Foundations as “Amigos o Rebeldes”: The Influence of Philanthropic Funding on Local Immigration Policy Outputs

Abstract Research on foundations public policy influence, traditionally, focuses on policy reform. Largely unexplored is the influence that philanthropic funding has on recipient communities. Unlike previous research, this study uses a newly compiled dataset on immigration-related giving to study how funding for political citizenship services, integration services, and government-related advocacy influences the identification of deportable immigrants across the continental United States. The quantitative analysis indicates foundations exert indirect influence within local immigration policy outputs through the use of targeted philanthropic grants. However, the effect of the indirect influence depends upon the policy activities receiving funding. Philanthropic foundations’ providing funding for political citizenship and integration services lead to decreases in immigration enforcement. While funding for government-related advocacy can help increase immigration enforcement, it can also help address issues of equity in immigration enforcement. Interviews with foundation grantees provide further insights into how the funding of these policy activities can alter the relationship between the philanthropic community and the government agencies implementing U.S. immigration policy.

immigration status, and national origin" (GCIR 2017). As overt discriminatory nationalism increased, the statement stood as a signal that these American foundations were taking a political stance against the new presidential administration's immigration policy, rhetoric, and enforcement practices. Were these foundations' words an empty promise? Can foundations' resources influence immigration policy within the U.S.?
With the use of Secure Communities, the Obama administration ushered in a heightened time of immigration enforcement that led to the deportation of over one million non-citizens. During this time, American foundations awarded about 1.85 billion dollars to nonprofit organizations across the U.S. for immigrants or immigration related services. These grants support a range of "policy activities" (Tompkins-Stange 2016) such as health and human services, citizenship courses, legal fees for representation to educational rights workshops. The description of these grants went beyond the traditional focus of national policy reform. Leading to the question, what influence does philanthropic funding have on local immigration policy outputs? Furthermore, as the U.S. federal government provides limited to no funding for immigrants (De Graauw and Bloemraad 2017), might the effect of philanthropic funding alter the relationship between the philanthropic sector and government?
Research on philanthropic foundations provides in-depth case studies detailing the roles and strategies used to achieve policy reform (Anheier and Daly 2006;Anheier and Hammack 2010;Dowie 2002;Fleishman 2007;Nielsen 2001;Tompkins-Stange 2016). This line of literature leaves unaddressed how funding influences local policy outputs within the communities receiving foundations' grant funding. This research contributes to the literature by examining how grant dollars influence the local public policy process. Another contribution of this research is the testing of the theoretical arguments within the punitive context of U.S. immigration policy using a newly collected dataset on foundation grantmaking related to immigrant services paired with Secure Communities immigration enforcement outputs, i.e. total identified deportable immigrants.
The long-standing and anecdotal argument about the (indirect) influence of philanthropic funding on public policy finds support in this research. Philanthropic grantmaking for immigration-related policy activities has an indirect effect on local immigration enforcement efforts. However, funding's effect depends upon which policy activities receive grant dollars. To further explore how funding affects policy, the next section provides a brief overview of previous studies focusing on the role of foundations within public policy. Following the literature review, the theoretical arguments about the role that funding for immigration-related policy activities has on local immigration enforcement efforts is presented. Next, the data and research methods section provides details on the dataset created for this quantitative analysis. To conclude, I provide a discussion using interviews with nonprofit grantees on how their work within each of the immigration-related policy activities relates to influencing immigration policy outputs and philanthropy's relationship with the federal government.

Literature Review
As autonomous organizations free from constituencies and government, foundations and their "philanthropists have not typically engaged directly in politics without the cushion of intermediaries" (Reckhow 2012, 13). This independence allows foundations the freedom to extend funding to nonprofits for issue advocacy or public service provision in intentional and instrumental ways aligning with their missions and values Frumkin 2008;Nielsen 2001). However, foundations independence and mostly unregulated autonomy within the American socio-political system has led to criticism about their role in creating an undemocratic society further entrenched by capitalistic elitism without true regard for the preferences and interests of the public. Their role in society helps silence pluralism and activism through the funding of programs and services even from liberal progressive foundations (Anheier and Daly 2006;Domhoff 2009;Dye 2000;Reich 2018;Roelofs 2003;Tompkins-Stange 2016).
Scholars have attempted to classify foundations' roles into strategic patterns and actions by examining foundations' influence as intermediary organizations (Scott and Jabbar 2014), institutional entrepreneurs (Quinn, Tompkins-Stange, and Meyerson 2014) and responsive actors (Mosley and Galaskiewicz 2015). Foundations act as funders or institutional entrepreneurs to promote innovation and research for market interventions, think tanks, policymakers, and the broader public (Clemens and Lee 2010;Scott and Jabbar 2014). Weissert and Knott (1995) find that the research and process knowledge generated from foundation grantees is useful in the public policy process.
The quantitative research on foundations centers on studying what influences grantmaking patterns. Of the studies on grantmaking, Reckhow (2012) is one of the only authors to focus on following the grant money. After finding that foundations are more likely to provide grants to organizations with centralized organizational structures, she examines how these grants shape the subsequent educational policy reform. Her analysis indicates that for foundations to have longer-lasting victories in policy reform, the type of recipient organization plays a large role in their influence. In funding organizations with centralized power structures such as school districts, foundations' influence was short-lived and unsupported by the community. By funding advocacy organizations and nonprofits whose power structures look different from a bureaucratic school district, policy changes, while slower, had time to build support and legitimacy to provide a long-term policy reform victory. Her research illustrates that philanthropic foundations can influence policy reform, but which organizations receive funding matters in influencing policy reform. Building off Reckhow (2012), this study seeks to delve further into how the funding of policy activities influences local policy efforts within immigration policy.

