GAINING POLITICAL CAPITAL THROUGH SOCIAL CAPITAL : POLICY-MAKING INCLUSION AND NETWORK EMBEDDEDNESS OF MIGRANTS ’ ASSOCIATIONS IN SPAIN

This article analyzes the relevance of network embeddedness and social capital in allowing migrants’ associations to gain political capital and access to policy making in the cities of Barcelona and Madrid. With data from a survey of migrants’ associations in both locations, we examine the degree to which embeddedness in networks of links with other migrants’ associations and with autochthonous Spanish civil society organizations are consequential for the inclusion of migrants’ organizations in policy-making processes. The results show that migrants’ organizational social capital is critical in facilitating their intermediation function vis à vis political institutions and decision makers, above and beyond their access to financial and human capital.

The integration of immigrants into the societies where they come to settle has become a widely contested topic in most western democracies.Both media commentators and scholars frequently express concern over what they consider a limited degree of social, cultural, and political integration of immigrant-origin communities in receiving societies.Recently, discussion has become particularly intense within political science due to the pessimistic outlook portrayed by two very prominent scholars (Huntington 2004;Putnam 2007).
Underlying these debates is a common preoccupation about-as well as a certain desire to shape-the policies that governments at various levels should enact to encourage migrants' integration into the receiving society (see Ireland 1994;Soysal 1994;Favell 1998;Heckmann and Schnapper 2003;Koopmans, Statham, Giugni, and Passy 2005;and Givens 2007).To what extent should collective needs and identities be addressed with public policies that actively promote integration through sponsoring migrants' associations and their capacity to selforganize?Is it better-in the long run-to avoid having explicit integration policies and let immigrants seek their own resources to form associations? Or are active multicultural policies that encourage certain forms of immigrant organizing more effective in reducing conflict and empowering migrant groups?Scholars are far from reaching an agreement as to whether multicultural policies are more effective than assimilationist ones in empowering migrant populations and incorporating them into the polity of the countries where they live (see, for example, Koopmans 2003Koopmans , 2010;;Parekh 2008;Duyvendak, Pels, and Rijkschroeff 2005).Some argue that multicultural policies are _______________________________ * The data for this article have been collected, processed, and analyzed with the support of the LOCALMULTIDEM project (CIT5- CT-2005-028802) funded by the European Commission's 6th Framework Programme, and the CAPSOCINMIG Project (SEJ2005-07733/CPOL) funded by the Spanish Ministry of Science and Education.Further information regarding these two projects can be found on their respective websites: http://www.um.es/localmultidem and http://www.um.es/capsocinmig.A previous version of this article was presented at the 2008 WPSA conference with the support of the research development fund of the Institute for Social Change at the University of Manchester.We thank the participants in the conference panel, and especially Jonathan Hiskey, as well as the two anonymous reviewers of Mobilization for their thoughtful comments to previous versions of this article.† Laura Morales is a Research Fellow in the Institute for Social Change at the University of Manchester.Luis Ramiro is an Associate Professor in the Department of Political Science at the University of Murcia.Please direct all correspondence to Laura Morales, Institute for Social Change, 2nd floor, Humanities Bridgeford Building, University of Manchester, Oxford Rd, Manchester M13 9PL, United Kingdom, or by email at laura.morales@manchester.ac.uk.beneficial for the incorporation of migrants and ethnic minorities to the extent that they nurture the self-organization of migrants in associations and provide institutional channels for their participation in policy making (see Fennema and Tillie 2001;Bloemraad 2005Bloemraad , 2006)).
While these questions lie at the core of current popular and scholarly debates on the public management of immigration, this is not the place to address them properly (see Morales and Giugni 2011).Yet, what all these models of migrant integration seem to share is an implicit understanding that migrants' social and cultural integration is-at least to a certain extentfavored by their political incorporation to the public arena (see Jones-Correa 1998).Active engagement in electoral politics is among the most successful mechanisms to promote the full consideration of migrant minorities as citizens worthy of politicians' attention.Because parties and candidates are vote maximizers, the dynamics of electoral competition make migrant and ethnic minorities an attractive new pool of voters whose electoral preferences can be molded to form loyal constituencies-especially when these groups are large enough to affect an election's outcome (Leighley 2001).In this regard, the literature on political transnationalism has shown that the sustained transnational links among migrant groups contribute to the emergence of migrant political entrepreneurs who help bridge the politics of the homeland with local politics (Basch, Glick-Schiller, and Szanton-Blanc 1994;Guarnizo, Portes, and Haller 2003;Portes and Rumbaut 2006).
But electoral politics is not the only mechanism through which this political incorporation takes place.In fact, what happens between elections should be regarded as equally important as the elections themselves.Through lobby and advocacy activities, migrant communities that actively engage in the public arena are able to shape the real political choices that elites make.This political activity is exercised by various types of civic and advocacy organizations that act as intermediaries between elites and migrant groups (see Knoke andLaumann 1982 andKnoke 1990).Attending meetings at the local council, joining committees that debate policy decisions, and participating in crisis teams or blue-ribbon panels to solve episodes of social conflict are some of the ways that migrants' associations exert a continued impact on the politics of the countries where they come to settle.
