Abstract

The paper argues for an extension of both institutionalist as well as interest-based approaches to the study of comparative industrial relations. Taking concessionary bargaining in Germany and the United States as a comparative test case, it can be shown that a purely institutionalist analysis that stresses the persistent structural differences and predicts divergent IR outcomes is not sufficient to explain the high degree of similarities as well as the processes of convergence. The paper therefore proposes an extension of the institutionalist framework by explicitly incorporating bargaining power as a central variable in explaining international differences in IR policies and practices.

Introduction

In an insightful article Wailes, Ramia, and Lansbury reviewed the strengths and weaknesses of "the two main approaches to the impact of international economic change on national policy matters" (Wailes et al. 2003, 617), that is, the "new-institutionalism" and the "interest-based" approach. While both of these approaches have their virtues in explaining policy outcomes in the social sciences, Wailes et al. argue that each provides only partial explanations and is open to criticism. To summarize their argument, they blame the by now hegemonic new institutionalist paradigm of adopting a form of structural determinism that is unable to explain the sources of change. Insofar as the institutionalist framework stresses the persistence of diversity in policy outcomes that primarily result from path dependencies, it tends to understate the degree of similarities and the processes of convergence. While on the other hand Wailes et al. express sympathy toward the much less-dominant interest-based approach to the study of comparative industrial relations (IR) (Pontusson and Swenson 1996), they still acknowledge its shortcomings in that "interest-based models tend to be excessively parsimonious in deriving interests . . . and ignore the role that institutions may play in structuring politics and determining preferences" (Wailes et al. 2003, 623). Based on their review of the literature, Wailes et al. argue for an integration of the two approaches, as they see them as complementary, so that "the adoption of insights from one is likely to improve the explanatory power of the other" (Wailes et al. 2003, 623).

This paper argues for a further extension of a theory to the study of comparative IR that incorporates bargaining power as a central variable in explaining international differences in IR policies and practices. While institutionalist variables are necessary for defining the constraints and opportunities within which the actors adopt their strategic choices, the "interest" variable even in Wailes et al.'s approach remains somewhat ambiguous. Specifically, when referring to "interests" as an explanatory variable, the authors underemphasize the fact that industrial relations are characterized by antagonistic interests between management and labor. It remains unclear as to which of the competing interests among management and labor should be expected to prevail. This paper proposes to incorporate a power dimension to the international study of IR that might offer a remedy to this problem because it rests on the insight that IR outcomes are the result of strategic interaction among rational actors. This approach is applied to a U.S.–German comparison of concessionary bargaining in the context of economic adversity and employment cutbacks. It can be shown that differences in the extent of employee concessions as well as employer quid pro quos can be explained by a power approach that is informed by institutionalist analysis. Other than a purely "interest-based" approach, the bargaining power approach is capable of identifying systematic power asymmetries that help to explain which of the often diametrically opposed material interests among employers and employees prevail in different institutional contexts.

The paper is structured as follows: It starts with a presentation of recent survey data on concessionary agreements in Germany, which will be compared to the literature on concession bargaining in the United States (section 2). Section 3 sketches out a possible "Varieties-of-Capitalism"-type institutionalist explanation for the dissimilarities in the bargaining outcomes and criticizes it as being overly simplistic and incapable of explaining change in the actors' behavior that is potentially damaging to a purely institutionalist analysis. Instead, a concept of bargaining power is presented in section 4 that relates asymmetric power relations to underlying asymmetries in the actors' dependence in an exchange relationship. Section 5 demonstrates that a modified approach that takes account of both institutional differences as well as power asymmetries is fruitful for explaining concession bargaining in the United States as well as the decentralization of collective bargaining and the emergence of concessionary pacts in Germany.

