manufacturing not only mobile handsets, but also telecommunication ICs. Its competitor, HTC, is a company that focuses solely on the production of mobile phones and not on telecommunication ICs. However, both Samsung and HTC procure an input, the telecommunication IC in a handset, from Qualcomm.
Matsushima (2004) considered a Hotelling-type product differentiation model to investigate the location strategies of upstream and downstream firms. Aiura and Sato (2008) analyzed the location equilibrium with a duopoly spatial competition model in which there exists a raw material site located at the center of the linear market. Matsushima (2009) set up a Hotelling-type location model with upstream and downstream duopolies to investigate the relationship between the strategies of downstream firms for product differentiation and vertical integration. Kourandi and Vettas (2010) discussed the location equilibrium with upstream and downstream duopolies in a linear city model. Matsumura et al. (2010) investigated the location equilibrium by allowing firms to exchange the cost-efficient production technology via royalties in the spatial framework.
If \(\alpha = 0\), then it implies that the marginal cost of two firms are the same. Therefore, the location equilibrium would be the same as D’Aspremont et al. (1979), and the principle of maximum differentiation holds. If \(\alpha = 1\), then it implies that there is cost asymmetry between two firms, similar to the models in Liang and Mai (2006) and Matsushima and Matsumura (2009).
Under the case of \(c_{v }> c_{u}\), the purpose of input outsourcing is to seek cheaper suppliers, and it does not incorporate strategic behavior. Therefore, we do not consider this case.
Substituting the equilibrium prices of the duopoly case (i.e., 4.1 and 4.3) into the inequality: \((p_{d}-p_{v}) < \varPhi _{1 } \equiv t(x_{d}-x_{v}) (2 - x_{d}-x_{v})\) for the condition of the duopoly market structure, then we arrange the inequality as: \(\alpha ( w-c_{v}) < \varPhi _{2} \equiv t(x_{d}-x_{v})(4 - x_{d}-x_{v})\). Similarly, we also can have the condition of the monopoly market structure: \(\alpha (w-c_{v}) \ge \varPhi _{2} \equiv t(x_{d}-x_{v}) (4 - x_{d}-x_{v})\).
After considering that the upstream firm compares the profits under duopoly and monopoly, we can get the “globally” optimal wholesale price which maximizes the upstream firm’s profit. However, the math is complicated and it does not bring any new insights. Thus, we treat the market structure as exogenous determined to show our main results in the paper. The detailed mathematics in “globally” optimal wholesale price is available upon request.
The assumption \(c_{v}\le c_{u}\) requires that \(\alpha \ge 6 / (10 - x_{d}-x_{v})\) in order to make sure the duopoly market condition, i.e., \(\varPhi _{3}\) is positive.
The competition effect is negative for firm \(v: ( \partial \pi _{v}^{D}/ \partial p_{d}^{D}) = [(p_{v}^{D}-w^{D}) + \alpha (w^{D}-c_{v})] /2t(x_{d}-x_{v}) > 0\) from (3.1) and \((\partial p_{d}^{D}/ \partial x_{v}) = ( - 2t/3)(2-x_{v}) < 0\) from (4.4). Similarly, the competition effect is positive for firm \(d: ( \partial \pi _{d}^{D}/ \partial p_{v}^{D}) ( \partial p_{v}^{D}/ \partial x_{d}) = (p_{d}^{D}-w^{D}) (1+x_{d})/3 > 0\). The hinterland effect for firm v is \(( \partial \pi _{v}^{D}/ \partial x_{v})=q_{v}^{D} [(p_{d}^{D}-p_{v}^{D})+t(x_{d}-x_{v})^{2}]/(x_{d}-x_{v}) > 0\) and that for firm d is \((\partial \pi _{d}^{D}/ \partial x_{d}) = - [(p_{d}^{D}-w^{D})/2] \{1 - [(p_{d}^{D}-p_{v}^{D}) / t(x_{d}-x_{v})^{2}]\} < 0.\)
The strategic effect and outsourcing effect for firm v are \((\partial \pi _{v}^{D}/ \partial p_{d}^{D}) (\partial p_{d}^{D}/\partial w^{D})( \partial w^{D}/ \partial x_{v}) = - (3 - \alpha ) [3 - \alpha (1+x_{v})][(p_{v}^{D}-w^{D})+\alpha (w^{D}-c_{v})]/6\alpha ^{2}(x_{d}-x_{v}) < 0\) and \((\partial \pi _{v}^{D}/ \partial w^{D})( \partial w^{D}/ \partial x_{v}) = [t(1 - \alpha ) q_{v }^{D}] [3 - \alpha (1+x_{v})] /\alpha ^{2 }> 0\), respectively. Similarly, the strategic effect and outsourcing effect for firm d are \([(\partial \pi _{d}^{D}/ \partial p_{v}^{D}) ( \partial p_{v}^{D}/ \partial w^{D}) ( \partial w^{D}/ \partial x_{d})] = (p_{d}^{D}-w^{D}) (3 - 2 \alpha ) [3 - \alpha (1+ x_{d})] / 6 \alpha ^{2 }(x_{d - }x_{v}) > 0\) and \([( \partial \pi _{d}^{D} / \partial w^{D})( \partial w^{D}/ \partial x_{d})] = - tq_{d}^{D} [3 - \alpha (1+x_{d})] / \alpha ^{2 }< 0\), respectively.
