Sexual Minority and Employee Engagement : Implications for Job Satisfaction

Despite the increasing attention given to the construct of work engagement in the workplace, it remains under-researched in the academic literature. Using Kahn’s conceptual foundation of work engagement, this study examines whether high levels of work engagement lead to equally satisfying work experiences for members of the workforce regardless of their sexual orientation. Using the 2012 Federal Employee Viewpoint Survey (FEVS), authors find that while active engagement at work had positive influence on employee job satisfaction regardless of one’s sexual orientation, high level of engagement at work among LGBT employees was less strongly associated with job satisfaction than it was for those non-LGBT employees. Implications are discussed.

Work engagement, defined as the employment of oneself physically, cognitively, and emotionally during role performance (Bakker & Demerouti, 2008;Kahn, 1990), has received much attention recently as having positive consequences for employees and organizations (Jeung, 2011;Saks, 2006). The benefits of being engaged include increased job satisfaction (Schaufeli & Bakker, 2004). Similarly, research has shown that disengaged employees cost organizations a heavy financial burden (Frank, Finnegan, & Taylor, 2004). Therefore, work engagement is argued in the literature to cure many organizational problems (Banihani, Lews, & Syed, 2013).
Currently, the image of an ideal worker in most organizational studies is an "engaged" one, regardless of one's social or cultural identities, to help improve the organization's outcomes (Banihani et al., 2013;Wilson, 1998). This identity-neutral view of work engagement assumes that men and women, heterosexuals and gay men and lesbians, or white and those of racial minority can equally demonstrate their engagement in the workplace. Although empirical studies have not specifically examined the moderating role of employee's sexual orientation on the relationship between work engagement and its outcomes, researchers in social demography indicate that diversity characteristics of group members may moderate the impact of work engagement on important work-related outcomes (Banihani et al., 2013;Milliken & Martins, 1996;Williams & O'Reilly, 1998). We argue that ignoring sexual orientation in organizational research and theory contributes to the perpetuation of inequalities in the workplace.
The aim of this study, therefore, is to respond to the demand for incorporating sexual orientation into organizational research and to work engagement. The main question raised in this study is: Does sexual orientation influence work engagement and organizational outcomes?? Despite the recognition in the literature that there is no 'one size fits all' model of engagement and that the effects of work engagement may vary with one's sexual orientation, to our knowledge no empirical studies have been conducted on the extent to which one's engagement with work impacts employees of sexual minority. Furthermore, despite the increased attention to strengthening the principle of equal employment opportunity in the public sector for lesbian, gay bisexual, and transgender (LGBT) individuals, most notably by the President Obama who declared the promotion of workplace diversity and workplace inclusion a 4 top priority through his recent Executive Order, research related to sexual orientation has been restricted by scarcity of appropriate data (Leppel, 2014). Therefore, understanding the impact of sexual orientation on the relationship between employee engagement and work-related outcomes will be one of the critical first steps for developing more inclusive human resource management practices. Previously there has been no comprehensive empirical research that examined the extent to which LGBT employees feel (dis)satisfied with their job compared their counterparts. As such, it stands to reason that understanding their responses to organizational attachment may help clarify the extent to which their LGBT status affects them emotionally. Therefore, this research is of particular practical importance to public managers who are concerned with employee satisfaction.
The study is structured as follows. In the next section, we start with the review of relevant literature on the work engagement and organizational outcome relationships, followed by a theoretical context of integrating sexual orientation as a moderator between work engagement and employee job satisfaction. We then explain the data and method used, describe the variables analyzed, and outline key findings. We close with both implications of the findings and recommendations for practice and future research.

Work Engagement and Job Satisfaction
An important first step in understanding how sexual orientation influences the relationship between work engagement and job satisfaction is to determine the direct theoretical linkages between work engagement and its proposed outcomes. To explain these relationships, we turn to several extant theories in social and organizational psychology.
Cognitive dissonance theory suggests that individuals have an inner drive to hold all our attitudes and beliefs in harmony and avoid disharmony (or dissonance) (Aronson, 1969;Festinger, 1957; also see Blanchard, Welbourne, Gilmore, & Bullock, 2009). According to Festinger (1957), we hold many cognitions about the world and ourselves; when they clash, resulting in a state of tension known as cognitive dissonance. As the experience of dissonance is unpleasant, we are motivated to reduce or eliminate it, and achieve consonance (i.e., agreement). In the workplace, employees want their behaviors and attitudes to be aligned. The basic premise is that because attitudes are usually easier to change than behaviors, employees will change their attitudes to justify their behavior (Blanchard et al., 2009). For example, employees with high level of engagement in work by taking ownership of projects and creatively approaching their tasks are more likely than others to justify their behavior by believing that their jobs are both "good" and "worth the extra effort" (Blanchard et al., 2009, p. 115).
Similarly, in his pioneering research on followership theory, Kelley (1992) explains that active engagement in work role fulfills important personal needs for employees as it provides for comradeship with co-workers, satisfying one's social needs (Howell & Costley, 2006). In addition, by actively engaging in one's work, employees are more likely to identify with the leader, which enhances one's self-concepts. This reinforcement of one's self-concept then helps satisfy individual needs for self-esteem. While research surrounding followership theory is relatively sparse, Howell and Costley (2006) suggest that active work engagement is likely to result in increased motivation, satisfaction, and feelings of empowerment (also see Macey & Schneider, 2008;Jin & McDonald, forthcoming). Several empirical works support the argument that engagement is a causal antecedent of several vocationally relevant outcomes, including job 6 organization (Kahn, 1990;May, Gilson, & Harter, 2010). According to social dominance theory (Sidanius & Pratto, 1999), every organization is organized by systems of social group-based hierarchies in which at least one social group has dominance over others. It is in this hierarchy that some individuals (i.e. those who identify as LGBT) may experience disproportionate privilege and power due to their membership in a social group that has low social status within the organization. Individuals with higher levels of social dominance orientation, therefore, display discriminatory behavior towards the group to which they do not belong (Sidanius, 1993). The question, then, is whether the so called "heterosexism" exists in today's workplace. Research studies have demonstrated the existence of sexual stigma (i.e., the shared knowledge of society's negative regard for any behavior, identity, or community that is not heterosexual), heterosexism (the cultural ideology that perpetuates sexual stigma), and sexual prejudice (negative attitudes based on sexual orientation), as well as the effects that such attitudes have on the everyday experiences of LGBT employees (Herek, 2002). Therefore, it can be said that LGBT individuals and their characteristics are less valued and regarded as less useful than heterosexuals. As a result, several scholars have argued homosexuals will expect lower job satisfaction because their workplace often does not allow them to present more of their selves to the jobs (e.g., Drydakis, 2015). Taking this into account, heterosexuals are in general expected to have better status and better influence than homosexuals in work environment which makes them to have fuller confidence and higher expectations (and thus higher level of consistency in their beliefs) that their dedication to work will lead to high level of job satisfaction (Festinger, 1962). Subsequently, understanding the challenges that come from the existence of sexual stigma in the workplace, it is reasonable to assume that those of sexual minority will have lower expectations that their engagement at work will lead to equally satisfying work experience compared to their heterosexual counterparts. Also, given the empirical evidence that gay men and lesbians are more likely to report lower job satisfaction than heterosexuals (e.g., Drydakis, 2015), it is not out of line to expect that the relationship between work engagement and job satisfaction will be weaker for homosexuals.
Furthermore, psychological safety is referred to as "feeling able to show and employ one's self without fear of negative consequences to self-image, status, or career" (Kahn, 1990, p. 708). Similarly, it is argued that situations that are characterized by more psychological safety allow people to be more engaged (Nembhard & Edmondson, 2006). As mentioned earlier, research studies ranging from population surveys and statistics to experimental studies and self-report questionnaires provide support for the existence of heterosexism in the workplace (e.g., Blandford, 2003;Herek 2002). In fact, according to the U.S. Equal Employment Opportunity Commission (2011), the number of discrimination charges received during the 2010 fiscal year increased by seven percent from the previous year. Lim and Cortina (2005) argue that these mistreatments and workplace deviances are often directed at individuals of sexual minority. As for psychological availability, the final element in Kahn's theory, which refers to "the sense of having the physical, emotional, or psychological resources to personally engage at a particular moment" (Kahn 1990, p. 714), Kahn posits that engaging in role performance depends on how employees cope with the various demands of both work and non-work aspects of their lives. However, organizations are built on the image of the "traditional worker", an image which is mostly emulated by heterosexuals (Acker 1990). Although some units of government are making progress in supporting LGBT rights, unfortunately, there has been no consistent set of policies or systems in place to assure universal protection of LGBT workers (Norman-Major & Becker, 2013). For example, while various federal laws (e.g., Civil Rights Act of 1964) provide protection against discrimination based on race, sex, religion or national origin, or even disability and age, there is no parallel federal law prohibiting discrimination in the public or private sector on the basis of sexual orientation (Norman-Major & Becker, 2013). Based on the theoretical argument and extant empirical evidence, we explore the following hypothesis: H3: The positive relationship between work engagement and job satisfaction will be less strong among lesbians, gay men, bisexuals, and transgender employees, as opposed to heterosexual men and women.

Data
The data used in this study come from the Office of Personnel Management's (OPM) 2012 Federal Employee Viewpoint Survey (FEVS) (OPM 2012). For the first time since it began as the Federal Human Capital Survey, the Federal Employee Viewpoint Survey attempted to reach all permanent full-and part-time civilian federal employees covering 97 percent of the executive branch workforce. A total of 687,687 eligible respondents completed rendering a response rate of 46 percent, making it the largest and most diverse response to the FEVS to date.
Most importantly, this survey marks the first time the Office of Personnel Management added an LGBT identifier to a survey of federal employees, which provides a wealth of data on employment patterns, job satisfaction, perceptions of discrimination and turnover intentions, resulting in a sample of over 13,599 LGBT federal employees to study, not counting those who "preferred not to identify" themselves as LGBT (65,562 cases) and missing cases (78,686 cases). This LGBT release contains the first comprehensive national data on LGBT public employees (Norman-Major & Becker, 2013), which makes our analysis critical for future studies.
The FEVS survey was further condensed into several constructs by scaling similar questions. Table 1 shows the descriptive statistics of variables used in the study. The scaled variables were all mean centered in the actual analysis in order to reduce multicollinearity in the interaction terms.

Dependent Variables
The survey items did not contain a tested scale for job satisfaction, so we utilize a proxy index for this measure. An additive index was developed with three items (alpha = 0.72) capturing satisfaction with job, pay, and organization: (1) "Considering everything, how satisfied are you with your job?" (2) "Considering everything, how satisfied are you with your pay?" and (3) "Considering everything, how satisfied are you with your organization?"

Independent Variables
Work engagement is measured by an index of five items (Cronbach's alpha = .63). While they generally reflect the degree of one's engagement at work (e.g., Kelley, 1992), specifically, two items tap employees' physical engagement (e.g., Kahn, 1990): "My talents are used well in the workplace," and "The people I work with cooperate to get the job done. Two other items measured cognitive engagement (e.g., Kahn, 1990;Rothbard, 2001): "I know how my work relates to the agency's goals and priorities," and "I am held accountable for achieving results." And one item taps emotional engagement (e.g., Harter et al., 2002;Kahn, 1990): "I am constantly looking ways to do my job better." Supervisory status, gender, age, tenure in federal agency, turnover intention and sexual orientation are included as control variables. As briefly described earlier, due to the OPM's strategy for ensuring confidentiality, some of the demographic characteristics were suppressed to prevent identification of individuals. Supervisory status variable, initially with two separate leadership groups between manager and executive, was given as a dummy variable (1=manage/supervisor/executive, 0=non-managerial). Gender (female=1), turnover intention (leaving=1), and sexual orientation (LGBT=1) were included as dummy variables. Table 2 presents the results of hierarchical regression model. The demographic variables in step 1 of the job satisfaction model show interesting relationships with the dependent variable. Results show that those in the non-managerial positions were less likely than supervisors to report high levels of job satisfaction (p < .001). Female workers were more likely than males to report high levels of job satisfaction (p < .001). Age was also a positive predictor of job satisfaction (p < .001). Tenure, turnover intention, and sexual orientation were negatively related. For example, an increase in tenure in the federal agency was associated with decrease in job satisfaction (p < .001), while those with intention to leave were more likely than others to report low levels of job satisfaction (p < .001).

