«David Sally L ike an upraised sword glinting in the merciless sun of the Colosseum, the Roman Empire—its pageantry, cruelty, power, and ...»
LESSONS FROM REPUBLICAN ROME
ike an upraised sword glinting in the merciless sun of the Colosseum,
the Roman Empire—its pageantry, cruelty, power, and imperiousness—
has dazzled the eyes and captured the minds of the modern crowd.
When we think of ancient Rome today, due in no small part to the box
office triumph of movies such as Ben Hur and Gladiator, we focus on the luster of the Empire and neglect Rome's early history as a republic. This is unfortunate, since the twenty-first century manager has far more to learn from Republican than Imperial Rome. In particular. Republican Rome had a successful system of co-leadership that lasted for over four centuries. This structure of co-leadership was so effective that it extended from the lower levels of the Roman magistracy to the very top position, that of consul.
Many companies today are evolving in a direction opposite to that of Rome, as departmental empires are transformed to republics, divisional emper- ors are replaced by consuls, and organizational imperial rule is changed to shared leadership. For these firms and those that might soon follow them, and for those paired managers in consular-type positions, there is good reason to review the history of the Roman Republic and gather the relevant lessons and best practices from more than 400 years of experience in co-leadership. These best practices are supported by modern research in social cognition and teams, and together they help answer the question: How do you establish a system in which two heads really are better than one?
From Emperors to Consuls There are five key reasons that co-leadership is important to a huge variety of companies: job sharing, teams, families, technology, and mergers.
84 CALIFORNIA MANAGEMENT REVIEW VOL 44, NO. 4 SUMMER 2002 Co-Leadership; Lessons from Republican Rome First, to accommodate the growing need for employees to balance work requirements and personal/family responsibilities, many firms have instituted formal job-sharing arrangements.' More than half of the large firms surveyed by Watson Wyatt in 1997 had employees sharing a job. Such arrangements occur not only on the plant fioor, sales department, or back office, but also in an increasing number of managerial and leadership positions.
Second, teams—as numerous authors have pointed out—are character- ized by shared leadership structures, and the increasing usage of teams at lower levels of companies has naturally created pressure for less imperial rule at higher levels.^ The rising generation of corporate leaders is far less likely to have passed through a "command and control" environment alone. Therefore, this cohort is less likely to expect solo, unalloyed leadership and is more likely to have the skill set necessary to share authority. As a policy matter, companies have also realized thatthe logic of teams may equally apply to the boardroom: Raymond Miles, the former dean of Berkeley's Haas School of Business, says, "If you go down into an organization, everybody agrees that having self-managing teams expected to share leadership and engage in self-governance is the way to tap the full resources of an organization. If that argument holds at the lower level, why doesn't it hold at the top?"' After experiencing great success with the shift to team production in Champion International's paper mills in the early 1990s, CEO Andrew Sigler felt that it was vital to create a team out of the firm's senior executives and find a successor, Richard Olson, who had team-oriented style and skills.* Co-leaders are a uniquely structured team of two people, and co-leadership is a much shorter and more natural step away from shared (but usually unequal) team leadership than it is from a hierarchical single commander.
Third, if the logic of teams is merely suggestive in this matter, then the logic of the modern American family is compelling. Economists have found that as a woman's share of the household income increases so does her internal bargaining power.' Where once only father knew best, now it is much more likely in two-parent families that decision making is explicitly shared and divided equally. Therefore, this change in the structure of the family combines with the utilization of teams within the firm to create an expectation for, and a naturalness to, co-leadership. Family also promotes co-leadership in the area of succession and transition in family-owned businesses. A 1997 survey of 3,000 American family businesses conducted by Arthur Andersen and MassMutual found that 10% of these companies had co-CEOs and 42% were contemplating co-leadership as a serious succession strategy."^ When the members of the third generation of Nordstroms retired in 1995, six members of the next generation were named co-Presidents.
