The Leadership Challenge

In today's world there are countless opportunities to make a difference. And more than ever there is a need for people of all ages, from all backgrounds, with all types of life experience to seize those opportunities that lead to greatness. More than ever there is a need for leaders to inspire us to dream, to participate, and to persevere. The Leadership Challenge offers everyone the chance to do just that – to take the initiative and make a difference. The Leadership Challenge is about how leaders mobilize others to want to get extraordinary things done. It's about the practices leaders use to transform values into actions, visions into realities, obstacles into innovations, separateness into solidarity, and risks into rewards. It's about a climate in which people turn challenging opportunities into remarkable successes. OBJECTIVES  Identify their leadership strengths and weaknesses  Communicate their fundamental values and beliefs  Set the example for others by aligning their actions with their shared values  Express their image of the future  Inspire others to share a common vision  Search for opportunities to take the risks needed for growth  Build collaboration, teamwork, and trust  Strengthen the ability of others to excel  Recognize the accomplishments of others  Apply The Five Practices of Exemplary Leadership® to a current organizational challenge


Introduction
Widespread malaise characterizes the progress of enterprise transformation (ET) in Eastern Europe (e.g. Nellis 1999). At the same time, scholars express a sense of dissatisfaction with agency-based corporate governance analysis as the main avenue for analysing ET (Stiglitz 1999, Kogut and Zander 2000, Spicer, McDermott and Kogut 2000. Notably, few performance differences can be explained by the differences in ownership and governance structure, especially in the former Soviet Union (Estrin and Wright 1999).
Case evidence points to an aspect scarcely considered in the literature: the role of individuals at the helm of the enterprise. This paper presents a theoretical logic that can explain the progress in ET, independent of the principal-agent logic of corporate governance that has dominated the transition literature so far. Successful ET requires to solve a major coordination game. Yet coordination games frequently fail, even in simple experimental settings (Ochs 1995).
A mechanism is thus needed that induces agents to choose routines that provide the mutually best outcome. Leaders can resolve coordination games not only by changing the structure of the game but by creating common knowledge among players on which routines shall be pursued in the future.
They can therefore overcome the extensive coordination problems faced by firms during ET.
The next section introduces the coordination problems of economic transition, which are analysed in a game-theoretic perspective in section three. Section four concludes with the implications for policy and further research.

