Nonprofit Economic Development Organizations and the Institutional Arrangement of Local Economic Development

In the United States, local economic development is increasingly being managed by nonprofit organizations. However, the institutional arrangement of local economic development is an understudied topic in the scholarly literature on nonprofit management and leadership. This paper examines why communities select nonprofits to manage economic development and the effect this institutional arrangement has on local development policy. We hypothesize that the form of local government and the population size of a community are variables affecting the likelihood that a community will select a nonprofit organization for economic development. Additionally, we argue that nonprofit organizations manage economic development differently than agencies directly controlled by local governments. Thus, organizational types influence economic development policy outcomes. To examine the paper’s hypotheses, we use data from the International City/County Management Association’s (ICMA) 2014 economic development survey. The paper’s analysis provides evidence that smaller cities, compared with larger communities, are more likely to select nonprofit organizations to manage economic development, and it appears the selection of a nonprofit to manage economic development influences the type of development tools used by communities.

In the United States, local economic development is increasingly being managed by nonprofit organizations. However, the institutional arrangement of local economic development is an understudied topic in the scholarly literature on nonprofit management and leadership. This paper examines why communities select nonprofits to manage economic development and the effect this institutional arrangement has on local development policy. We hypothesize that the form of local government and the population size of a community are variables affecting the likelihood that a community will select a nonprofit organization for economic development. Additionally, we argue that nonprofit organizations manage economic development differently than agencies directly controlled by local governments. Thus, organizational types influence economic development policy outcomes. To examine the paper's hypotheses, we use data from the International City/County Management Association's (ICMA) 2014 economic development survey. The paper's analysis provides evidence that smaller cities, compared with larger communities, are more likely to select nonprofit organizations to manage economic development, and it appears the selection of a nonprofit to manage economic development influences the type of development tools used by communities.

Keywords: Local Economic Development, Nonprofit Economic Development Organizations
The institutional arrangement of local economic development is changing. Compared with the past, nonprofits are more involved in managing the numerous public and private organizations engaging in local economic development. Even with this increased role for nonprofits in local economic development, we know little about the reasons why communities select nonprofits for economic development and the effect that this institutional arrangement has on local policy and outcomes. In the scholarly literature on public administration, nonprofit management, and economic development, few studies examine the decisions beyond institutional arrangements of local economic development and the effect of these administrative decisions.
Thus, our paper analyzes the institutional arrangement of local economic development in the United States. For the purposes of the research, we view institutional arrangement in local economic development as the type of organization, nonprofit, or public agency empowered to be primarily responsible for local economic development. We are interested in knowing the characteristics of communities (in particular the form of government and the size of the community) that have entrusted nonprofit development corporations to manage economic development and the effect of this institutional arrangement. To examine this research problem, we use data from the International City/County Management Association's (ICMA; 2014) Economic Development Survey. The survey asks city and county officials to report whether a nonprofit or public agency is primarily responsible for local economic development. The findings of this research may inform administrative design of economic development in the future by uncovering factors influencing the decision to select a local government agency or a nonprofit to coordinate economic development and the effect of institutional arrangement on the type of economic development policies used by communities. Given that our topic is understudied in the research, we view our study as an exploratory one. The research may also serve as a foundation for future work in the understudied area of nonprofit management and economic development.
Communities turn to nonprofits to manage economic development in the hopes of removing political interference from economic development. Many local officials also view nonprofits as being more effective in achieving economic development for communities (Krumholz, 1999). Knowing more about why communities empower nonprofits to manage economic development and the effects of this administrative decision will inform both scholarship and practice. Our exploratory research can be used to guide future work in describing and explaining the administrative features of local economic development. Having access to this knowledge will help practitioners of economic development, as they work on creating evidence-based organizations. This is a crucial goal for research because communities are focused on being selfsufficient and sustainable as a result of their economic policies.

Nonprofit Economic Development Corporations (NEDOs)
Nonprofit development corporations are 501(c)(3) organizations created to promote economic development (Stokes, 2017;Sullivan, 2004). The use of nonprofit development corporations to manage economic development, or what is usually referred to as nonprofit economic development organizations (NEDOs) in the literature, has increased in recent decades (Sullivan, 2004). NEDOs are empowered to manage economic development because these organizations bridge the gap between public and private development actors. As described by Sullivan (2004), the NEDOs "use resources from both the public and business sector to promote economic growth in a city or region" (p. 59). Accordingly, NEDOs are often considered public-private partnerships for economic development. The organizations are viewed as being efficient due to fewer regulations compared with pure public agencies, but the organizations are still able to use public resources, such as bonding authority, public buildings, and governmental staff (Sullivan, 2004). The growth of NEDOs is an important trend in the practice of community development and nonprofit management. This trend demonstrates the increasing importance of publicprivate partnerships to local economic development.
