Worker perceptions of representation and rights in Germany and the USA

Germany and the USA have very different systems of legal representation and rights at work, but these differences and their effects may have lessened. We draw on a large-scale telephone survey to explore worker perceptions of these systems, and find that perceptions of German workers are more favourable than those of their US counterparts, but not by as much as might be expected. Our findings could in part be explained by cross-national differences in both worker ideologies and the way the different systems function, but they also point to the importance of perceptions in understanding and assessing cross-national institutional differences, and have implications for the future of workplace representation and rights in both nations.


Introduction
German workers enjoy both substantial individual legal protections (ILO, 2008;OECD, 2011) and strong collective representation rights (Behrens, 2013). Germany is a coordinated market economy (CME), and trade unions tend to serve a positive function for employers by negotiating industry-wide agreements (Hall and Soskice, 2001). Employers therefore typically have reason to accept rather than resist them. Trade union representation is complemented by works councils, which by law enjoy strong participation and consultation rights, enabling them to represent workers effectively and to play a 'partnership' role (Höpner and Waclawczyk, 2012). In the USA, a liberal market economy (LME), workers have limited individual rights and protections and collective representation is weak. This is compounded by ineffective labour laws, which make it difficult for workers seeking union representation to overcome employer resistance (Gross, 2012). Where unions are established, they are relegated by law to a largely adversarial role and lack meaningful participation or consultation rights, lessening their ability to represent their members and reinforcing employer opposition (Friedman, 2013).
How far does actual practice, as viewed by workers themselves, reflect these institutional differences? We might expect workers in Germany to view both individual rights and collective representation systems as far more effective than in the USA. Yet some writers have questioned how far representative institutions in Germany remain strong and effective, particularly in growing segments of the economy that do not conform to the traditional model (Addison et al., 2007b;Doellgast and Greer, 2007;Streeck, 2009). There has also been some weakening of individual rights and protections. Conversely, some US scholars have identified a trend to stronger rights for workers and more judicious employer practices, complemented by a growth in alternative, non-union forms of representation (Estlund, 2010;Piore and Safford, 2006). Thus, institutional contrasts in the two countries may not accurately indicate how far worker perceptions differ. This may be compounded by differing institutional norms in these two countries, which may cause Germans to expect more of their institutions than do their US counterparts.
We explore this question by drawing on data from a random household telephone survey of 1000 employees in each country that we commissioned in November 2009. Participants were over 17 years of age, worked more than 15 hours per week and had been with their current employer for more than six months. The surveys were conducted by professional polling firms in each country. Participation rates are difficult to establish: most non-respondents hung up before their eligibility could be determined. However, excluding such cases, our best estimates are 59 percent for the response for the German sample and 51 percent for the American.
Although worker perceptions of their institutions and how they function can suffer from problems of imperfect and even systematically distorted information, they are useful in at least two respects. First, workers' perceptions can have important implications for the functioning, legitimacy and even long-term future of these institutions (Jackson, 2005). Second, such perceptions can complement more objective analysis. In particular, they can provide information and insights into how institutions appear to function 'on the ground'.
In this article, we hope not only to contribute to knowledge of how German and US institutions appear to function from the point of view of workers, but also to shed light on broader issues of relevance to comparative analysis. The potential of worker surveys for addressing such issues is often overlooked in the comparative literature, and even where such data have been used, authors have typically had to rely on incompatible datasets, limiting the basis for meaningful comparison. Such datasets have also typically been restricted to LMEs. In contrast, our dataset is compatible across our two samples and allows us to explore differences in worker perceptions between archetypical LMEs and CMEs. The comparison of these perceptions in two very different economies requires very cautious analysis: not only may respondents have imperfect information at the institutional level, so that their perceptions may not accurately reflect the actual functioning of institutions, but differences in the predominant institutional traditions and norms of these countries mean that what is generally considered to be effective or desirable will also differ (Frege and Godard, 2010). Nevertheless, our study enables us to shed light on how far, and how, the very substantial institutional differences in the two countries matter in practice. Thus we also contribute to the debate on the 'convergence thesis' (Katz and Darbishire, 2000) in comparative industrial relations.

Findings
In presenting the results we address six issues: the prevalence of alternative forms of representation in each country; the perceived effectiveness of these institutions; representation gaps; perceived levels of management opposition; respondents' evaluations of the relationships between their representatives and employer; and their confidence in their rights at work.