Findings
LGBT employees were more likely than non- LGBT employees to report low levels of job satisfaction and thus support hypothesis 2.
As hypothesized, work engagement had significant positive effect on employee job satisfaction (p < .001) (hypothesis 1). Specifically, one unit increase in work engagement was associated with .852 unit increase in job satisfaction scale (see step 2). Adding work engagement in step 2 created slight changes among the effects of demographic variables.
LGBT employees remain less satisfied than their heterosexual counterparts, but the coefficient becomes smaller by more than half. This indicates that some of the dissatisfaction reported by LGBT employees can be attributable to low work engagement (Pitts, 2009). Once work engagement was controlled, the dissatisfaction is reduced, signifying the relevant role of work engagement. Other demographic variables also show slight changes. Managerial authority, tenure, and turnover intention all remain negatively associated, but their coefficients are reduced to -.043, -.009, and -.312, respectively. More interestingly, when taking into account the impact of work engagement, the direction of relationship changed for gender and age; female workers were less likely than males to report high job satisfaction and an increase was associated with lower job satisfaction.
In step 3 of the model, we added an interactive term. A statistically significant result would indicate that the effect of work engagement on job satisfaction is moderated by the employee's sexual orientation. The result indicates that sexual orientation moderates the relationship between work engagement and job satisfaction, and that its effect on job satisfaction was less positive among LGBT employees, supporting hypothesis 3.
In summary, because the interaction effect is statistically significant, the regression coefficient (β = .853) of work engagement represents its conditional effect on job satisfaction when employees are heterosexuals (i.e., sexual orientation = 0), controlling for all other variables. The effect of interaction, depicted in Appendix, shows that while work engagement had generally positive effect on job satisfaction regardless of one's sexual orientation, the effect was stronger among non-LGBT employees.

Discussion and Limitations
Despite the evidence of heteronormativity in the workplace (Pringle 2008), no direct studies have been conducted on examining the differential effects of work engagement regarding employee's sexual orientation. Additionally, most of the work engagement studies focus predominantly on how it affects organizational performance as its primary outcome. While we recognize a few studies that focused on its impact on psychological affectivity, findings were mixed (e.g., Alarcon & Edwards, 2011;Thian, Kannusamy, & Klainin-Yobas, 2013). As mentioned in Lewis and Pitts' (2009) study, lack of data involving LGBT individuals in the public sector has been a primary obstacle. This study responds to these needs, analyzing the effects of work engagement on job satisfaction from the lens of sexual orientation. The results of this study add to recent research that indicates the importance of social identities for understanding the effects of work engagement on organizational outcomes (Banihani et al., 2013;Pringle, 2008). Moreover, this study is the first large-N study using public sector survey data, which fielded identifying information about LGBT employees.
There are three primary findings that warrant highlighting. First, work engagement matters in the workplace. In each model (i.e., step) in which it was included, work engagement was positively and significantly related to job satisfaction. Second, the moderated regression model demonstrated that work engagement affects job satisfaction more positively and significantly among heterosexual employees than it does LGBT employees, resulting in lower job satisfaction for those of sexual minority. It can be seen that the psychological conditions that lead to work engagement can be more emotionally challenging on LGBT employees because of organizational culture and the belief that heterosexual characteristics should be emulated by all workers (Banihani et al., 2013).
It is expected that, when other things are equal, engaged workers will generally gain greater satisfaction than those who are less engaged or completely disengaged from work. However, in reality, those "other things" are not always equal. As suggested in our findings, the impact of work engagement in regards to job satisfaction differs with one's sexual orientation. A related research by Banihani et al. (2013) has shown that work engagement is a gendered construct where it is easier for men to be engaged than women. These findings together suggest that one's social identity is an important construct that needs to be examined in future engagement studies. For public sector organizations to be successful in both recruiting and retaining talented workers who are sexual minority, they must look beyond their salary and pecuniary incentives which are unavoidably lower than their private sector competitors (Ingraham, Selden, & Moynihan, 2000). Instead, they should focus more on strengthening the inclusiveness of their human resource management practices by making the process of work engagement a psychologically, cognitively, and physically safer experience for LGBT employees. In addition, for organizations to remain competitive in retaining talented employees, more efforts on understanding the drivers of job satisfaction, other than performance, are needed. More efforts on understanding how work engagement affects employees of biological and social minority and how leadership can reduce the potential gap in job satisfaction would further clarify their contributions in the workplace.
Several limitations of this study must be acknowledged that point to the direction for future research. First, despite the large sample size which helps control for many treats to validity (Yang & Kassekert, 2009), the cross-sectional nature of the data still poses concerns about causality. Although engagement literature generally supports the direction of relationship examined in the present study, several researchers also showed that job satisfaction can be an antecedent of work engagement. For example, scholars of organizational support theory (Eisenberger & Stinglhamber, 2011) argue that employees who are satisfied with their job through perceived organizational support develop felt obligation to 'return the favor' by engaging in their work. Thus, longitudinal data will be able to provide a more nuanced approach to determining their causal relationship.
The second limitation comes from using a secondary dataset. The scales for job satisfaction and work engagement for this study were developed post hoc. Although values for Cronbach's alpha below even .7 can realistically be expected and acceptable when dealing with psychological constructs (Kline, 1999), using proven scales (e.g., Schaufeli et al., 2002) would allow readers to compare with other studies and greatly enhance the understanding of engagement and job satisfaction relationship and the role of sexual orientation in the workplace.

Disclosure Statement
This research was supported by the National Research Foundation of Korea Grant by the Korean Government (NRF-2013S1A3A2055108).
Fleet expenses, whether through purchase, routine maintenance, or even repair are considerable for all levels of government annually. Considering the number of services disseminated through the use of automobiles, solid waste transport trucks, dump trucks, utility vehicles as well as the traditional fire, police and rescue vehicles, it is not surprising that many local governments are strategic in both determining the time of vehicle replacement as well as payment method. Since so many services are delivered on wheels, or even with tracks, payment disbursement can take place either through traditional operating budget requirements or even debt service. For smaller governments, this process is not as practical due to available funds, with replacements occurring only when there is a situation of exceptional need.
The traditional budgetary and cash management literature primarily focus on the basic aspects of costs inclusive of personnel and traditional supplies. Fleet financing is a critical element in not only budget formulation, but also in general public finance. Costs associated with initial purchase and operation can have an impact on not just the purchasing department, but can impact the revenue requests of several departments and in some cases, the entire government unit. Thus far, there has been very little literature discussing this topic.
This exploratory study examines the various methods associated with the financing of fleets among local governments in North and South Carolina in conjunction with appeasing service provider demands. Although there are numerous types of vehicles used to distribute services, the focus of this article is primarily automobiles and half-ton pickup trucks, including of SUVs and emergency medical services (EMS) vehicles. A primary model was constructed to determine the type of financing used by local governments for vehicle purchase, with additional models which tests need assertion by public officials. Preliminary findings indicate that staff with direct purchasing responsibilities have much influence in both financing methods as well as the number of vehicles purchased.
This study advances the literature in several ways. First, it provides an illustration of the finance practices associated with fleet purchases along with the type of benchmarks used to facilitate replacement decisions. Second, the study definitely advances the traditional budgeting literature by testing for the influence of various actors on actual purchasing decisions. Among the actors in the process, the 'gatekeeper' role of the manager and the finance officer are emphasized as they balance the need assertions of elected officials and staff while maintaining fiscal stability. The paper is organized as follows: the next section examines the literature surrounding models and methods for fleet replacement in addition to some of the pivotal actors in the budget decision making process. In the following section, a data and methods section provides a breakdown of the variables used for measurement followed by models testing for fleet financing methods and the number of vehicles purchased at a given time. Finally, there will be a findings section with discussion.

Previous Literature
The literature on government fleet purchasing methods is quite limited. Upon initial investigation, most studies focused on alternative models of determining optimal machine replacement with varying levels of demand, costs, numbers and types of machines purchased (Chen, 1998;Hopp, Zydiak, & Jones, 1993;Jones, Zydiak, & Hopp, 1991;Tang & Tang, 1993). More recent studies have added the elements associated with firm budget constraints along with the question of whether it is best to purchase or even lease (Hartman, 2000;Hartman & Lohmann, 1997;Karabakal, Bean, & Lohmann, 2000). None of the above studies were applied to a local government environment where the demand and service levels vary considerably, nor were they inclusive of numerous actors all vying for fleet replacement dollars with subjective interpretation of need. Moreover, many local government policies involving fleet replacement have basic numerical or even categorical values which determine replacement, and are based on type of use in service delivery. The policies all have some degree of basic life cycle costing indicators such as the optimal number of miles, type of service performed, reliability, maintenance costs, varying levels of fixed costs and safety issues (City of Gillette, Wyoming, 2011;Messera, 2007;Michel, Bell, Bronson, Owens, and Roylance, 2000). Ammon (2003) has been the only source to point out the three primary fleet financing methods used by local governments: cash, reserve funds, and debt financing. Many governments still purchase vehicles with cash, usually from unreserved fund balance monies. Purchasing vehicles with cash is advantageous for smaller governments with limited cash reserves, since the interest associated with borrowing is negated; however, the unexpected costs associated with a sudden purchase can offset efforts at maintaining expenditure smoothing trends, not to mention possible conflict due to the competition for funds from other areas.
A second financing method consists of earmarked revenue use, more commonly known as the vehicle replacement fund. With this application, local governments set aside a portion of undesignated revenue in a reserve fund to be used exclusively for fleet purchases. In some cases, all expenditures take place through the fund with annual replenishment while other uses consist of the fund acting as a buffer or as an internal 'line of credit' in case expenditures exceed expectations.
The third method is through borrowing. Methods associated with this strategy include bond issuance, bank financing, or even leasing (Ammon, 2003). For larger governments, this may be a preferred method due to the large amount of vehicles in or out of service, the ability to have a sustained line of credit, and more control over large budget fluctuations with the elimination of outright purchases. The major drawback to this option is continual payment along with the associated interest payments, especially if the local government has a low or no bond rating.

Applicable Theory
Traditional finance theory usually is not inclusive of specific equipment used in the distribution of services. However, these same services are part of departments and in some cases, specific programs. Thus, the bureaucratic expansion model provides much of the explanation for these finance decisions and practices (Craswell, 1975;Downs, 1967;Wildavsky, 1984). In this model, bureaucrats are continuing trying to increase departmental, agency, or even program budgets by imploring politicians to increase funding for their areas due to continual or expanding service demand. The dire consequences for insufficient funding are at times invoked, especially by emergency services personnel. Police and fire officials within municipal governments and sheriff departments at the county level frequently warn of substantial increases in crime and potential loss of life in the event of revenue losses. For example, a sheriff in Illinois recently cited budget cuts that could have been used for personnel as the reason for a jailbreak. According to the sheriff, the inmates recognized fatigue in officers that were working extra shifts due to a lack of personnel (Associated Press, 2015). The more traditional departments such as social services, environmental services, and planning departments, can substantiate requests due to service demands that result from changing demographics.

