Mayors, Governance Coalitions, and Strategic Capacity

A comparison of two German cities responding in the 1980s to de industrialization directs our attention toward factors that explain how leaders build governance and strategic capacities: a “crisis” situation, higher level government aid, party ties and low party competition, solidarity, friendship,andmomentum.Factors that donot explain variation in these cases include the formal resources of strong-mayor city charters and the existence of a governance coalition. Mayors are in a unique position to increase governance capacity through in formal means, and if they do, they often also increase strategic capacity. This effect is, however, short term at best.

governance capacity and strategic capacity. Governance capacity is the ability to act at all, given barriers to action. Strategic capacity reflects on the ability of a city to act in a planned, coordinated, and rational manner.
The governance and strategic capacities of cities, the operationalization of which is discussed in greater detail in the second section of this article, vary from city to city. Although governance capacity is a necessary condition for strategic capacity (a city has to be able to act in order to act strategically), governance capacity is certainly not a sufficient condition for strategic capacity. Even as most cities are doing something in the field of economic development, economic development leaders in cities find it difficult to respond strategically to particular problems (Spindler and Forrester 1993). Thus, we expect that the set of cities that "shoot anything that flies" (Rubin 1988) in economic development policy making is larger than the set of cities that engage in precise targeting and other forms of strategic decision making. Some cites are better than others at gathering information, activating resources, coordinating the efforts of actors and organizations throughout a long period of time and a large geographic area, and self-critically evaluating their own policies. As the incidence and causes of strategic capacity have been neglected in the urban governance literature, the goal of this article is to identify factors that are potentially associated with the strategic capacity of urban policy-making systems through the use of empirical cases. The third section discusses a range of potential factors suggested by the existing urban politics literature, with a special focus on leadership issues.
Determining whether any particular factors are empirically associated with strategic capacity is an exercise in systematic comparison. The standards for governance studies have been raised in recent years as researchers recognize the need to be able to go beyond single-case studies to middlerange theory building, or the effort at discovering the rules of the game of governance that hold for many (but not necessarily all) cities, across international barriers, and at different points in time (Mossberger and Stoker 2001, 814;Sellers 2002b;DiGaetano and Strom 2003;Gissendanner 2003). The two-case case study on which this article draws, which is discussed in the fourth section, compares the German cities of Dortmund and Augsburg. These cases differed in terms of their strategic capacity in responding in the 1980s to similar problems of industrial decline, and the task at hand is to demonstrate covariation between particular influence factors and strategic capacity in these two cities.
Although the findings are drawn from only two German cities, these cases are not gadflies; many cities in Germany and elsewhere have attemptedusing similar ideas with varying degrees of success-to ameliorate the problem of unemployment in the context of deindustrialization. Moreover, the study presented here fills a gap in our knowledge of German urban governance. To date, international comparative work on urban governance has been dominated by comparisons among cities in the English-speaking world, apart from the growing attention to France and a few other exceptions (Levine 1994;Cole and John 2001;Kantor, Savitch, and Haddock 1997;Le Galès 2002;Sellers 2002aSellers , 2002bDiGaetano and Strom 2003). Cases from Northern and Central Europe have been neglected in governance studies, a gap that is surprising given the degree of attention paid to the UK. In terms of the variation of governance contexts on dimensions relevant to theories of governance, the United States and Germany are similar in many ways that distinguish both from the UK and invite comparison. Moreover, any convincing explanation for the emergence of a particular pattern of governance in the United States and Britain must make clear why identical patterns do or do not emerge elsewhere. Germany, in particular, is interesting because it has a large number of cities, its state (Länder) and county (Kreis) governments are both salient for urban policy, its local governments have responsibility for a relatively broad range of government functions, its cities have separately elected councils and mayors, and its urban leaders have been confronted with problems similar to those faced by leaders in the United States and elsewhere (Clarke and Mayer 1986;Wollmann 2000;LeGalès 2002). Governance structures in Germany, although showing some unusual features (Clarke and Mayer 1986, 403;Gabriel 2000, 14;Wollmann 2000, 52;Strom 2001, 8;Sellers 2002b, 622), do at least allow for variation on key variables for governance theory such as governance and strategic capacities, policy output, local party systems, mayoral power, interest group influence, forms of public-private partnership, and state-level involvement. All of these qualities make German cities an interesting empirical source for theory building.
Based on the two case studies, it will be argued in the fifth section that what leaders-mayors in particular-do makes a difference in whether a city can respond strategically. Rather than simply being a case in which "agency matters," the situation is much more ironic. Leaders are increasingly powerless in a Weberian sense of "power over" other actors in a hierarchical chain of command, yet this individual impotence is completely compatible with a high influence over strategic capacity. It is the coalitional power of the group, not the command power of individuals, that is central for achieving significant policy outcomes, and leaders matter in the accumulation and use of coalitional power.
The article concludes with two broad questions for further research. Particularly intriguing is that after the mid-1980s, the strategic capacity of Dortmund declined even as its governance capacity in economic development policy remained high; this pattern calls for an explanation. The findings also raise concerns regarding the democratic responsiveness of strategic cities, for it seems that strategic capacity is more easily built in closed policy systems, at least in the early response phases. We are confronted with the question of how to reconcile democratic responsiveness with governance and strategic capacities.

GOVERNANCE CAPACITY AND STRATEGIC CAPACITY
Although governance capacity is a widely used concept and has been defined as the ability to "generate and manage resources necessary to govern" (DiGaetano and Klemanski 1993, 368), no standard indicators exist for measuring it. For reasons of expediency and because the governance approach has always centered on the ability of actors to act, governance capacity is widely held to be closely related to policy output. The higher the governance capacity in a city, the greater the power of decision makers to overcome barriers to action, as evidenced by particular actions.
The concept of strategic capacity, in contrast, taps into the process by which policies are made, implemented, evaluated, and adjusted. The idea of strategic decision making originated in business schools and then spread into public administration discourses (see Honadle 1981, 577;Bowman and Kearney 1988, 342;Kaufman and Jacobs, 1993, 13). Recent work on strategy makes the link to local governance (Jacobs 2000) and local leadership (Leach and Wilson 2000). Synthesizing these approaches and emphasizing rationalistic elements, strategic capacity is understood in this article as the ability of a governance network to make choices that (1) are based on a consideration of alternatives; (2) follow from an evaluation of market circumstances allowing for some form of calculation of the costs and potential benefits of a given policy; (3) are subject to an evaluation of implemented policies based on information about subsequent changes in local market needs, opportunities, and problems; (4) include coordination of the actors and organizations in policy making in a wide enough geographic region so as to avoid wastage of resources; (5) encompass a plan of implementation; and (6) are long term, which is defined as longer than the election cycle of elected officials who are part of the governance network. This is a maximal definition. It would be surprising for the same decision-making network to demonstrate every aspect of strategic capacity throughout many years, but, given any one particular point in time, some cities will be more strategic than others. Strategic capacity is a matter of degree, not kind.
Governance and strategic capacities were operationalized in the study discussed below using four measurable characteristics of local policy systems.
The first factor is the number of economic development areas in which a city was active and successful. Only those policies were included that were identified as "successful" by at least two local actors. Based on studies of the typical tools and targets of local economic development activity (e.g., Organization for Economic Cooperation and Development 1993;Clarke and Gaile 1998), five potential areas of activity were selected: physical infrastructure, "soft" infrastructure, business acquisition, taking care of existing businesses, and aid for small businesses. This first measurement reflects strongly on local governance capacity. The remaining three measures are designed to provide information about the policy system and the process by which policies were made. These are the specificity of the city's policy targets, the number of economic development organizations considered to be important by local actors, and the degree to which these important organizations coordinated their activity. For each measure, economic development policy systems were ranked using the preset guidelines indicated in Table 1. 2 Although a "hard" ranking for any or all measures does not equate to strategic capacity, it is plausible to suggest that a city able to generate relatively hard responses is in a better organizational position to make policy decisions strategically. These criteria yield only an imperfect approximation of strategic capacity, but no better measurements exist as yet.