Forms of representation
Union density and bargaining coverage have been in almost continuous decline in the USA since the 1960s and in Germany since the early 1990s. In 2009, union density was 18.8 percent and bargaining coverage 62.0 percent in Germany; in the USA the figures were 11.8 and 13.6 percent (Visser, 2011).
With only a small portion of the US labour force covered by collective bargaining, there has been a growing literature focusing on alternative forms of representation (Kaufman and Taras, 2000). One involves management established non-union representation systems, which may serve as substitutes for unions (Kaufman, 2000). Another involves representation by associations, especially those based on 'identity' groups, which typically aim to establish stronger rights and protections for their members at law, but can also serve as supports and advocates at the workplace level (Piore and Safford, 2006).
In Germany, despite union density decline in the last two decades, collective bargaining coverage remains extensive. In addition, almost half those in private sector jobs are covered by a works council (Elguth and Kohaut, 2010). There is considerable overlap between works council and union coverage, but roughly three quarters of all workers in the German private sector are covered by at least one of these two forms of representation (Addison et al., 2002;Gerlach and Meyer, 2007). Accordingly, management established, non-union systems of representation and identity group representation play no significant role. However, occupationally based associations may be widespread, and may have some representational role.
Our survey asked respondents whether they were union members; covered by a collective agreement; and a member of another association that assisted with work-related matters. German respondents were asked if they were covered by a works council (or Personalrat in the public sector), and US respondents who were not covered by a collective agreement were asked whether a management-established non-union system existed.
The American results for these questions have been discussed elsewhere (Godard and Frege, 2013); here, we compare them to those for the German questions. They are presented in Table 1.
In the German sample, almost a quarter of respondents report that they are union members, and two-thirds that they are covered by a collective agreement. Two-thirds also report works council representation. This is higher than estimates of 45 percent in the large-scale IAB-Betriebspanel survey, and reflects the fact that we excluded temporary workers and those working under 15 hours a week. In addition, unlike the IAB we include public employees; when these are excluded from our sample, 60 percent report coverage. Only one in ten report that they are represented by another association.
While fewer US respondents report union membership, half report some form of representation. A third of the non-union sample report that there is a 'non-union, management established system, where worker representatives meet with management'. One in seven report that they are 'a member of another type of association to assist you with work-related matters'. Of those with neither bargaining coverage nor a managementestablished system, 11 percent (6% of the total sample) report membership in an association.
Both US and German respondents indicating that they were members of an association were asked about its main characteristics. In Germany two-thirds of these, and in the USA three-quarters, report that the association is occupationally based. Only 7 percent in each case report membership of an identity-group based association. Overall, 84 percent of German respondents report at least one form of representation: more than one and a half times that for the US sample. Moreover, eight in ten have legal representation, which is almost five times that for the US sample.
These results are largely consistent with available statistics and research findings and do not warrant extensive discussion here. The main possible exceptions are for associations, which are based largely on occupations and not as widespread as some American authors have conjectured, and for management-established systems in the USA, which appear to be widespread and so may fill much of the gap left by union decline (Godard and Frege, 2013). This latter finding means that the differences between USA and Germany may not be as great as suggested when one looks only at legal representation. Yet any such conclusion would be premature, because much depends on the actual effectiveness of these forms of representation.

Evaluations of alternative forms of representation
A conventional assumption is that extra-legal representation mechanisms are inferior to legal forms, if only because they do not enjoy the support of, and enforcement by, the state. Yet it would be mistaken to regard employer-established systems as complete shams; a number of studies document their potential benefits for workers (Estlund, 2010;Kaufman, 2000Kaufman, , 2005. Conversely, legal forms of representation may not always be highly effective in representing member interests and concerns. Moreover, workers in different countries can have substantially different expectations and 'tastes' for representation; thus differences in the strength of representation in Germany and the USA may not be fully reflected in worker perceptions of the effectiveness with which they are represented. To address this possibility, respondents who indicated that they had each form of representation system were also asked about the extent to which representatives in that system 'can be counted on to stand up for members, even if this means a disagreement with management' and the extent to which they 'actively consult workers about their ideas and concerns'. Respondents were offered a three-point scale, where 1 = to a great extent, 2 = some extent, 3 = very little if at all. We have discussed the US results elsewhere (Godard and Frege, 2013), with reference to a different research question. Here, we compare them with the German responses, as presented in Table 2.