Theory
Philanthropic grantmaking for policy activities can directly or indirectly influence the public policy process (Fleishman 2007;Frumkin 2008;Sandfort 2008;Tompkins-Stange 2016). In providing funding to intermediary organizations engaging in the provision of these policy activities, foundations impact policy outputs and outcomes. The following theoretical arguments apply this view of grantmaking influence to immigration-related giving by examining how funding for three major immigration policy activities affects local immigration enforcement efforts, i.e., the identification of deportable immigrants. The arguments conclude with a discussion of the proposed effect that the outcome will have on the relationship between the philanthropic sector and government.
Nonprofit organizations engage in policy activities that can affect non-citizens' level of deportability. Immigrants' political citizenship status, level of integration, and the receptiveness of their community affects their deportability (De Genova 2002). The funding of policy activities within each of these areas affects the identification and subsequent deportation of immigrants. In funding these policy activities, the philanthropic community can alter their relationship with the government agencies implementing immigration enforcement efforts. For example, the previously referenced GCIR statement alludes to a developing adversarial relationship between these foundations' upcoming grantmaking patterns and the presidential administration. Table 1 summarizes the theoretical arguments about the influence of funding for political citizenship services, integration services, and government-related advocacy.
Foundation grants for citizenship services to the immigrant community fund a range of policy activities from political status adjustments to universal representation. Grant funding for political citizenship services leads to changes in immigrants' deportability level by adjusting their formal political status. Thus, status adjustments and representation during immigration court proceedings culminate in decreasing the size of the deportable immigrant population. Increasing grants targeted for policy activities to facilitate political citizenship services will be related to a decrease in the identification of deportable immigrants. As these policy activities work to ensure that individual non-citizens cannot or are not deported, these strategies alter their relationship with the punitive and social immigration agencies implementing immigration policy. By decreasing the size of the deportable immigrant population, the philanthropic community is creating an adversarial relationship with the government due to their resistance in helping facilitate the deportation of immigrants. However, their work on status adjustments helps reduce the compliance and learning administrative costs (Herd and Moynihan 2019) for both immigrants and the bureaucrats of U.S. Citizenship and Immigration Services. In helping alleviate the administrative burden of government, the philanthropic community works as a supplementary community partner (Najam 2000;Young 2000) to increase the number of processed immigration visas and citizenship applications.
With no bureaucratic organization focused on helping immigrants integrate into life within the U.S., nonprofits have stepped in to fill the void with financial help from foundations (Clemens 2006;De Graauw and Bloemraad 2017;Douglas 2001;Steuerle and Hodgkinson 2006). Foundations aid in legitimizing the funding of nonprofit services for all immigrants regardless of formal political status. This funding supplements government grants and contracts, which can have strict criteria for when and what clients receive services. With an unmet need for services, foundation grantmaking supports nonprofits in creating environments and relationships where the immigrant community is active and unafraid to seek out services compared to an environment where immigrants' fear of deportations isolates them from reaching out to the local government for aid (Cordero-Guzman 2005; De Graauw 2016; Frasure and Jones-Correa 2010; Leyro and Stageman 2018;Nichols et al. 2018;Watson 2014). Furthermore, integration policy activities fund a variety of services from language training to health care access (Andersen 2010;Bloemraad 2006;Cordero-Guzmán et al. 2008;De Graauw 2008De Leon et al. 2009;Gleeson 2015;Jones-Correa 2008a, 2008bModarres and Kitson 2008;Wilson 2013) that help ensure immigrants meet the requirement of having "good moral character" and "not likely to become a public charge." This multi-dimensional service approach focuses on providing holistic changes to an immigrant's deportability from the micro-level to the macro-level. Therefore, increasing the funding for integration services will be related to decreasing the identification of deportable immigrants. If this occurs, funding will lead to a similar multi-dimensional relationship between the philanthropic community and the agencies implementing U.S. immigration policy. Depending on which policy goal is of focus, deportations, or visa processing, the philanthropic community's funding creates a government relationship that can fluctuate between adversarial and supplementary.
Government-related advocacy is the last instrumental pattern of giving. This funding category captures policy activities related to advocacy efforts by nonprofit organizations. Funding for government-related activities is different from grants made to integration and political citizenship services because funding in those areas supports front-line programs and services for the immigrant community. It is unlikely that a foundation with an anti-immigrant stance would give funds to nonprofits serving the immigrant community. With funding for government-related advocacy, "there are large foundations which overtly promote right-wing activities … or which devote their resources to … the promotion of 'conservative values'" (Horowitz and Horowitz 1970, 227) which can lead to restrictive and punitive immigration policy. As Brown (2013) points out, the Tea Party's growth in 2008 spurred conservative foundations grantmaking to fund nonprofits focusing on anti-immigrant rhetoric that advocates for policies aimed at making immigrants' ability to reside in the U.S. difficult to stricter border enforcement. However, as this instrumental category is a catch-all for advocacy, increased funding will be related to changes in immigration policy outputs. However, there is no direction hypothesized as grants can be given to either maintain the status quo of current laws, advocate for policy reform, or simultaneous funding of both conservative think tanks or progressive immigrantserving nonprofits. Unlike the previous funding categories, the effect of funding related to government advocacy can lead to either an adversarial or cooperative relationship between the philanthropic community and government. If the influence of advocacy funding leads to a decrease in the identification of deportable immigrants, this is reflective of an adversarial relationship with the government regarding immigration policy. However, if the influence of advocacy funding has a positive effect on the identification of deportable immigrants, the possibility of a cooperative or co-opted relationship could occur as nonprofits' research could produce policies that increase the deportability of immigrants, even if not intentional.

