Uncorrected Proofs: Proof of Your Article Attached, Please Read Carefully

JOHN WILEY & SONS, LTD.
, THE ATRIUM, SOUTHERN GATE, CHICHESTER P019 8SQ, UK
*** PROOF OF YOUR ARTICLE ATTACHED, PLEASE READ CAREFULLY ***
After receipt of your corrections your article will be published initially within the online version of the journal.
PLEASE NOTE THAT THE PROMPT RETURN OF YOUR PROOF CORRECTIONS WILL
ENSURE THAT THERE ARE NO UNNECESSARY DELAYS IN THE PUBLICATION OF
YOUR ARTICLE
READ PROOFS CAREFULLY
ONCE PUBLISHED ONLINE OR IN PRINT IT IS NOT POSSIBLE TO MAKE ANY FURTHER

CORRECTIONS TO YOUR ARTICLE
This will be your only chance to correct your proof

Please note that the volume and page numbers shown on the proofs are for position only
FS
ANSWER ALL QUERIES ON PROOFS (Queries are attached as the last page of your proof.)
List all corrections and send back via e-mail to the production contact as detailed in the covering e-mail,
or mark all corrections directly on the proofs and send the scanned copy via e-mail. Please do not send
O
corrections by fax or post
O
CHECK FIGURES AND TABLES CAREFULLY
Check sizes, numbering, and orientation of figures
PR
All images in the PDF are downsampled (reduced to lower resolution and file size) to facilitate Internet
delivery. These images will appear at higher resolution and sharpness in the printed article
Review figure legends to ensure that they are complete
Check all tables. Review layout, titles, and footnotes
D
COMPLETE COPYRIGHT TRANSFER AGREEMENT (CTA) if you have not already signed one
Please send a scanned signed copy with your proofs by e-mail. Your article cannot be published
TE
unless we have received the signed CTA
OFFPRINTS
EC
25 complimentary offprints of your article will be dispatched on publication. Please ensure that the
correspondence address on your proofs is correct for dispatch of the offprints. If your delivery address
has changed, please inform the production contact for the journal – details in the covering e-mail.
Please allow six weeks for delivery.
R
Additional reprint and journal issue purchases

R
Should you wish to purchase a minimum of 100 copies of your article, please visit
http://www3.interscience.wiley.com/aboutus/contact_reprint_sales.html
O
To acquire the PDF file of your article or to purchase reprints in smaller quantities, please visit
http://www3.interscience.wiley.com/aboutus/ppv-articleselect.html. Restrictions apply to the use of
reprints and PDF files – if you have a specific query, please contact permreq@wiley.co.uk.
C
Corresponding authors are invited to inform their co-authors of the reprint options available
To purchase a copy of the issue in which your article appears, please contact cs-journals@wiley.co.uk
N
upon publication, quoting the article and volume/issue details

Please note that regardless of the form in which they are acquired, reprints should not be resold, nor
further disseminated in electronic or print form, nor deployed in part or in whole in any marketing,
U
promotional or educational contexts without authorization from Wiley. Permissions requests should be
directed to mailto: permreq@wiley.co.uk
SMJ697
Strategic Management Journal

Strat. Mgmt. J., 29: 000–000. (2008)
Published online in Wiley InterScience (www.interscience.wiley.com) DOI: 10.1002/smj.697
Received 16 August 2005; Final revision received 24 February 2008
A BARGAINING PERSPECTIVE ON STRATEGIC

OUTSOURCING AND SUPPLY COMPETITION
CATHERINE C. DE FONTENAY and JOSHUA S. GANS*
Melbourne Business School, University of Melbourne, Carlton, Victoria, Australia
FS
This article considers the outsourcing choice of a downstream firm with its own upstream
production resources or assets. The novelty of the approach is to consider the outsourced
function as involving resources consistent with the resource-based view of the firm. From a
bargaining perspective, we characterize a downstream firm’s decision whether to outsource to
O
an independent or to an established upstream firm. In so doing, the downstream firm faces a
trade-off between lower input costs afforded by independent competition, and higher resource
O
value associated with those who can consolidate upstream capabilities. We show that this trade-
off is resolved in favor of outsourcing to an established firm. Copyright  2008 John Wiley &
Sons, Ltd.
1
2
INTRODUCTION PR
This article takes this new agenda seriously and
applies it to a consideration of the outsourcing
27
28
D
3 Recent work in strategic management has advo- decision of the firm.2 Outsourcing is generally con- 29
4 cated the use of formal models of bargaining ceived of as a decision by the firm not to make a 30
TE
5 to understand the strategic impacts of firm deci- service/product internally, but to purchase it exter- 31
6 sions (Brandenburger and Stuart, 1996; Lippman nally. In doing so, however, it is often taken as a 32
7 and Rumelt, 2003). The idea is to consider the starting point that the function to be outsourced is 33
8 resources that a firm controls and to model how not yet being conducted, or, at the very least, is not 34
EC
9 those resources translate into bargaining power— being performed using an existing resource base, 35
10 as a means of modeling value appropriation.1 including physical assets and human capital, whose 36
11 expenditures (in investment or development) have 37
12 already been sunk.3 While considering outsourc- 38
R
13 Keywords: outsourcing; resource-based view; property ing as if no key resources already exist allows one 39
rights; bargaining; Shapley value
14 to highlight many of the advantages and disadvan- 40
R
*Correspondence to: Joshua S. Gans, Melbourne Business

15 School, University of Melbourne, 200 Leicester Street, Carlton, tages of outsourcing, it also may disguise other 41
16 Victoria 3053, Australia. E-mail: J.Gans@unimelb.edu.au important strategic issues. In particular, where such 42
O
1
17 Foss and Foss (2005) argue that this leads naturally to a 43
relationship between the resourced-based view (RBV) and the
18 economics of property rights (EPR); the latter being a way of 44
C
2
19 modeling control over resources or assets in a bargaining context. Other applications of the new agenda include Gans, McDonald, 45
They present a strong case for viewing resources as bundles and Ryall (2005); Adner and Zemsky (2006); and Ryall and
20 Sorenson (2007). 46
N
of property rights, property rights that can be reallocated in a

21 restructuring of a firm or organization. However, they emphasize 3
Cachon and Harker (2002) is a good representative example of 47
22 the importance of modeling transaction costs, something not this, where firms negotiate over outsourcing contracts that may 48
U
23 done in this article. As such, their approach would appear to or may not result in outsourcing and then all costs associated 49
be complementary to our own. with production are incurred.
24 50
25 51
26 52
Copyright  2008 John Wiley & Sons, Ltd.

2 C. C. De Fontenay and J. S. Gans
1 resources do exist, firms must consider the value of while maintaining branding and distribution. One 58
2 those existing resources that may become part of option was for GE to turn to an existing competi- 59
3 the bundle of property rights to be outsourced. tor—Matsushita—to perform its production. The 60
4 We take an alternative approach that is more nat- other—and the one GE ended up pursuing—was 61
5 urally consistent with a situation where resources to outsource to Samsung, who did not have a 62
6 do exist and firms must evaluate what prop- strong position in microwave oven production. GE 63
7 erty rights they should hold over them. It begins had been concerned about the risks associated with 64
8 with the status quo, whereby a firm has already dealing with a competitor and so chose an indepen- 65
9 invested in physical assets and resources that per- dent entity. Today, Samsung is the market leader.5 66
10 form certain functions and is considering outsourc- Motorola had a similar experience more recently 67
11 ing those functions. In this setting, outsourcing with the Taiwanese mobile phone maker BenQ. 68
12 represents an important strategic issue because it Motorola farmed out design and manufacturing 69
13 involves the transfer of the ownership of those to it only to find BenQ entering the large Chi- 70
FS
14 assets and resources to another firm. Moreover, nese market under its own brand.6 In both cases, 71
15 given the longer-term nature of such a divesti- these firms were criticized for allowing outsourc- 72
16 ture, this decision has a time frame longer than ing to provide a source of competition. Our anal- 73
17 the length of any particular outsourcing contract. ysis addresses these issues and finds the criticisms 74
O
18 While those contracts may be longer-term, as to have some foundation.7 In particular, managers 75
19 Domberger (1998) (among others) has pointed need to consider some important trade-offs when 76
O
20 out, outsourcing contracts—particularly, specifi- deciding how to outsource. 77
21 cation and pricing—tend to be renegotiated over The key trade-off is that a firm that outsources 78
PR
22 time. Consequently, in considering outsourcing as to an independent entity creates additional compe- 79
23 a divestiture, a firm must have regard to the amount tition for the supply of inputs to it. However, the 80
24 it will appropriate ex post. Of course, that appro- value of the productive assets to an independent 81
25 priation will determine the value that a purchaser firm is simply what they earn in competition. In 82
26 of those assets places upon them. Therefore, we contrast, when outsourcing to an established firm, 83
D
27 model a firm’s outsourcing decision taking into additional competition is not created, but there is 84
28 account both the value a firm expects to appro- implicit consolidation as more assets are controlled 85
TE
29 priate in an ongoing way and the value of the by a single firm. That consolidation is valuable 86
30 resources or assets under the control of another to the established firm—akin to the difference 87
31 owner. between monopoly and competitive profits—and 88
32 This approach allows us to consider an important so will increase the amount paid for the assets in 89
EC
33 strategic issue associated with outsourcing: whom the resulting acquisition.8 However, it also repre- 90
34 to outsource to. Some major firms have outsourced sents a longer-term issue for the outsourcing firm 91
35 by spinning off separate, independent corporations who will likely pay more to procure inputs. 92
36 that supply them with outsourced functions, but 93
R
In this article, we show how a proper consid-

37 they are also free to access the market for those eration of a bargaining approach to value appro- 94
38 services. This is seen as a way of bringing compet- 95
R
priation can resolve the trade-off between asset

39 itive pressure to bear on the efficiency with which value and competitive input supply terms. In doing 96
40 those functions are performed. Other firms wish to 97
O
41 tap into the competencies of established operators 98

5
42 and outsource to them. To date, however, we are For an account of this see Jarillo (1993). 99
C
6
43 not aware of any study that considers this decision: ‘Outsourcing Innovation,’ BusinessWeek, 21 March, 2005: 53. 100
7
44 whom to outsource to?4 Contrast these choices with Apple, who in recent times has 101
N
favored established firms for component manufacturing; los-

45 Consider the case of GE and the production of ing ongoing advantages from competition. In 2005, it famously 102
46 microwave ovens. Having entered that market in moved its processing chips in its computer products to an Intel 103
U
47 the 1970s, GE faced competitive pressure and a platform rather than to the weaker manufacturers of Motorola, 104
IBM, and AMD. Of course in this case, Apple was not consid-
48 reduction of market share. In 1983, it considered ering a spin-off, as was the case with GE and Motorola, but it 105
49 outsourcing microwave oven production entirely did make a decision that impacted competition in its suppliers’ 106
50 markets. 107
8
It is typically the case that the sum of profits of two competitors
51 4
For a review of recent work in strategy on outsourcing see working together exceeds the sum of them in competition with 108
52 Espino-Rodriguez and Padron-Robaina (2006). one another (see Gilbert and Newbery, 1982). 109
53 110
Copyright  2008 John Wiley & Sons, Ltd. Strat. Mgmt. J., 29: 0–0 (2008)
54 DOI: 10.1002/smj 111
55 112
56 113
57 114
Strategic Outsourcing 3
1 this, we consider a situation where bargaining dic- be used to quantitatively assess the factors driving 58
2 tates the supply terms arising between upstream these decisions. 59
3 and downstream firms. Following de Fontenay While the research question of this article is 60
4 and Gans (2005), the natural bargaining solution novel, the use of this type of bargaining envi- 61
5 for this is the Shapley value. They demonstrate ronment to describe strategic integration or out- 62
6 that when upstream and downstream firms bargain sourcing has important antecedents in the litera- 63
7 bilaterally in a noncooperative fashion, outcomes ture. Inderst and Wey (2003) and de Fontenay and 64
8 are value maximizing and each firm appropriates Gans (2005) provide models of bilateral oligopoly 65
9 its Shapley value as would arise in the underly- (again with two upstream and two downstream 66
10 ing coalitional game. Such bargaining naturally firms) to analyze horizontal and vertical integra- 67
11 describes situations when supply agreements are tion respectively. Each of these studies, however, 68
12 formed from face-to-face negotiations rather than concentrates on the social welfare implications of 69
13 anonymous wholesale markets. It is for this reason alternative market structures and does not consider 70
FS
14 that we will not model the noncooperative game the decision of a firm currently holding upstream 71
15 here and simply consider its bargaining analog.9 assets, to outsource. Here, our focus is on the 72
16 We use this approach to consider the set of nego- purely strategic aspects of a private outsourcing 73
17 tiated supply terms that would arise under alterna- decision when outsourcing changes asset owner- 74
O
18 tive ownership structures. As noted, there is no ship and bargaining terms. 75
19 double marginalization problem and, indeed, sup- There is also literature on strategic outsourc- 76
O
20 ply chain management is efficient. Hence, whether ing that deals with related considerations. Baye, 77
21 functions are outsourced or not and whom they Crocker, and Ju (1996) do not consider outsourc- 78
PR
22 are outsourced to does not matter for the overall ing per se, but rather a conglomerate’s decision 79
23 production of value. to operate autonomous or semiautonomous divi- 80
24 In this situation, we show that a firm will want sions. They demonstrate that such divisionaliza- 81
25 to outsource to an established firm—creating a tion, while costly for individual firms, is driven 82
26 monopoly—rather than to an independent player. by competition with other conglomerates as firms 83
D
27 While a monopoly does raise input costs for compete to establish stronger downstream market 84
28 the firm, those higher costs are partly borne by positions. Baye et al. (1996) consider the costs of 85
TE
29 other downstream firms. So when we consider outsourcing and how increased upstream competi- 86
30 tion might motivate it regardless. Such incentives 87
the value of the asset to an established upstream
31 for outsourcing are also present in our model; how- 88
firm relative to an independent one, it is suffi-
32 ever, this question is not the primary focus of our 89
EC
ciently greater in that it compensates the outsourc-