US respondents tend to evaluate the quality of both union and management-established systems relatively favourably on these two dimensions, especially in comparison to the evaluations of their German counterparts. American members of associations tend to evaluate these somewhat less favourably. Moreover, while their German counterparts rate such associations about the same when it comes to standing up for workers, they rate them much more favourably when it comes to consultation. This probably reflects a stronger associational structure for professionals, as one would expect in a CME.
Overall, the results suggest that American workers with representation are happier with this representation than are their German counterparts. This may in considerable measure reflect differences in the nature of representative systems and in worker knowledge and expectations as to their functioning. In Germany, works councils have been designed as representative delegation systems which function independently from workers' activism and are embedded in a cooperative system with management, playing by clearly defined rules that encourage cooperative solutions. Trade unions on the other hand are in most cases outside the workplace and not very visible in their routine activities. Thus, both unions and works councils tend to be bureaucratic and removed from workers on a day-to-day basis. Moreover, because they are more institutionalized within the economy as a whole and able to act proactively, levels of conflict may be reduced, thus rendering their achievements less visible. German workers may thus take their union and works councils representatives for granted, much as citizens often take their civil rights for granted when these rights are not in question. The German responses may also reflect a more critical culture, one that encourages the expression of discontent, and an institutional environment that is conducive to doing so without fear of reprisal.
In contrast, American unions have in recent decades faced increasingly insecure and conflictual circumstances: innate employer opposition, weak labour laws and economic liberalization. As such, they more often have to both stick up for and consult their members (and be seen to do so). Representatives in management-established systems face similar pressures, albeit for different reasons. These systems can always be criticized as shams, designed to prevent workers from seeking union representation; or where this is not the case, they may whet the appetite of workers for the 'real thing' (Taras and Copping, 1998). So, as for their union counterparts, it is important that representatives at least appear to stand up for workers and consult them. Indeed, because these systems tend ultimately to be set up to serve management objectives, much of what they do may be to manipulate the information available to workers and the way in which it is perceived, strengthening the likelihood of positive evaluations. American workers are also much more likely to accept employer authority than are their German counterparts and are highly receptive to management established systems (Freeman and Rogers, 1999), consistent with institutional norms and traditions in support of strong property rights and minimal state (and legal) involvement in the economy (Ely, 2008;Godard, 2009).
It follows that our findings may largely reflect differences in both the functioning of representation systems and the expectations and values of workers in each country. In addition, German workers have much higher levels of representation to begin with. If one multiplies the coverage results in Table 1 by the evaluations in Table 2, it becomes clear that a higher percentage of the total German sample reports that they have representation and that it is of high quality. Using this method reveals that 9 percent of the total US sample report high quality union representation with regard 'standing up' for workers, while 7 percent report this with respect to consultation. The comparable statistics for Germany are 30 and 20 percent. Similarly, 14 percent of the total US sample report high quality employer established representation with regard to 'standing up' for workers, while 15 percent report this with respect to consultation. The comparable statistics for German works councils are 18 and 25 percent. (The low levels of alternative association membership mean that there is little difference with regard to this form of representation.) So a simple comparison of the results in Table 2 may be misleading. Nonetheless, these suggest that, for those with representation, American workers generally perceive this more favourably than do their German counterparts and that the latter may even be somewhat disillusioned with their institutions. Moreover, adjusting for representation levels may be misleading if it includes those who not only lack a particular type of representation but also do not want it (see below). For these respondents, an absence of high quality representation does not matter. When we further adjusted our results accordingly, we found little difference across the two samples (available on request).