Budget Actors
There are several actors in the budget formulation and implementation process which can influence the purchase of various assets, including county fleet purchases. First, there is the elected body, which is the elected commission or council. In nearly every case, commissioners make the final decision concerning budget ratification. Previous studies have found that commissioners see themselves as very involved in the budget process and that service provision for the county as a whole has the highest priority versus individual departments or programs (Modlin, 2008). In rare cases, some elected bodies will actually examine individual account codes to determine if there has been excessive spending (Modlin and Stewart, 2014).
County officials elected to other offices have varying levels of influence on expenditures. For an elected Register of Deeds or a district attorney, county funding is usually routine and without much conflict; whereas, county sheriff departments utilize a substantial amount of equipment and provide many services which require transportation. During budget hearings, a sheriff can cite increased call volume, inmate transportation, detention center staffing, warrant distributions, increased levels of equipment usage, and technological changes as reasons to increase departmental budgets inclusive of additional vehicles. Preferences can also be based on quality and performance. In contrast, the finance officer, elected trustee or treasurer has the responsibility to ensure that the primary government unit has sufficient funds for many types of service provision while simultaneously maintaining substantial cash reserves.
County administrative staff usually bear the responsibility of providing aninitial version of the budget. The budget administrator or analyst (for larger counties that have these positions) get budget requests from department heads and other elected officials and present the requests to the finance officer who, if well experienced in the job or has intimate knowledge of the entire government unit, will prepare an initial version of the budget usually with the endorsement of the manager. Basically, the finance officer and the manager act as gatekeepers for local governments, trying to provide the best service provision as possible without overspending (Morgan, Robinson, Strachota, & Hough, 2015). At the state level, Goodman (2008) found that legislative and executive budget analysts have this role as well.
Modlin (2011a) found that county commissioners frequently endorse the county manager's proposed budget; this same study found that while county managers often agree with the suggested budget recommendations of department heads, managers also question the origins of revenue for any additional expenses. Obviously, there is inevitable conflict between the manager's budget recommendations and that of other elected officials, particularly the sheriff. Commissioners decide the outcome in these cases (Modlin, 2011a).
This study attempts to examine the various ways of financing vehicles used in county service delivery in North and South Carolina. Unlike other studies, this study actually examines empirical evidence demonstrating payment method as well as need assertion. A primary model will be predicated on a number of financial and institutional factors that may have a relationship with method of payment. An additional behavioral model is also introduced to determine how successful actors are at determining need and how this influence affects the number of purchases and financing.

Data and Methods
Examining fleet financing methods and need assertion required the solicitation of information from county government finance personnel. Surveys were sent to all county finance officers in North and South Carolina. After multiple rounds of dissemination, responses were received from 33% of counties with about the same percentage from each state. All counties in the survey were from professionally administered county governments in which there is the presence of a county manager/administrator. Among the responses, each financial classification of government is represented from those with budgets of less than $25 million to counties with budgets which exceeded $100 million.
The primary dependent variable is the primary payment method counties utilize for fleet, displayed as PURCHASE. The variable is coded "3" if the county pays for vehicles through multiple methods. Predictors used to test for purchasing method are both exploratory and from the literature. The presence of a vehicle replacement policy, given as VRP, in conjunction with how the need is determined, NEEDDT, a vehicle replacement fund, VRF, along with a requirement that there is specific cash available, CASHREQ, prior to purchase all had expectations of influencing purchasing method. Depending on quality perception, the resale value, RESALE, and the presence of a county garage, GARAGE, all are influential in both purchasing and activity costing schemes. The presence of a county garage enables counties to set up an internal service fund to account for indirect costs associated with vehicle repair (Modlin, 2011b). Finance officers were also asked about the make preferences for any official -elected or otherwise, MAKEPREF, and final decision maker, PURFINAL, concerning purchases. The size of the budget was also used as a variable BUDGET.
The secondary dependent variable testing the amount of influence in asserting need is VEHPUR, which is an ordinal variable based on the number of vehicles purchased at any one time. Finance officers were asked to provide a self-based interpretation of fleet needs by department heads, FINDEP, as well as the sheriff, FINSHERIFF. Finance officers were also asked to determine how well department heads influenced the manager or administrator concerning fleet needs, DEPCM, how well the sheriff asserted needs to the manager, SHERIFFCM, and how well the sheriff influenced the county commissioners/council, SHERIFFCOM. There was also an examination of make preferences for both the sheriff, MAKESHERIFF, and department heads, MAKEDEP. All of the information for the variables was obtained through survey data with the exception of budget information which was obtained from the North Carolina State Treasurer's  (2012). Table 1 provides a list of variables used in the analysis.
The absence in the literature of any form of fleet financing model has provided a prompt for a regression model which encompasses many of the factors associated with fleet financing practices. Based on the survey data as well as the observation of fleet transaction activity, ordered logistic regression models were created to examine these factors for possible relationships. In the first two models, the type of purchase method, whether exclusively by cash or some form of borrowing is influenced by the financial variables in the model and as described in table 1. A second model will test official need assertions and make preferences against the number of fleet purchasing methods.  8  3  1  4  Total  23  9 3 13 In the final two models, the predictors associated with need assertions and make preferences will again be used against the actual number of vehicles purchased to determine the effectiveness of official fleet desires. Below is the model which will test these assumptions.
The final decision for vehicle purchases was included in the final model for the purposes of including as many actors as possible in the model. The garage variable was included for two reasons. First, many larger counties have a fleet manager along with other personnel with significant responsibilities and input into purchasing decisions. The same can also be said for smaller counties with a county garage for its own fleet. Second, contracting out, especially to a dealer, can have very disparaging costs, thus influencing the number of vehicle purchases by decision makers.

Results and Discussion
Service provision among county governments is fairly comprehensive with varying levels of service. Finance officers were first asked about all fleet purchases despite the type of vehicle.
Overall, larger governments were in a much better position to purchase a variety of vehicles for service provision compared to smaller counties. The other interesting finding was the number of SUV purchases. In every budget classification, there were a substantial amount of SUVs purchased, especially among larger governments where the number of responses rivaled that of the traditional automobile. Governments that were not purchasing EMS, fire, or solid waste trucks were more than likely contracting out to a service provider. In the case of fire departments, donations are a primary source of funding. For smaller counties, this is a growing trend with both financial and personnel implications. Although not specifically discussed in this study, many county governments purchase backhoes, excavators, traditional bulldozers, and other heavy construction equipment. This equipment is primarily used for solid waste in landfills but has other uses as well, such as general debris cleanup including abandoned building/home demolition.
Financing methods for traditional automobiles including SUVs takes place in a number of ways, as illustrated in table 2. The majority of counties used a combination of methods which includes short-term borrowing from a local bank which had the asset capability to handle the capacity. Banks that handle county business provide a wealth of services that reduce float and enable more flexible cash flow (Modlin, 2014; Modlin and Stewart, 2012). Most counties that borrowed financed for three to four years with an interest rate under 3%. Overall, the amount of borrowing was generally between $100,000 and $500,000. However, approximately half of the counties in the study purchased vehicles outright with cash reserves, especially those with smaller budget sizes. This group was also more likely to utilize multiple methods of payment if necessary, as compared to the other groups. Only three counties stated that they used debt service for fleet payment.
The majority of counties in the study stated did not employ a vehicle replacement policy or a vehicle replacement fund (VRF). For the counties that had a vehicle replacement fund, the amount could be as much as $750,000. Interestingly enough, two of these counties had budget sizes of less than $25 million. An on-hand cash requirement was necessary for most counties prior to purchase. Table 3 suggests the average amount was approximately 200,000. Again, two counties in the smallest budget group required an on-hand cash balance of $1 million prior to purchase. Auto purchases were considered capital budget items for the majority of counties in the study.
The findings surrounding need determination produced some of the more interesting results. Of course, mileage was the primary determinant of replacement, followed by quality issues. The findings for this particular category had considerable variation and also somewhat of a state divide. No county in South Carolina provided a mileage number of less than 150,000 miles before replacement would be considered, with two counties requiring 200,000 miles prior to replacement consideration. In North Carolina, respondents which provided a mileage number cited between 125,000 and 150,000 miles, with just one county stating 200,000. In addition to mileage, performance issues surrounding repair were also a major reason for replacement. One county finance officer stated that if the repair costs exceeded 50% of National Automobile Dealers Association (NADA) value of the automobile, then replacement was considered. Resale values were usually up to 25%, with autos designated as surplus disposed of by internet or public auction. Some counties rotate vehicles between departments in order to extend service life. For instance, older automobiles within administration may be sent to departments which require less travel in order to get additional service.
Administrators played an important role in purchase decision making. In nearly half of the counties surveyed, there was a combination of elected officials and the county administrator/manager with the manager having primary responsibility in ten counties. For the most part, this took place in larger counties. Finance officers had marginal assessments of sheriff and department head vehicle needs; however, the sheriff had a slightly higher impact on asserting need to both the manager and commissioners (table 3). For make and model preferences, it was fairly minimal for department heads, but county sheriffs were very opinionated. For counties that cited a make and model preference, it was evenly divided between the Dodge Charger and the Ford Interceptor, with other specialized requests including Chevrolet Tahoes.
An ordered logistic regression model was created to test institutional independent variables against purchase methods and the number of vehicles purchased. A link test was performed to ensure that there was no specification error or a nonlinear combination of independent variables. In the first purchase model, the presence of a vehicle replacement policy (VRP) provided the strongest relationship with purchasing method. According to the odds ratio, counties with a vehicle replacement policy were seven times more likely to use multiple methods of purchasing inclusive of debt service. The vehicle replacement policy represents a more standardized way of purchasing indicating less probability of a request to delay these debited encumbrances. The other significant variable within the model was resale or salvage value. Counties that received low resale value (usually 25% or less) were more likely to use alternative methods of payment. The vehicle replacement model had additional significant findings. Larger counties were the major beneficiaries. For every unit increase in budget size, counties were three times more likely to purchase anywhere between 1-5 additional vehicles. Additionally, as the number of methods used to establish need (NEEDDT) decreased, the number of vehicle purchases increased, but the residual odds ratio of less than 1 indicates that this relationship has modest implications on actual purchases compared to using additional indicators for replacement. The other surprising finding is how the presence of a county garage led to increased vehicle purchases. The ability of a county to diagnose, repair, and maintain a fleet provided some explanation for increased purchases. Considering the indicator, the McFadden's Pseudo-R² demonstrated an acceptable level of variance between vehicle purchases and the predictors. The model was also significant at the .05 level.
The next two models focus on finance officer interpretation of need and the level of influence of other officials (table 5). In Purchase Model 2, higher levels of agreement with the sheriff concerning assertion of need led to increased methods of financing. In this case, the odds ratio is high enough that it suggests finance officers can play a pivotal role in the initiation of intended vehicle purchases through encumbrances. Again, the presence of a county garage was significant in purchasing methods. This time, the study is attempting to capture the assertions of fleet managers, shop supervisors, and even mechanics.
The opinions and assertions of officials had significant levels of influence on the number of vehicles purchased. The dynamic between the sheriff, the county manager, and the commissioners produced the most interesting findings. If the sheriff was successful at convincing the county manager that vehicles were needed, the odds of obtaining additional vehicles increased 23 times; however, this same assertion did not appear to exist with the commissioners. The findings indicated less budgetary assertion by the sheriff leads to an increase in vehicle purchases. It appears to be more advantageous for department heads not to assert requests concerning vehicle replacement. More vehicles are usually ordered when department heads do not have a preference. Department heads may be the beneficiaries of surplus vehicles in many of these cases. A test of the full model with all 9 predictors against a constant-only model was statistically reliable χ 2 (9, N=48) = 25.87, p < .005, indicating that the predictors, as a set, reliably distinguish between the different levels of vehicle purchases.
Statistical findings among the models provided some explanation for fleet purchasing methods and need assertions, but there are additional conclusions that can be drawn from more isolated responses. On the positive side, the findings indicated that many counties want to obtain a maximum amount of use within the automobile fleet, especially in South Carolina, where some counties do not consider replacement until the 200,000 mile mark. Even in the abovementioned NADA response associated with repair cost, an eight-year-old Ford Crown Victoria with 150,000 in good condition would need substantial engine, body, or transmission damage in order to exceed 50% of its NADA value. In addition, counties not only are avoiding debt in purchases, but also attempt to isolate costs with the use of county-owned garages.

