Tacit Knowledge and the Structure of License Contracts: Evidence from the Biomedical Industry

• Author: Deepak Hegde
• DOI: http://doi.org/10.1111/jems.12060
• Date: 01 September 2014
• Published date: 01 September 2014

Tacit Knowledge and the Structure of License

Contracts: Evidence from the Biomedical Industry

DEEPAK HEGDE

Stern School of Business

New YorkUniversity

New York,NY 10012

dhegde@stern.nyu.edu

How do contracts deal with the exchange of tacit knowledge, which is difficult to observe,

enforce, and verify? This study compares the sample characteristics of 505 license agreements

between inventors and developers of biomedical inventions with theoretical prescriptions about

contractual terms when the parties’ tacit knowledge is required to commercialize the inventions.

The analysis reveals that (a) three- and four-part payments are common in licenses of biomedical

inventions; (b) the importance of inventors’ tacit knowledge in development activities is directly

related to the magnitude of royalty rates in the licenses; and (c) the importance of developers’

tacit knowledge in development is associated with minimum annual payments when the license

includes measurable benchmarks to address the shelving of inventions, and to upfront payments

in the absence of such benchmarks. Firms use a complex mix of upfront and state-contingent

payments, broadly consistent with theoretical prescriptions, to cope with the incentive issues

inherent in the exchange of tacit knowledge.

1. Introduction

The market for ideas involves the exchange of tacit knowledge. Tacit knowledge is hard

to specify in contracts, and the parties involved in its exchange (typically the inventors

of a new idea and those who seek to commercialize it) may have differing incentives to

apply the knowledge for the idea’s success. How do contracts deal with the exchange of

tacit knowledge?

Classic economic models recommend contracting upon tacit knowledge with two-

part payment schemes consisting of upfront fees and output-based royalty rates. Ac-

cording to the models, when inventors’ tacit knowledge is important for developing

the idea, the optimal contract specifies royalty rates, which tie the inventors’ income to

revenues; on the other hand, when developers’ tacit knowledge is critical for commer-

cialization, the contract specifies higher upfront payments (e.g., Arora, 1995; Choi, 2001;

Jensen and Thursby,2001). Recent models recommend more complex payment schemes

with output-contingent lump-sum payments to the inventor—such as milestone

I am grateful to Thomas Varner for data on license contracts and David C. Mowery for his guidance. I thank

Bronwyn Hall, Steven Tadelis, Brian Wright, Catherine Wolfram, Bhaven Sampat, Ernesto Dal Bo, Jennifer

Walske, Rui de Figueiredo, Pablo Spiller, Pedro Gardete, Justin Tumlinson, Mohan Turaga, Orie Shelef, and

seminar participants at the University of California, Berkeley, Johns Hopkins University, Emory University,

Universityof British Columbia, University of Toronto, New YorkUniversity and the University of Pennsylvania

for their suggestions. I am also indebted to three anonymous referees and the coeditor for a number of helpful

suggestions. The Kauffman Dissertation Fellowship program and the Bradley Foundation generously funded

this study.

C2014 Wiley Periodicals, Inc.

Journal of Economics & Management Strategy, Volume23, Number 3, Fall 2014, 568–600

Tacit Knowledge and License Contracts 569

payments and minimum annual fees—to address the distortive effect of royalty rates

(e.g., Crama et al., 2008; Dechenaux et al., 2009, 2011).

Despite the theoretical prescriptions, we know relatively little about how parties in

the real world deal with the exchange of tacit knowledge, because of the lack of contract-

level data.1To address this gap, I assemble a sample of 505 licenses between inventors

and developers of biomedical inventions and compare the contracts’ characteristics with

those of their theoretical counterparts. The licenses, which were negotiated between 1995

and 2008, were reported as material transactions by public corporations in their U.S.

Securities and Exchange Commission (SEC) filings. These high-value transactions—the

mean upfront payment to inventors is $1.5 million in my sample—among sophisticated

parties are likely to adopt contractual safeguards against information problems and,

thus, be informative regarding the practices adopted by value-maximizing agents.

License agreements between inventors and developers in the biomedical industry

provide an ideal context in which to assess the implications of contract-design theories

because (i) the development process for new drugs and devices is expensive, lengthy,

and uncertain (according to DiMasi et al., 2003, the mean drug-development process

costs $350 million, takes 7.5 years, and has a success probability of 18%);2(ii) inventors

of drugs and medical devices often lack downstream capabilities and license their inven-

tions to specialist developers who apply their tacit knowledge in downstream activities

(according to Pharmaprojects, a leading drug-development database, 51% of the drugs

launched during 1981–2008 were associated with at least one license agreement); (iii)

often, the commercial success of inventions depends on the inventor’s application of

tacit knowledge in downstream activities (according to a survey of biomedical licensors

and licensees by the Licensing Executive Society, 2008, 70% of licensed biomedical in-

ventions require significant redesign and development and, thus, communication with

the inventor after the agreement date); and (iv) patents are known to effectively facilitate

the exchange of ideas in the biomedical industry (Cohen et al., 2000), thus allowing me

to focus on the effects of tacit knowledge on license structure, rather than on issues such

as expropriation.