Based on past research, we know several reasons for the growth of NEDOs. First, communities may turn to nonprofit development corporations to isolate economic development from politics (Krumholz, 1999). The organizations may serve as a buffer between economic development and politics. Second, according to Sullivan (2004), the growth in NEDOs may be explained by "entrepreneurial city governments" wanting to build closer relationships with local businesses, thus creating the quasi-entities to help manage economic development. Third, NEDOs are popular administrative designs in housing policy (Goetz, 1992). This is most likely the case because nonprofits are important players in the housing policies of most communities. The increased emphasis by policymakers in offloading housing programs from government to nonprofits has contributed to the growth in NEDOs. Lastly, scholars, such as Stokes (2007), find that local governments use NEDOs and other forms of development corporations to foster collaboration among local organizations and gain a broader approach to local economic development. The scholarly literature has thoroughly explored the factors explaining the growth in NEDOs. Still, no study in the literature examines the characteristics of communities that empower NEDOs to be primarily responsible for local economic development.
Furthermore, few studies explore how the growth in NEDOs has affected development policy outcomes for local communities. The only clear effect is that NEDOs limit public participation. In perhaps the most comprehensive study, research by Sullivan (2004) examining survey data on approximately 500 NEDOs shows the organizations to be thoroughly integrated in the process of local economic development. However, it appears citizens are less likely to participate in the decision-making processes of NEDOs, compared with the levels of public involvement in public agencies.
Citizen and neighborhood organizations do not actively engage NEDOs. The organizations tend to follow the wishes of the local business community. We are interested in extending this literature by learning why communities select NEDOs to manage development and how the choice of using an NEDO to coordinate development influences the type of policy tools used by communities.

The Institutional Arrangement of Local Economic Development
While researchers examining the administrative design of local economic development have not studied why local governments select nonprofits to coordinate local development policies, scholars do know the effect of the institutional arrangement in other policy areas. For instance, work by Feiock and Jang (2009) examines why some cities may incorporate nonprofits into the service delivery of programs for the elderly. The authors found that cities with council-manager systems are more likely than mayor-council systems to contract with nonprofits for the delivery of services to the elderly. Feiock and Jang (2009) presents evidence that council-manager cities were more likely than mayor-council cities to contract out to nonprofits because city managers are more likely than their elected mayor counterparts to view nonprofits as being efficient and effective providers of services. Additionally, local government structure, in general, influences economic development policies (Cox & Mair, 1988;Feiock & Kim, 2001;Hawkins & Andrew, 2011). Based on this past research, we argue that form of government may affect the decision to have a nonprofit manage economic development. As a result, we hypothesize that a community form of local government affects the likelihood that nonprofits are given the task of helping to coordinate local economic development.
The variation in political culture throughout different regions may also influence the selection of nonprofits to manage local economic development. Region is an important variable in the study of economic issues (Feldman & Florida, 1994). In the United States, there are key regional differences in manufacturing and innovation (Florida, 2014). Thus, regional differences may affect the selection of a nonprofit to coordinate development and, as we discuss later, local development policy. Based on this, we hypothesize that the region in which a community is located affects the likelihood that nonprofits are given the task of help coordinate local economic development.
Urban areas are more likely than rural communities to use nonprofits in the administration of public policy (Feiock & Jang, 2009;Pender, 2015). Urban communities use nonprofits more than nonurban communities due to the observations that urban areas tend to have larger concentrations of professional nonprofits, compared with nonurban areas (Feiock & Jang, 2009), and nonprofits in urban areas are more likely to receive federal grant funding, compared with nonprofits in nonurban areas (Cohen, 2014;Pender, 2015). Therefore, for the purposes of our study, we view metropolitan areas as being more likely to empower NEDOs because they have a larger pool of professional nonprofit organizations, compared with those in medium and small cities. Based on this, we hypothesize that urban areas are more likely to select nonprofits to help coordinate local economic development, compared with nonurban areas.