33 study.10 90
ing downstream firm for those higher input costs.
34 In contrast, McGuire and Staelin (1983) con- 91
This baseline result does, however, suggest that GE
35 sider outsourcing a retail function—to a common 92
and Motorola might have erred in their decisions
36 retailer—as a means by which competing suppliers 93
R
previously discussed. Thus, the majority of the arti-

37 might mitigate price competition between them.11 94
cle is devoted to enriching our baseline model to
38 Unlike our model below, they utilize a linear pric- 95
R
consider when independent outsourcing might be a ing model and so only find this to be optimal—in
39 more appropriate strategic choice than established 96
40 terms of total industry profits—if product differen- 97
O
firm outsourcing. Of course, as individual cases tiation is sufficiently low. Our study starts from the
41 may have other factors in play, we do not propose 98
42 that our model decisively explains or renders inap- 99
C
10
43 propriate the decisions taken by these firms. Our
Another set of papers considers horizontal subcontracting 100
between rival firms (see, e.g., Spiegel, 1993; Nickerson and
44 goal here is to inform managers of the trade-offs 101
N
Vanden Bergh, 1999; and Shy and Stenbacka, 2003). The focus
45 they face and the modeling techniques that can of these studies is the ability of subcontracting between rivals to 102
46 soften competition between them. For the most part our study 103
U
does not consider the impact of outsourcing on downstream

47 competition (that is the focus of de Fontenay and Gans, 2005).
104
48 9
Use of the Shapley value for this type of bargaining was advo- However, we do discuss how these considerations relate to our 105
49 cated by Lippman and Rumelt (2003) and has been employed in primary research question in the Robustness section herein. 106
the analysis of firm boundaries in the EPR approach by Hart and 11
50 McGuire and Staelin (1983) do not examine explicitly the deci- 107
Moore (1990). While we do not model it here, similar results sion of whom to outsource to. However, their broader analysis of
51 would arise if the coalitional approach of MacDonald and Ryall industry structure considers whether firms might jointly operate 108
52 (2004) and Brandenburger and Stuart (2007) were used. through a single retailer or through private franchisees. 109
53 110
54 DOI: 10.1002/smj 111
55 112
56 113
57 114
1 perspective of a supply rather than a demand out- are located in separate geographic markets. Thus, 58
2 sourcing decision, but establishes a related result downstream assets compete for inputs from up- 59
3 under bargaining that applies regardless of the stream firms and upstream firms compete to sup- 60
4 degree of product differentiation. We demonstrate, ply them, but downstream firms do not compete 61
5 in that case, that outsourcing to an established directly with one another. Clearly, this assump- 62
6 firm can also soften competition; something that is tion does not correspond to some of our motivating 63
7 costly, but to the outsourcing firm may be worth- examples. As such, having built intuition and anal- 64
8 while because it can appropriate sufficient value ysis based on an independent final good market 65
9 for the outsourced assets. As described below, our assumption, we will (in the Robustness section) 66
10 stronger conclusion rests on the fact that we con- amend our analysis to include head-to-head com- 67
11 sider asset market transactions that are bilateral petition between downstream firms. 68
12 (rather than cooperative) involving externalities on It is assumed that inputs produced from either 69
13 other market participants. upstream assets can be used by downstream firms 70
FS
14 The article proceeds as follows. The next section in producing final goods for consumers. Our anal- 71
15 sets up our 2x2 model and discusses our approach ysis here does not specify whether the inputs are 72
16 to outsourcing. The restriction to two upstream and perfectly substitutable or not, or whether one set 73
17 two downstream firms is for analytical ease and of inputs is more specialized to one downstream 74
O
18 expanding those numbers would leave our main asset or not. The model can flexibly accommo- 75
19 qualitative results unchanged. Next, we derive date all these possibilities with homogeneous and 76
O
20 results for the bargaining approach and construct nonhomogeneous products alike. What is key is 77
21 baseline results; followed by our examination of that each downstream firm can potentially pro- 78
PR
22 the robustness of our baseline results to changes in cure inputs from either of the upstream firms or 79
23 the model, including allowing for internal bargain- from both simultaneously. Thus, we need impose 80
24 ing, competition among downstream firms, compe- no artificial constraints on who trades with whom. 81
25 tition with integrated entities, and double marginal- As a baseline case, we also allow for asymme- 82
26 ization. We then explain independent outsourcing tries between firms at the same functional level. So, 83
D
27 and conclude with a final section. say, UA may have substantially lower costs than UB 84
28 or D1 may have a significantly larger market than 85
TE
29 D2 . However, to obtain cleaner results, we may at 86

30 MODEL SET UP times assume that functional assets are symmet- 87
31 ric with the same costs, capabilities, and market 88
32 Our approach is to consider an industry with dis- potential. 89
EC
33 tinct upstream and downstream functions. The sim- Each independent firm bargains with each firm 90
34 plest example would be that of manufacturing ver- on the other side of the market. Thus, we do 91
35 sus retailing (or distribution). For either to perform not allow independently owned upstream or down- 92
36 productively requires resources. The downstream stream firms to collude. However, integrated firms 93
R
37 resources or assets may be retail outlets and distri- that own one or more assets simply set internal 94
38 bution networks, while the upstream resources or supply conditions efficiently. 95
R
39 assets may be production facilities. Denote by qij the quantity of inputs sold by Uj to 96
40 While our analysis here could apply to func- Di . Payments from downstream firm i to upstream 97
O
41 tions that require many assets (physical, human, firm j are a function, pij (qij ). We consider a 98
42 and other) and permit many collections of assets general form where supply contracts are a pair 99
C
43 in the industry (corresponding to arbitrary num- (pij , qij ); that is, a lump sum payment of pij 100
44 bers of upstream and downstream firms), for sim- in return for the quantity, qij . Notice that this 101
N
45 plicity, we focus on the case where each func- is a distinct assumption from the literature on 102
46 tion requires a single asset and there are only vertical relationships in the absence of bargaining. 103
U
47 two such assets available for each function. Thus, In those models, upstream firms set simple per- 104
48 there are two downstream assets—denoted Di for unit (or linear) prices to downstream firms. This 105
49 i ∈ {1, 2}—and two upstream assets—denoted Uj results in familiar issues of double marginalization. 106
50 for j ∈ {A, B}. In contrast, when firms negotiate over price and 107
51 Another simplifying assumption that we make quantity pairs, the resulting supply contracts are 108
52 initially is to assume that the downstream assets efficient in the sense that they avoid the double 109
53 110
54 DOI: 10.1002/smj 111
55 112
56 113
57 114
1 marginalization distortion. This is an appropriate owner; or (ii) the owner of the other upstream 58
2 outcome given that we allow firms to engage in asset, UB . 59
3 direct negotiations over supply terms rather than DATE 1 (Price formation/bargaining): Each down- 60
4 anonymous wholesale market trading. stream firm engages in bargaining with each 61
5 By avoiding double marginalization, we remove upstream firm over the procurement of inputs. 62
6 a cost of outsourcing; that is, relative to ver- If a downstream firm owns an upstream asset, 63
7 tical integration, outsourcing creates the double it need not pay for or bargain over that asset’s 64
8 marginalization distortion and reduces profits. input supply. Bargaining outcomes are based on 65
9 Instead, to consider the nonbargaining benefits to the Shapley value. 66
10 outsourcing, we adopt a ‘black box’ approach to DATE 2 (Production): Production takes place, 67
11 the net benefits of outsourcing. We assume that if downstream output is sold, and all payments are 68
12 an upstream asset, Uj , is not owned by a down- made. 69
13 stream firm there is an additional benefit, j , for 70
FS
14 that asset that is earned by its owner. This ben- This timeline reflects the basic premise of this 71
15 efit may arise from the additional competencies study—that the determination of asset ownership 72
16 of independent ownership or from some private is a longer-term decision than the determination 73
17 or other benefits of control of that asset.12 We of input prices. Outsourcing contracts are rarely 74
O
18 will assume, however, that the benefits are only set in stone and have many elements that are 75
19 realized if the asset is utilized in production. This renegotiated over time. In contrast, it is more 76
O
20 captures the notion that the resources of the firm difficult to move assets in and out of a firm. As 77
21 are more valuable being productively utilized by asset ownership is more ‘sticky’ as a decision 78
PR
22 an outsourcing partner. It will turn out that, in it has the greater strategic importance. As such, 79
23 general, a downstream firm will suffer a pricing outsourcing decisions (at Date 0) will have regard 80
24 disadvantage if it outsources production. In this to the expected prices that will be negotiated ex 81
25 case, the benefit of placing resources elsewhere, post (at Date 1) as well as the value of the assets 82
26 j , will compensate for it. Of course, if j < 0 to the outsourcing firm. Given its importance, we 83
D
27 it may not be profitable to outsource at all.13 The comment in detail on the relationship between 84
28 main purpose of this assumption is to focus not on outsourcing and asset ownership. 85
TE
29 the decision of whether to outsource or not, but on 86

30 the strategic decision of whom to outsource to. Integration and asset ownership 87
31 88
32 D1 currently produces a service in-house, requiring 89
EC
33 Timeline the use of one upstream asset—assumed here to 90

34 be UA . Notice that, just because D1 provides this 91
We begin with the status quo where D1 owns an
35 service, it is not precluded from providing that 92
upstream asset, UA . The other downstream firm
36 service to others (namely, D2 ). As D1 and D2 93
R
(D2 ) may or may not own the other upstream asset.

37 do not directly compete, such provision can, at 94
We explain what asset ownership means below.
38 times, be efficient.14 Nonetheless, the decisions 95
R
The timeline for our model is as follows:

39 over inputs supplied from UA , (q1A , q2A ), lie with 96
40 D1 . Thus, D2 will have to negotiate with D1 97
O
DATE 0 (Outsourcing): D1 decides whether or over input supply; a negotiation that will take into
41 98
not to outsource. Outsourcing means selling account D1 ’s expected use of that asset. In this
42 99
C
its upstream asset to either (i) an independent case, D1 and D2 will still, of course, negotiate
43 100
44 with UB . Figure 1 depicts the graph of bargaining 101
N
45 12
These might be incentives for independent owners to under- relationships that corresponds to this status quo 102
46 take noncontractible actions; something that would not occur if (the black lines representing lines of negotiation 103
U
these managers were employees of a larger, vertically integrated rather than flows of inputs per se).
47 structure (Grossman and Hart, 1986). It may also account for a 104
48 reduction in transaction costs (Foss and Foss, 2005). If D1 sells the asset, it could choose to sell it 105
49 13
Here j is a fixed benefit. The results of this study will be to an independent owner (let’s call that firm by its 106
50 unchanged if j is a marginal benefit (depending on quantity 107
produced by the upstream firm) or a specialized benefit impact-
51 ing favorably on trade with one downstream firm more than 14
Indeed, it can be desired even if D1 and D2 do directly compete 108
52 another. (de Fontenay and Gans, 2005). 109
53 110
54 DOI: 10.1002/smj 111
55 112
56 113
57 114
Finally, we consider situations where D2 oper- 58

Color Figure - Online only
UA UB
ates UB in-house. This means that when D1 con- 59
siders outsourcing to an established firm, it will 60
give D2 control over an upstream monopoly (see 61
Figure 3). Moreover, because it is integrated, that 62
D1 D2 decision will not create the additional benefits (A 63
and B ). 64
Figure 1. Status quo. This figure is available in color In each case, there will be a potential gain from 65
online at www.interscience.wiley.com/journal/smj trade for D1 and a respective potential purchaser of 66
its upstream asset (UA ). Below we model the asset 67
acquisition game as a (discriminatory) auction and 68
1 associated asset, UA ).15 D1 and D2 will then nego- evaluate the choices that D1 makes in this context. 69
2 tiate with UA on equal terms; specifically, UA will 70
FS
3 know the price and supply it expects to make to 71
Coalitional notation
4 D2 when dealing with D1 . This corresponds to the 72
5 left-hand side of Figure 2. However, each down- As noted, the bargaining solution that we employ 73
6 stream firm will also consider its position with UB is the Shapley value. In its direct application or 74
O
7 when agreeing to supply terms. As will be seen, as derived from more primitive game-theoretic 75
8 UA and UB will compete with one another in this assumptions (de Fontenay and Gans, 2006), the 76
O
9 regard. From D1 ’s perspective, selling the asset Shapley value says that each firm receives a weight 77
10 to an independent owner generates a competitive sum of coalitional values as its profits. For that 78
reason, it is useful to define a notation for those
PR
11 value for the asset but also lower expected costs for 79
12 future input supplies driven by additional upstream coalitional values. 80
13 competition. A coalition in a strategic setting, such as that 81
14 In contrast, if D1 sells the asset to UB , the employed here, is a collection of agents who are 82
15 two upstream assets are commonly owned. In this able to trade with one another. The coalitional 83
D
16 situation, both D1 and D2 will negotiate with a value is the maximum surplus they can achieve 84
17 single entity for all of their input supply. Not through that trade. However, since the real prim- 85
TE
18 surprisingly, this gives the upstream firm a stronger itives in determining surplus are the resources or 86
19 assets involved in the trade, we will denote coali- 87
position. This strength will be reflected in its asset
20 tional values based on those assets as a notational 88
value, but the higher expected input prices will be
21 convention and trade the net benefits to outsourc- 89
something D1 takes into account in its outsourcing
EC
22 ing (should it occur) as an additional variable. 90

decision. Figure 2 depicts the outsourcing choice
23 For example, the grand coalition where all 91
for D1 .
24 agents and their assets participate in trade can gen- 92
25 erate value denoted by (D1 D2 UA UB ). (D1 D2 93
R
26 UA UB ), which is literally the maximal level of 94