At the very least, regardless of the citizenship or incorporation model the receiving country follows, the migrant communities that engage in such activities through their own organizations will be in a better position to further their interests and well-being.In this sense, the ability to voice the concerns of the community is the very first step to being able to shape policy outcomes.But voice and access are not equally distributed among social groups, or-for that matter-among civil society organizations.Some associations actively engage in public affairs, while others refrain from politics; and even among those interested in lobbying and advocacy, some are able to build political coalitions that facilitate access to policy making while others remain marginalized.Learning more about the processes that lead some civil society organizations to gain greater access to policy makers while others are less privileged is thus key to an understanding of the biases within the policy process.This is particularly important in a policy sector like immigrant integration where governments-national, regional, and local alikearound Europe and elsewhere are increasingly willing to actively engage with migrants' associations when designing their policy strategies. 1 Yet, most of the scholarly literature has concentrated instead on describing and theorizing about what structural and institutional aspects lead to immigrants being able to successfully self organize.In the process it has neglected those factors that lead immigrant organizations to be more or less influential. 2 What helps migrants' organizations influence policy making?What resources-economic, human, and social-condition their capacity to influence decision-making processes?This article draws on the experience of organizations in two Spanish cities to underscore the role of social networks-or social capital-in advancing the empowerment of migrants' organizations.
Social capital-in the form of links with other migrants' organizations, and ties to mainstream political, advocacy, and civil-society organizations-enhances the voice of migrants' organizations in the policy process, at least in certain policy contexts.As we will show, social capital turns out to be as strong as-or stronger than-economic and human capital as a determinant of policy inclusion.Financial and bureaucratic resources-though clearly important-are not the main drivers of access to the policy domain for migrants' organizations in Spain.This can be contrasted with the position originally advanced by resource mobilization theory (McCarthy and Zald 1977), and also supported by past research on the political economy of associations (Knoke 1990).As our findings suggest, the social capital of migrants' organizations decisively determines their access to policy makers and decision-making opportunities.This is, most likely, a consequence of the additional information, influence, and credentials that social capital provides to these social agents (see Lin 2001: 19-20).These assets happen to be crucial in a context where public administrators have limited and imperfect information about migrant communities' relevant interlocutors.
This article contends that in contexts where the policy process is characterized by the lack of clear procedures, structures, and practices of social consultation and intermediation-such as the immigration policy domain in Spain-social capital will emerge as one of the most relevant resources to achieve policy inclusion.Rather than relying on vast financial resources or highly skilled human resources, organizations can instead focus on building and mobilizing contacts, thus acquiring centrality, visibility, and "prestige."Their ties to a wide range of organizations and actors will establish them as "relevant" in the eyes of policy makers, who have not (as yet) instituted formal mechanisms or practices of interest intermediation.
In this sense, our use of two case studies-of the policy inclusion of migrants' organizations in Spain-illustrate situations in which a policy field is in the making.Substantial migration flows to the two cities considered in this study-Barcelona and Madrid-are relatively recent phenomena, as is the case all around Spain more generally.During the last decade, both cities have become major immigration magnets in Spain, and Europe more broadly.This is especially true of Madrid for Latin Americans from Andean countries (Bolivia, Colombia, Ecuador, and Peru).High rates of economic growth between 2000 and 2008 attracted a large number of immigrants-mostly economic or labor migrants-whose entry to the country was facilitated by a very open visa policy that was common for many Latin American citizens travelling to Spain. 3 The unprecedented nature of massive immigration to these two cities specifically, and the country more generally, has meant these inflows have taken place in the absence of well-defined policies and actions in most areas of immigration management and immigrant integration (see Morales, González, and Jorba 2009).
Moreover, immigrants of Latin American origin benefit from considerable privileges in many respects.The most relevant is the easier and quicker access to Spanish citizenship, as they are only required to document two years of continued legal residence in the country when the general rule-applying, for example, to Moroccans, one of the largest immigrant groupsrequires ten years.Additionally, the fact that Latin American immigrants, for the most part, also speak Spanish as their mother tongue goes a long way toward facilitating their integration individually and as a group, since this endows them with easier access to various social and public resources-for example, access to the mainstream media.
With regard to policy making, while border controls and entry policies are the sole responsibility of the national government, regional governments are allowed to regulate immigration issues within the limits of their own powers (health care, education, social services, etc.).However, these powers have not been extensively developed thus far, and regional governments and local governments mostly manage daily integration issues.Moreover, whereas access to social services and welfare rights are generously granted, even to undocumented immigrants, by all levels of government, voting rights at the local level have not yet been recognized for most non-EU citizens, and local governments have no powers to extend these rights. 4 Overall, the situation is one characterized by unclear distribution of powers over core immigrant integration policies, with multiple overlapping policies implemented by various levels of government.