Frustrated demand for representation
If American workers are uninterested in obtaining increased representation, then it might be argued that the relatively low level of representation in that country is not a problem and, indeed, that higher levels would even be contrary to their wishes. In contrast, if a significant percentage of German workers express frustrated demand for representation, then there may be reason to think that their de facto rights are not as strong as they appear to be at law. Table 3 reports the results for questions about frustrated demand for the two main forms of representation in each country. (We did not ask about alternative associations.) It reveals that only one in five non-union workers in the US sample would vote for a union in a certification ballot, although another one in ten are unsure. This result differed very little between non-union workers with and without an employer-established system. In turn, two in five workers without a union or a management-established system report that they would, at least to some extent, prefer the latter system to union representation or no representation at all. Thus, although these figures are not strictly comparable, frustrated demand would appear to be higher for employer-established representation than for union representation. In total, three in five workers without either a union or an employer-established system, representing 30 percent of the total sample, report that they would like to have one or both. These findings are generally consistent with those of surveys in 1994 and 2001 (Freeman, 2007: 9;Freeman and Rogers, 1999: 140-148), but they differ with respect to the frustrated demand for unions. Most polls have estimated this to amount to roughly a third of non-union American workers, which is higher than our findings. However, this difference may reflect the scapegoating of unions during the economic crisis of 2008 and the overall decline in approval for unions since then (Jones, 2011). The results for Germany suggest at first sight that there is more unfilled demand for both works councils and collective bargaining coverage than there is for unions in the USA. A third of respondents without a collective agreement would like to be covered by one, while three in ten without a works council would like one. However, the high coverage of both of these forms of representation means that, of the total sample, only one in ten workers expresses frustrated demand for either. Among the 20 percent of workers in the sample with neither form of representation, close to half would like one or both forms. Thus, only one in ten German workers would like one or both forms of representation but have neither. This is far lower than for the US sample and suggests that frustrated demand is far greater in the USA.

Explaining the gaps: Are employers the problem?
It is generally assumed that German employers do not explicitly oppose either works council or union representation, not only because of national traditions of cooperation and a general pro-union public sentiment, but also because of strong laws and because these institutions can and do serve positive functions for employers. In the USA, by contrast, low levels of union coverage and the representation gap between those who would like to have a union and those that actually have one are in the final instance typically attributed to the opposition of employers and their ability to take advantage of weak labour laws to undermine an organizing drive should one occur (Godard, 2004;Weiler, 1990). There is also substantial objective evidence of employer anti-unionism (Human Rights Watch, 2000).
Yet once again, these depictions may be oversimplified. Strains on the German system in recent years may have meant that employer opposition has become a factor, with evidence of increasing anti-unionism and at times illegal practices, contributing to the growing proportion of workplaces that are not covered by a collective agreement and/or lack a works council. Conversely, there may be many American employers who would be relatively tolerant of a union organizing drive but never face one because of positive HRM practices. Table 4 reports respondent perceptions of the extent to which the absence of a particular form of representation could be attributed to their employer. German respondents without collective bargaining coverage or a works council were asked if employer opposition was the main reason. American respondents were asked the extent to which employer opposition explained their lack of union representation. We class US responses 'to a great' or 'some' extent as 'yes', responses 'to very little extent' or 'not at all' as 'no'.
In the German sample, one in four workers without collective bargaining coverage attribute this to employer opposition. This is higher than in the USA, although it should again be understood in the context of generally much higher coverage in Germany. As such, it accounts for only one in 12 of all German respondents, as compared to one in five of all American respondents. Yet the finding that three in four German respondents without coverage do not attribute this to their employer is surprisingly high, because employers generally decide whether to adhere to a sectoral agreement (or to negotiate a company agreement). It is possible that many of these respondents were in workplaces where there was no industry agreement. Some respondents may also have opted out of coverage, even though this would be of questionable legality. Others could be managerial or professional workers, who may fall outside of the agreement covering others in their workplaces. Our survey included questions directly asking about the first two of these possibilities, and also included variables allowing us to explore the third possibility. With respect to the former, 40 percent responded that a lack of an industry agreement was the primary reason they were not covered, while 8 percent responded that they had simply opted out. Preliminary analysis (regression and zero order correlations) suggested little support for the third possibility. With regard to works councils, hardly any workers without works council coverage attribute this to employer opposition. This is consistent with the widespread perception that German employers generally embrace works councils. It may be that many German workers simply do not view works councils as effective and choose not to exercise their right to establish one. This would be somewhat at odds with our finding that three in ten workers without a works council would like one. One possibility is that this reflects general values that support works councils in principle, yet a corresponding belief that it is not worthwhile to spend the time or resources necessary to act on these values. Such a conclusion would be consistent with the rather tepid evaluations works councils receive from those who already have them.
It is also possible, however, that works councils are perceived as less necessary or even viable in some workplaces, even though workers may prefer them as an ideal. This may be especially likely where there is already extensive informal and direct consultation, either because the workplace is small or because the employer has adopted high performance practices, which can include such consultation. It may also be that workers in professional jobs already have considerable autonomy and control. These possibilities were borne out in part by subsequent analysis of our data set. Although workplace size did not bear a statistically significant association with the desire for a works council, both high performance employer practices and managerial/professional work bore significant negative associations. It may even be that new HRM practices are displacing works councils -an important possibility but one that we cannot explore here.
