The Costs and Benefits of Volunteering Programs in the Public Sector: A Longitudinal Study of Municipal Governments

Abstract

Public sector volunteering is an increasingly common phenomenon that has attracted public administration scholars’ attention for several decades. Previous literature suggests that one of the main advantages of citizen involvement in public service delivery is related to subsequent cost savings. However, a lack of longitudinal data has limited our ability to test this proposition and constrained our understanding of the subject. The main goal of this study is to fill this lacuna and analyze the relationship between volunteering and the level of spending as well as the number of paid employees in city governments across 10 years. The empirical findings indicate that volunteer involvement is indeed significantly and negatively associated with the level of spending and the number of paid employees, with the number of services unaffected. The cost savings, however, may come with delays most likely because it takes time and resources to properly train volunteers. This may also drive up administrative costs.

Keywords

volunteer management local government cost savings budget size public sector volunteering

Introduction

The efficient and effective delivery of public services constitutes a core purpose of public administration (Neshkova & Guo, 2012). While bureaucratic expertise is an important element for successful policy implementation, citizen involvement during various phases of the policy cycle is increasingly seen as a way to enhance bureaucratic decision-making and accountability (Nabatchi & Leighninger, 2015). The use of volunteers to co-deliver public services constitutes one such alternative. Although volunteers are generally associated with nonprofit organizations, there is an increasing recognition of the contribution that volunteers have in public service delivery (Brudney, 1999; Dover, 2010; Gazley & Brudney, 2005).

Potential benefits from volunteer involvement may include improvements in service efficiency and effectiveness, better incorporation of citizen preferences, and enhancement of democratic goals such as greater bureaucratic accountability (Brudney, 2005; Brudney & Kellough, 2000). A major proposition in the literature is that volunteers can either generate cost savings or enhance service levels with a given amount of resources (Brudney, 1984; Parks et al., 1981). However, despite several efforts to analyze the costs and benefits of volunteering in the public sector (e.g., Brudney & Duncombe, 1992; Brudney & Gazley, 2002; Brunet, DeBoer, & McNamara, 2001; Handy & Srinivasan, 2004; Hilke, 1986), there have not been many longitudinal studies with representative samples to provide more robust empirical evidence about the alleged relationship between volunteering, cost savings, and paid-employees in the public sector.

The main goal of this article is to address this gap in the literature and explore the effects of public sector volunteers on cost savings and paid staff. The study uses a panel dataset on all cities in the State of Georgia with different levels of volunteer engagement in public service delivery across a 10-year period. This study examines two main hypotheses derived from pertinent literature. First, the study tests the proposition that volunteers are associated with cost savings holding the level of services constant. Second, it tests the hypothesis that volunteer use may be associated with a reduction in the number of paid employees. Empirical findings from the panel data analysis indicate that volunteer programs are associated with cost savings and may lead to reductions in the number of paid employees, with the number of services not affected. In addition, I find that volunteer engagement is associated with significant increases in city expenditure on the administrative function, which is responsible for volunteer recruitment, training, and management. These relationships have not been sufficiently explored in previous studies, and the findings provide a valuable empirical contribution to the literature with important practical implications.

This article is organized as follows. The next section lays out a theoretical framework describing how public sector volunteers may influence cost savings as well as paid employees. This section concludes with a set of hypotheses derived from the literature to be tested in this study. A relevant literature review is summarized in the third section. The next section presents the data, variable description, and the research method. Empirical results are discussed in the fifth section. The article concludes with relevant implications for research and practice.

Theoretical Framework: Advantages and Disadvantages of Public Sector Volunteering

Volunteering in the Public Sector

Researchers commonly consider volunteers in the context of nonprofit and voluntary organizations. Most national surveys on volunteering trends such as those produced by the Department of Labor (DOL) and Bureau of Labor Statistics (BLS) do not distinguish between public and nonprofit sectors. However, scholars have documented a large scope and magnitude of volunteering in public organizations and argued for studying public sector volunteering separately. For instance, Brudney, Harris, and Sink (in press) suggest that between 20% and 30% of all volunteers in the United States donate time to public organizations, an estimate that is consistent with a prior study (Brudney, 1999), which found that nearly a quarter of all volunteers donated their time to a government organization in 1992. The International City/County Management Association (ICMA) Alternative Service Delivery (ASD) Survey for 2007 also found that 27% of cities and counties used volunteers to deliver at least one type of service (Nesbit & Brudney, 2013).

