Transcript Pharmaceutical-Biotechnology R&D: Technological

Pharmaceutical-Biotechnology R&D:
Technological Performance Implications
JONGWOOK KIM
Western Washington University
College of Business and Economics
Introduction: Research question

How do different governance mechanisms – both formal
and informal – impact alliance performance?



How do formal governance mechanisms that seek to
mitigate information asymmetry impact alliance
performance?
How do social ties matter for alliance performance?
How do different levels of uncertainty matter for alliance
performance?
Introduction

Why biotechnology alliances?


It is difficult to measure alliance performance, particularly
in R&D alliances
Biotechnology alliances provides an empirical context
where a fairly objective indicator of alliance performance
exists:

Because of FDA regulations, information about intermediate
stages in drug development are made public
Biotechnology industry

The empirical context



Small biotechnology R&D firms usually have capabilities in
product R&D activities (upstream activities)
Larger pharmaceutical firms (and some established
biotechnology firms) supply capabilities in
commercialization activities (downstream activities, i.e.,
manage the regulatory process, large-scale manufacturing,
marketing & distribution, etc.)
Lengthy and uncertain regulatory (FDA) approval process
Biotechnology industry

Key milestones in the drug development process

Preclinical development: animal testing

Phase I clinical trials: human testing (for toxicity)

Phase III clinical trials: large-scale controlled experiments
on potential patients

(most expensive and time-consuming, 33-42% of total cost)
Introduction

Strategic alliances



Firms enter alliances to pool complementary resources to
achieve common goals, but at the same time, partner firms
are seeking access to types of resources not available to
them
Information asymmetry problems for firms acquiring R&D
on the part of client firms (usually pharmaceutical firms or
larger biotechnology firms)
The level of information asymmetry and overall
uncertainty of the drug development project changes as the
drug development process progresses
Strategic alliances

Informal responses to information asymmetry


Social ties (i.e., reputation and repeated ties) (Granovetter,
1985; Gulati, 1995; Powell, Koput, & Smith-Doerr, 1996;
Shane & Cable, 2002)
Formal responses to information asymmetry


Direct oversight through participation in R&D firm’s
management or through formal contracts
Client firms can invest in stages by setting up intermediate
milestones (Noldeke & Schmidt, 1995; Sahlman, 1990)
Theory development & hypotheses

Repeated ties

Information asymmetry problems in alliance formation are
more severe in the early stages of development



The client firm cannot accurately ascertain the potential
economic value of an R&D firm’s research output
Repeated ties between the client firm and the R&D firm is
an indication of the R&D firm’s product quality
H1: All else held constant, alliances with partner firms who had
prior alliances with one another will more likely reach Phase I
clinical trials
Theory development & hypotheses

Reputation effects


Reputation effects (past alliances by the R&D firm) may
also signal the R&D firm’s product quality
H2a: All else held constant, the greater the number of prior
alliances that the R&D firm had been involved in, the more likely
the alliance is to reach Phase I clinical trials.
Theory development & hypotheses

Reputation effects


There may be negative effects of prior alliance experience
of R&D firms: R&D firms may exploit asymmetric
information by out-licensing less-promising drug
compounds (Pisano, 1997; “lemons” hypothesis)
H2b: All else held constant, the greater the number of prior
alliances that the R&D firm had been involved in, the less likely
the alliance is to reach Phase I clinical trials.
Theory development & hypotheses

Deal size


Alliances where the expected payoff is relatively large will
lead to greater commitment by alliance partners
In particular, greater commitment suggests that the client
firm is more likely to continue funding such an alliance


The initial commitment to maintain relationship can be
viewed as a real option (Kogut, 1991; Mahoney, 2005)
H3: All else held constant, larger the amount of funds committed
to the alliance, the more likely the alliance is to reach Phase I
clinical trials.
Theory development & hypotheses

Oversight




R&D firms may be tempted to use research funds from the client
firm to fund projects other than the specified ones (Pisano, 1990;
Lerner & Malmendier, 2005)
R&D firms may derive private benefits from carrying marginally
beneficial trials forward in early stages (Guedj & Scharfstein,
2004)
Direct oversight by the client firm should reduce such alliances
from going forward
H4a: All else held constant, oversight by the client firm over the
development process will decrease the likelihood of the alliance reaching
Phase I clinical trials.
Theory development & hypotheses

Milestone payments



Client firms fund R&D projects in stages so that future
funding is contingent on performance by the R&D firm
(Sahlman, 1990; Noldeke & Schmidt, 1995; Kaplan &
Stromberg, 2004)
H4b: All else held constant, the presence of milestone payments
will increase the likelihood of the alliance reaching Phase I
clinical trials.
H4c: All else held constant, the presence of milestone payments
will increase the likelihood of the alliance reaching Phase III
clinical trials.
Theory development & hypotheses

Control rights – intellectual property



Well-defined intellectual property rights (i.e., property
rights on publication of research findings, patents, etc.) will
make clearer each partner’s gains from the alliance
Intellectual property rights become more important in the
later stages of development (information asymmetry
dominates in early stage)
H5a: All else held constant, well-defined intellectual property
rights will increase the likelihood the alliance reaching Phase III
clinical trials
Theory development & hypotheses

Control rights – manufacturing




The client firm is a pharmaceutical firm or an established
biotechnology firm that is fully-integrated into
manufacturing and distribution
Manufacturing is an especially critical control right because
of strict FDA regulations (Lerner & Merges, 1998)
Efficient division of innovative labor suggests that the
client firm should control this aspect of the alliance
H5b: All else held constant, the client firm’s control over
manufacturing will increase the likelihood of the alliance reaching
Phase III clinical trials.
Theory development & hypotheses