Concessionary Bargains in the United States and "Pacts for Employment" in Germany

Concession Bargaining—The U.S. Experience

Concession bargaining in the United States as defined by Bell to include "some kind of 'giveback' from workers to management in the terms of some earlier specified agreement" (Bell 1995, 260) reached its heyday in the first half of the 1980s (D. Mitchell 1994). Under pressure stemming from a prolonged recession, increasing international competition, and deregulation, workers in such union strongholds as the airline, automobile, or tele communications industries had to agree to concessions including wage and benefits cuts or freezes, the introduction of two-tier wage plans, or the modification or elimination of cost-of-living allowance (COLA) clauses. Instead of basic wage increases, many unions could only secure lump sum bonuses or deferred compensation packages such as gain-sharing and stock option plans that only provided a payout in case the company survived and reached a certain level of future profitability. In many instances work rules were relaxed and job classifications reduced in an effort to improve labor productivity by increasing internal flexibility. Based on Bureau of Labor Statistics (BLS) data, Bell estimates that in each year from 1982 to 1988 between 48 and 76 percent of all workers were affected by at least one of these concession instruments (Bell 1989).

In exchange for often significant wage concessions, workers generally received only marginal (if any) quid pro quos from management (Cappelli 1983; Chaison and Plovnick 1986). For the second half of 1982, Cappelli estimates that management concessions could be secured in about one third of concession cases (Cappelli 1983, 298). Furthermore, employer givebacks were often merely symbolic in that management promised to share in the sacrifices or that union leaders were put on company boards of directors without giving them voting rights and the power to effectively influence corporate strategic decision making. In those instances where job security improvements were negotiated, management for the most part only agreed to suspend employment cutbacks instead of giving explicit employment guarantees for the future (Cappelli 1983). The general picture of asymmetric concession bargaining trade offs holds true during the most recent negotiations in the airline industry. Facing the imminent threat of bankruptcy and liquidation, unions repeatedly agreed to wage and benefit cuts as well as work rule concessions at carriers such as, among others, US Airways and United Airlines that drastically reduced workers' total compensation and lowered benefits for present and future retirees. Even with these massive cutbacks, though, the workforces did not get formal employment guarantees in return. If anything, management executed additional employment reductions and outsourcing in an effort to further trim labor costs.

Germany

Compared to the United States, company-specific concession bargains in Germany are strikingly different with respect to the extent of workers' concessions and the give-and-take symmetry. Although worker representatives in Germany today face similar economic adversity as U.S. unions did in the 1980s, they are considerably more successful than their U.S. counterparts in securing quid pro quos from management. As a recent representative survey of works councils in 2,477 establishments in the private sector demonstrates, nearly one in four (23 percent) establishments with a works council applied a concessionary agreement in 2003 (Massa-Wirth and Seifert 2005). But whereas U.S. concession bargains are predominately targeted toward cost cutting by lowering wages and benefits, the majority of German concessionary contracts aim at increasing labor productivity and efficiency, thus (at least up until recently) minimizing the negative financial consequences for employees. Working time measures such as reducing or extending contractual work hours or flexibly aligning work hours to demand fluctuations by means of introducing working time accounts play a prominent role in these agreements. Overall, survey data indicates that 76 percent of the companies applying a concessionary pact use one or more instrument for modifying working time policies whereas only 42 percent adjust compensation. Moreover, the average amount of financial concessions is considerably smaller than in U.S. concession bargains. Instead of cutting basic wage rates, company-specific bonuses are reduced or scheduled wage increases are delayed. Multivariate analyses show that the choice of concession instruments significantly depends on the economic environment, with companies experiencing demand fluctuations using measures to adjust work hours while establishments that suffer from weak profits predominately resort to direct monetary instruments that reduce labor costs in the short run (Seifert and Massa-Wirth 2005).

Speaking in terms of employer givebacks, one-sided concession agreements, which prevail in the United States, are fairly rare in Germany. Where worker representatives agreed to concessions, they could secure explicit employment and/or production guarantees in 87 percent of the cases. By far the most widespread guarantees have been explicit no-layoff and no-plant-closure guarantees, which could be found in 71 and 44 percent of all concession cases, respectively. For example, workers at Volkswagen and DaimlerChrysler agreed to concessions in 2004 only after the companies offered long-term employment and production guarantees until 2011 (Volkswagen) and 2012 (DaimlerChrysler).

