The accurate measurement of company performance is the foundation of corporate governance and compensation planning. Yet despite its importance, performance measurement is poorly understood. Many people believe that total returns to shareholders (TRS)—that is, share price appreciation plus dividends—is the cleanest way to measure performance. Though TRS has many merits, we believe that it cannot be uniformly applied to all companies in all situations. Incorrectly used, as it often is, it can give rise to misunderstandings about performance that in turn distort management incentives, lead to bad decisions, and alienate outstanding managers.

The analysis of corporate performance cannot be boiled down to a single number, although it can be a systematic and rigorous process. It should consider the financial market's assessment of a company (which includes but goes beyond TRS), the company's underlying performance, and its expected future performance as reflected in its market value. When a company has taken all of these elements fully into account, it can proceed to build a solid foundation for the governance and compensation decisions that are so critical to success.

The flaws in TRS

Any performance measure must incorporate a company's share price performance. If they do nothing else, investors will look at how well a stock has performed for them. But a performance measure must do more than simply record how much a stock goes up (or down). It must cut through the noise of the market and provide an accurate picture of exactly how and why managers are creating value. Seen from this perspective, TRS has obvious shortcomings.

To begin with, share prices are driven by many factors other than management performance. During the period of one to three years over which TRS is usually measured for the purpose of evaluating performance, much of the movement in a company's share price will be driven by the market as a whole or by the industry sector in which it operates. Analysis of total shareholder returns for a sample of nearly 400 companies since 1962 shows that on average over 40 percent of the returns during any one- or three-year period can be explained by market and sector movements. It follows that if performance is measured on the basis of TRS alone, managers are in effect being rewarded or penalized for events outside their control. Yet traditional share option schemes do just that, and the bull stock market of the past decade has rewarded option-holding employees in all but the most woefully underperforming companies.

The other side of the coin is a growing problem for the volatile high-technology industry. When a sector re-rating made share prices plummet, companies found they had to reprice their employee share option packages to retain key staff.

It is the delivery of "surprises" that produces higher or lower total shareholder returns compared to the market

To make things worse, share prices in the short term are driven more by differences between actual performance and market expectations and by changes in these expectations than by the level of performance per se; it is the delivery of "surprises" that produces higher or lower total shareholder returns compared to the market. As a result, companies that consistently meet high performance expectations but do no exceed them can find it hard to deliver high TRS. The market may believe that management is doing an outstanding job, but its approval has already been factored into the share price.

One way to understand the problem is by analogy with a treadmill. The expectations for future financial performance implicit in the share price are represented by the speed of the treadmill. If managers are able to beat these expectations, they accelerate the treadmill and so deliver above-average shareholder returns. However, as performance improves, the expectations treadmill turns more quickly. The better managers perform, the more the market expects from them; they have to pound the treadmill ever faster just to keep up.

For outstanding companies, the treadmill is moving faster than for anyone else. It is difficult for management to deliver at the expected level without faltering. Accelerating the treadmill will be hard; continuing to do so will eventually become impossible (see sidebar, "Tina's dilemma").

This explains why extraordinary managers may deliver only ordinary share price increases in the short run. If their compensation is based substantially on TRS through stock options, they are likely to be insufficiently rewarded. This predicament illustrates the old saw about the difference between a good company and a good investment: in the short term, good companies may not be good investments, and vice versa.

In the case of companies of which less is expected, on the other hand, TRS-driven measures may overcompensate managers. During the early years of a recovery, for example, beating expectations may be relatively easy: the expectations treadmill is not moving fast. Since the market also gives credit for future performance by re-rating the share price to reflect changes in the performance expected in all future years, the net effect is that managers can deliver high TRS even when they have improved performance only marginally.

When a company is re-rated to reflect higher expectations of future performance, there is a considerable multiplier effect: the movement in share price reflects the present value of all the changes in expectations for all future years' cashflows. As a result, TRS could be well over 50 percent. Merely to announce a new CEO can be enough to shift a share price by more than 10 percent before he or she has even arrived, and certainly long before there has been any improvement in performance. On the day in 1996 that Credit Suisse announced the appointment of Lukas Mühlemann as CEO, for example, the bank's share price rose by about 20 percent, causing shareholder value to soar by about $3 billion.

A complementary measure

An alternative market-based performance measure, market value added (MVA), has gained popularity recently, especially with the publication of financial consultancy Stern Stewart's MVA rankings in Fortune magazine in the United States and in other financial publications around the world. MVA is calculated as the difference between the market value of a company's debt and equity and the amount of capital invested. A related measure expressed as a ratio is market-to-capital, the ratio of a company's debt and equity to the amount of capital invested.¹

While MVA and market-to-capital ratio do pose difficulties of comparison and measurement because they rely on accounting data and are affected by asset ownership decisions, they provide a useful complement to TRS because they measure different aspects of a company's performance. TRS can be likened to the speeding up or slowing down of the treadmill; it measures performance against the expectations of financial markets and changes in these expectations. In effect, it is a measure of how well a company beats the target set by market expectations—a measure of improvement, in other words. MVA and market-to-capital, on the other hand, can be likened to the current speed of the treadmill. They measure the financial market's view of future performance relative to the capital invested in the business, and therefore assess the expectations of the absolute level of performance.

