The Paradox of Decorporatization

The Paradox of Decorporatization Why Less Corporatism Increases Inequality in Interest Group Access - Not to be quoted Peter Munk Christiansen Department of Political Science and Government Aarhus University Bartholins Allé 7 DK-8000 Aarhus C pmc@ps.au.dk Paper for presentation at Dansk Selskab for Statskundskabs årsmøde 23.-24. oktober 2014 på Hotel Vejlefjord 1

Abstract: According to a standard definition, corporatism is a system of interest intermediation in which privileged interest groups are integrated into the government decision making machinery. Offhand, this would imply that decorporatization as witnessed in many formerly corporatist countries during the last few decades increases access for more non-privileged groups. In other words, if corporatism is privileging, decorporatization must imply deprivileging. However, this appears not to be the case in Denmark in the period 1975-2010. While access to decision makers through the remaining corporatist committee system has become less unequal, access through other forms of contacts has become more unequal. One interpretation is that that the regulation of access that is implied in a corporatist structure also supports some of the less resourceful interests while a free market for access leaves more room for large and resourceful interest groups and less for the small and poor. The analysis is based on large scale surveys to interest groups in 1976 and 2010 and on data on all government committees with interest group representation in the same two years. Introduction In many countries with a corporatist heritage, corporatist structures have been weakened (Crepaz 1994; Blom-Hansen 2000; Häusermann, Mach and Papadopoulos 2004; Christiansen et al. 2010; Anthonsen, Lindvall and Schmidt-Hansen 2011); however there is no agreement on the degree of change, and in some cases not even about the direction of change (Molina and Rhodes 2002). If we understand corporatism as a particular system of interest group representation in public policy making, an important question is what replaces corporatism when it withers away. An obvious answer is more lobbyism (Lewin 1994; Christiansen and Rommetvedt 1999; Rommetvedt et al. 2013) or at least some other form of interest group involvement in policy making (Richardson 2000). But how will decorporatization affect the chances that different types of interest groups are heard? This is a key issue for former corporatist partners as well as for groups that were not recognized as close partners in earlier corporatist exchange relations. Will former partners find it increasingly difficult to gain access to decision makers and reversely will formerly peripheral groups find it easier? Or are other forms of dynamics involved in interest group access? Put in another way: Does decorporatization decrease or increase inequality in interest group access to political decision makers? These questions touch upon a classic theme with far-reaching implications. Interest groups play an important role in a democracy because they tie together the political and administrative decision making machinery on the one hand and representatives of the surrounding society on the other, but they do so on very different conditions in terms of resources (Baumgartner and Leech, 1998). Interest groups play an important role for deliberation on issues of potential conflicting interests on 2

the condition that they can be heard (Mansbridge 1992; Öberg 2002). Interest groups also play an important role for the protection of minorities, because they can voice intensive interests that may be outnumbered if decisions are made on the basis of majority voting (Schmitter and Karl 1991; Lewin 1994). Corporatism is far from a perfect system for diversity in the voicing of interests because a representation bias is genetic to corporatism, but so is pluralism because of the strong upper-class accent" of the chorus in the pluralist heaven (Schattschneider 1960: 35; cf. also Schlozman, Verba and Brady 2012). Corporatism and pluralism both represent biased systems even if it is notoriously difficult to define a non-biased system (Baumgartner and Leech 1998: 52, 111-12, 117; Schlozman and Tierney 1976). Corporatism is biased because it privileges some groups at the expense of others (Christiansen et al. 2010). Pluralism although based on a notion of perfect competition among groups (Kariel 1961: 1-4; Baumgartner and Leech 1998: 52) and therefore in principle a perfect system of interest group representation is biased because interests do not come easy into being (Olson 1965) and because resources mean so much for access (Schattschneider 1960; Beyers and Kerremann 2007). Both systems produce bias, but which one corporatism or pluralism produces most bias, i.e. creates the highest level of access inequality among interest groups? This is the question I try to answer in this paper. In the following section, I argue that from a theoretical point of view we will counterintuitively expect pluralism to produce more inequality than corporatism. However, the proof of the pudding is in the eating! Consequently we must test our hypothesis against the empirical evidence. It is difficult to compare corporatist and pluralist countries, because different levels of inequality of interest group access may be produced by other factors than the interest mediation system itself. I therefore pursue another strategy, namely a longitudinal (double) comparison of Denmark in 1976 and 2010. This is probably the best possible design in terms of answering the research question. I first compare the level of inequality in access to public committees in the two years. These seats are distributed on the basis of corporatist norms and their distribution is a good proxy for inequality in corporatist structures. I then compare the level of inequality in the intensity of contacts with decision makers. Such contacts are not or at least less guided by corporatist norms than committee representation. The distribution of contacts on interest groups is used as a proxy for inequality related to a decorporatized or pluralist system. A comparison of the level of inequality in the distribution of seats and the distribution of contacts allows us to say something about the effect of decorporatization. The empirical analyses provide support for our basic expectation. For all groups corporatist participation as measured by committee representation has become less unequal, while the distribution of contacts to decision makers has become more unequal. Even if a few group types do not conform to this pattern, I conclude that while corporatism offers very unequal access 3