Researchers note volunteering “ranked second only to the use of contracting with respect to the number of services where the method is applied and the percentage of local governments that have adopted it” (Brudney, 1990c, p. 11). Volunteers become particularly important for public administration during times of fiscal distress, when they comprise a major service delivery alternative (Brudney & Warren, 1990; Levine & Fisher, 1984). Instead of serving as a cheap form of labor and functioning as ancillaries to paid staff, many volunteers are involved in capacities that are central to the public organization (Nesbit & Brudney, 2010). For instance, volunteer firefighters or reserve/auxiliary police officers in many jurisdictions are trained, organized, and equipped to assume roles that are nearly identical to those of paid employees. Some volunteers in the Peace Corps and AmeriCorps VISTA are involved in managerial and policymaking roles (Reeves, 1988).

Because of the extensive volunteer engagement and the central roles they may assume in various public functions, scholars have argued that government volunteers merit separate inquiry from volunteering in nonprofit and voluntary organizations (Brudney, 1990b; Brudney & Warren, 1990). Public sector volunteering programs may offer distinct advantages and disadvantages for public service delivery that may be overlooked when only considered in the context of the nonprofit or voluntary sectors. The following subsection reviews potential effects of public sector volunteering on the size of public budget and the level of paid employees as discussed in the literature.

Potential Costs and Benefits of Public Sector Volunteers

Relevant literature suggests that volunteer involvement in public services may be associated with several potential benefits, such as improvements in service efficiency and effectiveness, better incorporation of citizen preferences, and even enhancement of democratic goals such as greater bureaucratic accountability (Brudney, 2005; Brudney & Kellough, 2000). Volunteers may bring specialized skills to public organizations that current employees may not possess. Examples include lawyers who provide legal services on a volunteer basis (Stebbins, 2009), or former military or police personnel with specialized skills volunteering as instructors for in-house training or at academies (Dobrin & Wolf, 2016). In other cases, as volunteers may have greater social and service motivations (Coursey, Perry, Brudney, & Littlepage, 2008; Stukas, Hoye, Nicholson, Brown, & Aisbett, 2016), they can bring skills such as counseling and outreach that enable a better understanding of the needs of clients and citizens. This may allow public organization to improve relations with the community and increase public support for government.

One of the major propositions in the literature is that volunteers can generate cost savings or enhance the level of services with a given amount of resources (Brudney, 1984, 1990b, 1993; Parks et al., 1981). Volunteers can either engage in supportive activities or assume core functions, which allow governments to provide essential services with fewer employees who must be available at any one time. In theory, because volunteers consist of unpaid labor, their involvement may help reduce the need for paid employees. This in turn can lower the cost of service inputs while holding the quantity of service outputs constant. Personnel expenditures generally take a sizable share of public resources, so, according to this proposition, greater interchangeability between volunteers and paid employees may lower the size of public budgets.

Alternatively, public sector volunteers may not necessarily be a substitute for paid employees, but instead complement and augment their efforts. According to this scenario, volunteering may be associated with higher levels of service outputs holding the number of paid employees and budget size constant. The extent to which volunteers substitute or complement paid employees is hard to measure, but it may depend on the nature of the organization mission, the type of services provided, union regulations, and other factors (Handy, Mook, & Quarter, 2008; Russel, Mook, & Handy, 2017). Volunteers may be more likely to supplement rather than replace paid employees during peak load periods, fiscal challenges or service expansion (Sowa & Word, 2017), or because of opposition from paid employees (Ellis & Noyes, 1990; Fischer & Schaffer, 1993; Pynes, 2008). The optimal mix of volunteers and paid labor also depends on different levels of volunteer administrative costs (Duncombe & Brudney, 1995).

Although the common perception is that volunteers are not paid for their work and therefore are “free,” researchers note that volunteering programs may not be completely cost-free (Kiser & Percy, 1980). Despite the expectation that volunteers can assume core activities, volunteer programs still require resources to recruit, train, supervise, coordinate, and retain volunteers (Brudney, 1990a; Brudney & Kellough, 2000; Emanuele, 1996; Hatry, 1983). This may drive up administrative costs (Handy & Mook, 2011). The programs that involve citizens in more institutionalized capacities can impose additional costs for liability insurance. While some service costs may be more important than others, scholars note that volunteer accounting and recordkeeping are notoriously poor (Brudney & Duncombe, 1992), thereby making it difficult to assess the full set of administrative costs associated with volunteer management.

