|
Article Excerpt
1. Introduction
The governance of interfirm collaboration is the subject of large body of research that spans multiple disciplinary areas. A recurring focus in this work is on the roles played by contracts in these settings, both as project planning tools and as external enforcement devices. To the extent that such agreements facilitate smooth collaboration and avoid costly misunderstandings, they enhance the performance of the interfirm activities they govern. (1) In this sense, contractual content is a managerial choice variable with strategic implications.
Recently, a growing stream of literature has considered the issue of contract design in repeated relationships. (2) Collaborators in repeated partnerships experience shared learning, build trust, establish reputations, and use relational mechanisms to facilitate successful outcomes. Thus, anticipation of an ongoing relationship should influence the design of the formal contract. Although some progress has been made in understanding the individual effects of these factors, little is presently known about their joint, interactive effects.
Our purpose in this paper is to contribute new empirical findings on formal contract design when firms repeatedly interact, exploring in greater depth the specific details of a particular class of contracts and analyzing how these details vary when relational considerations are at play. Specifically, we examine 52 contracts for joint technology development in the telecommunications equipment industry to determine whether and how their content is affected by relational considerations. Three categories of salient terms emerge from our analysis: those that (i) define the rights and obligations of the parties ("contract detail"), (ii) facilitate observation of performance ("monitoring"), and (iii) provide enforcement prescriptions ("penalties"). We investigate whether the inclusion of term types varies with the relational status and deal experience of the partners. We find that contracts become more detailed and include stronger enforcement terms when at least one of the firms has prior deal experience. These effects are stronger when collaborating firms have prior relationships with each other.
Our study sits at the intersection of two important streams of literature, one in economics and the other in management. On the economics side, an extensive body of formal theory examines various aspects of both relational and formal contracting (Baker et al. 2002, is a central contribution). (3) Recent work on the interplay between these two classes of mechanisms identifies a number of theoretical interactions between them--from substitutes to complements to outright collision (e.g., Bernheim and Whinston 1998, Levin 2003, Kvaloy and Olsen 2005). In contrast, the relevant management literature presents a rich body of empirical findings on contracting in a variety of settings (e.g., Mayer and Argyres 2004, Argyres et al. 2007, Poppo and Zenger 2002). (4) These contributions cover a wide spectrum and, not infrequently, offer alternative explanations to those provided in the economics literature.
Given the breadth and variation in prior findings, both theoretical and empirical, we do not test any specific model, but instead take an inductive approach by providing descriptive results of direct relevance to several of the key ideas put forth in each of these streams. This has implications for how we organize our paper: We lead with our empirical analysis and then follow up with a careful discussion relating our findings to the prominent ideas in both management and economics. It is in this discussion that we review the relevant literatures--both theoretical and empirical--and contrast them against what we deem our most interesting results. We discover both consistencies and gaps between our results and existing theories, with the latter suggesting several potentially fruitful directions for theoretical development. To the best of our knowledge, our examination of monitoring and penalty clauses with respect to prior deal experience is entirely new.
In technical terms, the approach taken here contributes to the existing empirical literature on several dimensions. Our categorization scheme goes beyond the discrete-form analysis that has been a hallmark of the literature examining, for example, alliance structure choices (Gulati 1995, Oxley 1997, Sampson 2004). In this sense, we add to the growing stream analyzing how contract terms vary with deal-specific factors (e.g., Crocker and Reynolds 1993, Lerner and Merges 1998, Mayer and Argyres 2004, Argyres et al. 2007). By proceeding inductively, we generate a broad set of terms that permits more comprehensive comparisons than traditionally available in studies of this kind. Moreover, our approach uncovers several ancillary, previously unreported stylized facts--most notably that the majority of firms in our sample waive their rights to courtroom dispute resolution. This observation suggests that the role of contracts goes beyond the legal function of providing evidence for third party dispute resolution and is further support for the claim that contracts may influence deal performance. In this sense, we take a more holistic approach to examining contracts in our context with a view to providing some empirical regularities that may be used to advance theory development in the area. Our hope is that this broader design may provide a useful guideline in future studies.
