• What Is It?
• Who Can Benefit?
• How Can It Help You?
• Various Approaches

MARKOWITZ AND EFFICIENT FRONTIER BASED MODELS

Portfolio theory, as first conceived in the 1950s by Dr. Harry Markowitz, provided a classic model for managing risk and reward. Markowitz realized that stocks and bonds interact in a predictable manner (i.e., when one class of stock went down, others tended to go up), and by managing these interactions, he could diversify risk.  Central to the method was the efficient frontier, a plot of portfolios showing reward on one axis and a risk function on the other.  The efficient frontier is a line drawn through the minimum risk points, illustrating the least amount of risk possible for each level of expected reward.

This model, developed for stock portfolios, spawned many of the ideas that were incorporated into project portfolio management, but not all of the concepts translate well.

Pros:

The efficient frontier provides a method of comparing portfolios on the basis of risk, a dimension that is often overlooked.

Cons:

Evaluation of risk and reward, and application of the efficient frontier in project portfolio management is very different than in stock portfolio implementations. In stock portfolios, reward is measured by a single metric (rate of return) and expressed at a single point in time. In a project implementation, reward is measured in terms of multiple metrics that are tracked for multiple time periods. In stock portfolios a single measure of risk is used. In a project portfolio, alternative risk measures are required. 

CROSS PLOTS AND BUBBLE CHARTS

Portfolio cross-plots were introduced and used in many industries in the 1980s. Cross-plots are two-dimensional plots with one business attribute scaled in the vertical axis and another in the horizontal axis. The attributes are often based on composite measures with names such as “attractiveness” and “competitiveness.” Assets or groups of assets are plotted to show their relative strength across the two business attributes used.  Bubble charts use this same basic structure but represent third and forth variables by the color and the size of the points.

 

Pros:

This model is visual and allows users to compare and contrast assets easily.  Use of cross plots and bubble charts forces practitioners to organize their data and think about their potential investments in terms of more than one attribute at a time.  These methods are very descriptive of certain aspects of a portfolio.

Cons:

The cross-plots do not relate directly to strategic performance goals, incorporate timing and duration, deal with risk and uncertainty in a robust manner, or capture the interactions among projects.  A single investment can look very good on one bubble chart and very poor on another.  This conflicting information often makes decisions difficult. 

 

RANK TABLES

Rank tables are commonly referred to as portfolio management tools. Rank tables consist of a sorted list of projects with the “best” projects at the top of the list. The sort is based on a ranking variable that may be a simple measure or a complex, weighted measure. The user simply selects projects from top to bottom until some critical resource is exhausted (usually capital).

Pros:

Ranking is easy to understand and explain. This model is excellent for selecting from a list of very similar projects but breaks down as the projects become less similar. 

Cons:

Rank tables fail to capture the complex interactions that exist in most project portfolio situations. They provide a poor link to most companies’ strategic goals and do not deal with risk and uncertainty in an explicit form.  Timing and materiality can be difficult to incorporate into a rank table.  Furthermore, the results can be extremely sensitive to the input data.  Small changes in one opportunity's description can cause it to rise or sink significantly on the rank table.

PROJECT ROLL-UPS

Project roll-ups constitute one more model for portfolio management. Roll-ups are done by first selecting the projects for the roll-up and then summing all their appropriate measures. This gives an aggregate picture of what this deterministic portfolio will look like relative to the measures used. Often the measures consist of multi-year profiles for the metrics identified in a company’s strategy.

Pros:

Roll-ups are useful in answering the question: “What would our business look like if we invest in this set of opportunities?”  They incorporate multiple metrics in multiple time frames, thus allowing the user to consider the impact of timing, duration, and materiality of projects.

Cons:

There is limited opportunity to manage the interactions and dependencies among projects in a roll-up. It can be time consuming to manually change the timing and composition of projects in a portfolio to see the impact on performance. Furthermore, this type of model rarely captures risk and uncertainty in an explicit form.

CONCLUSION

Our approach, using goal seeking Strategic Portfolio Management, incorporates the best of each of these methods, but it is much more powerful than any of them.   We know that not every company or every situation needs this kind of horsepower, but if you are in a complex business and are exposed to risk and uncertainty, this approach is the right one for you.

Optimization can help us pick the best selection of projects and start times to ensure that all the goals are met. Rather than answering the question, “What are the forecasted results from my selections?”, optimization allows us to ask, “Which selections do I need to obtain the performance I want?”

Interpretation of the data is visual, but all results can be investigated and explained, even if they appear non-intuitive.

Analysis is fast and flexible, and the model can be adapted to fit your business needs rather than altering the problem to fit the model.

Multiple strategic, operational and financial goals can be considered over all the time periods that are of interest.

Comparison among very dissimilar assets is based on each asset’s contribution to aggregate goals.

Results are not sensitive to small changes in the input data.  When optimization is used, projects are chosen based on their gross characteristics and how they interact with the other projects in the portfolio.

The operational realities, interdependencies and complexities of your business can be taken into consideration.

Risk and uncertainty can be integrated into the analysis and measured in terms of the probability of achieving goals and the potential magnitude of the deviation from those goals.