OVERVIEW:This considerably simplified method for finding optimal R&D portfolios requires a minimum of information about the individual projects, namely the probability of a successful/unsuccessful project and the monetary returns in each case. The model requires only the assumption that the decision maker is risk averse. Linear programming, which was required in an earlier version, is replaced with a simple tree diagram that allows the user to readily uncover the nondominated solutions. This methodology is applied to an artificial five-project portfolio and to a real-world 30-project portfolio.

To date no single model has been published which fully satisfies the needs for a practical R&D project selection technique. Some earlier models cannot handle risk well, while others do not provide efficient portfolios. This paper will present a model, adapted from the literature of financial portfolio optimization, which provides a practical means of developing preferred portfolios of risky R&D projects. The method is simple and highly intuitive, requiring estimation of only two parameters, the expected return and the Gini coefficient. The Gini coefficient essentially replaces the variance in the two-parameter mean–variance model and results in a superior screening ability. The model that we present requires estimates of only these two parameters and, in turn, allows for relatively simple determination of stochastic dominance (SD) among candidate R&D portfolios. We apply our model to a simple artificial five-project set and then to a set of 30 actual candidate projects from an anonymous operating company. We demonstrate that we can determine the stochastically non-dominated portfolios for this real-world set of projects. Our technique, appropriate for all risk-averse decision makers, permits R&D managers to screen large numbers of candidate portfolios to discover those which they would prefer under the criteria of SD.

Businesses today are experiencing profound pressures to reform and improve stakeholder-related practices and their impacts on stakeholders and the natural environment-in short, to manage responsibly as well as profitably. Pressures for expanding the emphasis on profits to managing responsibly derive from three general sources: primary stakeholders such as owners, employees, customers, and suppliers; secondary stakeholders such as non-governmental organizations (NGOs), activists, communities, and governments; and general societal trends and institutional forces. The latter include a proliferation of "best of" rankings, the steady emergence and development of global principles and standards that are raising public expectations about corporate responsibility, and new reporting initiatives emphasizing the so-called triple bottom lines of economic, social, and environmental performance.To respond to these pressures, many multinational corporations (MNCs) in particular are developing what we have called total responsibility management (TRM) systems approaches for managing their responsibilities to stakeholders and the natural environment. In this article we outline the dominant pressures pushing the evolution of total responsibility management and present a managerial framework that highlights the three main components of TRM approaches-inspiration (vision), integration, and improvement/innovation-with the indicators inherent to a responsibility measurement approach.

This paper examines tradeoffs between two means of improving system reliability. These are: (1) increasing component reliability through more stringent acceptance sampling, and (2) improving system reliability through increases in component redundancy. The paper also demonstrates methods for generating probability distributions on system reliability. This is done by specifying a prior distribution on the number of defects in a lot, revising this distribution based on information in the acceptance sample, then transforming the posterior distribution on the number of defects into a distribution on reliability for a given system. Finally, the paper demonstrates the use of stochastic dominance decision criteria to analyze the tradeoffs between redundancy and component reliability. All computational work for this paper was done using EXCEL 4.0.

This paper describes a methodology for the selection of research and development (R&D) projects to add to or remove from an existing R&D portfolio, The analysis uses the criterion of conditional stochastic dominance to make selection recommendations. This criterion takes into account the effect of a given project on the risk and return of the existing portfolio. We use a methodology previously employed to analyze stock portfolios; however, we apply it using simulation in an R&D portfolio context. We apply the methodology to the portfolios of two actual companies and find that it generates priorities very close to those developed by internal company heuristics, We conclude that this methodology can be applied appropriately in these circumstances and that its recommendations are consistent with observed derision maker behavior. Our results suggest that an R&D manager should not consider project selection decisions in isolation, but, following this methodology, should take into account the context of the existing portfolio.

R&D managers express displeasure with the state of the art in portfolio decision models and are often doubtful about their firm's current portfolio of projects. Many feel that conventional portfolio decision models are impractical, requiring data that are almost impossible to estimate, or fail to take into account the risk-mitigating effects of diversification across a portfolio. The new approach presented here is scientifically rigorous, yet simple and practical. It can be implemented easily using only an EXCEL spreadsheet, and requires as inputs only the estimated success probabilities and payoffs of each R&D project. The model develops priorities for each of the R&D projects, which fake into account the joint risk of the entire portfolio. The project priorities developed by this model are closely correlated with the internally established priorities for 54 R&D projects in two large ethical pharmaceutical firms.