Explaining Dissimilar Bargaining Outcomes—A Case of Institutional Variation?

The considerable differences in the exchange relationship between U.S. and German concessionary bargains—that is, the virtual absence of explicit and long-lasting employment and production guarantees in the former and the dominance of two-sided give-and-take agreements in the latter IR sys-tem—might at first sight suggest an institutionalist analysis as Germany and the United States represent "most different cases" with respect to their IR institutions and policies (Katz and Wheeler 2004; Keller 2004). As Germany and the United States represent two distinct "social systems of production" (Hollingsworth 1997a), with Germany being a quasi-prototype for a "coordinated market economy" whereas the United States represents an "uncoordinated" or "liberal market economy" (Hall and Soskice 2001), dissimilarities rather than similarities in collective bargaining outcomes would be expected. Quasi-symmetric employment pacts in Germany might be explained by the fact that companies are embedded in quasi-corporatist mechanisms of non-market-based economic coordination at the national, industry, and company level that serve as "beneficial constraints for rational voluntarism" (Streeck 1997) and provide incentives for adopting long-term business strategies based on "diversified quality production" (JŸrgens 2003; Streeck 1992). As German companies heavily rely on highly skilled and motivated workers and benefit from the continued cooperation of workers and their representatives, they refrain from demanding excessive and unconditional concessions during economic slumps and instead provide for contractual guarantees for future employment. Companies are afraid of breaking up the fragile "productivity coalition" (Windolf 1989) by playing hard ball and opportunistically terminating implicit contracts. Much of the literature on the rationale behind nominal wage rigidity and on how employers actually benefit from upholding the quasi–gift exchange in the "psychological contract" by providing continued stable employment follows this kind of reasoning (Bewley 1999; Franz and Pfeiffer 2002; Guest 2000).

The U.S. political economy, on the other hand, with its voluntaristic and market-based logic of economic coordination, is said to favor business strategies based on Tayloristic mass production with a short-term focus on cost cutting rather than long-term considerations like productivity and quality improvement (Hollingsworth 1997b). Compared to German employers, U.S. companies rely much less on the continued cooperation of highly qualified workers with firm-specific skills, so that aspects such as fairness in employment relations and a long-term commitment to the workforce only play a marginal role in corporate strategic choices. Even during temporary cyclical slumps, U.S. companies resort to (mostly seniority-based) layoffs rather than keeping the staff level constant (Abraham and Houseman 1993). In the unionized sector of the economy (as this is the only sector where collective concession bargaining by definition can occur), the focus on arm's length employment relations is exemplified by adversarial labor management relations and union avoidance strategies by management (Kleiner 2002). Because collective bargaining exclusively takes place at the company level, the cyclical variation in union wages is much higher in the United States than in Germany, because during economic slumps management does not hesitate to press for concessions while during economic upswings unions use their bargaining leverage to demand significant wage improvements. Quasi-symmetric employment pacts where both sides share in the sacrifices would thus contradict the confrontational logic prevailing in the United States (Block and Berg 2003).

But a purely institutionalist approach to an analysis of U.S. and German collective bargaining outcomes during economic downturns fails to provide satisfactory answers to some puzzling questions. For example, why did German employers and business associations recently launch such full-blown attacks on collective bargaining and codetermination institutions if they purportedly benefit so much from cooperative labor-management relations (Thelen 2000)? And why, according to recent survey data, are concessionary pacts in Germany significantly more often found at large as compared to small establishments, and why have large corporations been the forerunners in negotiating concessionary bargains (Massa-Wirth and Seifert 2005)? According to conventional wisdom, large corporations are the backbone of the German IR system as they are union strongholds with self-conscious and professionalized works councils and militant workforces. Why did labor not succeed in fighting back concession demands at these companies? And why on the contrary did management at large corporations press so hard for concessions even though they risk the breakdown of consensual labor relations on which their business strategy of "diversified quality production" crucially depends? It becomes obvious that these anomalies are potentially damaging to purely institutionalist concepts that do not take account of the power resources and the microeconomic preferences of the relevant actors. The following section sketches out a power concept that specifies the circumstances under which actors with diverging preferences are able to achieve their goals.