UA UB UA-UB
27 industry profits that can be generated by efficiently 95
R
28 employing both upstream and both downstream 96

or ? assets. For example, if πi (qiA , qiB ) are Di ’s prof-
29 97
O
30 its net of payments for inputs and Cj (q1j , q2j ) are 98

31 D1 D2 D1 D2 99
C
32 100
Color Figure - Online only
UA UB UA-UB
33 Figure 2. D1 ’s outsourcing choice 101
N
34 102
35 or ?
103
U
15
36 Note that there is a difference between an independent out-
sourced firm and an upstream market entrant. An entrant, while
104
37 independent, also brings resources into the industry. An inde- D1 D2 D1 D2 105
38 pendent outsourced firm does not and, in this article, acquires 106
39 ownership over existing resources. Market entry could be accom- 107
modated in this framework but it would add complexity in Figure 3. D1 ’s outsourcing decision (D2 integrated).
40 attempting to account for the additional resources and separating This figure is available in color online at www. 108
41 that out from bargaining effects. interscience.wiley.com/journal/smj 109
42 110
43 DOI: 10.1002/smj 111
44 112
45 113
46 114
47
48
49
1 Uj ’s costs of input supply (where Cj (.) is assumed all i; (D1 D2 UA UB ) ≥ (D1 D2 Uj ) for all j , 58
2 to be convex),16 then: and (D1 D2 Uj ) ≥ (Di Uj ) and (Di UA UB ) ≥ 59
3 (Di Uj ) for all i and j . Second, if an asset at 60
4 (D1 D2 UA UB ) ≡ max{qij } i=1,2 π1 (q1A , q1B ) the same functional level can be removed with- 61
j =A,B
5 out reducing coalitional value (that is, the previous 62
6 + π2 (q2A , q2B ) − CA (q1A , q2A ) inequalities hold with equality), then that asset and 63
7 − CB (q1B , q2B ) (1) its functional neighbor are perfectly substitutable. 64
8 For instance, this might arise for upstream assets 65
9 Notice that this value can be achieved regard- that have unlimited capacity and the same costs. 66
10 less of whether D1 and UA are integrated, or 67
11 if UA is outsourced and owned by UB , or is 68
Ownership notation
12 independently operated. However, in the latter 69
13 two cases, the grand coalition surplus is, in fact, We rely heavily on a notation for the payoffs each 70
FS
14 firm receives under various ownership structures. 71
(D1 D2 UA UB ) + A + B , while in the former
15 In this regard, we grouped commonly owned assets 72
case it is (D1 D2 UA UB ) + B .
16 The bargaining outcome also depends upon together. For instance, say that D1 owned UA 73
17 but D2 and UB were independent. This would be 74
O
value that would arise should agreements between
18 described by the state (D1 UA , D2 , UB ). If D1 sold 75
firms break down, resulting in some firms being
19 UA to an independent owner the state would be 76
excluded from selling or being sold to by others.
O
20 (D1 , UA , D2 , UB ), while if those assets were sold 77
In this situation, alternative ‘coalitions’ or sup-
21 to UB the state would be (D1 , D2 , UA UB ). 78
ply structures form. These might involve a single
PR
22 Given this, the expected payoff to firm i, is 79
upstream asset supplying both downstream assets
23 given by vi (.). So if the ownership structure was 80
(monopoly):
24 (D1 UA , D2 , UB ), then D2 ’s expected payoff would 81
25 be vD2 (D1 UA , D2 , UB ), while the integrated D1 and 82
(D1 D2 Uj ) ≡ max{qij }i=1,2 π1 (q1j )
26 UA ’s payoff would be vD1 UA (D1 UA , D2 , UB ). 83
D
27 + π2 (q2j ) − Cj (q1j , q2j ) (2) 84
28 85
TE
29 a single downstream asset supplied by both up- BASELINE RESULTS 86

30 stream assets (monopsony): 87
31 We now derive our results regarding the strate- 88
32 (Di UA UB ) ≡ max{qij }j =A,B gic outsourcing decision of D1 . We begin with the 89
EC
33 status quo whereby D1 performs upstream oper- 90

πi (qiA , qiB ) − CA (qiA ) − CB (qiB ) (3)
34 ations itself, owning UA . In deciding whether to 91
35 outsource and to whom, D1 will compare its profits 92
or alternatively a situation where there is only a
36 from the status quo with its profits and the profits 93
R
single upstream and downstream asset (bilateral

37 of other firms in various scenarios: (a) outsourcing 94
monopoly):
38 to an independent firm and (b) outsourcing to an 95
R
39 established firm. In this regard, we initially assume 96

(Di Uj ) ≡ maxqij πi (qij ) − Cj (qij ) (4)
40 that D2 and UB are independently owned firms; we 97
O
41 explore the implications of this assumption below. 98

It is worthwhile noting some important proper-
42 99
C
ties of these coalitional values. First, adding assets

43 100
into the potential trade pool increases coalitional The outsourcing decision
44 101
value. That is, (D1 D2 UA UB ) ≥ (Di UA UB ) for
N
45 The focus of this article is how bargaining con- 102

46 siderations impact outsourcing and, in particular, 103
U
16
47 Note that this assumption rules out scale economies. This how outsourcing is achieved. We begin, therefore, 104
assumption is made for simplicity only and to avoid some tech-
48 nical issues associated with non-interior solutions (de Fontenay by considering D1 ’s decision to outsource in the 105
49 and Gans, 2005). All of the conclusions of the model would first place and what might drive it. Thus, we will 106
50 continue to hold if upstream production technologies exhibited assume here that D1 does not have a choice as 107
scale economies. This would mean that it was optimal to have
51 a single upstream supplier—thereby adding to the bargaining to whom to outsource to. We will endogenize that 108
52 reasons as we derive below. choice below. 109
53 110
54 DOI: 10.1002/smj 111
55 112
56 113
57 114
1 To consider D1 ’s outsourcing decision, we con- Proposition 1: Regardless of whether it out- 30

2 sider what payoffs each firm gets in the status quo sources to an independent or to an established 31
3 (as represented by Figure 1). If D1 owns and confirm, D1 ’s ongoing costs of input supply will 32
4 trols the assets of UA , in negotiations with UB , D1 only fall if A is sufficiently large. 33
5 is assured of supply from UA while in negotiations 34
6 The proofs of all propositions are in the Appendix. 35
with D2 , D1 can effectively exclude the use of UA .
7 This result says that just because A is positive 36
This impacts the possible outcomes that can arise
8 (which means that outsourcing is efficient from an 37
if a full set of supply agreements is not reached: industry perspective) does not mean that a deci- 38
9
10 namely, there cannot be a situation where D2 is sion to outsource will result in those costs being 39
11 supplied by UA without D1 also being supplied. passed through to D1 . Instead, UA incurs some 40
12 Similarly, there cannot be a situation where UB of these costs in negotiations with D1 . But, more 41
13 supplies D1 without UA also supplying D1 . Thus, interestingly, D2 is a beneficiary of those gains 42
FS
14 D1 is never at risk of facing an upstream monopo- (see Table 1). This is because when D1 outsources 43
15 list that it does not own. Table 1 lists the (Shapley (whether to an independent or an established firm), 44
16 value) payoffs to the three firms in this scenario. D2 now negotiates with that firm and so can cap- 45
17 Outsourcing means two things. First, the out- ture some of the productive benefits flowing from 46
O
18 sourced firm can realize the benefit A . Second, D1 ’s outsourcing decision. Consequently, D1 does 47
19 D1 is no longer assured of supply in certain cir- not fully internalize the benefits of outsourcing in 48
O
20 terms of ongoing reductions in its supply costs. 49
cumstances; namely, if bargaining between it and
21 That said, D1 ’s outsourcing decision does not 50
a supplier breaks down, it may face a monopoly or
PR
22 purely rest on its ongoing returns (that is, vD1 51
23 no supply at all. This latter effect may mean that (D1 , D2 , UA , UB ) − vD1 UA (D1 UA , D2 , UB )). 52
24 D1 is in a diminished bargaining position and may Instead, another firm (independent or established 53
25 be facing higher ongoing costs. as the case may be) benefits from the acquisition 54
26 Given this, we can state the following proposi- of the assets, UA , and from the ongoing abil- 55
D
27 tion: ity this gives it to earn rents through bargaining. 56
28 57
TE
29 Table 1. Shapley value outcomes 58
Ownership structure Payoffs (Shapley values)

EC
UB independently held
4(D1 D2 UA UB ) + 2(D1 D2 UA )
D1 owns UA v D 1 UA = 1 + 1 B ,
12 +2(D1 UA UB ) + 4(D1 UA ) − 4(D2 UB) 6
vD2 = 12 1 4(D1 D2 UA UB ) + 2(D1 D2 UA ) + 61 B
−4(D1 UA UB ) − 2(D1 UA ) + 2(D2 UB )
R
vUB = 12 1 4(D1 D2 UA UB ) − 4(D1 D2 UA ) + 32 B

+2(D1 UA UB ) − 2(D1 UA ) + 2(D2 UB )

R
3(D1 D2 UA UB ) + (D1 D2 UA ) + (D1 D2 UB )

Complete nonintegration vD1 = 12 +(D1 UA ) + (D1 UB ) − (D2 UA ) − (D2 UB ) + 61 (A + B )
1
+(D1 UA UB ) − 3(D2 UA UB )
O
3(D1 D2 UA UB ) + (D1 D2 UA ) + (D1 D2 UB )

vD2 = 12 −(D1 UA ) − (D1 UB ) + (D2 UA ) + (D2 UB ) + 61 (A + B )
1
+(D2 UA UB ) − 3(D1 UA UB )
C
3(D1 D2 UA UB ) + (D1 D2 UA ) − 3(D1 D2 UB )

v UA = 1 + 2 A
12 +(D1 UA ) − (D1 UB ) + (D2 UA ) − (D2 UB ) 3
N
+(D U U
1 A B ) + (D U U
2 A B )
3(D1 D2 UA UB ) + (D1 D2 UB ) − 3(D1 D2 UA )
vUB = 12 1 −(D U ) + (D U ) − (D U ) + (D U ) + 2
U
1 A 1 B 2 A 2 B
3 B
+(D1 UA UB ) + (D2 UA UB )

Nonintegrated upstream monopoly vD1 = 61 2(D1 D2 UA UB ) + (D1 UA UB ) − 2(D2 UA UB ) + 61 (A + B )

vD2 = 61 2(D1 D2 UA UB ) − 2(D1 UA UB ) + (D2 UA UB ) + 61 (A + B )

vUA UB = 61 2(D1 D2 UA UB ) + (D1 UA UB ) + (D2 UA UB ) + 32 (A + B )
DOI: 10.1002/smj
1 If an independent firm purchases these assets, it acquirer outsourcing may not occur. This result is 58
2 receives vUA (D1 , D2 , UA , UB ). If there were many not surprising given our earlier observation that 59
3 such independent firms, each one would be will- some of the efficiency benefits from outsourcing 60
4 ing to pay this full value. Thus, D1 , by outsourcing, by D1 will be passed along to D2 through bar- 61
5 could earn: gaining. 62
6 The second part of the proposition looks at 63
7 vUA (D1 , D2 , UA , UB ) the converse. When D1 outsources to an inde- 64

8 Asset Acquisition Payment pendent firm, unless A > 0, it will not find 65
9 this outsourcing profitable. The same cannot be 66
10 (vD1 UA (D1 UA , D2 , UB ) 67
− (5) said of outsourcing to an established firm. There
11 −vD1 (D1 , D2 , UA , UB )) it is possible that outsourcing could be prof- 68

12 Ongoing cost increase itable even if it is productively inefficient (that is 69
13 A < 0). The reasons for this will become clear 70
FS
14 That is, to choose outsourcing, D1 ’s ongoing cost when we compare the two types of outsourcing 71
15 increase (if any) must be weighed against the pay- below. 72
16 ment it could receive for its outsourced division. For the moment, it is interesting to note the 73
17 Similarly, were D1 to outsource to an established empirical implications of this. First, in examining 74
O
18 firm, its total return could be as high as: outsourcing by spin-off, if this is observed it is 75
19 efficient. Thus, measures of industry productivity 76
O
20 vUA UB (D1 , D2 , UA UB ) − vUB (D1 UA , D2 , UB ) will increase as a result of observed outsourcing. 77