Fourth, the increasing specialization of science and advancement of technology make conversations between different scientific sub-fields—and certainly between scientist and layperson—difficult at best and impossible much of the time.' The dialogue between engineer and marketer, or scientist and salesperson, may be facilitated by having a dedicated line of communication between two CALIFORNIA MANAGEMENT REVIEW VOL 44, NO. 4 SUMMER 2002 85 Co-Leadership: Lessons from Republican Rome partners who gradually create the vocabulary necessary to understand each other. Monsanto, for example, in its "two in a box" program, paired a scientist with a marketing or finance person and made them co-leaders of a global product division.^ The last and most visible reason that co-leadership is becoming more essential in the modern economy is the mega-merger. Co-CEOs were created by the merger of Citicorp and Travelers, Daimler and Chrysler, Novo and Nordisk, Ernst & Whinney and Arthur Young, and Time and Warner, among others. Having the CEO of each of the combining firms share the top post can be a strong signal of equal footing to the new organization. Mads Ovlisen, one of the coCEOs of Novo Nordisk, a combination of pharmaceutical firms of very different sizes, comments, "It was essential that the merger not be perceived as a takeover, but rather as a true merger of equals.... The best way to preserve what was best in both companies was to demonstrate that you could indeed share the helm."' However, as some of the combinations mentioned above (e.g., Citicorp and Travelers) portend, simply placing two leaders at the helm does not guarantee that both will grasp the tiller and steer in the same direction, or that one will not be swiftly pitched overboard. Smooth sailing, as the Romans understood, can only happen if the co-captains are embarked on the right kind of voyage, are supported by certain organizational structures, and are acting in appropriate ways.
A Brief Lesson in History Understanding Rome's keys to successful co-leadership requires a brief review of ancient history.'" After its founding in 625 BCE, Rome was ruled for over a century by a succession of kings." With the expulsion of the last king in 510 BCE, the Republic was founded and the first pair of consuls were appointed a year later. The principal challenge facing the new republic was the conquest of the Italian peninsula—the acquisition of the lands of the Etruscans and other native peoples and the repulsion of the Gauls and other invaders. After achieving this goal in two centuries, the leaders of Rome turned their eyes to the conquest of the greater Mediterranean, where they encountered Carthage in the three Punic Wars. Over these four hundred years, then, Rome was a state constantly at war.
With martial victories came territory and peoples, and the Roman state expanded accordingly. From the perspective of battlefield-as-marketplace and government-as-firm metaphors, Rome can be seen as a high-growth company that had achieved sufficient scale during this period to warrant a more complex organizational structure. In addition to the pair of consuls (CEOs), the bureaucracy by 270 BCE was headed by two censors (CFOs), two praetors (COOs), four aediles (SVPs of production and operations), four quaestors (comptrollers), and numerous provincial governors (regional VPs).
86 CAUFORNIA MANAGEMENT REVIEW VOL 44, NO. 4 SUMMER 2002 Co-Leadership: Lessons from Republican Rome Internally, the pre-eminent social conflict occurred between the orders— the patricians (nobles) and the plebians (common people). The latter group were full-fledged citizens of Rome but were denied certain privileges, including holding any of the political offices listed ahove. Intermarriage hetween the orders was formally outlawed in 450 BCE. However, the patricians' monopoly on power could not withstand the growing numbers, motivation, and wealth of the plehs, as they organized their own bureaucracy, laws, and rights—a state within a state. In 367 BCE, a merger was effected, as the Licinian laws advanced a number of economic reforms and gave wealthy plebians access to the consulship. For the next two hundred years, almost all consulships were composed of one patrician and one plebian, and within decades, plebians had attained all of the magistracies, including a guaranteed spot as co-censor.
The Roman Way of Co-Leadership Given the amount of change, growth, and conflict over such a long stretch of years, how did the Romans manage to achieve a workable system of co-leadership? There are a number of rules, norms, structures, and behaviors that history reveals as critical, that correspond to modern research, and that translate directly to organizations dealing with co-leadership today.