To understand the contrast between MVA and TRS, consider the example of US retailers Sears and Wal-Mart. Over the five years ending December 31, 1997, Sears achieved an average TRS of 22 percent a year, while Wal-Mart managed 5 percent a year. Is Sears creating more value? Is it performing better?

The MVAs and market-to-capital for Sears and Wal-Mart are shown in Exhibit 1. On December 31, 1997, Wal-Mart's market capitalization (debt and equity) was $101.3 billion and its invested capital $32.1 billion. This resulted in an MVA of $69.2 billion, one of the highest anywhere in the world. At the same time, Sears' MVA was $11.8 billion, based on a market value of $42.5 billion and invested capital of $30.7 billion. If we then look at the market-to-capital ratio, Wal-Mart scored 3.2, Sears 1.4. In other words, every dollar that Wal-Mart had invested was valued by the market at $3.20, while every dollar Sears had invested was valued at $1.40.

Wal-Mart creates more value, so it has a high market-to-capital, but it was not able to exceed the market's performance expectations because its treadmill was already moving fast. Sears does not create as much value, so it has a lower market-to-capital, but during the substantial restructuring it has undergone over recent years it has beaten market expectations. Its treadmill was moving slowly, and has speeded up. It could be argued that both companies have performed well over the five years, given their different starting points.

Combining TRS and market-to-capital can provide interesting insights into the dynamics of a company's performance, especially when the period examined is shorter than 10 years. To illustrate, Exhibit 2 plots a number of leading retailers in terms of market-to-capital ratio and TRS. The companies fall into four quadrants.

Quadrant 1 companies are the corporate elite. They include US clothing retailer The Gap, US supermarket chain Kroger, and French supermarket chain Carrefour. These companies have earned exceptionally high TRS over the past five years and have high market value in relation to the amount of capital invested in them. Quadrant 3 companies are the opposite; they face a considerable performance challenge. They include US supermarket Great Atlantic and Pacific, US discount retailer Kmart, and German retailer Karstadt. In each case, TRS is low or negative, and market-to-capital is lower than for most retailers. Companies in this quadrant (and in quadrant 1) are easy to evaluate because both measures are low (or high).

Evaluating companies in quadrants 2 and 4 is more difficult. Quadrant 2 companies are recovering underperformers. This group includes Sears, US drugstore chain Rite Aid, and UK brewer and pub and restaurant chain Whitbread. These companies have high TRS but low relative market-to-capital. Five years ago, when expectations of their performance were very low, their market-to-capital was even poorer. They have since demonstrated better than expected performance, accelerating their treadmill, but their market-to-capital ratios are still nowhere near those of excellent competitors.

Companies in quadrant 4 may have suffered from unrealistic market expectations, or they may be underachievers. They include Wal-Mart and Nordstrom. These companies have high relative market-to-capital but low TRS. Although highly valued, they have not exceeded—indeed, in some cases have not met—market expectations. Without detailed analysis, it is impossible to say whether this is the result of unrealistic performance expectations by the market at the beginning of the period, or of managers' inability to realize their companies' potential. The treadmills were simply moving too fast, and the companies have been unable to keep running at the required pace.

In these assessments, we used the relative measure of market-to-capital, but we can also use the absolute measure of MVA. Exhibit 3 shows the performance of the same retail companies using both absolute and size-adjusted measures. Relative to the amount of capital invested, the top retailer in our sample is The Gap. On an absolute basis, the winner is Wal-Mart. The Gap creates more value for each dollar invested, but Wal-Mart creates more absolute wealth. Which is better? It is impossible to say, and probably irrelevant. Both are star performers.

Underlying financial performance

No stock market-based measure truly reflects a company's underlying financial performance. As we have seen, market measures tend to lump together actual performance and expectations of future performance. To separate historical and anticipated results, it is necessary to analyze underlying financial results with the aid of accounting-based measures. While such measures clearly have their limitations, we generally find they yield useful insights.

It is now generally accepted that market valuations tend to reflect a company's ability to generate cashflow over the long term. It is also possible to identify shorter-term indicators of a company's ability to generate value. There are two key drivers of cashflow (and ultimately value): the growth rate at which a company increases its revenues, profits, and capital, and its return on invested capital relative to its cost of capital.