21 Asset Acquisition Payment Second, the same cannot be said of outsourcing 78
PR
22 to other established firms. In this case, industry 79
23 (vD1 UA (D1 UA , D2 , UB ) 80
− (6) productivity might well decline as a result. How-
24 −vD1 (D1 , D2 , UA UB )) ever, if the assets, UA , held little market value, 81

25 Ongoing cost increase and D1 procured from an established firm, this 82
26 decline would only occur if A were sufficiently 83
D
27 Notice that, in this case, the maximum payment large (Proposition 1). In this case, outsourcing 84
28 that UB could make for the assets, UA , would be would be associated with rises in industry produc- 85
TE
29 its ongoing profits post-acquisition less its prof- tivity. 86

30 its pre-acquisition. We say maximum because UB Finally, it is instructive to examine how the 87
31 may not have to pay that much to ensure it being overall scarcity of supply might impact on the 88
32 the outsourced supplier to D1 to still be attractive 89
EC
returns to outsourcing. Supply scarcity could be

33 relative to the status quo. The actual asset acquisi- modeled by consideration of the convexity of 90
34 tion payment would be the subject of a negotiation upstream costs. If these were not convex and 91
35 between D1 and UB . firms had unlimited capacity, then, for example, 92
36 When we take into account the full poten- (D1 D2 UA UB ) = (D1 D2 Uj ) 93
R
and
37 tial benefits of outsourcing (acquisition price plus (Di UA UB ) = (Di Uj ); i.e., adding an upstream 94
38 ongoing impacts), we find the following: 95
R
asset would add little to industry profits. On the

39 other hand, with highly convex costs and limited 96
40 Proposition 2: A > 0 is not a sufficient con- upstream capacity, (D1 D2 UA UB ) = (Di UA UB ) 97
O
41 dition for outsourcing to be profitable for D1 . and (D1 D2 Uj ) = (Di Uj ); that is, additional 98
42 A > 0 is a necessary condition for D1 to find demand could not be fulfilled so that additional 99
C
43 it profitable to outsource to an independent firm. downstream markets would add little to industry 100
44 profits. 101
N
45 The first part of the proposition says that just In our general set up we prove the following: 102
46 because outsourcing is efficient does not mean 103
U
47 that D1 will find it profitable. Given Proposi- 104

48 tion 1, clearly this will be the case if D1 can- Proposition 3: The returns to outsourcing to 105
49 not appropriate UA ’s full asset value in spinning an established firm fall as upstream capacity 106
50 off or selling the division. However, the proposi- becomes scarce. The returns to outsourcing to 107
51 tion makes no assumption as to that, and so even an independent firm rise as upstream capacity 108
52 when D1 can appropriate UA ’s full value to its becomes scarce. 109
53 110
54 DOI: 10.1002/smj 111
55 112
56 113
57 114
1 Thus, the scarcity of upstream capacity has what each firm will bid and how this impacts the 58
2 markedly different impacts on the returns to estab- seller.17 59
3 lished and independent firm outsourcing. Prior To begin, note that from D1 ’s perspective, it 60
4 to outsourcing, D1 has the option of procuring earns more net of any payments for assets by 61
5 internally or externally. It can negotiate favor- selling to an independent firm. That is, 62
6 able external returns when it has its own inter- 63
7 nal supply. This option is more valuable when vD1 (D1 , D2 , UA , UB ) − vD1 64
8 capacity is scarce. However, when outsourced 1 65
9 to an established firm, that capacity is concen- × (D1 , D2 , UA UB ) = 66
12
10 trated in the hands of a single firm that even   67
11 (D D U U )
1 2 A B
68
in the absence of such consolidation would earn  +(D1 UA UB ) − (D2 UA UB ) 
12 scarcity rents. Hence, the value of outsourcing   69
 +(D1 D2 UA ) + (D1 D2 UB )  > 0 (7)
13 is reduced. In contrast, an independent firm will   70
+(D1 UA ) + (D1 UB )
FS
14 value a scarce upstream resource highly relative 71
15 −(D2 UA ) − (D2 UB ) 72
to not owning any upstream assets at all. Thus,
16 it will increase its bid for those assets accord- where the inequality follows as (D1 D2 UA UB ) ≥ 73
17 ingly. It is the fact that the value of being able (D2 UA UB ) and (D1 D2 Uj ) ≥ (D2 Uj ) for each 74
O
18 to earn scarcity rents is greater for independent j . Thus, D1 will earn lower ongoing profits by 75
19 than established operators that changes their rela- selling its asset to UB . For this reason, D1 will 76
O
20 tive returns. not simply want to accept the highest bid in this 77
21 auction. Instead, it will expect UB to bid a certain 78
PR
22 amount above the bids of an independent purchaser 79
23 in order to secure the asset. 80
24 Independent versus established firm 81
outsourcing In this situation, the maximum amount an inde-
25 pendent firm would bid for the assets is vUA 82
26 (D1 , D2 , UA , UB ). Thus, in order to beat an inde- 83
We now consider D1 ’s decision choices if it were
D
27 pendent bid for the assets, and assuming it has 84
to outsource: would it outsource to an indepen-
28 complete information regarding these values, UB 85
dent firm (say, by spinning off its division) or
TE
29 would have to bid an amount, b, so that: 86

to an established firm (by selling off its upstream
30 87
division). There are two forces at work in consider-
31 b + vD1 (D1 , D2 , UA UB ) 88
32 ing this decision. First, by creating an independent 89
EC
upstream supplier, competition is produced that D1 s Returns from Established Outsourcing

33 90
34 lowers the cost of procuring inputs. Second, the vUA (D1 , D2 , UA , UB ) 91
value of upstream assets is diminished relative to a > (8)
35 +vD1 (D1 , D2 , UA , UB ) 92
36 situation whereby they are consolidated in a single 93
R
upstream entity. D1 s Returns from Independent Outsourcing

37 94
38 It turns out that the latter (asset value) effect 95
R
Will UB bid this amount? If it does, knowing that

39 dominates the former (ongoing cost) effect. To 96
the alternative is independent outsourcing, its bid
40 demonstrate this, we suppose that D1 puts its UA 97
O
must satisfy:
41 asset up for auction. To model this, we assume 98
42 that there is a continuum of independent own- vUA UB (D1 , D2 , UA UB ) − b 99
C
43 ers who might purchase the asset (so that each 100
UB s Returns from Winning
44 bids up to their willingness to pay) while there 101
N
45 is a single owner of UB who might also bid. As ≥ vUB (D1 , D2 , UA , UB ) (9) 102

46 D1 ’s future surplus depends upon the owner, it UB s Returns from Losing 103
U
47 will not treat such bids equally; thus, the auction 104
48 is discriminatory. Moreover, UB is aware of the The following proposition demonstrates that (8) 105
49 different implications associated with not becom- and (9) will both hold, so D1 will receive strictly 106
50 ing the owner and will take this into account 107
51 in its bidding. Thus, the auction produces exter- 17
Gans (2005) provides a comprehensive analysis of such auc- 108
52 nal effects, so care must be taken in specifying tions in the context of ownership. 109
53 110
54 DOI: 10.1002/smj 111
55 112
56 113
57 114
1 more than the value of the outsourced assets to they have no real productive value. Thus, even if 58
2 an independent firm because UB bids enough to an established firm already has high efficiency and 59
3 compensate it for ongoing losses from a lack of strong capabilities and does not need (potentially 60
4 competition. overlapping) resources from another company, an 61
5 independent firm would not be in a position to pay 62
6 Proposition 4: Suppose that A is sufficiently for those assets given the presence of that sup- 63
7 high so that D1 always wants to outsource. plier.18 This possibility reinforces the prediction 64
8 Under a discriminatory asset auction, D1 will that outsourcing will occur to an established firm. 65
9 always prefer to outsource to an established 66
10 rather than to an independent upstream firm. 67
11 ROBUSTNESS 68
12 In this situation, UB will be willing to set b 69
13 just above vUA (D1 , D2 , UA , UB ) + vD1 (D1 , D2 , UA , Our baseline model provides a clear prediction that 70
FS
14 UB ) − vD1 (D1 , D2 , UA UB ); which is the bid of the based on bargaining position considerations, firms 71
15 independent firm. Thus, while considering just the should outsource to established upstream suppli- 72
16 cost implications, D1 would choose to set up an ers rather than create competition for them. The 73
17 independent UA rather than outsource to an estab- baseline model has some built-in assumptions that 74
O
18 lished UB , the value of A’s assets in UB ’s hands is simplify exposition. Consequently, in this section 75
19 greater than their value with an independent owner. and the next, we consider alternatives. This section 76
O
20 This is because the higher costs D1 faces also flows looks at alternatives that reinforce the baseline 77
21 on to D2 with a consequent increase in asset value result, while the next looks for explanations of 78
PR
22 that more than compensates D1 for ongoing losses. when independent outsourcing might be profitable. 79
23 Having options of where to outsource raises 80
24 the returns to outsourcing. Established firms bid 81
Internal bargaining
25 more because of competition in the auction from 82
26 independents. Thus, the overall incentive to out- de Fontenay and Gans (2005) assumed that a man- 83
D
27 source is higher than if only one type of firm ager who was somewhat irreplaceable was asso- 84
28 is being considered. In addition, what is driving ciated with each physical asset. What this meant 85
TE
29 established firm outsourcing is that the auction pro- was that if that asset was owned and operated by 86
30 cess takes into account the ‘before and after’ prof- another firm, the manager concerned would have to 87
31 its of D1 , UB , and the independents. However, it be an employee of that firm. That manager would 88
32 ignores D2 . Because D2 is harmed more by estab- have some bargaining power over their wage as 89
EC
33 lished firm outsourcing than independent outsourc- the asset would be less or unproductive without 90
34 ing, this drives the incentives of others to extract them. 91
35 those rents. Specifically, D1 and UB are able to Allowing for internal bargaining of this kind 92
36 share in the rents accruing from UB ’s improved changes the returns to outsourcing. Specifically, if 93
R
37 bargaining position with respect to D2 . D1 were to outsource UA to an independent com- 94

38 These effects can be starkly demonstrated when pany, it would not have to negotiate with UA ’s 95
R
39 it is the case that the divested assets add lit- manager directly but instead through the indepen- 96
40 tle to the ability of an established firm to sup- dent firm. Not surprisingly, UA ’s manager would 97
O
41 ply the industry; that is, upstream capacity is have more bargaining power if they more closely 98
42 plentiful so that (D1 D2 UA UB ) = (D1 D2 Uj ) controlled the physical asset (as they would under 99
C
43 and (Di UA UB ) = (Di Uj ). In this case, the outsourcing) than if they were just an employee of 100
44 gain from trade of making A independent is D1 . It is not clear that this increases the cost of 101
N
45 − 16 (D1 UA ) + 65 A while the gain from trade of outsourcing per se because integration carries its 102
46 own cost (paying for employment), but its presence 103
selling A to an established UB is 61 ((D1 D2 UA ) −
U
47 104
48 (D1 UA )) + 65 A . Simply put, while indepen- 18
Note that if it were the case that the divested assets were 105
49 dent outsourcing reduces the value appropriated by much less efficient than the established upstream assets, they 106
50 upstream firms to zero, the threat of this is enough would still provide competitive discipline but earn no long-term 107
value. The threat of that competition will drive a more efficient
51 for the established supplier to be willing to pay a established firm to pay to have those assets removed from the 108
52 very high cost for the divested assets even though industry. 109
53 110
54 DOI: 10.1002/smj 111
55 112
56 113
57 114
1 changes the potential return. This is especially the value of UA ’s downstream assets. Consequently, 58
2 case if we follow de Fontenay and Gans (2005) the bargaining implications would be isomorphic 59
3 and assume that, under outsourcing, the manager to the outsourced supply case. 60
4 becomes the owner of the independent firm. Where these might differ is when there is inter- 61
5 In contrast, by outsourcing to an established nal bargaining. As de Fontenay and Gans (2005) 62
6 firm, the manager of UA remains an employee show, internal bargaining means that the status quo 63
7 regardless. It is just that their employment transfers point without outsourcing involves differing profits 64
8 from D1 to UB . This may change their wage. for the integrated firm (D1 − UA ). This is because 65
9 de Fontenay and Gans (2005) did not examine depending upon who owns whom (i.e., whether 66
10 outsourcing per se. Their focus was on vertical we consider this backward or forward integration 67
11 integration, so they did not consider D1 ’s choice from the viewpoint of the economic property rights 68
12 between independent and established firm out- view of the firm), the implications for firm prof- 69
13 sourcing. Doing so with internal bargaining yields its are different. Hence, in comparing those with 70
FS
14 the following result: outsourcing, the returns to outsourcing may them- 71
15 selves be different. However, even in this case, the 72
16 Proposition 5: Suppose that A is sufficiently choice between independent and established firm 73
17 high so that D1 always wants to outsource and outsourcing will be driven by similar considera- 74
O
18 that there is internal bargaining. Under a dis- tions to Proposition 5. 75
19 criminatory asset auction, D1 will always prefer 76
O
20 to outsource to an established rather than to an 77
Competing downstream firms
21 independent upstream firm. 78
PR
22 Thus far we have assumed that downstream firms 79
23 Thus, internal bargaining considerations—while compete in separate markets (geographic or other- 80
24 placing a cost of an established upstream firm in wise). However, in the case of GE with micro- 81
25 terms of creating a powerful employee—do not waves and Motorola with mobile phones, they 82
26 overturn the baseline result. This is because out- could be regarded as competing directly with the 83
D
27 sourcing to an established firm still consolidates established firms that they considered outsourcing 84
28 competitive outcomes, the returns of which are to. In each case, the anecdotal reports listed this as 85
TE
29 shared between D1 and UB , and otherwise merely a reason for outsourcing to an independent firm (at 86
30 transfers the powerful employment relationship least one that was independent at the time). For this 87
31 and its bargaining costs between D1 and the estab- reason, it is natural to explore downstream compe- 88
32 lished upstream firm. Moreover, as the manager of tition as a motivation for independent outsourcing. 89
EC
33 UA has some bargaining power in either case, UB ’s It turns out, however, that downstream compe- 90
34 incentives to bid for the upstream assets, UA , are tition only reinforces the rationale for established 91
35 themselves diminished.19 firm outsourcing. To see this, we cannot rely on 92
36 a simple application of the Shapley value as a 93
R
37 model of bargaining between upstream and down- 94