To summarize, the use of volunteers may potentially lead to lower costs of service provision most likely by reducing the number of paid employees and the cost of payroll. Volunteering programs, however, may not be totally free as resources are needed for volunteer recruitment, training, and management. This could potentially reduce or even negate potential savings. Another possibility is that the use of volunteers may help the entity to increase its service capacity with a given number of paid employees and budget size. Whether and to what extent cost savings or increased service production occur remains unclear. The main purpose of this article is to test empirically the theoretical propositions put forward in the literature and provide more evidence on the relationship, if any, between volunteer use, government expenditures, and the level of paid employees. This discussion leads to the following hypotheses:

Hypothesis 1: Volunteer use is negatively associated with expenditure levels.

Hypothesis 2: Volunteer use is negatively associated with the number of paid employees.

The alternative in both cases is no effect. The next section provides a review of the relevant literature and specifies gaps that deserve further exploration.

Relevant Literature Review

Most discussions of volunteers in public service delivery have focused on the potential for generating cost savings, regardless of whether cost savings were actually achieved (Brudney & Gazley, 2002). The underlying basis is that the addition of citizen inputs either decreases the costs devoted to producing services or increases service levels with a given budget, ultimately resulting in a net decrease in service expenditures (Brudney, 1993; Gazley & Brudney, 2005).

Among prior studies exploring the fiscal consequences of volunteer usage, Hilke (1986) finds that the use of volunteer firefighting units reduces overall expenditures on firefighting activities, and thus lower spending and taxes for local government in general. Brudney and Duncombe (1992) develop a methodology for comparing the costs of using paid, volunteer, and mixed staffing to examine different levels of cost-effectiveness in municipal fire departments and which staffing arrangements work better than others.

Brunet et al. (2001) find that professional fire departments are more cost-effective at high levels of fire protection, whereas volunteer fire departments are more cost-effective at low levels of fire protection. Donahue (2004) emphasizes the importance of managerial influence in determining the costs of production. Specifically, she finds that volunteer firefighter agencies with more centralized management structures tend to be associated with higher per capita costs of production. Handy and Srinivasan (2004) explore the costs and benefits of volunteering within the context of hospitals and find that the use of volunteers provides a net return of an average of US$6.84 in value from volunteers for every dollar spent. Brudney and Gazley (2002) conduct a longitudinal analysis of a single agency, the Small Business Administration (SBA), and did not find a reduction in the agency budget when adding volunteers to the workforce.

While these studies provide valuable insights about the association between volunteers and cost savings, their shortcomings include the cross-sectional nature of the data or limited generalizability of the findings. Several questions with respect to the effects of volunteering on government operations and expenditures merit further investigation. To address this gap, I build upon extant research to assess the effect of volunteer involvement in public service provision on the level of spending and paid employees. The costs of public services have always been a matter of high salience for governments, and although volunteering may be an important determinant in reducing costs, this issue has not been sufficiently addressed in longitudinal studies. As a result, public managers have been left with limited knowledge about the impact on public budgets of their decisions to engage volunteers in public service delivery. Thus, this study seeks to improve our knowledge and expand understanding of the benefits and consequences of citizen involvement in government operations. The following section explains the research design used to test our hypotheses in the context of municipal governments.

Research Design

Data

I test the hypotheses using panel data on municipal governments in the state of Georgia covering the decade 2006 through 2015. There are currently 524 cities and towns in Georgia that possess a charter of municipal incorporation approved by the General Assembly (Georgia State Government, 2017). There is no legal distinction among cities, towns, or municipalities in Georgia.

As some cities lack complete data for some years of the study, our panel data are unbalanced and “have missing years for at least some cross-sectional units in the sample” (Wooldridge, 2013, p. 491). However, unlike previous studies that relied on cross-sectional analysis, our data allow longitudinal analysis and provide new insights into causal relationships. Focusing on the cities within a single state also enables us to eliminate any heterogeneity in the outcomes of interest caused by variation in state institutions. The study includes 522 out of 524 cities in Georgia.¹ The analysis of the missing data revealed that 78.5% of the observations (n = 4,037) had complete data for all years. The most common missing data pattern, in about 14% of all observations (n = 734), had to do with the number of employees and volunteers. I regressed an indicator variable for whether a city responded to the Wage & Salary survey (yes = 1, else = 0) on the observable city characteristics using a logit model to assess potential bias from self-selection. The model classified correctly zero nonresponses suggesting that the variables used in this study are not correlated with the factors responsible for whether a city responded to the survey. Tests of means equality of population indicated that there is no statistical difference in terms of total population (p > .10) between the sample used in the analysis and the population. Overall, the sample includes cities with population ranging from a few dozen to more than half a million residents. The descriptive statistics for all the variables used in the study are presented in Table 1.