Enterprise Transformation as a Coordination Problem
ET requires a radical change of the enterprise (Newmann andNollen 1998, Meyer 1998) changing not only the production process but also the nature of its external relations and thus its 'inner logic'. Enterprises are complex social organizations that bring together a large number of individuals with very diverse and potentially conflicting interests. This holds especially true in enterprises undergoing the transformation from plan to market as not only internal but also many external stakeholders aim to influence its restructuring strategy (Berglöf andv.Thadden 1999, Mygind 1999). The success or failure of ET depends on the combined effort of inter alia, ' employees, who are a very heterogeneous group, and may have ownership rights too, ' shareholders (after privatization), who may be dispersed, or linked to the firm in ways other than ownership, Formal models on markets in transition are also presented by Aghion and Schankerman (1999), who 1 analyze how investment in infrastructure development can, by reducing transaction costs, intensify competition and thereby induce industrial restructuring, and by Atkeson and Kehoe (1996) who model labour market frictions.
Page 2 ' providers of non-equity capital, ' providers of technological and managerial knowledge, ' suppliers and customers, some related by a long-term cooperation, ' government bureaucrats and politicians, whose support is necessary not only in firms in state-ownership, but also for private firms with a restructuring plan that depends on the regulatory environment, or direct or indirect financial support.
Case research on privatization and restructuring provides plenty of evidence of ET being derailed by conflicts between multiple agents. In Hungary, Antal-Mokos (1998) observed that internal 'politicking' between different groups and the involvement of political interest groups has prevented the implementation of a coherent strategy. In Poland, the need to obtain support from employees and their elected worker council has inhibited many restructuring plans (Carlin et al. 1995, Brada and Singh 1999, Bak and Kolawcuk 1998. In the Baltic states and Russia, insiderownership necessitates building broad support within the organization for major strategic decisions.
Throughout the region, conflicts between multiple stakeholders, or with individuals holding restitution claims, delayed substantive investment in restructuring.
Scholars have recognized coordination problems as a major cause of market failure during transition. Swaan (1997) argues that several types of transaction costs are involved in establishing new relationships because agents have few contacts and reference points for future business.
Agents not only have to engage in search and negotiations, but they have to develop new competences required for market-based transactions. This transaction problem has been formalized by Blanchard and Kremer (1997) who show that the output drop in transition economies can be largely explained by coordination failure, which they model based on information asymmetries. Roland and Verdier (1999)  The complex bargaining situations with multiple diverse agents related to any firm are hard to analyse with principal-agent theory, because agency relationships are poorly defined, or nonexisting. The policy advice of principal agency theorists thus focusses on the creation of clear Moreover, the concept of Nash-equilibrium may be unsuitable to identify an optimal strategy for a 2 firm because it is by definition stable under the condition of non-cooperative behaviour. However, most firms would be far better off if they can achieve cooperative behaviour. This requires a mechanism that encourages cooperation, which our approach provides.
Page 3 governance structures that define agency relationships. Yet this is easier said then done.
Formalizing the negotiation setting is complex and, while a Nash-equilibrium may exist, finding it may take far longer than the pressures of competitive markets permit. 2 Case evidence shows, however, that some firms have been successful over the past decade.
We have collected in-depth case studies on ET and identified those positive outliers that according to the respective authors, excel in terms of performance (table 1). These firms developed a proactive strategy pursuing a niche market strategy and/or acquiring complementary assets internationally. As would be expected, they have some valuable technological assets and a workable governance structure with hard budget constraints.
***** insert The importance of bringing in new managers, rather than creating stronger incentives for incumbents is also highlighted in empirical studies. Barberis et al. (1996), who analyse 452 shops in Russia, find that change of the manager stimulates restructuring. Hence they argue that "restructuring requires new people, who have new skills more suitable to a market economy", and that "equity incentives for old people might not be particularly effective in bringing about significant change" (1996:488). Also, Claessens and Djankov (1998)

find that performance in the Czech
Republic is improved by changing managers, but not by providing managers with incentives in form of equity stakes. All their performance indicators are negatively correlated with the length of tenure of the general manager of the firm, but positively correlated to the external recruitment of Moreover, Dyck (1997) presents a theoretical model that presumes that a change of managers is a 3 key objective for privatization.
We adapt the following terminology: strategy refers to the action of the firm, and routine to the 4 actions by individuals within the firm.
Page 4 managers. Djankov (1999) finds that general management training, independent of its functional 3 specialization, has a substantive positive effect on firm performance.
What distinguishes a leader from an administrative head of a business unit, as many of them were before 1989? First, they have to develop a corporate strategy. They have to identify business opportunities through continuous analysis of the business environment and of the position of the firm and its resources. On this basis, strategies have to be developed, assessed, and adjusted, which requires both analytical skills and creativity. The strategy may be expressed in a business plan that shows how the firm shall be repositioned. It establishes how the resources of the firm shall be combined and organized in innovative ways to create a competitive edge in whatever markets the firm chooses to be in. Among the infinite number of possible restructuring strategies, the leader has to identify one path that the enterprise shall pursue.
Secondly, the strategy has to be implemented. This requires coordination of the activity of all the stakeholders that provide resources for the operations of the firm. They, or at least a critical mass of them, have to be convinced of the path to pursue. An important step in doing so is to create appropriate incentive structures for the individual agents. Notably, those who would loose out, may have to be compensated by side payments. If the incentives faced by individual agents are not compatible, e.g. due to prisoners' dilemma type problems, non-cooperative games can emerge.
To resolve such incentive conflicts, (Dewatripont and Roland (1995) propose sequential coalition building. This has been essential at several stages of the Russian reform process as stakeholders have been expropriated or co-opted (Shleifer and Treisman 1999).
Yet, even if all stakeholders would become better off with the new strategy, they may fail to coordinate their individual routines accordingly. They may face cognitive barriers to 4 understanding the structure of the game and other players' likely actions. Leadership can overcome the coordination failure and bring all members of an organization on a common path of changeby creating the expectation that everyone else is pursuing the same path (Foss 2000). We analyse a 'coordination game' to demonstrate this role of leadership. Coordination games may appear simplistic relative to non-cooperative games, yet they provide a powerful tool to analyse For a recent review of coordination games, formal models and macroeconomic applications see 5 Cooper (1999).
Page 5 organizational realities, and the emergence or design of institutions in particular (Camerer andKnez 1994, Cavert 1995). To focus the argument, we thus assume henceforth that incentive-5 compatibility problems have been resolved.