38 Forward versus backward integration stream firms. This is because that model allows 95
R
39 While outsourcing considered here is outsourcing competing downstream firms to reach agreements 96
40 of supply, the considerations would apply equally that collusively lead to monopoly outcomes in the 97
O
41 to outsourcing demand; that is, where UA chooses industry—that is, it assumes away the very com- 98
42 to divest D1 . In this case, the choice would be petitive forces we are trying to take into account. 99
C
43 between an independent D1 or selling D1 to D2 . Instead, the bilateral bargaining approach devel- 100
44 As D1 and D2 do not directly compete, the choice oped by de Fontenay and Gans (2005, 2006) pro- 101
N
45 between the two options is similar and consolida- vides a noncooperative game theoretic approach to 102
46 tion of demand is attractive because it raises the bargaining that allows for competitive externalities 103
U
47 to be present and to matter. This means that the 104

48 19
This type of force is explored in more detail in Gans (2005). bargaining outcome—while being Shapley value- 105
49 There, outsourcing to an established firm may bring about like in terms of what firms appropriate—has value 106
50 an additional cost in terms of the incentives for the manager determined in a noncooperative fashion. Specifi- 107
of UA to undertake certain productive actions. Such potential
51 inefficiency would, of course, mitigate the incentive toward cally, industry profits are not maximized and may 108
52 established firm outsourcing, but not necessarily eliminate it. potentially change with outsourcing. 109
53 110
54 DOI: 10.1002/smj 111
55 112
56 113
57 114
1 For our purposes, it is useful to consider a and industry profit dissipation. D1 will also favor 58
2 simple version of the de Fontenay and Gans that outcome. Hence, the possibility that outsourc- 59
3 (2005) setup to illustrate why downstream com- ing encourages future competition is another rea- 60
4 petition will drive established firm outsourcing. son to favor established firm rather than indepen- 61
5 Suppose that when both downstream firms are dent outsourcing. It suggests that in GE’s case, 62
6 present, they compete and industry profits are and perhaps Motorola’s as well,22 the choice of 63
7 reduced to (Dˆ 1 D2 UA UB ) < (D1 D2 UA UB ) and independent outsourcing may have been poorly 64
8 ˆ
(D1 D2 Uj ) < (D1 D2 Uj ), respectively. The idea considered or at least based on factors other than 65
9 is that the presence of competing downstream firms bargaining position and potential competition.23 66
10 leads to rent dissipation to final consumers, but 67
11 the magnitude of this depends on the cohort of Outsource and break up 68
12 upstream resources present in the industry. 69
13 With this setup, we can demonstrate the follow- In addition, we consider outsourcing by a ‘large’ 70
FS
14 ing. firm with many productive assets. Such outsourc- 71
15 ing may involve creating a separate entity or alter- 72
16 Proposition 6: Suppose that A is sufficiently natively breaking up production between differ- 73
17 high so that D1 always wants to outsource and ent partners. British Petroleum (BP) did the latter 74
O
18 that downstream firms directly compete. Under when it outsourced its information technology ser- 75
19 a discriminatory asset auction, D1 will always vices in the 1990s (Cross, 1995). 76
O
20 prefer to outsource to an established rather than Given this, as a final exercise we consider D1 ’s 77
21 to an independent upstream firm. outsourcing decision if it is the owner of all 78
upstream assets. The Shapley values of each firm 79
PR
22
23 The intuition is as follows. When both downstream in this status quo are: 80
24 firms outsource their production, industry profits 81
25 are driven by competition between them. Even vD1 UA UB (D1 UA UB , D2 ) 82
when there is an upstream monopoly, industry
26 1 83
= (D1 D2 UA UB ) + (D1 UA UB ) (10) 84
D
27 rents are dissipated by downstream competition.
2
28 Indeed, a key result of de Fontenay and Gans 85
vD2 (D1 UA UB , D2 )
TE
29 (2005) is that this rent dissipation is the same, 86

30 regardless of whether upstream assets are inde- 1 87
pendently held or consolidated.20 Thus, the choice = (D1 D2 UA UB ) − (D1 UA UB ) (11)
31 2 88
32 between outsourcing to an independent or estab- 89
EC
33 lished firm rests on pure bargaining effects; pre- In this situation, the decision is not to whom to 90
34 cisely the same as those derived in Proposition outsource but whether to outsource as a consoli- 91
35 4. As such, allowing for downstream competition dated supplier or to break up the upstream assets 92
36 does not change this choice.21 into competing entities. 93
R
37 Similarly, suppose that in outsourcing to an D1 ’s incentive to outsource and break up versus 94

38 independent firm, D1 creates a platform for future continue as an integrated entity depends on the 95
R
39 downstream entry by that firm. This is arguably sign of the following inequality: 96
40 what happened with GE and Samsung with respect 97
O
41 to microwave ovens. In this situation, future indus- vD1 (D1 , D2 , UA , UB ) + vUA (D1 , D2 , UA , UB ) 98
try profits will be dissipated further by indepen-
42 + vUB (D1 , D2 , UA , UB ) ≥ vD1 UA UB (D1 UA UB , D2 ) 99
C
43 dent firm outsourcing as all downstream firms face 100

44 future competition. While this will raise the inde- 101
22
N
pendent firm’s bid for D1 ’s upstream assets, this In that case, BenQ’s entry was arguably into a separate
45 geographic market and so less of a direct competition issue. 102
46 possibility also raises UB ’s bid price as it recog- However, it did involve BenQ becoming integrated, which is 103
U
47 nizes that its successful bid would prevent entry something we consider in the next section.
104
23
For instance, one referee suggests that established upstream
48 firms might represent stronger entry threats into downstream 105
20
49 This is a general outcome in economic models of vertical markets than independent ones. This would be a reason to favor 106
contracting (McAfee and Schwartz, 1994; Rey and Tirole, 2007). independent outsourcing. However, to fully rationalize GE’s and 107
50 21
de Fontenay and Gans (2005) do demonstrate that an integrated Motorola’s decisions, it would have to explain why the observed
51 upstream monopolist can achieve higher industry profits. This entry from the independent firms was weaker than what might 108
52 will change the returns to that particular structure. have occurred from established ones. 109
53 110
54 DOI: 10.1002/smj 111
55 112
56 113
57 114

1 3(D1 D2 UA UB ) − (D1 D2 UA )  an important dimension. We consider such dimen- 37
2 1  −(D1 D2 UB )  sions next. 38
 
3 ⇒  +(D1 UA ) + (D1 UB )  39
4 12  −(D U ) − (D U ) 
40
2 A 2 B
5 −3(D1 UA UB ) − (D2 UA UB ) EXPLAINING INDEPENDENT 41
6 5 OUTSOURCING 42
7 + (A + B ) ≥ 0 (12) 43
6
8 Our baseline results give a clear prediction that 44
9 However, D1 could also choose to outsource but faced with a choice between outsourcing to an 45
10 keep the upstream assets as a single entity. Its independent firm or an established firm, a down- 46
11 incentive to do this is given by the following: stream firm will choose the established firm. While 47
12 the ongoing costs of doing so are higher, the down- 48
13 vD1 (D1 , D2 , UA UB ) + vUA UB (D1 , D2 , UA UB ) stream firm will appropriate enough asset value in 49
FS
14 ≥ vD1 UA UB (D1 UA UB , D2 ) the sale to completely offset them. 50
15 This prediction from the bargaining perspective 51
16 1 stands in contrast to observations that spin-offs 52
⇒ (D1 D2 UA UB ) − (D1 UA UB )
17 6 and independent outsourcing do occur. Indeed, 53
O
18 this was the case with our motivating examples 54
5
19 −(D2 UA UB ) + (A + B ) ≥ 0 (13) in the introduction. Consequently, given the clar- 55
6
O
20 ity of our prediction on this, it is worthwhile 56
21 In comparing (12) and (13), it is readily appar- to consider some amendments to the baseline 57
PR
22 ent that the incentive to outsource and keep the model that might generate an opposite conclu- 58
23 upstream assets together is greater than the break sion. 59
24 up strategy. This is because as their initial owner, Clearly, if it were the case that independent 60
25 D1 gains more by selling an upstream monopolist outsourcing was more productively efficient than 61
26 than upstream competitors, with D2 being harmed established firm outsourcing (we had assumed 62
D
27 as a result of this decision. they were symmetric in this regard), then we 63
28 Thus, BP’s choice to break up its outsourc- might observe independent outsourcing. In addi- 64
TE
29 ing arrangements stands alongside that of GE and tion, if D1 ’s upstream assets had no value in 65
30 Motorola as a puzzle from a purely bargaining and of themselves, then D1 would prefer to out- 66
31 perspective. This suggests that either other, non- source to an independent rather than to an estab- 67
32 bargaining factors were at work in those cases, lished firm. However, our model was flexible 68
EC
33 or alternatively, the setup here is restrictive on enough to consider a range of environments. 69

34 70
35 Table 2. Shapley value outcomes 71
36 72
R
UB owned by D2

vD1 UA = 1 (D1 D2 UA UB ) + (D1 UA ) − (D2 UB ) ,
R
D1 owns UA
2
vD2 UB = 1 (D1 D2 UA UB ) − (D1 UA ) + (D2 UB )
2
O

UA nonintegrated vD1 = 12 1 4(D1 D2 UA UB ) + 2(D1 D2 UB )
−4(D2 UA UB ) − 2(D2 UB ) + 2(D1 UA )
C
vUA = 12 1 4(D1 D2 UA UB ) − 4(D1 D2 UB ) +

+2(D
2 UA UB ) − 2(D2 UB ) + 2(D1 UA )
4(D D U U ) + 2(D D U )
v D 2 UB = 1
N
1 2 A B 1 2 B
12 +2(D2 UA UB ) − 4(D1 UA ) + 4(D2 UB )

vD1 = 1 (D1 D2 UA UB ) − (D2 UA UB )
U
D2 owns UA and UB
2
vD2 UA UB = 1 (D1 D2 UA UB ) + (D2 UA UB )
2

D1 owns UA and UB vD1 UA UB = 21 (D1 D2 UA UB ) + (D1 UA UB )

vD2 = 21 (D1 D2 UA UB ) − (D1 UA UB )
DOI: 10.1002/smj
1 In each case, however, this outcome did not As can be seen from the above inequality, the 58
2 arise. impact on D2 is, in fact, a positive one and its bar- 59
3 What we consider here, therefore, are explana- gaining position is enhanced. Thus, D1 will only 60
4 tions that rely on bargaining outcomes rather than choose to outsource if the benefits from so doing 61
5 on other environmental factors. That is the focus of are sufficiently high. Nonetheless, if it were going 62
6 this article and is appropriate given that it was the to outsource, D1 would outsource to an indepen- 63
7 relative changes in bargaining position that drove dent firm in this instance.24 64
8 Proposition 4. Thus, if the starting position in an industry were 65
9 that other downstream firms owned their own sup- 66
10 Outsourcing to an integrated supplier pliers, then D1 would prefer to outsource and cre- 67
11 ate an independent firm rather than to rely on 68
12 The first alternative specification—and one we supply from one of these others. Notice that this 69
13 anticipated in setting up the model—arose if D2 − occurs even though downstream firms do not com- 70
FS
14 UB were themselves integrated, with UB already pete in the final product market. It is a pure effect 71
15 owned by D2 . Table 2 states the Shapley value arising from the change in bargaining position that 72
16 outcomes for this situation. outsourcing will bring as well as the potential 73
17 In this situation, we obtain the following result: to realize any productive benefits from outsourc- 74
O
18 ing. 75
19 Proposition 7: Suppose that A is sufficiently 76
high so that D1 always wants to outsource and
O
20 Double marginalization 77
21 that D2 and UB are integrated. Under a discrim- 78
inatory asset auction, D1 will always prefer to Our bargaining approach here models upstream
PR
22 79
outsource to an independent rather than to an and downstream firms negotiating supply agree-
23 80
established upstream firm. ments that are efficient. In particular, when down-
24 81
25 stream firms make their decisions as to input vol- 82
Outsourcing to an integrated supplier is less prof- umes to purchase, their marginal decision reflects
26 itable than outsourcing to an independent one; a 83
D
27 the marginal costs faced by upstream firms in 84
reversal of the prediction of Proposition 4. producing those inputs. This is reasonable when
28 The intuition is as follows. First, note that 85
TE
29 upstream and downstream firms negotiate directly 86

there are no advantages from outsourcing to the with one another. Indeed, supply agreements con-
30 existing upstream player. Specifically, the sum 87
31 tain rebates, discounts, and volume bonuses that 88
of payoffs to the two integrated chains prior to are all designed to ensure efficient procurement.
32 89
EC
the merger is (D1 D2 UA UB ), and if UA moves In contrast, when supply agreements are deter-
33 from D1 to D2 , both affected parties are part 90
34 mined at arms length, it is sometimes argued 91
of the transaction and so there are no external that efficient pricing does not result. Instead, sup-
35 effects; it is a pure transfer of value. Hence, the 92
36 ply arrangements involve downstream firms being 93
R
two structures involve different ways of split- offered a simple linear price (i.e., price per unit)
37 ting (D1 D2 UA UB ) between D1 and D2 , and 94
38 and then choosing the quantity they would like to 95
R
hence there are no gains from this type of asset purchase at that price. In this situation, outsourcing
39 exchange. 96
40 results in the well-known double marginalization 97
O
In contrast, if D1 sells UA to an independent problem whereby supply prices are above marginal
41 owner, there may be gains from trade as this 98
42 costs and downstream firms themselves procure 99
impacts D2 .
C
43 and price on the basis of this inflated price. The 100

end result is less production and, indeed, profits in
44 vD1 (D1 , UA , D2 UB ) + vUA (D1 , UA , D2 UB ) 101
N
45 the industry. 102