This lack of a systematic and clearly defined model of immigration and migrant-integration policy coexists with the absence of a clear model of interest mediation.Whereas trade unions and business organizations have been fully institutionalized and incorporated into the policy process through various coordinating mechanisms, interest intermediation in Spain is generally not well-structured for policy domains other than labor and economics. 5Furthermore, the lack of structuration and the habit of consultation with a wider range of organizations in most policy arenas has resulted in the omnipresence of trade unions and business associations as the main actors.This is also the case for integration policies in Madrid, and in Spain more generally.The case of Barcelona diverges slightly in this regard, as the local government of this city has enacted a longstanding approach of wider consultation mechanisms through a multiplicity of consultation and participation bodies or committees that are much more inclusive, and which incorporate a wider range of civil society organizations (see Morales, González, and Jorba 2009 for a detailed description).
In this article, our main aim is to assess the impact of migrant organizations' social capital on their incorporation into the policy process.To this end, we explicitly compare the impact of social capital to the effect that other organizational resources-financial and human capitalhave for achieving the same goals.First, we discuss the role that the literature on associations has assigned to social networks and ties.Structures of social connection are generally regarded in the scholarly literature as instrumental to collective action, insofar as they foster the flow of communication and information, facilitate recruitment, and increase access to other financial resources and infrastructure.
After describing the data and the methods employed to collect them, we present our empirical findings.With multivariate regression models we show that the social capital of migrants' associations in Barcelona and Madrid is of utmost relevance for getting access to the host country policy process.Economic capital turns out to be an asset of limited value for gaining policy-making access in these two Spanish cities.In contrast, the position of migrants' organizations in the network of contacts with other migrants' associations and with local autochthonous organizations has a crucial impact on their capacity to have a say in decisionmaking processes.We conclude the article with a discussion on why this is the case, and how further research can shed more light on this topic.

CAPITAL, NETWORKS AND POLITICAL INCORPORATION
Scholarship on the political economies of associations has demonstrated the relevance of three forms of capital in supporting their political mobilization and advancing their capacity to influence the policy process (Knoke 1983(Knoke , 1990)).Financial resources-especially in contexts where lobbying requires intense pressure-are necessary for organizations to have a chance at mobilizing supporters.Bureaucratic structures, organizational differentiation, and professionalized staff are frequently necessary to mobilize the organizational and human skills that the complexities of policy intervention require.How organizations' social capital works to foster or hinder access to the policy process is less well understood (see Edwards and McCarthy 2004).Knoke (1983) shows that the type and diversity of links with sponsors have clear effects on members' evaluations of goal effectiveness through different perceptions of sponsor reputations.In turn, these evaluations are important for individual members' commitment and involvement in mobilization processes.Similarly, Ansell (2003) illustrates how the nature of interorganizational links between associations has an impact on their capacity to engage in collaborations with public authorities: associations that are more inclined to reach out to organizations beyond the strictly local realm are also more inclined to collaborate with governmental elites.
Yet, while various scholars acknowledge that networks and connectivity matter for organizational capacity and influence (see Diani 2003 for a summary), the underlying processes at work are unclear.Most existing studies point to information and communication flows.Fennema andTillie (1999, 2001) have argued that connections among ethnic organizations are positively affects migrants' participation in the host polity. 6The primary reason for this is that these links act as channels of communication that promote greater collective information and trust, and link the newcomers' organizations to the native political elite.
High connectivity and centrality are also mobilization resources (see various chapters in Diani and McAdam 2003).Better connected organizations are better at mobilizing supporters and forming coalitions that will enhance their ability to influence policy makers.In particular, ties with the political parties and the governmental institutions increase an organization's capacity to influence policy making (Klandermans 1989).However, given that-in one way or another-social networks primarily act as resources, and that the value of specific resources is dependent on the cost-benefit structure of action, the impact of connectivity on organizational action and policy influence is likely to be very sensitive to the mobilization context in which these networks are embedded (see Gould 1993).
A different, but related, debate focuses on the types of linkages that are more effective in promoting integrated cooperation.The debate about "bridging" and "bonding" social capital (Putnam 2000) has also been extended to analyses of the effectiveness of organizational activity and the implications for integration and social cohesion.Bonding social capital or "strong" ties favors internal group cohesion and identity formation (Granovetter 1973(Granovetter , 1983)), while bridging social capital or "weak" ties allows communication to expand farther and to promote wider (or macro) social cohesion and exchange. 7Therefore, different types of embeddedness in organizational networks are likely to produce different outcomes in terms of the integration and capacity of migrants' organizations (Jacobs and Tillie 2004).
However, it is unclear whether bonding relationships will necessarily be detrimental for the political incorporation of migrants' organizations to the host polity.While a narrow focus on identity formation might prevent them from engaging in wider political action, the scholarship on social movements has also clearly established that strong collective identities can be an asset to mobilize less advantaged social groups (Taylor and Whittier 1992;Friedman and McAdam 1992).Consequently, a high degree of connectivity among migrants' organizations-even if only one of the multiple indicators of "bonding" ties-multiplies their social capital and is likely to have an impact on their political action and effectiveness.