Data & Research Methods
This research tests the theoretical link between philanthropic funding and public policy by focusing on the influence grant dollars have on local enforcement outputs. To test this indirect link, a newly created dataset spanning Secure Communities' tenure under the Obama administration is matched with grant data collected by the Foundation Center (2016) for all the counties within the 48 continental states and the District of Columbia. In addition to the quantitative analysis, additional insight into the way funding translates to front-line services and programs comes from supplementary interviews with foundation grantees.

Influence of Philanthropic Funding
The dataset for the quantitative analysis uses four primary data sources. The first data source provides the immigration policy output measures. These immigration policy measures come from interoperability reports released by the Department of Homeland Security's Office of Immigration and Customs Enforcement (IDENT 2014). A detailed discussion of the operationalization of the two measures of immigration policy continues below. The second data source is the Foundation Center, which provides immigration-related grant data from 2008 to 2014. To ensure the results capture the influence of philanthropic funding, a set of controls for factors that influence immigration enforcement comes from the American Community Survey This quantitative analysis uses a seemingly unrelated regression analysis 1 at the county level. The immigration policy outputs are mutually exclusive as an immigrant cannot be both a high priority deportable immigrant and a low priority deportable immigrant. Unlike a traditional ordinary least squares regression, 2 a seemingly unrelated regression analysis can account for the interdependent nature of the two policy outputs that can lead to the possibility of correlation among the error terms (Moon and Perron 2006;Zellner 1962). The sections that follow will detail the operationalization of each immigration policy output, instrumental grant category, and control variables. Furthermore, a discussion of the grantee interviews provides information on the selection of counties and organizations invited.

Immigration Policy Outputs
There are two immigration policy outputs of interest. The goal of Secure Communities was to locate deportable immigrants from the interior of the United States. Secure Communities was touted as a program centered on priority level enforcement, its implementing agencies were to focus their efforts on high-priority deportable immigrants. The program notes that high priority deportable immigrants pose a threat to the safety of the United States by 1 Models with state fixed effects and year fixed effects are not reported but were run. The relationships between philanthropic funding and policy outcomes remain unchanged when state fixed effects or year fixed effects are included. 2 As another robustness check, a traditional OLS was run for each of the two models included in the SUR results shown. The results hold, and the R-square remains the same for both individual models. The only changes are in the size of the beta coefficient, the coefficients of interest increase in the OLS models between 0.0005 and 0.001. Additionally, there are no changes in the direction of the effect. having a felony or at least three misdemeanor convictions on their record. Less of a focus was their second-order priority enforcement goal, immigrants deportable due to unlawful entry into the United States or with up to two misdemeanor convictions on their record. With the priority levels in mind, the first immigration measure focuses on capturing a local law enforcement agency's strength to act as a force multiplier for the federal government. The force multiplier is calculated using the total number of matched deportable immigrants divided by the total noncitizen 3 population. This measure captures how many immigrants per capita are identified as deportable by the local sheriff's department. The larger this measure, the more strongly a county acts as a force multiplier. Due to the overdispersion of zeros, this measure is logged to help normalize the distribution for analysis. On average, a county had a noncitizen per capita identification of 0.07 with a 0.40 standard deviation. Table 2 contains the summary statistics of each policy output.
Despite having deportation priorities, Secure Communities disproportionately identified higher rates of immigrants with only misdemeanor convictions or a civil violation of unlawful presence on their records rather than identifying and deporting individuals with felony criminal records (TRAC 2014). An enforcement disparity 4 measure captures the ratio of low priority to high priority deportable immigrants identified as a means of identifying the philanthropy's influence in creating policy equity. In the sample data, the average disparity in enforcement is about two low priority immigrants to every one high priority immigrant with a standard deviation of about three low priority immigrants to every one high priority deportable immigrant.

Immigration-Related Philanthropic Funding
The Foundation Center catalogs grants from the largest foundations in the United States, making this the best data source for U.S. grant data. 5 Individual grants must meet specific parameters within the nonprofit population field and the grant activity fields for inclusion in the dataset. These two fields must contain either the words "immigrants," "immigrants and migrants," "undocumented immigrants," or "immigration." The Foundation Center provided about 27,000 grants related to immigration from 2008 to 2014. Of these grants, close to 26,000 grants were provided to domestic nonprofits focusing on immigration-related services, integration, or advocacy. The following discussion on grants provides detailed information on each of the funding categories operationalization and a brief overview of foundations' funding pattern for immigration-related giving within the United States. As foundations leverage their private resources to help fund nonprofits promoting their private values, immigration was not a top funding policy priority of the philanthropic community. During the Secure Communities (2008)(2009)(2010)(2011)(2012)(2013)(2014), there was an average of 81,000 grant-making foundations in the United States, which gifted a little over 378 billion dollars across nonprofits for a range of policy issues. From this $378 billion, 6 only 1.85 billion in philanthropic grants provided aid for immigrationrelated services within the 48 continental United States and the District of Columbia. Figure 1 shows the total aggregate funding distributed across U.S. counties from 2008 to 2014. The average county received about $84,000 in total for immigration-related giving, with a standard deviation of about $1.3 million. This figure illustrates a pattern of unequal funding distribution where many communities with immigrants received even less per immigrant if any funding at all.
