Patrick Greenlee - Academia.edu (original) (raw)
Papers by Patrick Greenlee
Federal Regulatory Guide, 2020
Social Science Research Network, 2004
A backward ownership interest held by a downstream firm yields a partial rebate of the upstream m... more A backward ownership interest held by a downstream firm yields a partial rebate of the upstream margin. Input demand increases with backward ownership, and the upstream firm optimally responds by raising price. With symmetric costs, every downstream firm's equilibrium input/output choice is invariant across a class of ownership profiles, including uniform ownership. Moreover, equity trading results in uniform holdings, so partial vertical ownership may have no real effects. With asymmetric costs ex ante, equity trading amplifies the asymmetries and shifts output toward lower-cost firms. With homogeneous goods, this improves producer and total surplus. With differentiated goods, it may harm consumers.
Social Science Research Network, 2006
Consider a monopolist in one market that faces competition in a second market. Bundled loyalty di... more Consider a monopolist in one market that faces competition in a second market. Bundled loyalty discounts, in which customers receive a price break on the monopoly good in exchange for making all purchases from the monopolist, have ambiguous welfare effects. To analyze such discounts as predatory pricing is incorrect. In some settings, they act as tie-in sales. Existing tests for whether such discounts violate Section 2 of the Sherman Act do not track changes in consumer surplus or total surplus. We present a new test and use it in an illustrative example based on SmithKline that assumes the "tied" market is a homogeneous good. If the tied market is characterized by Hotelling competition, bundling by the monopolist causes the rival firm to reduce its price. In numerical examples, we find that this can deter entry or induce exit.
RePEc: Research Papers in Economics, 1999
Research sharing is an important objective of many research joint ventures. When partners share R... more Research sharing is an important objective of many research joint ventures. When partners share R&D but do not maximize joint profits, large consortia are more profitable than small ones, and joint ventures prefer dispersed rivals. For much of the spillover space, a coalition formation game that permits limited membership predicts that at most, three joint ventures form. Research-sharing joint ventures improve welfare when spillovers are low, and banning research sharing joint ventures is beneficial for high spillovers. With imperfect research sharing and low spillovers, allowing only research sharing is the best industry-wide joint venture alternative for consumer surplus.
Social Science Research Network, 2000
Research sharing is an important objective of many research joint ventures. When partners share R... more Research sharing is an important objective of many research joint ventures. When partners share R&D but do not maximize joint profits, large consortia are more profitable than small ones, and joint ventures prefer dispersed rivals. For much of the spillover space, a coalition formation game that permits limited membership predicts that at most, three joint ventures form. Research-sharing joint ventures improve welfare when spillovers are low, and banning research sharing joint ventures is beneficial for high spillovers. With imperfect research sharing and low spillovers, allowing only research sharing is the best industry-wide joint venture alternative for consumer surplus.
RePEc: Research Papers in Economics, Aug 11, 2004
We consider a model where bidders in an auction own passive partial claims over their rivals’ auc... more We consider a model where bidders in an auction own passive partial claims over their rivals’ auction profits. While the cross ownership confers no ability to directly affect bidding behavior, the claims on rival profits dampen bidding competition. It is not uncommon for enforcement agencies to investigate such an arrangements as a possible violation of antitrust laws. When bidders hold the same aggregate level of shares in rival bidders in a first-price independent private-value auction, we find that such cross ownership has an effect similar to reducing the number of bidders while holding constant the distribution of the highest value in the auction. A similar decrease in competition occurs in independent private-value second-price auctions with two bidders and in first-price complete information auctions. In each of these cases, the competitive effect disappears as the level of cross ownership goes to zero. English auctions in complete information environments, however, generate strikingly different outcomes. There, for any positive level of cross ownership, the unique subgame perfect equilibrium generates the same result as the perfectly collusive outcome: the highest valued bidder wins the object at the lowest possible price
RePEc: Research Papers in Economics, Jul 1, 2008
Virginia for useful comments. The views contained herein are the authors' and do not necessarily ... more Virginia for useful comments. The views contained herein are the authors' and do not necessarily reflect the views of the U.S. Department of Justice. The views expressed herein are those of the author(s) and do not necessarily reflect the views of the National Bureau of Economic Research. NBER working papers are circulated for discussion and comment purposes. They have not been peerreviewed or been subject to the review by the NBER Board of Directors that accompanies official NBER publications.