FACTORS POTENTIALLY ASSOCIATED WITH STRATEGIC CAPACITY
What might enable a coalition to respond to crisis and, when it acts, to do so strategically? As I assert that governance capacity is a necessary (although not sufficient) condition for strategic capacity, our attention is drawn to the idea of governance coalitions as found in the urban regime and governance literatures (Stone 1989, 179;John 2001, 9). These are coalitions of actors who control financial and political resources-also among governmental levels-and agree to pool them in common cause. The existence of such coalitions should be a necessary condition for creating governance capacity.
Yet, because governance capacity is not sufficient for guaranteeing strategic capacity, our attention is drawn to factors shown by previous comparative urban governance research to have been relevant to governance processes. DiGaetano and Strom (2003) productively suggest grouping these factors under the rubrics of structure, culture, and agency. Structural factors are both political and economic. Relevant political structures are those that set boundaries on what local actors can do legally and those that make financial and other resources available to local governments. The political structures considered in this study are city government charters and the availability of aid from higher level governments. Economic structures are relevant insofar as they act as a trigger or constraint on collective action and include levels of local employment and unemployment, the sectoral mix of the local economy, the tax base, and social welfare expenditures. Local actors are especially sensitive to changes in economic structures, whether real or only perceived. Some work (Moss 1993;Levine 1994;Clarke and Gaile 1998) argues that crisis-the (perceived) rapid worsening of structural conditions-can be an important trigger for action and innovation in networks, and thus it is given special attention here. Crisis can also function as a trigger for innovation insofar as it leads local actor networks to abandon old problem definitions, redefine their cognitive frameworks, and find new common goals; this can happen, for example, when decision-making networks are opened to new actors who bring in different perspectives and ideas (Benz and Fürst 2002, 27). Other authors are more skeptical about the positive effects of crisis. Political leaders often rely on old solutions in crisis situations, in which consensus building is made difficult by the lack of information about the possible effects of policy innovation (Häußermann and Siebel 1994, 40;Benz and Fürst 2002, 27). A smaller body of literature focuses on the effect of local political culture on governance. For governance studies, it is useful to conceptualize political culture in terms of the effects it has on network actors. First, network actors cannot change political culture. Second, the local political culture can serve as both a barrier to action as well as a resource for leaders. For example, mayors can use patronage in nonreform cities of the northeastern United States but not in reform cities (Ferman 1985). The element of local political culture explored in this study is the local electoral (or party) system, which harbors a number of potential problems and resources for local leaders.
Agency factors in governance studies are, in fact, closely tied to more traditional issues of power and leadership. Here, our attention is called repeatedly to the role of mayors. 3 Although governance studies have a more sophisticated understanding of the nature of power-it is thought of as a creative process involving the interaction of coalition members-no empirical study of governance coalitions has yet to suggest that coalitions are leaderless. 4 It seems that the creation and maintenance of coalitions both depend on successful leaders.
The literature on leader-follower relations directs our attention to leadership skills, resources, and processes. Common kinds of leadership skills identified more or less specifically include persuasion, marshalling resources, keeping order, bargaining, being visible, manipulating conflict, coercion, creating common cause "for" or "against" something, co-opting rival actors, appealing to common ethnicity or other forms of group identity, concealment, adaptability to unexpected situations, appealing to friendship, and winning solidarity through charisma (Kotter and Lawrence 1974, 67;Ferman 1985;Reese 1997, 228). This list is by no means complete, and other authors emphasize leadership resources, defined by Dahl (1961, 226) as "anything that can be used to sway the specific choices or the strategies of another individual." These can include finances and staff in the city government, city jurisdiction over policy areas in which the mayor wants to act, mayoral jurisdiction within the city government in these areas, a salary sufficient for full-time work, personal staff, friendly newspapers and broadcasters, groups of people who can be mobilized, and solidarity (Dahl 1961;Pressman 1972, 512). The empirical work on leadership suggests that leaders use skills and resources in resolving problems in four basic, typical processes: (1) agenda setting and policy initiation, (2) policy implementation, (3) the maintenance of internal cohesiveness (especially of the governance network), and (4) representing the local government to the outside (Kotter and Lawrence 1974;Leach and Wilson 2000, 13-17).
A final, common theme in the leadership literature is that of interaction effects. Actions at one point in time change actors'motivations and resources later on, and, moreover, local coalition members can often anticipate these changes and generally be assured of future interaction with other coalition members. Aside from the opportunities for deal making across time and issue areas made possible by reiterative interaction, the fact that the same actors interact in different situations throughout protracted time periods means that the elements of leadership mutually reinforce one another in a dynamic fashion. A skillful mayor builds momentum by converting success in one process or policy area into a resource in another; success is a resource. Dahl (1961) showed how Mayor Lee of New Haven, Connecticut, used influence in one policy area to secure power in another, thus eventually securing a central coordinating role. An unlucky mayor, on the other hand, may find that actions designed to accomplish goals in one policy area interact with other actors and arrangements, causing problems in other areas (Ferman 1985, 13).
In sum, if mayors or other leaders influence strategic capacity at all, they do so through the accumulation of resources and their skillful application in the four basic processes of agenda setting and policy initiation, policy implementation, coalition maintenance, and external representation. Leaders, followers, skills, and resources are interactive, and the outcomes of this interaction are also relevant to local governance and strategic capacities.