The Foundation Center dataset provides specific details about each immigration-related grant made by U.S. foundations. Among the grant details is information on the foundation, the nonprofit grantee, and specific details about the reason for the grant awarded. Of specific interest was the location of the grantee, the populations served by the nonprofit, the amount of the grant, the detailed grant description, and the overriding grant activity funded. Using the grant information, each grant amount is allocated into one of the three instrumental categories that best matches the overriding grant activity. 7 The grants are 6 All dollar amounts have been adjusted for inflation and represent 2016 constant dollars. 7 Each grant received up to five activity codes illustrating the various programs or services the grant will help fund. However, operationalization was based on the separate variable provided by the foundation center that identified the overriding grant activity providing the top description for the basis of the grant.

Influence of Philanthropic Funding
not mutually exclusive within the funding mechanisms of interest. However, the reason to use the overriding activity is that this data column captures the primary focus of the funding intention. 8 The full grant description provides insights into all the goals the grant intends to accomplish, whereas, the overriding activity selects the primary policy activity targeted by the funds. The following instrumental funding categories detail the coding of individual grants: Political Citizenship Services represents grants with an overriding grant description that includes any of the following terms: "Immigrant Rights", "Immigrant Services", "Immigration & naturalization", "Immigration Law." 9 There were 7615 individual grants that were awarded under theses overriding immigration policy related activities. For example, these grants had grant descriptions that included, "For the Worcester Refugee and Immigrant Assistance Center to provide legal representation to LGBT asylum seekers and educate the community about international human rights conditions", "Increase family stability for mixed status immigrant children by increasing the number of families who can start down the path to citizenship", and "To provide legal, health, education and emergency services to Bay Area immigrant low-wage workers, day laborers, and families." Not all grants had specific descriptions or any description at all. Some grants were awarded to help with general operating or unrestricted support.
Integration Services grant category is composed of over 13,500 individual grants with overriding grant activities such as "Education", "Health", "Human Services", "Community Development", "Community Food System", and "Mental Health" etc. 10 Integration services is the largest category of individual grants. These grants fund services and projects such as but not limited to, "Work on policy, research and convening including ELL populations, dropout rates and opportunities to learn math and science," "For Latina Health Initiative, providing health education, gynecology, and mental health care to Spanish-speaking women in Chicago," and "To increase Muslim community participation in the city of Fremont's Make a Difference Day, as a means to bring together Muslim and non-Muslim residents to jointly address community issues and needs through community service." Government Related Advocacy is comprised of individual grants with overriding activities such as but not limited to "Freedoms", "Minority Rights", "Community Policing", "National Security", "Police Agencies," etc. 11 This category is the smallest of the three in terms of number of grants with 4,666 grants. This category represents policy activities related to advocacy work that helps ensure the protections of individuals to communities. For example, full grant descriptions include funding, "To promote positive images of new immigrants in ethnic and mainstream media," "For the national security and human rights program," "To support the public safety and law enforcement program." All individual grants receive an instrumental category code using their overriding grant activity. Before aggregating all the grants in each category up to the county level, the individual grant amounts are adjusted to 2016 constant dollars. Each county receives a value for all three instrumental categories, but if the county received no philanthropic funding, that value is zero. In order to normalize the distribution, one is added to each grant amount and logged. Based on yearly giving patterns, foundations averaged larger grant funding allocations for integration services for the immigrant community over citizenship services or government advocacy. The average county received about $39,500 for integration services with a standard deviation of under $500,000. In comparison, the average country receives about $24,000 for political citizenship services with a standard deviation of less than $480,000. Government advocacy had the lowest average funding of about $20,000, with a standard deviation of about $411,000.

Policy Environment
The implementation and enforcement efforts of U.S. immigration policy occur through bureaucratic agencies. In this analysis, the bureaucratic agency is local law enforcement, the county sheriff's department. So, when evaluating immigration policy outputs, it is necessary to control for local law enforcement's capacity for implementation and the influences within the policy environment that can affect the identification of deportable immigrants (Brewer and Selden 2000;Boyne and Meier 2009). First, to capture local law enforcement's capacity to identify deportable immigrants, it is important to understand its resource capacity. In the analysis, the research model includes a variable capturing the size of the total police protection force. This measure includes the total employment of officers within the county and city law enforcement officers. 12 The second measure is the aggregated total operating budget of each county's local law enforcement agencies. Jaeger (2016) and Farris and Holman (2017) identified that local sheriff's budgets are related to their willingness to comply with Immigration and Customs Enforcement. These two organizational capacity measures are logged to normalize their distribution as the sizes of counties vary across the continental U.S. The second component of organizational capacity is the inputs that an organization can process. In this area, organizational inputs are the total number of arrests that occur within the county. The total submissions of fingerprints within a county reported by the Interoperability Reports capture the total number of inputs into the local county jail. Secure Communities' activation occurred through a rollout activation pattern where counties across the U.S. activated in waves from late 2008 through the end of 2012. This activation pattern causes an overdispersion of structural zeros, 13 Before logging total submission, a one is added to each county's submission total. This operationalization of the total submissions helps normalize the distribution of inputs. In 2011, Director of Immigration and Customs Enforcement, John Morton, issued a memo implementing prosecutorial discretion. The memo gave the directive to agents to exercise more discretion by prioritizing deportable immigrants who pose a real danger. The directives in the Morton memo focus on calling for enforcement agencies to redirect their efforts on the task of priority level enforcement. This call to refocus efforts should lead to a decrease in the enforcement disparity between high priority deportable immigrants and low priority deportable immigrants. As such, a dummy variable to capture this policy shift's implementation is coded one for the years 2012 through 2014.