This paper proceeds as follows. Section 2 describes our sample and the contracts in detail. In [section]3, we empirically examine the relationship between contract structure and repeated interactions. In [section]4, we contrast our results with existing work--both empirical and theoretical--in some detail. Concluding comments are presented in [section]5.
2. Sample and Description of Contracts
Technology development across firm boundaries is a fertile environment for enforceability problems in formal agreements. Often, research and development (R&D) projects are passed back and forth between firms at various stages, making it difficult to observe the behavior of one's partner. Moreover, project outcomes can be highly uncertain, making it difficult to infer unobserved partner performance (see, e.g., Holmstrom 1989). Unintended leakage of intellectual property may also be an issue in this context. The contracts in our sample capture the attempts of partnering firms to address these issues.
2.1. The Contract Sample
Our source of joint technology development contracts is SEC filings. Public firms, under SEC disclosure requirements, submit "material contracts" as part of their 8K, 10K, 10Q, and S-l filings, including joint technology development contracts. To obtain the largest sample possible, our contracts are collected for the years 1991 to 2000, inclusive. Because the SEC requirement is to file material documents and not joint development contracts specifically, filing of these contracts is somewhat discretionary. Thus, the limitation of this data is that we observe only contracts of public companies and likely only the largest and most important deals for these firms.
We collected contracts in the telecommunications equipment and microelectronics industries. The convergence of the telecommunications equipment with computer and microelectronics markets in the late 1980s substantially accelerated the pace of technological development (e.g., The Economist 1997). Product life cycles shortened while the cost of development increased. To gain access to different technologies, realize economies of scale in R&D, and spread the risk and expense of development, firms in these industries frequently collaborated in their R&D activities. (5)
The firms included in our sample range in size from the largest players in each industry, such as Motorola and IBM, to much smaller, more specialized firms like Global Village Communication and Positron Fiber Systems. (6) Collaborations take many forms, including cross-licensing arrangements, joint technology development agreements, and formal joint ventures for development and manufacturing. Projects in our sample focus on end-product- or manufacturing-process-driven R&D.
To collect a sample of contracts, we began with a list of firms with primary operations in SIC 366 or 367 from Compustat (approximately 625 firms). From this list, we downloaded all SEC filings from the EDGAR database and supplemented these filings with those from Disclosure, Inc. (a company that provides SEC filings on request for a fee). (7) From approximately 12,500 filings, we obtained over 120 technology development contracts. Screening for deals involving some form of joint development reduced our sample size to 52 contracts. Thus, we explicitly exclude those deals that are solely licensing arrangements. Licensing arrangements--typically, defined technology transfers or usage agreements with royalties attached--are an interesting class of contracts to consider, but do not involve the same contracting issues as joint development agreements between firms. Because our focus is on the contract heterogeneity that may result from repeated interactions, we try to control for other sources of heterogeneity by limiting our scope of examination and, thus, excluding licensing contracts. This approach is similar to that taken by earlier research on contract terms (e.g., Robinson and Stuart 2007).
These 52 contracts involve 78 organizations, when we count both parties to the deal. However, the number of firms for which we have the SEC filings is 36; that is, each of these 36 firms represents one party to the contract. The number of sample contracts per firm is set out in Table 1. The majority of firms in the sample (78%) have only one contract. The remaining firms have anywhere from two to five contracts in the sample.
Table 1 Number of Sample Contracts per Firm
Contracts per firm Frequency Percentage (%)
1 28 78
2 4 11
3 1 3
4 2 6
5 1 3
Total 36 100
Note. Firms included are only those with SEC filings.
These deals, even when confined to those for joint technological development, cover a broad spectrum of purposes, from development of new microprocessor cores based on existing technology to developing a "next-generation" ferroelectric chip. For more information on these sample deals, see the online supplement (provided in the e-companion). (8)
For data on the frequency of deals between firms, we relied upon the Securities Data Company (SDC) Database on Alliances and Joint Ventures. The SDC database compiles information on a firm's alliance activity from news reports, SEC filings, and industry and trade journals. Ideally, we would measure the full contracting network among firms in our sample to better assess reputational mechanisms that may be at work; unfortunately, a full sample of all contracts and deals for each firm in the sample is not available. Although coverage by SDC is inevitably incomplete, because firms are not required to report alliance activity, this database represents one of the most comprehensive sources of information on interfirm deals. Using SDC data, we captured all repeated deal activity for a firm and broke this information down into two components: (1) prior deals for a firm where the partners are the same as partners in our sample contracts for that firm, and (2) prior deals for a firm where partners differ from those in our sample contracts for the firm.