Table 1.

Descriptive Statistics.

Variable	M	SD	Minimum	Maximum
Dependent variables
City expenditures (ln)	14.30	2.07	8.86	21.46
City employment (ln)	2.93	1.67	0.00	9.12
City general government expenditures (ln)	12.50	1.60	7.38	19.72
Volunteering measures
Ratio of volunteers to paid employees	0.18	0.83	0.00	15.00
Ratio of volunteers to paid employees truncated at 1	0.09	0.27	0.00	1.00
Dummy indicator (volunteers involved = 1)	0.15	0.36	0.00	1.00
Workload
Managerial functions provided (count)	11.35	2.08	2.00	13.00
Services provided (count)	15.96	3.03	4.00	19.00
Facilities provided (count)	10.03	4.47	0.00	17.00
Employees per capita	0.01	0.01	0.00	0.20
Revenue structure
IGR (percent of total revenue)	0.08	0.13	0.00	0.94
Charges for services (percent of total revenue)	0.36	0.23	0.00	0.97
Political institutions
Number of seats on the board excluding CEO (count)	4.95	1.11	2.00	16.00
CEO elected by vote of commission (yes = 1)	0.03	0.17	0.00	1.00
CEO elected by popular vote (yes = 1)	0.97	0.17	0.00	1.00
Board members elected at-large (yes = 1)	0.72	0.45	0.00	1.00
Board members elected by district/at-large (yes = 1)	0.07	0.25	0.00	1.00
Board members selected by district (yes = 1)	0.20	0.40	0.00	1.00
Board members’ election not applicable (yes = 1)	0.01	0.09	0.00	1.00
Population
Population total (ln)	7.41	1.61	3.14	13.20
Population growth (relative to the previous year)	0.01	0.05	−0.47	1.18
Population density (per square mile)	797.80	806.12	19.26	7,429.90
Population in labor force (percent)	45.23	5.75	21.75	64.29
Wealth
Population in poverty (percent)	20.76	7.08	4.50	48.10
NPV (US$1,000s per capita)	2,106.39	5,964.72	24.91	90,007.43
MRPV (US$1,000s per capita)	16.07	11.28	3.67	53.77
Political factors
Voter turnout (percent)	74.24	4.44	54.82	86.65
Votes for democrat (percent)	38.31	14.99	12.07	84.47
Year dummies
2007	0.10	0.30	0.00	1.00
2008	0.10	0.30	0.00	1.00
2009	0.10	0.30	0.00	1.00
2010	0.10	0.30	0.00	1.00
2011	0.11	0.31	0.00	1.00
2012	0.11	0.31	0.00	1.00
2013	0.10	0.30	0.00	1.00
2014	0.09	0.29	0.00	1.00
2015	0.09	0.28	0.00	1.00

Note. N = 4,037, T = 10 (from 2006 to 2015). IGR = intergovernmental aid; NPV = net property taxable value; MRPV = mean residential property value.

The main source of our data is the Georgia Department of Community Affairs (DCA), which collects information on the number of paid employees and volunteers involved in government operations through its annual Local Government Wage & Salary survey. All financial data also come from the Department. The data on services, facilities, and managerial functions performed by each city were extracted from the Government Management Indicators Survey (GOMI) also conducted by the Department. The GOMI survey also reports data on municipal political institutions, such as the number of seats on the board, whether the CEO is appointed by the commission, and how board members are elected (by district, at-large, or by some combination of the two). The data on total city population and population in labor force were provided by the BLS. Data on the poverty level were extracted from the Small Area Income and Poverty Estimates (SAIPE) of the U.S. Census Bureau. Net property taxable value (NPV) and mean residential property value (MRPV) data were obtained from the Georgia Department of Revenue. Finally, presidential political indicators were obtained from the Georgia Secretary of State website.

Dependent Variables

The main dependent variables of interest are the total level of city expenditures and the number of paid city employees, part-time and full-time combined. Based on the extant theory discussed in the previous sections, I hypothesize that volunteer engagement is negatively associated with the size of the budget and the number of paid employees. I use the natural log of both outcomes in this study. Before taking the natural log, dollar amounts were adjusted for inflation using the Consumer Price Index from the BLS.

