(FORTUNE Magazine) – Richard F. Vancil, Lovett-Learned professor of business administration at Harvard Business School, became interested in how power is passed on in a corporation when he heard Reginald H. Jones, chairman of General Electric from 1972 to 1981, tell how he chose John F. Welch Jr. to be his successor. Deciding to write a book on the subject, Vancil persuaded 28 other chief executives to describe the transfer of power in their own companies. In addition, he analyzed the history of 227 large corporations between 1960 and 1984, exploring how top management and the board worked together to choose a successor. Out of this material, Vancil has crafted Passing the Baton, a fascinating book recently published by the Harvard Business School Press and from which this article is adapted.

ON THE BROAD LANDSCAPE of management there is one small piece of turf that has not been systematically explored. The reason is simple: It is forbidden territory for all but a few selected members of the corporate tribe. There, almost like witch doctors retreating to a secret cave to conjure up a new elixir, those few develop and act out a ritual that ultimately produces a new CEO. The secrecy preceding the event is appropriate. A corporation is not a democracy that chooses its leader by popular demand. The board proceeds discreetly, in partnership with the incumbent CEO, because the stakes are large. The most common pattern of CEO succession in large U.S. corporations is to select an overt heir apparent several years before the incumbent CEO is expected to step down. These two executives work together until the CEO, as in a relay race, passes the baton. The promotion of the heir apparent is almost a nonevent. The other succession process, less common but more widely reported in the business press, is a horse race, which yields a winner -- and several losers. An overt contest of two or more contenders entails a heavy psychic cost to everyone involved, but there may be benefits as well. One CEO who staged such a contest put it this way: ''During the horse race I could almost see the competitors growing -- sometimes two or three inches a week.'' The single most striking trend in corporate succession during the past 25 years is the increasing willingness of corporations to appoint an outsider as CEO. In the late 1960s only 8% of new CEOs were outsiders; by the early 1980s the figure was up to 25%. I define outsiders as executives who had been with the company five years or less when they got the top job. As the tasks facing a company's top management have grown in size and complexity, the load has typically been parceled out among more executives. The most common type of top management structure in large U.S. corporations during the past 25 years has been two executives holding between them some combination of the five top management titles -- chairman, chief executive officer, vice chairman, president, and chief operating officer. For example, one man might serve as chairman and CEO, the other as president and COO. In 1960 the second most common structure involved a single executive, an all- powerful chairman sitting alone atop the company in what I call the ''solo'' mode. But by 1984 the second most common structure consisted of three or more executives, each carrying one or more of the top five titles -- the ''team'' mode. A company might have a CEO and two vice chairmen, for instance. The solo mode of management dropped from 36% to 21% of the total from the early 1960s to the early 1980s, while the team mode increased from 8% to 25%. ANOTHER CONCLUSION from my study: The incumbent CEO is not an uninterested bystander in the succession process; he manages the process and must accept responsibility for finding his successor the moment he becomes CEO. His power in the selection stems from the simple fact that he knows the candidates better than any other member on the board of directors. The important role played by the board is to provide counsel to the CEO, unless and until the unfortunate moment arrives when they lose faith in him and are forced to take stronger action. I met Reg Jones of GE in April 1982, and he piqued my curiosity about the succession process. Jones was the first CEO to tell me a succession story -- how Jack Welch was selected as GE's new CEO. Here is that story: ''I asked the executive manpower staff as early as 1974 to prepare a first cut of those who might be on the slate to succeed me. The initial pool of 96 quickly melted down to ten. Interestingly, in that set of ten, Jack Welch was not one of the contenders. At that point, he was only a 37-year-old division manager. ''Guided by our tradition of early retirement for the chairman, in my own mind I had established a target date for my retirement of 1980 and laid out a road map for the intervening years. The road map was a very personal and private document. It was far too detailed and contingent for me to burden the board of directors with it in the very early years, but I did open the dialogue on succession in a very general way in 1974, and the board became progressively and very deeply and