Additional Findings
In-depth responses from finance officers provided indication that some officials are not overly concerned about the cost aspect associated with vehicle purchases and maintenance. Responses indicated requests for full-size vehicles such as Chevy Tahoes, usually for sheriff departments. In many of these cases, it is definitely a "want versus need" situation. There have actually been requests to use these same vehicles for K-9 transportation, which severely accelerates the depreciation of a vehicle that is initially priced near $50,000 without extra wiring for law enforcement purposes. Some requests included one-ton chassis pickup trucks. Brand new trucks in this category have a payload capacity in excess of a ton and the dually models can tow in excess of 20,000 lbs., with some manufacturers boasting 30,000 lbs. 1 Very few counties have this kind of need capability on a regular basis. Pickup trucks with a half-ton chassis can satisfy more than 95% of daily service demands. Dump trucks and other heavy use trucks can easily be modified with a compatible receiver combination to satisfy additional demands. However, few vehicles can compare to the excessive costs associated with the acquisition of a Hummer-based vehicle. Caution should be exercised when higher levels of government provide such a vehicle to a local government at no cost. The break-even analysis for this vehicle fails very quickly. The gas mileage is usually just as poor as the one-ton pickup (around 10 miles per gallon), not to mention a very limited use capability. Even with frequent use, the gas usage creates a major performance issue in addition to insurance costs and the problem of finding parts, especially for chassis and other drive train related issues.

Conclusion
This exploratory study examined the fleet purchasing practices of county governments in North and South Carolina. In addition to examining method of payment, the study also examined the level of need assertion by county officials. Findings indicated that when counties decided to make a purchase, most chose direct cash payments, especially counties that received low salvage values. More vehicles were usually purchased among counties with additional fiscal capacity and if there was a county garage that could be used for repairs. If mileage was a primary indicator for replacement, only 1-5 vehicles are purchased.
The influence of the actors provided some of the more interesting findings. A significant relationship was found between finance officer's assessment of sheriff needs and multiple financing methods, as well as the presence of a county garage. However, vehicle purchasing activities had mixed results. When the sheriff was successful at communicating needs to the county manager in addition to a make preference, the odds of additional vehicle purchases increased 23 times. Conversely, this number decreased when the sheriff appealed to commissioners. Overall, the sheriff appeared to be much more successful if there was a good working relationship with staff, especially the county manager. While there was a negative relationship between a sheriff's assertion to commissioners and vehicle purchases, commissioners tried to adhere to the requests and more than likely directed the manager to satisfy as many requests as possible. The findings associated with sheriffs can easily influence the department head findings. If county staff are spending additional time with the financing and purchasing of the law enforcement fleet, it becomes very challenging to address additional fleet needs. In this study, the only unilateral finding was that the sheriff's departments had the highest vehicle turnover compared to other departments.
There are also some limitations to the study from both an institutional standpoint as well as from a behavioral standpoint. First, these are professionally administered county governments with county administrators that place some emphasis on spending restrictions. In other forms of governments, such as commission, decision making by elected officials that have parochial responsibilities (such as public works) and do not have any fiscal responsibility, are not prioritized by fund balance requirements and substantial cash reserves. Second, there are many personnel characteristics which influence need assertion. Employee position, tenure, comprehensive knowledge of government needs, and relationships with other officials, especially those involved in the final decision concerning fleet purchases, all have influence in the number of vehicle purchases. Third, the latitude in purchasing decisions also plays a role. For instance, if county ordinances require drug seizure monies to be debited to the general fund, this revenue source can be distributed to several areas; whereas, if the sheriff controls this revenue stream, spending becomes a unilateral decision.
This study has also supported findings from previous research. First, the request for vehicle replacement is not that unusual, but the more specialized request for heavy duty vehicles, especially SUVs, does add credibility to the bureaucratic expansion model (Craswell, 1975;Downs, 1967;Wildavsky, 1984). These vehicles are larger and more expensive than what would have been purchased twenty years ago. Second, the findings also indicated that finance officer rate of need assertion is lower than the amount of assertiveness placed on commissioners and the manager by the sheriff indicating that these administrators do indeed act as gatekeepers for public finances verifying previous writings (Morgan et al., 2015). The findings also verify the amount of influence the county manager has on overall policy decisions not to mention budget directives (Modlin, 2011b). In this study, county elected bodies turn to the manager to find ways to satisfy requests and simultaneously maintain a sufficient fund balance.
Overall, there are optimal solutions that can sustain many arguments. It is advantageous for county fleets to be representative of the population if possible. For instance, if the average taxpayer operates a vehicle approximately five years old, the fleet should be somewhat representative of that figure as well. To account for many of the life cycle costing issues surrounding the operation of a vehicle for more than 150,000 miles, a policy adoption of engine and transmission replacement at this interval versus total vehicle replacement would provide additional years of usage with a major reduction in direct costs associated with those vehicles.
In any case, with the advancement of on-board diagnostics to assist with automotive repair, each county should be receptive to the 200,000 mile standard with finance officers continually examining various fleet costs for improvement.

Disclosure Statement
The author(s) declare that there are no conflicts of interest that relate to the research, authorship, or publication of this article.

Survey of County Finance Officers
For each question, please mark your response unless otherwise stipulated (questions are related to the most recent purchases of county fleet)

1.
Does your county currently have a vehicle replacement policy?

2.
How is automobile need primarily determined (Please check the most common response or rank answers)? Excessive mileage/age (please identify approximate mileage number C d ) Poor performance (Engine, transmission, suspension, gas mileage) Not enough vehicles for service Vehicles not large enough/cannot meet certain specifications Individual just requesting (no legitimate reason). Other (please specify)

3.
Are vehicle purchases considered capital project items? No Yes, for some departments (please identify) Yes, for all departments

4.
When obtaining vehicles, how are they purchased? Outright on a cash basis Installment purchases/Funds borrowed from primary bank Debt Service Other (please specify)

5.
Do you have a vehicle replacement fund? No Yes, with an amount of less than 100K at the beginning of the fiscal year. Yes, with an amount between 100-500K at the beginning of the fiscal year. Yes, with an amount between 500-750K at the beginning of the fiscal year. Yes, with an amount between 750K-1M at the beginning of the fiscal year. Yes, with an amount of more than 1M at the beginning of the fiscal year.

6.
How much cash do you require to be on hand prior to a purchase? Less than 100K 100-500K 500-750K 750K-1M More than 1M

7.
Does the make/model make a difference on purchasing requests? Yes-(Please explain) No

8.
In general, is there a specific request by department heads for a particular brand and/or model? Yes-(Please identify) No

9.
Is there a specific request by the sheriff for a particular brand and/or model? Yes-(Please identify) No

11.
On average, how would you rate department head vehicle need assertion based on a scale of 1 to 5, with 5 being considered a real need as determined by you.
1 2 3 4 5 Not Really Needed At All Really Needed

12.
On average, how would you rate the sheriff's vehicle need assertion based on a scale of 1 to 5, with 5 being considered a real need as determined by you.
1 2 3 4 5 Not Really Needed At All Really Needed

13.
On a scale of 1 to 5, with 5 being the highest level of influence, how influential are department heads at convincing the county manager that replacement vehicles are needed? 1 2 3 4 5 Not Influential At All Very Influential

14.
On a scale of 1 to 5, with 5 being the highest level of influence, how influential is the sheriff at convincing the county manager that replacement vehicles are needed? 1 2 3 4 5 Not Influential At All Very Influential

15.
On a scale of 1 to 5, with 5 being the highest level of influence, how influential is the sheriff at convincing the commissioners that replacement vehicles are needed? As the American population has become increasingly diverse in the past few decades, minority civic participation, including volunteering, has gained growing attention among scholars and practitioners. Understanding the factors that encourage minorities to give their time to organizations or to friends and family is critical to a country that wishes to foster voluntary participation and to maintain its strong civic tradition. Asian Americans are one of the fastest growing minorities in the United States. From 2000 to 2010, the Asian population increased by 46%, reaching 14.7 million, or 4.8% of the total population in the country (U. S. Census Bureau, 2000Bureau, , 2010. Much of this growth comes from foreign-born Asian immigrants. Close to 62% of single-race Asian Americans are foreign-born or immigrants, who account for approximately 28% of the total foreign-born population in the U.S. (Grieco & Trevelyan, 2010). Additionally, Asian Americans are the most culturally diverse minority group in the country. Among the over twenty Asian-American subgroups, the Chinese, Asian-Indian, Filipino, Vietnamese and South Korean are the largest, each with a population over one million.  Wang, 2007). Most of the existing studies on Asian American volunteering are qualitative and regional (Smith, Shue, Vest, & Villarreal 1994, 1999Smith, Shue, & Villarreal, 1992). While these qualitative studies provide rich accounts of Asian Americans' experiences of volunteering and motivations to volunteer, their small sample size and lack of systematic comparison across Asian subgroups prevent a generalization of the findings.
Sundeen, Garcia, and Wang (2007)  Additionally, Chinese and Filipino volunteerism positively associates with their household income, while that of Asian Indians positively relates to their educational attainment. These findings suggest that future studies of Asian volunteering should consider examining Asian subgroups separately, as their diverse cultural backgrounds may shape their volunteering behavior in unique ways.
Due to insufficient sample size, Sundeen et al. (2007) did not examine the volunteering activities of other major Asian subgroups, such as immigrants from South Korea and their native counterparts. Korean Americans are notably different from other Asian-American subgroups in their culture, religious belief, language and other socioeconomic characteristics. For example, the majority of Korean Americans in the United States are Protestants (61%), while most Filipino Americans are Catholic (65%), about half of Asian Indians are Hindu (51%), and half of Chinese Americans (52%) describe themselves as religiously unaffiliated (The Pew Forum on Religion & Public Life 2012). Given these differences, it is unclear whether Korean Americans would share a similar pattern of volunteering with any of the Asian subgroups that Sundeen, Garcia and Wang studied: for example, the types of organizations for which they volunteer ; how personal, social and cultural resources shape their volunteering; how many hours they volunteer, etc. This study will expand our knowledge of Asian Americans' volunteering by focusing on Korean Americans, a subgroup left out in Sundeen et al. (2007).
Additionally, Sundeen et al. (2007) did not examine the influence of religiosity on volunteering, as the dataset they used had no information on respondents' religious beliefs and religious attendance. Numerous studies on immigrant and minority philanthropy, however, have shown that religiosity significantly influences the volunteering behavior of African Americans, Hispanics, Asian Americans and immigrants (Ecklund & Park 2005;Wang, Yoshioka, & Ashcraft, 2013). Given the importance of religion and religious attendance in Korean Americans' lives, we will extend the literature by examining how religiosity shapes Korean Americans' voluntary participation in community service.
Most of the existing empirical studies on Asian American volunteering (Eckland & Park, 2005;Sundeen et al., 2007) examined formal volunteer activities (doing unpaid work in the context of a formal organization), but did not include informal volunteering (doing unpaid work for the well-being of family, friends, and the community at large). While ethnic minorities may volunteer less for formal organizations, they are more likely to contribute their time to help friends, families and neighbors (Carson, 1999;Wang, 2011). Therefore, it is significant to study informal volunteering to gain a better understanding of ethnic minorities' behavior.
A recent study by Lee and Moon (2011) examined Korean immigrants' volunteerism for ethnic and mainstream organizations in the United States. The study did not separate religious and secular volunteering, which could be considerably different between minority immigrants. As Handy and Greenspan (2009) found, Canadian immigrants were more likely to volunteer for religious congregations that are ethnically homogenous than for secular organizations that are more diverse. Consequently, we will extend the literature by examining whether Korean Americans are more inclined to volunteer for religious or secular organizations. In addition, we will study how decisions to volunteer for each type of organization are shaped by the level of acculturation and other personal characteristics and social resources.
Using data from the Johnson Center for Philanthropy's (2009) Charitable Giving and Volunteering among Korean Americans Survey, this study bridges the gap in the literature by exploring the informal and formal volunteering activities of Korean Americans and the influence of acculturation on their decisions to volunteer. Specifically, we seek to answer the following questions: 1) Do Korean Americans participate more in secular and religious than in informal volunteering activities? 2) To what extent does acculturation affect Korean Americans' secular, religious and informal volunteering, respectively? 3) What other factors determine Korean Americans' likelihood of secular, religious and informal volunteering? In the next section we review the literature and develop a theoretical model on the ways acculturation and other socioeconomic factors structure Korean Americans' formal and informal volunteering.