This review of past research on the impact of the forms of capital and, in particular, of interorganizational networks and social capital leads us to the following expectations: • Migrants' organizations with greater financial and human resources are more likely to be included in the policy process.• Migrants' organizations with higher levels of social capital-in the form of collaborative ties with other migrants' organizations and with other mainstream autochthonous organizations-are more likely to be included in the policy process.• While both higher degrees of "bridging" (with autochthonous organizations) and "bonding" (with other migrants' organizations) social capital are expected to increase the chances of inclusion in policy making, we expect that "bridging" connections will be more consequential than "bonding" ones in this regard.
In the next section we describe the data and methods used to test these propositions in the empirical section of this article.

DATA AND METHODS
Our definition of "migrants" includes those of foreign birth, plus their immediate descendants, who are commonly referred to as first and second generations. 8Two additional definitions are fundamental in determining our selection of the organizations were the objects of analysis.First, we defined an association as "a formally organized named group most of whose members-whether persons or organizations-are not financially recompensed for their participation" (Knoke 1986: 2).Thus, family, clans, groups of friends, commercial organizations, foundations, or government agencies were not included.Second, we defined an association as a "migrant" organization when, at a minimum, about half of its members or half of its leadership were of migrant origin, including first and second generations.We therefore used a relatively broad definition of the term "migrants' organization," but we did not include associations mainly devoted to work with or for migrants if mostly composed by autochthonous Spaniards.
The collection of the data on migrants' associations in Barcelona and Madrid proceeded in two stages between 2007 and 2008.First, because no reliable single list was available in any Spanish city, we produced a list of all existing migrants' organizations from multiple sources. 9Second, we organized interviews with organizational leaders and administered a face-to-face questionnaire of approximately one hour of duration.We excluded from the study those organizations that were not formed by migrants, for example, promigrant advocacy organizations, and those that no longer existed.
In total, we detected close to 900 migrants' organizations in the two cities during the listing phase.Many had disappeared, were ineligible according to our criteria, or were "dormant" by the time we started fieldwork.After sifting the initial list, more than 400 organizations-223 in Barcelona and 199 in Madrid-were confirmed as active, and we arranged face-to-face interviews with all eligible organizations that agreed to participate in the study, completing 163 questionnaires: 100 in Barcelona and 63 in Madrid. 10The response rates were about 45 percent in Barcelona and 32 percent in Madrid (table 1). 11In most cases, nonresponse was due to the organizational leaders failing to keep with the appointments, refusing to participate in the study, or to the impossibility of getting accurate contact information for the organization. 12The final results are, thus, likely to provide a relatively accurate picture of the migrants' organizations with some activity and visibility in both cities.
Our fieldwork results suggest that the associational density of the migrant population isas might be expected-lower when compared to that of the autochthonous population: around 1.6 associations per 1,000 migrants in Barcelona and 0.8 in Madrid, as opposed to an average 5.7 per 1,000 inhabitants when the total population is considered. 13However, as in the case for the autochthonous population, associational density is larger in Barcelona than in Madrid, a pattern consistent with the view of a more "vibrant" associational field in the former city.
Furthermore, although Latin American organizations dominate the associational field of migrants in both cities, they are not grossly overrepresented when compared to their share in  the foreign resident population (see table 2). 14Indeed, non-Maghrebi African associations are more numerous than the share of the corresponding population would lead us to expect, and there is also a substantial proportion of associations in both cities that combine members and leaders of multiple world regions.Thus, even if Latin American organizations are more numerous, the field of migrants' associations in both cities is quite heterogeneous in its origin and composition.

RESULTS
Our survey asked organizational leaders about their involvement in a number of forms of policy making at the three main levels of government in Spain: local, regional, and national.
The questionnaire asked first whether the organization or their leaders had received an invitation to participate in the policy process through any of the multiple mechanisms that were available at each level of government and, when invited, whether they had eventually participated (see table 3, next page).Only a limited number of migrants' organizations received requests to participate in the policy process in the two years prior to the survey, and these invitations were much more frequent at the local level of government than at the regional or national level.Yet, those that are invited are very likely to accept and participate in policy decision making, as almost all associations that received an invitation reported having participated.Overall, however, 48 percent of associations in Barcelona and 68 percent in Madrid were invited to participate in decision-making structures at some level of government and similar proportions had eventually participated.Overall, the degree of policy inclusion of migrants' organizations is fairly frequent in Barcelona and even more common in Madrid.Moreover, these invitations are often linked to stable and formal procedures-for example, invitations to join a committee or advisory council as a permanent member-than through more informal and ad hoc mechanisms-such as occasional invitations for consultation.In any case, above all, what determines participation in decision-making structures is being invited to participate in the first place. 15So next we will concentrate on examining how different forms of capital influence the likelihood of migrants' organizations being invited to join the policy process.