Social Science Research Network, 2017
Conditional pricing practices are pricing strategies in which a seller conditions its prices on f... more Conditional pricing practices are pricing strategies in which a seller conditions its prices on factors such as volume, the set of products purchased, or the buyer's share of purchases from the seller. This short primer provides a unifying overview of the economic literature that addresses these practices.
Managerial and Decision Economics, May 1, 1999
In this paper we extend the existing literature on research and development (R&D) investments... more In this paper we extend the existing literature on research and development (R&D) investments and research joint ventures (RJVs) in two important ways. First, we analyze and compare the case where firms collude in the product market to the benchmark case of competition in the output market. Second, we allow firms to form coalitions endogenously as a separate stage in the game. We develop profit functions that depend on the partition of firms into joint ventures and the nature of product competition between venture partners. Our results illustrate the restrictive nature of some assumptions made in the literature. Typically multiple RJVs of different sizes form in equilibrium. In general, RJVs should not be promoted if they entail product market collusion. Given the information available to policy-makers, it is unlikely that an R&D policy more refined than analyzing and allowing RJVs on a case-by-case basis is feasible. Copyright © 1999 John Wiley & Sons, Ltd.
European Economic Review, May 1, 2006
A backward ownership interest held by a downstream firm yields a partial rebate of the upstream m... more A backward ownership interest held by a downstream firm yields a partial rebate of the upstream margin. Input demand increases with backward ownership, and the upstream firm optimally responds by raising price. With symmetric costs, every downstream firm's equilibrium input/output choice is invariant across a class of ownership profiles, including uniform ownership. Moreover, equity trading results in uniform holdings, so partial vertical ownership may have no real effects. With asymmetric costs ex ante, equity trading amplifies the asymmetries and shifts output toward lower-cost firms. With homogeneous goods, this improves producer and total surplus. With differentiated goods, it may harm consumers.
SSRN Electronic Journal, 2004
EAG Discussion Papers are the primary vehicle used to disseminate research from economists in the... more EAG Discussion Papers are the primary vehicle used to disseminate research from economists in the Economic Analysis Group (EAG) of the Antitrust Division. These papers are intended to inform interested individuals and institutions of EAG's research program and to stimulate comment and criticism on economic issues related to antitrust policy and regulation. The analysis and conclusions of EAG Discussion Papers are solely those of the author(s) and do not represent the views of the United States Department of Justice.
We consider a model where bidders in an auction own passive partial claims over their rivals’ auc... more We consider a model where bidders in an auction own passive partial claims over their rivals’ auction profits. While the cross ownership confers no ability to directly affect bidding behavior, the claims on rival profits dampen bidding competition. It is not uncommon for enforcement agencies to investigate such an arrangements as a possible violation of antitrust laws. When bidders hold the same aggregate level of shares in rival bidders in a first-price independent private-value auction, we find that such cross ownership has an effect similar to reducing the number of bidders while holding constant the distribution of the highest value in the auction. A similar decrease in competition occurs in independent private-value second-price auctions with two bidders and in first-price complete information auctions. In each of these cases, the competitive effect disappears as the level of cross ownership goes to zero. English auctions in complete information environments, however, generate strikingly different outcomes. There, for any positive level of cross ownership, the unique subgame perfect equilibrium generates the same result as the perfectly collusive outcome: the highest valued bidder wins the object at the lowest possible price
SSRN Electronic Journal, 2001
When a monopolist sells an input to an oligopoly, consumer and total surplus frequently are invar... more When a monopolist sells an input to an oligopoly, consumer and total surplus frequently are invariant to changes in passive ownership of the monopolist by downstream firms. Within broad classes of ownership profiles, strong invariance holds: the input and output choices of downstream firms are invariant to a change within the class. While passive ownership raises input demand, the upstream firm responds by raising price. Strong invariance always holds for bilateral monopoly. For a broad class of models with fixed proportions technologies, aggregate output is invariant across all passive ownership profiles. Passive ownership is privately profitable because it shifts sales from rivals.
