Stanford Business

MAY 2006


The Half-Truths of Leadership

Leaders have far less control over organizations than people believe, but they can be more effective if they understand leadership myths and use them to their institutions’ advantage.

by Jeffrey Pfeffer and Robert I. Sutton


Illustrations by John Ritter

Excerpted by permission of Harvard Business School Press. Hard Facts, Dangerous Half-Truths, and Total Nonsense: Profiting from Evidence-Based Management, by Jeffrey Pfeffer and Robert I. Sutton. Copyright 2006 Jeffrey Pfeffer and Robert I. Sutton. All rights reserved.

After Andy Grove, the cofounder and former CEO and chairman of Intel, retired, he was asked how he kept his people motivated. “Well, part of it is self-discipline and part of it is deception,” Grove answered, “deception in the sense that you pump yourself up and put a better face on things than you start off feeling. But after a while, if you act confident, you become more confident. So the deception becomes less of a deception.”

We are obsessed with leadership. Thousands of studies and books are devoted to it, and we still want more. So much is written about leaders because we believe that our fate, and the fate of our organizations, is in their hands and ought to be. We talk and act as if leaders are all-powerful deities who wield complete command over even the largest organizations and that the organizations are better off for this fact.

It is not just that people believe leaders have almost total control of their organizations. Many people believe that leaders should have complete control. People in higher positions are presumed to know what should be done and how to do it better than their underlings. As social psychologist Stanley Milgram famously proved, we as human beings almost instinctively obey authority, and our very obedience reinforces the idea that leaders ought to be making the decisions that we then obey. But the idea that we are well served to have leaders in control of their organizations is a half-truth. In some cases, it is helpful, but in many instances, organizations have failed because of excessive centralization and too much influence and control on the part of the leader.

Each of these half-truths—that leaders are in control and that they ought to be—is sometimes true, but also sometimes false. In understanding these half-truths, we can provide a more nuanced view of leadership and offer some useful guidelines for those who occupy leadership roles in organizations.

History shows that leaders can make a big difference in the world. Gandhi, Martin Luther King Jr., and Winston Churchill all made positive contributions through their leadership. Other leaders had profoundly negative effects—Joseph Stalin and Adolph Hitler each consigned millions of people to their deaths.

Systematic quantitative research also demonstrates that leadership can influence organizational performance. For example, Jeffrey Pfeffer and Alison Davis Blake, PhD ’86, studied all National Basketball Association teams over a four-year period, and found that although simply changing coaches didn’t matter, bringing in a new coach enhanced team performance when that new coach had prior professional coaching experience, a strong historical win-loss record, and a track record of improving past teams.

One of the most sophisticated studies of the leadership-performance link was conducted in the automobile industry by Marvin Lieberman, Lawrence Lau, and Mark Williams. The three Stanford researchers estimated productivity equations to ascertain the growth in both labor and capital productivity in six companies over a 40-year period. The study found effects of top management on all of the companies except Toyota (because it had a system that made performance robust and largely independent of who occupied senior leadership positions) and that these effects were significant. For instance, Don Petersen, Ford’s CEO during much of the 1980s, increased productivity growth some 3.1 percent annually above the average for all Ford executives. This is perhaps because, as Peterson, MBA ’58, told us, Ford was so desperate at the time that they decided to put someone in charge who actually knew something about cars and trucks, rather than a bean-counter.

Leaders not only affect financial indicators of performance, they also affect their organization’s interpersonal climate and the satisfaction and mental well-being of those they lead. The evidence strongly suggests that leaders can certainly make a big difference to individuals when they are abusive or ineffective, or both. Researchers routinely find “that 60 to 75 percent of the employees in any organization—no matter when or where the survey was completed and no matter what occupational group was involved—report that the worst or most stressful aspect of their job is their immediate supervisor,” Robert Hogan, Gordon Curphey, and Joyce Hogan wrote in American Psychologist. These studies also show that “abusive and incompetent management create billions of dollars of lost productivity each year.”

We have witnessed firsthand how poor leaders can drive skilled and motivated people out of their organizations and into the arms of competitors, or perhaps even worse, cause people to withhold discretionary effort even while remaining in their jobs. We once interviewed a senior executive who was so demoralized by his boss’s stubbornness and poor decisions that he gave up trying to argue and instead carefully implemented every decision exactly as his controlling and detail-oriented superior instructed. This executive learned to take pleasure in how badly things turned out—he called it “engaging in malicious compliance.” Such effects can be particularly disastrous for young or small organizations that do not have large financial cushions, accumulated brand equity, or customer bases to fall back on. In short, the fact that leaders affect both morale and financial performance seems self-evident.

Except that it isn’t entirely true. Managers often have far less influence over performance than most people think. A classic study of the performance of 167 companies over a 20-year period sought to allocate variation in performance to the effects of industry, year, company-specific effects, and the impact of changes in leadership. Not surprisingly, the conclusion of authors Stanley Lieberson and James F. O’Connor was that company and industry had much larger effects on variation in sales, profits, and profit margins than did changes in leadership.