A key challenge for testing the contractual implications of noncontractible effort is

that they are identical to those of another,frequently coincident, asymmetric information

problem: private information about the ideas’ value, which leads to adverse selection

(Hagerty and Siegel, 1988). According to contract-design models, privately informed

inventors can signal their ideas’ superior value by offering high royalty rates and low

upfront fees, whereas privately informed developers compensate superior ideas with

high upfront fees (e.g., Gallini and Wright, 1990; Beggs, 1992). Hence, it is plausible that

when a party’s tacit knowledge is important for development, the party is also privately

1. A small number of recent studies analyze large samples of license contracts. For example, Anand and

Khanna (2000) describe interindustry differences in the structure of licenses; Elfenbein (2009) investigates the

relationship between the presence of differentpayment terms and the probability of license termination using

a sample of Harvard University licenses; Macho-Stadler et al. (1996) and Kotha et al. (2011) analyze licenses

and report that when inventor assistance is required for development, licenses are more likely to specify

performance-based payments to the inventor. However, the above studies do not analyze variation in the

magnitude of contractual terms as a function of both parties’ noncontractible effort, which is the focus of this

study.

2. The development of a new drug typically starts with the filing of an “Investigational New Drug” (IND)

application, followed by Phase-1 clinical trials, and ends with the approval of a “New Drug Application”

(NDA) by a regulatory authority (such as the U.S. FDA). The statistics cited heremask substantial therapy-class

level variation in development costs, times, and success probabilities. Most medical devices and instruments

are exempt from clinical trials but require either a premarketnotification (510K) or premarket approval.

570 Journal of Economics & Management Strategy

informed about the idea’s value; thus, empirically identifying the contractual implica-

tions of noncontractible effort requires controlling for the effects of private information.

In general, measuring the importance of tacit knowledge and private information are dif-

ficult because they are observed imperfectly, even by the parties in the transaction. I tackle

this challenge by using novel proxies for the importance of tacit knowledge and private

information, and by controlling for the influences of other variables such as the contract-

ing parties’ bargaining power, capital constraints, and risk preferences in the analysis.

I find that the sample licenses commonly specify three- and four-part payments

to inventors. A total of 57% of the licenses in my sample stipulates either milestone

payments or minimum royalty payments in addition to upfront fees and royalty rates.

The importance of inventors’ tacit knowledge in development is associated with higher

royalty rates in the licenses. In contrast, when the inventor’s knowledge can be codified

and specified, developers favor upfront fees or milestone payments to the inventor. The

relationship between the importance of the developer’s tacit knowledge and payments

is more nuanced. When the inventor can stipulate measurable benchmarks to safeguard

against the developer’s tendency to shelve the invention, noncontractible developer

effort is associated with minimum annual payments, and, in the absence of such stipu-

lations, with upfront payments. These findings are consistent with the prescriptions of

recent contract-design models and suggest that the payments related to the transfer of

knowledge depend on its observability and the nature of the development process.

My findings also highlight other information-related challenges and practices as-

sociated with licensing in the biomedical industry.For example, the sample distribution

of royalty rates does not display the variance one might expect with the wide disper-

sion of expected revenues and other payments in the sample. The licensing executives

I interviewed revealed that they decide royalty rates by relying on rules-of-thumb or

“comparables” because of the difficulty in predicting and verifying downstream rev-

enues. Yet, a single percentage difference in royalty rates can map to a difference of

millions of dollars in payments, and the effects of such industry practices on tasks

compensated with royalty rates—including the transfer of inventors’ tacit knowledge—

merit further investigation. Milestone fees appear prominent in licenses of early-stage

inventions, and at least some variation in minimum annual payments may be related

to inventors’ private information about their inventions’ quality. Taken together, the re-

sults suggest that while several factors affect contractual payments, the licensing parties’

concerns associated with the application of tacit knowledge explain an important share

of the variation in the payments.

The remainder of the paper is organized as follows. Section 2 reviews the empirical

implications of contract-theory models that relate the exchange of tacit knowledge to

the design of license payments. Section 3 describes the study sample and sampling

issues. Section 4 describes the payment terms in my sample and other variables. Section

5 presents regression specifications, findings, and the limitations of my analysis. Section

6 concludes with a discussion of the findings.

2. Empirical Implications of License-Design Theories

2.1 Framework and Assumptions

Consider an inventor who lacks the downstream capabilities to test, manufacture, and

market her invention. The inventor has three options to profit from her invention: ac-

quire the downstream capabilities; develop the invention in alliance with another firm;