The Measurement of Economic, Human, and Social Capital
Unlike Edwards and McCarthy (2004), who distinguish among five forms of resourcesmoral, cultural, social-organizational, human, and material-we focus only on three main forms of capital that we consider the key potential drivers of inclusion in policy making: economic or material, human, and social capital.Table 4 illustrates the limited amount of economic, human, and social capital of migrants' associations in both Spanish cities. Migrants' organizations have generally a very limited amount of financial resources available: while the average budget is around €22,000 in Barcelona and €59,000 in Madrid, the total budget of the median organization is around €3,000 in Barcelona and €7,000 in Madrid.Furthermore, 13 percent of the interviewed associations in Barcelona claim to have no operating budget whatsoever.This large gap between the average and the median values is due to the existence of a few organizations with very large budgets.
This lack of financial resources is reflected in the limited availability of qualified human resources in the form of permanent and professionalized full-time staff.Only 7 percent of all migrants' organizations in Barcelona and 24 percent in Madrid have at least one full-time staff member.The small size of the typical migrants' associations explains both staff size as well as the lack of financial capital.Average membership is around 700 members in Barcelona and 1,600 in Madrid, but this figure is significantly skewed by a few large organizations, concealing a median membership size of around 50 individuals for the migrants' organizations in Barcelona and 160 in Madrid.When it comes to material and human resources, migrants' organizations are substantially better off in Madrid than in Barcelona (see Morales, González, and Jorba 2009).
If we turn our attention to the social capital of migrants' organizations, we notice that ties with autochthonous associations prevail over links with migrants' organizations in both cities.Unlike Fennema and Tillie (1999) and Fennema (2004), our primary source of information is not interlocking directorates, nor do we have detailed information on overlapping memberships (as in Diani 2003).Our respondents were presented with a set of questions on contacts and collaborations with other organizations.First, we asked them to provide the names of the ten organizations in their respective cities-migrant or autochthonous-with which they had most contact in the past two years.In addition to this open ended question, we asked two additional questions, one based on a list of several local autochthonous Spanish organizations (parties, unions, environmental groups, human rights associations, antiracism organizations, NGOs, charities, neighborhood associations, and religious organizations), and the other based on a list of all the migrants' organizations that we had located during our census-creation process. 16For both sets of organizations, we asked about contact and close collaboration, but here we only report responses regarding contact (weak ties), since the information that might become useful for policy inclusion flows farther through weak ties.The average migrants' organization is in contact with more than 9 other migrants' organizations in Barcelona (median value of 6) and 7 in Madrid (median value of 4), while it is "weakly" linked to almost 10 Spanish autochthonous organizations in Barcelona (median value of 7) and with 11 in Madrid (median value of 6.5).Closer collaborations are, logically, much less frequent in all cases.In general, thus, migrants' organizations in Barcelona would seem to be slightly better connected both with other migrants' organizations and with autochthonous civil society.However, the picture is less clear than it would seem at first glance, given that 10 percent of all migrants' associations in Barcelona mention no contact whatsoever with other migrants' organizations and 8 percent are not linked to any autochthonous organizations, whereas the respective figures for Madrid are 3 and 13 percent.
Along with the economic, human, and social capital of migrants' organizations, four additional organizational properties are thought to be relevant for the policy inclusion of these associations, and we also provide their descriptives in table 4. On the one hand, we mentioned that immigrants of Latin American origin benefit from a certain advantage in Spain due to sharing the same language and to the favorable nationality rules that apply to them.It would, thus, be reasonable to expect that Latino organizations might have a greater advantage when it comes to influencing the policy process in these two cities.And, as we see in table 4, around 70 per cent of all migrants' associations we interviewed in Barcelona are of Latino origin, and so are 65 percent of those questioned in Madrid.
Additionally, the number of years the association has been active, as well as the nature of their policy aims, have commonly been considered important factors for the policy action and impact of organizations (see Knoke 1990, andDiani 2003).Organizations that have been in the public arena for a longer time are more likely to be considered adequate intermediaries for the migrant constituencies.Equally, organizations whose main purpose is to influence decisionmaking processes are also likely to allocate the necessary organizational resources to influence policy and to be regarded as potential representatives of migrants' preferences and demands by policy makers.
Given that massive migration flows into Spain and the two cities we study are a relatively recent phenomenon, most migrants' organizations in the city are relatively young.More than 50 percent of them were founded after 2000, and the average age of the organizations is 7 years in Barcelona and 10 in Madrid.Furthermore, only a limited number of migrants' organizations (less than 30 percent in both cases) have lobbying and other political activities as one of their main organizational goals.Many migrants' organizations in Madrid are primarily cultural-musical, artistic, and so on-associations that rarely engage in political activities; and while many others focus on issues and topics related to the migration process, they mostly regard themselves as service providers rather than as advocacy groups.Thus, the politici-zation of these organizations is relatively limited.

The Role of Economic, Human, and Social Capital for the Policy Inclusion: Madrid
Table 5 presents the results of the multivariate logistic regression models of policy inclusion on the several organizational attributes that reflect the economic, human, and social capital of the interviewed migrants' associations.The table presents five different models of  Localmultidem andCapsocinmig projects (2007-2008).Notes: Values are the logit coefficients, with standard errors within parentheses.Significance marked as * for p≤ 0.10; ** for p≤ 0.05 the variable that measures whether the association has received any invitation to participate in the policy process at any level of government.The first four models are nested models that show the results when different sets of variables are incorporated, while the fifth model is more parsimonious as it excludes indicators of lobbying activities as well as measures the budget available to the organizations. 17All the models pool the cases for Barcelona and Madrid because separate analyses yielded very similar results for both cities, both in terms of the size and direction of the coefficients, and merging the cases makes our estimations more robust and efficient given the relatively limited number of cases.