These financial value drivers reflect common sense: a company that earns high profits for every dollar invested in the business will be worth more than one that earns low profits for every dollar invested. Similarly, a fast-growing company will be worth more than a slow-growing company that earns the same return on invested capital (providing the return on capital exceeds the cost of capital). In practice, we make a number of adjustments to accounting figures to calculate return on invested capital so that they better reflect economic performance, and to increase comparability across companies.²

Exhibit 4 plots the revenue growth and return on invested capital for Sears and Wal-Mart between 1995 and 1997. (Before 1995, Sears' performance was distorted by its ownership of Dean Witter Discover and Allstate Insurance.) Over the period, Wal-Mart's revenue growth averaged 12.7 percent a year compared with 7.8 percent for Sears, and its return on capital averaged 13.1 percent against Sears' 8.2 percent. Both companies' cost of capital was around 8 percent. Wal-Mart achieved both higher growth and higher returns on capital.

Growth and return on capital do not take account of the companies' relative size. To do that, we use a measure called economic profit. Economic profit, also known as economic value added or residual income, measures the value created during a single year by taking a company's after-tax operating profit and deducting a charge for the capital employed. Growth in economic profit allows us to combine growth and return on invested capital into one measure. Exhibit 5 shows the calculation of economic profit for Sears and Wal-Mart. Wal-Mart's economic profit totals almost $4 billion for the years 1995-97, four times that earned by Sears. Moreover, Wal-Mart's economic profit is trending up, while Sears' is trending down.

How can it be that Sears earned higher TRS than Wal-Mart when its underlying performance was much poorer? The answer goes back to the treadmill. Sears was not expected to do well, but did better than expected. Wal-Mart, on the other hand, was the victim of high expectations. It probably earned more economic profit than any other retailer in the world, while sustaining its high growth. But the market expected even better.

Market expectations

Ultimately, a measurement system must find a way to link a company's stock market performance to its underlying financial performance. The way to do this is to start by estimating what the market's expectations for future performance are by reverse-engineering the current market value using a discounted cashflow model. This entails developing cashflow projections that when discounted at the cost of capital equal the share price in the market.

We can then draw a line showing the combinations of future growth and return on invested capital that are consistent with today's market value. These lines represent the level of performance needed to meet market expectations. If a company delivers this level of performance, its share price should rise in line with its cost of equity less the dividend yield (assuming the market as a whole moves in line with expectations). If it exceeds expectations, its share price should rise more quickly.

Exhibit 6 shows these lines for Sears and Wal-Mart, and also reveals how market expectations compare with recent performance. Wal-Mart is expected to perform considerably better than Sears, and even better than it has done recently. For Sears, the market appears to expect a similar level of performance as in the past.

Using the performance analysis

Once a company understands market expectations and its underlying financial performance, it has a solid foundation for making governance and compensation decisions. These need to be based on two dimensions: stock market measures (such as TRS and MVA) and underlying financial performance.

Consider first the stock market dimension. To eliminate the effects of general movements in the market or sector, all measurements should be relative to the company's sector or to a peer group of companies. In addition, given that market measures such as MVA and TRS can award credit before performance is delivered, any executive compensation scheme linked to these metrics should allow for a period of vesting and clawback to ensure that credit is given only for results delivered and not for share price management.

To set companies with a high market-to-capital ratio and high market expectations (such as Wal-Mart) the goal of delivering a TRS higher than their peers are called upon to achieve may be unrealistic; their expectations treadmill is already moving fast. Instead, they should be made to focus on keeping their market-to-capital high against that of their peer group. In other words, they should be measured by whether they are continuing to deliver the performance that the financial community expects. Companies with lower market-to-capital ratios and market expectations than their peers (such as Sears) are running on a slower treadmill. Speeding it up is more realistic for them, and so TRS is an appropriate measure.

Managers must be judged by what they have actually delivered, not just on the basis of what the market believes they will deliver in future

In addition to the stock market dimension, we also need to think about underlying performance, measured by return on capital, growth, or growth in economic profit. Because of their short-term focus, these measures can, if used on their own, lead companies to make decisions that will not necessarily create value over time. Used properly, however, they can ensure that managers are measured on the basis of what they have delivered, as well as on the basis of what the market believes they will deliver in the future. In the case of an underperforming company, they can measure whether managers are closing the gap with its peer group. For higher-performing companies, staying ahead of the peer group may be a more appropriate target.

Performance management is a complicated subject, but it is essential that companies understand it. Managers must be judged by what they have actually delivered, not just on the basis of what the market believes they will deliver in future. Measures such as TRS are insufficient on their own. They must be combined with a number of other ingredients: additional metrics such as market-to-capital, an understanding of the expectations of future performance that are implicit in the company's market value, and an assessment of underlying financial performance.