When Jeffrey Pfeffer published a review of research on leadership back in 1977, he found that although leaders do have some impact, their actions rarely explain more than 10 percent of the differences in performance between the best and the worst organizations and teams. More recent studies confirm that the link between leadership and performance is modest. Studies of leaders from large samples of CEOs in public companies to university presidents to managers of college and professional sports teams show that organizational performance is determined largely by factors that no individual—including the leader—can control. Even the most powerful executives have little influence over macroeconomic trends, the price of international currency and oil, wars and terrorism, organizational history, and the weather.

There is also evidence that leadership effects are modest because the people who are allowed to hold and keep leadership positions are pretty similar to each other. In theory, different leaders could have a big impact if they saw the world in different ways. In practice, however, leaders are selected for similarity in education and outlook. And organizations do not have unfettered access to any leader who might be potentially available. Leaders who appear to be successful will be more highly sought after, and are more likely to take positions at larger, already more successful organizations. So, another reason why leaders may make less difference in practice than in theory is that poorly performing organizations may have limited access to those leaders who would be most able to make a big difference.

Many organizations also choose internal candidates, people who have worked their way up the ranks. As renowned management theorist (and Business School professor emeritus) James March pointed out, this too drives out differences: “Assuming that all promotions are based on similar attributes, each successive filter further refines the pool, reducing variation among managers. On attributes the organization considers important, vice presidents are likely to be significantly more homogeneous than first-level managers.”

What the evidence indicates is that leaders can and do make an important difference in organizational and group performance, although the effects are not as large as usually assumed nor as important as many other factors. It seems clear that leaders have some chance of making things somewhat better, but they can also make things much worse by taking actions that increase employee turnover and diminish employee motivation, as well as encourage lying and stealing, and by causing numerous other organizational problems. This all suggests that avoiding bad leaders may be a crucial goal, perhaps more important than getting great leaders.

Leadership certainly matters. But the belief that leaders have massive influence over performance turns out to be a half-truth. Why does this irrational faith in the power of potent individuals persist?

One reason is a matter of perception. More than 30 years ago, Gerald Salancik conducted a simple experiment: A person was asked to control a model train as it traveled around a track. An observer watched the person try to control the train. Unbeknownst to both, the experimenter kept changing the power going to the train, making it speed up and slow down unexpectedly, which caused it to derail. The person running the train soon recognized that he or she had little control. The observer perceived something different. The observer could not and did not see the fluctuations in speed that were outside of the control of the person running the train, but instead saw a person who couldn’t keep a model train on the tracks. As the person running the train was visible and salient, the observer attributed the train’s performance to that person’s effort and ability, not to the unexplained and invisible factors that actually caused the derailments. In just this fashion, when we look at organizations, we see the people who are in charge; we don’t see the constraints that affect both their behavior and the company’s performance.

Another reason we give leaders too much credit or blame is that all humans need to make sense of the onslaught of confusing information thrown at us. So we use cognitive shortcuts to interpret what we see and experience in comforting and efficient ways. Our culture’s romance with leadership may reflect flawed beliefs, but it helps people translate this mess into simple terms that we can understand, cope with, and communicate to others. Even a renowned leader like former General Electric CEO Jack Welch couldn’t change his firm’s history, nor could he be everywhere at once. A while ago, we were talking with former GE executive Spencer Clark, who had led a large business during Jack Welch’s reign. Clark joked, “Jack did a good job, but everyone seems to forget that the company had been around for over a hundred years before he ever took the job, and he had 70,000 other people to help him.”

A related bias is that we often generalize from the performance of the unit to the qualities of the leader, and then infer that since performance was good (or bad), the leader must be also. People conclude that good companies have good managers, and bad companies have bad managers, regardless of other facts. But leaders who claim “It isn’t my fault” and “I couldn’t have done anything about it” aren’t doing themselves or their organizations any favors over the long haul. Ducking the heat shatters the illusion of control.

Experiments by Fiona Lee and her colleagues show that hypothetical managers who took responsibility for bad events like pay freezes and failed projects were seen as more powerful, competent, and likeable than managers who denied responsibility. Lee and her colleagues also examined yearly stock price changes in 14 companies over a 21-year stretch. They found that in years when senior management blamed their firm’s troubles on internal and controllable factors, stock prices were consistently higher the next year, compared to when executives denied responsibility for setbacks.

Dell CEO Kevin Rollins illustrates how to take blame. When the company missed quarterly revenue projections by several hundred million dollars—even though profits were up 28 percent and revenue grew 15 percent—he blamed himself, stating that “we executed poorly on managing overall selling prices.” Rollins then went on to explain how he and his team were taking steps to fix the problem.

An important reason leaders get too much credit or blame for what goes on boils down to money, power, and prestige. Leaders have enormous incentives for perpetuating the myth that they are in control, including the nearly $10 million per year that the average CEO of the largest 200 U.S. firms was paid in 2003. Regardless of their actual influence over organizational performance, senior managers still have great discretion over how their firms’ resources are distributed. They spend their days among people who have every incentive to flatter and fawn over them, further fueling the half-truth that leaders are lords and masters of all that transpires around them, including how well their organizations do.