for interest groups, a decorporatized system offers even more unequal access for interest groups to decision makers. Decorporatization and equality of access. The theoretical argument By corporatism we understand a specific way in which interest groups are involved in decision making on public policies. More precisely we define corporatism as institutionalized and privileged integration of organized interests in the preparation and/or implementation of public policies... (Christiansen et al. 2010: 24). That integration is privileged implies that some interest groups have a more prominent position than others. That it is institutionalized implies that privileged groups tend to keep their privileged positions over long periods of time. In a corporatist system there is thus a limited number of core interest groups with a persistent and intense presence in each policy area, and a large number of non-privileged groups with significantly decreasing persistent access to decision makers (Fraussen and Beyers 2014; Binderkrantz, Christiansen and Pedersen 2014). Privileges are not earned through a one shot game, but are built up through consecutive interactions which involve political exchanges between interest groups and decision makers. Corporatist structures are built over long time spans as iterative games between state actors and interest groups in pursuit of solutions to shared problems (Christiansen and Nørgaard 2003; Öberg et al. 2011). State actors exchange influence to interest groups for technical information and political support. The output from corporatist policy making must be functional to the problem in question although seldom a perfect solution because it is negotiated between actors with different interests and it must be accepted by the actors that are involved in its implementation, i.e. corporatist policies must be legitimate. Legitimacy in corporatist policies is created in the interface between representation structure and policy making. Corporatism presupposes relatively strong interest groups with a high membership rate because otherwise they are not able to enter credible commitments with state actors and other interest groups. Groups need to be legitimate representatives of their section of societal life. This implies that groups who enter corporatist negotiations as privileged groups tend to be large and resourceful. However, there are exceptions from this general tendency. In order to legitimize political decisions state actors must secure some minimum balance between interest groups. In some cases this balance is reached through incorporation of small and less resourceful groups. The balance will depend on the structure of interests in a sector and on the political commitment to different considerations (Christiansen, Nørgaard and Sidenius 2004: 276ff.). In a policy area characterized by interest group politics (cf. Wilson 1980), such as labor market politics, privileged groups may all be 4

large and resourceful because this type of politics is characterized by the presence of large groups that outbalance each other. In areas characterized by entrepreneurial politics and client politics (cf. Wilson 1980) groups do not balance each other in the same way. Entrepreneurial politics is difficult to carry out because policy proposals will be opposed by strong groups. This may induce policy makers to ally themselves with pro-reform groups in order to outbalance reform resistance. Some of these groups may be relatively small at least compared to the anti-reform groups. An example is environmental regulation where one or more green groups win a position as a privileged group (e.g. Christiansen, Nørgaard and Sidenius 2004: 218ff.). Client politics is easy to bring about because policy benefits are acquired by strong groups. For decision makers the problem is not to carry policy through but to stand their ground and reduce demands from the clients. In order to include other considerations than those of the clients, policy makers may ally themselves with groups that represent alternative views. Industrial subsidies may be an example where decision makers seek the support from others besides the subsidized industries and therefore include smaller and less resourceful groups (Christiansen, Nørgaard and Sidenius 2004: 167ff.). Another example of client politics is public sector regulation where the privileged clients are typically well organized and resourceful groups of public employees. In order to supplement the inputs from these traditional suspects, decision makers may wish to include other considerations in policy making, for instance by privileging selected user groups (reference). In sum: Corporatism involves the privileging of some groups at the cost of others. In order to win a privileged status, groups must have the necessary resources in terms of members and manpower to participate in the exchange with other groups and decision makers. Many privileged groups are therefore large and resourceful. However, since corporatist structures reflect many different political considerations, some privileged groups are relatively small and less resourceful. This by no means indicates that corporatist representation structures are balanced; they are per definition unbalanced because of the privileging effect. The relevant question is how unbalanced a corporatist system of interest representation is compared to a non-corporatist system. Let s do the reasoning in a case in which a single country over some time is transformed from a relatively strong corporatist system to a relatively weak corporatist system. This will have several implications for the way interest groups are involved in the policy-making and policy-implementing process. First, decorporatization implies by definition less privileging. Without a guarantee of being invited to negotiations or consultations, groups must consider from case to case how they can earn influence in an exchange with decision makers and other groups. 5

