If most companies aren’t delivering demonstrably exceptional performance, what is the justification for granting demonstrably exceptional compensation to senior executives? Rare indeed is the pay-for-performance contract that seeks to pay only for performance by separating out the effects of luck from the contributions of the skill and effort of the executive.

In many companies, business unit managers are rewarded solely on the basis of their performance against return on investment targets. Unfortunately, that often leads to denominator management because executives soon discover that reductions in investment and head count—the denominator—“improve” the financial ratios by which they are measured more easily than growth in the numerator: revenues. It also fosters a hair-trigger sensitivity to industry downturns that can be very costly. Managers who are quick to reduce investment and dismiss workers find it takes much longer to regain lost skills and catch up on investment when the industry turns upward again.

You can’t do better than what Warren Buffett said to the people at Salomon Brothers many years ago: “If you lose money for us, we will be forgiving. If you lose reputation for us, we will be ruthless.” You make the situation clear by stating your intentions and you back them up in the design of your compensation program. If there’s any suggestion of bad behavior, the money goes back to the company. That’s the only fair and credible way. Any CEO who won’t come in on that basis is somebody you don’t want to bet on because he is not willing to bet on himself.

Ordinarily, the US public attitude is generous and open minded toward profits and compensation. If somebody makes a lot of money the response is, "Great, someday it might happen to me." What drives people crazy is when you make a lot of money at people's expense. This is a powerful political form of resentment. It's resentment at being exploited. Americans want companies to make a profit but to make it by doing some good for others, not just themselves.

When I look at stock options, I am more and more convinced that in a basic sense, the trouble is not whether you expense or not - stock options are just a bad instrument. They're so subject to abuse, you want to get rid of them. There ought to be better ways of compensating people. There are better ways. Because the results are so capricious, so fortuitous, depending upon what the overall market is doing.

The logic behind the use of options as managerial incentive is flawed once you consider what behaviors are actually rewarded. Roger Martin, Dean of the University of Toronto's business school and one of the co-founders of the strategy consulting firm Monitor, noted the problems of mixing the measuring and rewarding of performance in an expectations market - the stock market - with the measuring and rewarding of performance in the real market of sales, earnings, and productivity. As he noted, in the National Football League, players would never be permitted to profit from beating the point spread - the expectations market - because it would encourage all kinds of nefarious activity. Martin argued that, "stock-based compensation is an incentive to increase expectations, not performance."

While financial returns may be a good measure of how well executives are managing the company's existing assets, they do not accurately reflect executive performance in areas with deferred returns-for example, strategic planning, growth opportunities, business initiatives, or investments in the discovery and development of new products and technologies. It is clear, then, that incentive plans based solely on accounting measures can induce senior management to make short-term business decisions, and compromise investment opportunities with the promise of deferred or non-monetary returns.

Conventional incentive plans work only with current value, leaving future value-which is often the larger amount-up for grabs.

Often, performance-based plans are a measure not of an executive's ability to reach a specific goal, but his ability to negotiate favorable, easily attainable goals with his superiors or the board.

History will look back on the last decade of executive compensation in the United States as an atrocity. The levels of pay exceeded any historical precedent, bore little comparison with compensation in other countries, and - most importantly - failed to have any correlation with the creation of value for shareholders.