J. Co-leaders arrive and depart together.
Consuls and other magistrates served fixed terms. They assumed office together and, in normal circumstances, quit on the same day. If, as sometimes happened, one was injured in battle or died of old age, then he was usually not replaced. For example, the Roman historian, Livy, remarking upon the religious transgression created by the mid-term replacement of a deceased censor and the subsequent capture of the city by the Gauls, writes, "Thereafter no censor who died in office has ever been replaced."'^ The purpose of this rule is to prevent an asymmetric pairing between an entrenched incumbent and a novice (Steve Ross and Nick Nicholas at Time Warner), or between a lame duck and a rising star (Mitchell Kertzman and John Chen at Sybase). Power asymmetries such as these have been shown to affect the level of cohesiveness in small groups." Learning about the job should be intertwined with learning about each other and learning to work together. So, if one of the co-leaders departs unexpectedly, the other must remain in the position, but only as long as it takes the firm to place him or her in a new position and to recruit an entirely new duo. Of course, if a co-leader is temporarily off the job, the other forges on and the organization benefits from this flexibility and continuity: "for when news came that a huge army of Gauls had encamped in Latin territory, Scipio was seriously ill, and the conduct of the war was given by special enactment to Popilius."'* The viability of Sandy Weill's and John Reed's co-leadership of Citigroup was threatened at the outset by their failure to agree on an exit strategy. Reed CALIFORNIA MANAGEMENT REVIEW VOL. 44, NO. 4 SUMMER 2002 87 Co-Leadership: Lessons from Republican Rome says, "Neither one of us knew just when we should bring it to an end. But we both knew it would end, and we chose not to agree on anything because we felt that anything we agreed on would turn out to be a mistake and having agreed on it would create a problem, as the whole world waited for the date to arrive."" The Romans would reply that the far bigger problem was this duo's inability to specify a common exit.
2. Co-leaders must have no chance of immediately and permanently ascending to solo leadership.
If Rome was severely threatened by an enemy, the Senate was allowed to appoint a dictator who would serve for a term of six months or until the threat was parried. Such a move was partially symbolic, a clarion meant to galvanize the public and army, but it had real consequences as well, as special powers were granted to the dictator. By statute, the current consuls were not eligible for dictatorship: "The choice was made from among the ex-consuls, for this was a provision of the law that was passed concerning the dictatorship."'* This stricture is essential, and it is the one usually violated in co-CEO situations, especially those resulting from a merger. If a co-leader has some hope of abandoning the other, then cooperation swiftly turns to contest, productive disagreement to struggle, and teamwork to back stabbing. The possibility of sole succession, because of the political maneuverings of the winner and the damaged career prospects of the loser, causes co-leadership to be transitory and to be similar to a one-shot prisoners' dilemma, a situation in which we know defection is far more likely than cooperation.'^ The violation of the first and second rules was responsible for the transformation of the Roman republic to an empire, as first Sulla in 80 BCE, then Pompey in 52 BCE, and Caesar in 45 BCE, were sole consuls as well as "perpetual" dictators. Similarly, the co-equal status of the Daimler Benz and Chrysler organizations was undermined by the planned ascension of Jurgen Schrempp to solo leadership upon Bob Eaton's retirement. Schrempp admits, "Me being a chess player, I don't normally talk about the second or third move. The structure we have now with Chrysler [as a standalone division] was always the structure I wanted. We had to go a roundabout way but it had to be done for psychological reasons. If I had gone and said Chrysler would be a division, everybody on their side would have said: 'There is no way we'll do a deal.' But it's precisely what I wanted to do."'* To implement this system in an organization requires clarity, firmness, and discipline on the part of the board of directors with respect to co-CEOs, and on the part of senior managers with respect to co-leaders further down in the company. A single exception might be enough to ignite the competitive ambitions of all the joint leaders in a firm. An example of rules one and two in action occurred at Accor SA, the French hotel group, in 1996, when the co-chairmen, Paul Dubrule and Gerard Pelisson, left office together after thirty years of running the company, in favor of a single outsider, Jean-Marc Espalioux."
CALIFORNIA MANAGEMENT REVIEW VOL 44, NO. 4 SUMMER 2002 Co-Leadership: Lessons from Republican Rome
3. Co-leaders' assignments must be "lot-worthy."