Bargaining Power Explanations

Unfortunately, off-the-shelf concepts of bargaining power in the field of IR are difficult to apply in concession bargaining contexts. Bargaining power has traditionally been conceptualized as the "relative willingness and ability to strike and take a strike" (Livernash 1963, 18). This may hold true in "normal" bargaining settings where unions press for improved wages and working conditions by threatening to withhold production. But during economic slumps or in a context where management threatens to cut back on employment or even close down the facility, the union's strike weapon dramatically loses its effectiveness. Sociological approaches to power relations in the tradition of Emerson's "power dependance" theory (Emerson 1962) offer a possible solution to this conceptual dilemma. They view bargaining power in more general terms as a function of the relative dependency of the two parties in an existing exchange relationship. Here "the power of one party is based on the opponent's dependence on that party" (Bacharach and Lawler 1981, 60), where dependence is itself a function of both a party's alternative outcome sources as well as her commitment to the exchange relationship. Given these assumptions, "a party's bargaining power should be greater, the lower the opponent's alternatives and the higher the commitment of the opponent to the outcomes at issue in the relationship" (Bacharach and Lawler 1981, 63) so that "an increase in the ratio of A's alternatives or commitment to B's alternatives or commitment increases B's relative bargaining power" (Bacharach and Lawler 1981, 210).

The dependence approach does not pretend to offer a deterministic concept by which bargaining outcomes could unambiguously be inferred from "objective" constraints but rather leaves room for bounded rationality. Bacharach and Lawler acknowledge that "the cognitive use and manipulation of 'objective' conditions is more important to an understanding of bargaining than the objective condition themselves" so that "conveying an impression of power will yield the same consequences as having 'real' power; manipulating an opponent's perception of the power relationship will have the same effect as an actual change in that relationship" (Lawler and Bacharach 1986, 195). So the by now familiar concepts developed by new institutional economics such as opportunistic behavior stemming from information asymmetries, incomplete contracts, and principal-agent problems can be easily incorporated into the basic dependence approach to bargaining power (Furubotn and Richter 1997). Indeed, information asymmetries play an important role in concession bargaining contexts as management has an incentive to overstate the economic crisis and labor cost disadvantages as well as its willingness to cut back on employment or relocate production.

A Modified Power-Based Analysis

The two concepts briefly summarized in sections 3 and 4 are now applied in an attempt to explain concession bargaining outcomes in dissimilar institutional and socioeconomic contexts. While differences in IR systems in particular and the wider political economy in general define the institutional "constraints and opportunities," bargaining power theory makes us aware of the strategic choices of the key actors and the relevant factors that determine an actor's bargaining power in an exchange relationship. The subsequent sections briefly discuss the relevant factors that determine each party's dependence in terms of alternatives and commitment.

Alternatives/Exit Options

Alternatives can be regarded as the extent and the attractiveness of exit options for both management and workers. Based on the insights gained from bargaining power theory, labor's bargaining power decreases with the attractiveness of management's exit options and vice versa.