Second, interest groups will more often have to take the initiative to contact decision makers (Christiansen and Nørgaard 2003: 108ff.). The logic of privileging implies a continued relationship with frequent, multifaceted, and reciprocal contact between decision makers and interest groups. Easing the logic of privileging implies less frequent, less multifaceted, and less reciprocal contacts. Consequently, relatively more initiatives for contact will come from interest groups in the pursuit of access and influence because there is no guarantee of automatic inclusion. Third, since access to administrators is less privileged and intense, groups will try to compensate through the (further) exploitation of other venues (Christiansen and Rommetvedt 1999) such as increased parliamentary contacts (Rommetvedt et al. 2013), increased media appearance (Binderkrantz, Christiansen and Pedersen 2014) or even using the judiciary (Varone, Ingold and Jourdain 2014). Fourth and finally, since decorporatization does not mean that decision makers do not need to interact with interest groups, decorporatization leaves room for access for hitherto non-privileged groups. More groups will have the chance to affect political decisions, because less regulation of access will be exploited by new groups trying to make their case. The three first consequences of declining corporatism the demand to earn access from case to case, more initiative on behalf of the groups, and the demand to exploit more venues point in one direction in terms of demands on groups: For formerly privileged groups access to decision makers will be more complicated, and it will be more demanding than earlier to deliver in the exchange relation with decision makers. This point in the direction of increased bias in group access, if we presuppose that the ability to deliver increases with increased interest group resources. The last consequence openings for new groups points in the opposite direction. It is most likely that groups with improved chances of access are on average less resourceful than the groups under a corporatist regime. Since we cannot theoretically determine whether a corporatist or a pluralistic system implies the highest level of equality/inequality, we must answer empirically. The proof of the pudding is in the eating. Design and data There are many possible ways to test how equality of interest group access is affected by decorporatization. One way would be to compare corporatist and non-corporatist countries. However, we would not know to what extent differences in equality are produced by the absence of corporatism in non-corporatist countries or if they are wholly or partially produced by something 6

else. The solution chosen here is a longitudinal study of a single country. I compare Denmark in 1975 and 2010 and some intermediate years. This does not make up a perfect most similar systems design. On the one hand, we control for a large number of factors that are unchanged for Denmark in this period, such as electoral system, basic party structure, structure of the central administration, basic interest group structure etc. On the other hand, Denmark has not remained unchanged over the 30 years: The very basic tradition for privileged access has changed in favor of more ad hoc integration of interest groups in policy formation (Öberg et al. 2011); political agendas were dominated by expansion of policy programs in the 1970 and by retrenchment after the turn of the century (Christiansen 2014); and political communication has become increasingly mediatized (Binderkrantz 2014). These changes may have consequences for the distribution of access to decision makers. However, we expect that declining corporatism is the most important of the three variables to affect equality in interest group access. To deal with possible problems of third variables affecting the dependent variable, I use two measures of interest group access. The first measure is membership of public committees. Although corporatism has weakened in the period under observation, a considerable number of committees with members from interest groups remains. Membership of committees is an important indicator of corporatism, because committee membership expresses a privileged position. The functions of the committees have changed dramatically in this period: In 1975, 194 committees were mainly occupied with the preparation of political decisions and had interest group representatives as members. This number had dropped to 14 in 2010 (Christiansen 2014: 127). The number of groups with administrative tasks has not dropped nearly as much and there has been a steady growth in the number of committees with advisory tasks. The committee system is nowadays only rarely used for policy formulation purposes, but the system is by no means dead. I will compare equality in committee membership in 1975 and 2010. Since committee composition is decided by the ministry or the minister membership in 1975 as well as in 2010 we get a measure of the level of equality in an institutional setting in which corporatist norms are still at work even if the general level of corporatism has been reduced. We have a full registration of interest group members of public committees in 1975 and 2010 and the groups to whom they belong. The 1975-data has been collected by Jacob Buksti, Ole P. Kristensen, and Lars Nørby Johansen (Buksti and Johansen 1979; Johansen and Kristensen 1982). Each group has been assigned an ID number in order to trace representatives across committees. In 1975, 374 committees had 1748 interest group seats distributed on 378 different groups. The same researchers conducted a survey of all nationwide interest groups in 1976. From this survey we know group income, the number of members and the number of employees for most of the groups. 7

A comparable data set for 2010 has been established. We compiled a list of all national public committees in work on December 31, 2010. The list was based on many different sources: The Ministry of Finance s homepage, other ministries homepages, news outlets, interest groups homepages, etc. All committee members whom we could clearly identify as representatives of a specific national interest group were coded and their affiliation with interest groups registered. 1964 interest group seats were in 2010 distributed on 273 committees, and they represented 417 different groups. Each group was given an ID number for tracing across committees. Information on group income and the number of political employees was collected via a survey conducted in 2010-11 to all nationwide Danish interest groups. The second measure of access is the frequency and intensity of contacts to different tiers of government. These contacts are not, or at least much less than seat representation, guided by corporatist norms. Contacts thereby measure group interference with decision makers when interference is not restricted by norms controlled fully by the minister or the ministry s civil servants. We naturally expect a positive correlation between access through the committee system and the intensity of contacts between groups and the relevant political actors (Cf. Christiansen et al. 2010), but there is no 1:1 relation. We expect to be able to conclude something about the effects of decorporatization exactly by comparing the structure of access through the committee system and the structure of contacts. We are in the lucky situation to have measures of the frequency of individual groups contacts with bureaucracy and parliamentary actors in both years. In 1976, Jacob Buksti and Lars Nørby Johansen conducted a survey of all nationwide interest groups in Denmark (State Archives DDA-0053). The survey contained questions on the frequency of every group s contact to ministries and parliamentary actors. These questions were repeated in slightly different versions in 1981, 1993, and 2000. A survey conducted in 2010/2011 used some of the 1976-questions again, and this allows us to measure the intensity of the groups contacts to ministries and parliamentary actors. The 2010 list of nationwide interest groups contained 2,543 groups. 1,645 groups answered the questionnaire and of these 1,109 groups declared that they pursued political influence in a broad sense. There is no central registry of interest groups in Denmark. For every survey conducted, the researchers have had to collect the names and addresses of every nationwide interest group from multiple sources, such as existing survey lists, sectoral records on interest groups, newspapers, and from the 1990s and onwards also the Internet. Even if it is probably easier to deliminate the population of interest groups in a country with a corporatist heritage (Christiansen 2012; Fisker 2013), it is not possible to collect records with 100 percent coverage. However, we are quite sure 8