If we focus our attention on the control variables, several interesting findings emerge.First, in contrast with the expectations set by the literature on the political economy of associations, the age or length of existence of migrants' organizations is generally irrelevant for their policy inclusion.Older organizations are not any more likely to receive invitations to participate in policy-making bodies than those of more recent formation.This indicates that policy makers are not biased towards those migrants' organizations that have been in the public arena for a longer period of time, and that older organizations are not able to mobilize more resources of any sort into gaining access to the policy process.
Second, contrary to what we expected-and what most observers would predict-there is no positive bias in favor of the organizations of Latin American migrants, but instead quite the opposite.Associations whose leadership or membership is primarily composed of Latin Americans are less likely to be invited to participate in policy-making mechanisms than other migrants' organizations.This is because Latin American associations are much more heterogeneous with regard to their main purpose and goals.In general terms, Latin American associations are more frequently devoted primarily to ensuring the survival of the cultural customs and traditions of these migrants.The other reason for this result lies in the fact that policy makers are very aware of the need to diversify representation in consultative bodies and mechanisms by migrant origin.Due to the limited number of seats in these bodies, they are likely to underrepresent Latin American organizations relative to their share in the associational field and the migrant population.
Third, the fact that some migrants' organizations are primarily pursuing political and lobbying activities while other associations do not view these areas of action as one of their primary concerns is not a significant factor in policy inclusion.In fact, migrants' organizations that declare "lobbying" as one of their main activities are not more likely to receive invitations to participate in the policy process than the rest; and those that view politics as one of their main concerns are more likely to receive such invitations but just because they are also more likely to be better connected to autochthonous organizations. 18 Finally, although the descriptives of the variables of policy inclusion presented in table 3 suggest that migrants' associations in Madrid are more often invited to participate in the policy process, the multivariate regressions in table 5 introduce some doubts about the nature of this gap.Throughout models 1 to 4, the dummy variable for Madrid remains statistically insignificant, and it only approaches the standard levels of significance in models 4 and 5 once all the social capital indicators are included.Thus, though these results are not very clear-cut, it seems that the more frequent inclusion of migrants' organizations in Madrid than in Barcelona is partly related to the different distribution of resources of these associations in the two cities.
Turning our attention to the variables that are at the core of our concern, as we have discussed in previous sections, the resource mobilization theories put forward by social movements scholars (McCarthy and Zald 1977;Edwards and McCarthy 2004) and the literature on the political economies of associations (Knoke 1990) led us to expect that financial and human capital would be crucial resources for policy inclusion and impact.Yet, our results show that financial or material resources have no significant effect on the likelihood of migrants' organizations receiving invitations to join the policy process in either of the two cities or when the data for both are combined. 19 However, the two indicators of human capital are significant, and are critical resources for policy inclusion, especially the size of the membership base.As figure 1 shows, migrants' associations with no full-time staff have a probability of 0.55 of being invited to participate in the policy process, while having at least one member of staff devoted full-time to this job increases the likelihood to 0.9. 20Additionally, the chances of being invited to participate in decision making gradually increase as the membership size does, but a very modest increase in the number of members from very few (2-3 members) to a still-small group (50 members) yields a substantial increase in the probability of being invited to participate from 0.19 to 0.49.
In sharp contrast to the limited impact of economic capital, social capital indicators are clearly relevant in accounting for the patterns of policy inclusion of migrants' organizations  in Barcelona and Madrid.Both the indicator that taps into "bonding" social capital (contacts with other migrants' organizations) and that which relates to "bridging" social capital (contacts with autochthonous organizations) significantly increase the chances that a migrant association will be invited to have a say in policy making, as expected.Indeed, also as hypothesized at the beginning of this article, "bridging" contacts are slightly more consequential than "bonding" (see figure 2).In fact, an equivalent increase in the number of contacts from 0 to 40 results in a 2.6 times increase in the probability of being invited to policy-making processes when the ties are with other migrants' organizations, and in a 4.4 increase of that same chance when the links are with autochthonous organizations.
Nevertheless, we note that the impact of human capital factors is much more substantial than that of the social capital indicators we have taken into account.If we use as the baseline a "resourceless" migrants' organization-that is, one with no full-time members of staff, just 3 members, no contacts with other migrants' associations, and no links to autochthonous organizations-this association will have a 0.01 probability of being invited to participate in the policy process.If this organization becomes "resourceful" in human capital and employs a full-time member of staff and has the membership size that marks the lower bound of the last quartile (317 members), while still having no contacts whatsoever with other migrants' and autochthonous organizations, then the corresponding probability increases to 0.47.If, instead, the organization remains "resourceless" in human capital but "resourceful" in social capital-with a number of contacts with other migrants' and autochthonous organizations that places them in the higher quartile: 13 and 15 respectively-the chances of policy inclusion only rise to 0.35.Consequently, migrants' organizations that are in the higher end of human capital resources but fail to develop a large contact networks are more likely to be included in policy making than those that develop a large network of contacts but have limited human resources.Of course, the ideal situation is to have both, as that increases the probability of inclusion to near certainty (0.97).