RESEARCH DESIGN AND CASE SELECTION
Following a research design approach that has proven its value (Feagin 1990;Fuchs 1992), two cities were chosen for comparison that differed on the dependent variable, strategic capacity. As part of a larger study that included U.S. cases (Gissendanner 2001), Dortmund and Augsburg were selected from a pool of 22 large, deindustrializing cities in former West Germany. 5 Written and telephone inquiries were made with a local development official in all 22 cities to assess the relative degree of strategic capacity of each. Officials named the policy areas in which the city was active in economic development, identified targeting priorities in the 1980s, and reported on the degree to which the city was active in economic development in comparison to other German cities by their own estimation. The interviews documented variation in the degree of activity, the number of active policy areas, and the targeting specificity among German cities responding to industrial decline. Both Dortmund and Augsburg were among the active cities in economic development. Dortmund was selected to represent those that responded relatively strategically; Augsburg represents the group of relatively less strategic cities.
The logic of comparison dictates that because these two cities demonstrated different levels of strategic capacity, any of the factors identified above that were present in Dortmund but not in Augsburg are potentially important for fostering strategic capacity generally. To obtain such information, persons in each city reputed to have been central decision makers in economic development from 1980 to 1999 were interviewed in 1999. A standard questionnaire gathered interviewees' opinions regarding the most influential actors, most central institutions, and most significant policies in their city's economic development responses to decline. Initial interviewees were selected from lists of influential actors compiled by two local experts. In a snowballing technique, each interviewee was asked to compile a list of who they considered to be the most important actors in economic development in the city, and an effort was made to interview those mentioned repeatedly. This information allowed the creation of reputational rankings for each city based on the number of times a particular actor was identified as influential by another actor, weighted by the absolute number of individuals that each interviewee named. Due to resource constraints, the number of interviews conducted was limited to 14 in Dortmund and 10 in Augsburg. Secondary literature, newspaper articles, official documents, and talks with academic experts, journalists, and higher level governmental officials were used to verify and complement the interview data.
Dortmund in 1970, with 645,000 residents in an area of 271 square kilometers, was the fourth most populous city of the federal state of North Rhine Westphalia (NRW). In 1970, 42% of the city's jobs were in the manufacturing and mining sectors, 6 1.17 times the national average for cities of 100,000 population or higher. Table 2 documents the loss of manufacturing and mining jobs through 1984, whereby rising unemployment rates reflect the problem that these lost jobs were not fully compensated by service-sector growth. Further aggravating the situation was the gradual shrinkage of Dortmund's population beginning in the mid-1970s. The last concerted battle to prevent deindustrialization in Dortmund was lost in 1981, when the local steelmaker, Hoesch, announced plans to cut local production in an episode of Dortmund history styled as the Hoesch crisis.
The smaller city of Augsburg, with a population of 211,566 living in an area of 86 square kilometers, was Bavaria's third most populous city in 1970.
That year, 47% of its jobs were in manufacturing and mining sectors, or 1.31 times the national big-city average. The city lost more than 15,500 manufacturing jobs between 1970 and 1984, and unemployment in the region had risen steeply by 1983. Most of the lost jobs were in the textile sector. Despite these poor economic conditions, job losses were gradual, and other cities in Germany were always worse off than Augsburg. There was even some job growth in manufacturing from 1985 to 1989, and the city's population had Gissendanner   also expanded in the 1970s due to annexation. Nevertheless, even though all German cities were losing manufacturing jobs in this period, Dortmund and Augsburg were losing them faster, and the decline was actually steeper in Augsburg. 7 The salience of differences in the economic situations of the two cases studies is discussed below.

Local Leadership
Dortmund's government was formally shaped by the North German city charter model's weak-mayor system until reforms took effect in 1999. Only the council was directly elected, for five-year terms. The council selected the nonsalaried mayor from among its own membership by majority vote for a term of eight years. The mayor's main functions were serving as the ceremonial head of the municipality and preparing council meetings in consultation with the city manager. The weak mayor shared leadership duties with a salaried city manager and the majority party caucus leader; in fact, because this arrangement created a troika of leaders and thus invited conflict, NRW later reformed its city charter system to strengthen the mayor's position. The city manager, appointed by the council for a term of eight years, was the chief executive of the local administration and served as the legal representative of the city in negotiations with other political units. In German local councils, the caucus is an organizational unit of parliament with formal decision-making functions. As a rule, caucuses are party based. Because the caucus organization is the main instrument for enforcing party discipline in the council, the majority caucus leader possesses leadership resources for formal decision making in excess of those available to the mayor.
The mayor of Dortmund for the entire period of the study was Günter Samtlebe. He had held the post since 1972 and retired in 1999 as Germany's longest serving mayor. Samtlebe was a nationally known figure, charismatic in working-class circles but an often unloved, old-fashioned Social Democrat. He embodied the typical intermeshing of Social Democratic (SPD) party officials, local and state government leaders, and traditional industry that was typical of the Ruhr valley from the 1960s onward. In many Ruhr cities, the troika of the highest officers of city government-the mayor, the leader of the SPD caucus in the city council, and the nonelected city manager-served previously, simultaneously, or subsequently as functionaries in the local SPD organization and as representatives on the managing boards of important local industrial giants. The overlapping functions turned each top politician into a kind of self-contained governing coalition, or a "political multifunctionary" (Gerszewski and Thull 1998, 68;Goch 2001, 226). Dortmund's most prominent multifunctionaries were Hoeschianers, for many of them had connections with the local steel firm Hoesch. This group included but was not limited to Samtlebe, the SPD caucus leader Horst Zeidler, and Hermann Heinemann, who was Dortmund's representative in the state parliament. Heinemann also chaired the local party organization and served as the state's labor minister. When the Hoeschianers were of one mind, as they usually were on economic issues, they easily maintained a hegemony over Dortmund politics (Bovermann 1995, 19).

Party Politics and the Intergovernmental Context
Party organizations and federal state activism are two central conditions framing local responses to economic decline in Germany. Elections in all German cities are partisan (although the major parties do not always completely control candidate selection). The SPD held the majority in Dortmund's city council continuously between 1949 and 1999, and every Dortmund mayor since 1949 has been a Social Democrat. Until recently, Dortmund's resident population and its civil service boasted one of the highest proportions of SPD members in Germany, and it was among the most important SPD support bases for state and national elections.
At the state level, the SPD has held the reins of government either alone or in coalition continuously since 1966. The SPD's influence at the state level reached its apex between 1980 and 1995, when the SPD held the absolute majority in state government under the leadership of the popular Johannes Rau. Beginning in 1981, the state government vigorously targeted coal and steel regions like Dortmund and continuously increased its yearly budget for local economic development in these areas through the 1980s (Goch 2001).

Projects in Response to Decline
Dortmund's response to deindustrialization in the 1980s can be divided into three areas corresponding to the three projects that were considered to be successful by actors interviewed for this study: the Dortmund Technology Center and related projects (initiated in 1984 and completed in 1985-1986), labor market programs including direct aid for laid-off manufacturing workers (ongoing in the 1980s), and the renovation of the city's river port (1980)(1981)(1982)(1983)(1984)(1985)(1986)(1987). The technology center combines technology transfer and business incubator functions and was built on a greenfield site adjacent to the University of Dortmund with the intent of increasing cooperation among city officials, private firms, and the university. Within a few years after its opening in 1985, the technology center was praised nationally and is the principle reason why Dortmund was considered in the 1980s as having successfully restructured its economy ("Drei Städte: Ein Problem" 1986). Simultaneously, other Dortmund actors combined private and public funding for aiding laid-off industrial workers. This effort included the expansion of job creation schemes, early retirement, and retraining (Hennings, Kahnert, and Kunzmann 1991, 7-8). A small number of actors pointed to river port renovation as an important project in the 1980s. This was a classic hard-infrastructure project that intended to aid traditional local industries and that attracted a significant amount of federal funding. Table 3 summarizes the policy successes in Dortmund as well as the criteria used to assess the city's strategic capacity. Augsburg is discussed in the next section.