that all large and politically relevant interest group figure in our records, but some minor groups may have escaped our efforts. To sum up, the design of the study implies a double comparison in order to measure the level of equality in interest group access to decision makers in Denmark in the mid-1970s and in 2010. The first comparison measures inequality in committee membership, which is still guided by corporatist norms even if they have somehow changed during the 35 years in focus. This allows us to measure corporatist inequality in both years. The second comparison measures inequality in the intensity of (self-reported) non-institutionalized contacts to decision makers, i.e. contacts that are not (or less) affected by corporatist norms. We expect that inequality has increased more for the latter type of contact than for membership of public committees. Analysis: Inequality in committee seats and contacts We start our analysis by looking at the committee system. As mentioned in the design and data section above, committees with representatives from interest groups are still part of the Danish administrative system, which has changed somewhat over the years: Most importantly, the number of committees with interest group representation has dropped and maybe the most important development committees today are much less used to preparing political decisions than in 1975. Political decisions are today prepared in the ministries with or without interest group participation. In the latter case, it rarely takes place in a committee setting (Christiansen et al. 2010). However, there is still a significant number of committees mainly occupied with advisory and administrative tasks, and they still constitute a significant venue for interest group access. Committees were and are an important part of the Danish politico-administrative system, although their prevalence has been reduced in general and in relation to decision preparation in particular. Figure 1 shows an accumulated distribution of committee seats in 1975 (1748 seats in 374 committees from 378 groups) and 2010 (1964 seats in 273 committees from 417 groups). The figure has two messages: The first is that corporatist norms imply a very uneven distribution of seats. An estimate of the population of nation-wide interest groups yields 2100 groups in 1975 and 2500 groups in 1980 (Fisker 2014). Some of these groups are not politically relevant because they do not have political goals. If we say, based on our experience with collecting lists of group populations, that two out of three groups have political ambitions, we have 1400 and 1700 groups with political ambitions in the two years. This means that only one out of four groups gains representation in a committee. Furthermore, the distribution of represented groups is much skewed. In both years a few interest groups, 27 in 1975 and 20 in 2010, occupy half of the seats. The most privileged interest groups have around 100 seats, whereas a large part of the groups represented have only one seat; 9

46.3 percent in 1975 and 49.9 percent in 2010 were without committee representation! And not to forget: Most of the relevant groups do not have any committee representation. Figure 1: Accumulated distribution of the percentage of seats held by different groups 100 90 80 70 60 50 40 30 20 10 0 1 26 51 76 101 126 151 176 201 226 251 276 301 326 351 376 401 Note: The black line represents 1975 and the grey line 2010. Figure 1 does not give a precise answer to our question about the development of inequality in interest group access. On the one hand the distribution seems to be more skewed in 2010 than in 1975. The curve is simply a little steeper in 2010. On the other hand, more groups are represented in 2010. In Table 1 we have regressed interest group resources on a number of committee positions for each group. As a measure of resources we use the number of employees in each group. Since we expect the relation between employees and committee representation to be non-linear, the number of employees has been logarithmically transformed. 1 As expected and in line with recent research (Fraussen and Beyers 2014; Binderkrantz, Christiansen and Pedersen 2014), we find a positive and strongly significant effect of number of employees on the number of committee seats. We also see that the coefficient is somewhat lower in 2010 than in 1975, i.e. resources have come to play a 1. A few descriptives on the (non-transformed) independent variable. In 1975 the average number of employees among groups with contact with public authorities was 10.53 and the median was 1.0. 42.4 percent of the groups reported no employees. The maximum was 524 employees. For 2010 the corresponding figures were an average of 12.9, median 1.1. 40.4 reported less than one employee and 29.8 reported no employees at all. 10