Alliance management



Compared to early stages of the development process, the
financial stakes for later stages are higher (33-42% of total cost of
drug development is spent on Phase III trials), but the
technological uncertainty decreases
The client firm’s capabilities in managing the alliance process
(collaborating with the R&D firm, dealing with regulatory
agencies, etc.) matter more for performance in the later stages
H6: All else held constant, the client firm’s capabilities in managing the
development process will increase the likelihood of the alliance reaching
Phase III clinical trials.
Empirical results

Data




rDNA database (Recombinant Capital)
SEC filings (10-K, 10-Q, etc.)
N = 169, research alliances (dyads only, 1990-1999) that were
analyzed by Recombinant Capital (contract analyses)
Dependent variables:


Technological performance (early stage): conditional on the drug
compound having reached preclinical trials, whether it reached
Phase I trials (“success” if Phase I within 3 years)
Technological performance (later stage): conditional on the drug
compound having reached Phase I trials, whether it reached Phase
III trials (“success” if Phase III within 5 years)
Empirical results

Independent variables (and control variables)







Repeated ties: number of prior alliances between the two alliance
partners
Reputation effects: number of prior alliances by the R&D firms
Deal size: total estimated dollar amount (including upfront
payments, equity investments, loans, milestone payments, etc.)
that the client firm as agreed to transfer to the R&D firm
Milestones: dummy variable of whether the alliance has milestone
payments specified in the agreement
Oversight: whether board seats (on R&D firm’s board) and/or joint
committees are specified in the agreement
Intellectual property rights: whether patent rights and/or publication
policies are specified in the agreement
Control variables: client firm size (sales), biotech-biotech alliance,
functional scope of alliance
Early Stages: Preclinical – Phase I (Panel probit, N = 99)
(1)
(2)
(3)
(4)
(5)
Deal size
0.4269*
(0.2188)
0.4463*
(0.2327)
0.4422*
(0.2373)
0.7408*
(0.3913)
0.8839*
(0.4813)
Reputation
effects
-0.3254*
(0.1701)
-0.4112**
(0.2080)
-0.4276**
(0.2127)
-0.5661*
(0.2950)
-0.7424*
(0.4048)
0.9171
(0.5681)
0.9315
(0.5690)
1.2464*
(0.6669)
1.3021
(0.7814)
0.2366
(0.4024)
0.3238
(0.4720)
0.2838
(0.5217)
-1.3009*
(0.7024)
-1.2853*
(0.7512)
Repeated
ties
Milestones
Joint
committee
Board seats
-1.9612
(1.5535)
Rho
0.2791
(0.3722)
0.4412
(0.3366)
0.4404
(0.3421)
0.5959**
(0.3210)
0.7045**
(0.2630)
Log L
-58.2820
-56.6391
-56.4626
-53.3178
-51.7303
5.66
6.19
6.14
5.78
4.80
Wald chi-sq
Later Stages: Phase I – Phase III (Panel probit, N = 70)
(1)
(2)
Publication policy
(3)
0.8899*
(0.5133)
Patents
0.4526
(0.7103)
Client
manufacturing
1.0207*
(0.6033)
1.0284*
(0.5651)
Milestones
1.0498*
(0.5832)
Rho
0.6563**
(0.2468)
0.6911***
(0.2095)
0.6350***
(0.2187)
Log L
-42.7177
-42.0969
-40.62994
4.10
4.39
6.23
Wald chi-sq
*p < 0.10, **p <0.05, ***p < 0.01
Empirical results

First model (early stages): Preclinical to Phase I

Repeated ties: inconclusive

Reputation effects: negative impact on performance


Consistent with Pisano’s (1997) “lemons” hypothesis
Different specifications (past alliances in same technology area only, past
research alliances only) are also negative

Deal size: positive (supported) impact on performance

Direct oversight: partially supported


Only joint committees negatively impacts performance
Unobservable attributes of client firm (rho) from panel probit model:
mixed results


Mixed results of the impact of unobserved individual client firm
characteristics on performance
Client-firm effects seem to be weak, but present
Empirical results

Second model (later stages): Phase I to Phase III

Milestones: positive (supported) impact on performance


Property rights: partial support


Publication policy clause has positive impact on performance, but
patent clause is not supported
Control over manufacturing by client firm: positive (supported)
impact on performance


Incentives in the form of milestone payments
Division of innovative labor where the client firm has more expertise
Unobservable attributes of client firm (rho) from panel probit
model: supported

Stronger support in second model of unobservable client firm-effects,
suggesting the importance of alliance management capabilities on
the part of the client firm
Conclusions


Overall, as level of uncertainty inherent in the
alliance setting changes, different variables seem to
impact performance in different ways
Data suggest information asymmetry as key driver
in early stages

R&D firms are inherently optimistic about the prospects of
their products and client firms that are able to best mitigate
possible opportunism resulting from asymmetric
information seem to have better alliance performance
Conclusions

In later stages where alliances are in Phase I clinical
trials, how the alliance is managed seems to be more
important


Data suggests well-defined property rights and division of
innovative labor, which are aspects of alliance management,
are important for performance
Also, unobservable client firm effects are strong in later
stages suggests that client firm heterogeneity (which may
include alliance management capabilities) also partially
account for variations in alliance performance