Literature Review and Hypotheses
An individual's decision to volunteer, giving time freely for the benefit of another person, group or cause, and the level of voluntary involvement are jointly influenced by his or her socioeconomic, cultural and human capitals (Wilson, 2000). The extant literature of minority volunteerism indicates that minorities and immigrants have a different propensity or level of voluntary participation compared to their majority or native counterparts (Sundeen et al., 2007;Sundeen, Garcia, & Raskoff, 2009;. This could be attributed to various resources, such as social networks, income and education, which encourage minorities and immigrants to volunteer formally and informally.

Acculturation
Acculturation is one of the most important explanatory factors in the study of ethnic minority populations, including their volunteerism (Sundeen et al., 2007;Sundeen et al. 2009;Tucker & Santiago, 2013). It reflects the extent to which individuals acquire the values, behaviors, lifestyles` and language of the host culture (Zane & Mak, 2003). Numerous measures have been developed to assess behaviors and attitudes related to acculturation, including language use: preferences and proficiency; social affiliations; daily living habits; cultural traditions and customs; perceived prejudice and discrimination; cultural identification and pride; and generational status (Berry, 1997;Fletcher, Campbell, & Fast, 2007;Tucker & Santiago, 2013).
The acculturation literature maintains that ethnic groups possess distinctive cultures (values, beliefs, norms and attitudes) and exhibit distinctive behavioral patterns. When ethnic minorities and immigrants encounter the mainstream culture, they experience varying degrees of cultural conflicts. The greater the cultural distance (that is, the dissimilarity of the two cultures in language, religion, etc.), the greater the cultural conflict and the greater the need of cultural learning and adjusting (Berry, 1997). Language proficiency facilitates the cultural adaptation and learning process. Research on minority and immigrant volunteerism shows that inability to communicate fluently in the host language acts as a principle barrier to discourage participation in volunteer or civic activities (Fletcher et al., 2007;Tucker & Santiago, 2013;Wang & Handy, 2013). Additionally, the perceived cultural distance and language barrier could influence the types of organizations that attract minorities and immigrants to work for them. Portes and Sensebrenner (1993) found that immigrants were motivated to establish their own community organizations and participate in ethnic-oriented secular or religious groups that provide them a sense of belonging and a degree of familiarity by interacting with people from the same cultural background or speaking the same language. Being connected to ethnic-oriented groups also helps minorities and immigrants build social networks that could offer informal help or provide information on jobs and other opportunities (Lee & Moon, 2011). As a result, immigrants, particularly those who have limited language proficiency, find more comfort volunteering for ethnic-oriented organizations or religious congregations in the early stage of migration (Handy & Greenspan 2009). According to the Asian American Legal Defense and Education Fund (2013), compared to other ethnic groups surveyed, Korean Americans have the highest level of limited English proficiency. Approximately 67% of Korean Americans speak English less than "very well," compared to 37% of all Asian Americans nationally. In view of the potential impact of the language barrier on volunteering and the overall level of English proficiency of Korean Americans, we expect:

H 1: Korean Americans who speak English at home to be more likely than their counterparts to engage in formal secular and religious volunteering, but not in informal volunteering.
Acculturation experiences also differ by generation and citizenship. First-generation immigrants have been socialized to their own culture of origin and must adapt to the culture of the host society after immigration (Zane & Mak, 2003). U.S.-born ethnic minorities, in contrast, are exposed to a multicultural context beginning from birth that comprises the parents' ethnic culture and American culture. Their values, life styles and behaviors, including volunteer work, would be similar to those of the mainstream culture. In a study of Asian American volunteering, Sundeen et al. (2007) find that second-generation native-born Asian Americans, Chinese, Asian Indians, Filipinos and naturalized Asian immigrants, as well as immigrants from India and the Philippines, are more likely to volunteer than other non-naturalized first-generation Asian immigrants. We anticipate this finding will apply to Korean Americans as well. Therefore, we posit: H 2: Korean Americans who are citizens or second-generation immigrants are more likely to engage in formal volunteering than noncitizen first-generation immigrants.
Cultural identity, practices of cultural traditions and the cultural stress of respondents also indicate the level of acculturation. Wang and Handy (2013) found that cultural identity affects Canadian immigrants' participation in local organizations. For Korean Americans we expect those who identify themselves as Korean, who practice Korean traditions and holidays more often, and who experience stronger cultural conflict to be less likely to engage in formal volunteering and more likely to volunteer informally to help family, friends and neighbors. Therefore, H 3: Korean Americans who identify themselves as Koreans are less likely to engage in formal volunteering than their counterparts.

H4: Korean Americans who practice Korean traditions and often observe Korean holidays are less likely to engage in formal volunteering than their counterparts.
H5: Korean Americans who experience stronger cultural conflict are less likely to engage in formal volunteering than their counterparts.

Religiosity
Religion shapes people's values and beliefs. It is one of the main institutions that foster American civic life. Since most religions encourage believers to care for others and give to the needy, they can promote pro-social behaviors, such as helping, volunteering in local communities, and charitable giving (Cnaan, Kasternakis, & Wineburg 1993;Musick, Wilson, & Bynum, 2000;Wilson & Janoski, 1995). Furthermore, religious congregations provide a venue for volunteering. Handy and Greenspan (2009) found that immigrants who have been members of a congregation longer and who attend service more often are more likely to volunteer and to volunteer more hours. Ecklund and Park (2005) found that religious attendance significantly increases community volunteering among Asian Americans. Moreover, Protestant Asian Americans are more likely to volunteer than their non-religious counterparts. Since over 60% of Korean Americans are Protestants (The Pew Forum on Religion & Public Life, 2012), we expect those with this religious affiliation to be more likely to volunteer for both secular and religious organizations.
H 6: Korean Americans with religious affiliations are more likely to volunteer for both secular and religious organizations.

Control Variables
An individual's socioeconomic characteristics, including educational attainment, household income and homeownership, determine that person's social status. The dominant status theory argues that social status is positively associated with voluntary-organization participation (Smith, 1994;Wilson & Musick, 1997). Educational attainment is the strongest and most consistent predictor of volunteerism among ethnic-minority groups, immigrants and the general population (Smith, 1994;Wilson, 2000). For minorities and immigrants education in the United States provides opportunities to become integrated into the mainstream culture, to learn about the values of civic engagement and to practice volunteering (i.e., through school-based community service). Accordingly, immigrants who graduated from American educational institutions are more likely to engage in formal volunteering than those who completed the highest degree in their home country. Statistics show that Asian Americans have the highest educational attainment of any racial or ethnic group in the United States. Close to 90% of Asian Americans graduated from high school, and 52% have a bachelor's degree or an advanced degree, a rate much higher than that of whites (33%) and other minority groups (17%) (U.S. Census Bureau, 2011), and the highest rate among all foreign-born racial/ethnic groups (Crissey, 2009). Among Asian subgroups Korean Americans have a higher percentage of college graduates than some other subgroups, such as Cambodians and Hmong (Zhou & Kim, 2006). The overall high educational attainment indicates great potential for Asian Americans, including Koreans, to engage in formal volunteering.
Household income tends to be positively associated with voluntary participation among the general population (Smith, 1994), as higher income provides a sense of financial stability that allows or encourages people to volunteer or contribute their time for free. In a study of Asian-American philanthropy, Chao (2001) argues that Asian-American immigrants go through three stages of philanthropic activity: the survive stage, the help stage and the invest stage. They become more involved in mainstream philanthropy by helping and investing as they gain more financial stability and economic security. We extend this argument to Korean Americans' decision to volunteer. Korean Americans with higher household incomes could be more likely to volunteer as they become less constrained by the need to earn money. Data suggest that Asian Americans overall, but particularly Koreans, have greater financial resources to support their decisions to volunteer than other minorities. Over 60% of Asian American households have an annual income of $50,000 or more, compared to 54% of non-Hispanic white households and 36% of those of other racial/ethnic groups (U.S. Census Bureau, 2011). However, in terms of the level of involvement, Freeman (1997) finds that the higher the income, the lower the number of hours volunteered in general. The negative relationship between income and volunteering can be explained by the rational-choice theory, which views volunteering as the opportunity-cost of working. People with higher income are likely to volunteer for fewer hours, since their opportunity-cost of giving up work is high. This applies to Korean Americans as well.
Homeownership not only indicates a person's wealth, but also reflects the degree to which one is integrated into the community and has a stake in its amenities, services and general quality of life (Smith, 1994). Homeowners are inclined to live in a community longer than renters, which helps them develop a sense of connection and, build social networks, while increasing their chance of being asked to volunteer by either friends or formal organizations (Rotolo, Wilson, & Hughes, 2010).
Family is an important socialization forum for various aspects of civic engagement, including volunteering, and we found multiple perspectives to understand the impact of family status on volunteering decisions. For example, Wilson and Musick (1997) reported that being married and having children increase the chance of volunteering. People with children may learn about volunteering opportunities through their interaction with other parents in the neighborhood or with schools and organizations that serve youth and children. This is particularly important for immigrant parents, who may have limited social connections to other types of formal institutions . The volunteerism literature also shows that married people are more likely to volunteer than single people (Wilson, 2000). Nonetheless, single people without children may volunteer more hours since they have greater discretionary time for non-family activities and may want to develop social networks through formal volunteering (Sundeen, 1990). Likewise, married Korean Americans, particularly those with children, could be more likely to volunteer, while single Korean Americans without children could volunteer more hours.
Employment may broaden Korean Americans' social connections and thus increase their chances of being asked to volunteer, as a growing number of corporations encourage and organize their employees to volunteer for local nonprofits. Still, employment could also reduce the free time available for unpaid voluntary work (Markham & Bonjean, 1996;Wilson, 2000). Sundeen et al. (2007) find that Asian immigrants, employed part-time, are significantly more disposed to volunteer than those who are unemployed, while full-time employed Asian immigrants are similar to the unemployed in their volunteering propensity. Therefore, in this study we control for employment status in our examination of Korean Americans' volunteering. The gender difference in the inclination to volunteer may vary across cultures (Gaskin & Smith, 1997). Dominant status theory argues that individuals who are characterized by a more dominant set of social positions and roles are more likely to volunteer (Smith, 1983). In a traditionally male-dominant culture, like that of many Asian countries, women may be encouraged to take a domestic role or to volunteer only for certain types of organizations, such as those serving youth or disadvantaged populations, but not for political activities. In a study of Asian-Indian volunteerism, Kurien (2002) finds that in some Pan-Indian voluntary organizations, like the Indian Association of North Texas, the majority of individuals who had served on the board were men, which reflects the male-dominant culture of the Asian-Indian community. However, Sundeen et al. (2007) find that Asian-American females and Filipino females are more likely to volunteer, but overall the gender differences are not significant among Chinese, Asian Indians or other Asian immigrants. Studies also show that formal volunteering peaks at an individual's middle age. In this study, we control for both gender and age.