intimately involved. ''One technique that I used is what may be called the airplane interview. I sat down with each of the seven or eight candidates. They didn't know the purpose of the meeting, and I made sure they didn't tell the others, so that everybody came in surprised. You call a fellow in, close the door, get out your pipe, and try to get him relaxed. Then you say to him, 'Well, look now, Bill, you and I are flying in one of the company planes, and this plane crashes. (Pause) Who should be chairman of General Electric Co.?' ''Well, some of them try to climb out of the wreckage, but you say, 'No, no, you and I are killed. (Pause) Who should be the chairman?' And boy, this really catches them cold. They fumble around for a while and fumble around, and you have a two-hour conversation. And you learn a great deal from that meeting. Because they've not had any warning that this was going to strike, they blurt out things that you damned well better remember.'' Jones held four such interviews with each candidate. By the end of the third round it was apparent that he was asking them to define feasible teams. So the contenders began talking to each other, and some decided they would rather be a vice chairman on a winning team than simply another loser. The effect was a severe politicization of the organization. By mid-summer 1979 the lid was ready to blow. My guess is that, by that time, three or four of the contenders had made it clear to Jones that they would depart if not selected as the next CEO. The compromise was to elect Welch and two of the other contenders, John F. Burlingame and Edward E. Hood Jr., who had agreed to work with him, to be vice chairmen, but to maintain the race for some period while the board became more comfortable with Welch as the next CEO. The rationale for the compromise was that Welch was a risky maverick. That gamble produced a winner, as we now know. My assessment of the succession process at GE is to deplore the trauma and to admire the outcome. WILLIAM A. SCHREYER became the third CEO in Merrill Lynch's history in June 1984. Large by any standard, Merrill Lynch was not quite 15 years old as a publicly traded corporate entity, and in sharp contrast to General Electric, it had very few traditions regarding succession. I met with Bill Schreyer in August 1985, 14 months after he became CEO. Here is his story: ''One of Chairman Roger Birk's first statements to the board of directors on becoming CEO in 1981 was that he intended to serve as CEO only until mid-1985, when he would be 55 years old. My promotion to president in 1982 made me the logical successor. I was 53 at the time. In late May 1984, Birk asked if I could spare him some time over the Memorial Day weekend, and we met on Friday and again on Monday. Birk said he was ready to pass the CEO title to me immediately. The plan was that he would continue as chairman for another year before retiring. ''I hit the ground running, because there was a great deal to be done at that point. We were in the middle of a major cost-cutting exercise, and were also restructuring a number of our business units. Then, on July 1, I was playing a hard set of tennis and began to have chest pains. Just what I needed, having been CEO for only three weeks! The doctors said there was no real urgency, but that I would have to have bypass surgery as soon as it was convenient. I decided to postpone it until January 1985 and didn't even tell my wife of the impending surgery. ''I needed to get the house in order sooner rather than later. I moved quickly to make several organizational changes. My intent was to have all of the candidates who should be considered as the next president reporting directly to me. This meant moving several people. The executive vice president in charge of the capital markets was a particularly important appointment, and I had an executive recruiting firm do some scouting for me before giving it to Jerry Kenney ((an insider)). The three key candidates, as I saw it at that time, were Kenney, 43, Bob Rittereiser, 45, the chief administrative officer, and Dan Tully, 52, running our consumer markets. ''With that in place by mid-September, I decided to go ahead and have the heart surgery in October, right after the board meeting that month. The board did want to know who I would recommend as the next CEO in the event that something happened to me. I recommended Tully. My operation was a success, obviously. ''In late May, on the Friday before our May board meeting, Rittereiser told me he was accepting a job as president of E.F. Hutton. A few days after that, in early June, I asked Tully to spend a day with me at my home. We talked all morning. As lunchtime approached, I asked him if he would like a drink and he said he was ready for one. That was when I told him that at the June meeting I would recommend to the board that he be promoted to president. He laughed then and said, 'You bastard, you just had to rake me over the coals one more time.' And we both laughed. I also made it clear to him that becoming chief operating officer did not necessarily mean that he would someday become