H 1: A community's 1 form of local government affects the likelihood that economic development is conducted by a nonprofit or local government, controlling for metropolitan area and region.
The decentralized nature of the nation's federal system often encourages competition instead of cooperation. Local governments may not cooperate with one another because of leadership issues, suspicion of other localities, and resource inequalities (Lackey, Freshwater, & Rupasingha, 2002) but are likely to cooperate on policies when they share common goals and are faced with resource constraints (Post, 2002). Communities with more regional cooperation may be more likely to empower nonprofit development corporations to conduct local economic development. Cooperative communities recognize the importance of using the resources of local nonprofits. Uncooperative communities may want to retain control of economic development in a public agency. Furthermore, communities may turn to nonprofit corporations to develop a broader approach to economic development (Stokes, 2007). Thus, the institutional arrangement of local economic development may be influenced by a community's level of intergovernmental cooperation. Based on this, we hypothesize that communities in regions with high levels of cooperation among local governments are more likely to select nonprofit development organizations to help coordinate local economic development, compared with communities in regions with high levels of competition.
Communities following evidence-based practices for economic development, such as having a written plan, may be likely to empower local nonprofits to manage development. The theoretical reason is that communities that plan are more likely to be cooperative. When communities plan, they engage in a process of collaboration. Furthermore, these communities are likely to form private, nonprofit, and public partnerships that serve as the foundation for planning, and, through these integrative processes, the communities may be likely to create NEDOs, which merge private and public features into one development agency. Based on this, we hypothesize that communities with written economic development plans may be more likely to select nonprofit development organizations to help coordinate local economic development, compared with communities without written economic development plans.
H 2: Communities with written economic development plans may be more likely to select nonprofit development organizations to help coordinate local economic development, compared with communities without written economic development plans.
The administrative design of local economic development most likely influences policy outcomes. Feiock and Kim (2001) wrote that "[t]he type of agency that plays the lead role in local development may directly and indirectly shape policy choices" (p. 35). Even with such potential effect, the influence of organizational type on local economic development policy receives little attention in the scholarly literature (Sharp, 1991). Past work shows how form of government (Feiock & Kim, 2001) influences development policy, but the literature fails to explain whether the structure of local governments influences the likelihood that a community will select a nonprofit to manage economic development. In the literature on the topic, there is disagreement regarding how institutional arrangement of nonprofits affects economic development policies. Fleischmann, Green, and Kwong (1992) demonstrate how organizational type has an influence on local development policies, but a study by Feiock and Kim (2001) provides evidence of no relationship between organizational type and local development.
When it comes to the influence of organizational type on actual policies, local governments and nonprofits produce different policy outcomes (Feiock & Andrew, 2006). State and local government officials turn to nonprofits to take advantage of their flexibility (Feiock & Andrew, 2006) and the motivations of their workers. Employees in nonprofits tend to be motivated by the desire to effect social change through their organizations (Moore, 2000;Weisbord, 1988). Furthermore, nonprofits may focus on effectiveness, whereas government is concerned with fairness in a policy (Lipsky & Smith, 1989). Empirical research demonstrates that employees of nonprofits are more satisfied with their jobs than employees in government and private firms (Mirvis, 1992), and nonprofit managers highlight the "social purposes" of their organizations (Moore, 2000). These key differences between nonprofits and local governments may cause nonprofits handling local development to utilize different policies and tools than agencies controlled directly by the government.
We know little about the factors driving the use of difference types of policies for economic development. The variation in local development policy is an understudied area of the literature. As Sharp (1991) noted, "little research has been done to account for variation in cities' economic development policy activities" (p. 129). There is evidence that local governments are more responsive to the public on local economic development issues than are private organizations and quasi-public ones, such as chambers of commerce (Sharp, 1991). Mayor-council systems respond to economic pressures in a more aggressive manner than do other forms of local government. Fleischmann et al. (1992) found that development organizations directly controlled by the government are less likely to seek expansive policies than quasi-governmental agencies and nonprofits. Work by Feiock and Kim (2001) shows no evidence of this link between institutional arrangement of development and the likelihood a community will seek expansive development policies. However, we know little about what influences communities to select one policy alternative over others. We argue how institutional arrangement may be an important factor influencing the types of economic development policies used at the local level. Based on this, we hypothesize that the type of organization (nonprofit or local government) responsible for local development in a community affects the type of economic development tools used by local governments, controlling for metropolitan area and region.