In addition to the influence of local law enforcement's capacity in immigration enforcement, the policy environment within which agencies operate affect their implementation and enforcement efforts. To capture the influence of the policy environment, a set of controls capturing the social, political, and economic climate are in each model. The first set of controls accounts for the policy targets and non-policy targets of the community. According to past research (Cox and Miles 2013;Wong 2012), the activation 12 The local law enforcement presence, unfortunately, is not reported at each county level. Rather, the Annual State and Local Government Employment Survey reports aggregates to the state level of all local law enforcement agency's budgets and personnel. 13 Structural zeros represent counties who have not yet activated. Activation began late in 2008 with full implementation across all U.S. counties and territories as of 2013. Six counties did not experience activation because they were grouped with nearby counties. and enforcement of Secure Communities occurred in counties with high levels of specific populations, Hispanics, and noncitizens. Compared to the contemporary immigration narratives which push Hispanics as the modernday face of immigration, historical immigration narratives focused on the Asian population within the United States as the face of immigration (Daniels 2005). In addition to Hispanic and Asian populations, the model includes the population percentage of Whites and African Americans. The final social measure captures the level of education with a county. As the percentage of individuals with less than a high school education increases, so will the pressure for immigration enforcement.
Immigration policy is a politically polarizing issue with both sides of the aisle arguing for different enforcement policies and reform typically fueled by rhetoric concerning economic distress caused by immigration. During times of crisis, anti-immigrant rhetoric spikes with the Republican party calling for stricter immigration policies (Calavita 1996(Calavita , 2010 The model includes variables to account for the economic conditions that could lead to stricter immigration enforcement. The first measure captures the percentage of non-homeowners representing a lack of economic vitality in the community. As unemployment rates rise, calls for increased immigration enforcement increase. To capture the influence of unemployment on immigration enforcement, the model contains the county-level unemployment rate. The third measure captures the education level within the county. This measure of education captures the percentage of individuals with up to a high school degree. The economic variables focus on controlling for the narrative of immigrants as public charges (Moloney 2012). As the percentage of the population on public assistance rises, so will the immigration enforcement in a community. The final control measure captures whether the county lies along the U.S.-Mexico Border. These counties not only have local law enforcement engaging in immigration enforcement efforts, but also have the presence of U.S. Customs and Border Patrol within 14 Voting percentages at the county level were collected from the Data.gov website managed by the U.S. General Administration, Technology Transformation Service. The website follows the Project Open Data schema that requires all datasets to report: title, description, tags, last update, publisher and contact name.
Influence of Philanthropic Funding their local communities. A border county likely has higher enforcement than counties located away from the U.S. Mexico border.

Grantee Interviews
The quantitative analysis helps establish the connection between philanthropic funding and public policy. However, it cannot explain how these dollars become the mechanisms that alter immigrants' deportability across our communities. To further shed light on how funding translates into policy activities, foundation grantees were interviewed about the philanthropic grants received and the services the grants funded. The interview data assists in shedding light on the mechanism gap established by the theoretical arguments. Although the interviews with Executive Directors, Directors of Programs, or Directors of Legal Services took place years after the grant funding was received and used to fund services, nonprofit service delivery is slow to adapt to changes implemented to increase their effectiveness. As such, it is unlikely that immigrant-serving nonprofits introduced radical changes to their service delivery between receiving the grant and the interview. The interviews are analyzed using a theoretical thematic analysis (Braun and Clarke 2006) linking the themes through direct manifestation or through an interpretative connection that associates the work by nonprofits to changes in immigrants' deportability 15 (Saldaña 2015).
The interview data includes 43 semi-structured interviews. These interviews took place within three states and the District of Columbia, spanning 12 counties. The counties where the grantees are located have foreign-born populations of more than 17% and high philanthropic funding, more than the average of $4 million. 16 The only location not fitting this description is the District of Columbia, which is included due to it being the nation's capital, the hub for advocacy, and the headquarters for many regionally located immigrant-serving nonprofits. An email invitation with a phone call follow-up was sent nonprofits meeting specific criteria within the grant dataset. The interview inclusion criteria for organizations in the grant data is that the organization must identify "immigrants" as one of the top three populations served or are the organization is a recipient of a grant with an 15 A limitation of this research is that only nonprofits with a pro-immigration mission were interviewed. 16 This average was calculated by only including the counties that received funding from 2008 to 2014.
"immigration services" overriding grant activity. When responses were low for the nonprofit grantees found in the dataset for the counties of interest, 17 a search on GuideStar identified additional immigrant-serving nonprofits to recruit for interviews. The search parameters for nonprofits in GuideStar mirrored the population parameters used for the grant data. Of the nonprofits interviewed, 71% of the interviewees identified in the dataset are immigrant-serving nonprofits. Of the interviewees, 81% are philanthropic grantees between 2008 and 2014. Additionally, of the interviewees 37% did not receive grants for political citizenship services, 20% did not receive grants for integration services, and 42% did not receive grants for government-related advocacy.