the CEO of Merrill Lynch. Almost exactly one year from the date I became president and CEO, Birk retired, I became chairman and CEO, and Tully was named president and chief operating officer.'' Bill Schreyer had to move fast in dealing with the succession issue at Merrill Lynch, in sharp contrast to the leisurely pace Reg Jones enjoyed. Schreyer's first action, after three months in office, was to reorganize the top level of the corporation so that the three primary candidates reported - directly to him. The result was to create a horse race among the three candidates. Schreyer intensified whatever competition there may have been among them by disappearing from the office for 2 1/2 months. How old should the new heir apparent be? Tully was four years younger than Schreyer, whereas Rittereiser and Kenney were roughly a dozen years younger. Selecting Tully as the new president, with the implication that Schreyer might serve only a five-year term as CEO, would provide the opportunity to appoint a successor to Tully within four or five years, but might cause one or both of the other candidates to depart in frustration over the delay. Selecting one of the younger candidates as Schreyer's heir apparent would mean that the next president would not be appointed for eight years or so. This might be a demoralizing prospect for some of the next generation. Rittereiser helped make the decision for them by accepting a position at E.F. Hutton, and the board easily decided that Tully should be promoted to president and CEO. The younger Kenney remains president of Merrill Lynch Capital Markets. OF ALL NEWLY appointed CEOs, roughly 10% are fired, with half of them failing within the first three years. That's how Paul F. Oreffice, whom I first met in 1979, got his job as president and CEO of Dow Chemical Co. A year earlier, Oreffice's predecessor, Zoltan Merszei, had been ousted by the board. I particularly wanted to understand the Dow situation because at the time of Merszei's removal that company had the highest percentage of inside directors among large industrial companies in the United States. In effect, a bunch of Dow executives had decided to get rid of one of their own. Here's the company's story, first from Oreffice's point of view: ''My colleagues and I know that Dow is an anomaly on the landscape of corporate governance in the United States. How we make it work is complex, as team management usually is. The CEO must give up that post at age 60, except that if he is appointed CEO after age 55, he shall have a tenure of five years. An analogous policy applies to all officers who have been successful enough to be elected to the board of directors. We call it 'deceleration' and it applies to all officers and directors beginning at age 60. At that point, the officer or director is supposed to start turning over his duties to younger managers. To remind those officers that deceleration is real, their compensation drops 10% each year, so that when they turn 65, they are only making 50% as much as they were five years earlier.''

When I asked Oreffice to comment on the two-year tenure of his predecessor, Zoltan Merszei, he begged off but helped me get in touch with Earle Barnes, who had served as chairman under Oreffice. Barnes had subsequently retired. Here is what he told me: ''Zoltan Merszei, an emigrant from Hungary, had started with Dow in Canada but had spent his entire career working for Dow in Europe. Dow President and CEO Ben Branch put him in charge of the European operations. I was in charge of Dow U.S.A., so Merszei and I were peers, and candidates to succeed Branch. Merszei was not your typical Dow executive; he was a very brassy individual. I had the feeling that he didn't trust me, and the feeling was mutual. ''In July 1975 my wife and I had just arrived at our summer home in Wyoming for a brief visit when Branch called to say that Zoltan Merszei had been selected as the next CEO of Dow. I was disappointed, of course, and surprised by the timing, because Branch would not step down as CEO until May 1976. With plenty of time to think it over, I finally decided that I could not run Dow U.S.A. under Merszei, and I informed Branch of that. When the succession occurred, I became executive vice president with responsibility for the technical side of the company, and Paul Oreffice succeeded me as the president of Dow U.S.A. ''When Merszei took over, he made Dow U.S.A. a target for change in the image of Dow Europe. He wanted to be a dictator and refused to form an operating committee with two or three other executives in his cohort. Merszei's style was one of intimidation and fear, and he had more than one serious confrontation with Oreffice. The atmosphere in the company began to sour. ''After it was all over, we referred to the episode as 'Seven Days in May,' after the title of a melodramatic novel of that time. In fact, the events occurred during the ten days preceding the board of directors' meeting in April 1978. A few weeks prior to that, I had been flying with Oreffice and had never seen him so downhearted. The harassment that he was taking from Merszei