Purchasing together or alone? Tradeoffs for information sharing

Joint purchasing (JP) agreements are negotiated amongst competing OEMs to enable them to obtain rebates from their supplier(s). The iterative interactions involved in reaching such JP agreements result in disclosure of private information that the OEMs might prefer to keep private by procuring individually rather than jointly. This paper investigates how the information sharing dimension affects (1) OEMs’ motivations towards joint purchasing agreements, and (2) consumer surplus, specifically in industries characterized by market demand and technology level uncertainties. Past literature has looked at competitors’ incentives in sharing information, however, there is a fine line between sharing information and collusion/anti-competitive practices. Thus it is important to understand when information sharing, which is inevitable in joint purchasing agreements, will be beneficial to both the OEMs and consumers (and hence likely to be acceptable to regulators). To answer these questions we investigate a multi-period duopoly model with Cournot competition in which OEMs decide on purchasing jointly (or not) in the first stage, and decide on their production quantity in the second stage, under uncertainty and information asymmetry on demand and product technology. We find that a joint purchasing agreement tends to be preferred by OEMs and is also beneficial to consumers when either the substitutability between the two OEMs products is sufficiently low; or when technology level uncertainty is in the intermediate range for a high level of product substitutability; or when the procurement cost saving through JP agreement is high enough. Otherwise, a JP agreement may harm either the OEMs or the consumers. For OEMs bringing periodical updates to the market we provide a framework to analyze how their product level technology uncertainty (i.e., the potential for the product be a “breakthrough”) can affect their strategic decision on purchasing jointly with a competitor.