In summary, our results show that the social capital migrants' organizations forge with other migrants' associations and with the receiving society actors has a crucial impact on their capacity to become engaged in policy making processes.Yet, social capital is not the most valuable resource for migrants' organizations in Barcelona and Madrid to have a say in the host country's public arena-human resources are.While financial and material capital are of very little importance for getting access to policy making, the size of an organization's human base and the distance of an organization's reach to other civil society organizations are the most significant determining factors for having a say in the policy process.

DISCUSSION
Migrants' organizations in Barcelona and Madrid are only moderately included in the policy process.Yet connected organizations benefit to a great extent from their social capital: the interconnections with other migrants' and autochthonous Spanish organizations.Hence, migrants' associations that are centrally located in the broader network of all migrants' organizations in the city, and that are well connected to autochthonous civil society organizations, are particularly advantaged for policy action and impact.
These results are especially interesting because, unlike what is expected from resource mobilization and political economy theories of organizations, older and materially resourceful associations are not particularly privileged in their access to the policy process in these two cities.Our evidence indicates that, rather than requiring the mobilization of vast economic resources, gaining a voice in the decision making process in these Spanish cities requires being well-connected and having one person dedicated full-time.
How can we explain these results?Massive immigration from other countries is only a recent phenomenon in these two cities and more generally in Spain.The political elites have only lately begun designing clear policy approaches for the management of immigration and migration related social problems.Immigration policies are, thus, in the making in these cities and in Spain as a whole, and there is a lack of clarity about how these policies should be designed, and what actors need to be included in decision-making processes.Because most immigrants are not granted the right to vote in any elections-unless they are also Spanish nationals-the mechanisms of representation rely on migrants' organizations as the only intermediaries between the host political elites and the immigrants.Yet, the local elites and the public servants have a limited knowledge of the field of migrants' organizations because these are usually small, quite informal, and frequently invisible.
Consultation mechanisms in Spain often operate without clear rules and procedures about who is entitled to participat and who is not.Often, decisions about which organizations will be consulted are made on an ad-hoc basis; and when rules exist-as for some of the more established immigration forums at the national and regional levels-the government has absolute discretion in the nominating process.In this context of loosely structured processes of consultation, and in a situation of limited and imperfect information about the immigrant associational field, political elites need to rely on the available proxies to make judgments about which organizations are relevant and worthy of being consulted when designing immigration policies.Migrants' organizations who are centrally located in the network of ties among all migrants' associations are the most likely to be visible to policy makers; and so are the migrants' organizations that are better connected to autochthonous organizations, and especially to those that have traditionally been more active in advocacy and service provision on immigration issues. 21 In the absence of clear and formal procedures for public consultation on immigration policy issues, strategic networking and dedicated staff provide visibility and secure a voice in the policy process.Moreover, because the immigration policy field is in the making, migrants' organizations do not need a lot of financial resources to be granted a place in decision-making processes.What they need is to be visible and prestigious.This is clearly not the case where policy making is subject to formal lobbying rules and procedures-as at the European Union level, or at the federal level in the United States-and getting one's voice heard requires mobilizing a substantial amount of economic and human resources.In other contexts, in which coordinating and consociational practices are in place, the existence of clear rules for participating in the consultation process will require-above all-the mobilization of a sizeable constituency, or the capacity to form umbrella organizations (see Hooghe 2005 on Belgian practices).
Our results suggest that the importance of the various forms of organizational resources and capital of migrants' organizations will depend on the political context in which those associations operate.In contexts where the policy process is well structured and consultation mechanisms and practices are consolidated and formalized, migrants' organizations are likely to need a substantial amount of economic and human capital to have their voices heard.In contexts where the policy field is still in the making and consultation is unstructured and follows relatively arbitrary processes, who you know and who knows you will play a vital role.We expect that the latter situation will not be exclusive to the Spanish case, but is likely to be applicable to other European countries like Italy and Portugal that-as Spain-combine a situation of recent massive migration inflows with a tradition of loosely structured policy consultation processes.Future research that compares different contexts of policymaking and consultation procedures, or that analyzes longitudinal changes in the policy field, should shed additional light on this topic.