MANAGEMENT

The company's exit options are twofold. First, management may shift actual or future production and employment to alternative locations with lower labor costs and employment regulations. Second, the company may try to terminate or circumvent the existing exchange relationship by getting rid of the union and switching from collective to individual bargaining. Concerning the first alternative, globalization and especially regional economic integration has clearly increased the number and attractiveness of alternative production locations abroad. The possibility of international plant relocation and outsourcing has caused a shift in bargaining power in management's favor in both Germany and the United States. But while foreign direct investment comes with uncertainty as companies have to cope with the problems of potential political instability and cultural differences in the host country, U.S. companies, unlike their German counterparts, have the possibility to lower labor costs not only by shifting production internationally but also domestically to non-union locations. Given the by now large non-union sector and the significant union wage premium in the United States, domestic regime competition is a viable option and an important source of employer bargaining power in the United States.

This coincides with comparatively weak union security in the United States that makes unions vulnerable to employer opposition. There is by now a large literature that documents employers' escape from collective bargaining, union avoidance, and union busting activities and how these strategies are facilitated by an unfavorable institutional environment (Bodah and Cutcher-Gershenfeld 1997; Klein and Wanger 1985; Kleiner 2002). In contrast, collective labor representation still has much more institutional support in Germany. Works councils are mandated by law and possess legitimation and power resources external to and independent from the specific employer. Furthermore, escaping from collective bargaining is still much more difficult for German compared to U.S. employers. Abandoning collective bargaining coverage (for example, by exiting the employers' association that negotiates the union contract for the whole sector) is unattractive especially for medium-sized and large companies. Given the still considerable organizational strength of German trade unions, large companies that opt out of industry-wide bargaining coverage might well become the object of a union organizing drive that forces them to accept a company-specific union contract that provides for wage and employment conditions comparable to the respective sectoral contract (Thelen 2000).

WORKERS

The exit options for workers are also twofold. Either they immediately change to an alternative job or they temporarily or permanently opt out of the labor force and collect non-wage income, that is, most importantly unemployment or (early) retirement benefits. The relative attractiveness of these exit options can be evaluated by comparing the expected future benefit streams from these alternative employment/income options relative to the present job.

In general, the attractiveness of alternative employment options varies pro-cyclically with few attractive exit options during economic recessions. In comparative perspective, the expected cost of job loss and especially of losing a union job is much higher for U.S. than for German workers. With collective bargaining coverage as low as 9 percent in U.S. private industry and structural employment decline in those sectors that are heavily unionized, the likelihood of finding a comparable job at a unionized company is very low, especially during economic slumps where most unionized companies have workers on layoff that have preferential hiring rights. Moreover, the expected wage loss that occurs when switching from a union to a non-union job is much higher in the United States compared to Germany. Recent econometric analyses for the United States conclude that the union wage premium stands at around 15 to 18 percent (Belman and Voos 2004; Brats-berg and Ragan 2002; Card 2001); that might be as high as 24 percent after correcting for match and misclassification biases (Hirsch 2003). In contrast, the union/non-union wage gap for Germany is estimated to be an insignificant 4 percent (Blanchflower and Bryson 2003). The overall cost of losing a union job in the United States is even higher when taking into account that fringe benefits such as health and pension benefits are much more generous in the union as compared to the non-union sector (Mishel, Bernstein, and Allegretto 2005). Benefit differentials are much less significant in Germany with its system of nationalized health care and pension insurance.

Still, this comparison may understate the real costs of losing a job at a large German company. In 2002 close to one half of all west German unionized companies but less than one in five companies in eastern Germany paid wages that on average were about 10 percent higher than that specified in the respective union wage contract (Kohaut and Schnabel 2003). The fact that the likelihood of effective compensation in excess of the union wage rises with firm size explains the puzzling observation that union strongholds such as DaimlerChrysler, Volkswagen, and Siemens have recently been among the forerunners in concessionary bargaining. Those companies could enforce hitherto unprecedented concessions as the prospect of losing company-specific bonuses and a lack of attractive exit options for the affected workers provided management with considerable bargaining power.