The Structure of the Governance Coalition
The ranking of Dortmund economic development actors is shown in Table 4. We see that the individuals responsible for economic development in Dortmund included elected officials from local and state governments, local and state administrative officials, private-sector representatives, and a university official. Within the Dortmund coalition, there are three groups of actors. The most prominent group is the old-guard elite (Mayor Samtlebe, majority caucus leader Zeidler, state representative Heinemann, and city manager Harald Heinze), whose influence extended beyond the single policy field of economic development. All of these individuals held top positions before the 1981 Hoesch crisis. Contrasted to this group of leaders are individuals who are specialized within the policy area of economic development as practiced in Dortmund, a group that can be further differentiated into those individuals who were active before the Hoesch crisis in 1981 and those who entered the network thereafter. The precrisis specialists stemmed from the venerable institutions responsible for local business issues in German cities: a union local, the chamber of commerce, the chamber of trades, and the city Economic Development Department. Some new specialists in the 1980s also originated in these same institutions, but due to new project activities, five new kinds of institutions came to be represented in the ranking of important economic development actors: the technology center through its director, the Fraunhofer Technology Transfer Institute, the University of Dortmund, the state government, and the steelmaker Hoesch.

URBAN AFFAIRS REVIEW / September 2004
The person reputed to have the most influence over economic development policy was Burkhard Dreher, recruited in 1985 from Berlin for the post of Dezernent 8 for economic development. Dreher arrived after the groundwork for innovation had been laid. The mayor is the only actor identified by almost every interviewee as one who was influential and who was present before 1981.

Leader-Follower Processes
Decision making in economic development occurred in a series of unplanned steps or stages in which the four processes of policy initiation, policy implementation, coalition maintenance, and outward representation are identifiable. After Hoesch announced layoffs in 1981, Mayor Samtlebe declared the situation a crisis and called together the representatives of a large number of social and economic institutions for the purpose of generating a response. These meetings became known as the Hoesch conference, and they represent the first step in the process of innovation. A "plan" was produced by the participants of the Hoesch conference that provided general guidelines intended to steer the city's future economic development efforts ("Wirtschaftsförderungsforum setzte inhaltliche Akzente" 1983;Hennings, Kahnert, and Kunzmann 1991, 5). The problem facing Dortmund leaders in the second step was one of maintaining coalition cohesion while simultaneously expanding the network. Leaders "solved" this problem by opening the network to new policy subleaders, whereby this form of delegation was probably not an intentional tactic. Rather, the plan functioned as a signal to actors working in established institutions-such as the chamber of commerce-to activate resources for policy development in accordance with the new guidelines. In this way, new members entered into the economic development coalition and championed innovative reforms. Once specific ideas materialized, old-guard leaders supported innovating policy subleaders and placated the conservative faction of the SPD, which opposed using resources for policies that did not directly benefit their working-class constituency. Leadership of the SPD neutralized internal opposition by promising compensation for laid-off workers and with the argument that although most workers would not get their jobs back through the new policy, at least their children might thereby gain a chance at finding employment in Dortmund.
In the third step of the process, old-guard leaders activated latent resources of public-sector institutions for the funding of new policy ideas.
The speed at which funding was made available for the technology center attests to the political skills of local leadership, the unity of the coalition, and the generosity of state programs. By 1985, policies had been implemented in the spirit of the Hoesch plan. Furthermore, old-guard leaders mobilized resources from the private sector and from higher levels of government, much of which was used for compensatory benefits for laid-off workers. The process continued in a fourth step as early project ideas, most especially the technology center, were outwardly represented as "successful." The perceived success of earlier policies became an important resource for later projects, as this aura made it easier for actors to attract funding for similar or complementary projects from higher level governments. Such ideas included a business park and a venture capital fund. Stage 5 was a phase of coalition consolidation. Through a bandwagon effect, successful policies reduced internal conflict within the local SPD, among the parties in the city council, and among citizen groups. Once the technology center was acknowledged as a success, groups who had previously opposed the new idea dropped overt opposition to later, similar policies.
In the ending phase of innovation, the old-guard leadership allowed policy subleaders to take over. After Burkhard Dreher entered the network in 1985, he brought a number of smaller projects to completion that built on previous successes; this is how he was able to build his reputation as the most influential person in the network. With time, this iteration structured the network disadvantageously as policy making became routine. Actors found it increasingly easy to overcome internal opposition and to generate higher level government resources for new projects similar to or connected with earlier innovations due to their growing reputation for competence and success. The old guard retreated from its intense interest in policy innovation and left no legacy of organizations with the authority to critically evaluate existing policies.
The processes of innovation in Dortmund can be condensed and summarized as follows. Although roughly chronological in order, they are not truly discrete steps.

Policy initiation, or agenda setting
Coalition members, including especially old-guard leaders, acknowledged the necessity of renewing existing policy in 1981. In consultation with representatives from a broad set of social and economic institutions, general guidelines for a response were created.

Coalition maintenance I: Strengthening and widening the governance coalition
The old guard allowed new specialists to develop policies in accordance with new guidelines. In this way, the coalition was expanded. Old-guard leaders supported innovating policy subleaders and placated blue-collar conservatives in the SPD.

Policy implementation
Resources were mobilized for implementation from the private sector and from higher levels of government, enabling policy specialists to implement agreed-upon policies.

Outward representation
Policies and projects already implemented were sold as "successful" to a number of relevant groups, including most (especially higher level) government officials.

Coalition maintenance II: Jumping on the bandwagon
Once early projects were acknowledged as successful, groups who had previously opposed the new policies found no advantage in continuing their opposition.

Reiteration
Steps 3 through 5 were reiterated as policy making becomes easier, routine, and, with time, less strategic because no new agenda-setting process took place.