The 43 interviews conducted are not meant to be a nationally representative sample. However, this research works with the assumption that the interviewed immigrant-serving nonprofits share similarities with broader immigrant-serving nonprofits across the U.S., especially in traditional-destination states. Potential interviewees who were grantees were identified from the Foundation Center data. Non-grant recipient interviewees were selected from GuideStar. The selection from GuideStar deviates slightly from previous research on immigrant or migrant-serving nonprofits. Previous research selected cases by focusing on organization specific traits such as the ethnic composition of leadership, the name of the organization (ethnic-sounding), or the percentage of immigrant/ migrant clientele (Cordero-Guzman 2005;Cortés 1998;De Graauw 2016;Hung 2007;Martin 2012). In most cases, this research's coding scheme overlaps with previous research, especially for immigrant-serving nonprofits. Immigrantserving nonprofits prioritize the immigrant community as their main or only clientele by listing immigrants or immigrants and migrants as the top served population and, usually, have leadership with ethnic or immigrant identity in positions of leadership. Except for one organization, 18 all organizations have an active 501(c)(3) status. 19 Lastly, unlike the previous research, which can limit inclusion to nonprofits with total annual revenue of $25,000, this research did not begin with this limitation when selecting nonprofits for participation. This limitation is an obstacle to getting the voices of grassroots nonprofits for immigrants. Grassroots organizations may not always meet the total revenue requirement, especially in low-income communities (Gleeson and Bloemraad 2013;Roth and Allard 2016;Roth et al. 2015).

Results
This research centers on testing the indirect link of philanthropy on public policy within a new policy area, immigration. For the analysis, grant amounts are aggregated into one of three categories that address different aspects of an immigrant's deportability and identification for deportation: political citizenship, integration, and advocacy. Based on the analysis of the three different patterns of  immigration-related funding, the results provide support for the hypothesis that philanthropic funding is indirectly related to policy outputs. Additionally, the results support the argument that the effects on policy outputs depend on the type of policy activities funded. Table 3 contains the results of the effect of philanthropic funding on local immigration enforcement. The analysis illustrates support for the theoretical argument of philanthropy's indirect influence on local immigration policy outputs. There is a statistically significant negative relationship between citizenship funding and the two immigration enforcement policy measures, total immigration enforcement per capita, and reducing disparity issues in immigration enforcement at the local level. A one percent increase in funding for political citizenship negatively influences immigration enforcement per capita by 0.03%. Furthermore, funding decreases the disparity in identifying more low priority deportable immigrants to high priority deportable immigrants by 0.01%, all else being equal. For example, if the average foundation increased their citizenship funding by 100% above the mean of roughly $24,000 to about $48,000, the average effect within a county would be to decrease the identification rate by 3% and the disparity in enforcement by 1%, all else being equal. This effect means that the average increase of about $24,000 could lead to protecting between 2 and 3 non-citizens and about one low priority immigrant less than the high priority immigrants identified. To understand the magnitude at a larger scale, the Vera Institute of Justice helped launch the New York Immigrant Family Unity Project funded by the New York City Council in 2013. With a budget of about 10 million dollars, by 2016, the representation project has helped almost 1800 non-citizens during its tenure (Stave et al. 2017). Philanthropic funding provides representation during immigration court proceedings. These services and programs available for immigrants create an adversarial relationship between the philanthropic community and Immigration and Customs Enforcement and local law enforcement agencies enforcing immigration law at the county level. It is possible that this funding helps increase services and programs that assist in preparing any type of visa, including the applications for Deferred Action for Childhood Arrivals, and naturalization applications. There is a supplementary relationship between the philanthropic community and U.S. Citizenship and Immigration Services. Both of these relationships occur simultaneously as each government agency is responsible for implementing two separate but not mutually exclusive immigration goals of enforcement or status adjustments.
The second category of immigration funding, integration services, captures funding relating to the integration of immigrants into their broader community. The first five years of arrival for immigrants are the most important to establish their path to integration. During this period, immigrants, especially undocumented immigrants, do not have access to any public benefits only those provided by nonprofits (De Graauw and Bloemraad 2017). Integration funding has a statistically significant and negative relationship with the per capita total enforcement and leads to a reduction in the enforcement disparity. A one percent increase in the funding for integration services decreases the total enforcement per capita in a county by 0.05%. Furthermore, increases in funding decrease the identification of low priority to high priority deportable immigrants by 0.01%. In a similar manner to citizenship funding, if the average foundation granted 100% more funding to a county already receiving the average amount of about $39,500 for integration services, it would see a similar pattern of identification of deportable immigrants as citizenship funding. These results indicate that the philanthropic community has a multi-dimensional relationship with the government. As integration services go hand in hand with citizenship services, the relationship between the philanthropic community fluctuates between adversarial and supplementary depending on which policy output is of focus. The U.S. government provides limited to no social service aid for immigrants. Instead, it is the philanthropic community supplements everything from health clinics to language acquisition, but this also acts as protection for immigrants and can lead to an adversarial relationship with the punitive enforcement arm carrying out deportations and removals from the U.S.