Formal Volunteering and Informal Volunteering
Informal volunteering is more personal and is not organized, and the donors and recipients are likely to already have a relationship that results in an obligatory response. In this situation helping behavior is dependent more on factors such as opportunity, resources and ability (Wilson & Musick, 1997). The literature on volunteering suggests that ethnic minorities and immigrants are more likely to volunteer informally than to volunteer for formal organizations (Wang et al., 2010). Sundeen et al. (2007Sundeen et al. ( , 2009) recommend continued research on informal volunteering, acculturation and other demographic variables on ethnic and immigrant groups in order to fully understand the conceptual framework of volunteering. For Korean Americans, we expect acculturation factors to impact the level of activity in informal volunteering that helps family, friends and neighbors.

Data and Methods
Data for the empirical test were derived from a web-based survey of Charitable Giving and Volunteering among Korean Americans conducted in 2009. This web survey was administered by the newspaper Korea Daily and Joong Ang Broadcasting Corporation, a major Korean-American news media organization in California. It was first posted on the main page of the Korea Daily website (www.koreadaily.com); newspaper (both web and print) and radio advertisements followed. The survey targeted Korean Americans residing in California and was administered both in English and Korean to minimize the impact of language difficulty on our sample. It was online for a month in January 2009; we obtained a total of 1,505 responses. After removing responses from outside California and cases with substantial missing data, our final sample for this study included 769 Koreans who are legal residents of California, including 427 (55%) immigrants with U.S. citizenship, 329 (43%) with permanent residency, and 13 (2%) with the permanent residency equivalent investor visa (EB-5).
The questionnaire included extensive questions about Korean Americans' philanthropic giving and volunteering. We address volunteering in three major philanthropic areas: (1) formal, secular nonprofit volunteering; (2) formal, religious nonprofit volunteering; and (3) informal ethnic-based volunteering. The respondents were asked whether they engaged in any of the three forms of volunteering in 2008 and the hours volunteered per month for each area. As shown in Table 1, 50 % of respondents participated in formal, secular volunteering, 70% participated in religious volunteering, and 65% volunteered informally for friends and relatives. The survey also asked questions regarding multiple dimensions of acculturation: cultural identity, generation, religion, education, occupation, marital status, gender, age, annual household income, housing situation, and so on.
We employed binary probit models to predict whether Korean Americans participated in secular, religious or informal volunteering activities in 2008. Volunteering in the three philanthropic areas is coded 1 if the respondent volunteered and 0 if not. For more meaningful interpretation of the results, we also used the Clarify program to produce predicted probability of each of the statistically significant independent variables (Tomz, Wittenberg, & King, 2001). 1 For expository purposes we focus on four significant acculturation variables (Language difficulty, Korean identity, Korean tradition, and Citizenship) from the probit model and illustrate the effect of each variable on the secular volunteering in figure 1.

Independent variables
Acculturation. Acculturation is a key explanatory factor in our research; we measured this social concept with multiple indicators including language difficulty, stress from American culture, Korean culture identity, ethnic traditions, and generational and citizenship status. Language difficulty is an indicator of the degree of difficulty in communicating in English in any aspect of the respondents' lives. Cultural stress, measured by examining three survey questions (see table 1), is an indicator of stress from living in the U.S. because of its foreign customs, culture and social norms. The Confirmatory Factor Analysis shows a high internal consistency of these three questions (Cronbach's Alpha=.71). Korean identity and Korean tradition are scale variables obtained from survey questions, "How much do you identify yourself as a Korean?" and "How important is it for you to preserve Korean culture?", with the answer to each question measured on a five-point scale. U.S. Citizenship affects the level of acculturation into American society and is measured dichotomously: if he/she is a U.S. citizen, coded 1, otherwise, 0. Generational status is a dichotomous variable measured by asking the question "Which generation of Korean-Americans are you?" Respondents who in the second generation-those who were born in the U.S. with at least one parent born outside the U.S.-and in the 1.5 generation-those who were born outside the U.S. and immigrated as a minor to the U.S.-are both coded 1, and those in the first generation-those who were born outside the U.S.-are coded 0. Religion is measured by three dichotomous variables: Protestant, Catholic and Buddhist. Each is coded 1 if a respondent self-identifies as Protestant, Catholic, or Buddhist, and 0 otherwise. This research controls for the socioeconomic status of respondents. Education is included as an eight-scale nominal variable, and measure of Education includes education from Korea as well. Age is measured as an ordinal variable with six categories (see table 1). Household income is measured as a seven-scale ordinal variable. Employment status in 2008 is coded 1 if the respondent was employed full time and 0 otherwise. Homeownership is coded 1 if a respondent was a homeowner and 0 otherwise. Family structure is added by asking whether the family has any children. This family variable Children is coded 1 if the respondent has one or more children, and otherwise coded 0. We also included two dichotomous variables, Marital status, with 1 being married, and Gender, with 1 being male. A possible multicollinearity issue among independent variables has been tested by measuring Variance Inflation Factors (VIFs). The measures of VIFs for our variables are all quite moderate (the largest is 2.91). Thus, there is no evidence of serious multicollinearity in our model. Table 2 provides the results of binary probit analyses regarding participation in volunteering activities in three philanthropic areas: secular, religious and informal. It is noteworthy that the measures of acculturation matter for participation in secular volunteering activities. Additionally, the estimation results reveal that religion, education, employment and age all influence Korean Americans' secular and religious volunteering. We find that difficulty in English communication (language difficulty) is negatively associated with the likelihood of secular volunteering. For relationships that are statistically significant, table 3 reports predicted For example, the probability of secular volunteer participation is .38 if a respondent expresses the highest level of language difficulty, compared to .52 when a person experiences a lower level (1 st quartile) of language difficulty. As predicted, Korean Americans who identify themselves strongly as Korean (Korean identity) are less likely to participate in secular and religious volunteering. The probability of secular volunteering slowly falls from .53 to .44 when Korean identity increases from its 1 st quartile up to its median point. It is noteworthy that the Korean tradition indicator, which measures how important Korean tradition is to a respondent, is positively associated only with participation in informal volunteering activities, not with secular or religious formal volunteering. Moreover, the binary results indicate that a respondent who is born in the U.S. or immigrated as a minor to the U.S. presents a higher probability (.59) to volunteer for secular organizations than a person who stays in the U.S. temporarily (.47). This positive association is consistent with our acculturation hypothesis.

Results
As expected, Protestant and Catholic Korean Americans are more likely to engage in secular volunteering, and all three religion variables (Protestant, Catholic and Buddhist) show significant positive likelihood of religious volunteering. The probability of secular volunteering rises .02 if a responder is a Protestant Christian (see table 3). Furthermore, the binary probit results suggest that most of the control variables, such as Education, Age and Income, are positively associated with voluntary participation, especially in secular nonprofits, which is consistent with our expectation. Interestingly, the negative association of Employment status reveals that Korean Americans employed full-time are less likely to volunteer for secular philanthropy. Full-time employment reduces the probability of secular volunteering by .10 (see Table 3). Our findings also revealed that Children and Gender make a significant difference in participation only as concerns informal volunteering. We show that families with no children are more likely to help people informally, compared to families with one or more children. The negative estimate of the Gender variable suggests that female Korean Americans are more likely to volunteer for family and friends, than are male Korean Americans. The probability of informal volunteering rises if a respondent is a female (.77) rather than a male (.59). Figure 1 shows an increase in Language difficulty from the 1st quartile (.086) to the maximum (.061) reduces the probability of secular volunteering by .025. Figure 1 also shows that Citizenship decreases the probability of volunteering in secular philanthropy by .019.

Discussion and Conclusion
This research empirically examines the volunteer behavior of one Asian-American subgroup in the United States. The respondents of Korean-American descent were asked to complete a webbased survey with initial contact by a major Korean-American news media organization. The assumption is that the data reflects the Korean native-born and immigrant population residing in California. Despite their limitations, the data provide a comprehensive insight into the personal, social and acculturation factors of recent immigrants and native-born citizens that influence their secular, religious and informal volunteering.
The findings indicate three distinct types of volunteering behavior for Korean Americans. Formal volunteering, including secular and religious, as expected, is strongly associated with acculturation factors and with personal and social variables such as religion, education, age and employment status. Informal volunteering, on the other hand, appears to be independent of and not complementary with the other two types of volunteering. Acculturation, religion and most of the other control variables did not have an association with informal volunteering behavior. Age and a strong desire to preserve Korean traditions had a positive influence, while male and number of children negatively impacted informal ethnic volunteering. Finally, religious volunteering appeared more similar to secular volunteering, with the socio-demographic factors of religiosity and education being significant predictors for both.
In regards to secular volunteering, the significant association of acculturation, as the process by which individuals change in adapting to demands of a new environment, suggests implications for nonprofit organizations seeking ethnic participation (Berry, 1997). Lack of English-language proficiency and strong identification with Korean culture result in less secular volunteering. Programs of language education, continuing adult education, and religion-related instruction may encourage immigrants to become engaged in community-based programming, increase their social-capital resources, and enhance their likelihood to volunteer. Surprisingly, citizenship lessens the likelihood of secular volunteering. This might be due to recent anxiety related to illegal immigration and the desire of some individuals, despite becoming citizens, to volunteer less for formal secular organizations. Sundeen et al. (2007) found that citizenship produced different results on volunteering among three Asian subgroups. Chinese Americans who were naturalized citizens were less likely to volunteer than non-citizens. An alternate explanation might be the bi-dimensional acculturation strategies reported by Lee, Sobal, and Frongillo (2003). They found that separate or marginalized Korean Americans contrasted with more assimilated Korean Americans, who tend to be younger, married, more educated and integrated, and more likely to volunteer. Those with citizenship might be in the marginalized group who have lived long enough in the United States to obtain citizenship but maintain their cultural heritage and have not yet sought to become involved in the greater community. More research is needed on the impact of citizenship and to further explore the bi-dimensional acculturation model for Korean Americans.
A determinant of formal volunteering for Korean Americans is participation in religious institutions, which supports the finding for Asian Americans by Ecklund and Park (2005). Accordingly, religiosity is the factor associated with religious volunteering for Korean Americans. Of interest, is the limited association of acculturation variables impacting religious volunteering. Korean identity is a statistically significant factor negatively associated with religious volunteering. A large percentage of Korean Americans are Protestant (Smith et al., 1999), which may explain why acculturation has very little association with religious volunteering for this Asian-American group. Still, the literature is ambiguous concerning the impact of diverse eastern religion faiths, including Catholicism and Protestantism on the Asian-American immigrant's volunteering behavior for formal secular organizations (Sundeen et al., 2007). More research is required on the relationship between secular and religious volunteering for Korean and other Asian-American groups. In particular, will Korean Americans who attend a Protestant church in an ethnic neighborhood with other similar immigrants, increase their social capital and networking, which support greater formal volunteering, or will such attendance increase cultural capital, thereby supporting the isolation of specific ethnic values and desires and resulting in less volunteering to formal organizations? Since churches play an important role in the adjustment of ethnic and immigrant groups into mainstream society, a better understanding of the underlying foundations of volunteering for religious and secular organizations by ethnic and immigrant groups is essential.
This research examined the informal helping behavior that is important to fully understanding the conceptual framework of volunteering for the mainstream population (Wilson & Musick, 1997) and uniquely critical for ethnic and immigrant subgroups (Sundeen et al., 2007(Sundeen et al., , 2009). Variables such as religion and education were found to be significantly associated with secular and religious volunteering, but not with helping friends, family, and community members. Significantly associated with informal volunteering was age (obligations increased with age, peaking with children and extended-family obligations), children (children at home decreased the likelihood of helping others informally), gender (women in particular helping, nurturing and providing care to hold the family together increased informal volunteering, as contrasted to men). Acculturation variables (particularly language difficulty, Korean identity, citizenship, and generational status) were significantly associated with formal volunteering and to a lesser extent religious volunteering; these variables were not associated with informal volunteering. Surprisingly, the Korean tradition was the only acculturation variable that increased the likelihood of helping family and friends. This independent and unrelated behavior for informal volunteering, as compared to formal acts of philanthropy, calls for additional examination of Korean Americans' and Asian Americans' volunteering practices.