The Role of the Mayor in Increasing Governance and Strategic Capacities
Neither Mayor Samtlebe nor any other single individual was the boss of the Dortmund coalition. Samtlebe claims that leadership in the network was more spontaneous than planned, more a product of the political culture of the Ruhr than of rational strategizing (Reichmann 1996, 400). In at least two phases of the process, however, Samtlebe led the network like no other thenactive individual could have done. He declared convincingly that the Hoesch layoffs were a crisis situation. This move was by no means a necessary or even a desirable one, for crisis necessitates action and by thus catapulting the situation rhetorically into a crisis, Samtlebe put his own administration under a great deal of political pressure. That his reputation and administration were so endangered, however, may have been just the kind of signal he needed to activate the latent regional network for new resources.
At the outset of the process, Samtlebe may have intended to rally resources to save the city's old industries, as predicted by observers of Ruhr political networks (Grabher 1993). Soon after 1981, however, this option was no longer feasible, and by calling the Hoesch conference, Samtlebe had effectively opened a strategy-making process that expanded into alternative economic development policies. These alternatives were indeed implemented.
Samtlebe and other leaders continued to play an important role in the next phase, when the redirecting of funds for new priorities caused political conflicts. Samtlebe bridged conservative (old industry) and progressive (new technology) SPD factions, whereby it was particularly difficult to make laidoff workers understand that a technology center, located near the university, was the appropriate response to industrial decline. Rank-and-file SPD members of the older generation maintained a skeptical distance from universities, which in Germany have an elitist tradition. Samtlebe was immune to this instinct, perhaps because of his friendship with his next-door neighbor, who just happened to be rector of the University of Dortmund.
Samtlebe really shined in the fourth phase, when local projects were sold as successful to Dortmund residents and to outsiders. Samtlebe was acknowledged as an exceptionally good salesman for the city. As the figure most closely associated with the city in the public's eye, he had a clear personal interest in selling Dortmund's activities as successful, and, as the city's official figurehead, he had the greatest opportunity for doing so.
During the first years of innovation, new organizations in economic development were created, such as the governing board of the technology center. Existing organizations such as the chamber of commerce, the local savings bank, unions, and Hoesch all gained personnel with immediate experience in the tasks of economic development. Through the realization of new projects, representatives of state government and other organizations were brought into the local policy system. As the data give an indication of when these new actors entered into the governance coalition, they provide some support for the argument that communication in Dortmund over economic development issues increased and included actors from quite a new mixture of institutions as new projects were conceptualized and implemented. This may have aided in the generation of new ideas and in the exploitation of new resources, which would represent an increase of governance capacity in the policy arena of economic development.
By the late 1980s, innovation stagnated because city policy making remained focused on issues related closely to the technology center and operations within the center became increasingly routine. Even as Dortmund stayed active in the policy areas of economic development, the original innovation effort had left no new organizations for economic development with incentives to critically evaluate previous policies and to open new agendasetting processes under routine conditions. New policy initiatives in the 1990s reverted in large part to standard economic development practices common in Dortmund in the 1970s, focusing on land redevelopment projects led and funded by the public sector. In sum, governance capacity and strategic capacity grew simultaneously in the 1980s, as evidenced by the expansion of the governance coalition to include new members with new ideas, the addition of new active policy areas and organizations in economic development, the increasing specificity of targeting, and cooperation among economic development organizations. Yet, even as the city's governance capacity remained high into the 1990s and the policy system had realized many elements of strategic capacity, it was clear by the late 1980s that the system's strategic capacity had waned. To use a distinction made by Gernot Grabher (1994), the system had adjusted well to new conditions, but it had not permanently increased its ability to adjust to new conditions.

Local Leadership
Augsburg uses a form of the South German city charter model. The South German system creates the strongest mayors in Germany (Cusack 1999), uniting all of the four main functions of the local executive under a popularly elected mayor. The mayor is the voting chair of the separately elected council and all its committees, the agenda setter in the council, the city's chief administrative officer, its ceremonial figurehead, and its legal representative (Gunlicks 1986). In all other systems, the mayor has to share these areas of responsibility with other individuals. In Augsburg, both council members and mayors are directly elected for terms of six years. Mayoral power is balanced by professional Referenten, who are political appointees of the council and supervise a group of administrative departments. Because the mayor has but a small personal staff, his policy initiatives are dependent on the expertise and assistance of the Referenten, although the mayor always has the last word. Because the Referenten are selected by the council to match the partisan distribution of council members, there is always a chance that some of the mayor's leading civil servants will be members of a competing party (Gunlicks 1986, 79); in Augsburg, this institution contributed to the generally high fragmentation of political leadership.
From 1978 to 1984, the SPD's Hans Breuer was mayor while the council was led by a Christian Social Union (CSU) majority. The elections in 1984 returned Breuer to the mayoralty, while a splinter group of the CSU joined with the SPD to form a majority council coalition. In 1991, the CSU won the council and also were able to see Peter Menacher installed in the mayor's office.

Party Politics and the Intergovernmental Context
The state of Bavaria was less active in regional economic development than NRW in the 1980s. Although the Bavarian government did have a hand in the industrialization of the state economy after the war, this transformation was completed by the 1980s. Moreover, all large Bavarian cities were economically more stable than Augsburg in the 1980s, which may be one reason why Bavaria's economic development aid for local governments did not approach NRW levels until the early 1990s.
Relations between Augsburg and the state government have been clouded by a combination of regional and party differences. Augsburg is the symbolic capital of the Bavarian region of Swabia, whereas the seat of the Bavarian government, Munich, also happens to be the symbolic capital of the region of Upper Bavaria. The rivalry between these two regions stretches back throughout many centuries. Until 1991, regional rivalries were exacerbated by political differences, because the blue-collar town of Augsburg was a traditional stronghold for the SPD whereas the state government always has been dominated by conservatives.
Augsburg's party system is very competitive by German standards. Although the SPD had the largest single caucus in Augsburg's city council and held the mayorship until 1991, it never dominated the council as in Dortmund. The two largest parties, the SPD and CSU, also held diametrically opposed positions on economic development. Although the SPD was interventionist, the CSU argued that the chamber of commerce should take the lead in economic development issues and shunned economic development activism on promarket ideological grounds. Leaders of the SPD were willing to increase economic development activities, but this activism implied the relegation of chamber leaders from their leading role in the agenda-setting process, a role that local business interests confidently played. Thus, chamber members may have been wary of local Social Democrats. For similar reasons, union and business leaders were also split, in stark contrast to their corporatist familiarity in Dortmund.

Projects in Response to Decline
Projects considered successful by the central decision makers in Augsburg economic development in the 1980s included the 1982 construction and 1988 expansion of a convention center (a public-sector initiative) and a venture capital fund organized by a prominent banking executive and member of the chamber. The convention center was the only project considered successful by a majority of interviewees. This was, however, a low-risk project organized by standard economic development institutions on the public-sector side and was funded mostly by the state government. The Augsburg effort is summarized above in Table 3.
The Economic Development Department and the chamber of commerce claimed successes in the 1980s in routine tasks, including two large acquisitions. Because these acquisitions were initiated by private firms, however, they are not considered here to be part of the city's economic development policy. Several risky and complex policy initiatives failed in some way or another. The city successfully guided an effort to save SWA, a large textile firm, from bankruptcy during the early 1980s. The recapitalization of SWA required a complex financing arrangement that involved all levels of government, but the firm went bankrupt a few years after the massive effort. In another disappointment, plans made in 1988 to convert a historic factory building into a business incubator never advanced to implementation.