smaller role for committee representation. Corporatism is still strongly discriminating in favor of resourceful interest groups but less so in 2010 than in 1975. When we split the material on different group types, we get a more nuanced picture. The general tendency holds for a majority of groups, business groups, professional groups and civil society groups, but not for unions and institutional groups. For the two latter groups two of the traditional corporatist partners resources seem to have become more important for committee membership. An interpretation of this result is that the general result less effect of resources is a consequence of the widening of committee membership. More seats are distributed in 2010 than in 1975, which means less competition. Furthermore, new group types have entered the competition with success: Civil society groups have a significantly more seats in 2010 than in 1975 (although few compared to unions and business groups) and they tend to have fewer staff resources. The finding that unions need more resources than earlier in order to win committee seats may be explained by a significant drop in the number of committee seats held by unions. Less seats means greater competition. It s the opposite for business groups, which take more seats in 2010 than in 1975. Less competition means less effect of resources. Table 1: Negative Binomial Regression with number of committee positions as dependent variables, 1975 and 2010 1975 2010 Coefficients Std. error N Coefficients Std. error N All groups Employees (ln) 7.33***.467 1404 6.54***.349 Constant -1.993.110 11.95***.117 Pseudo R 2.118.138 Unions Employees (ln) 5.13 ***.753 205 7.03 ***.782 149 Business Employees (ln) 8.39 ***.751 478 7.42 ***.653 275 Institutional groups Employees (ln) 6.69 *** 1.441 54 7.22 *** 1.12 69 Professional groups Employees (ln) 8.71 *** 2.331 259 4.07 4.31 211 Civil society groups Employees (ln) 6.81 *** 1.41 408 6.08 ***.713 461 Note: Levels of significance: * = 0.05, ** = 0.01, *** = 0.001. We now turn to contacts between interest groups and different actors in two different venues. Interest groups try to access decision makers in different ways: One is to establish different types of contacts: meetings, formal and informal, telephone talks, contacts through email (not in 1975!) and letters, participation in common conferences etc. Contacts may be institutionalized, but typically less than committee membership. Contacts in this broad sense are of course part of a corporatist system, but they are also a sign of other types of state-group relations (Christiansen and Rommetvedt 1999). 1076 11

Our surveys asked respondents about the frequency of their contacts with administrative and political actors: How often do you have contacts with the following actors? Response options: Daily, At least weekly, At least monthly, At least yearly, or Even less /never. The question was asked concerning contacts to MPs, parliamentary committees, party groups, parliament as such, and departments and agencies. For the purpose of this paper the four parliamentary venues have been recoded into one venue. We thus have a single variable measuring contact frequency to parliament. As far as the administrative venue, the 1976 questionnaire asked about the frequency of contacts to departments and agencies, 2 whereas in 2010 the two venues were split. For the purpose of comparison, the 2010 data were recoded into one variable. Table 2 reveals the basic pattern of contacts in the two years and there are several interesting developments: Administrative contacts are still more frequent than political contacts. This is not strange, since there are only 179 parliamentarians compared to some 20,000 administrators. 16 percent of all groups with contact to public authorities report at least weekly contacts to administrators compared to 9 percent for political contacts. However, the lead of administrative contacts has been severely reduced. It is actually quite remarkable how few political contacts interest groups had in 1975. Close to 60 percent of groups with some contacts did not have political contacts compared to 11 percent today. The increase from 11 to 27 percent in the share of groups with rare or no administrative contacts is also remarkable. Table 2. Distribution of political and administrative contacts, 1975 and 2010. Percent a) Daily At least weekly At least monthly At least yearly Rarely/ never Total N 1975 Political contacts 2.0 1.9 12,1 24.9 59.2 100.0 800 1975 Administrative contacts 9.3 12.4 35.2 31.9 11.2 100.0 800 2010 Political contacts 1.7 7.7 22.7 40.7 27.1 100.0 881 2010 Administrative contacts 3.1 12.7 27.0 31.9 25.2 100.0 880 a. Groups were screened somewhat differently in the two years: In 1976, groups were asked whether they had any contact with the political or administrative system. Only groups answering yes to this question are included in the table. In 2010 groups were asked a number of questions regarding their activities. Groups that denied working to affect the views of the population, the media agenda, the political agenda, or the issue of legal instructions or the administration s decisions are excluded from the table. The changes in the pattern of contacts are not uniform for all group types. The pattern for different group types is shown in Table A1 in the appendix: The rise in political contacts has been particularly large for unions, business groups and institutional groups; precisely the groups that were the 2. In the Danish administrative system, most departments are the ministers secretariats and agencies are the implementing bodies. 12

traditional close corporatist partners. It is therefore a reasonable interpretation that the changes in the corporative structures related to the preparation of political decisions have pushed for more political (and administrative) contacts in order to compensate for the loss of access through the committee system (cf. Rommetvedt et al. 2013; cf. also Binderkrantz, Christiansen and Pedersen 2014). If this is a valid explanation, unions appear to have been successful only in the political arena. The fall in the intensity of administrative contacts is particularly strong for unions, professional groups, and civil society groups. For unions this may be a sign of a weakened position in relation to the political and administrative system during the last couple of decades (Klitgaard and Nørgaard 2014). The increase in political contacts appears to be related to especially center-left parties rather than with all parties in parliament (cf. Binderkrantz and Christiansen 2013: 194). The drop in administrative contacts for civil society groups may be an effect of the sharp increase in the number of these groups (Binderkrantz and Christiansen 2014). In Table 3 we run regressions for political and administrative contacts. The dependent variable is the intensity of political and administrative contact. 3 For all groups and all contacts, the coefficient is somewhat higher in 2010 than in 1975, which indicates an increased importance of resources. This is also the result when we split on political and administrative contacts, but with a clear tendency of a bigger difference between the coefficients for political contacts than for administrative contacts. Also here do we find differences between the groups. The general cluster is found for a majority of the groups, i.e. for unions, business groups, and civil society groups, whereas the two smallest types of groups, institutional and professional groups, behave differently: for them contact intensity is less dependent on resources today than in 1975. For unions and business groups the rise in the coefficients for political contacts is particularly remarkable. They are both large group types and they have increased their political contacts significantly in the period. The result is increased competition for political attention and that increases the cost of being in the lobby business. In sum, while resources play a smaller a role in terms of attainment of committee seats, they play an increasing role in terms of contacts to administrations and particularly to political actors. 3. The variable has been constructed in this way: Daily contacts are given a value of 10, weekly 7, monthly 3, yearly 1 and more rarely/never 0. For the variable for all contacts the sum is halved so that this variable can also take values from 0 to 10. 13