The results suggest that, in the USA, employer opposition may be less important than commonly assumed: only one in five non-union respondents attribute the absence of union representation to such opposition. This may simply reflect the low percentage who would vote for a union: workers who do not perceive the need for a union are less likely to have negative perceptions of their employer and more likely to consider their nonunion status as the outcome of their own free choice. However, we observed a correlation of only 0.26 between propensity to vote for a union and perceived employer opposition, so this is only a partial explanation. We also observed a correlation of only 0.23 between a measure of 'good' HR practices and perceived employer opposition, suggesting that the HRM explanation may also be only partial. A further possibility is that, with union density so low, the thought of organizing a union never enters the minds of most workers, and so employer opposition is not considered. But whatever the case, this opposition is generally not seen by non-union workers as a major reason why they do not have a union. Employer opposition may thus not be as great a factor as the US literature on labour law and union organizing would lead us to expect.

Relations with employers
In Germany, as noted above, collective bargaining is generally considered to have largely positive functions for employers, effectively taking wages -and most other distributive issues -out of competition and serving a coordinating function across employers. Thus distributive conflict has been limited in Germany, reflected in a much lower level of strike activity than in LMEs (Hale, 2008: 33). Works councils are likewise perceived to have a largely integrative function: given the high skills of the workforce, coupled with the stronger quality orientation of employers, they add value to the employer (Streeck, 1992). Although research confirming this has been mixed in its results (Addison et al., 2004(Addison et al., , 2007a, there is little evidence of employer resistance to works councils, and our results are consistent with this. Yet as noted above, there have been a number of strains on the German system in recent years, as employers have increasingly sought flexibility at workplace level, and distributive issues have increasingly shifted into the domain of works councils. In the USA, collective bargaining, and the role of unions in general, are generally considered to be adversarial. As discussed earlier, this may in part reflect both employer ideologies and the lack of a positive economic function for unions in LMEs. But some also argue that it is actually worsened by a system of labour law that seems premised on the assumption of such relations (Adams, 2001). Moreover, the requirement that a majority of workers must support having a union, coupled with lengthy organizing 'campaigns' during which employers are able to engage in a variety of anti-union tactics, virtually ensures that this assumption is borne out in many cases. To make matters worse, certification only grants unions the procedural right to bargain on behalf of their members, with few substantive consultation or co-decision rights. As a result, unions are largely in a reactive position even when recognized. Table 5 reports workers' perceptions of the relations between their representative institutions and their employers. As with Table 4, the format of the questions differed, so comparisons should be made with caution. German workers were asked whether union and works council representatives had a good or a bad relationship with management and provided with the four response options: good, bad, neither, and not sure. American respondents were asked the extent to which union representatives had a positive relationship with management, and were provided with three response options which we have classed in the same manner as for Germany.
Although exact comparisons are difficult because of the response formats, only a third of German respondents with bargaining coverage rate the union-management relationship as good, fewer than in the USA. This may reflect the distance between workers and the level at which bargaining takes place, so that many workers are unfamiliar with the quality of the relationship and thus perceive it in neutral terms. Indeed, of German respondents covered by an agreement, six in ten answered 'neither' or 'not sure'.
However, when the 'not sure' responses are removed from the sample, the percentage of German respondents rating the relationship as good is still only 37 percent. Conversely, seven out of ten German respondents with a works council rated the relationship between management and their representatives as 'good', and none rated it as 'bad'. One in six answered that it was neither good nor bad, which could reflect recent strains indicated above. But even so, it would not appear that these strains are particularly widespread or serious enough to give rise to bad relations. This is consistent with our finding ( Table 4) that there is virtually no perceived employer opposition to works councils among respondents not covered by one, and with the belief that works councils are effective at helping to find cooperative solutions where conflict might otherwise occur and as such have generally positive relations with employers. However, in view of the rather tepid evaluations of works councils reported earlier, it may also be that many works councils have come, or are perceived to have become, largely co-opted to serve managerial interests -something that has long been considered a possible risk of the German system (Gumbrell-McCormick and Hyman, 2010).