We did not discriminate between the types of prior deals; that is, we did not differentiate among prior deals for manufacturing, marketing, technology licensing, or R&D, because our measure is intended to capture the presence of relational mechanisms, which we expect may develop over prior interactions of different types (i.e., not just R&D). Furthermore, we expect that firms gain experience in contracting or with partnering (either with a specific partner or more generally) from all types of deals. For example, it appears that for R&D collaborations, prior experience of any type is positively correlated with performance (Sampson 2005).
Information on prior deals was collected for the years 1985 to present. Note that our prior deals measures are time variant. We attempt to capture experience at the time the contract was drafted. Thus, we only count those prior deals that occurred prior to the contract signing date. We supplement this data with financial information (including R&D spending) where available from Compustat, collected for the year prior to the contract date.
2.2. Descriptive Overview
Several patterns are revealed in the descriptive summary of the joint technology deals and partner firm characteristics presented in Table 2. Most of the joint development deals involve R&D only (i.e., no manufacturing or marketing). A slight majority are cross-border deals, involving firms headquartered in different countries. The average reported deal value is about $91 M.
Table 2 Deal and Firm Descriptives
N Mean SD Min Max
Average deal value (in 17 91,147 178.198 1 700
millions)
Technology breadth (0 to 52 0.404 0.495 1
1, with 1 being next
generation)
Average deal length in 47 3.426 7.711 50
years (where specified)
Prior deal exists--same 52 0.096 0.298 1
partner (dummy variable)
Contemporaneous deal 52 0.173 0.382 1
exists--same partner
(dummy variable)
Prior deal 52 0.808 0.398 1
exists--different partner
(dummy variable)
Average prior deal 52 31.480 61.454 370.5
activity for firm, not
including same partner
(since 1985)
Average prior deal 52 25.049 54.784 362.5
activity for firm, not
including same partner
(five years prior to
contract date)
Manufacturing (dummy 52 0.481 0.505 1
variable)
Marketing (dummy 52 0.192 0.398 1
variable)
Cross-border (dummy 52 0.558 0.502 1
variable)
Firms physically colocate 52 0.115 0.323 1
for joint development
(dummy variable)
Development is primarily 52 0.423 0.499 1
vertical (one firm for the
other) (dummy variable)
Equity stake exists 52 0.308 0.466 1
(cross, majority, or
minority equity)
(dummy variable)
Intellectual property 52 0.212 0.412 1
shared equally (dummy
variable)
Joint venture formed 52 0.115 0.323 1
(dummy variable)
Termination date set in 52 0.808 0.398 1
contract (dummy variable)
Firm size (total assets) 64 10,955.84 21,120.97 0.419 81,091.00
(in millions)
Firm size (net sales) (in 64 10,236.11 19,062.16 0.094 75,094,00
millions)
R&D spending (in 64 748.95 1,311.00 0.28 5,094,00
millions)
Note. Data for 78 organizations involved in 52 joint development deals
are shown.
Partners do not typically have a history together; we see repeat dealings with a specific partner in only ten contracts in our sample. However, over 80% of our firms do have prior deal experience of some kind with different partners. Colocation of joint development activities occurs in only 11.5% of our contracts. Equity holdings between partners (whether majority, minority, or cross-equity) occur in about one-third of the deals; joint venture formation (i.e., formation of a new firm in which each partner holds equity) occurs in just over 11% of the cases. Intellectual property is typically not shared equally, but is split between partners according to input contributions, each firm's area of technological expertise, and end-product market. Finally, firms tend to set a fixed termination date for the joint development, on average, just over three years from the contract date.
There are several contractual...
|