Table 3: OLS Regression with contact frequency as dependent variable and log transformed number of employees as independent variable, 1975 and 2010. Standardized coefficients 1975 2010 Coefficients Std. error N Coefficients Std. error N All contacts, all groups Employees (ln).475 ***.051 740.526 ***.019 858 Political contacts, all groups Administrative contacts, all groups Constant 1.195 ***.088.596 ***.055 Adj. R 2.226.276 Employees (ln).314 ***.060 742.489 ***.019 859 Constant.431 ***.087.735 ***.053 Adj. R 2.098.238 Employees (ln).448 ***.075 748.464 ***.016 858 Constant 1.977 ***.130.716 ***.056 Adj. R 2.201.216 Political contacts, unions Employees (ln).375 ***.102 151.548 ***.050 122 Administrative contacts, unions Employees (ln).386***.190 151.396*** Political contacts business Employees (ln).280**.095 225.623***.031 211 Administrative contacts business Political contacts, institutional groups Administrative contacts, institutional groups Political contacts, professional groups Administrative contacts professional groups Political contacts, civil society groups Administrative contacts, civil society groups Employees (ln).471***.144 230.579*** Employees (ln).529***.124 40.421*** Employees (ln).641***.273 40.428*** Employees (ln).421***.082 100.279*** Employees (ln).350***.212 102.229*** Employees (ln).241***.111 222.330*** Employees (ln).377***.120 221.393*** Note: Levels of significance: * = 0.05, ** = 0.01, *** = 0.001. Conclusion Interest groups are important players because they affect the way politically relevant problems are understood, reacted upon, implemented and eventually revised. Interest groups are important democratic players because they bring demand and support to the political system and the political process. However, they also bring bias into the system because they represent specific causes or specific groups interests. Interest groups are not neutral to the political process and to the political output and outcome. That s why the structures that regulate access to political decision making are important. One such structure is corporatism. Among its main attributes is that it privileges some.045.025.067.053 069 032.030.30 122 211 55 54 98 98 369 369 14

groups at the cost of others. In its modern form, corporatism has existed in a number of Western countries for more than 100 years. Corporatism has been a major institution behind the establishment and maintenance of political solutions to societal problems in the labor market and in industrial policies for the first many decades of corporatism and later also for many other policy sectors such as health, education, and environmental policy (cf. Christiansen and Nørgaard 2003). Corporatism was and is biased in favor of privileged groups, which tend also to be the larger groups, but the system was also legitimate and produced fairly effective results. At some point the outputs produced by corporatist institutions became less legitimate and less effective. Particularly state actors became skeptical of parts of the corporatist institutions and their oucomes, particularly those that related to policy formation (Öberg et al. 2011; Christiansen 2014). As a consequence, some corporatist structures in Denmark especially those related to policy formation have been weakened. The question is now how this institutional change has affected interest group access to decision makers. Intuitively the answer is that since corporatism implies the privileging of strong groups, decorporatization implies less privileging and thus more equality of access. Theoretically the question is not easy to solve. On the one hand, decorporatization is more demanding for groups: They have to fight for access from case to case, they have to take the initiative more often, and they have to exploit more venues in order to compensate for loss of access (cf. Binderkrantz, Christiansen and Pedersen 2014). These three factors point in the direction of more resource demanding strategies, which will imply more inequality. On the other hand, new groups or formerly peripheral groups may get their chance as corporatism withers away. All this point towards more equality in group access and calls for an empirical answer. However, the question is not easy to answer empirically because changes in corporatism take place over long periods of time in which changes in other political institutions may have affected interest group access. In this study I exploit the fact that corporatism never regulated all interest group access in Denmark, nor has it yet totally withered away. Denmark has a layered system in which corporatist structures co-exist with other mechanisms that regulate interest group access. By comparing the still existing corporatist structures with other access structures it is possible to compare the levels of equality/inequality at the same point in time. The results are fairly clear: Access to public committees has become less dependent on resources in the form of staff. Despite fewer committees there are more committee seats for interest groups, and new groups have accessed the committee system. It is still strongly biased in favor of resourceful groups, but less so in 2010 than in 1975. We find the opposite development for contacts to political 15

and administrative decision makers, which are less regulated by corporatist norms. It has become more resource demanding for groups to gain frequent contact to decision makers. My interpretation goes like this: Particularly the traditional corporatist partners that used to have significant political influence through the committee system have been forced to develop alternative strategies. They have lost committee seats related to policy formation, and they have lost some of their administrative contacts and they have tried to compensate by establishing more political contacts (cf. also Rommetvedt et al. 2013). The result is increased competition for political contacts and maintaining frequent political contacts has become more resource demanding. Consequently inequality has increased. The results are not uniform for all groups. Among the traditional corporatist partners, unions seem to be more weakened than businesses. At least they take fewer committee seats than before, and gaining access to the remaining corporatist venues as well as the contact venue is more challenging. The development from a strong to a weaker corporatist system implies a sort of deregulation of interest group access. Deregulation means more market regulation. Since free economic markets tend to favor the resourceful, it is maybe not strange that the same thing happens when we establish a (more) free market for interest group access. 16