Using footnote 9, we sum up the strategic effect and outsourcing effect for firm d, finding that \([( \partial \pi _{d}^{D}/ \partial w^{D}) ( \partial w^{D }/ \partial x_{d})]+ [( \partial \pi _{d}^{D}/ \partial p_{v}^{D}) ( \partial p_{v}^{D}/ \partial w^{D}) ( \partial w^{D}/ \partial x_{d})] = \{[3 - \alpha (1+x_{d})]/\alpha ^{2}\}\{ - tq_{d}^{D} +[ (p_{d}^{D}-w^{D}) (3 - 2 \alpha ) / 6 (x_{d - }x_{v})]\} = \{[3 - \alpha (1+x_{d})]/\alpha ^{2}\}\{ - \alpha [ - \alpha ( w^{D}-c_{v})+ t(x_{d}-x_{v})(4 - x_{d}-x_{v})]\}\). Using (5.3), it can be further arranged to \(\{ - 6 tq_{d}^{D} [3 - \alpha (1+ x_{d}) ]/9\alpha \} < 0\) due to \(0<\alpha < 1\). This implies that the outsourcing effect always dominates the strategic effect for firm d.
From (11.2), we derive the second-order condition as: \(( \partial ^{2}\pi _{d}^{D}/ \partial x_{d}^{ 2}) = - 6t + 10 \alpha t - 6 \alpha tx_{d}\). Substituting (13.1) into it fulfills the second-order condition: \((\partial ^{2}\pi _{d}^{ D }/ \partial x_{d}^{ 2}) = - 2\sqrt{t^{2}(5\alpha -3)^{2}+3\alpha ^{3}(c_{u}-c_{v})}< 0.\)
It fulfills the second-order condition as: \((\partial ^{2}\pi _{v}^{M}/ \partial ^{2}x_{v}) = - 2t < 0\).
References
Arya A, Mittendorf B, Sappington D (2008) The make-or-buy decision in the presence of a rival: strategic outsourcing to a common supplier. Manag Sci 54:1747–1758
Aiura H, Sato Y (2008) Welfare properties of spatial competition with location-dependent costs. Reg Sci Urban Econ 38:32–48
Brekke KR, Straume OR (2004) Bilateral monopolies and location choice. Reg Sci Urban Econ 34:275–288
Buehler S, Haucap J (2006) Strategic outsourcing revisited. J Econ Behav Organ 61:325–338
Chen Y, Ishikawa J, Yu Z (2004) Trade liberalization and strategic outsourcing. J Int Econ 63:419–436
Chen Y, Dubey P, Sen D (2011) Outsourcing induced by strategic competition. Int J Ind Organ 29:484–492
D’Aspremont C, Gabzewicz J, Thisse J-F (1979) On Hotelling’s stability in competition. Econometrica 47:1145–1150
Egger H, Egger P (2007) Outsourcing and trade in a spatial world. J Urban Econ 62:441–470
Friedman JM, Thisse J-F (1993) Partial collusion fosters minimum product differentiation. RAND J Econ 24:631–645
Hotelling H (1929) Stability in competition. Econ J 39:41–57
Jehiel P (1992) Product differentiation and price collusion. Int J Ind Organ 10:633–643
Kourandi F, Vettas N (2010) Endogenous spatial differentiation with vertical contracting. CEPR Discussion Papers, 7948
Liang WJ, Mai CC (2006) Validity of the principle of minimum differentiation under vertical subcontracting. Reg Sci Urban Econ 36:373–384
Lin YJ, Tu KI (2013) Vertical outsourcing and location choice. Procedia Econ Finance (Int Conf Appl Econ 2013) 5:494–501
Mai CC, Peng SK (1999) Cooperation vs. competition in a spatial model. Reg Sci Urban Econ 29:463–472
Matsumura T, Matsushima N, Stamatopoulos G (2010) Location equilibrium with asymmetric firms: the role of licensing. J Econ 99:267–276
Matsushima N (2004) Technology of upstream firms and equilibrium product differentiation. Int J Ind Organ 22:1091–1114
Matsushima N (2009) Vertical mergers and product differentiation. J Ind Econ 57:812–834
Matsumura T, Matsushima N (2009) Cost differentials and mixed strategy equilibrium in a Hotelling model. Ann Reg Sci 43:215–234
Neven DJ (1986) On Hotelling’s competition with non-uniform customer distributions. Econ Lett 21:121–126
Pierce A, Sen D (2014) Outsourcing versus technology transfer: Hotelling meets Stackelberg. J Econ 111:263–287