When comparing workers' alternative non-wage income sources after being laid off, German workers on average fare considerably better than U.S. union workers. Both the duration of unemployment benefits as well as the replacement rate are much more favorable to workers in Germany as compared to the United States (Werner and Winkler 2004). Furthermore, German social policy up until recently provided strong financial incentives to resort to early retirement as an effective instrument to reduce employment by way of "cooperative downsizing" (Streeck 2001, 4). Externalizing adjustment costs to the social security system helped stabilize existing wage and employment standards so that companies could reduce labor costs while at the same time uphold cooperative labor relations and offer job security for the remaining workforce. Recent reforms that curtailed the options for early retirement and made this exit option less attractive may be responsible for the fact that concessionary bargaining in Germany has since experienced an upswing while the terms of recent concessionary contracts have become markedly less favorable for workers (Massa-Wirth and Seifert 2005). To be sure, exit options into nonemployment for U.S. unionized workers are much more attractive than for non-unionized workers. Contractual provisions such as SUBs and extended benefit coverage while on temporary layoff or company-specific early retirement programs and severance pay packages provide for income security for workers on layoff (BNA 1995). But while those programs are on average still less generous than public social security benefits in Germany, they are also subject to renegotiation in U.S. concession bargaining and do not represent an external fallback position on which workers can rely. As these measures are generally funded by the employer, they are not an effective means to shift part of the burden onto the public and lower the adjustment costs to be shared among the parties to a concessionary agreement.

Commitment

The concept of commitment as the second major dimension of bargaining power represents the degree to which the actors have a stake in and are locked into the existing exchange relationship. All else equal, labor's bargaining power increases with the degree of management's commitment in the relationship.

MANAGEMENT

The degree of management's commitment is determined by both direct as well as indirect costs (that is, the transaction and opportunity costs) of terminating the relationship. Prominent among direct costs are redundancy costs that occur when the company announces layoffs. Due to stricter job security legislation and legal provisions that provide works councils with significant codetermination rights in case of mass layoffs, employment adjustment costs are considerably higher in Germany as compared to the United States (Jahn 2004; Kittner and Kohler 2002). Even though the "employment-at-will" doctrine does not fully apply to the U.S. union sector where job security provisions curtail management's rights in layoff decisions, these contract clauses still do not restrict the company's ability to reduce employment levels as such. Rather than raising the transaction costs of employment separation as in Germany (where employers are mandated to negotiate "social plans" to attenuate the negative consequences of job loss), job security provision in U.S. labor contracts typically provide for seniority rules in layoff decisions in order to prevent arbitrariness. Strict job security plans that contain outright no-layoff clauses or even set guaranteed employment levels such as in the U.S. automobile industry are the rare exception rather than the general rule (Block 2001). Furthermore, while for German management job security legislation and the associated costs are a fixed given, contractual job security provisions are itself subject to renegotiation and become a bargaining chip in concessionary negotiations. In fact, prominent among the non-wage issues in U.S. concession bargains have traditionally been the relaxation of strict work rules and job security clauses.

Besides direct transaction costs associated with employment cutbacks, management also has to consider indirect or opportunity costs that arise from workforce reductions. Laying off experienced workers with firm-specific skills causes the loss of firm-specific human capital that the company needs to stay competitive in the future. Management's commitment to uphold the existing exchange relationship (that is, to provide for job guarantees and to refrain from redundancies) should therefore be greater the more the company relies on the continued utilization of workers' competencies in production and product innovation. While broad generalizations in a cross-national comparison are not feasible as the value of human capital depends on the company-specific business strategy, conventional wisdom in the literature on "social systems of production" and "varieties of capitalism" maintains that German companies predominately follow "high road" business strategies based on "diversified quality production" whereas institutional constraints and opportunities in the United States favor "low road" strategies based on "cost competition" (Turner, Wever, and Fichter 2001). While based on opportunity cost considerations German companies should tend to have more commitment to uphold existing employment relationships, a valid cross-national analysis would clearly have to be conducted on the industry or even company level as variation within countries is probably considerable larger than between countries (Katz and Darbishire 2000). The increasing focus of German companies on cost considerations (not least because of the adoption of shareholder value philosophies) and the introduction of cost competition in quality markets may have contributed to the recent increase in concessionary bargaining in Germany.