was almost unbearable, and it sounded to me like he might resign. So I went to two members of the board, Carl Gerstacker, former chairman, and Ted Doan, former CEO, telling them how serious the situation was. Gerstacker then organized a process to deal with the matter. ''At the March board meeting, Gerstacker suggested that it might be useful for the board to have an opportunity to reaffirm its confidence in Merszei. He proposed that two of the directors who were not standing for reelection the next month conduct a private poll of each director and report back to the board in April. Merszei did not object at all; in fact, he really didn't take it seriously because of his supreme self-confidence. Promptly after the meeting he took off for a 2 1/2-week trip to the Pacific. ''A majority of the votes at the April board meeting indicated that Merszei should step down as CEO. Merszei got up and left the room. The next item on the agenda was to elect a new CEO. Oreffice was elected. Branch agreed to step aside as chairman so Merszei could take that post. But that didn't work for very long either, since Merszei was completely ostracized by company management. Merszei finally resigned from the company at the February 1979 board meeting, after he had obtained an offer from Armand Hammer to join Occidental.'' If the prior CEO is no longer a member of the board, it is extremely difficult for a group of outside directors to reach the judgment that the incumbent CEO should be asked to resign. In such situations, the president and chief operating officer who takes the initiative by approaching one of the directors is at great risk personally. A new insight here, for me, was the realization of the important role a prior CEO can play if he stays on the board. I acknowledge the argument in favor of stepping out in order to clear the decks for the next CEO, but I believe keeping a prior CEO on the board is relatively cheap insurance for the one time in ten when it is necessary to disappoint the incumbent. A BRIEF PARAGRAPH in the Wall Street Journal on November 22, 1985, reported: ''Emhart Corp. named Peter L. Scott chairman and chief executive in a surprise move. Mr. Scott, head of a venture capital firm, once was in line for United Technologies' top job.'' It was clear that there was a story here. I got it first from Michael S. Scott Morton, a professor of management at MIT and an Emhart director, and then from T. Mitchell ''Mike'' Ford, the outgoing chairman and CEO. Scott Morton's story: ''To understand the recent succession process, you've really got to understand how the board works at Emhart. I joined the board in 1976 and found that I liked all the outside directors. Two of us adopted the practice of driving down from the Boston area to Hartford together and having dinner on the evening before a board meeting. Ford and all of the outside directors were invited and eventually all of them came. Ford legitimized the group, calling it the Committee on Directors and making me chairman. The dinners ran from 6 to 9 P.M. in a private dining room, and 90% of the conversation was strictly business. ''Initially, the agenda was purely strategic. Emhart had been successful in a hostile takeover of United Shoe Machinery Co. in 1976, but two years later we were still in the process of digesting it; it was twice our size and operated all over the world. We also had the nice problem of a huge cash flow, so there were a lot of investment proposals for us to review. Ford would raise an opportunity for us to discuss one when it was still a gleam in his eye, and one of our roles was the constructive killing of most such proposals. Ford is very good at identifying talented directors and then making use of their talent. However, many times it was hard for us to know where he stood on an issue. Viewed from one perspective, he appeared indecisive; from another, he could be seen as trying to stimulate ideas and build consensus within the group. In any event, Mike's style did create a very active board of directors. BY 1980 things were pretty well in hand and we started to devote some of our attention to the issue of succession. We knew we had two good inside candidates. We only had a brief discussion about bringing in an outsider; we all wanted to avoid the disruption that would cause. Peter Scott's name did come up in 1983 when he resigned from the succession rat race then under way at United Technologies. We discovered, however, that in raising money for his new venture capital firm he had made commitments to investors and could not leave the company for ten years. ''By December 1984 we were all becoming quite frustrated. The committee had been seriously discussing succession at six-month intervals for nearly five years. If Ford had strongly pushed his choice between the two insiders and recommended one of them to the board at that point, we would have all agreed with a sigh of relief. Instead, he continued to listen and the tenor of our discussions began to change. It was becoming increasingly clear to all of us that our two insiders hadn't had the experience we were going to require.