Strategic managers are consistently faced with the decision of how to allocate scarce corporate resources in an environment that is placing more and more pressures on them. Recent scholarship in strategic management suggests that many of these pressures come directly from sources associated with social issues in management, rather than traditional arenas of strategic management. Using a greatly improved source of data on corporate social performance, this paper reports the results of a rigorous study of the empirical linkages between financial and social performance. Corporate social performance (CSP) is found to be positively associated with prior financial performance, supporting the theory that slack resource availability and CSP are positively related. CSP is also found to be positively associated with future financial performance, supporting the theory that good management and CSP are positively related.© 1997 by John Wiley & Sons, Ltd

This paper presents methods for generating stochastic, nondominated solutions for multiobjective math programming problems. The methods are based on the assumption that the objective function coefficients are random variables with probability distributions which are known or can be approximated. The methods then generate solutions that are nondominated in terms of the expected value of each objective and the probability that each objective meets or exceeds a target value. Heuristics for generating these solutions and choosing the preferred solution are presented and illustrated with examples. The paper also discusses computational issues and issues of nonlinearity.

This paper presents improved methods for measuring a consumer's and producer's risk in acceptance sampling. We define Bayesian risks for both the consumer and producer. A Bayesian consumer's risk is defined as the probability that a lot which is accepted will contain more than a designated level of defectives, as opposed to the traditional measure of consumer's risk: the probability that a lot which contains a designated number of defectives will be accepted. A Bayesian producer's risk is defined as the probability that a lot which is rejected will contain less than a specified level of defectives, as opposed to the traditional measure: the probability that a lot which contains a designated number of defectives will be rejected. We conduct sensitivity analyses to examine the response of these risk measures to changes in the probability distribution on the number of defectives in the lot and to the variance of these distributions. We conclude that Bayesian consumer's risk gives better information to the decision maker than does a conventional consumer's risk and should be considered the preferred measure. We give practical equations for assessing these risks.

In this study, we hypothesize that institutions invest more heavily in companies with strong corporate social performance. Analysis indicated a significant, positive relationship between social per...

In the multiobjective problem it is not uncommon that solution methodologies produce a large number of nondominated alternatives. The decision maker is then left with the difficult task of choosing from this set. In this paper we present five methods for assisting the decision maker in this choice by reducing the set of all nondominated solutions to a manageable number. First we review three methods which have been previously published. We then present two new methods. The first builds on the earlier approaches. The second uses stochastic techniques to eliminate solutions which exhibit too much risk, leaving only a potentially small number of solutions with acceptable risk characteristics.

OVERVIEW: As mergers and acquisitions continue throughout the pharmaceutical industry, firms grow larger. Yet, in pharmaceutical research, bigger is not necessarily better. Evidence from a 19-year study of new product output across 31 firms suggests that the number of new chemical entities produced per R&D dollar is not enhanced by ever-increasing firm size. As a result, large pharmaceutical firms should consider restructuring and rationalizing their R&D efforts with an eye toward trying to create a smaller firm atmosphere within their larger firm environment. Tactics available to help realize this goal include decentralization, nurturance of researchers, and strategic alliances with other firms.

In acceptance sampling, producer's and consumer's risk are traditionally based on assumed fixed values of p , the proportion of the lot which is defective. A more useful definition of producer's risk would be the probability of rejecting a lot in which the proportion defective falls within some range of acceptable values. Similarly, a more useful definition of consumer's risk would be the probability of accepting a lot in which the proportion defective falls within some range of unacceptable values. In this paper, we construct measures of these more useful definitions of risk by assuming that p follows either a uniform or triangular probability distribution. The proposed measure yields consumer's risk values, β', which are smaller than the traditionally computed values by a factor of up to twenty times. The proposed measure of producer's risk, α', gives values smaller than traditional values by a factor of two to four times. Decision makers who adopt the proposed measures may be able to reduce sample sizes substantially while maintaining given risk levels.