Independent Variables

The main predictor of interest in this study is the extent of volunteer involvement in public service provision. These data come from the question in the Local Government Wage & Salary survey asking the respondents to enter the total number of “volunteer or non-paid workers serving as employees.” The main independent variable of interest is operationalized in three ways. First, the level of volunteer engagement is operationalized as a ratio of citizen volunteers to the sum of full-time and part-time paid employees. Five of the cities in the sample with population from 191 to 432 residents in some years reported up to 20 volunteers, while having only one or two paid employees. In some cases, this results in a ratio of 15. Hence, to limit such extreme values and to reduce the chances that outliers have a disproportional effect, I use an additional measure of volunteering winsorized at 1, which means that all observations with ratios above 1 are recoded as 1. About 4% of the cities reported having more volunteers than paid employees, and about 9% had half as many volunteers as paid employees. It is worth noting that such high ratios indicate that in a number of cases volunteer engagement in our study most likely was associated with major undertakings, where volunteers played a substantial role in government operations rather than assisting with trifling housekeeping routine. The third measure is an indicator variable coded as 1 for all city-years that involved at least one volunteer and 0 for those that had zero volunteers. About 93% of cities that did not engage in volunteer programs in the first year did not have any volunteers next year. Meanwhile about 57% of cities with volunteers in the first year continued to engage volunteers in the next year.

First, I test the relationship between the use of volunteering and our dependent variables in the current year. In additional models, I lag the volunteering variables by 1 year. As mentioned in the theoretical discussion, proper management of volunteering programs may require substantial investment for recruitment and training. Because of that, noticeable cost savings may come with a delay. In addition, volunteers may be able to perform more important functions with more experience, which also takes time.

Controls

The study utilizes several clusters of controls capturing essential city characteristics that theoretically might affect the outcomes in our study. Cities provide different numbers of services, maintain different facilities, and perform various managerial functions. All of these can affect the levels of spending and employees. Hence our models include three count variables indicating the total number of services, facilities, and managerial functions undertaken by the city. The types of functions, services, and facilities are listed in the appendix. I also account for the total number of paid employees per capita in the expenditure models. This variable helps to tease out the effects of volunteering holding the number of paid employees constant.

Local government institutions are known in urban affairs literature to affect various aspects of local policymaking (e.g., Carr & Karuppusamy, 2010; Feiock, Jeong, & Kim, 2003; Frederickson & Johnson, 2001; Karuppusamy & Carr, 2012; Svara, 2005). Hence, a set of indicator variables controls for variation in municipal political institutions. I account for whether the chief executive is elected by the commission, as well as the method of electing board members (by district, at-large, or some combination of the two). About 97% of the CEOs are elected by popular vote, and in 72% of the cities board members are elected at-large. These two groups serve as omitted comparison categories. I also account for the total number of seats on the board excluding the CEO.

I control for the proportion of local revenue that is received from intergovernmental aid (IGR) and from charges for services. Previous studies find that local officials with a higher proportion of their budgets funded from local taxes rather than IGR tend to be more parsimonious, holding all else equal (Duncombe & Yinger, 1997). The number of residents served, which is one of the main drivers of the level of expenditures and paid employees, is accounted for with the natural log of city population. Population growth is measured as a percent change relative to the previous year, and population density is expressed as the total number of residents per square mile of the city area in a given year.

In addition, I implement several controls for city wealth and ideological leaning. Wealthier communities with a higher fiscal capacity may be more tolerant of inefficient government operations (Eom & Rubenstein, 2006; Grosskopf, Hayes, Taylor, & Weber, 2001; Hayes, Razzolini, & Ross, 1998; Leibenstein, 1966, 1978). These data are not available at the city level across time, so I proxy them by county-level data. Thus, I use total net taxable property within the county, MRPV (assessed at 40% of the fair market value in Georgia), and percent population in poverty. Theoretically, cities located in wealthier counties as measured by these indicators should also have a wider property tax base and a higher revenue raising capacity. An estimation of Pearson’s correlation coefficients between these variables at the county and city level was made using data from the U.S. Census available for all cities for 1 year. The correlation matrices indicated that MRPV in the county is significantly correlated with the city median housing value (0.65) and median household income (0.63). The county level of poverty is also significantly and positively correlated with city poverty (0.60), and negatively correlated with income (0.62) and housing value (0.59). Hence, I believe these county-level variables are reasonably good controls for the purposes of this study.

Another variable controls for the percent of population in the labor force. A higher proportion of population in the labor force may indicate less demand for social services and smaller budgets. These data are not available at the city level, so I also proxy them by county-level data. As for ideological leaning, I use the percent of votes for a Democratic presidential candidate in the most recent elections. Thus, the data from 2004 are used for 2006 and 2007 with the 2008 election data used for 2008-2011, and the 2012 one for 2012-2015. There is a substantial variation in ideological preferences, from 18% to 73% of voters leaning toward a Democratic presidential candidate. I also control for voter turnout at these elections as a proxy for civic engagement. Ideological leaning and voter turnout are also available at the county-level only.

