Legislation targeting companies has the best of intentions - preventing another Enron by Sarbanes-Oxley is an example - yet the punishment clauses have made it difficult to attract anyone except risk takers and has pushed CEO salary costs through the roof.

The opposite of the salary-heavy CEO is the one who takes options instead. On the exterior, it may seem to make sense because companies are paying solely for performance. More often than not, companies are paying no matter what.

A new study by Donald Hambrick of Penn State and W. Gerard Sanders of Brigham Young finds that CEOs with stock option-heavy compensation packages tend to lead their companies to more extreme performance yet they are more big losses than big gains.

In the October/November issue of the Academy of Management Journal, the authors note that the basic purpose of options has been to promote managerial aggressiveness in top executives, even if they sometimes led them "to undertake large-scale risky investments that tended to deliver extreme company performance." What happened instead, they found, "was that the extreme performance delivered by option-loaded CEOs was more likely to be in the form of big losses than big gains."

Hambrick and Sanders arrived at their conclusions through an analysis involving 950 companies selected at random from the Standard & Poor's 500, Mid-Cap, and Small-Cap indexes. For each company, they measured the proportion of CEO compensation paid in stock-option grants over three-year periods and its relationship to the magnitude of company investments in the fourth year and to the company's financial performance in the fifth year.

The authors examined three types of investments; research-and-development funding, capital investments and acquisitions. They assessed financial performance in terms of extremeness - how far two key measures (a company's return on assets and total shareholder returns) deviated from the performance expected of the company in a given year based on "a comprehensive set of control variables known to affect firm performance", such as the company's size, its international scope, current general economic conditions, current industry conditions, and the company's performance in the previous year.

Hambrick and Sanders discovered that the higher the percentage of CEO pay represented by stock-option grants, the higher the level of company investment spending and the more extreme the firm's financial performance, particularly stock performance. This extremeness, they found, is not an automatic outcome of high investment spending. But high levels of CEO stock options coupled with high levels of investment produce what the authors call a "combustible combination" that "bring[s] about very extreme outcomes" in companies' share prices. This probably happens, the authors surmise, because "option-loaded CEOs undertake big projects that are long-odds in nature."

In general, Sanders and Hambrick find, "big losses were more common than big gains under high levels of stock-option pay." For example, in companies where stock-option grants constituted half or more of the CEO's pay, 10.1 percent sustained big shareholder losses, while only 6.8 percent enjoyed big gains, a significant difference. In addition, 6.9 percent suffered extreme losses in return on assets, while only 3.9 percent reaped extreme gains, again a significant difference.

In contrast, no such disparity existed when stock options constituted less than 20 percent of CEO pay; in fact, extreme shareholder gains outnumbered extreme losses.

In seeking to account for the tendency of option-loaded CEOs to sustain big losses, Sanders and Hambrick conjecture that the answer rests with the basic nature of options.

"Because option-loaded CEOs benefit from share price increases but lose nothing if share prices drop, they can be expected to sort investment alternatives according to the expected values of gains while paying little attention to the likelihoods or magnitudes of losses," they write. "If we accept the commonsense idea that the projects with the biggest possible upside are likely to also have the biggest possible downside, and then couple it with the assumption that option-loaded CEOs have little concern with the size or probabilities of downside outcomes, it is straightforward to expect that option-loaded CEOs have a relatively high likelihood of delivering big losses."

In sum, "option-loaded CEOs are riveted on upside possibilities, with little concern for downside. Not only does this asymmetry affect the selection of strategic initiatives ... but it may also cause CEOs to be inattuned to early signs of project failure and generally careless about risk mitigation."

In striking at a feature of CEO options intrinsic to their basic purpose-fostering managerial aggressiveness-the study sets itself apart from many critiques that find fault with other failings of companies' options use. For example, it has been pointed out that options can reward top executives for share-price increases that are owed more to a rising stock market than to the executives' own efforts, or that companies may award options to executives so liberally that CEOs are satisfied with growth that fails to match the true cost of capital.

Such critiques tend to produce fixes to spur CEO aggressiveness-for example, measures that make option payoffs contingent on the amount a company's share-price rise exceeds industry or market averages, or requirements that CEOs partially meet the cost of options out of their own pockets.

In contrast, the problem identified by the new study is not that options may fail to achieve their basic purpose, to promote aggressive risk-taking, but that they often achieve it all too well.

And this problem cannot be fixed by making complex adjustments in the way these instruments are administered. If there is a solution, it is eminently simple: Award options on a very limited basis or don't award them at all, supplanting them with restricted stock.

"These findings may help put the nail in the coffin of executive stock options," comments Hambrick. "And even if not," he adds, "they certainly ought to give the corporate world pause in using them nearly as extensively or heavily as they have been used in the recent past."

While the professors allow that executive stock options may have their place in moderation or in special situations, they conclude by wondering if a better course would be a widespread supplanting of options by grants of restricted stock-that is, stock that can only be sold after a certain amount of time passes or a certain goal is achieved. While the study did not test the effect of restricted-stock grants specifically, it did assess the effect of CEO stock ownership on risk-taking.

"CEOs who held large amounts of stock delivered results that were not as lopsidedly negative" as those recorded for option-loaded chiefs, they report. "Thus, CEO shareholdings seemed to promote a more prudent type of risk-taking than was generated by stock options."

The professors add in conclusion: "Stock ownership causes CEOs to be equally concerned about gains and losses, whereas stock options encourage CEOs to think primarily about upside potential and little about downside. ... The current trend toward motivating CEOs with restricted stock may be generally sensible."

"Swinging for the Fences: The Effects of CEO Stock Options on Company Risk-Taking and Performance" will appear in the October/November issue of the Academy of Management Journal.