The Structure of the Governance Coalition
The Augsburg ranking (Table 5) includes the same kinds of institutions and organizations as in Dortmund. The number of institutions represented in the Augsburg network, however, is smaller, and no institutions appear other than those traditionally responsible for economic development policy in all German cities. This suggests that economic development in Augsburg was an area of routine policy activity for the city, an impression strengthened by the nature of the development projects that the city implemented. Yet the network includes, in contrast to Dortmund, three private business leaders who were recognized as influential independent of their membership in the chamber of commerce. Half of the reputedly influential members of the coalition were private-sector based, which substantiates the reported dominance of the chamber over agenda setting in economic development. All other ranked persons had positions in government.
For the time period of this study, the Augsburg coalition was composed of two subnetworks. This stands in contrast to the Dortmund network, which, although consisting of three distinct kinds of actors, functioned as a single coalition as it pursued common (if only vaguely connected) goals. In Augsburg, one network centered in the chamber of commerce, while the other formed around city council actors, as evidenced by the fact that business interviewees typically identified businesspeople as more influential whereas politicians identified politicians as more influential. The only actors consistently acknowledged as influential by both business and government actors were Economic Development Department Director Hannes Buss and Mayor Hans Breuer.

Leader-Follower Processes
There were no processes of policy innovation and implementation in Augsburg comparable to those in Dortmund. A coalition existed that united politics and business in the city, but this coalition focused on routine tasks. Roundtable agenda-setting efforts did occur in Augsburg. University professors, union leaders, and city officials came together in 1981 and again in 1987 in "labor market workshops." Their initiator was not the mayor but Martin Pfaff, an SPD official and faculty member of the University of Augsburg. Yet these efforts remained without effect. Due to the competitiveness in the party Gissendanner / URBAN GOVERNANCE IN GERMANY 65  system, the task of uniting an economic development coalition around common goals and projects was particularly difficult. In this situation, the chamber dominated agenda setting in economic development policy, but it ran into implementation problems because of the divisiveness of the political side. The economic downturn of the 1990s changed this constellation gradually. By 1997, on the heels of elections in 1996 in which economic development was a central issue, development policy had become the top priority in political circles for both parties and across the public-private divide. Yet, only as the time frame of the study came to an end were successful efforts at policy initiation observable. The maintenance of coalition cohesion was blocked by the same difficulties as those affecting agenda setting and policy implementation. Moreover, parallel to the party differences that split the network in the 1980s and 1990s, there were open conflicts among residents between prodevelopment interests and antidevelopment neighborhood groups over projects such as airport expansion and the redevelopment of the city's dilapidated textile quarter ("Textilviertel" 1992; "Industrie soll nicht weiter wachsen" 1992). Thus, although economic development issues in Augsburg were contested with more democratic openness than in Dortmund, this brought the disapproval of business leaders and prevented political leaders from delivering city resources for development projects. The task of outward representation in the field of economic development by and large was conducted by the chamber of commerce, whose ties to state politicians were much better than those of local political leaders in the 1980s, when the city had an SPD mayor.
During the time period of this study, Augsburg engaged in economic development activity, but in routine forms and without the need for publicprivate cooperation. The public sector and the private sector engaged in separate activities. New organizations were created and the Economic Development Department was reorganized, but no organizational form emerged that successfully bridged the public-private divide. As a consequence, although there was a network or governance coalition in the policy field of economic development, this coalition was fractured and demonstrated relatively low governance and strategic capacities.