NOTES
3 Visa policies drastically changed during the 2000s due to the massive inflow of Latin American immigrants since the late 1990s. 4There is a constitutional impediment to granting voting rights at the local level to non-nationals in the absence of bilateral international agreements for the mutual recognition of these rights.Until 2011, only EU citizens and Norwegians effectively had such rights.However, the Socialist government of José Luis Rodríguez Zapatero has completed the legal procedures that have enfranchised a substantial number of, mostly Latin American, immigrants to vote in the local elections starting with those of May 2011.Reciprocity agreements have been signed with-among others-Bolivia, Chile, Colombia, Ecuador, Peru, and Paraguay. 5See Siaroff (1999) for a comparative overview of interest representation systems in OECD countries, where Spain scores lowest-together with Portugal-in what regards both corporatist and consensual/concertation arrangements of interest intermediation.Lane and Ersson (1999: 235) also score Spain with the lowest value on a similar index of corporatism. 6However, see van Heelsum (2005) for a discussion of the generalizability of Fennema and Tillie's results to other cities in The Netherlands. 7See Baldassarri and Diani (2007) for an interesting extension of this distinction to the nature of the relationship that is exchanged: transactions versus social bonds. 8Although, strictly speaking, they are also migrants, we will not consider in our study foreigners born in the EU-15 countries and other advanced industrial democracies like the US, Canada, or Japan. 9The sources included all available official registers at the local, regional, and national level, a number of civil society directories, information provided by the embassies and consulates, systematic internet searches, lists provided by strategic informants (such as large pro-migrant NGOs, and trade unions), and snowballing from the first contacts with listed organizations and from completed questionnaires. 10A pilot study in Madrid conducted in 2003-2004 rendered 33 additional questionnaires.However, given that the measurement of interorganizational network ties changed substantially between the pilot questionnaire and the 2007-2008 questionnaire, and because the opportunities for associations to engage in policy making mechanisms at the local level had changed somewhat (especially in Madrid), we have chosen not to include the information on those organizations in the analyses presented in this article. 11This response rate is acceptable when compared to a similar (but postal) survey carried out in 2003 to autochthonous associations in two districts each in Barcelona and Madrid, which only obtained a response rate of between 12 and 21 percent-depending on the district-and for which the response rates were also substantially higher in Barcelona.That study is the only comprehensive one of civil society organizations in Spain and was part of the Citizenship, Involvement, and Democracy project in Spain.For a description of the methodology of that survey see Font, Montero, and Torcal (2006: Annex 2). 12Of the remaining associations, only eleven organizations that were not interviewed were frequently mentioned-by at least five interviewed organizations.The rest were never mentioned or were mentioned by only one or two organizations.We made every possible effort to get an interview with these organizations, but were not able to achieve it due to failed appointments or refusals to collaborate with the study. 13This latter figure is the average of the estimation provided by the study to local associations in the four districts of Barcelona and Madrid previously mentioned. 14The table shows the distribution by world region for all the associations listed as "active" in the listing stage of fieldwork.Thus, these results reflect the general outlook of the whole target population of migrants' associations. 15Contrary to what one could expect, migrants' associations that are more active politically or employ lobbying as one of their main repertoire actions are not substantially more likely to be invited to join policy making mechanisms.In fact, the correlation between receiving an invitation and being active in politics is 0.09 and with lobbying it is 0.12, none of them statistically significant. 16The list of Spanish organizations included some 75 names, and the list of migrants' organizations included 280 in Barcelona and 210 in Madrid. 17These two exclusions respond to the fact that the coefficients for these two variables are clearly not significant from the first step of their inclusion in the nested models and to the variable on budgetary capacity being missing in 12 cases.The exclusion of these two variables does not substantially change the results when models 4 and 5 are compared but, as it increases the number of cases, it results in some coefficients becoming statistically significant. 18This is shown in the fact that this variable loses statistical significance in models 4 and 5 when the variable that measures contacts with autochthonous organizations is included. 19We report on the results for the separate models on Barcelona and Madrid, though they are not shown here due to space constraints. 20All the estimations of probabilities have been done using a baseline "typical" organization that is 6 years old, is formed by Latin American immigrants, is not active in politics, is based in Madrid, has no full-time members of staff, has 100 members, is named by 6 other migrant organizations, and names 7 autochthonous organizations.Thus, from the "typical" organization that adopts the median values in all variables, probability changes are computed by changing only the value of the variable of interest and keeping the rest unaltered. 21See Caponio (2005) for a description of the Italian case, which is in many respects similar to the Spanish situation.
Contacts w/ migrant orgs (indegree)number of contacts from 0 to 40 ones

Figure 2 .Figure 1 .
Figure 2. Comparing the Independent Effect of Two Elements of Social Capital on Invitations to Participate

Table 1 .
Summary of the Fieldwork Process and the Migrant Population Source: Census list created during fieldwork with Immigrants' Associations,Localmultidem and Capsocinmig  projects (2007-2008), and local population register (Padrón, as of January 1, 2007).Notes: 1 Not active over total confirmed. 2terviews over total confirmed active.3Eligibleassociations over total foreign-born population.

Table 2 .
Associations by Main World Regions of Origin of the Constituency (percentages) Source: Census list created during fieldwork with Immigrants' Associations, Localmultidem and Capsocinmig projects(2007)(2008), and local population register (Padrón, as of January 1, 2007).

Table 5 .
Regressions of Invitation to Policy-Making Processes on Various Measures of Organizational Resources (logistic regressions)