New Challenges in Alliance Portfolio Management - Innovation and Marketing in the Pharmaceutical Industry - page 156

(continued)

the measure of product market diversity, we apply an entropy measure for
technology diversity. Profi tability is measured as Return on Assets on the
basis of COMPUSTAT databases.
We fi nd that product market diversity and technology diversity are related,
as their correlation is 0.49. Regression analysis showed, however, that they
are distinct as both exert a separate positive signifi cant log-linear effect on
profi tability. The effects remain unchanged when we include control variables
(advertising, R&D, and capital intensity) or when we estimate more complex
system models where both types of diversity are endogenized.

5.4 Technological Diversity of Alliance Portfolios in the

Pharmaceutical Industry: Three Challenges

Having addressed the multifaceted nature of diversity and singled out technology diver-
sity, I now turn to three challenges that need to be addressed to advance our knowledge
on technology diversity in alliance portfolios in the pharmaceutical industry.

5.4.1 Challenge 1: Competing Theoretical Perspectives

In order to understand the consequences of a diverse alliance portfolio, insight into
fi rm motivations is essential, especially in case different theoretical perspectives
lead to different normative recommendations. As much of the prior alliance litera-
ture has relied on learning theory, several authors source from learning theory to
study alliance portfolios. The learning perspective suggests that an alliance should
be arranged such that knowledge transfer and integration are optimized. That per-
spective has important consequences. If each alliance should allow for knowledge
transfer, all ties should be strong given the scientifi c nature of knowledge.
Building portfolios of strong ties, however, severely constrains the level of diver-
sity one fi rm can manage: close ties require time and resource commitments, and
mobilizing and coordinating knowledge transfer is diffi cult (Koka and Prescott
2008 ). In addition, on top of time and resource constraints, assimilating knowledge
from very diverse partner fi rms creates problems of overload and complexity (Ahuja
and Lampert 2001 ). Further, learning theory has highlighted the importance of
knowledge integration and recombination, which again constrains the level of diver-
sity any fi rm should pursue: assimilating and integrating highly diverse knowledge
from technology domains, more diverse than a single fi rm can manage, can pose
insurmountable diffi culties (Fleming and Sorenson 2001 ). If learning theory applies
and these hindrances add up, a diverse alliance portfolio creates unwieldy manage-
ment structures (Goerzen and Beamish 2005 ). Following this line of logic, Phelps