SSRN Electronic Journal, 2005
Loyalty discounts, offered to customers that meet purchase thresholds, can shift share from rival... more Loyalty discounts, offered to customers that meet purchase thresholds, can shift share from rival firms. They emerge as equilibrium strategies in each of three types of market structure that we study. In a differentiated single product duopoly, only one firm employs a loyalty program that customers adopt in equilibrium. Such discounts increase producer surplus whenever consumers strongly prefer the product of the firm whose loyalty program is satisfied. When a firm requires customers to meet thresholds in multiple rivalrous markets, the equilibrium loyalty discount and consumer surplus may fall. We identify a test to distinguish competitive and exclusionary conduct based on the characterization of equilibrium in these two market structures. Finally, when a "branded" product monopolist faces competition for a separate "generic" product, the monopolist in equilibrium raises the branded spot price and offers a loyalty discount to customers that purchase generic product only from the monopolist.
SSRN Electronic Journal, 2017
Conditional pricing practices are pricing strategies in which a seller conditions its prices on f... more Conditional pricing practices are pricing strategies in which a seller conditions its prices on factors such as volume, the set of products purchased, or the buyer's share of purchases from the seller. This short primer provides a unifying overview of the economic literature that addresses these practices.
The Antitrust Bulletin, 2005
Loyalty rebates have sprung into prominence as a potential anticompetitive practice by dominant f... more Loyalty rebates have sprung into prominence as a potential anticompetitive practice by dominant firms due to several recent cases on both sides of the Atlantic. Examples include Michelin' in Europe, LePage'5 2 in the United States, and Virgin-BN in both. While courts have grappled with the potential exclusionary effects of loyalty rebates, the economics literature until recently has lagged somewhat behind in examining loyalty rebates as a competitive strategy. Thus the backdrop of legitimate pricing strategy against which courts must attempt to discern exclusionary behavior has been correspondingly murky. In their simplest form, loyalty rebates offer a reward, in the form of a rebate, free products, or lower prices, to customers that make suf-" Economists,
In a recent issue of Antitrust Law Journal, Timothy J. Muris and Vernon Smith summarize laborator... more In a recent issue of Antitrust Law Journal, Timothy J. Muris and Vernon Smith summarize laboratory experiments performed by Smith and several co-authors that were reported in Caliskan et al. 1 Muris and Smith claim that these experiments demonstrate that anticompetitive effects from bundling are un-likely. Consequently, they argue that "regulation of bundling based upon theo-retical models of exclusionary harm is premature and misguided." 2 In discussing theoretical models of bundling, Muris and Smith refer primarily to our earlier paper as well as several papers by Barry Nalebuff. 3 We welcome empirical work that throws light on the applicability of vari-ous theories. However, the experimental design implemented by Caliskan et al. makes key assumptions that do not match the framework of our earlier paper or Nalebuff's. Therefore, the inferences that Muris and Smith attempt to draw from the Caliskan et al. results have very little to say about theoretical, Exclusionary...
Virginia for useful comments. The views contained herein are the authors' and do not necessarily ... more Virginia for useful comments. The views contained herein are the authors' and do not necessarily reflect the views of the U.S. Department of Justice. The views expressed herein are those of the author(s) and do not necessarily reflect the views of the National Bureau of Economic Research. NBER working papers are circulated for discussion and comment purposes. They have not been peerreviewed or been subject to the review by the NBER Board of Directors that accompanies official NBER publications.
Journal of Industrial Economics, 2005
Research sharing is an important objective of many research joint ventures. When partners share R... more Research sharing is an important objective of many research joint ventures. When partners share R&D but do not maximize joint profits, large consortia are more profitable than small ones, and joint ventures prefer dispersed rivals. For much of the spillover space, a coalition formation game that permits limited membership predicts that at most, three joint ventures form. Research-sharing joint ventures improve welfare when spillovers are low, and banning research sharing joint ventures is beneficial for high spillovers. With imperfect research sharing and low spillovers, allowing only research sharing is the best industry-wide joint venture alternative for consumer surplus.
Managerial and Decision Economics, 1999
In this paper we extend the existing literature on research and development (R&D) investments... more In this paper we extend the existing literature on research and development (R&D) investments and research joint ventures (RJVs) in two important ways. First, we analyze and compare the case where firms collude in the product market to the benchmark case of competition in the output market. Second, we allow firms to form coalitions endogenously as a separate stage in the game. We develop profit functions that depend on the partition of firms into joint ventures and the nature of product competition between venture partners. Our results illustrate the restrictive nature of some assumptions made in the literature. Typically multiple RJVs of different sizes form in equilibrium. In general, RJVs should not be promoted if they entail product market collusion. Given the information available to policy-makers, it is unlikely that an R&D policy more refined than analyzing and allowing RJVs on a case-by-case basis is feasible. Copyright © 1999 John Wiley & Sons, Ltd.
