Big-time pay ... in a big-time economy.

LAST year, 58 CEOs of U.S. public companies were paid more than $20 million each. Can any one person ever be worth that much money?

The widening gap between CEO and worker pay makes these huge paychecks even harder to swallow. The median ratio of senior-executive to worker pay has grown from about 20 to 1 in 1970 to about 100 to 1 in 2005. Even worse, many workers believe that their real wage growth is slowing and their job security is low and dropping.

This has led to calls in Congress for various executive-compensation reforms, ranging from eliminating tax deductibility for deferred benefits above $1 million, to requiring specific shareholder votes on executive compensation, to setting actual limits on the maximum ratio of CEO to worker pay.

While the energy behind a lot of these proposals arises from the fear that the recent run-up in CEO pay is being created by a bunch of insiders rigging the game, it's not only crypto-socialists who get queasy when they see a CEO get a $200 million payout the same day that he's fired for incompetence. The current chairman of the Federal Reserve and his immediate predecessor, as well as the current secretary of the Treasury, have all highlighted growing income inequality as a serious issue.

Peter Drucker, who holds oracular status for thoughtful businessmen, is always cited by proponents of reduced executive compensation, because he argued for limiting the pay of any CEO to no more than 25 times that of the lowest paid of the company's workers. But it's instructive to go back and read Drucker's explanation for high CEO paychecks, back in 1977: "Every executive knows perfectly well that it is the internal logic of a hierarchical structure that explains them. The foreman has to get $15,000 after fringes or $20,000 including fringes. And each level above the foreman has to get at least 40 percent more, it is believed. If there are 30 levels, then the top man has to get $500,000 plus."

Which is quaint: From the use of a factory foreman as the standard example of a front-line supervisor, to the idea that there are 30 levels in a large company, it reads like something from a time capsule. Relying on Drucker from 1977 for guidance on reforming executive pay is like relying on Trollope for guidance on reforming Parliament. At some level of abstraction the insights about human nature are eternal, but the circumstances on the ground have changed so much that the practical proposals are no longer relevant.

Drucker's basic argument was that CEO compensation had to be set so high because there were so many levels between the shop floor and the CEO. But over the past 30 years new information technology and management practices have removed large swaths of middle management. There are many fewer layers between the CEO and the lowest-level worker in a typical public corporation today than in 1977; by Drucker's logic, this should have led to a partial convergence of CEO and "foreman" pay. So why has the reverse occurred?

Think back to the bureaucratic, inward-looking, and risk-averse corporate world of 30 years ago; it's important to remember just how bad it was. The Soviet Union was still seen as a realistic economic competitor in the 1970s, and there was widespread speculation throughout the 1980s that the U.S. could not compete with Japan and the rest of Asia. Continental Europe was held up as a model of developed-economy success.

If the U.S. of that time wanted to escape economic stagnation, the corporation of that time had to go. As I started my business career in the mid-1980s I had a ringside seat for the titanic struggle to reform American business. While an economy is constantly evolving, there was an economic Twenty Years' War that began in roughly 1980; by 2000 it was largely complete, and we now live in the new economy that it created. This wasn't a pretty struggle. It was trench warfare to force corporations to be more aggressive, entrepreneurial, and risk-taking, and it was fought out one company at a time. Change was fiercely resisted by incumbent managers, who often draped their resistance in appeals to tradition, corporate loyalty, or the public interest.

A lot of money was made by those with the wit, audacity, and stamina to force these changes. When successful, both the entrepreneurs who started new challenger companies and the partners in leveraged-buyout firms who bought control positions in existing companies became very rich. This process also created enormous financial risks and opportunities for management teams that were willing to transform companies. Ironically, by focusing on maximizing shareholder value in order to increase personal wealth, they did the hard work on the reforms that have renewed the American economy's overall position as the envy of the world. It was the invisible hand in action.

The implicit compensation paradigm in this new environment relies on markets rather than corporate diktats to set pay, accepts a much higher degree of income disparity based on market-denominated performance, and expects that most people will exploit the resulting market for talent by moving between companies many times during a career. The practical implications of this for compensation in the upper reaches of the American economy are staggering.

The astronomical incomes of professional athletes are often cited as a justification for high CEO pay, but this seems pretty fatuous: A typical budding executive does not really decide between an MBA and the NBA. But he often does decide between a career in business operations and one in investment banking, management consulting, or corporate law. The median Forbes 500 CEO, i.e., the 250th highest paid CEO among the 500 biggest U.S. public companies, made about $5.5 million in 2005. A successful senior partner in a management-consulting or corporate-law firm today makes about $1 to $2 million per year. A senior investment banker makes more: Last year the average partner at Goldman Sachs made more than $5 million. The take-home pay of a partner in a good private-equity firm or hedge fund makes all of these numbers seem like lunch money. The 400th highest paid public-company CEO in America makes about $2 million per year, but the 400th richest American is a billionaire.

There are several other factors driving the rapid recent increase in CEO pay: the deployment of new technologies, the growth in the size of the average public company, the increasing concentration of power in the CEO role, changes in tax policy, and so forth. Changes in some of these factors could reduce the pay gap between the CEO and the average worker. But seen in the context of the relevant job market, millions of dollars of annual income is what it takes to attract the right talent for the job.

It's because of this that, in the contemporary American context, attempts to directly control the price of executive labor always backfire. When Congress limited tax deductibility for cash compensation above $1 million, many companies simply increased the proportion of compensation delivered in the form of options. Options are now the primary element in most of the mega-payouts to CEOs. When corporate reformers tried to shame corporate boards by publicizing the highest paid CEOs, the CEOs reacted by being even more competitive with one another--and this drove up CEO pay even further. These reformers misunderstood the current world of executive compensation. Markets don't respond to shame, only to incentives.

Radically reducing the CEO-to-worker pay gap would require unwinding the whole new world of pay-for-performance that has created disproportionate income gains for the winners across the U.S. economy. To which populist critics say, basically, "Exactly": The real point of this whole movement is to repeal the 1980s. The problem, of course, is that this would likely create greater equality within an economy that could not grow--and not many Americans really want to trade our economy for Germany's.

But to say that this cure would be worse than the disease is not to imply that all's well in the new, market-based world of executive compensation. Markets are unsentimental about forcing change, and they are efficient at directing resources to their highest and best use, but for the participants they are a school for selfishness. Unbridled compensation markets consume a kind of social capital that any real company, or society, requires for success: solidarity with others and identification with a cause greater than self.

The real challenge is to achieve the economic performance that the market ethos enables without undermining the culture that enables the market. While no economic arrangement is permanent, we will probably be living with a high ratio of CEO to worker pay for a long time. Measures extreme enough to be effective in reducing CEO pay would be self-destructive. Fortunately, the CEO-to-worker ratio is not a very good summary measure of what creates feelings of inequity. More important to most people are: (1) the growth rate of their income, (2) the security of their standard of living, and (3) a belief that the game is not rigged against them. A practical political program should focus on these issues.

We live in an imperfect world, and life will always be to some degree unfair. Luckily, mature adults understand this, and intuitively accept the implication that in practical life the perfect is often the enemy of the good. The political salience of a voodoo-doll issue like CEO pay will rapidly wither if we focus on providing some baseline security, and on ensuring that a wealth of opportunities remain realistically available to a substantial majority of the population.

Mr. Manzi is the CEO of an applied-artificial-intelligence software company.