46 ≥ vD1 UA (D1 UA , D2 UB ) 103

U
24
47 1 Interestingly, if D1 were to outsource but then choose to auc- 104
48 ⇒ (D1 D2 UA UB ) − (D1 D2 UB ) tion off the assets, D2 ’s highest bid would match that of an
105
6 independent owner. That is, taking into account D2 ’s highest bid,
49 D1 would receive vD1 (D1 , D2 UA UB ) + vD2 UA UB (D1 , D2 UA UB ) 106
50 −(D2 UA UB ) + (D2 UB ) − vD2 UB (D1 , UA , D2 UB ) in this instance and D1 (D1 , UA , 107
D2 UB ) + vUA (D1 , UA , D2 UB ) if it sold to an independent firm.
51 5 108
−(D1 UA ) + A ≥ 0 (14) It is easy to see that these values are identical. In this case, other
52 6 factors would determine who the owner is. 109
53 110
54 DOI: 10.1002/smj 111
55 112
56 113
57 114
1 Upstream competition is a force that mitigates BANDWAGON EFFECTS 58

2 the double marginalization problem. Hence, it is 59
3 useful here to consider that alternative model rather The above analysis has demonstrated that (i) when 60
4 than our bargaining model and what it implies UB is independently owned, D1 prefers to out- 61
5 for the established firm versus independent firm source to it rather than to an independent firm; and 62
6 choice. Recall that the two considerations driv- (ii) when UB is integrated, D1 prefers to outsource 63
7 ing our baseline result—that established firm out- to an independent firm rather than to it. Given this, 64
8 sourcing raises ongoing procurement costs but it is a natural question to ask whether D1 ’s incen- 65
9 affords higher value from the sale of assets—still tives to outsource are higher if UB is independently 66
10 occur in the double marginalization environment. owned or not? That is, will outsourcing by D2 raise 67
11 The issue is how the trade-off between them incentives for D1 to outsource? 68
12 plays out. To analyze this, consider a situation where both 69
13 Analyzing linear pricing models in a vertical D1 and D2 own an upstream asset and D2 moves to 70
FS
14 contracting situation can become very complex. outsource first. Observing this, D1 decides whether 71
15 However, in a simple setting we can demonstrate to outsource. Working backward, if D2 has out- 72
16 the following: sourced (creating an independent UB 25 ), then D1 ’s 73
17 74
O
incentives to outsource are driven by the same con-
18 siderations as Proposition 4; that is, it will prefer 75
Proposition 8: Suppose that A is sufficiently
19 to outsource to an established rather than to an 76
high so that D1 always wants to outsource and
O
20 independent firm. If D2 does not outsource, D1 ’s 77
(a) downstream demand is linear; (b) upstream
21 incentives are given by Proposition 7. Comparing 78
and downstream outputs are perfect substitutes;
PR
22 these two, it is easy to see that D1 ’s incentives to 79
and (c) upstream and downstream firms are sym-
23 outsource at all are greater when UB is indepen- 80
metric. Under a discriminatory asset auction,
24 81
D1 will always prefer to outsource to an inde- dently owned. Thus, outsourcing elsewhere tends
25 82
pendent rather than to an established upstream to raise incentives for outsourcing by D1 .
26 83
firm. But given this, will D2 choose to outsource?
D
27 84
After all, it will recognize that by outsourcing, it
28 85
will not face an independently owned UB but a
TE
29 This proposition demonstrates that it is possible to 86

find cases where, under linear pricing, outsourc- commonly owned upstream monopolist. This may
30 87
ing imposes a cost on the entire industry, leading be an acceptable situation if D2 can appropriate
31 88
to greater double marginalization. This problem is sufficient value from the assets outsourced to com-
32 89
EC
less pronounced when there is upstream compe- pensate for ongoing costs associated with outsourc-
33 90
tition than when there is an upstream monopoly. ing. However, we have not specified in sufficient
34 91
For this reason, if outsourcing were to occur, a detail the ‘asset trading game’ that might arise in
35 92
downstream firm would prefer it to occur in a this situation. In many respects, that is a topic well
36 93
R
competitive way rather than a monopolistic one, beyond the scope of this article. Nonetheless, in
37 94
even if it could appropriate the rents from that its outsourcing decision, should D1 choose to out-
38 95
R
monopoly. source regardless, D2 will be comparing a situation

39 96
40 Double marginalization is a mechanism by where it stays integrated to one where there is a 97
O
41 which consolidating upstream assets reduces over- nonintegrated upstream monopolist. It may choose, 98
42 all industry profits. As a result, that effect can therefore, to outsource for the same reason as D1 99
C
43 counteract the bargaining position advantages that did in Proposition 1. 100

44 D1 can achieve with UB by established firm This discussion highlights not only some of the 101
N
45 outsourcing. However, the issue empirically is interdependencies between outsourcing decisions, 102
46 whether this type of pricing is an ongoing restric- but also the complexity that arises as a result of 103
U
47 tion of a sufficiently ongoing nature to impact those interdependencies. This suggests that exten- 104
48 long-term value creation in the industry. It sug- sions to the model to capture those complexities 105
49 gests that in investigating the nature of outsourc- are a fruitful area for future research. 106
50 ing, empirical researchers need to closely examine 107
51 the overall form of contracting that is present in 25
This is its only option in this game as there is no established 108
52 the industry. firm to outsource to. 109
53 110
54 DOI: 10.1002/smj 111
55 112
56 113
57 114
1 CONCLUSION intuition, when efficient bilateral contracts can be 58

2 written the reverse is true. The incentives to out- 59
3 Applications of bargaining theory to strategy and source (as opposed to retain production in-house) 60
4 supply chain management are in their infancy. are stronger when an upstream monopoly is pos- 61
5 Models based on bargaining have only been sible than when there is upstream competition. 62
6 recently developed. This article uses one such 63
7 model to evaluate the type of outsourcing choices 64
8 that might be made by firms; specifically, whether ACKNOWLEDGEMENTS 65
9 a firm wants to create an independent competitor or 66
10 to outsource to an established firm. In the absence This article was previously circulated under the 67
11 of any other motive for choosing between them title ‘Supply Competition and Outsourcing.’ We 68
12 (say, based on already developed competencies), thank Marianne Broadbent, Ken Corts, Michael 69
13 a downstream firm must trade off lower expected Ryall, participants at the CRES Foundations of 70
FS
14 input prices against lower asset value when choos- Business Strategy Conference (St. Louis, 2005) 71
15 ing between an independent and an established and two anonymous referees for helpful comments. 72
16 firm. Responsibility for all errors lies with the authors. 73
17 Utilizing a bargaining perspective, we demon- 74
O
18 strate that industry value is not impacted by these 75
19 choices. Based on pure bargaining effects, a down- REFERENCES 76
O
20 stream firm earns more by outsourcing, or even Adner R, Zemsky P. 2006. A demand-based perspective
77
21 better, by auctioning off its assets to an estab- on sustainable competitive advantage. Strategic 78
PR
22 lished firm. Basically, an upstream monopoly is Management Journal 27(3): 215–239. 79
23 more valuable to an asset owner and hence that Baye MR, Crocker KJ, Ju J. 1996. Divisionalization, 80
24 owner will have an incentive to create that out- franchising, and divestiture incentives in oligopoly. 81
American Economic Review 86(1): 223–236.
25 come. Brandenburger AM, Stuart HW. 1996. Value-based busi-
82
26 The analysis here is admittedly specialized to ness strategy. Journal of Economics and Management 83
D
27 focus on the bargaining aspects of outsourcing. As Strategy 5(1): 5–24. 84
28 such, we do not explicitly take into account other Brandenburger AM, Stuart HW. 2007. Biform games. 85
TE
29 explanations for outsourcing or imperfections in Management Science 53(3): 537–549. 86

Cachon G, Harker P. 2002. Competition and outsourcing
30 capital markets that might play a role in observed with scale economies. Management Science 48(10):
87
31 ownership structures. Our approach does not rule 1314–1333. 88
32 out these alternative factors, indeed it is comple- Cross J. 1995. IT outsourcing: British Petroleum’s 89
EC
33 mentary to them, but in their absence we do not competitive approach. Harvard Business Review 90
34 claim to have provided a complete set of condi- 73(May–June): 94–102. 91
de Fontenay CC, Gans JS. 2005. Vertical integration in
35 tions that might drive one form of outsourcing over the presence of upstream competition. RAND Journal 92
36 another. 93
R
of Economics 36(3): 544–572.

37 There is a sense in which establishing indepen- de Fontenay CC, Gans JS. 2006. Bilateral bargaining 94
38 dent upstream ownership as a long-term outcome with externalities. Working paper, Melbourne Business 95
R
39 relies upon the operation of antitrust laws. Hav- School, Melbourne Australia. 96
Domberger S. 1998. The Contracting Organization: A
40 ing divested its upstream assets to an independent 97
O
Strategic Guide to Outsourcing. Oxford University

41 owner, a downstream firm and the other upstream Press: Oxford, UK. 98
42 firms will still have an incentive to merge ex post. Espino-Rodriguez TF, Padron-Robaina V. 2006. A 99
C
43 Antitrust protection may limit this at this stage in a review of the outsourcing from the resource-based 100
44 way that it cannot at the earlier outsourcing stage. view of the firm. International Journal of Management 101
N
Reviews 8(1): 49–70.

45 A more comprehensive model of the dynamics of Foss K, Foss NJ. 2005. Resources and transaction costs: 102
46 asset ownership is something for further work. how property rights economics furthers the resourced- 103
U
47 Finally, it is useful to note that our results based view. Strategic Management Journal 26(6): 104
48 also shed light on the role of upstream com- 541–553. 105
49 petition per se. Outsourcing is often prescribed Gans JS. 2005. Markets for ownership. RAND Journal of 106
Economics 36(2): 433–455.
50 as a valuable outcome where upstream markets Gans, JS, MacDonald G, Ryall M. 2005. Operational- 107
51 are more competitive (Domberger, 1998; Spulber, izing value-based business strategy. Working paper, 108
52 1998). While the standard approach supports this Melbourne Business School, Melbourne, Australia. 109
53 110
54 DOI: 10.1002/smj 111
55 112
56 113
57 114
 
1 Gilbert RJ, Newbery DM. 1982. Preemptive patenting (D1 D2 UA UB ) + (D1 D2 UA ) 58
2 and the persistence of monopoly. American Economic
1  −(D1 D2 UB ) + (D1 UA UB )  59
Review 72(3): 514–526.  
3 =  +3(D2 UA UB )  60
4
Grossman S, Hart OD. 1986. The costs and benefits of 12  +3(D U ) − (D U ) 
61
ownership: a theory of vertical and lateral integration. 1 A 1 B
5 Journal of Political Economy 94(4): 691–719. +(D2 UA ) − 3(D2 UB ) 62
6 Hart OD, Moore J. 1990. Property rights and the theory 1 63
7 of the firm. Journal of Political Economy 98(6): − A (15) 64
1119–1158. 6
8 65
9
Inderst R, Wey C. 2003. Bargaining, mergers and tech- vD1 UA (D1 UA , D2 , UB ) − vD1 (D1 , D2 , UA UB ) 66
nology choice in bilaterally oligopolistic industries.
10 RAND Journal of Economics 34(1): 1–19. 1 67
= 2(D1 D2 UA ) + 4(D2 UA UB )
11 Jarillo JC. 1993. Strategic Networks: Creating the 12 68
12 Borderless Organization. Butterworth-Heinemann: 69
Oxford, UK. 1
13 +4(D1 UA ) − 4(D2 UB ) − A (16) 70
Lippman SA, Rumelt RP. 2003. A bargaining perspective 6
FS
14 71
on resource advantage. Strategic Management Journal
15 24(11): 1069–1086. 72
16 It is easy to see that the first (bracketed) term in 73
MacDonald G, Ryall MD. 2004. How do value creation
17 and competition determine whether a firm appropriates each of these equations is positive. Hence, out- 74
O
18 value? Management Science 50(10): 1319–1333. sourcing will only result in cost reductions (these 75
19 McAfee RP, Schwartz M. 1994. Opportunism in multilat- equations being negative) if A is sufficiently 76
O
eral vertical contracting: nondiscrimination, exclusiv- large.
20 ity, and uniformity. American Economic Review 84(1): 77
21 210–230. 78
PR
22 McGuire TW, Staelin R. 1983. An industry equilibrium 79
Proof of Proposition 2
23 analysis of downstream vertical integration. Marketing 80
24 Science 2(2): 161–191. Using the values in Table 1, the total return for 81
Nickerson J, Vanden Bergh R. 1999. Economizing in a
25 independent outsourcing is: 82
context of strategizing: governance mode choice in
26 Cournot competition. Journal of Economic Behavior 83
D
27 & Organization 40(1): 1–15. vUA (D1 , D2 , UA , UB ) + vD1 (D1 , D2 , 84
28 Rey P, Tirole J. 2007. A primer on foreclosure. 85
TE
29 In Handbook of Industrial Organization (Vol. 3), UA , UB ) − vD1 UA (D1 UA , D2 , UB ) 86

Armstrong M, Porter RH (eds). North Holland:
30 1 87
31
Amsterdam; 2145–2220. = 2(D1 D2 UA UB ) − 2(D1 D2 UB ) 88
Ryall MD, Sorenson O. 2007. Brokers and competitive 12
32 89
EC
advantage. Management Science 53(4): 566–583.

