Selecting among dissimilar assets: A portfolio solution

Jim DuBois, Portfolio Decisions International LP, Houston

As a portfolio management consultant, I am often asked: How can a company compare and choose among investment opportunities that are fundamentally different in character, or that occur in different time frames? For example, when building an upstream portfolio, how do you compare resource plays with their large scales and low margins with deep water exploration, or conventional step-outs, or frontier exploration?

Historical methods

Earlier generations of analysts attempted to address this question by developing single value economic indicators that, to some degree, accounted for the time dimension in a cash flow profile. For example, net present value (NPV) uses the concept of the time value of money to reduce a cash flow stream to a single, discounted value. Internal rate of return (IRR) and investment efficiency (IE) are variations on this theme. Some measures like payout and breakeven price are more simplistic, but in the end, the goal was always to have criteria for comparing (and often ranking) dissimilar opportunities using a common measure. Unfortunately, that strategy often fell short when applied to real world situations.

The problem with single value indicators is that the time character of investments is, in fact, important, and factoring it out of the investment decision causes as many problems as it solves. A small project with a quick return and a larger project with a long lead time may have similar NPVs, but their effect on company performance will be markedly different. The first may be useless to a company bent on significant growth, while the second may be out of reach to a company with cash flow issues. One investment can be declared superior to another solely on the basis of a single value indicator only if we ignore context.

Portfolio management

How, then, should you compare and decide among dissimilar assets? Using a portfolio approach, we don't compare them one against the other in the conventional sense. We choose the sets of projects that, together, give us the best chance of achieving all of our goals in all of the time frames of interest. All projects compete for funds, but not as directly as a ranking would indicate. The competition involved in building a portfolio is much more like that in making a team than in running a race. Yes, players compete to see who will fill which rolls: some will be starters while others will end up as role players or back-ups, while yet others are cut completely. Still, every player does not compete directly with every other player. The coaches select the players that they believe, together, have the best chance of winning. And the selection is not the end of the process. The games still have to be played, and during the course of the season injured players are replaced, new starters will emerge, and the team wins and loses based on a combination of skill, planning, and chance.

In a portfolio system, we select investments on the basis of how they help us achieve our desired overall performance. Companies typically have a number of different performance expectations or aspirations over a number of time periods. It is not unusual, for example, to have performance goals that cover, say, earnings, cash flow, and production. There are also resource goals: capital spending limits must be honored, debt covenants maintained, etc. These goals generally must be met year by year; investors are wary of promised performance sometime in the future with no intervening milestones along the way. The trick of choosing among dissimilar projects is to choose the combinations of projects that allow us to maximize our chances of meeting all of our goals in every specified time period. Generally, this will mean a mix of projects that generate short, medium and long term results that aggregate into a cohesive and successful whole. Instead of thinking about projects falling into rank order by some single criteria, we instead think in terms of sets of projects that work together to most efficiently achieve the performance our stakeholders expect. Ultimately, the best combination is entirely dependent upon factors such as the performance expectations that the company has set with its stakeholders, the size and financial shape of the company, and the status of previous investments.

When we have a number of projects to choose among and are evaluating the results using a number of metrics over a number of years, it becomes quite difficult to perform the required balancing act without help from computational methods. Linear programming techniques allow us to vary a number of inputs to find the maximum (or minimum) value of one variable while placing constraints on a large number of others. This method is ideal for our portfolio exercise, where we want to vary which projects are chosen and when, in order to maximize value, subject to a number of constraints on performance and on resources. Linear programming allows us to approach portfolio management from the point of view of goal seeking: finding portfolio combinations that meet all of our goals in all time periods of interest.

We sometimes see resistance to the idea that ranking on value is not the best way to choose a portfolio; especially from recent MBA graduates. Perhaps some of this confusion comes from the doctrine that the firm's sole goal should be the maximization of shareholder value. While this is more or less true (within ethical and legal boundaries, of course) it does not follow that NPV and shareholder value are equivalent. Shareholder value is best measured by sustained stock performance (growth plus income). Certainly the value of the assets is an important component of stock price, but equally important is sustained performance over a number of different measures, in line with company promises and forecasts. There is usually little in an annual report that encourages or allows a traditional long-term cash flow analysis. Investors and analysts must look to other measures to infer value and management competence.

Simple example

Let's assume that we have a base-level business that will decline over time. The predicted five year capital, production, and cash flow performance are shown in Figure 1. There are capital requirements to sustain the project, and production and cash flow are steadily declining. The NPV of the base is 75.