Salop SC, Scheffman DT (1983) Raising rival’s costs. Am Econ Rev 73:267–271
Salop SC, Scheffman DT (1987) Cost-raising strategies. J Ind Econ 36:19–34
Sappington D (2005) On the irrelevance of input prices for make-or-buy decisions. Am Econ Rev 95:1631–1638
Shy O, Stenbacka R (2003) Strategic outsourcing. J Econ Behav Organ 50:203–224
Tabuchi T, Thisse J-F (1995) Asymmetric equilibria in spatial competition. Int J Ind Organ 13:213–227
Ziss S (1993) Entry deterrence, cost advantage and horizontal product differentiation. Reg Sci Urban Econ 23:523–543
Acknowledgments
We are grateful to Professor Fu-Chuan Lai, Wen-Jung Liang, Hong Hwang, two anonymous referees and seminar participants at 3rd International Conference on Regional Science for their valuable comments, leading to substantial improvements of this paper. The usual disclaimer applies. Financial support from the Ministry of Science and Technology, R.O.C (NSC 101-2410-H-431-006) is gratefully acknowledged.
Author information
Authors and Affiliations
Corresponding author
Electronic supplementary material
Below is the link to the electronic supplementary material.
Appendices
Appendix 1: Equilibrium prices of final goods under a monopoly market
Under the monopoly downstream market case, we have the price relationship between two final goods by using \(\hat{{x}}=1\):
Substituting (4.2) and (4.4) into (2.3), we derive the demand of final goods for firm d:
Letting \(q_{d}^{D }= 0\), we have \(\alpha (w-c_{v})=t(x_{d}-x_{v}) (4 - x_{d}-x_{v})\). Substituting \(\alpha ( w-c_{v})\) into (4.2), we arrive at the equilibrium price of firm v under the monopoly case:
Substituting (19) into (17), we have the equilibrium price of firm d: \(p_{d}^{M }=w\). Substituting these prices into (3.2), we then have the profits of firm v when it monopolizes the whole downstream market: \(\pi _{v}^{M} = \alpha \) \((w-c_{v})-t(x_{d}-x_{v}) (2 - x_{v}-x_{d})\).
Appendix 2: Equilibrium input price under a duopoly market
Under the duopoly market case, the total profits of upstream firm u are: \(\pi _{u}^{D}= \pi _{u}^{D}(q_{v}^{D} (w), q_{d}^{D} (w)\), w). Differentiating it with respect to w, we obtain the equilibrium input price, which is (8.1). From (7.1), we find that ( \(\partial \pi _{u}^{D}/ \partial q_{v}^{D}) = (w-c_{u})(1 - \alpha ) > 0, ( \partial \pi _{u}^{D}/ \partial q_{d}^{D}) = (w-c_{u}) > 0\), and \(( \partial \pi _{u}^{D}/ \partial w)=q_{d}+ (1 - \alpha )q_{v}\). Using (5.1) and (5.3), we have \((\partial q_{v}^{D}/ \partial w)= - ( \partial q_{d}^{D}/ \partial w) = \alpha /6t(x_{d}-x_{v})\). We then can derive the equilibrium input price under the duopoly market case when \(0 < \alpha < 1\), which is (9) in the context.
Equilibrium input price under a monopoly market
Under the monopoly market case, upstream firm u’s profit is (7.2). Differentiating (7.2) with respect to w, we find that \((\partial \pi _{u}^{M }/ \partial w) = (1 - \alpha ) > 0\). It implies that the upstream firm would like to set the wholesale price as high as possible to increase its profit. However, we should consider that all of the consumers are served in the market; it requires that the utility of the farthest consumer in the market be positive. When \(0 \le x_{v} \le 1/2\), the farthest consumer locates at 1, his utility should be positive after purchasing the final good: \(U(x =1) = k - p_{v}^{M}-t(1 - x_{v})^{2} \ge 0\). When \(1/2 < x_{v} \le 1\), the farthest consumer locates at 0, his utility should be positive after purchasing the final good: \(U(x=0) = k - p_{v}^{M} - tx_{v}^{2 } \ge 0\).