The final category captures funding related to government advocacy. In aggregating grant funding up to the county level, the "partisan leanings" of foundations and nonprofits become obscure. This grant category is the only funding category where funding can support anti-immigrant or restrictive immigration policy advocacy. Thus, it is hard to detail the exact direction of the effect that funding can exhibit on policy outputs. The results indicate that instrumental giving related to advocacy has a statistically significant and positive influence on the total per capita identification of deportable immigrants but is negatively associated with the enforcement disparity ratio. It is interesting to see that the influence of philanthropic funding depends on the type of policy output. A one percent increase in government advocacy funding increases the total per capita enforcement within a county by 0.05%. The influence of government-related advocacy is the largest effect of any of the instrumental giving strategies even as the funding for advocacy is typically the least funded in total grants awarded and aggregated funding amount. On the other hand, a one percent increase in advocacy funding is related to closing the disparity between the identification of low priority immigrants to high priority immigrants by 0.01%, all else being equal. These results indicate a funding effect that leads to increases in enforcement but decreases the inequities in enforcement that targets low priority, non-threatening, immigrants for removal. Here we see that the nature of the relationship between the philanthropic community and government shifts compared to the other funding categories. The results indicate that an increase in advocacy funding for programs fighting for a variety of rights and freedoms can lead to an increase in the identification of deportable immigrants. However, the positive effect of advocacy funding does not indicate a clear cooperative relationship but rather a co-opted relationship due to the negative effect advocacy has on the disparity ratio. So, funding can increase total identifications, but it also helps protect low priority immigrants. If both had been leading to increased enforcement and disparity in policy outputs, one could conclude the possibility of a cooperative relationship between philanthropy and government.

Discussion and Implications
The quantitative analysis illustrated that funding policy activities related to integration and citizenship services can protect deportable immigrants. Foundations' funding can help influence local immigration policy outputs through intermediary organizations, but how do these grants translate into influencing policy? Interviewing nonprofit organizations who have received philanthropic grants for their work with the immigrant community sheds light on the mechanisms and processes that lead to influencing policy. The interviews provided insights into why, despite the receiving about 1.5 billion dollars in support for immigrant and immigration services, the results indicate small substantive effects. Despite desires for significant change, nonprofits directors understand that "programs might not necessarily move at a rate [needed] to be able to make a difference everywhere" 20 due to a variety of limitations from funding to current policies and executive orders. Creating policy change takes time, effort, extensive resources especially when the enforcement agencies' budgets receive unprecedented government funding. As the philanthropic community engages in multi-dimensional service provision and advocacy efforts, their relationship with the government becomes dynamic, moving between supplementary, cooperative/co-opted, and adversarial. As nonprofits become more effective in their work protecting immigrants, especially in hostile policy environments, the government can attempt to restrict the philanthropic community through special rules or regulations (Young 2000). As the immigration cases brought to the nonprofit legal service providers become increasingly complicated, the costs of representation and integration services needed per individual increases, even if nonprofits provide services at reduced costs compared to private firms.
Most immigrant-serving nonprofits, especially the grantees interviewed who offer legal services, expressed that their caseloads' demands are more extensive than the supply of attorneys or certified legal representatives available to assist in providing aid. Most nonprofits have waitlists and can only do intake forms and provide general information or referrals to other nonprofits or immigration attorneys. These legal services have risen in popularity due to their reduced prices compared to that of private immigration attorneys. Additionally, legal services have increasingly gotten popular as notario fraud has increased across the U.S. With all the services provided by nonprofits, the political citizenship programs are acknowledged as the best protection to avoid deportation. One grantee noted, "I include those [pro-immigrant] perspectives when I talk to people in public meetings, but I'd be deluding myself if I think that is going to be overarchingly effective. The most empowering thing I do for the kids is I win them [formal naturalization, permanent residency, asylum] status." 21 He points out that advocacy without reform only goes so far. Adjustment of immigrants' political status is the best protection. However, the amount of resources needed is continuously getting more challenging to attain or pay for by immigrants without aid from nonprofits, particularly when immigration enforcement only continues to heighten.
With the current backlog of applications at U.S. Citizenship and Immigration Services, nonprofits help ease the administrative burden by helping ensure accurate applications are submitted. The application services help decrease the probability of the denial of applications from status adjustments to naturalizations. With today's climate, nonprofits are seeing "more complicated cases in [our] workshops." 22 Immigrants may not realize how imperative it is for their applications to contain every detail of their lives, no matter how small. At nonprofits, attorneys can stress that knowing the full history of individuals can be the defining moment between status or deportation. "People are coming in who have more tickets, more violations, more things we need to be aware of … given this president sharing everything now is just really crucial because withholding anything would definitely be counted against you. So, as an attorney, if I know it now, we can try to deal with it and get the assistance they need as opposed to having it [withheld information] come up during the interview, then they are stuck." 23 However, the more complicated cases, which can include misdemeanors such as speeding tickets, can lead to increased time and financial resources needed to fight for naturalization or in any immigration court proceeding. In the long run, the services funded here reduce the deportable immigrant population but, in the short run, help ensure that applications are accurate and error-free for bureaucratic review. This dynamic nature of the policy activities is reflective of the relationship between philanthropic funding and government agencies responsible for implementing and enforcing immigration policy. Where services providing citizenship can be adversarial, the same services that help reduce government costs by assisting in the application process are a supplementary service for immigrants. The supplementary services often double as integration services.