HOW LEADERS INCREASE GOVERNANCE AND STRATEGIC CAPACITIES IN GERMAN CITIES
The cases of Dortmund and Augsburg in the 1980s suggest some propositions about how governance and strategic capacities are built in German cities responding to social problems, if we concentrate on factors predominant only in the more strategic city of Dortmund. Of particular concern are economic conditions and crisis, the political-structural context created by city charters and intergovernmental aid, cultural factors centering around party ties and low party competition, agency factors including the existence of governance coalitions as well as the resources of solidarity and friendship, and, finally, momentum.
First of all, the study's negative findings are quite interesting. The comparison supports the conclusion that governance capacity did not depend solely on formal institutional resources stemming from city charter systems, because Augsburg's mayors were clearly in a stronger institutional position than the more successful Mayor Samtlebe. Also, the mere presence of a governance coalition seems to have made little difference. As evidenced by the reputational rankings, each city had a coalition of the sort predicted by the governance approach. The group of most influential persons in both systems consisted of representatives of the same kind of organizations, led by the director of the Economic Development Department and with both mayors holding exactly the same high position in their influence rankings.
Turning to the structural factors differentiating the two cities, we see that economic decline in Dortmund was certainly more serious than in Augsburg through the mid-1980s (Table 2). Unemployment rates and welfare expenses were higher in Dortmund, whereas tax revenues were lower. Additionally, Dortmund benefited from more state aid in the 1980s. Given the research design of this study, there is no way of gauging whether these structural factors were more important than cultural or agency factors; for this purpose, it would be necessary to show that all cities in the Ruhr valley, for example, had higher governance and strategic capacities than cities in Bavaria due to state aid and more intensive crises. Yet the two-city comparison does show that structural factors are not the only ones at work.
There were clear differences in the local political cultures of Dortmund and Augsburg. In the SPD-dominated Dortmund, cooperation among top network elites was eased by common party membership and a similarly prointerventionist attitude toward economic development. Because the SPD controlled the state government as well, party ties were useful in getting information about state aid and garnering a share of it. Moreover, party ties bridged the institutional barriers that separate the key political actors in the North German city charter system: the mayor, the majority caucus leader, and the city manager. Dortmund's mayor faced no serious electoral competition. Party loyalty also helped coalition leaders push through economic development policies that did not serve the strictly defined interests of the SPD blue-collar rank and file. Several Dortmund interviewees also pointed to another aspect of local political culture-a mutual solidarity stemming from decades-long corporatist bargaining among government officials, labor leaders, and the steelmaker Hoesch.
The examination of agency factors in the two cities supports the proposition that broad economic structures and political culture affect local governance and strategic capacities only through local agents. Although economic conditions were certainly poorer in Dortmund, local leadership was required to turn these poor conditions into the politically useful resource of crisis. Mayor Samtlebe was on the forefront of efforts in Dortmund to turn layoffs and structural decline into a political issue in 1980 by styling the Hoesch layoff announcement as a crisis, implying that concerted action was necessary to avert a catastrophe. Crisis was a lasting resource for mobilizing key institutional actors, the media, and the public in the early 1980s in Dortmund. As a man draped in local traditions, Mayor Samtlebe was also well positioned to evoke the bonds of solidarity in garnering support for new policy directions. Another important resource for the Dortmund network was friendship. There was a personal form of friendship exemplified by Dortmund University Rector Paul Velsinger and Mayor Samtlebe. Their status as neighbors broke down the subtle class animosity that normally distances many non-university-educated SPD leaders from university officials. This aided coalition cohesion in Dortmund, because the success of the technology center depended on the cooperation between city hall and the university. There was also a more social form of friendship that helped bridge those party divisions that did exist in the network. For example, Velsinger and chamber of commerce CEO Walter Aden were CDU members but were linked socially into the Dortmund network through the Lions' Club and other avenues.
Momentum was a final and very important resource for Dortmund leaders, and the case illustrates well how momentum is created through reiteration and within the four leader-follower processes discussed above. Success, in particular, is a central element drawing our attention, for it helps to explain why the (failed) SWA bailout in Augsburg did not make a long-term contribution to governance capacity there. Dortmund experienced early success in its processes of innovation, and these successes snowballed into later processes such that positive outcomes in the policy initiation and implementation phases made the process of internal coalition maintenance and outward representation much easier. Importantly, these processes led to the implementation of the technology center, which in turn was widely acclaimed. Early success made it easier for the governance coalition to initiate and implement policies that could be styled as building on earlier ideas, which then resulted in the generation of more resources and further placated dissent within the coalition. Had early policy initiatives been viewed as failures, SPD traditionalists would have been confirmed in their views and may have tried to change the goals of the coalition.
In the contrasting "control" case of Augsburg, the barriers to collective action were higher, and mayors' resources for overcoming them were fewer. The lack of a clear crisis in Augsburg helps explain why the city did not mobilize a broad, strategic response to deindustrialization. Yet, hypothetically at least, Augsburg leaders could have used the bailout of a major textile plant in the early 1980s in the same way that Samtlebe used the Hoesch layoffs. It seems reasonable to insist that economic structures alone do not explain the lower strategic capacity of Augsburg in the 1980s. Economic problems must be converted by local actors into a politically mobilizing issue, calling for a general reevaluation of the network's goals. Whether a leader is able to create a salient issue out of economic decline depends on local political culture and local agency. Augsburg's political culture did not provide the kinds of resources for redefining economic development as in Dortmund. Although party loyalty, friendship, and solidarity were certainly present in Augsburg, these did not serve to bridge institutional divides. High party competition weakened the institutionally strong mayor, made the task of generating support from the state government (which was controlled by the mayor's partisan opponents) more difficult, made the policy area of economic development into a hive of controversy with high political risks for policy innovators, and prevented politicians from delivering resources for the implementation of projects supported by business leaders. To cap off these problems, several Augsburg interviewees argued that the high value set by local residents on self-sufficiency hindered efforts to create new partnerships for economic development across class and professional lines.
The comparison of Dortmund and Augsburg supports some generalizing propositions about the role of local agency in creating strategic capacity. The first proposition is that governance capacity depends not on institutions alone but also on the ability of leaders, including especially the mayor, to create and apply resources for the purpose of persuading other important individuals to support common policies or projects. 9 None of the structural and cultural resources noted above operated independently of individuals. On one level, particular individuals did make a difference in Dortmund. Resources like state government aid, party ties, solidarity, friendship, and momentum were latent until exploited by a skillful leader in a particular decision-making process. On another level, Dortmund's coalition was powerful not because of the power of one or more leaders; rather, the power of the group was built up over time through mutually reinforcing, successfully undertaken leader-follower processes. Power builds, if at all, through interaction among leaders and followers, and this explains why differences in governance and strategic capa-cities do not depend on economic structures, static resources like institutional position, or the leadership skills of single individuals.
If we look closely at the kinds of informal resources that were important for the development of governance capacity in Dortmund in the 1980s, we see that the mayor in particular had a central role to play, even in the presence of other leaders. This supports a second generalizing proposition: Informal resources available only to mayors, which emerge from their symbolic role as local figureheads, enable them to make unique contributions to the creation of governance capacity in local policy-making systems. The formal institutional scholarship is misleading in giving the impression that the position of ceremonial head is the least important of all of the mayor's functions. This exact function is the most important power resource of the mayor. Not only are symbols and gestures generally central to the creation of group power (Wolman 1988;Elkin 1987), but when mayors also happen to somehow symbolize local traditions through their personality, they are in the best position to make use of symbols and gestures. Symbolic figureheads can better appeal to party loyalty and solidarity, creating for themselves the legitimacy to get involved in everyone else's business to bind different interests, broker between rival groups, and sell the coalition's decisions to the public. Not every mayor is interested in leading policy innovation, but because of their visibility, they do stand to reap the greatest political gains from an increase of local governance capacity.
All of this suggests the intriguing possibility that mayors endeavoring to support innovation also tend to increase strategic capacity unintentionally. This unintentional effect arises from the fact that, to increase governance capacity, mayors activate and integrate new people, resources, and ideas within the existing coalition. Brokering, bridging, and activating are essential skills in coalition formation and are necessary to get things done. If mayors are successful in these efforts, a necessary result is that they expand the membership and the cognitive horizon of a previously more specialized and routinized policy system. When local coalitions are expanded in this way, they are outfitted with more information, can select from a wider variety of policy ideas, and have access to more financial resources. Thus, they are more likely to act strategically in subsequent policy-making situations than at previous times, regardless of whether the mayor wants this or not. Moreover, mayors are also among the few actors with an interest in the adoption of a strategic plan for the city, which is an important element in strategic capacity. Mayors like such plans, because, if they are accepted, they legitimate mayoral intervention in all areas of a city's activity (Leach and Wilson 2000, 77). Thus, given sufficient resources and the luck of early success, activist mayors will probably make a significant contribution to both the governance and strategic capacities of their coalitions.
The tendency of mayors to build strategic capacity is subject to limitations and is likely to be short term at best. One important limitation is that leaders are not always free to choose policy areas for innovation, despite the fact that the freedom of executives to focus efforts in response to new problems is an important component of strategic action. The Augsburg experience, in which the mayor was not able to rally other actors around a policy response to decline, shows how the ability to pick issues can be limited by the local party system. Although economic development would have been an attractive issue for the purpose of rallying the governance coalition around a common cause for the Social Democratic mayor Breuer, as it was for his colleague in Dortmund, Breuer faced the problem that party ideology-coupled with high party competition and a traditional cleavage between the public and private sectors over the control of economic development policy making-made this a politically divisive issue in Augsburg. Emphasizing economic development would have made governance more difficult in Augsburg.
A limitation on Mayor Samtlebe's contribution to strategic capacity was that, as the incumbent mayor, he had no incentive to collect or process information that did not confirm the successful image of the city. The differentiated and critical assessment of one's own policies is necessary for strategic action but not particularly useful in winning elections, nor does it help in building momentum within a coalition.
One last proposition suggested by the two-case comparison of Augsburg and Dortmund is that if the party and ideological cohesion of political actors in local coalitions is high, the participation of private-sector actors in governance coalitions is made easier. If the political sphere is fractionalized, intersectoral cooperation is likely to suffer. This reflects the urban regime argument that both public and private sectors must be able to deliver resources for common goals if governance is to take place. At the same time, this finding modifies the received wisdom about local economic development politics in Germany, which suggests that because German local politicians are not as dependent on local businesses, they tend to insulate themselves from private-sector demands. Dortmund shows that this same political power can also be used to open local governance coalitions to new kinds of business interests.