In addition to these controls, I account for any time trends in the data by including in each model 9-year dummy variables with 2006 being the omitted comparison category.

Estimation Methodology

Given the continuous nature of the dependent variables, linear models are estimated. The Hausman specification tests pointed to a potential unobserved heterogeneity bias indicating that the data do not support a random effects model. Therefore, a fixed-effects within-estimator model is used, which seeks to explain temporal variance with cities and accounts for time-invariant unobserved heterogeneity that may bias the random effects estimators (Halaby, 2004). Additional testing indicated potential presence of heteroskedasticity and autocorrelation. Hence, all models are estimated using standard errors clustered by city (Wooldridge, 2013). The variance inflation factor (VIF) check for multicollinearity indicated that multicollinearity presents no serious concerns (Gujarati, 2007).²

Empirical Results

First, I estimate the relationship between volunteering and the natural log of the total level of expenditures (Table 2). Controlling for important city characteristics, such as the number of functions performed, revenue structure, political institutions, population density and growth, fiscal capacity, and ideological leaning, I find no evidence that volunteer programs help cities obtain significant cost savings in the current year (Table 2). It appears that cities with a higher ratio of volunteers relative to paid employees maintain the same level of total spending. The results are stable across different operationalizations of volunteering. Some of the city spending drivers appear to be population size, the number of paid employees, availability of IGR, reliance of charges for services, political institutions, and the percent of population in labor force. Budget size also varies significantly from year to year.

Table 2.

Volunteering and the Level of Public Spending.

Variables	Model 1		Model 2		Model 3
Volunteering measure	Ratio		Ratio truncated		Dummy
Volunteering	−0.01	(−0.96)	0.00	(0.21)	0.02	(1.26)
Paid employees per capita	2.94**	(3.12)	2.99**	(3.11)	2.99**	(3.12)
Managerial functions	0.00	(0.16)	0.00	(0.22)	0.00	(0.21)
Services provided	−0.01	(−1.11)	−0.01	(−1.13)	−0.01	(−1.13)
Facilities provided	0.00	(0.81)	0.00	(0.80)	0.00	(0.80)
IGR	0.30***	(4.43)	0.30***	(4.39)	0.30***	(4.39)
Charges for services	0.39**	(2.73)	0.39**	(2.71)	0.39**	(2.71)
Number of seats on the board	0.06	(1.47)	0.06	(1.44)	0.06	(1.42)
CEO elected by commission	−0.11	(−1.06)	−0.11	(−1.07)	−0.11	(−1.10)
Board elected by district/at-large	0.04	(0.60)	0.04	(0.59)	0.04	(0.58)
Board elected by district	0.05	(1.43)	0.05	(1.42)	0.05	(1.41)
Board elections not applicable	−0.14***	(−4.74)	−0.16***	(−5.14)	−0.16***	(−5.62)
Population total	0.18*	(2.49)	0.18*	(2.49)	0.18*	(2.49)
Population growth	−0.18*	(−2.57)	−0.18*	(−2.58)	−0.18**	(−2.62)
Population density	−0.00	(−0.29)	−0.00	(−0.27)	−0.00	(−0.28)
Population in labor force	−0.00*	(−2.54)	−0.00*	(−2.57)	−0.00*	(−2.57)
Population in poverty	0.00	(0.67)	0.00	(0.65)	0.00	(0.65)
NPV	−0.00	(−0.71)	−0.00	(−0.74)	−0.00	(−0.77)
MRPV	0.00	(0.07)	0.00	(0.10)	0.00	(0.12)
Votes for Democrat	0.12	(0.55)	0.12	(0.54)	0.12	(0.53)
Voter turnout	−0.00	(−1.02)	−0.00	(−1.01)	−0.00	(−1.00)
Year = 2007	0.05**	(3.22)	0.05**	(3.25)	0.05**	(3.24)
Year = 2008	0.03*	(2.25)	0.03*	(2.26)	0.03*	(2.23)
Year = 2009	0.05**	(2.62)	0.05**	(2.62)	0.05*	(2.58)
Year = 2010	0.04*	(2.14)	0.04*	(2.19)	0.04*	(2.19)
Year = 2011	0.03	(1.48)	0.03	(1.51)	0.03	(1.50)
Year = 2012	0.04	(1.64)	0.04^†	(1.65)	0.04	(1.63)
Year = 2013	0.03	(1.30)	0.03	(1.33)	0.03	(1.33)
Year = 2014	0.03	(1.26)	0.03	(1.27)	0.03	(1.27)
Year = 2015	0.06*	(2.32)	0.06*	(2.33)	0.06*	(2.33)
Constant	12.82***	(19.66)	12.83***	(19.70)	12.83***	(19.73)
N	4,037		4,037		4,037
R² within	.05		.05		.05
R² between	.86		.86		.86
R² overall	.85		.84		.84

Note. IGR = intergovernmental aid; NPV = net property taxable value; MRPV = mean residential property value.