Federal Regulatory Guide, 2020
Social Science Research Network, 2004
A backward ownership interest held by a downstream firm yields a partial rebate of the upstream m... more A backward ownership interest held by a downstream firm yields a partial rebate of the upstream margin. Input demand increases with backward ownership, and the upstream firm optimally responds by raising price. With symmetric costs, every downstream firm's equilibrium input/output choice is invariant across a class of ownership profiles, including uniform ownership. Moreover, equity trading results in uniform holdings, so partial vertical ownership may have no real effects. With asymmetric costs ex ante, equity trading amplifies the asymmetries and shifts output toward lower-cost firms. With homogeneous goods, this improves producer and total surplus. With differentiated goods, it may harm consumers.
Social Science Research Network, 2006
Consider a monopolist in one market that faces competition in a second market. Bundled loyalty di... more Consider a monopolist in one market that faces competition in a second market. Bundled loyalty discounts, in which customers receive a price break on the monopoly good in exchange for making all purchases from the monopolist, have ambiguous welfare effects. To analyze such discounts as predatory pricing is incorrect. In some settings, they act as tie-in sales. Existing tests for whether such discounts violate Section 2 of the Sherman Act do not track changes in consumer surplus or total surplus. We present a new test and use it in an illustrative example based on SmithKline that assumes the "tied" market is a homogeneous good. If the tied market is characterized by Hotelling competition, bundling by the monopolist causes the rival firm to reduce its price. In numerical examples, we find that this can deter entry or induce exit.
RePEc: Research Papers in Economics, 1999
Research sharing is an important objective of many research joint ventures. When partners share R... more Research sharing is an important objective of many research joint ventures. When partners share R&D but do not maximize joint profits, large consortia are more profitable than small ones, and joint ventures prefer dispersed rivals. For much of the spillover space, a coalition formation game that permits limited membership predicts that at most, three joint ventures form. Research-sharing joint ventures improve welfare when spillovers are low, and banning research sharing joint ventures is beneficial for high spillovers. With imperfect research sharing and low spillovers, allowing only research sharing is the best industry-wide joint venture alternative for consumer surplus.
Social Science Research Network, 2000
Research sharing is an important objective of many research joint ventures. When partners share R... more Research sharing is an important objective of many research joint ventures. When partners share R&D but do not maximize joint profits, large consortia are more profitable than small ones, and joint ventures prefer dispersed rivals. For much of the spillover space, a coalition formation game that permits limited membership predicts that at most, three joint ventures form. Research-sharing joint ventures improve welfare when spillovers are low, and banning research sharing joint ventures is beneficial for high spillovers. With imperfect research sharing and low spillovers, allowing only research sharing is the best industry-wide joint venture alternative for consumer surplus.
RePEc: Research Papers in Economics, Aug 11, 2004
We consider a model where bidders in an auction own passive partial claims over their rivals’ auc... more We consider a model where bidders in an auction own passive partial claims over their rivals’ auction profits. While the cross ownership confers no ability to directly affect bidding behavior, the claims on rival profits dampen bidding competition. It is not uncommon for enforcement agencies to investigate such an arrangements as a possible violation of antitrust laws. When bidders hold the same aggregate level of shares in rival bidders in a first-price independent private-value auction, we find that such cross ownership has an effect similar to reducing the number of bidders while holding constant the distribution of the highest value in the auction. A similar decrease in competition occurs in independent private-value second-price auctions with two bidders and in first-price complete information auctions. In each of these cases, the competitive effect disappears as the level of cross ownership goes to zero. English auctions in complete information environments, however, generate strikingly different outcomes. There, for any positive level of cross ownership, the unique subgame perfect equilibrium generates the same result as the perfectly collusive outcome: the highest valued bidder wins the object at the lowest possible price
RePEc: Research Papers in Economics, Jul 1, 2008
Virginia for useful comments. The views contained herein are the authors' and do not necessarily ... more Virginia for useful comments. The views contained herein are the authors' and do not necessarily reflect the views of the U.S. Department of Justice. The views expressed herein are those of the author(s) and do not necessarily reflect the views of the National Bureau of Economic Research. NBER working papers are circulated for discussion and comment purposes. They have not been peerreviewed or been subject to the review by the NBER Board of Directors that accompanies official NBER publications.
