Companies in industries prone to significant swings in profitability present special difficulties for managers and investors trying to understand how they should be valued. In extreme cases, companies in these so-called cyclical industries—airline travel, chemicals, paper, and steel, for example—challenge the fundamental principles of valuation, particularly when their shares behave in ways that appear unrelated to the discounted value of their underlying cash flows.

We believe, however, that cyclical operations can be valued using a modified discounted-cash-flow (DCF) method similar to an approach used to value high-growth Internet start-ups.¹ First, though, we will explore the underlying relationships between the cash flows and share prices of cyclical companies, as well as the role securities analysts may well play in distorting market expectations of performance.

When theory and reality conflict

Suppose that you are using the DCF approach to value a cyclical company and have perfect foresight about its industry cycle. Would you expect the value of the company to fluctuate along with its earnings? The answer is no; the DCF value would exhibit much lower volatility than earnings or cash flow because DCF analysis reduces expected future cash flows to a single value. No individual year should have a major impact on the DCF value because high cash flows cancel out low ones. Only the long-term trend matters.

Company A, for example, has a business cycle of ten years and a highly volatile cash flow pattern that ranges from positive to negative (Exhibit 1, part 1). These cash flows can then be valued on the basis of the forecast from any one year forward. Discounting the free cash flows at 10 percent produces the DCF values in Exhibit 1, part 2. Exhibit 1, part 3, which brings together the cash flows and the DCF value (indexed for comparability), shows that the DCF value is far less volatile than the underlying cash flow. Indeed, there is almost no volatility in the DCF value, because no single year’s performance affects it significantly.

In the real world, of course, the share prices of cyclical companies are less stable. Exhibit 2 shows the earnings and share values (indexed) for 15 companies with four-year cycles. The share prices are more volatile than the DCF approach would predict—suggesting that theory and reality conflict.

Are earnings forecasts the culprit?

How can theory and reality be reconciled? On the assumption that the market values of companies are linked to consensus earnings forecasts, we examined these forecasts for clues.

What we found was surprising: consensus earnings forecasts appeared to ignore cyclicality entirely by almost always showing an upward trend, regardless of whether a company was at the peak or the trough of a cycle. Apparently, the DCF model is consistent with the facts, but the earnings and cash flow projections of the market are not (assuming that the market followed the analysts’ consensus).

This conclusion was based on an analysis of 36 cyclical companies in the United States between 1985 and 1997. We divided these companies into groups with similar cycles (three, four, or five years from peak to trough, for example) and calculated indexed average earnings and consensus earnings forecasts for each. We then compared actual earnings with forecast earnings over the course of the cycle.²

Exhibit 3 plots the actual and forecast consensus earnings for 15 of the companies, in the primary-metals or transportation equipment manufacturing industries. All have four-year cycles. As the exhibit shows, the consensus forecasts don’t predict the earnings cycle at all. In fact, except for the next-year forecasts in the years at the bottom of the trough, earnings per share are forecast to follow an upward path, with no variation. The forecasts don’t acknowledge even the existence of a cycle.³

Academic research has shown that earnings forecasts have a generally positive bias. Sometimes this is attributed to the pressures faced by equity analysts at investment banks.⁴ Analysts might fear that a company subjected to negative commentary would cut off their access, for example, or that a pessimistic forecast about a company that is a client of the bank they work for could damage relations between the two. In light of these worries, it is reasonable to conclude that analysts as a group are unable or unwilling to predict the business cycle for these companies.

The market is smarter

Business cycles, and particularly their inflection points, are hard for anyone to predict. It is not surprising, then, that the market fails to get its predictions exactly right. But we would be disappointed if it failed entirely at the task, as the consensus earnings forecasts do. This takes us back to the question of how the market ought to behave. Should it be able to predict the cycle and thus avoid fluctuations in share prices?