But, ask almost any organizational leader in the private or the public sector, “Do you have as much power, control, or influence as you think you need over your organization?” and the answer is almost invariably “no.” The leaders themselves, living and managing as best they can, recognize all too well the limits on their ability to make things happen.

Which raises the question: Should leaders be in more complete control of their organizations?

From their point of view, more control is almost always seen as better. That’s because of a self-enhancement bias—the idea that, in order to see ourselves in the best possible light so we can feel better about ourselves and feed our egos, we consistently overestimate our own skills and abilities. But the belief that leaders ought to be in control is a dangerous half-truth, because when they wield too much influence and control over followers, bad things often happen to their companies.

Leaders make mistakes—all people do. But when they exert tremendous control over their organizations, there are few checks or balances to rein in the errors. Most corporate disasters and financial scandals, including those perpetrated by Jeff Skilling and Andy Fastow at Enron, Al Dunlap at Sunbeam, Hank Greenberg at the large insurer AIG, and Dennis Kozlowski at Tyco, happened not simply because these people were greedy, immoral, or unethical, but because such people were in leadership positions with so much control that no one could challenge them.

Another problem arises when leaders have excessive control. One of the most persistent and powerful social psychological processes is that of commitment—we are more likely to carry through on decisions we have made and are therefore committed to. When leaders make decisions, they think the decisions are better (after all, they made them), but the rest of the people have no investment in actually carrying through on these choices. So, ceding too much control to leaders may leave the organization confronting substantial challenges in getting things implemented.

Finally, a huge behavioral science literature shows that having both perceived and actual control over what happens in our lives is essential to the mental and physical health of all human beings, and it causes us to try harder too. So while leaders may think things are better if they have more control over their organizations, the essence of what it means to be a healthy human being suggests that too much control in the hands of too few people is probably not good for either organizations or the people who work in them.

Despite evidence to the contrary, boards of directors, search committees of all stripes, human resource management departments, and just about anyone who completes performance evaluations will likely succumb to the belief that leaders are in control and ought to be. So they hire, praise, hold on to, and promote leaders who seem to be in control of events. This means that you will never get or keep a leadership position if you can’t convince others that you are in control—even when you don’t have much control. It also means that pretending you are in control of organizational performance can help you as a leader gain actual control over at least some aspects of that performance.

Part of a leader’s job then is to behave in ways that cause others to believe in the possibility of success of both the organization and the leader. To maintain the impression that you, as a leader, are in control—and to take some actual control as well—you need to start and sustain what we call the “leadership control cycle.” The cycle begins when leaders talk and act as if they are in control, persuading key players that they have a strong impact on the organization. When changes happen to the organization, these changes are attributed to the leaders. The leaders believe they actually helped shape these changes, which gives them added confidence to complete the cycle and talk and act even more persuasively. Confidence thus becomes self-fulfilling, setting in motion behaviors that in fact make things better, whereby the leaders’ confidence is justified. In other words, “the deception becomes less of a deception,” as Andy Grove suggested.

Leadership is a difficult craft because the expectations are always so high, the blame so swift and harsh, and leaders have less impact over what happens to their organizations than most people imagine. But it is a craft that people can develop over time and that some are better at than others. By considering the various myths and half-truths about leadership, leaders can have a more positive effect on their organizations. The big lesson is that the best leaders are smart enough to act like they are in charge but wise enough not to let their power go to their heads or to take themselves too seriously.

Jeffrey Pfeffer, left, is the Thomas D. Dee II Professor of Organizational Behavior at the Graduate School of Business, and Robert I. Sutton is a professor of management science and engineering at the School of Engineering and by courtesy at the Business School. He also codirects the Center for Work, Technology, and Organization. Photograph by John Leschofs

 

 

Back to Top

FEATURES IN THIS ISSUE

 

What Should Good Leaders Do?

We have seen that leaders neither have as much control over things as many people believe, nor should they. And although leaders aren’t omnipotent, what they do does matter. Here are some sensible suggestions that can guide behavior, or more accurately, show how to design and think about behavior. These guidelines emerge from four paradoxes that leaders face.

1. Everyone expects leaders to matter a lot, even as they have limited actual impact. Leaders need to act as if they are in control, project confidence, and talk about the future, even while recognizing and acknowledging the organizational realities and their own limitations.

2. Because leaders succumb to the same self-enhancement tendencies as everyone else, magnified by the adulation they receive, they have a tendency to lose their behavioral inhibitions and behave in destructive ways. They need to avoid this trap and maintain an attitude of wisdom and a healthy dose of modesty.

3. Because the desirability of exercising total control is itself a half-truth, effective leaders must learn when and how to get out of the way and let others make contributions. Sometimes the best leadership is no leadership at all.

4. Leaders often have the most positive impact when they help build systems where the actions of a few powerful and magnificently skilled people matter least. Perhaps the best way to view leadership is as the task of architecting organizational systems, teams, and cultures—as establishing the conditions and preconditions for others to succeed.