Social Science Research Network, 2017
Conditional pricing practices are pricing strategies in which a seller conditions its prices on f... more Conditional pricing practices are pricing strategies in which a seller conditions its prices on factors such as volume, the set of products purchased, or the buyer's share of purchases from the seller. This short primer provides a unifying overview of the economic literature that addresses these practices.
Managerial and Decision Economics, May 1, 1999
In this paper we extend the existing literature on research and development (R&D) investments... more In this paper we extend the existing literature on research and development (R&D) investments and research joint ventures (RJVs) in two important ways. First, we analyze and compare the case where firms collude in the product market to the benchmark case of competition in the output market. Second, we allow firms to form coalitions endogenously as a separate stage in the game. We develop profit functions that depend on the partition of firms into joint ventures and the nature of product competition between venture partners. Our results illustrate the restrictive nature of some assumptions made in the literature. Typically multiple RJVs of different sizes form in equilibrium. In general, RJVs should not be promoted if they entail product market collusion. Given the information available to policy-makers, it is unlikely that an R&D policy more refined than analyzing and allowing RJVs on a case-by-case basis is feasible. Copyright © 1999 John Wiley & Sons, Ltd.
European Economic Review, May 1, 2006
A backward ownership interest held by a downstream firm yields a partial rebate of the upstream m... more A backward ownership interest held by a downstream firm yields a partial rebate of the upstream margin. Input demand increases with backward ownership, and the upstream firm optimally responds by raising price. With symmetric costs, every downstream firm's equilibrium input/output choice is invariant across a class of ownership profiles, including uniform ownership. Moreover, equity trading results in uniform holdings, so partial vertical ownership may have no real effects. With asymmetric costs ex ante, equity trading amplifies the asymmetries and shifts output toward lower-cost firms. With homogeneous goods, this improves producer and total surplus. With differentiated goods, it may harm consumers.
SSRN Electronic Journal, 2004
EAG Discussion Papers are the primary vehicle used to disseminate research from economists in the... more EAG Discussion Papers are the primary vehicle used to disseminate research from economists in the Economic Analysis Group (EAG) of the Antitrust Division. These papers are intended to inform interested individuals and institutions of EAG's research program and to stimulate comment and criticism on economic issues related to antitrust policy and regulation. The analysis and conclusions of EAG Discussion Papers are solely those of the author(s) and do not represent the views of the United States Department of Justice.
We consider a model where bidders in an auction own passive partial claims over their rivals’ auc... more We consider a model where bidders in an auction own passive partial claims over their rivals’ auction profits. While the cross ownership confers no ability to directly affect bidding behavior, the claims on rival profits dampen bidding competition. It is not uncommon for enforcement agencies to investigate such an arrangements as a possible violation of antitrust laws. When bidders hold the same aggregate level of shares in rival bidders in a first-price independent private-value auction, we find that such cross ownership has an effect similar to reducing the number of bidders while holding constant the distribution of the highest value in the auction. A similar decrease in competition occurs in independent private-value second-price auctions with two bidders and in first-price complete information auctions. In each of these cases, the competitive effect disappears as the level of cross ownership goes to zero. English auctions in complete information environments, however, generate strikingly different outcomes. There, for any positive level of cross ownership, the unique subgame perfect equilibrium generates the same result as the perfectly collusive outcome: the highest valued bidder wins the object at the lowest possible price
SSRN Electronic Journal, 2001
When a monopolist sells an input to an oligopoly, consumer and total surplus frequently are invar... more When a monopolist sells an input to an oligopoly, consumer and total surplus frequently are invariant to changes in passive ownership of the monopolist by downstream firms. Within broad classes of ownership profiles, strong invariance holds: the input and output choices of downstream firms are invariant to a change within the class. While passive ownership raises input demand, the upstream firm responds by raising price. Strong invariance always holds for bilateral monopoly. For a broad class of models with fixed proportions technologies, aggregate output is invariant across all passive ownership profiles. Passive ownership is privately profitable because it shifts sales from rivals.
SSRN Electronic Journal, 2005
Loyalty discounts, offered to customers that meet purchase thresholds, can shift share from rival... more Loyalty discounts, offered to customers that meet purchase thresholds, can shift share from rival firms. They emerge as equilibrium strategies in each of three types of market structure that we study. In a differentiated single product duopoly, only one firm employs a loyalty program that customers adopt in equilibrium. Such discounts increase producer surplus whenever consumers strongly prefer the product of the firm whose loyalty program is satisfied. When a firm requires customers to meet thresholds in multiple rivalrous markets, the equilibrium loyalty discount and consumer surplus may fall. We identify a test to distinguish competitive and exclusionary conduct based on the characterization of equilibrium in these two market structures. Finally, when a "branded" product monopolist faces competition for a separate "generic" product, the monopolist in equilibrium raises the branded spot price and offers a loyalty discount to customers that purchase generic product only from the monopolist.