Substituting (19) into \(U(x =1)\) and \(U(x=0)\), respectively, we find that the input price has an upper bound, which is \(w^{M} \le k -t(1 - x_{d})^{2}\) if 0 \(\le x_{v} \le 1/2\). Similarly, we also can have an upper bound of the input price, which is \(w^{M} \le k + t [2(x_{d}-x_{v})-x_{d}^{2}]\) if \(1/2 < x_{v} \le 1\).
Considering the above conditions, we obtain the equilibrium input price under the monopoly market case: \(w^{M}=k -t(1 - x_{d})^{2}\) when \(0 \le x_{v}\le 1/2\) and \(w^{M}=k + t[2(x_{d}-x_{v})-x_{d}^{2}]\) when \(1/2 < x_{v} \le 1\) respectively.
Appendix 3: Interior location equilibrium with a duopoly market condition for firm d
Substituting \(x_v^{{D}^{*}} =0\) and \(x_v^{{D}^{*}} =[t(5\alpha -3)+\sqrt{A}]/3\alpha t\) (i.e., (12) and (13.1)) into the duopoly market condition \((\varPhi _{3})\), we obtain that the two downstream firms can survive under the interior location equilibrium if \((c_{u}-c_{v}) \le \varPhi _{3}^{I } \equiv (t/^{3})(5 \alpha - 3)^{2}\), where superscript “I” denotes variables involving the interior location equilibrium case. Comparing \(\varPhi _{3}^{I }\) and \(\theta _{d}\), we find that \(\varPhi _{3}^{I} - \theta _{d} = (t/ \alpha ^{3})[(4 \alpha - 3)(8 \alpha - 3)] \ge 0\) if \(\alpha \ge (3/4)\). Moreover, \(\theta _{d}\) is positive if \(\alpha < (6/7)\). Thus, the interior location equilibrium holds and the two downstream firms survive when \(0 \le (c_{u}-c_{v}) < \theta _{d}\) and \((3/4) \le \alpha < (6/7)\).
Corner location equilibrium with a duopoly market condition for firm d
From (13.2), the corner location equilibrium occurs when \((c_{u}-c_{v}) \ge \theta _{d} \equiv (t/\alpha ^{2})(6 - 7\alpha )\) and \(\alpha < (6/7)\). Substituting \(x_{v}^{{D}^{*}}= 0\) and \(x_{d}^{{D}^{*}}= 1\) into the duopoly market condition \((\varPhi _{3})\), we have \((c_{u}-c_{v}) \le \varPhi _{3}^{\mathrm{C}}\equiv (t/\alpha ^{2})(9\alpha - 6)\) and \(\alpha > (6/9)\) due to \(\varPhi _{3}^{\mathrm{C} }> 0\), where superscript “C” denotes variables involving the corner location equilibrium case. Comparing \(\varPhi _{3}^{\mathrm{C}}\) and \(\theta _{d}\), we find that \(\varPhi _{3}^{\mathrm{C}} - \theta _{d}= (3t/\alpha ^{2})(5\alpha - 4) \ge 0\) if \(\alpha \ge (4/5)\). Therefore, the principle of maximum differentiation holds and two firms survive in the market when \(\theta _{d }\le (c_{u}-c_{v}) \le \varPhi _{3}^{\mathrm{C} }\) and (4/5) \(\le \alpha < (6/7)\).
Appendix 4: Equilibrium location under a monopoly market
From (15), we have that the equilibrium location of the integrated firm v is 1/2 for any \(\alpha \) when \(0 \le x_{v} \le 1/2\) and \(1 - \alpha \) for \(0 <\alpha < 1/2\) when \(1/2 < x_{v} \le 1\). It shows that the equilibrium location of firm v is 1/2 when \(1/2 \le \alpha < 1\). However, when \(0 <\alpha < 1/2\), we have to compare \(\pi _{v}^{{M}^{*}}(x_{v}^{{M}^{*}}= 1/2)\) with \(\pi _{v}^{{M}^{*}} (x_{v}^{{M}^{*}}= 1 - \alpha )\) so as to obtain the equilibrium location of firm v. Substituting (16.2) into (15.2), the profits of the monopolist firm v when \(x_{v}^{{M}^{*}}= 1/2\) as:
Substituting (16.1) into (15.1), the profits of the monopolist firm v when \(x_{v}^{{M}^{*}}=1 - \alpha \) as:
Comparing \(\pi _{v}^{M }(x_{v}^{{M}^{*}}=1/2)\) with \(\pi _{v}^{M }(x_{v}^{{M}^{*}} =1 - \alpha \)) when \(0 <\alpha < 1/2\), we have:
According to the above analysis, the equilibrium location of firm v is \(x_{v}^{{M}^{*}}= 1 - \alpha \) when \(0 <\alpha < 1/2\).