Their underlying assumptions about the nature of capital budgeting for research and development projects are strong, but their output is unwieldy. Coming from the perspective of someone who would be engaged in the capital budgeting exercise, I would want to have a model to which I assign my staff and expect a useable result. Interestingly, the authors appear to concur with my assessment. They offer the caveat that "the model we have presented…will not always be of immediate use for decision support." Which then begs the question of what the point of the exercise was. The issue of capital budgeting for R&D is known, and most certainly the authors are on the right track with regards to dealing with the problem. Capital budgeting for a complex R&D project surely would benefit from having a workable model to account for the risk of failure at the different stages along the way.

One point worth mentioning while we are discussing the impractical nature of the authors' output is that in industries where R&D projects are high risk, the capital budgeting process has a major variable that the authors have not accounted for. One is that the risk is typically spread out over many different firms. This is often the case in the pharmaceutical business, for example, where joint ventures are becoming increasingly common. This indicates that while capital budgeting may...

The piece begins with a good idea but takes it nowhere. The authors expend great energy to prove the obvious. They work towards a model that could just as easily have been built on an Excel spreadsheet. The writing quality is poor -- the authors are more engaged by their own work than by delivering an output that is practical in the field. They have confused their target market -- writing for theorists while working on a model for practitioners. On a theoretical level, there is not much improvement needed. What instead needs to be done is for the authors to take the next step and translate their findings into a model that allows decision makers to make reasonable inputs and receive insight. The authors have built the guts of the machine, now they need to give it an interface that will allow its intended users to benefit from their work.
Works Cited:

De Ryck, Bert & Leus, Roel. (2008). R&D Project Scheduling When Activities May Fail. IIE Transactions. Vol. 40, pp. 367-384.