Appendix: Table A1. Distribution of political and administrative contacts and interest group type, 1975 and 2010. Percent. Contacts a) 1975 Political contacts 1975 Administrative contacts 2010 Political contacts 2010 Administrative contacts Group type Daily At least weekly At least monthly At least yearly Rarely/ never Total N All groups 2.0 1.9 12.1 24.9 59.2 100.0 800 Unions.6 3.6 15.4 23.1 57.3 100.0 169 Business 2.9 1.2 12.2 24.1 59.6 100.0 245 Institutional.0.0 17.4 21.7 60.9 100.0 246 Professional.0.0 4.6 12.0 83.4 100.0 108 Civil society 3.4 2.6 12.1 39.2 48.3 100.0 232 All groups 9.3 12.4 35.2 31.9 11.2 100.0 800 Unions 15.4 19.5 37.3 20.7 7.1 100.0 169 Business 13.9 12.4 33.3 27.9 12.5 100.0 243 Institutional 10.9 13.0 50.0 21.7 4.4 100.0 246 Professional.9 4.6 38.9 44.4 11.1 100.0 108 Civil society 3.4 10.3 31.0 40.1 14.3 100.0 230 All groups 1.7 7.7 22.7 40.7 27.1 100.0 881 Unions 2.3 10.9 31.3 32.0 23.4 100.0 128 Business 1.4 12.5 19.9 42.6 23.6 100.0 216 Institutional 1.8 5.4 42.9 42.9 7.1 100.0 56 Professional.0.0 5.9 40.6 53.5 100.0 101 Civil society 1.7 7.7 22.7 40.7 27.1 100.0 380 All groups 3.1 12.7 27.0 31.9 25.2 100.0 880 Unions 6.3 14.8 28.9 37.5 12.5 100.0 128 Business 6.9 20.4 30.6 25.5 16.7 100.0 216 Institutional 1.8 23.6 52.7 14.5 7.3 100.0 55 Professional.0 5.0 16.8 40.6 37.6 100.0 101 Civil society.8 12.7 27.0 31.9 25.2 100.0 380 a. Groups have been screened a little differently in the two years: In 1976 groups were asked whether they had any contact with the political or administrative system. Only groups answering yes to this question are included in the table. In 2010 groups were asked a number of questions regarding their activities. Groups that denied working to affect the views of the population, the media agenda, the political agenda, or trying to affect the issue of legal instructions or the administration s decisions are excluded from the table. 17

Literature: Anthonsen, Mette, Jonas Lindvall and Ulrich Schmidt-Hansen (2011). Social democrats, unions and corporatism: Denmark and Sweden compared. Party Politics, 17(1): 118-34. Baumgartner, Frank R. and Beth L. Leech (1998). Basic Interests: The Importance of Groups in Politics and in Political Science. Princeton: Princeton University Press. Beyers, J. and B. Kerremans. 2007. Critical resource dependencies and the Europeanization of domestic interest groups. Journal of European Public Policy, 14(3): 460-81. Binderkrantz, Anne Skorkjær (2014). Medierne: kampen om spalteplads, pp. 177-200 in Anne Skorkjær Binderkrantz, Peter Munk Christiansen og Helene Helboe Pedersen (eds.). Organisationer i politik. Copenhagen: Hans Reitzel Publishers. Binderkrantz, Anne Skorkjær and Peter Munk Christiansen (2013). Interesseorganisationerne, pp. 171-214 in Jørgen Grønnegård Christensen and Jørgen Elklit (eds.). Det demokratiske system. 3rd ed. Copenhagen: Hans Reitzel Publishers. Binderkrantz, Anne Skorkjær and Peter Munk Christiansen (2014). Decades of change? Interest group representation in Danish public committees in 1975 and 2010. Unpublished paper: Aarhus. Department of Political Science and Government. Binderkrantz, Anne Skorkjær, Peter Munk Christiansen and Helene Helboe Pedersen (2014). Interest group access to the bureaucracy, parliament and the media. Governance. doi:10.1111/gove.12089. Blom-Hansen, Jens (2000). Still corporatism in Scandinavia? A survey of recent empirical findings. Scandinavian Political Studies, 23(2): 157-81. Buksti, Jacob and Lars Nørby Johansen (1979). Variations in organizational participation in government: The case of Denmark. Scandinavian Political Studies, 2(3): 197-220. Christiansen, Peter Munk (2012). The usual suspects: Interest group dynamics and representation in Denmark. In Darren Halpin and Grant Jordan (eds.). The scale of interest organization in democratic politics. Data and research methods (pp. 161-179). Houndmills, Basingstoke: Palgrave Macmillan. Christiansen, Peter Munk (2014). Organisationernes politiske rolle efter 1975, pp. 55-89 in Anne Skorkjær Binderkrantz, Peter Munk Christiansen and Helene Helboe Pedersen (red.). Organisationer i politik. Copenhagen: Hans Reitzel Publishers. Christiansen, Peter Munk and Asbjørn Sonne Nørgaard (2003). Faste forhold flygtige forbindelser. Stat og interesseorganisationer i det 20. århundrede. Aarhus: Aarhus University Press. Christiansen, Peter Munk and Hilmar Rommetvedt (1999). From corporatism to lobbyism? Parliaments, executives and organized interests in Denmark and Norway. Scandinavian Political Studies, 22: 195-221. Christiansen, Peter Munk, Asbjørn Sonne Nørgaard and Niels Chr. Sidenius (2004). Hvem skriver lovene? Aarhus: Aarhus University Press. Christiansen, Peter Munk, Asbjørn Sonne Nørgaard, Hilmar Rommetvedt, Torsten Svensson, Gunnar Thesen and PerOla Öberg (2010). Varieties of democracy: Interest groups and corporatist committees in Scandinavian policy making. Voluntas, 21: 22-40. 18