WORKERS

As with firms, workers' commitment to their present job depends on the direct as well as indirect costs of losing the job and leaving the current employer. The loss of non-vested fringe benefits can be regarded as the most important direct cost (O. Mitchell 1983). Company-specific defined benefit pension plans that are not transferable are much more important in the United States than in Germany and lock U.S. workers in the union sector into the existing employment match. Besides the loss of pension accruals, workers in both Germany and the United States may lose additional benefits that are linked to the current job such as accumulated seniority rights or future payouts of profit-sharing plans that were often negotiated as a quid pro quo in U.S. concession bargains. While most of these potential sources of "job stickiness" apply to workers in both Germany as well as the United States, they generally are considerably more relevant for unionized workers in the United States. For example, the fact that the pilots have generally been the first union to negotiate wage cuts in the recent wave of concessionary bargaining in the airline industry is in large part due to the large pension promises that would be lost in case of a carrier bankruptcy. While the Pension Benefit Guarantee Corporation (PBGC) takes over responsibility for the pension funds of liquidated carriers, it only secures pension payments up to an annual maximum of about $45,000, which is considerably less than what pilots, as the highest-paid labor group at a carrier, can normally expect when they retire.

Besides these direct costs, firm-specific human capital can be regarded as a match-specific investment that serves as an additional source of employee commitment to the current job. Losing the job means also losing the quasi-rents that come along with utilizing idiosyncratic skills and competencies in a specific work environment. While this holds true for German as well as U.S. workers, the relative share of firm specific compared to general skills is said to be larger for U.S. compared to German workers (Hollingsworth 1997a). While in the U.S. system on-the-job and task-specific training prevails, the unique dual apprenticeship system in Germany provides workers with transferable general skills (Franz and Soskice 1995). German workers with vocational training are therefore not as much committed to a specific employer and tend to possess a higher level of external mobility, just like skilled trades in the U.S. craft sector. Again, this preliminary cross-national comparison is unable to provide clear-cut explanations with regard to the level of workers' commitment; one would have to look closely at specific worker subgroups and their relative share of firm-specific versus general human capital. Again, the recent concessionary bargaining in the U.S. airline industry shows that the maintenance workers (mostly skilled trades) have generally been the employee group at each airline that has most vigorously fought against concessions. While pilots and flight attendants, because of their industry-specific skills, are locked into their current job, skilled machinists are much less committed to employment in the airline industry as they have considerably higher cross-sectoral mobility because of transferable skills.

Conclusion

This paper argues for an extension of both simple institutionalist as well as interest-based approaches in comparative IR by incorporating bargaining power considerations that shift the analytical focus onto power asymmetries in labor-management relations. While the "new institutionalism" framework tends to be overly simplistic as it infers IR outcomes from institutional settings and downplays the degree of variance as well as change in IR systems, purely "interest-based" approaches fail to provide satisfactory answers as to which of thecompeting interests among the actors to an IR system in general and to collective bargaining in particular should be expected to prevail.

The comparative analysis of concession bargaining is used as a test case to demonstrate the weaknesses of both purely institutionalist and interest-based approaches as well as the strength of an extended concept that incorporates power considerations. Although a more in-depth empirical analysis would be necessary to strengthen the argument, this preliminary discussion of concessionary bargaining in Germany and the United States shows that a power-based analysis is able to avoid the pitfalls of both institutionalist as well as interest-based explanations. The overall differences in bargaining outcomes (that is, more drastic employee concessions and less employer givebacks in the United States as compared to Germany) are explained by a greater degree of power asymmetry in favor of management in the United States as compared to Germany. Besides explainingthese cross-national differences, a power-based approach is also capable of explaining a substantial part of subnational variance, which is traced back to differences in the relative degrees of alternatives and commitment and ultimately dependence in labor-management relations.