''Finally, in September we decided that we must go outside for a new CEO, and we prepared a short list of two or three. One of them was Scott. After + considerable discussion Scott said he was interested, but there was no way he could renege on his commitments. That was all we needed, and we began to explore ways to make it work. We ended up buying out his investors for about $5 million, a fairly high price to pay for a CEO, but the other way of looking at it is that we acquired a venture capital firm and had a first-class CEO thrown in for free.'' Ford continued the story: ''Several factors help explain why our board has worked so well. First, the size and composition of the board are extremely important. Emhart's board has ranged from nine to 14, normally numbering 12, with only two or three insiders. That's a small enough group that you can still have a real conversation. I have tried to find young directors (two of them were under 40 when appointed), and especially directors who are comfortable with high technology, which is the future for us. Second, the style of the CEO is the primary variable that will affect the way a board operates. My style is to be informal. A third factor is one that I don't know how to describe without using the word ego. I really believe that I need all the help I can get, but to encourage that, I have trained myself to be a good listener. ''One thing is certain: We all agree that the successor produced by our board was better than any result I could have engineered on my own.'' The Emhart board could be effective in the CEO selection decision because, for more than five years before the event, it had been thoroughly involved with Ford on substantive, strategic issues. Mike Ford deserves all the credit for the board he created. His advice to his counterparts across the country is honest and sound -- and almost impossible to implement. The constraining characteristic is ego, and many CEOs, victims of CEO-itis, cringe at the thought of being viewed as indecisive. I have not attempted to construct a prescriptive theory powerful enough to specify how the process of succession should be managed. For healthy companies, the stories I have told clearly show that many paths will get you there. Each year, the custody of billions of dollars of assets is transferred to new CEOs in an unregulated but orderly fashion. The process is generally effective for two reasons. First, the incumbent CEO and his senior officers devote enormous effort to developing the talents of their younger colleagues. Second, over the past three or four decades, the mobility of top managers has increased dramatically. We now have a national market for executive talent. Where do we go from here? The turmoil of unfriendly takeovers and ''voluntary'' restructurings has caused almost all companies to rethink the roles and responsibilities of the board of directors. In the so-called good old days, a board of directors that found itself with a CEO whose performance was just barely tolerable would simply wait him out. No more. If a corporation is to survive, the directors cannot allow a failing CEO to stay around very long. My basic recommendation: The board of directors must become more independent of the CEO. Too often today directors are handpicked by the chief executive and end up beholden to him. The primary vehicle for increasing the independence of the board lies in the committee on directors, more commonly called the nominating committee. The committee would include all outside directors and would meet several times per year. It should perform an annual formal appraisal of the CEO's performance and, after discussing this appraisal with him, should go into executive session to discuss its conclusions. The appraisal should range more broadly beyond bonuses and awards to include the CEO's objectives for the coming year and to review his performance against those goals a year later. The committee should take the initiative in identifying and nominating new candidates to join the board. These do not sound like radical proposals because they are what nominating committees were set up to accomplish. In practice, however, the initiative ends up in the hands of the incumbent CEO. Almost everyone, including me, believes that the CEO should have a right to blackball a candidate for the board simply on the grounds that the chemistry between him and the nominee won't work. In every other respect, though, the present process of selecting new directors should be turned on its head. If the board is to achieve true independence, which it needs if it is to help pick the next CEO, it must begin by being able to determine its own membership.