†

p < .10. *p < .05. **p < .01. ***p < .001.

The next set of models estimates the lagged effect of volunteering. The delayed effect is possible because volunteers may require investment for recruitment and training and may not take on substantial roles and therefore free up resources in the first year. Once properly trained, however, volunteers may help to generate tangible cost savings. Because the data are available from 2006 only, I lose 1 year of observations in these models and some city-years with only 1 year of data (Table 3). Using a 1-year lag of the volunteering measures, I find support for this proposition. It appears that volunteer use is significantly and negatively correlated with the budget size in the following year. The effect is robust to different operationalizations of volunteering. These models suggest that 1% increase in the volunteers-to-employees ratio may generate a 0.02% to 0.07% decrease in the total level of city spending (Table 3, Models 4 and 5). On average, cities with volunteer programs manage to reduce their spending by 0.05% in the following year compared to their counterparts not engaging in such programs (Table 3, Model 6). In substantive terms, the results suggest that the average city budget in the sample may be lowered by about US$8,500 and up to US$12,000 with each additional volunteer per 10 paid employees, holding other variables constant. These findings provide initial evidence in support of the proposition that volunteers may need substantial time and resource investment for recruitment and proper training before they can perform important functions generating tangible savings.

Table 3.

Volunteering (Lagged) and the Level of Public Spending.

Variables	Model 4		Model 5		Model 6
Volunteering measure	Ratio (t – 1)		Ratio truncated (t – 1)		Dummy (t – 1)
Volunteering	−0.02*	(−2.48)	−0.07**	(−3.15)	−0.05**	(−3.09)
Paid employees per capita	2.21*	(2.51)	2.24*	(2.53)	2.23*	(2.52)
Managerial functions provided	−0.00	(−0.49)	−0.00	(−0.32)	−0.00	(−0.34)
Services provided	−0.01	(−1.10)	−0.01	(−1.14)	−0.01	(−1.18)
Facilities provided	0.00	(0.34)	0.00	(0.38)	0.00	(0.36)
IGR	0.28***	(4.20)	0.29***	(4.25)	0.29***	(4.26)
Charges for services	0.36**	(2.60)	0.36**	(2.60)	0.36**	(2.62)
Number of seats on the board	0.08*	(2.35)	0.08*	(2.44)	0.08*	(2.40)
CEO elected by commission	0.00	(0.04)	0.01	(0.07)	0.01	(0.08)
Board elected by district/at-large	0.06	(0.91)	0.05	(0.85)	0.06	(0.89)
Board elected by district	0.04	(0.68)	0.03	(0.62)	0.03	(0.65)
Board elections not applicable	−0.12***	(−4.02)	−0.11**	(−3.10)	−0.11**	(−3.16)
Population total	0.11	(1.19)	0.11	(1.20)	0.11	(1.22)
Population growth	−0.08	(−1.11)	−0.08	(−1.04)	−0.07	(−0.98)
Population density	0.00	(0.11)	0.00	(0.08)	0.00	(0.03)
Percent in labor force	−0.40*	(−1.98)	−0.41*	(−2.01)	−0.41*	(−2.00)
Population in poverty	0.00	(0.39)	0.00	(0.40)	0.00	(0.43)
NPV	−0.00	(−0.16)	−0.00	(−0.25)	−0.00	(−0.23)
MRPV	−0.00	(−1.15)	−0.00	(−1.11)	−0.00	(−1.13)
Votes for Democrat	0.14	(0.56)	0.11	(0.43)	0.12	(0.49)
Voter turnout	−0.00	(−0.50)	−0.00	(−0.50)	−0.00	(−0.49)
Year = 2008	−0.03^†	(−1.66)	−0.02	(−1.58)	−0.02	(−1.56)
Year = 2009	−0.00	(−0.23)	−0.00	(−0.10)	−0.00	(−0.11)
Year = 2010	−0.00	(−0.24)	−0.00	(−0.13)	−0.00	(−0.14)
Year = 2011	−0.01	(−0.48)	−0.01	(−0.41)	−0.01	(−0.41)
Year = 2012	−0.01	(−0.25)	−0.00	(−0.16)	−0.00	(−0.16)
Year = 2013	−0.02	(−0.80)	−0.02	(−0.71)	−0.02	(−0.71)
Year = 2014	−0.01	(−0.63)	−0.01	(−0.52)	−0.01	(−0.53)
Year = 2015	0.02	(0.58)	0.02	(0.67)	0.02	(0.64)
Constant	13.36***	(18.21)	13.37***	(18.33)	13.36***	(18.39)
N	3,253		3,253		3,253
R² within	.05		.05		.05
R² between	.71		.72		.71
R² overall	.71		.71		.71

