Science Policy Research Report: The Use of Research Portfolios in Science Policy

George Washington University, Washington DC

Investigators

Abstract

Funding organizations have been very successful at selecting individual projects under the guidance of peer review. However, the growth in size and scope of R&D across new thematic, sectoral and geographic areas has made it increasingly difficult to develop strategic thinking about the directions and de facto priorities of R&D funding. Research portfolio approaches take an R&D landscape perspective of project management to facilitate strategic thinking. Looking at the ensemble of funded projects, one may wonder if some type of relevant research is underfunded, or whether there is room for improving efficiency of the investments taking into consideration the high uncertainties of research. This project complements current management focus on R&D program quality with systemic considerations on project risk, synergies and responsiveness to societal needs. It provides tools to R&D program managers/decision makers to consider aggregate R&D portfolios. The policy report will interpret analytical methods for better handling project risk and synergies at the program level. It will also explain methods to analyze the alignment between R&D program missions and the expected societal outcomes of program portfolio. This science policy report is based on the notion that an R&D portfolio is more than the sum of the projects contained in the portfolio. The reason is that the risks of failure and project learning are not additive but multiplicative in complex ways: by considering the interactions among projects and economies of scale and scope, risks can be reduced and overall learning enhanced. Three specific approximations are shown for portfolio management. First, methods for single-objective portfolios are developed for cases such as energy consumption in an industry or within the logic of public R&D. These methods better handle technical and market uncertainty, as well as account for all stop-go decisions in the lifetime of a research project. They are based on the literature using decision trees and real options methods. Second, the project develops methods for R&D programs that have multiple goals such as citizen wellbeing, the environment, and security. In such cases, the project analyses multicriteria methods or Data Envelopment Analysis. Third, the project introduces data and visualization techniques developed in the past few years that facilitate the identification of knowledge gaps and help in making more transparent the current focus of R&D investment.