At any point, a company or an industry could break out of its old cycle and move toward a new one that was higher or lower

That might be asking too much; at any point, a company or industry could break out of its cycle and move to a new one that was higher or lower (Exhibit 4). Suppose, for example, that you are valuing a company that is apparently at a peak in its earnings cycle. On the basis of past cycles, you would expect the industry to turn down soon, but there might be signs that it was about to break out of the old cycle.

In this situation, a reasonable valuation procedure might be to build two scenarios and weight their values. Under the first scenario, you assume, with a 50 percent probability, that the cycle will repeat the past and that the industry will turn down in the next year or so. Under the second, you assume, also with a 50 percent probability, that the industry will break out of the cycle and follow a new long-term trend based on current improved performance. The weighted average of these two values is the company’s value. We found evidence that this is in fact the way the market looks at problems of this sort.

We valued the four-year cyclical companies in three ways:

1. With perfect foresight about the upcoming cycle
2. With zero foresight, assuming that current performance represents a point on a new long-term trend (essentially the consensus earnings forecast)
3. With 50 percent perfect foresight and 50 percent zero foresight

Exhibit 5 summarizes the results. As they show, the market follows the path of neither perfect foresight nor zero foresight but, rather, a middle path, closer to 50-50. It could be argued that this is the right place for the market to be.

How to value cyclical companies: A cookbook

No one can predict an industry’s cycle precisely, and any single forecast of performance has to be wrong. But managers and investors can benefit by explicitly following the probabilistic approach to valuing cyclical companies that is outlined above. This approach avoids the traps of a single forecast and makes it possible to explore a wider range of outcomes and their implications.

The following method of valuing cyclical companies involves creating two scenarios (though it is of course possible to create more than two). This approach provides an estimate of a company’s value and scenarios—an estimate that puts boundaries on the valuation. Managers can use the boundaries to think about how they should modify their strategies and possible ways of responding to signals that one scenario was more likely to materialize than another.

1. Construct and value the "normal-cycle" scenario using information about past cycles. Pay particular attention to the long-term trend line of operating profits, cash flow, and return on invested capital because this will affect the valuation. Make sure the continuing value is based on a "normalized" level of profits—that is, on the company’s long-term cash flow trend line.
2. Construct and value a "new trend-line scenario" based on recent performance. Again, focus most on the long-term trend line because it will have the greatest impact on value. Don’t worry too much about modeling future cyclicality, although it will be important for financial solvency.
3. Develop an economic rationale for each scenario, considering factors such as growth in demand, technological changes that will affect the balance of supply and demand, and the entry or exit of companies into the industry.
4. Assign probabilities to the scenarios and calculate their weighted value, basing it on your analysis of the likelihood of the events leading to each of them.

Managers exacerbate cyclicality by initiating big spending projects when prices are high and then retrenching when prices are low

Can managers do anything to reduce cyclicality or to exploit it? They do, after all, have detailed information about their markets and might thus be expected to do a better job than the stock market at predicting the cycle and reacting appropriately. In our experience, however, managers do exactly the opposite and exacerbate the problem. Cyclical companies often commit themselves to big capital-spending projects just when prices are high and the cycle is hitting its peak. They then proceed to retrench when prices are low. Some develop forecasts that are quite similar to those issuing from equity analysts: upward sloping, regardless of where in the cycle the company is. In so doing, these companies send the wrong signals to the stock market.

Rather than spreading confusion, managers should learn to exploit their superior knowledge. They could first improve the timing of capital expenditures and then follow up with a strategy of issuing shares at the peak of the cycle and repurchasing them at the trough. The most aggressive managers could take this one step further and adopt a trading approach, acquiring assets at the bottom of the cycle and selling them at the top. In this way, a typical company in a cyclical industry could more than double its returns.

Again, however, theory is at odds with reality. In the real world, companies are reluctant to take the contrarian view. It is hard for a CEO to persuade a company’s board and backers to expand when the outlook is gloomy and to retrench when the future looks good and competitors are building. Instead, companies are quite more likely to act in lockstep with others in their industries and to perpetuate cyclicality. So while it might indeed be possible for a company to break out of the business cycle, it is a rare CEO who can make this happen.