On the other hand, the results suggest that the level of acrimony in the US system may be lower than the stereotypical view assumes. Only 15 percent of respondents perceived the relation negatively, while four in 10 considered it to be good and a similar proportion considered it to be neither good nor bad. In view of the structure of the US system, and the role of unions within it, we might expect a 'mature' relationship to be one that fits this latter, middle category. Yet in addition to the limited number of poor relationships, the finding that four in 10 report harmonious relations is something of a surprise. It would seem that the standard portrayal of union-management relations in the US applies to far fewer workplaces than is commonly assumed. It is, however, possible that union weakness, coupled with job insecurity, have reduced the adversarial role of unions in many workplaces. It is also possible that union-management relations have, partly as a result, been undergoing a transformation to a more cooperative approach. Subsequent analysis revealed some, but limited support for these conjectures: a correlation of 0.25 between employee perceptions of the union-management relation and a variable in our dataset addressing job insecurity, and 0.29 between this variable and an index of high performance practices, also in our dataset.

Confidence in rights at work
Although representation at work is an important element of democracy in and of itself, worker perceptions of this representation also matter. Perhaps even more important, however, is the extent to which workers are confident in their rights at work. This confidence may derive from the belief that they can depend on representatives to stand up for them should they be unfairly treated or their rights violated. But it can also derive from the broader legal system and the institutional environment of the employment relationship. In some respects, this may be the most important 'test' of differences in national employment systems. In Germany, workers have traditionally enjoyed strong protections through their works councils and unions, based on a longstanding conviction of the benefits of industrial democracy (Frege, 2007). Although these institutions may have declined in strength, workers also have strong rights under the law. This has been especially true with regard to job security, but it has also applied to the provision of due process and of various protections against unfair or discriminatory treatment. Even where workers do not have representation or choose not to rely on their representatives, they can turn to a labour court should their rights be violated. Moreover, although there has been some weakening in recent years of job security laws, these are still much stronger than in the United States. In view of the German institutional context, we would therefore expect workers to have a high degree of confidence in their rights.
In the USA, minimal restrictions on the exercise of employer authority have traditionally meant that workers have limited statutory rights at work. Union representation has served as the primary means to the attainment of fairness rights, yet unions now cover only a small portion of the workforce. However, there have long been laws protecting employees against selected unfair practices, especially with regard to the payment of wages (the Federal Labor Standards Act), but also with regard to discrimination on the grounds of race or gender. There have also been laws mandating family leave and advance notice of layoffs, some limits imposed on the doctrine of employment-at-will and a variety of rights and protections passed at the state and local levels (such as living wage laws). According to some scholars (Estlund, 2010;Piore and Safford, 1997), employers have in response generated standard personnel policies and practices and developed a new strategy centred on private arbitration procedures that ensure some form of due process. It is therefore possible that workers have not only stronger rights, but also greater confidence in these rights than traditional depictions of the US employment regime would suggest.
We asked six questions pertaining to confidence in rights. These concerned due process ('able to get a formal hearing with representation if unfairly treated'); job security ('protections make layoff of permanent employees unlikely'); and four questions asking if 'it is likely something could be done to make things right' if a co-worker was unjustly dismissed; a female co-worker was discriminated against; a co-worker was bullied by a manager; or a co-worker was denied pay or bonus money to which they were entitled. If a respondent asked for clarification, they were told: 'this could involve going to a government body, a union representative, a manager or some other channel'. We asked about co-workers because some of these questions may not apply to the actual respondents. The results are presented in Table 6.
Predictably, German workers display considerable confidence in respect of most of the issues. Yet a significant portion did express doubt. Most striking is that one in three respondents were either uncertain or disagreed that something could be done should a co-worker be unjustly dismissed or denied pay or bonus money. Given the German system, respondents may have set a high benchmark for effective protection of rights. It is also possible that many of the less confident workers are not covered by a union or a works council and so may perceive little ability to enforce their formal rights. To explore whether this might be true, we examined the correlations between each of the rights perception items and each form of representation (details are available on request). The correlations were strongest for works councils, ranging from 0.20 to 0.25. However, the correlations with collective agreement coverage were also statistically significant (p ≤ .05), ranging from 0.12 to 0.16.
American workers also express relatively high levels of confidence in their rights, although these levels are generally lower than for their German counterparts. A third to a half agree strongly with statements about each of the six issues covered, and another quarter agree somewhat. As might be expected in view of the US legal system, they are most confident about co-workers being able to do something if they are denied pay or bonus money, and least confident about job security and unjust dismissal. Nonetheless, even in the latter case, a majority expresses confidence. Although there has been some evidence that American workers are often unaware of just how poor their rights are (Freeman and Rogers, 1999), these results provide some support for the argument that the employment relation may not be as authoritarian and arbitrary as the traditional stereotype suggests, at least for a significant portion of the US workforce.