H 3: The type of organization (nonprofit or local government) responsible for local development in a community affects the type of economic development tools used by local governments, controlling for metropolitan area and region. and arts/culture) dd. Tourism promotion ee. Public/private partnerships ff. Programs to promote age-friendly businesses for seniors

Methodology
The study uses data from the ICMA's (2014) survey on local economic development. The ICMA administered the survey by mailing a paper copy in June 2014 to a nationwide sample of 5,237 municipal and county governments. The survey response rate is 23% with 1,201 local governments completing it. Past studies (Feiock & Kim, 2001;Sharp, 1991) rely on the ICMA data to examine the administrative features of local economic development. Data from the ICMA survey allow for the analysis of our study's main hypotheses. First, the survey asks whether a nonprofit or government agency is primarily responsible for local economic development. The responses to this question allow for the analysis of the factors making certain communities more likely to empower a nonprofit to manage economic development. Additionally, the responses to the question allow for the analysis of the data as an independent Dummy variables created from four US Census Bureau regions: "Northeast", "North Central", "South" and "West" with "Northeast region" as the reference group Level of Cooperation Weak = 0 Strong = 1 variable to determine the effect of the institutional arrangement of economic development on local development policies.
The ICMA survey also includes questions related to form of government, population, region, and other useful questions to help build regression models exploring why communities select nonprofits to manage economic development. Additionally, the ICMA survey asks respondents to indicate the extent to which their organizations use policies focusing on small business, retention of current businesses, and community development. The questions dealing with economic development tools can be used to analyze the effects that institutional arrangements have on local economic development. As can be seen in table 1, the communities surveyed by the ICMA are using a variety of economic development tools. The categories were developed by the ICMA and included in the survey questions. We used the categories to develop indices of the economic development tools used by communities. With the information from questions dealing with economic development tools, we can build regression models explaining the effect of institutional arrangement on policies used by communities. Table 2 presents the operational rules for the variables that we included in our models.
Descriptions of the variables are provided in table 3. Economic development responsibility was measured by survey question asking communities to describe the entity that has primary responsibility for undertaking economic development activities. The response to the survey question consisted of "the local government has primary responsibility" = 1, "A nonprofit development corporation has primary responsibility" = 2, and "Other (please describe)" = 3. Analysis showed that most of the "other" response options were nonprofit entities and were combined with the "A nonprofit development corporation has primary responsibility." For purpose of this analysis, economic development responsibility was coded "nonprofit" = 0 and "local government" = 1. Local economic development tools were measured using 32 four-point scale items relating to communities' evaluation of their levels of use of different tools in three general economic development areas. The three general economic development areas were small business (eight survey items), business retention and expansion (12 survey items), and community development (12 survey items) activities. The response option to the survey questions consisted of "not at all" = 1, "low" = 2, "medium" = 3, and "high" = 4. For the purpose of this analyses, we constructed a composite score for each of the three general economic development areas using the sum of the response to the respective survey items. Accordingly, each of the constructs used to measure economic development tools consisted of the sum of each communities' response across the items used in measuring that particular economic development area. The variables for economic development tools were constructed so that higher scores indicate a higher level of use of the tool.
Other variables for the study were form of government, metro status, population, economic development plan, geographic region, and level of cooperation. Form of government was coded "Mayor-Council and Council-Elected" = 0 and "Council-Manager and Council-Administrator" = 1. Metro status was coded "large" = 0 and "small" = 1 with large comprising communities in urbanized areas with population of at least 50,000 people and small was for communities with at least one urban cluster that has a population of at least 10,000 people. Population was measured by the total number of people living in the community in 2010. Economic development plan was measured by survey questions asking communities to indicate whether they have a written economic development plan and was coded "No" = 0 and "Yes" = 1. Geographic region was measured by dummy variables created from the four population regions of the U.S. Census Bureau: "northeast," "north central," "south," and "west" with "northeast region" as the reference group. Finally, level of cooperation was measured by survey questions asking communities to indicate their level of cooperation for economic development and tax base among local governments in their region and was coded "Weak" = 0 and "Strong" = 1.
Next, we present the results of our paper's analyses. We examine the descriptive results and analyze the multivariate results-in particular, two collection of models examining the decision to select nonprofits to manage economic development and the effects of this organizational choice on the development tools used by communities.