Crepaz, Marcus L. (1994). From semi-sovereignty to sovereignty: The decline of corporatism and rise of parliament in Austria. Comparative Politics, 27(1): 45-65. Fisker, Helene Marie (2013). Density dependence in corporative systems: development of the population of Danish patient groups, 1901-2011. Interest Groups and Advocacy, 2(2): 119-38. Fisker, Helene Marie (2014). Gamle venner og nye bekendtskaber: udvikling i den danske interessegruppepopulation. Aarhus: Department of Political Science and Government. Unpublished. Fraussen, Bert and Jan Beyers (2014). Who's in and who's out? Explaining access to policy makers in Belgium. Article submitted to West European Politics. Häusermann, Silja, André Mach and Yannis Papadopoulos (2004). From corporatism to partisan politics: social policy making under strain in Switzerland. Swiss Political Science Review, 10(2): 33-59. Johansen, Lars Nørby and Ole P. Kristensen (1982). Corporatist traits in Denmark, 1946-1976. In Gerhard Lehmbruck and Phillippe C. Schmitter (eds.) Patterns of corporatist policy-making (pp. 189-218). London: Sage Publications Ltd. Kariel, Henry S. (1961). The Decline of American Pluralism. San Francisco: Stanford University Press. Klitgaard, Michael Baggesen and Asbjørn Sonne Nørgaard (2014). Structural stress or deliberate decision? How Governments have Disempowered Unions in Denmark, European Journal of Political Research, 53(2): 404-21. Lewin, Leif (1994). The rise and decline of corporatism: The case of Sweden. European Journal of Political Research, 26(1): 59-79. Maloney, William A., Grant Jordan and Andrew M. McLaughlin (1994). Interest groups and public policy: The Insider/Outsider model revisited. Journal of Public Policy, 14(1), 17-38. Mansbridge, Jane (1992). A deliberative perspective in neocorporatism. Politics & Society, 20(4): 493 505. Molina, Oscar and Martin Rhodes (2002). Corporatism: The past, present, and future of a concept. Annual Review of Political Science, 5: 305-31. Öberg, PerOla (2002). Does administrative corporatism promote trust and deliberation? Governance, 15(4): 455-75. Öberg, PerOla, Torsten Svensson, Peter Munk Christiansen, Asbjørn Sonne Nørgaard, Hilmar Rommetvedt and Gunnar Thesen (2011). Disrupted exchange and declining corporatism: government authority and interest group capability in Scandinavia. Government and Opposition, 46(3), 365-91. Olson, Mancur (1965). The Logic of Collective Action. Cambridge, MA: Harvard University Press. Richardson, Jeremy (2000). Government, interest groups and policy change. Political Studies, 48(5): 1006-25. Rommetvedt, Hilmar, Gunnar Thesen, Peter Munk Christiansen and Asbjørn Sonne Nørgaard (2013). Coping with corporatism in decline and the revival of parliament: interest group lobbyism in Denmark and Norway, 1980-2005. Comparative Political Studies. 46(4): 457-85. Schattschneider, Edgar E. (1960). The Semisovereign People: A Realist s View of Democracy in America. New York: Holt, Rinehart and Winston. Schlozman, Kay Lehman, Sidney Verba and Henry E. Brady (2012). The Unheavenly Chorus: Unequal Political Voice and the Broken Promise of American Democracy. Princeton: Princeton University Press. 19

Schlozman, Kay Lehmann and John T. Tierney (1976). Organized Interests and American Democracy. New York: Harper and Row. Schmitter, Phillippe and Terry Lynn Karl (1991). What democracy is and is not. Journal of Democracy, 2(3): 75-88. State Archives DDA-0053. Interesseorganisationer i Danmark. Odense. Varone, Frédéric, Karin Ingold and Charlotte Jourdain (2014). Venue shopping, coalition building and advocacy success: evidence from California interest groups. Paper for the Interest Group Politics conference, Aarhus University, June 2-3, 2014. Wilson, James Q. (1980). The politics of regulation, pp. 357-447 in James Q. Wilson (ed.). The Politics of Regulation. New York: Basic Books. 20