Composing New Product Development Portfolios with Internal and External Projects
with Moritz Fleischmann and Jochen Schlapp
Minor Revision from Management Science (under 4th round revision)
Abstract: The process of building portfolios with competing internal and external new product development (NPD) projects---a key task for NPD portfolio managers in many innovation-driven industries---comprises two daunting challenges: (i) the collection of relevant information about the projects under consideration and (ii) the selection of the most promising projects under uncertainty. In order to create optimal NPD portfolios, firms must implement resource allocation policies that carefully control the tensions that arise between these two challenges. Hence, we ask the following question: Which resource allocation policies can best align a firm's (ex ante) information acquisition efforts with its (ex post) project selection decisions when the firm contemplates investing in competing internal and external NPD projects?
When Delegating AI-Assisted Decisions Drives AI Over-reliance
with Mirko Kremer and Francis de Véricourt
Major Revision from Management Science.
Cited in The Economist: "Beware the dangers of data," January 2025.
Abstract: We empirically test a theory of AI over-reliance in a delegated and algorithm-assisted decision making environment that endogenizes institutional incentives for decision makers (DMs) to rely on algorithmic recommendations. Based on a simple behavioral model, we predict that managers blame DMs for overriding algorithmic recommendations, despite knowing that such overrides most likely reflect valuable human judgment. This prediction highlights the ambivalent role of private information in human-machine interactions. While tacit domain knowledge justifies keeping humans in the decision-making loop, the non-codifiable and non-communicable nature of such knowledge poses a challenge in delegated settings where managers try to infer decision quality from what they can observe: algorithmic recommendations and decision outcomes. As a result of managers’ blame-like behavior, DMs over-rely on algorithmic recommendations and thus make worse decisions. We design controlled laboratory experiments that support these predictions, and thus present a hurdle towards the efficient integration of AI technologies into organizational decision processes.
Optimal Stochastic Feedback in Asymmetric Dynamic Contests (2022)
with Jochen Schlapp and Jürgen Mihm
Abstract: Contests, in which contestants compete at their own expense for prizes offered by a contest holder, have become the foundational primitive of many theories of competition. Recently, the focus in contest research has turned to the role of in-contest performance feedback. The extant literature on feedback has focused on specific ad-hoc policies in symmetric contests and hence failed to more broadly characterize optimal feedback policies. In this paper we solve a general formulation of an asymmetric contest involving feedback, and thus characterize the optimal feedback policy in a very wide class of (stochastic) feedback policies. We find that, in many settings where informative feedback is useful, feedback is optimal when it is both truthful and fully informative.
Composing New Product Development Portfolios: To Share or Not to Share information? (2022)
Abstract: When composing their innovation portfolios, firms often select from a pool of internally developed and externally acquired initiatives. The potential value of the initiatives is very uncertain in the beginning. Senior management of the organization eventually obtains more refined information about the value of external projects. The question is: Should she subsequently reveal this information to the internal project managers, or not? We investigate senior management's optimal communication strategy in combination with financial incentives to see how portfolio composition and agency costs are affected.