A Coordination Game Analysis
Transition can be depicted as moving socialist firms from a low-level equilibrium to a higher level one. It should, in theory, be obvious to all agents involved that they can achieve higher returns for their firms and for themselves personally, if they change their individual routines to market-based ones. Yet why are they not changing? Figure 1 illustrates the dilemma of agents in a firm undergoing deep restructuring as a 'coordination game' (cf. Farrell 1988, Colman 1997. Two players, 1 and 2 both face a choice between two routines A and B. The pay-off matrix appears to suggest, at first sight, that both players ought to play strategy B, and collect the pay-off of two units each.

Figure 1: A coordination game
Note: in the initial situation, all agents play routine A (shaded area).
Yet, suppose the two players have played routine A for the past forty years, currently earning pay-offs of (1,1). They may become aware that due to changes in the environment, routine B becomes feasible, but they individually have incentives to stay with strategy A, if expectations over other player's strategy are formed based on past behaviour. No one has incentives to divert from (A,A), which is a Nash-equilibrium. Such a backward formation of expectations is commonly assumed in adaptive learning and evolutionary models of game theory (e.g. Fudenberg and Kreps 1993), and fairly realistic unless potential losses are small. Experiments of repeated games have There has not been any systematic experimental research on how shocks that change the pay-off 6 influence the likelihood of observing convergence to an equilibrium, let alone a specific equilibrium (Ochs 1995). The outcome is likely to be highly sensitive to how the information on the changes reaches the players.
Dependent on the structure of the game, it may suffice that agents share common beliefs rather than 7 common knowledge (Monderer and Samet 1989). Agents' successful coordination of routines depends on the beliefs they hold about each others beliefs. The more well-founded these beliefs are, the better they are able to coordinate (Foss 2000).
On the other hand, in certain situations, even almost common knowledge may not suffice to coordinate on the most efficient outcome: Suppose coordination occurs via e-mail, yet a message gets lost. As agents do not know whose message got lost, they may fail to coordinate on the most preferable outcome (Rubinstein 1989).
Analogous to convention in the principal agent literature, we refer to the leader as she, and to the 8 players (agents, employees, stakeholders) as he.
Page 6 shown strong path dependency: "learning commonly yields convergence to an equilibrium in the stage game, but the outcome is frequently history-dependent, and the effects of strategic uncertainty may persist long after it has been eliminated by learning" (Crawford 1997 other player has the same information. Neither do they know when the other player will move to routine B. The agents will only act if they share 'common knowledge' (Lewis 1969, Geanakoplos 1992) on the structure of the new game, and when the switch to the new routines is to occur.
Common knowledge refers to a situation where A knows that B knows that A knows that B knows, and so on, i.e. all agents know that others also share the knowledge, inclusive the fact that everyone else knows it too. Without common knowledge, on both the new pay-off structure and the timing of the switch, the higher-level Nash-equilibrium may never be reached. 7 Let us take the timing issue first. In our simple example, the coordination can be achieved through modifications in the assumptions, e.g. by allowing pre-play communication (Farrell 1988, Myerson 1989, Kim and Sobel 1995. Suppose, one player is appointed leader allowing her to 8 make (non-committing) announcements about the strategy. This suffices for her to lead the game to a Pareto-optimal Nash-equilibrium if she announces a strategy, from which she has no incentives to divert. Communicating it, she creates a focal point that becomes common knowledge.
The situation is more complex if both players are permitted to send messages to each other before 9 the game. If played infinitely, they too reach the superior Nash equilibrium eventually (Kim and Sobel 1995). Yet, as shown in the experiments by Cooper et al. (1994), coordination failure in the initial stages of the game is likely.