Analysis: The Selection of Nonprofits to Manage Economic Development
In most communities, a public agency directly controls local economic development efforts. Seventy-six percent of communities give primary responsibility for economic development to a public agency. Still, a notable percentage of communities (23%) off-load responsibility for coordinating economic development to a nonprofit development corporation. Given the growth in NEDOs since the 1970s (Sullivan, 2004), the small percentage of communities selecting nonprofits to be primarily responsible for economic development is surprising. However, communities are comfortable with involving NEDOs as players in the local development process. The ICMA data shows how 40% of the sampled communities report including economic development corporations in their local development strategies. The other noteworthy finding from the descriptive statistics is that a surprisingly large percentage of local governments (43.5%) in the ICMA data lack a written development plan. It appears that many communities are not utilizing one of the basic economic development tools. Next, we examine the factors influence local governments to empower nonprofit agencies to manage local development policy.
In our analysis, we first examine the characteristics of communities that are more likely to empower nonprofits to manage development, compared with the localities that rely on public agencies. As noted, we hypothesized that form of government will be one of the influences. The ICMA survey asks local governments (cities and counties) to identify their form of government. For the sampled cities, respondents identified one of the following forms: mayor-council, council-manager, commission, town meeting, and representative town meeting. For the sampled counties, respondents identified one of the following forms: commission, counciladministrator (council-manager), and council-elected executive. We combine the form of government questions to analyze all local governments. Only 15 cases identified as cities commission (CO) and none for counties commission (C). In effect, the combination of cities and counties for this category results in only 15 cases being removed from the analysis.
We recognize how the ICMA data oversamples council-manager forms of local government. However, researchers use the ICMA to examine a host of administrative issues at the local government level. The data set is a standard one in the public administration literature. For instance, researchers use the data set to explain the selection of nonprofits for social services delivery (Feiock & Jang, 2009) and the influence of nonprofit organization selection on development policy (Feiock & Kim, 2001;Sharp, 1991). These two studies are central to our research. We employed logistic regression to determine the effect of form of government on the likelihood that a nonprofit development corporation or the local government is primarily responsible for economic development, controlling for metro status, region of the country, level of cooperation among governments in the community's region, and the community having an economic development plan. The analysis produces an overall statistically significant model χ 2 (1)=58, p<0.05 and explains 0.15% (Nagelkerke R 2 ) of the variance in whether economic development is conducted by a nonprofit or local government and correctly classified 58% of cases. Form of government with council-manager and council-administrator does not have a statistically significant effect on whether a community selects a local government agency or a nonprofit development corporation to be primarily responsible for economic development. Table 4 presents interesting findings for the literature on economic development and nonprofit organizations. First, larger communities based on population are more likely to use NEDOs, compared with smaller communities. For example, smaller communities (communities with at least one urban cluster that has a population of at least 10,000) were 0.26 times less likely than urban communities (communities with at least one urbanized area that has a population of at least 50,000) to have local government as an entity with primary responsibility for economic development (Exp(β)=0.26; p=0.00). Second, form of government (council-manager and council-administrator or more political forms) has no significant effect on the selection of a NEDOs to be primarily responsible for economic development (Exp(β)=0.93; p=0.78).
Third, region has a significant effect on communities selecting NEDOs. The south is less likely than the northeast and the north-central region to have an NEDO primarily responsible for economic development. Lastly, the level of cooperation for economic development in the community's region (measured by self-reporting from the community) affects the decision to give primary responsibility for economic development to an NEDO (Exp(β)=0.62; p=0.05). In fact, communities that report cooperation with other communities in their region were 0.62 times more likely to turn to an NEDO.
Based on the analysis, metropolitan status, region of the nation, and level of cooperation with neighboring communities are all factors that influence the institutional arrangements of local economic development. In effect, holding all other independent variables constant, we expect that smaller communities will be using NEDO primarily for economic development than larger communities. Local public agencies in small communities may lack the expertise needed to directly coordinate economic development activities. It appears that small communities turn to nonprofits to help build policy capacity. Chambers of commerce and development authorities may be playing an important role in small cities and communities, compared with agencies directly controlled by government (Rubin, 1986). Research shows that the presence of a development organization in a community produces more economic development activity (Rubin, 1986). Small communities may be turning to nonprofits to help build capacity, incorporated local businesses, and increase economic development activity.