Since R&D projects typically last for a number of years, their quantitative analysis and comparison requires some rational means of adjusting project worth as a function of the time profile of project expenditures and returns. An additional adjustment must be made to account for the risk that is present in virtually all R&D activities. Since many projects last for as long as 10 years, these time and risk adjustments may be quite substantial, and may dominate the decision calculus. This paper examines the state of the art in such decision-making models, and recommends methods for improvement. We deal first with adjustments for time alone, ignoring risk. Later, we examine methods for adjusting for both time and risk. Adjusting for Time Alone The most common method of adjusting for time profiles of returns on capital investment projects is the present value technique (here we subsume both net present value and internal rate of return under the same heading). A 1978 survey suggested that more than 85 percent of firms use such techniques to make capital budgeting decisions (1). In another study, Rizzuto and Cook show that rate of return is the primary determinant in management choices abo'ut R&D spending levels However, a third, experimental, study shows that the manager's discounted cash flow is not always consistent with the classical present value approach (3). While present value techniques are sound and well-seasoned methods, they do require several restrictive assumptions about the decision maker's time preferences. These assumptions, which are usually unexamined, limit the applicability of the technique. The choice among cash flows with different time profiles is a multi-attribute decision theory problem of a special type. In choices of cash flows over time, the decision maker would like to maximize the cash flow in each year. Since it is rare that any choice will yield this utopian solution, the practicing manager must trade off one year's cash flow against another's. A decision theorist would model this tradeoff process with a multi-attribute value function. This may be very difficult in practice. Decision theory, however, allows us to describe the behavioral properties implied by different tradeoff techniques. The present value technique is a well-known and trusted method for accomplishing tradeoffs among cash flows. The decision maker who uses this technique, however, is assumed to possess the following preference characteristics: 1. Suppose the decision maker is to choose between two cash flow streams, x = (x1, x2, . . ., xn), and y = (y1, y2, . . ., yn). It is assumed that the present-value decision maker will prefer cash flow stream x to y if one year's cash flow in x exceeds the same year's flow in y while all the other years are equal. This assumption, called dominance, seems natural for virtually any rational decision maker. 2. Suppose that the decision maker is faced with two specific cash flow streams, x = (10,16,20,25,50,100), and y = (10,15,20,40,50,100). Notice that these two streams differ only in Years 2 and 4. Now further suppose that the decision maker preferred x to y due to the slightly higher return in Year 2 which would help in some temporary cash flow problem. If the decision maker is to satisfy the assumptions of present value, then the choice between x and y must not change if any of the cash flows (other than Years 2 and 4) are changed by an equal amount in both x and y. So assume that x = (100,16,20,25,50,100), and y = (100,15,20,30,50,100). With such high returns in Year 1, the short term cash flow problem in Year 2 is no longer a matter for concern, so the decision maker decides to choose y. This change in preferences is inconsistent with the assumptions about the present-value decision maker. 3. It is assumed that the present-value decision maker is indifferent between one dollar of cash flow in year t and 1 + r dollars in year t + 1, where r is the relevant discount rate. …

An alternative to the net present value (NPV) formulation of the objective in problems involving choice between cash flows over time is presented. This approach is illustrated in an R&D project selection problem. It is shown that the NPV formulation is a special case of optimizing a multiattribute value function. This special case requires restrictive assumptions about the decision-maker's prefere...

(1989). Why Costs Increase When Projects Accelerate. Research-Technology Management: Vol. 32, No. 2, pp. 16-18.

A multi-objective model of the project-selection problem is described. The model departs from an earlier goal-programming formulation of the problem, which suggested Delphic methods for selection of priorities and aspiration levels. It is shown that the multiobjective formulation yields multiple nondominated solutions for the same problem solved by goal programming, whereas the goal-programming fo...

This paper provides a literature review which describes the current state of knowledge regarding the tradeoff between completion time and development cost in Research and Development projects. A review of the theoretical literature suggests a strictly convex curve, with more severe compressions of project duration being purchased at increasingly high cost. Empirical results support this theory, showing, for example, that a one percent reduction in project duration may increase project cost by one to two percent.