Page 7
Experimental studies show that one-sided announcements increase the likelihood of reaching the efficient outcome substantially (Cooper et al. 1994). 9 Moreover, a leader can overcome coordination failure by creating common knowledge about the strategy to be pursued (Foss 2000). She takes the necessary decisions on corporate strategy and the future role of the agents. Since the strategy is new -strategy formulation is a creative act -we cannot assume, as most game-theoretic analysis does, that the structure of the new game is common knowledge. Two-person two-routine games are very simplifying approximations of the real world, where games are large, with imperfect recall, state-contingent uncertainty etc.
Agents are likely to have incomplete information (or, especially in a transition context, none at all) about other players and available strategies (Calvert 1995, Foss 2000. In other words, the common knowledge about the game has to be created. Strategic decisions have to be communicated to all stakeholders in such ways that everyone knows that this knowledge is shared with all other relevant parties. Thus the leader can facilitate the move to a higher level equilibrium by coordinating the complementary actions of agents through designing incentive compatible routines and by creating a common knowledge concerning their implementation. The techniques employed by leaders to create common knowledge may be culturally bound. Western business leaders and scholars stress the need to create a tangible vision to focus the organisation's activities and learning (e.g. Ireland andHitt 1999, Finkelstein andHambrick 1996). A shared vision impliescommon knowledge on the organisation's objectives. It may be created through, for example, public speeches to communicate in person and simultaneously to many agents, who thus know that they share the new knowledge with everyone else who also attended the event. Increasingly, video-conferencing and e-mailing to 'everyone in group x' provide alternative, though imperfect, means.
Participatory decision processes that involve public debate on the underlying issues facilitate the creation of common knowledge on challenges facing the organization, and alternative strategic responses (though not the decision itself). In Russia, Vlachoutsicos (1998) observed a traditional participatory process, which ends with a decision that is taken at the top and communicated 'topdown'. This process, even if ritualized and with little impact on the actual decision in socialist and This situation is similar to 'stag hunt games' where routines are associated with different efforts and 10 returns depend on the lowest effort by any player. Experiments have shown that, although groups often reach an equilibrium, failure to coordinate on the Pareto-optimal equilibrium was observed in several of the experiments reviewed by Crawford (1997) and Ochs (1995).
Page 8 post-socialist firms, provides an important means to share knowledge. Moreover, it informs the leader where she has to fill in gaps of the common knowledge of the organization. Figure 1 depicts the simplest form of a coordination game. Yet, the real world is more complex, even without conflicting interests. Complexity, however, reinforces the inertia, i.e. the stability of the inferior Nash-equilibrium. Let us consider some related challenges facing enterprises during transformation using coordination games as framework. Small variations can make the game a more appropriate reflection of the situation of firms facing ET: • Suppose, a sunk cost is required to change the routines, which may moreover be higher the earlier a player changes his routine. increases. This is caused by the fact that in most of the games analyzed, defection of a single agent significantly reduces returns for all others. Theoretical models show such failure for instance if mutations or inertia influence the selection of routines (Cooper 1999:14). Experimental research mostly uses 'weakest-link games', and find that groups of 10 persons or more mostly fail to coordinate on the optimal equilibrium (e.g. Camerer and Knez 1994).
However, such strict necessity of all agents co-operating is not an appropriate reflection of ET, which is the focus of this analysis. It is more fitting to assume that superior outcome is reached if many but not necessarily all agents cooperate. Figure 3 considers a game where many players need to coordinate their shift from an established but inferior routine <old' to a superior one called <new'. The new routine yields a higher pay-off if, and only if, a substantial proportion of the agents shift to the new regime. If only a small number of agents change, everyone may in fact become worse off. This leads to a critical mass problem: it is necessary that a sufficiently large number of agents believe that a critical number of others will shift to the new routine. This is illustrated in the dynamic coordination game with ten agents in Figure 3: In real life situations, such a transitional crises may undermine the credibility of the leader, or of the 11 announced strategy. If the structure of the game and its payoffs is not common knowledge, this can thus lead to a resistence to change, and in fact attempts to return to the original equilibrium, even though under the above assumptions the game would converge to the superior Nash equilibrium. In the initial situation, all ten players follow the old routine and collect a payoff of 3, as illustrated by the shaded area. If some agents change to the new routine, this will reduce the benefits for those staying with the old routine. If all ten switch to the new routine, they will all be better off.