Though the results indicate that, after controlling for the other predictor variables, the odds of using NEDO primarily for economic development by council-manager and counciladministrator government is 0.93 times that of mayor-council and council-elected government. This observed difference might have occurred by chance or some random events. A key finding in our research is that form of government does not influence the institutional arrangement of local economic development.
Compared with the northeast, the north central and west regions had higher odds of (3.19 and 3.58 times) of using NEDO primarily for economic development. It appears the south is less likely than other parts of the nation to use NEDOs or nonprofits in local economic development. The region's low levels of political participation may make it less likely that communities form separate nonprofits to coordinate economic development. Future research needs to tease out why the south is less likely to utilize nonprofits, compared with the northeast and other parts of the nation.
The odds of using NEDOs primarily for economic development for communities reporting strong cooperation was 0.62 times that of communities reporting weak cooperation, holding all other independent variables constant. Communities cooperate when citizens share common goals and are in research-constrained environments (Post, 2002). We argue that communities that identify as being in regions with more cooperation among public organizations are more likely to use nonprofits to coordinate development, compared with high-conflict areas.
Cooperative communities recognize the importance of multiple types of organizations working together to achieve policy goals. Stokes (2007) found that communities turn to nonprofit corporations to develop a broad approach to economic development. Cooperative communities may be more likely to want a diverse approach to development. We found evidence that cooperative communities are more likely to empower an NEDO to manage development.
Next, we examine the effect of institutional arrangement on the types of economic development tools used by communities.

Effect of Institutional Arrangement on Economic Development Policy
To examine the effect of organizational type on economic development policy outcomes, we use ordinary least square (OLS) regression to explain the institutional arrangement influence on the type of economic development tools employed by communities (see table 1). We used the constructed indices for small business tools, retention and expansion tools, and community development tools. We use these indices as the dependent variables in a series of OLS regression models explaining the influence that organization type has on the type of tools used by communities.  Source: ICMA (2014) Note: Statistical significance at the 0.01 level and 0.05 level is indicated by ** and *, respectively. Dependent variable in this analysis is business retention and expansion activities; sum of 12 business retention and expansion initiative as index score First, we analyze whether economic development being conducted by a nonprofit or local government influences the small business index, controlling for form of government, and whether metro status and geographic region influences the small business index. Table 5 presents the results of this analysis. The model is statistically significant, F(6)=6.71, p<0.05 and explains only 3.8% of the variation in small business index. Except for metro status, none of the other variables have a statistically significant effect on small business index. Specifically, small metro areas (urban cluster with a population of at least 10,000) are more likely to be involved in small business initiative than a large metro area (urbanized area with a population of at least 50,000) (b= 2.08, p=0.00).
Second, table 6 shows no relationship between economic development responsibility and the use of business retention and expansion tools. It appears that communities with nonprofits managing local economic development are just as likely to use business retention and expansion tools, as localities with public agencies directly managing economic development. Third, it appears that economic development responsibility affects the type of community development activities being conducted by cities and counties. Based on the analysis presented in table 7, local governments with nonprofits responsible for economic development were more likely to utilize community development tools, compared to local governmetns with public agencies being the primary organizations leading economic development.
While whether economic development being conducted by a nonprofit or local government appears to influence the small business index, the effect is masked by the presence of control variables. This finding suggests that, regardless if economic development is conducted by nonprofit or local government, it does not affect involvement in small business activities, but the evidence points to metro status affecting involvement in small business initiatives. This is most likely the case because small businesses are often crucial to the local economy in a small community, and owners of small businesses can also exert more influence in local political processes of small cities than they can in large cities. Furthermore, smaller cities are less likely than major urban areas to have large corporations.
Next, we construct an OLS regression model to determine whether using nonprofit or local government as a tool of economic development has an effect on business retention and expansion index, controlling for form of government, metro status, and geographic region. The model is not statistically significant, F(6)= 1.84, p>0.05, and none of the variables have a statistically significant effect on business retention and expansion. This is an interesting finding in that none of our variables were shown to have an effect on business retention and expansion policies, even form of government.
