Here is an excerpt from an article written by Jeffrey Pfeffer for the Harvard Business Review blog. To read the complete article, check out the wealth of free resources, and sign up for a subscription to HBR email alerts, please click here.

*     *     *

It’s never easy to decide to stop pursuing a strategy. Americans got a reminder of that in President Obama’s speech recently on Afghanistan; it was dispiriting to hear him describe the extended timetable required to remove even just the incremental troops who went in as the surge. But at least Obama did manage to make a decision to scale back. Many leaders faced with a strategy that isn’t working don’t get that far.

Even when things clearly aren’t going right, strong psychological tendencies keep the average leader from admitting it and correcting course. A pathbreaking study by Barry Staw in the 1970s helped to clarify why. In it, MBAs were asked to choose the best R&D investment strategy for a case company; then, they were shown how that strategy played out (disappointingly). In the next round they were asked what to do next: Should they switch R&D projects in midstream, or pour more money into the original strategy? Staw found that the answer differed substantially based on who made the choice in the second round. When the same person responsible for the disappointing first strategy was given the power to decide the next move, it was much more likely that they would choose to stay the course. They were predisposed to escalate the commitment because to do otherwise would be to admit a mistake. (Interestingly, Staw’s paper makes direct reference to the war in Vietnam as a situation where logic might fall prey to face-saving.)

None of us likes to admit to bad decisions, but imagine how much harder that is for someone who has been chosen to lead a large organization precisely because he or she is thought to have the power to see the future more clearly and chart a wise course. The faith of others not only creates pressure, it also infects the leader with the impression that he or she really is powerful enough to make things work out. For proof of how far this self-confidence can go, look to Ellen Langer’s research on the “illusion of control.”

She showed that people have such an inflated sense of the control they personally exercise over their circumstances that they are willing to bet more on gambles when it is their own hand that rolls the dice or pulls a card from a deck. Among leaders, whose beliefs in their powers to intervene effectively have had plenty of reinforcement, that tendency is surely amplified.

And here’s one last piece of research I can’t resist citing in a discussion of why leaders are so likely to double-down on bad bets. Leslie Perlow did a fascinating study of engineers doing software development and observed them frequently using their time in ways that could only make it more difficult to meet their aggressive deadline. Her perceptive interpretation of this behavior was that, while an action that prevents problems goes mostly unnoticed, pulling a flailing project from the fire garners plenty of attention and rewards. Since leaders — like doctors — earn reputations for being great by resolving crises, they may have the same incentive to create the need for heroics.

With all these psychological tendencies acting on a leader’s judgment, it’s a wonder anyone ever manages to pull the plug on an effort that is consuming resources but going nowhere. Only the best leaders can hold fast to the truth that their job is to set strategy and ensure effective execution of it — and that at least half of that job is deciding what not to do. They know they must be disciplined in thinking about what products not to pursue and bring to market, what geographies not to enter, what activities not to focus on at the moment.

Gary Loveman, CEO of Harrah’s Entertainment, is someone who gets this. Visiting Stanford one day, he told my class that when he entered the company as COO he reduced most executives’ job scope, because he believes that people don’t do very well processing complex agendas and that success mostly comes from effort focused on the most critical and achievable objectives. Scale this up and you have the notion of comparative advantage, which suggests that companies, regions, and even nations should prune their activities and focus on areas of relative strength.

*     *     *

Jeffrey Pfeffer is the Thomas D. Dee II Professor of Organizational Behavior at the Graduate School of Business, Stanford University, where he has taught since 1979. His latest book is Power: Why Some People Have It and Others Don’t.