Conclusions
This article has explored how workplace representative systems and rights are viewed by workers in two countries with very different institutional traditions: Germany and the United States. The findings provide insights into how well each system fares from the point of view of its main constituents and hence into possible issues and problems faced by each. Because these systems differ in a number of respects, and because worker values and expectations also differ, we cannot draw definitive conclusions as to how well each system currently functions in any objective sense, and we certainly cannot strictly compare them in this respect. However, the way in which workers view their institutions may have important implications for their legitimacy, and cross-country comparisons can provide referents for assessing worker perceptions in a particular nation. This is especially so if these comparisons include discussion of how both the institutional conditions and expectations of workers differ and hence may explain differences in perceptions, which we have been careful to do. Our results reveal that German workers still appear to fare better overall. There is some frustrated demand for both collective bargaining coverage and works councils, but the levels for both are considerably higher than for their US counterparts. Moreover, although there is some perceived employer opposition to collective bargaining, employers are still seen as highly receptive to works councils. In general, German workers are also more confident about their rights at work. Germans are often ambivalent, however, about the performance of their institutions, and indeed do not (on average) evaluate them as favourably as Americans do theirs. Indeed, there is little difference across the two samples for each type of representation we examined, if one excludes those who do not have that type and also do not want it. Although union coverage is low in the USA, half of all workers have some form of representation. They also tend to evaluate this representation as favourably or more so than do their German counterparts. Our results also reveal that the level of frustrated demand (especially for unions) is relatively low, that employers are seen by only a small minority of non-union workers as an obstacle to union representation, and that relations between unions and their employers are not as acrimonious as typically portrayed. Furthermore, workers have substantial confidence in a number of their rights at work.
These results might be interpreted as providing some support for the convergence thesis: either because worker representation and rights at work are somewhat stronger in the USA than might be expected given national traditions and law, or because they are somewhat weaker than expected given German traditions and law, or indeed both. This is consistent with arguments concerning recent developments in both countries, discussed earlier, and confirms the value of looking beyond formal institutions and law to establish how institutions are actually functioning 'on the ground'.
Ultimately, however, our results may simply show that the USA and Germany are two different regimes, each with its own particular institutional traditions and the norms and expectations to which these have given rise. If one adheres to German norms and expectations, then there is no doubt that the German system is to be preferred, if only because workers are much more likely to have legal representation and their perceived rights are stronger. These are, from a German perspective, an essential measure of democracy, and expectations for them tend to be high. Yet because they are so institutionally embedded, their performance is not readily visible. But if one adheres to American norms and expectations, there is much more room for doubt, because legal representation and rights at work would seem to be much less important as a measure of democracy (Frege, 2007;Frege and Godard, 2010), and expectations tend to be low. In this respect, our results may in considerable measure reflect the apparent tendency for Americans to be more deferential to authority and to management than their German counterparts, at least when it comes to workplace relations. This would explain not only why the American system is in a sense less 'democratic' than its German counterpart, but also why this makes less difference than might otherwise be expected.
Our findings may therefore be important less for any light they may shed on the convergence thesis than for illustrating the importance of institutional and cultural norms in understanding cross-national differences in industrial relations systems and how they matter to actors within them. Indeed, in subsequent work with our dataset we establish this to be the case. Yet even more important may be the implications for anyone who believes in the importance of legal rights and representation at work to the democratic fabric of nations. In the case of Germany, our findings suggest that the stronger system of legal representation and rights may be in some jeopardy, if only because workers seem to be either uninformed or disillusioned with its performance. This implies that the detached representative system characteristic of Germany may not be sufficient to cope with the challenges of the 21st century and that, if these institutions are to maintain their legitimacy, there is need to move towards a more transparent, directly accountable model with more grassroots involvement. In the case of the USA, our findings suggest that unions and collective bargaining have been largely replaced by systems created if not dominated by employers, and that these may constitute a form of managerial hegemony that offends basic precepts of democracy yet appears to be largely unchallenged, in large part because of institutional and cultural norms (Godard, 2009). This implies a need for political and academic activism and, in particular, a need to expose the dangers posed by company unionism as part of a potentially broader decline in the quality of American democracy.
In short, our results suggest that now is not the time for complacency in either nation.

Funding
The research in this paper was funded by a Social Sciences and Humanities Research Council of Canada standard research grant.