We construct a model to determine whether using nonprofit or local government as a tool for understanding economic development has an effect on community development index, controlling for form of government, metro status, and geographic region. The model is statistically significant (F(6)=3.85, p<0.05) but only explains 1.9% of the variation in community development index. Communities with economic development being coordinated by a public agency are less likely to be involved in community development activities than economic development by nonprofits (b=-1.43, p=0.01). Also, compared with the northeast region, the north central and west regions were less likely to be involved in community development activities. The south region was also less likely to be involved in community  Source: ICMA (2014) Note: Statistical significance at the 0.01 level and 0.05 level is indicated by ** and *, respectively. Dependent variable in this analysis is business retention and expansion activities; sum of 12 business retention and expansion initiative as index score development relative to the northeast region, but the difference was not statistically significant (b=-1.78, p=0.09). The results suggest that organizational type and geographic region affect involvement in community development activities.
To examine the size of a community in greater detail, we substitute the population size of the local government for the dichotomous variable of metro status. The next three OLS tables examine the effect of this change to our models. Table 8 presents the result from analysis of small business activities using population. In the original model for small business activities, the model was statistically significant, and only metro status was statistically significant. When population (instead of metro status) was used, the model remained statistically significant, and the agency responsible for economic development in the south region became statistically significant. Additionally, the proportion of variance in small business activities explained by the independent variables (adjusted R 2 ) decreased by 1.3% (from 0.04 to 0.03). In the original model for community development activities, the model was statistically significant, and the agency responsible for economic development, north central, and west regions was statistically significant. When population (instead of metro status) was used, the model remained statistically significant, and only the agency responsible for economic development and population was statistically significant (see Table 9). Additionally, the proportion of variance in community development activities explained by the independent variables (adjusted R 2 ) increased by 2.3% (from 0.02 to 0.04). Table 9 presents the result from analysis of business retention and expansion activities using population. In the original model for business retention and expansion activities, the model was not statistically significant, and none of the independent variables was statistically significant. When population (instead of metro status) was used, the model became statistically significant, and population also became statistically significant. Additionally, the proportion of variance in business retention and expansion activities explained by the independent variables (adjusted R 2 ) increased by 1.6% (from 0.01 to 0.02). Table 10 presents the results when using population size as the metro status variable. According to the analysis, communities with nonprofits managing economic development are still more likely to use community development tools, compared to localities with public agencies directly manaing economic development. In this model, population is also shown to have an effect with larger communities more likely to use community development tools.
In sum, in all three activities (small business, business retention and expansion, and community development), the model was statistically significant, and adjusted R 2 increased except for small business activities. Also, in all three activities, population had a statistically significant positive effect. The "marginal effect" (regression coefficient or "b") and the corresponding standard error for population are low, which is most likely due to unit of measurement. Additionally, except for business retention and expansion activities, the agency responsible for economic development had a statistically significant effect. Specifically, when local government has primary responsibility for economic development, it is less likely to engage in small business and community development activities.
The low R 2 values in this study demonstrate a limitation of our models and analyses. In effect, our models have low predictive powers and are not reliable for future forecasts (we cannot make good predictions based on the model). Decisions involving the determination of the entity with primary responsibility for undertaking economic development activities and tools to use for economic development activities are complex and might involve many uncontrollable and/or unknown factors that were not accurately captured in our model. A more complex model accounting for the factors affecting political decision-making might help to more accurately capture decisions on the entity with primary economic development responsibility and use of economic development tools. Besides, R 2 is affected by several factors, including the nature of the data, and it is possible that the nature of data used in this study, particularly the lower range of values of the independent variables, contributed to the lower R 2 . This appears to be supported by the changes in R 2 when population, instead of metro status, was used for our analysis of economic development tools. Furthermore, our models do provide information on what variables do or do not predict the likelihood that a community would select a nonprofit or local government agency and then the effectiveness of that institutional choice.
The low variation in our models is one of the reasons why the research should be exploratory in nature. We are starting the process of collecting evidence on the institutional arrangement of local economic development. Future models should explain the effect of politics on factors influence institutional arrangements for development and the effect that these administrative choices have on local economic development. Given that our research is an understudied area of public administration and economic development, we argue that, even with the low variation in our models, our findings contribute to the literature.

Findings and Discussion
In this paper, we analyze why nonprofits are selected to coordinate local development policies and the effect of this decision on economic development. The paper's analysis finds little evidence that form of government influences the decision of local governments to give responsibility for economic development to a nonprofit agency. We did find evidence that place matters when it comes to the institutional arrangements of local economic development. First, smaller cities are less likely to select a public agency to coordinate development policy, compared with larger cities. It appears small communities are relying more on NEDOs and nonprofits to help implement their economic development policies. Feiock and Jang (2009) found a different result. The authors argued that smaller communities have fewer nonprofits; therefore, smaller cities were more likely to turn to a local government agency to manage economic development. We found evidence that the opposite may be occurring. Small communities rely on local nonprofits.