33 Shy O, Stenbacka R. 2003. Strategic outsourcing.
−2(D1 UA ) + 2(D2 UB ) 90

34 Journal of Economic Behavior & Organization 50(2): 5 91
203–224. −2(D2 UA UB ) + A (17)
35 6 92
Spiegel Y. 1993. Horizontal subcontracting. RAND
36 93
R
Journal of Economics 24(2): 570–590.

37 Spulber DF. 1998. The Market Markers: How Leading Note that the first (bracketed term) is negative. 94
38 Hence, outsourcing will only be profitable to an 95
R
Companies Create and Win Markets. McGraw-Hill:

39 New York. independent firm if A is positive. However, this 96
40 is not sufficient for this to be the case. In con- 97
O
41 trast, for established firm outsourcing, the return is 98

42 APPENDIX: PROOFS OF potentially: 99
C
43 PROPOSITIONS 100
44 vUA UB (D1 , D2 , UA UB ) 101
N
45 102
46 Proof of Proposition 1 − vUB (D1 UA , D2 , UB ) − (vD1 UA (D1 UA , D2 , UB ) 103

U
47 1 104
Using the values in Table 1, note that the poten- − vD1 (D1 , D2 , UA UB )) = 2(D1 D2 UA )
48 12 105
tial ‘cost’ of outsourcing for the independent and
49 106
50
established firm cases respectively are: −2(D2 UA UB ) − 2(D1 UA ) 107

51 5 108
+2(D2 UB ) + A (18)
52 vD1 UA (D1 UA , D2 , UB ) − vD1 (D1 , D2 , UA , UB ) 6 109
53 110
54 DOI: 10.1002/smj 111
55 112
56 113
57 114

1 Here the bracketed term may be positive or nega- 1 58
2 tive. Hence, A being positive is neither necessary = 2(D1 D2 UA ) − 2(D1 UA ) + 2(D2 UB ) 59
12
3 nor sufficient for such outsourcing to be profitable. 60
5
4 −2(D2 UA UB ) + A (20) 61
5 6 62
6 Comparing (19) and (20) we have: 63
7 From (17), note that as capacity becomes less 64
scarce, 2(D1 D2 UA UB ) → 2(D1 D2 UB ) and
8 1 65
9 2(D2 UB ) → 2(D2 UA UB ) eliminating a nega- 2(D1 D2 UA UB ) − 2(D1 D2 UB ) 66
tive term and so increasing the returns to outsourc- 12
10 67
11 ing to an independent firm. 68
−2(D1 UA ) + 2(D2 UB ) − 2(D2 UA UB )
12 From (18), when capacity becomes limited, the 69
13 returns to outsourcing become 12 1 (−2(D U U ) 70
2 A B 5
+ A
FS
14 5
+ 2(D2 UB )) + 6 A while with unlimited capac- 6 71
15 72
ity they are: 12 1 (2(D D U ) − 2(D U )) + 1
16 1 2 A 1 A ≤ 2(D1 D2 UA ) − 2(D2 UA UB ) 73
17 5 . It is clear that the limited capacity return is 12 74

O
6 A
18 higher. Thus, the returns to outsourcing to an estab- 5 75
19 −2(D1 UA ) + 2(D2 UB ) + A 76
lished firm rise as capacity becomes more scarce. 6
O
20 77
⇒ (D1 D2 UA UB ) − (D1 D2 UA )
21 78
≤ (D1 D2 UB )
PR
22 (21) 79
23 Because a discriminatory auction is being used, the 80
24 choice between independent and established firm where the last inequality follows because the 81
25 outsourcing will depend upon the bilateral gains marginal return of UB in coalition value is highest 82
26 from trade given that outsourcing will occur. The when UA is not present. 83
D
27 gain from trade of an asset to an independent UA 84
28 is: Proof of Proposition 5 85
TE
29 86
30 vUA (D1 , D2 , UA , UB ) Using the payoffs in de Fontenay and Gans (2005; 87
Table 1: 551); independent outsourcing will win
31 UA s Gain 88
32 an auction over established firm outsourcing if: 89
EC
(vD1 UA (D1 UA , D2 , UB )
33 − 90
34 −vD1 (D1 , D2 , UA , UB )) vD1 (D1 , D2 , UA , UB ) + vUA (D1 , D2 , UA , UB ) 91

35 D1 s Loss − vUA (D1 , D2 , UA UB ) 92
36 93
R
1 > vD1 (D1 , D2 , UA UB ) + vUB (D1 , D2 , UA UB )

37 = 2(D1 D2 UA UB ) − 2(D1 D2 UB ) 94
12
38 − vUB (D1 , D2 , UA , UB ) (22) 95
R
39 −2(D1 UA ) + 2(D2 UB ) 96

40 5 This is equivalent to: 97
O
41 −2(D2 UA UB ) + A (19) 98
6 1
42 (D1 UB ) − (D1 D2 UB ) − 2(D1 UA ) 99
C
43 The gain from trade of selling A to an established 12 100

44 UB is: 101
1
N
45 −5(D2 UB ) − A > 0 (23) 102

vUA UB (D1 , D2 , UA UB ) 6
46 103
U
47 −vUB (D1 UA , D2 , UB ) which can never hold. 104

48 UB s Gain 105
49 106
(vD1 UA (D1 UA , D2 , UB ) Proof of Proposition 6
50 − 107
51 −vD1 (D1 , D2 , UA UB )) For expositional purposes we assume that each 108

52 D1 s Loss upstream and each downstream firm is symmetric. 109
53 110
54 DOI: 10.1002/smj 111
55 112
56 113
57 114
1 In this situation, we can use the payoffs of Table 1 an integrated upstream asset, i’s demand will take 58
2 amended for the value reductions from down- into account the cost of input supply. 59
3 stream competition. Given this, an established firm If D1 currently produces inputs in-house (own- 60
4 will outbid an independent firm if: ing UA ), it may also sell those inputs externally 61
5 to D2 . In this situation, under the assumption of 62
6 ˆ 1 D2 UA UB ) ≤ 2(D
(D ˆ 1 D2 Uj ) (24) Cournot quantity competition, equilibrium input 63
7 supplies (q̂1A , q̂1B , q̂2A , q̂2B ) are determined by the 64
8 Where this inequality continues to hold as in the following two equations: 65
9 Proof of Proposition 4. 66
10 maxq1A ,q2A 67
11 68
12 Proof of Proposition 7 (a − b(q1A + q̂1B ))(q1A + q̂1B ) 69
13 −(a − 2b(q1A + q̂1B ))q̂1B 70
Proposition 4 establishes that D1 would prefer to +(a − 2b(q2A + q̂2B ))q2A − α(q1A + q2A ) (27)
FS
14 outsource to an established rather than an inde- 71
15 pendent upstream supplier. However, what if the −β 12 (q1A + q2A )2 72
16 established supplier was integrated: that is, what maxq1B ,q2B 73
17 74
happens if UB is already owned by D2 ? In this (a − 2b(q̂1A + q1B ))q1B
O
18 75
case, the analog of (21) becomes: +(a − 2b(q̂2A + q2B ))q2B
19 (28) 76
−α(q1B + q2B ) − β 12 (q1B
O
20 vD2 UA UB (D1 , D2 UA UB ) 77
21 +q2B )2 78
−vD2 UB (D1 , UA , D2 UB ) − vUA (D1 , UA , D2 UB )
PR
22 79
23 Marginal Gain Asset Value for A from UB Ownership In this situation, it is easy to show that: 80
24 ≥ vD1 (D1 , UA , D2 UB ) − vD1 (D1 , D2 UA UB ) (25) 81
25 82
Marginal Cost Increase for D1 Outsourcing to UB q̂1A = 3(4b + 3β), q̂2A = (4b + β),
26 83
D
27 q̂1B = 3β, q̂2B = (4b + 7β) 84
Rearranging and using values from Table 2, notice
28 that: 85
TE
29 where = a−α . In this case, the 86

30 24b2 + 34bβ + 10β 2 87
vD2 UA UB (D1 , D2 UA UB ) + vD1 (D1 , D2 UA UB ) payoffs to each firm are:
31 88
32 =(D1 D2 UA UB ) 89
EC
33 vD1 (D1 , UA , D2 UB ) + vUA vD1 UA (D1 UA , D2 , UB ) 90

34 ≥ (26) 91
(D1 , UA , D2 UB ) + vD2 UB (D1 , UA , D2 UB ) 88b3 + 180b2 β + 126bβ 2 + 25β 2
35 = 92
36 =(D1 D2 UA UB )+A 2(b + β)2 93
R
37 vD2 (D1 UA , D2 , UB ) = 16b 94

which cannot hold if A > 0. Thus, it is prefer-
38 95
R
able to outsource to an independent firm in this vUB (D1 UA , D2 , UB )

39 96
instance. 16b3 + 60b2 β + 78bβ 2 + 25β 2
40 97
= +
O
41 2(b + β)2 98
42 Proof of Proposition 8 99
C
43 100
Let downstream demand for Di be pi = a − where = 12b a − α 2 . It is useful to note that,
44 + 5β 101
b(qiA + qiB ) and its costs be linear with marginal
N
45 in this case, D1 ’s output exceeds D2 by 50 percent. 102

46 cost, α < a. Finally, suppose that Cj = β 12 (q1j + It has two options. First, it could sell its assets to 103
U
47 q2j )2 . an independent firm, in which case we would have 104

48 The marginal value of an additional unit of a completely nonintegrated industry. Alternatively, 105
49 qij supplied by j to downstream firm i is a − it could sell the assets to UB , in which case we 106
50 2b(qiA + qiB ). Given the homogeneity of upstream would have a nonintegrated upstream monopolist. 107
51 inputs, this means that i’s (inverse) demand for We examine the equilibrium outcomes of each of 108
52 external inputs is pi = a − 2b(qiA + qiB ). If it has these in turn. 109
53 110
54 DOI: 10.1002/smj 111
55 112
56 113
57 114
1 If D1 chooses to outsource to an independent (a − α)2 58

2 firm, the two upstream firms solve: vUA UB (D1 , D2 , UA UB ) = + 2 59
8b + β
3 60
4 (a − 2b(q1A + q̂1B ))q1A 61
Note that, in this case, D1 will choose to out-
5 +(a − 2b(q2A + q̂2B ))q2A 62
maxq1A ,q2A (29) source to UB if:
6 −α(q1A + q2A ) − β 12 (q1A 63
7 +q2A )2 vD1 (D1 , D2 , UA UB ) + vUA UB (D1 , D2 , UA UB ) 64
8 (a − 2b(q̂1A + q1B ))q1B 65
9 ≥ vD1 UA (D1 UA , D2 , UB ) + vUB (D1 UA , D2 , UB ) 66
+(a − 2b(q̂2A + q2B ))q2B
10 maxq1B ,q2B (30) (a − α)2 b(1744b4 + 3880b3 β 67
−α(q1B + q2B ) − β 12 (q1B
11 +3593b2 β 2 + 1073bβ 3 + 57β 4 ) 68
12 +q2B )2 ⇒≥ (33) 69
(b + β)2 (8b + β)2 (12b + 5β)2
13 −α 70
which gives equilibrium outcomes q̂ij = 6ba +
FS
14 2β − α)2 if β = 0. 71
15 for all ij. In this case, which collapses to ≥ 109(a
576b 72
16 73
17 b(a − α)2 Comparing (31) and (33) it is apparent that 74
vDi (D1 , D2 , UA , UB ) =
O
18 (3b + β)2 if D1 were to outsource, it would earn more 75
19 by outsourcing to an independent firm than to 76
(2b + β)(a − α)2
O
20 vUj (D1 , D2 , UA , UB ) = + UB . This is because the latter case results in an 77
2(3b + β)2 upstream monopoly with a more pronounced dou-
21 78
ble marginalization problem than the more com-
PR
22 Note that, in this case, D1 will choose to out- 79
23 petitive upstream case. 80
source to an independent firm if:
24 This presumes, however, that D1 is choosing 81
25 vD1 (D1 , D2 , UA , UB ) + vUA (D1 , D2 , UA , UB ) between alternative bilateral transactions. Suppose 82
26 instead that it was to auction its upstream assets to 83
≥ vD1 UA (D1 UA , D2 , UB )
D
27 an independent firm or UB . Recall our assumption 84
28 (a − α)23 3b(36b4 + 62b3 β that there is a continuum of independent firms 85
TE
29 who might purchase those assets and that (31) 86