Disclosure Statement
The author(s) declare that there are no conflicts of interest that relate to the research, authorship, or publication of this article.
In 2015, the Department of Veterans Affairs Office of Policy and Planning convened the first of its kind forum to inform stakeholders about national policy needs to advance the outcomes for veterans and their families as they reintegrate back to civilian life after military service. This article reports of the proceedings of the forum, which brought together more than 30 participants from across the federal government, private sector, nongovernmental organizations, and academic institutions. During the forum, participants discussed the need for a conceptual framework and standard lexicon to support veteran family reintegration policy and strategy. Forum participants highlighted the importance of a collaborative relationship between researchers and policy makers, and identified research gaps and emerging topics that will help inform national reintegration outcomes.

Keywords: Veterans, Military Families, Reintegration, Transition, Veterans Affairs
Every year, hundreds of thousands of service members and their families leave the military and begin the process of permanently reintegrating back into their civilian communities. The Department of Defense forecasts that the rate of separations from active duty will result in over 1 million new veterans over the next 5 years (Chairman's Office of Reintegration, 2014). For many veterans and their families, reintegration is a complex, multifaceted process that involves finding a "new normal" in the realm of family relationships, wellness, and economic stability.
The fluctuating nature, frequency, and intensity of military conflicts coupled with an everchanging American society have presented incomparable challenges and complexities for both the veteran family and the community during the reintegration process. This complexity continues for veterans who have served since the attacks of September 11, 2001, many of whom experienced combat in Iraq and Afghanistan. Research conducted by the Department of Veterans Affairs (VA) estimates that about half of all Iraq and Afghanistan veterans perceive some difficulty in their own reintegration (Sayer et al., 2015). Post-9/11 veterans are more likely to experience head or neck trauma (30 percent) than those who served in Vietnam (16 percent) or World War II (21 percent) (Owens et al., 2008). Furthermore, while veteran unemployment is lower than non-veterans, more than half of all post-9/11 veterans will experience a period of unemployment upon separation from the military, often compounding the reintegration process (Bureau of Labor Statistics, 2016; Department of Veterans Affairs, 2015a).
What has not fluctuated, however, is our nation's resolve and determination to support veterans and their families as they reintegrate back into our communities. As the Civil War came to a close in 1865, Abraham Lincoln charged that we would "care for those who shall have borne the battle and for his widow, and his orphan." Nearly 100 years later in the wake of World War II and Korea, General Omar Bradley's Commission of Veterans Pensions affirmed that our main obligation is to ensure the successful reintegration of veterans into civilian life (President's Commission on Veterans' Pensions, 1956). This commitment continues today. Practitioners and academics have worked to affirm this commitment to veterans and their families by developing programs and research to (a) reduce veteran homelessness; (b) advance education by implementing and evaluating the Post-9/11 GI Bill; (c) partner on veteran hiring initiatives; and (d) identify opportunities to better connect veterans to their local communities (Johnston & Angell, 2013;McDonald, Jin, Camilleri, & Reitano, forthcoming). Despite these efforts, the individual nature of each veteran family's reintegration presents significant challenges for policy makers. We found that ongoing research and analysis is generally confined to specific areas, such as health care, employment, disability, or education. Furthermore, there are currently limited theories in the literature to support a multidisciplinary framework regarding the full cycle of the reintegration process or how these areas interrelate. This can present challenges for policy makers, service providers, and stakeholders to develop, agree on, and ultimately fund complementary strategies that enable holistic outcomes for veterans and families. Moreover, the family unit has often been overlooked in the literature as key parties that both support the veteran and experience the reintegration journey themselves.
VA's Office of Policy and Planning (OPP) recognizes that concentrated efforts are needed to advance the connection between multidisciplinary research and interdisciplinary policy analysis in order to inform policies which support positive reintegration outcomes for veterans and families. To this end, OPP has developed a policy research agenda which identifies veteran family reintegration as a priority research area (Office of Policy and Planning, 2015a). In June 2015, OPP convened a day-long forum in Arlington, VA with the purpose of informing stakeholders of policy needs in order to advance the outcomes for veterans and their families as they reintegrate back to civilian life after military service. Government experts from VA and other federal agencies, such as the Departments of Defense and Labor, were joined by experts from other government and non-government organizations. Participants were asked to leverage their practitioner experiences and academic research knowledge in order to explore key challenges facing veterans and families during reintegration, and to consider desired scenarios that address the key challenges from a broader conceptual perspective.
This article summarizes the proceedings of OPP's forum on veteran family reintegration and incorporates information from practitioners and researchers, a review of current literature, and OPP's ongoing veteran policy and research initiatives. More specifically, this article (1) explores the need for a conceptual framework and standardization in understanding of the reintegration process; (2) discusses the importance of stakeholder awareness through an understanding of programs, services, and partnerships involved in veteran family reintegration; and (3) identifies data gaps and understudied areas critical to further understanding veteran family reintegration.

Participants
OPP hosted the forum and invited the Department of Defense's Office of Military Community and Family Policy as an honorary co-host. Forum participants included individuals with a range of backgrounds, experiences, and knowledge. They were either involved in current veteran family reintegration efforts, and/or were experts who could address related policy and research areas. Attendees represented the organizations are identified in table 1.

Framework for Veteran Family Reintegration
Castro and Kintzle (2014) postulate that military transition, or the progression through which service members transition out of the military, occurs in three overlapping phases: (1) approaching the military transition, (2) managing the transition, and (3) assessing the transition. Forum participants did not come to a consensus on the definition of veteran family terms are often confused or used interchangeably in program implementation. When looking at certain government and nongovernment reintegration programs, the context of reintegration can refer to either post-deployment reintegration or permanently leaving the military. For example, Currie, Day, and Kelloway (2011) define reintegration as a post-deployment transition. The Defense Centers of Excellence (2015) defines periods of transition as being many things, such as a change in status, relocation, or the permanent return to civilian life. These different processes may present service members, veterans, and their families with different challenges. Post-deployment reintegration may present challenges towards reintegrating the service member back into the family unit, whereas permanent reintegration into the civilian community may present different challenges to the family unit, such as changes to the family's financial status (Doyle & Peterson, 2005;Sayers, 2011;Government Accountability Office, 2014). Furthermore, a review of the United States Code and Code of Federal Regulations finds no common legislative language that defines "veteran reintegration" or which governs all federal veteran reintegration support programs and strategies across multiple agencies.
Because the forum focused on veteran family reintegration, some participants also highlighted the changing understanding of what defines the veteran family. Participants stated that the veteran family is no longer limited to the veteran with an opposite-sex spouse and dependent children. Participants generally agreed that the definition should be broad and allow for any definition of family, as defined by the veteran and those closest to him or her. For example, participants stated that the veteran family can include caregivers, children, parents, partners, and siblings, because each plays an important role in the veteran family and can also be affected by reintegration themselves. The VA is already taking some efforts to recognize this changing definition of the veteran family. For example, the Secretary of Veterans Affairs has started socializing the subtle change in the language of VA's mission, which is based on Abraham Lincoln's "to care for him who shall have borne the battle, and his widow, and his orphan." VA (2014b), in Secretary Robert McDonald's core values message, states that VA's mission is "to care for those who shall have borne the battle and their dependents, and their survivors." Furthermore, in response to recent Supreme Court decisions regarding Fourteenth Amendment protections for same-sex marriage recognition, VA now recognizes all same-sex marriages without regard to a veteran's state of residence (Department of Veterans Affairs, 2015b).
With more than 218,000 service members projected to leave active duty military service in fiscal year 2015 alone, forum participants agreed that there is an urgent need for practitioners and researchers to have an understanding of a framework for reintegration, which should include a standardized lexicon surrounding veteran family reintegration (Office of the Assistant Secretary of Defense for Readiness and Force Management, 2014). Furthermore, an increased dialogue between researchers, practitioners, and policy makers within the context of a conceptual framework and standard lexicon was said to be required in order to ensure that stakeholders are able to broadly align their efforts in order to best understand and improve veteran family reintegration outcomes for each individual case.

Inventory of Programs, Services, and Partnerships
Forum participants represented the broad array of reintegration programs available for veterans and their families in government and non-government sectors. OPP ensured that a wide representation of VA offices dedicated to veteran family reintegration participated in the forum. For example, staff and leadership from VA's Veterans Health Administration, Veterans Benefits Administration, National Cemeteries Administration, and Voluntary Service Office attended the event. Further, government and non-government stakeholders with unique perspectives on women veterans, military and veteran families, and other veterans' needs were also invited to ensure a diverse representation of thought. Many participants had pre-existing relationships and were already working to develop or sustain partnerships around specific veterans' issues. For example, VA cooperates with the Departments of Defense and Labor, as well as the military services and Small Business Administration, to implement transition assistance programs (Veterans Benefits Administration, 2013). Despite the prominence and experiences of attendees, however, participants were not positioned to be aware of every possible program or service already supporting veterans and their families. In fact, many times throughout the day, a participant would highlight the need to develop a certain program only to be informed by another participant that such a program already existed.
Participants noted that this lack of complete awareness may be the result of the sheer number and scope of military and veteran-focused organizations, programs, and services. Some participants stated that this makes it a challenge to be fully aware of the entire spectrum of programs, services, and partnerships available for veterans and their families. Other participants stated that this is representative of a strong community of concern. However, still others noted that this uncharted and overwhelming community of concern can lead to reintegration programs and initiatives being disjointed or compartmentalized. According to the National Center for Charitable Statistics (2015), there were more than 41,000 registered nonprofit organizations alone serving the military and veterans. Furthermore, federal, state, and local governments, as well as academic institutions, local communities, and private citizens also provide a wide array of programs and services to veterans and their families (Mendieta & McDonald, 2013). As a result, even the most seasoned advocates participating in the forum did not know about all of the available reintegration programs and services for veterans and their families. This may present difficulties for veterans and their families as well (Carter & Kidder, 2013).