Second, the location of the community, i.e., its region, in the nation affects the institutional arrangement of economic development. Communities in the northeastern part of the United States are more likely to use nonprofits in managing economic development, compared with the rest of the nation. Additionally, as past work shows, citizens are less likely to participate in the decision-making processes of NEDOs (Sullivan, 2004). Due to the lack of research in this area, future work needs to be done to understand the factors leading to communities in the south being more likely to empower local nonprofits to manage economic development.
Third, cooperative communities are more likely to turn to a local public agency to manage economic development, compared with communities where economic developers indicated uncooperative behavior with neighboring communities. This is a useful finding for the literature in that it goes against the theoretical reasoning that cooperative communities are likely to involve many groups in policymaking so, therefore, will be more likely to empower nonprofits to manage development. Cooperative communities may have leadership characteristics that push them toward the more centralized control of keeping economic development in a public agency. These communities may work best with neighboring jurisdictions because there is not as much decentralization in their leadership and policymaking processes. In other words, being cooperative may be the dependent variable being explained by centralized institutional arrangements-not vice versa.
In additional to form of government's lack of effect on institutional arrangement, another important null finding is how communities with a written plan are just as likely to empower local governments to manage economic development as they are to select nonprofit organizations to be responsible for the policy area. As mentioned, the large percentage of ICMA communities without a written economic development plan is surprising, but the finding follows past research. Jennings and Hall (2012) examined the evidence-based practices of state agencies, and they found economic development to be one of the most politicalized policy areas. Perhaps local economic development agencies are just as likely as their state counterparts to follow politics over evidence and be weary of drafting a plan that may be politically risky. Our study, however, shows that institutional arrangement of development is not related to the presence of a written economic development plan.
Whether economic development is managed by a nonprofit or local government agency only had an effect on community development activities. Nonprofits managing economic development were more likely than local government agencies to utilize community development activities. Our analysis shows that communities with nonprofit organizations coordinating economic development are more likely to utilize policies promoting small businesses and community development.
Accordingly, our analysis differs from the finding of Feiock and Kim (2001) that organizational type does not affect development policy. Nonprofits reported higher usage of tools such as transit to promote community, job training for low-skilled workers, high-quality physical infrastructure, tourism promotion, the formation of community development corporations, and other community-development policies. It appears that communities with nonprofits taking the lead in development are more likely to recognize the importance of community-developmenttype policies. This finding is not striking because community-development corporations (CDCs) are often involved in community-development work and are organized as nonprofits.
Communities more likely to have CDCs are more likely to shift responsibility for economic development from a government agency to a nonprofit. Additionally, it appears that communities with more public-private partnerships are more likely than other localities to have a nonprofit primarily responsible for economic development. Future research is needed to explore the finding regarding nonprofits being more likely to engage in small business promotion and community development, compared with government agencies.

Conclusion
The analysis of the ICMA data shows that organizational type is an important administrative design component affecting local economic development policy. Metro status and geographic region appear to affect the likelihood that economic development is conducted by a nonprofit or local government agency. However, form of government does not affect the selection of nonprofits to manage local development. It appears that smaller cities (with populations below 50,000) are more likely than larger urban areas to invest in their small businesses. On the other hand, the analysis provided no evidence of what drives business retention and expansion activities in the ICMA communities. Lastly, we found that nonprofits managing development were more likely than local governments to use community-development activities, such as creating public-private partnerships, promoting tourism, and developing place-based infrastructure.
Our findings present implications for economic development managers working in local public agencies and nonprofits. First, the limited findings on the factors explaining the selection of nonprofits or public agencies in managing economic development can help inform local practitioners about the institutional arrangements of local development made by similar and different communities-in population size, local region, form of government. We learned how small communities are likely to turn to nonprofit organizations to help with economic development. One practical implication of this finding is that local public managers and public affairs programs need to focus on educating and training nonprofits in the area of economic development. Second, we found evidence supporting the hypothesis that communities with nonprofits managing local economic development are more likely to implement communitydevelopment policies. Community-development tools focus on transit, housing, and an overall higher quality of life through education, amenities, and recreation. For policymakers looking to implement these types of policies into their communities, based on our limited findings, we argue they should select a nonprofit to coordinate their local economic development.