Integration services seek to function as wrap-around services to create "ecosystems of social change," 24 However, funding is required to accomplish this, especially if there is a large low-income (undocumented) immigrant community. Until recently, it was up to the philanthropic sector to provide non-naturalized immigrants, including undocumented immigrants, with citizenship of noncitizens (Bosniak 2008), by including immigrants in their service provision, extending them the opportunity to access basic life necessities without regard to formal political status. Immigrants can access multiple services at their local nonprofits such as language courses to workforce development or family-focused services like food pantries. With the various programs and services provided, more than a 100% funding increase would be required to see substantive shifts across the communities. These services all help provide immigrants with access to mobility across the various measures of integration: social, economic, and cultural (Jiménez 2011;Jones-Correa 2011;Nee and Alba 2012). Immigrant integration is not a unidirectional process, but a symbiotic process between immigrants and the receiving community. The community encompasses all institutions from private to public, and nonprofits help facilitate the interaction with immigrants to promote inclusivity and advancement initiating the process of integration. One grantee noted, "the health and social services are open to everyone … we have a health fair in partnership with local clinics and universities to address basic health issues to mental health issues. In partnership with the local community college, we have instructors coming here, to provide a range of education, so if an individual wants [to get] their GED, we can provide that, but we also do [provide] ESL … to more empowerment and entrepreneurship [courses], so with adult education, we can certify individuals to open up businesses." 25 Through their service delivery, nonprofits attempt to resolve the tensions between their community norms and the immigrant community (Gilbert 2008). These services serve as a stepping stone for immigrants to understand the broader American culture and their community's norms as a way of helping them live, work, and socially engage with other members of their broader community. As the federal government increases regulations and limitations on who can access government services, philanthropic funding creates freedom and flexibility to serve all immigrant community members. Funding for these policy activities can fuel tensions between the federal government and philanthropic community even if these services and programs fill a void that does not exist through government services illustrating the multi-dimensional relationship between supplementary, cooperative, and adversarial.
Scholars have defined advocacy in multiple ways (Almog-bar and Schmid 2014). At its core, advocacy pertains to activities defining, educating, and promoting issues in attempts to mobilize or alter the public policy process. Advocacy is fundamental to helping produce policy changes through reform or the creation of new public policy (Chin 2017;De Graauw 2007;Kingdon and Thurber 1984;Le Roux and Goerdel 2009;Mintrom 1997;Mintrom and Vergari 1996;Mintrom and Norman 2009;Sabatier 1988;Schmid et al. 2008;Vaughan and Arsneault 2008). This category embodies the spirit of engaging in macrolevel changes from community education issue campaigns to political advocacy by nonprofits. Like policy change, advocacy may or may not have the desired effect despite the tireless work of nonprofits and community organizers because efforts must target multiple government venues. The complexity of working to change national policy while also educating officials tasked with implementing immigration policy was illustrated by a director who expressed, "we are doing more legislative advocacy … but we are [also] dealing with trying to engage local law enforcement to understand the U-visa system in order to get the needed signatures for [survivor] applications. They [law enforcement] have several misconceptions that [if they sign] they are helping [promote] 'illegal' immigration when, in fact, the [U-visa certification] component is designed to help law enforcement reach the criminal offenders." 26 Although nonprofit advocates and their trained immigrant advocates may host meetings and trainings, culture and structure within political institutions are difficult to overcome, which can still lead to cooperation between local law enforcement and federal immigration agents leading to increased enforcement.
While immigration reform would help provide the most protection and, ultimately, reduce unequal immigration enforcement, nonprofit advocacy does not engage in single-issue advocacy. Unlike the illegality trap (Jones-Correa and de Graauw 2013) most elected officials and the general public have fallen into, the philanthropic community focuses on service and advocacy for all immigrants, not just the undocumented community. Nonprofits engage in "a lot of coordination and information sharing between the local office staff. So, what they [regional partners] are seeing on the ground in their communities is being collected in examples and stories to do policy advocacy at the federal level, but sometimes in support of local and state too." These efforts in advocacy create opportunities for members of the organizations to share their testimony during policy development. However, their advocacy efforts do not always lead to policies inclusive of all immigrants needing aid or even the passage of legislation.

Conclusion
Identifying which issues to prioritize, to what extent, and where to devote time and resources remain vital questions for the philanthropic sector. During the Secure Communities era, there was an average of 81,000 grantmaking foundations in the United States. These foundations granted less than half a percent of their total giving directly for the immigrant community during a period where local law enforcement agencies identified and provided Immigration and Customs Enforcement with more than 2 million non-citizens' information for possible deportation (IDENT 2014). This half a percent of all philanthropic funding, almost 2 billion dollars, provided some protection for the immigrant community and funded integration services. The philanthropic protection of immigrants created an adversarial relationship with the federal government looking to deport all deportable immigrants. However, the same funding helped ensure that applications going to the U.S. Citizenship and Immigration Services offices across the U.S. were more accurate and as error free as possible. These application services help aid in reducing the application backlog and time required to review an immigration application creating a supplementary relationship between the philanthropic community and the federal government. This research indicates an association between philanthropic funding in both an adversarial and supplementary way to U.S. immigration policy. They are at best a prediction of what can happen long term if philanthropic funding continues to flow to front line organizations regardless of whether immigration is a salient policy issue in the national news or not. The limitation of this research is the overgeneralization of the motivation of instrumental giving of all foundations. With different types of internal structure and funding, foundations' motivations, strategies, and reasoning for giving will vary. It is hard to generalize to all American foundations and the results they intend to achieve. Instead, it concentrates on quantifying the influence of funding and local policy influence by following the money (Reckhow 2012). So, no matter what the reasons for giving might be, where foundations give their money has consequences for public policy.
Money influences politics, and philanthropic dollars are no different. The work that the philanthropic sector engages in takes time and requires substantial resources to reach its end goal of influencing public policy reform. During times of lull in the policy cycle, the funding provided to communities influences the local political environment and the lives of immigrants. Philanthropic funding indirectly affects immigration policy by funding intermediary organizations serving the immigrant community. With awarding of each immigration-related grant, foundations choose to either be the federal government's friend or foe when it comes to achieving U.S. immigration policy goals.