Data Gaps on the Veteran Family
Throughout the forum, many participants cited available data sets on military families, such as those found in the Department of Defense's Military Family Life Project's Active Duty Spouse Study and annual Support to Military Family Readiness Plans report. However, they noted the lack of available similar data on veterans' families. A few participants stated that this is primarily due to a lack of longitudinal data on veteran families. A forum held by the Department of Defense's Transition to Veterans Program Office (2015) reported similar concerns about longitudinal data gaps. While the Department of Defense and VA have collaborated on some research, such as on post-combat family reintegration, and many universities have established centers that dedicate research towards both the military and veteran family, the literature does not offer many examples of veterans and their families being the subject of research during both pre-and post-separation (Doyle & Peterson, 2005).
As a solution to gaps in longitudinal data, some participants recommended extrapolating information from the available research on military families. Other participants, however, cautioned against this activity to avoid making erroneous conclusions, as the needs and situations of military and veteran families may not be the same. They also highlighted the challenges with obtaining and sharing data across organizations, both government and nongovernment alike. Research being conducted outside of the United States, such as the Life After Service Survey conducted by Veterans Affairs Canada, may provide an example of how researchers can follow service members and their families as they become veteran families (Thompson et al., 2014). As part of the Life After Survey Study series, Veterans Affairs Canada will be conducting a study specifically on the outcomes of families as they transition from military to civilian life.
Through the forum and other veteran policy research and analysis initiatives, VA has found that the veteran family has not yet been the subject of a comparative amount of study and research attention that the military family community has received. For example, a simple internet search using variations of the phrase "veteran family research" results in information sources on military family research. In order to help make the veteran demographic more accessible to researchers and encourage interest in veteran topics, VA has identified some preliminary resources to help guide first-time veteran researchers (Office of Policy and Planning, forthcoming). Furthermore, VA has improved its efforts to make public data easily accessible and usable through the Open Data Initiative (Department of Veterans Affairs, 2014a). However, there is a continued need by government and non-government organizations to collect and share the information needed to support strategic policy, especially on non-healthcare topics such as veteran and families' choices, contemporary challenges, and their concerns.

Emerging Areas of Research
Throughout the day, forum participants cited that while there are data gaps relating to veteran family research, several understudied areas related to veteran families and reintegration are emerging but require additional research for further policy exploration.
Participants commented that service members are provided financial education and protections from predatory lending to mitigate financial instability for the military family, while veterans are not provided the same protections. For example, the Military Lending Act of 2006 provided specific protections for active-duty service members and their families in consumer credit transactions. In 2015, after new legislation and a 3-year study by the Departments of Defense and Treasury, the Federal Trade Commission, and financial regulators, service members and their families are further protected from committing to loans with excessive fees and charges (Cronk, 2015). 1 However, such legislation does not apply once the service member leaves the military and becomes a veteran. The potential for loss of income and non-transferability of protective measures may result in financial instability for certain veterans and their families, especially during the period of reintegration, where as many as 1 out of every 2 Iraq and Afghanistan veterans face a period of unemployment (Department of Veterans Affairs, 2015a; Government Accountability Office, 2014).
Participants also noted that geographic dispersion and other differences between the active duty and reserve and national guard (R/NG) component creates added challenges, including greater isolation and reduced access to the military community supportive network. This is particularly important as R/NG service members have had an increase in combat deployments over recent decades (Vogt et al., 2008). Given the frequency of mobilization, the experiences of the R/NG service member and their families in dealing with separation, relocation, and reintegration are distinct from their active duty counterparts. A review of literature has found that limited research has been done to look at how these differences post-deployment affect the reintegration outcomes of R/NG service members, especially when confronted with similar combat experiences as their active duty counterparts.
When active duty service members return from combat, they will return to their military bases, which oftentimes have support services available within their own communities. Many reintegration resources are available to R/NG service members when they return home (Defense Centers of Excellence, 2016). However, some participants stated that support to R/NG may not be as readily available and may require additional initiative or effort on the part of the R/NG service member or family, because they may live in communities that are a great distance from the nearest military base. Additional research on R/NG outcomes are required to understand the challenges faced and to evaluate the effectiveness of programs and support offered to R/NG families.

Conclusion
Government and non-government organizations provide resources, support, and services through programs and benefits in order to facilitate successful veteran family reintegration. However, the forum highlighted the need for further research and collaboration to ensure that there is a common understanding of veteran family reintegration. Based on these findings, practitioners and academics can take immediate action by exploring the development of a collective framework for reintegration with input from the community of veteran stakeholders. Such a framework should include a common understanding of the lexicon of definitions, outline broad strategies, and identify potential roles and responsibilities for the federal government and community stakeholders. Furthermore, the veteran family reintegration framework can help serve to identify which of the tens of thousands of military and veteran support programs in the United States today are stakeholders specific to the veteran family reintegration process.
Even with such a framework in place, researchers and policy makers need to take steps to address gaps in data, in particular, surrounding the veteran family. As a first step, VA is implementing veterans policy research and analysis as a process for analyzing issues related to the benefits and services needed for the overall health and wellbeing of veterans, service members, and their families (Office of Policy and Planning, 2015b). Innovative collaboration between researchers and policy makers, such as the veterans research and analysis process, is one example of how this can be achieved. Such collaboration, as exemplified by this forum, has already highlighted the need to focus on data gaps relating to the veteran family and to continue efforts on emerging research topics such as family financial literacy and R/NG reintegration outcomes. In the long-term, practitioners and academics must continue to work collaboratively to drive a more robust, veteran-focused dialogue around the myriad of policy issues affecting the lives of veterans and their families.
In the last few decades, few topics have been of more interest to public administration scholars than governance. The change of public service delivery paradigm from government to governance enables us to develop and test diverse service delivery mechanisms that are generally characterized by networking among multiple actors, public-private collaboration, and negotiation and persuasion as decision-making rules (Goldsmith & Eggers, 2004;Salamon, 2002). One of the areas where this current trend is well observed and studied is local administration. Appropriately, Teles' Local Governance and Inter-municipal Cooperation comes at a time when we are reviewing what we have done so far and considering what more we academics need to do in the future.
Teles points out briefly, but quite effectively, why studying local governance and intermunicipal 1 cooperation is important to contemporary public administration. He argues that "place and identity play a relevant role in public policy" (p. vii). Simply put, local is where we live every day with other people and where most fundamental services are provided to us, so it is an old, but fundamental, research theme. According to Teles, the relevance of studying interlocal cooperation, especially in a comparative context, appears in the diversity of institutional arrangements, cultural and administrative traditions, different societal functions, and dissimilar competencies in each case. In spite of its challenging diversity, Teles encourages scholars to pursue systematic and coherent research in this arena.
The book is comprised of three sections that address its objectives step-by-step: 1) discussing key issues related to inter-municipal cooperation through an in-depth literature review and the author's own development of the extant research (Chapters 1 to 3); 2) adding comparative lessons and implications to the literature by examining European countries (Chapters 4 and 5); and 3) identifying challenges that need to be addressed in future research (Chapter 6). As emphasized throughout the book, the underlying need for this study is the continuing trend of territorial change reform across European countries. The study of inter-municipal cooperation based on the comparative perspective is the book's biggest contribution to the literature.
Chapter 1 focuses on motivations for inter-municipal cooperation. To explain the increased use of cooperation between municipalities, Teles provides his own framework which includes four categories of drivers: a priori (nature of the problem, history of previous collaboration); intrinsic (identity and territorial context, power relations among municipalities, institutional context, organizational profile of each municipality); extrinsic (external influence and incentives); and a posteriori (expected outcomes). The combination of those four dimensions is effective and comprehensive for examining inter-municipal cooperation motivation. Teles also posits that while it provides efficiency gains as a major benefit, economies of scale pose a serious challenge to democratic control and accountability for participating local governments. This fundamental conflict between benefits and potential risks is repeated throughout the book, which helps the author maintains a balanced perspective when addressing inter-municipal cooperation.
In Chapter 2, Teles explains the diversity of inter-municipal cooperation arrangements. The nature of inter-municipal cooperation varies from a more voluntary and bottom-up approach to a more compulsory and top-down approach. It involves different combinations of actors comprising internal (local politicians, local bureaucracy, and other inter-municipal organizations) and external (civil society organizations, private companies, regional or central organizations, and citizens) components. There are also variations in its form among service delivery, contracts, coordination, and information sharing. In addition, different theories have been tested to study inter-municipal cooperation, including the political economy model, the network theory approach, the public choice argument, and the collective action approach. This chapter is the most theoretical part in the book and provides considerable information. Although these theoretical discussions are complex, Teles approaches each while putting the diversity of inter-municipal cooperation at the core.
Chapter 3 compares alternatives that municipalities can choose to increase the efficiency and scale of economy of local governance. While it is his primary focus, Teles does not simply praise inter-municipal cooperation, but balances his approach with discussion of other alternatives of consolidation and meso-governance arrangements. Teles suggests local governments can also work with upper level units such as national government and regional organizations when small is not beautiful. Another approach is corporatization, which takes advantage of market mechanisms in local service provision. From this chapter we learn there are multiple alternatives in local and regional governance, and finding the most relevant one is a challenging goal for local governments. 2 It would be beneficial, however, if the author examined more deeply the influence of regional organizations on inter-local cooperation in the European context. Since regional organizations may significantly influence voluntary inter-local cooperation efforts (Kwon, Feiock, & Bae, 2014;LeRoux, Brandenburger, & Pandey, 2010), a comparative study including European and U.S. regional organizations would be interesting.
Chapters 4 and 5 analyze European local governance, focusing on strategies that local governments have taken for successful territorial reforms (e.g., balancing economy of scale and democracy). In Chapter 4, Teles explains that a variety of cooperative arrangements have been used among European countries in conjunction with the countries' different territorial reform approaches. For example, while Denmark prefers amalgamation, this approach is not welcomed in Portugal. Chapter 4 is a brief summary based on an extensive review of literature that includes key characteristics of many European countries. In Chapter 5, Teles strongly suggests the need to consider the unique political contexts of each country when studying its territorial reform approaches. He specifically focuses on Portugal. He explains the political contexts of Portugal and examines the current status and relevance of each of the major approaches within that political system. An interesting challenge to choosing between consolidation and decentralized inter-municipal cooperation in Portugal is that, while the national government maintains strong political power, citizens have a culture of valuing local identity. Both of these chapters demonstrate that diversity in approaches to territorial reform among European countries means that local governance and inter-local cooperation are complex and challenging research themes that will be better understood when more comparative studies are done.
In the last chapter (Chapter 6), Teles discusses questions that need to be addressed in future studies. One fundamental question is how to improve the governance capacity of intermunicipal cooperation in a multi-level policy context. Each of three key domains (efficiency, democracy, and stability) generates specific questions. Teles argues that high levels of diversity and complexity in local governance and inter-local cooperation across countries, especially those caused by the different political systems, continue to increase the demand for more empirically evidence-based and more comparative studies. We need to bear his recommendations for future studies in mind.
Overall, the book accomplishes its aforementioned objectives by helping readers understand the academic and practical meanings of inter-local cooperation in local governance, the variation in the use of inter-local cooperation across countries (particularly in Europe), and the areas that we need to elaborate on in the future. Through these efforts, Teles shows that we have achieved much in the study of local governance and inter-local cooperation, but this field is so dynamic that more rigorous and comparative research should be added continuously.
An extended volume could include more detailed explanations of the issues covered in the text; however, it was not the author's intention to publish a textbook-type thick book. In spite of this, the current version has sufficient advantages, especially because it includes major critical issues in inter-local cooperation and expands non-U.S. contexts. This book will help practitioners and students and scholars of local governance and city management understand the dynamics in this field. This would also be a useful textbook for graduate seminar classes on local and metropolitan governance.

Disclosure Statement
The author(s) declare that there are no conflicts of interest that relate to the research, authorship, or publication of this article.