SSRN Electronic Journal, 2017
Conditional pricing practices are pricing strategies in which a seller conditions its prices on f... more Conditional pricing practices are pricing strategies in which a seller conditions its prices on factors such as volume, the set of products purchased, or the buyer's share of purchases from the seller. This short primer provides a unifying overview of the economic literature that addresses these practices.
The Antitrust Bulletin, 2005
Loyalty rebates have sprung into prominence as a potential anticompetitive practice by dominant f... more Loyalty rebates have sprung into prominence as a potential anticompetitive practice by dominant firms due to several recent cases on both sides of the Atlantic. Examples include Michelin' in Europe, LePage'5 2 in the United States, and Virgin-BN in both. While courts have grappled with the potential exclusionary effects of loyalty rebates, the economics literature until recently has lagged somewhat behind in examining loyalty rebates as a competitive strategy. Thus the backdrop of legitimate pricing strategy against which courts must attempt to discern exclusionary behavior has been correspondingly murky. In their simplest form, loyalty rebates offer a reward, in the form of a rebate, free products, or lower prices, to customers that make suf-" Economists,
In a recent issue of Antitrust Law Journal, Timothy J. Muris and Vernon Smith summarize laborator... more In a recent issue of Antitrust Law Journal, Timothy J. Muris and Vernon Smith summarize laboratory experiments performed by Smith and several co-authors that were reported in Caliskan et al. 1 Muris and Smith claim that these experiments demonstrate that anticompetitive effects from bundling are un-likely. Consequently, they argue that "regulation of bundling based upon theo-retical models of exclusionary harm is premature and misguided." 2 In discussing theoretical models of bundling, Muris and Smith refer primarily to our earlier paper as well as several papers by Barry Nalebuff. 3 We welcome empirical work that throws light on the applicability of vari-ous theories. However, the experimental design implemented by Caliskan et al. makes key assumptions that do not match the framework of our earlier paper or Nalebuff's. Therefore, the inferences that Muris and Smith attempt to draw from the Caliskan et al. results have very little to say about theoretical, Exclusionary...
Virginia for useful comments. The views contained herein are the authors' and do not necessarily ... more Virginia for useful comments. The views contained herein are the authors' and do not necessarily reflect the views of the U.S. Department of Justice. The views expressed herein are those of the author(s) and do not necessarily reflect the views of the National Bureau of Economic Research. NBER working papers are circulated for discussion and comment purposes. They have not been peerreviewed or been subject to the review by the NBER Board of Directors that accompanies official NBER publications.
Journal of Industrial Economics, 2005
Research sharing is an important objective of many research joint ventures. When partners share R... more Research sharing is an important objective of many research joint ventures. When partners share R&D but do not maximize joint profits, large consortia are more profitable than small ones, and joint ventures prefer dispersed rivals. For much of the spillover space, a coalition formation game that permits limited membership predicts that at most, three joint ventures form. Research-sharing joint ventures improve welfare when spillovers are low, and banning research sharing joint ventures is beneficial for high spillovers. With imperfect research sharing and low spillovers, allowing only research sharing is the best industry-wide joint venture alternative for consumer surplus.
Managerial and Decision Economics, 1999
In this paper we extend the existing literature on research and development (R&D) investments... more In this paper we extend the existing literature on research and development (R&D) investments and research joint ventures (RJVs) in two important ways. First, we analyze and compare the case where firms collude in the product market to the benchmark case of competition in the output market. Second, we allow firms to form coalitions endogenously as a separate stage in the game. We develop profit functions that depend on the partition of firms into joint ventures and the nature of product competition between venture partners. Our results illustrate the restrictive nature of some assumptions made in the literature. Typically multiple RJVs of different sizes form in equilibrium. In general, RJVs should not be promoted if they entail product market collusion. Given the information available to policy-makers, it is unlikely that an R&D policy more refined than analyzing and allowing RJVs on a case-by-case basis is feasible. Copyright © 1999 John Wiley & Sons, Ltd.