Abstract:We provide a model of investment in innovation that is dynamic, features multiple heterogeneous research projects of which only one potentially leads to success, and in each period, the researcher chooses the set of projects to invest in. We show that if a search for innovation starts, it optimally does not end until the innovation is found -- which will be never with a strictly positive probability.
What problem does this paper attempt to address?
The problem that this paper attempts to solve is how to optimally select research projects for investment in a dynamic, multi - project innovation investment model, especially when there are multiple heterogeneous research projects and only one project may succeed. The core issue of the paper is to explore, in this situation, if a search for innovation is initiated, whether this search will continue until innovation is found, or will be terminated prematurely under certain conditions. The paper analyzes this process by constructing a model and arrives at a key conclusion: if the search is initiated, it will, with a positive probability, never end until innovation is found or it is determined that there is no possibility of success. This conclusion is in contrast to the view in the existing literature that even if the optimal strategy recommends continuing the search, the actual search usually stops within a finite time.
Specifically, the paper focuses on how to optimally allocate resources for search when there are multiple research projects with different costs and success rates in an infinite time horizon. The paper takes into account information externalities (that is, the information learned from failures will affect the evaluation of the success probabilities of the remaining projects) and the heterogeneity of projects (that is, different projects have different costs and success rates). These factors work together, leading to the "never - ending search" phenomenon proposed in the paper, that is, under certain conditions, rational decision - makers will choose to keep searching and will not easily give up the search even if the marginal cost keeps rising.