Page 10
However, the initial position is a stable Nash-equilibrium: with adaptive expectations (based on other agents' past behaviour), no agent, not even a pair of agents, has an incentive to change. It requires a coordination of -in the example -at least six agents to play 'new' to create a situation where playing 'new' yields at least equal return of 3 units.
Considering the evolutionary dynamics of the game, it has three equilibria, of which two are stable. In repeated rounds of this game, with expectations formed based on past outcomes, the dynamics of the game will lead to a convergence to an equilibrium where all agents play the same strategy, 'old' or 'new', though it may temporarily rest at the inferior unstable equilibrium where four player pursue the new strategy. If at least five agents shift to 'new' (e.g. because they trust their leader), then the game will converge to the new equilibrium where all play 'new'. Notice, however, that during the adjustment, the return here falls from 3 units to 2 units before increasing to 6 units, a transitional crises as observed in many firms during radical change. 11 If less than four agents shift to the new regime, the dynamics of a repeated game will lead to a return to the original, low-level equilibrium. The leader thus does not need to convince all Page 11 stakeholders to adopt the new routine, but only a critical mass of agents. This critical mass is 'five out of ten' in the example, but may be anywhere from 0.1% to 100% of the workforce in real life, and some agents may carry more weight than others.
The critical mass game, however, requires less strict assumptions about agents sharing common knowledge on the structure of the game. It suffices that a critical mass of agents believes that a critical mass understands the game, and will thus switch. With evolutionary dynamics like those depicted in Figure 3, large groups are easier to coordinate. The leader thus to create common beliefs among a large number of her employees, but does not need to fulfil the strict assumptions associated with the theoretical concept of common knowledge.

Conclusion and Extensions
ET in transition economies requires radical changes not only at the organizational level, but for each individual related to it. To induce an agent to change his behaviour, he not only needs to learn the new routines, but he has to form the belief that other agents will also change their routines in such way that his new routine will make him better off. We have argued that this kind of coordination problem can be overcome by a leader of the organization, who performs the following tasks: • to define the strategy of the firm out of a subset of alternative transformation strategies.
• to create incentives that minimize conflicts of interest among the stakeholders, i.e. the new corporate strategy becomes a Nash equilibrium of individual routines.
• to create common knowledge about the new strategy among a critical mass of agents who will thus switch and initiate evolutionary dynamics which in turn will lead to a superior Nash-equilibrium.
This leadership challenge emerges not only in enterprise transformation. Similar dynamic coordination games emerge at several levels in transition economies (Shleifer and Treisman 1999).
For instance, firms are integrated into business networks, especially if there is limited entry. They will move from the old-boy survival and rent-seeking type of networks (Huber and Worgötter 1998) to entrepreneurial networks when a critical mass of firms finds it more beneficial to leave the old networks. Also the protectionism of regions within Russia suffers from a dynamic coordination game default as individual regions try to protect their local industry through trade Page 12 barriers (Broadman 1999). Most would benefit from trade liberalization, but -as in international trade -strong lobbying by local firms inhibits the liberalization. At every level, a critical mass of agents changing to the new set of routines, attitudes and business strategies is necessary to move the group as a whole to a new, better world. A leader can play a pivotal role in moving society from one stable equilibrium to another one, usually a superior one.
Further research should incorporate this crucial role of leaders. In particular, empirical studies on enterprises in transition should include vectors of personal characteristics of leadership, such as the prior experience and reputation of both the CEO and the top management team. The empirical research by Barberis et al. (1996) has pointed to the importance of managers as leaders of firms in transition, yet this needs to be taken further. Theoretical research should deepen the analysis of coordination problems in transformation processes, analysing for instance under which circumstances agents behave cooperatively and/or follow the direction proposed by a leader.
Moreover, how can one create selection mechanisms that bring individuals with coordination skills into leadership positions, and how can incentives be designed for top managers to act as leaders, and in the best interest of the organization?