QUESTIONS FOR FURTHER RESEARCH
Looking beyond the 1980s, we are confronted with the problem that Dortmund realized some elements of strategic capacity, peaking in the mid-1980s, but saw its strategic capacity wane thereafter even as it remained very active in economic development. Why did strategic capacity rise with governance capacity at first, only to fall in the context of continuing high governance capacity? The study spurs two paths of speculation. First, the structural conditions for economic development planning changed for Ruhr valley cities during the course of the 1980s and may have set new constraints on innovation in Dortmund. In 1985, a year after Dortmund's technology center was initiated, the state government explicitly favored a number of particular tools and targets in economic development in its mechanisms for distributing state aid, including more funds for technology centers. Moreover, special aid was made available to subsidize new job-creation schemes intended to expand on earlier economic development projects (Goch 2001, 266-67). Because Dortmund was well positioned to receive a portion of these funds and was an acknowledged leader in the field, the incentive to do something completely new would have been low. In sum, the case may be that the innovative standard that Dortmund helped to set was soon institutionalized at the state level; this would have served to stall innovation in Dortmund while supporting further activity there.
A second speculative explanation makes use of agency factors and follows from the hypothesis that both an initial increase of strategic capacity and its subsequent capping are codetermined by the motives and abilities of leaders and their interaction with the other members of coalitions. If this is the case, then mayors in particular play a central role in the timing of the increase and decrease of governance and strategic capacities due in large part to their high public visibility. Mayors have incentives to respond to urgent problems and thus aid in building local governance capacity. In doing so, they can open their networks to new actors, who bring in new ideas and thus indirectly increase strategic capacity in the early phases of response. The problem here is that mayors have no incentive to reevaluate successful policies or to reopen the agenda-setting process once the immediacy of the original crisis has evaporated. Governance capacity can stay high because the network finds it easier to do the same kinds of things over again after early projects experience success. Yet the strategic qualities of policy making evaporate.
Finally, the Dortmund experience gives cause for the disconcerting speculation that democratically responsive networks are less likely to act strategically in response to social problems. As in the cases of New York and Chicago's responses to fiscal stress in the 1970s, coalitions may be able to generate strategic capacity "only when they are protected from the vagaries of electoral politics and the demands of political interest groups" (Fuchs 1992, 279). We are also reminded of Stone's (2001) suggestion that governance requires the unification of a governance coalition around a central, unifying purpose. The more groups involved in a coalition, the more difficult it will be to find a unifying purpose. Nonetheless, the case of Dortmund suggests one potential way to mitigate this apparent dilemma. The Dortmund coalition of traditional coal-and-steel Ruhr valley elites proved to be willing to accept an expansion of the local governance coalition in economic development policy through the inclusion of actors from institutions not previously represented in the traditional network, including the university and high-tech businesses. Although this kind of representation is a far cry from democratic responsiveness, it did open the elite network to a somewhat greater spectrum of interests. Future governance research should look for governance coalitions that were even more representative but still able to preserve their governance and strategic capacities.

NOTES
1. The governance discussion within which this article is set focuses on the state and its actions, whereby governance is "defined as a process of coordinating actors, social groups, and institutions to attain particular goals, discussed and defined collectively in fragmented, uncertain environments" (Le Galès 2002, 15). Rhodes (1997, 15) uses governance to refer to "self-organizing, interorganizational networks characterized by interdependence, resource exchange, rules of the game and significant autonomy from the state." John (2001, 9) defines governance as a flexible pattern of public decision-making based on loose networks of individuals. The concept conveys the idea that public decisions rest less within hierarchically organized bureaucracies, but take place more in long-term relationships between key individuals located in a diverse set of organizations located at various territorial levels. Sellers (2002a, 9) uses the term "urban governance" to underscore processes delimited by the political economy of a territorial unit like a city, such that the focus is placed on actions and institutions that affect its regulation. Regulation can be imposed from outside the city or initiated from within, either formally or informally; its actors can come from many organizational bases including local and supra-local governments and other kinds of public organizations, pressure groups, businesses, and public-private partnership organizations. For more discussion and debate, see Pierre (2000).
2. The easy to very hard terminology is not original (cf. Mitchell-Weaver, Miller, and Deal 2000); it implies the argument that for each element of decision making, some outcomes are harder to realize or less commonly attempted than others. It should be "easier" or more common to engage in only one kind of development policy without any targets other than the general goal of job creation, conducted by only one organization, and with no need for coordination among actors or organizations. This, in fact, describes the traditional form of economic development activity of almost every large Western European and U.S. city. It is "harder" or less common for a city to engage in a diverse set of activities, focused on specific targets and enjoying the support of many organizations who coordinate their activities.
3. Conventional narratives of city politics document the influence of mayors in great detail (see Royko 1971;Puhl 1990). Academic studies of local governance also nearly always show mayors or their functional equivalent to be in the group of a city's most influential individuals, irrespective of the methodological approach used (e.g., Hunter 1953, 63;Dahl 1961;Stone 1989;Sbragia 1990, 57-58;John 2001). The influence of mayors and other top executives for policy output has been documented in cities responding to economic decline (Judd and Parkinson 1990), in German cities implementing administrative reforms (Banner 2001, 203), and in U.S. cities generally depending on their leadership style (Kotter and Lawrence 1974). Moreover, the potential influence of mayors is now more commonly recognized in Germany, Italy, and the UK due to institutional reforms and changes in the economic and political structures of governance (Le Galès 2002).
4. In fact, Clarence Stone (1993, 1) elaborated on his concept of urban regimes to emphasize the centrality of political leadership, which he defines as "a creative exercise of political choice, involving the ability to craft arrangements through which resources can be mobilized, thus enabling a community to accomplish difficult and nonroutine goals." 5. The pool of 22 cities comprises the only German cities that had a population of 100,000 or higher in 1970, had a higher-than-average concentration of employment in the combined sectors of manufacturing and mining in 1970, and fared worse than the national average in both manufacturing and total employment change between 1970 and 1984.
6. The German statistics aggregate jobs in utilities with mining. Mining employment in Augsburg was negligible but is reported for reasons of comparison. 7. This can be shown using the manufacturing and mining location quotient, which is the ratio between the proportion of employment attributable to manufacturing and mining locally to the same proportion in all cities of 100,000 population and above (cf. Flegg, Webber, and Elliot 1995, 549). Both Dortmund and Augsburg's manufacturing and mining location quotients declined from 1970 to 1984, which indicates that they were losing jobs in these sectors faster than other big German cities. 8. A Dezernent (the functional equivalent in Augsburg is a Referent) is a politically appointed administrator who is usually responsible for a small bundle of policy areas.