+43b2 β 2 + 12bβ 3 + β 4 )
30 ⇒≥ (31) holds. In this case, an independent firm will be 87
(b + β)2 (3b + β)2 (12b + 5β)2 willing to pay vUA (D1 , D2 , UA , UB ) for the assets.
31 88
On the other hand, UB will be willing to pay
32 − α)2 if β = 0. Out- 89
EC
33 which collapses to ≥ (a 12b vUA UB (D1 , D2 , UA UB ) − vUB (D1 , D2 , UA , UB ) as it 90

34 sourcing leads to a direct advantage, , but at the realizes that D1 will find it worthwhile to sell to 91
35 cost of the emergence of a double marginaliza- an independent firm as (31) holds. In this case, D1 92
36 tion problem reflected in the right-hand side of the will only sell to UB if: 93
R
37 above inequality. 94
38 On the other hand, if it outsources to an estab- vD1 (D1 , D2 , UA UB ) + vUA UB (D1 , D2 , UA UB ) 95
R
lished firm, that upstream monopolist will solve:

39 − vUB (D1 , D2 , UA , UB ) 96
40 97
O
41
max{qij } i=1,2 ≥ vD1 (D1 , D2 , UA , UB ) + vUA (D1 , D2 , UA , UB ) 98
j =A,B
42 (a − 2b(q1A + q1B ))(q1A + q1B ) (a − α)2 b(31b + 2β) 99
C
43 ⇒− ≥0 (34) 100
+(a − 2b(q2A + q2B ))(q2A + q2B ) (3b + β)(8b + β)2
44 −α(q (32) 101
1A + q1B + q2A + q2B )
N
45 1 which can never hold. Thus, even an auction that 102

46 −β 2 (q1A + q2A )2 + (q1B + q2B )2 103
would allow D1 to extract the monopoly rent that
U
47 would accrue to UB , it would still prefer to sell to 104

which gives equilibrium outcomes q̂ij =
48 a−α an independent firm. 105
49 2(8b + β) for all ij. In this case, 106
50 107
51 3b(a − α)2 108
vDi (D1 , D2 , UA UB ) =
52 (8b + β)2 109
53 110
54 DOI: 10.1002/smj 111
55 112
56 113
57 114
WILEY AUTHOR DISCOUNT CLUB
We would like to show our appreciation to you, a highly valued contributor to Wiley’s
publications, by offering a unique 25% discount off the published price of any of our
books*.
All you need to do is apply for the Wiley Author Discount Card by completing the
attached form and returning it to us at the following address:
The Database Group (Author Club)

John Wiley & Sons Ltd
The Atrium
Southern Gate
FS
Chichester
PO19 8SQ
UK
Alternatively, you can register online at www.wileyeurope.com/go/authordiscount
O
Please pass on details of this offer to any co-authors or fellow contributors.
O
After registering you will receive your Wiley Author Discount Card with a special promotion
code, which you will need to quote whenever you order books direct from us.
PR
The quickest way to order your books from us is via our European website at:
http://www.wileyeurope.com
D
Key benefits to using the site and ordering online include:
x Real-time SECURE on-line ordering
TE
x Easy catalogue browsing

x Dedicated Author resource centre
x Opportunity to sign up for subject-orientated e-mail alerts
EC
Alternatively, you can order direct through Customer Services at:

cs-books@wiley.co.uk, or call +44 (0)1243 843294, fax +44 (0)1243 843303
So take advantage of this great offer and return your completed form today.
R
Yours sincerely,
R
O
C
Verity Leaver
N
Group Marketing Manager

author@wiley.co.uk
U
*TERMS AND CONDITIONS

This offer is exclusive to Wiley Authors, Editors, Contributors and Editorial Board Members in acquiring books for their personal use.
There must be no resale through any channel. The offer is subject to stock availability and cannot be applied retrospectively. This
entitlement cannot be used in conjunction with any other special offer. Wiley reserves the right to amend the terms of the offer at any
time.
REGISTRATION FORM
For Wiley Author Club Discount Card
To enjoy your 25% discount, tell us your areas of interest and you will receive relevant catalogues or leaflets
from which to select your books. Please indicate your specific subject areas below.
Accounting [] Architecture []
x Public []
x Corporate [] Business/Management []
Chemistry [ ] Computer Science [ ]

x Analytical [ ] x Database/Data Warehouse [ ]
x Industrial/Safety [ ] x Internet Business [ ]
x Organic [ ] x Networking [ ]
FS
x Inorganic [ ] x Programming/Software [ ]
x Polymer [ ] Development
x Spectroscopy [ ] x Object Technology []
O
Encyclopedia/Reference [ ] Engineering [ ]
x Business/Finance [ ] x Civil [ ]
O
x Life Sciences [ ] x Communications Technology [ ]
x Medical Sciences [ ] x Electronic [ ]
x Physical Sciences [ ] x Environmental [ ]
PR
x Technology [ ] x Industrial [ ]
x Mechanical [ ]
Earth & Environmental Science [] Finance/Investing [ ]

x Economics [ ]
D
Hospitality [] x Institutional [ ]
x Personal Finance [ ]
TE
Genetics [] Life Science []

x Bioinformatics/ []
Computational Biology Landscape Architecture []
EC
x Proteomics [ ]
x Genomics [ ] Mathematics []
x Gene Mapping [ ] Statistics
x Clinical Genetics [ ] []
R
Manufacturing
[]
R
Materials Science
[ ]
Medical Science [ ] Psychology [ ]
O
x Cardiovascular [ ] x Clinical [ ]
x Diabetes [ ] x Forensic [ ]
C
x Endocrinology [ ] x Social & Personality [ ]

x Imaging [ ] x Health & Sport [ ]
N
x Obstetrics/Gynaecology [ ] x Cognitive [ ]
x Oncology [ ] x Organizational [ ]
x Pharmacology [ ] x Developmental & Special Ed [ ]
U
x Psychiatry [ ] x Child Welfare [ ]

x Self-Help [ ]
Non-Profit [] Physics/Physical Science []
Please complete the next page /

I confirm that I am (*delete where not applicable):
a Wiley Book Author/Editor/Contributor* of the following book(s):

ISBN:
ISBN:
a Wiley Journal Editor/Contributor/Editorial Board Member* of the following journal(s):
SIGNATURE: …………………………………………………………………………………… Date: ………………………………………
PLEASE COMPLETE THE FOLLOWING DETAILS IN BLOCK CAPITALS:
TITLE: (e.g. Mr, Mrs, Dr) …………………… FULL NAME: …………………………………………………………………………….…
FS
JOB TITLE (or Occupation): ..…………………………………………………………………………………………………………………
DEPARTMENT: ……………………………………………………………………………………………………………………………………………..
O
COMPANY/INSTITUTION: ……………………………………………………………………………………………………………………………
O
ADDRESS: ……………………………………………………………………………………………………………………………………………………
PR
………………………………………………………………………………………………………………………………………………………………………
TOWN/CITY: …………………………………………………………………………………………………………………………………………………
COUNTY/STATE: ………………………………………………………………………………………………………………………………………….
D
COUNTRY: …………………………………………………………………………………………………………………………………………………….
TE
POSTCODE/ZIP CODE: …………………………………………………………………………………………………………………………………
DAYTIME TEL: ………………………………………………………………………………………………………………………………………………

EC
FAX: ………………………………………………………………………………………………………………………………………………………………
E-MAIL: …………………………………………………………………………………………………………………………………………………………
R
YOUR PERSONAL DATA

We, John Wiley & Sons Ltd, will use the information you have provided to fulfil your request. In addition, we would like to:
R
1. Use your information to keep you informed by post of titles and offers of interest to you and available from us or other
O
Wiley Group companies worldwide, and may supply your details to members of the Wiley Group for this purpose.
[ ] Please tick the box if you do NOT wish to receive this information
C
2. Share your information with other carefully selected companies so that they may contact you by post with details of
titles and offers that may be of interest to you.
N
[ ] Please tick the box if you do NOT wish to receive this information.
E-MAIL ALERTING SERVICE

U
We also offer an alerting service to our author base via e-mail, with regular special offers and competitions. If you DO wish to
receive these, please opt in by ticking the box [ ].
If, at any time, you wish to stop receiving information, please contact the Database Group (databasegroup@wiley.co.uk) at John Wiley & Sons Ltd,
The Atrium, Southern Gate, Chichester, PO19 8SQ, UK.
TERMS & CONDITIONS

This offer is exclusive to Wiley Authors, Editors, Contributors and Editorial Board Members in acquiring books for their personal use. There should
be no resale through any channel. The offer is subject to stock availability and may not be applied retrospectively. This entitlement cannot be used
in conjunction with any other special offer. Wiley reserves the right to vary the terms of the offer at any time.
PLEASE RETURN THIS FORM TO:

Database Group (Author Club), John Wiley & Sons Ltd, The Atrium, Southern Gate, Chichester, PO19 8SQ, UK author@wiley.co.uk
Fax: +44 (0)1243 770154

Uncorrected Proofs: Proof of Your Article Attached, Please Read Carefully

Caricato da

Informazioni sul documento

Descrizione originale:

Titolo originale

Copyright

Formati disponibili

Condividi questo documento

Condividi o incorpora il documento

Opzioni di condivisione

Hai trovato utile questo documento?

Questo contenuto è inappropriato?

Copyright:

Formati disponibili

Uncorrected Proofs: Proof of Your Article Attached, Please Read Carefully

Caricato da

Copyright:

Formati disponibili

JOHN WILEY & SONS, LTD.

, THE ATRIUM, SOUTHERN GATE, CHICHESTER P019 8SQ, UK

*** PROOF OF YOUR ARTICLE ATTACHED, PLEASE READ CAREFULLY ***

ONCE PUBLISHED ONLINE OR IN PRINT IT IS NOT POSSIBLE TO MAKE ANY FURTHER

This will be your only chance to correct your proof

Check sizes, numbering, and orientation of figures

unless we have received the signed CTA

Additional reprint and journal issue purchases

upon publication, quoting the article and volume/issue details

Strategic Management Journal

A BARGAINING PERSPECTIVE ON STRATEGIC

*Correspondence to: Joshua S. Gans, Melbourne Business

of property rights, property rights that can be reallocated in a

Copyright  2008 John Wiley & Sons, Ltd.

In this article, we show how a proper consid-

priation can resolve the trade-off between asset

41 tap into the competencies of established operators 98

favored established firms for component manufacturing; los-

ciently greater in that it compensates the outsourc-

previously discussed. Thus, the majority of the arti-

does not consider the impact of outsourcing on downstream

29 D2 . However, to obtain cleaner results, we may at 86

29 the decision of whether to outsource or not, but on 86

33 Timeline the use of one upstream asset—assumed here to 90

(D2 ) may or may not own the other upstream asset.

The timeline for our model is as follows:

Finally, we consider situations where D2 oper- 58

22 ing (should it occur) as an additional variable. 90

26 UA UB ), which is literally the maximal level of 94

28 employing both upstream and both downstream 96

30 its net of payments for inputs and Cj (q1j , q2j ) are 98

29 a single downstream asset supplied by both up- BASELINE RESULTS 86

33 status quo whereby D1 performs upstream oper- 90

single upstream and downstream asset (bilateral

39 established firm. In this regard, we initially assume 96

41 explore the implications of this assumption below. 98

ties of these coalitional values. First, adding assets

45 The focus of this article is how bargaining con- 102

1 To consider D1 ’s outsourcing decision, we con- Proposition 1: Regardless of whether it out- 30

29 Table 1. Shapley value outcomes 58

Ownership structure Payoffs (Shapley values)

vUB = 12 1 4(D1 D2 UA UB ) − 4(D1 D2 UA ) + 32 B

3(D1 D2 UA UB ) + (D1 D2 UA ) + (D1 D2 UB )

3(D1 D2 UA UB ) + (D1 D2 UA ) + (D1 D2 UB )

3(D1 D2 UA UB ) + (D1 D2 UA ) − 3(D1 D2 UB )

29 its ongoing profits post-acquisition less its prof- tivity. 86

returns to outsourcing. Supply scarcity could be

asset would add little to industry profits. On the

47 that D1 will find it profitable. Given Proposi- 104

29 would have to bid an amount, b, so that: 86

upstream supplier, competition is produced that D1 s Returns from Established Outsourcing

upstream entity. D1 s Returns from Independent Outsourcing

Will UB bid this amount? If it does, knowing that

37 bargaining position with respect to D2 . D1 were to outsource UA to an independent com- 94

37 model of bargaining between upstream and down- 94

47 to be present and to matter. This means that the 104

29 (2005) is that this rent dissipation is the same, 86

37 Similarly, suppose that in outsourcing to an D1 ’s incentive to outsource and break up versus 94

43 dent firm outsourcing as all downstream firms face 100

33 or alternatively, the setup here is restrictive on enough to consider a range of environments. 69

vUA = 12 1 4(D1 D2 UA UB ) − 4(D1 D2 UB ) +

29 upstream and downstream firms negotiate directly 86

* PROOF OF YOUR ARTICLE ATTACHED, PLEASE READ CAREFULLY *

29 where = a−α . In this case, the 86

37 vD2 (D1 UA , D2 , UB ) = 16b 94