Note. IGR = intergovernmental aid; NPV = net property taxable value; MRPV = mean residential property value.

†

p < .10. *p < .05. **p < .01. ***p < .001.

Given these findings, it becomes particularly important to explore how volunteering is related to the level of paid employees. Is it the case that the cost savings observed in the following year are due to replacement of paid employees by volunteers with a subsequent drop in the payroll? Relevant models are summarized in Table 4 (current year) and Table 5 (volunteering lagged by 1 year). The empirical findings from the models suggest that volunteering is also significantly and negatively correlated with the level of paid employees, but in a different way. A 1% increase in the volunteers to paid employees ratio is associated with a 0.03% reduction in the number of paid employees, but there is no significant lagged effect. These results indicate that we may expect a significant decline in the number of paid employees in the first year, but not in the following year. The dummy indicator is insignificant in all models of employment potentially suggesting that the extent of volunteer involvement may be most important here.

Table 4.

Volunteering and the Number of Paid Employees.

Variables	Model 7		Model 8		Model 9
Volunteering measure	Ratio		Ratio truncated		Dummy
Volunteering	−0.03***	(−3.32)	−0.03	(−0.86)	−0.01	(−0.29)
Managerial functions provided	0.01	(0.85)	0.02	(0.97)	0.02	(0.98)
Services provided	0.00	(0.06)	0.00	(0.02)	0.00	(0.02)
Facilities provided	0.01	(1.00)	0.01	(1.00)	0.01	(1.01)
IGR	−0.07	(−1.13)	−0.07	(−1.16)	−0.07	(−1.15)
Charges for services	−0.15^†	(−1.70)	−0.16^†	(−1.79)	−0.17^†	(−1.82)
Number of seats on the board	−0.01	(−0.15)	−0.01	(−0.17)	−0.01	(−0.21)
CEO elected by commission	−0.12	(−0.96)	−0.12	(−0.97)	−0.12	(−1.00)
Board elected by district/at-large	0.15*	(2.01)	0.14*	(1.97)	0.14*	(1.97)
Board elected by district	−0.03	(−0.36)	−0.03	(−0.40)	−0.03	(−0.40)
Board elections not applicable	0.05	(0.37)	0.01	(0.09)	0.00	(0.02)
Population total	0.30**	(3.03)	0.30**	(3.01)	0.29**	(2.98)
Population growth	−0.14	(−1.19)	−0.13	(−1.14)	−0.14	(−1.16)
Population density	0.00	(0.20)	0.00	(0.25)	0.00	(0.28)
Percent in labor force	−0.17	(−0.56)	−0.21	(−0.68)	−0.21	(−0.69)
Population in poverty	−0.00	(−0.19)	−0.00	(−0.26)	−0.00	(−0.26)
NPV	0.00^†	(1.86)	0.00^†	(1.86)	0.00^†	(1.85)
MRPV	−0.00	(−0.78)	−0.00	(−0.74)	−0.00	(−0.74)
Votes for Democrat	0.78**	(2.89)	0.77**	(2.83)	0.77**	(2.85)
Voter turnout	−0.01*	(−2.20)	−0.01*	(−2.16)	−0.01*	(−2.16)
Year = 2007	0.05**	(3.16)	0.05**	(3.15)	0.05**	(3.16)
Year = 2008	0.04*	(1.98)	0.04*	(2.07)	0.04*	(2.05)
Year = 2009	0.02	(1.03)	0.02	(1.08)	0.02	(1.07)
Year = 2010	0.01	(0.20)	0.01	(0.32)	0.01	(0.34)
Year = 2011	−0.02	(−0.71)	−0.01	(−0.59)	−0.01	(−0.57)
Year = 2012	−0.01	(−0.22)	−0.00	(−0.11)	−0.00	(−0.10)
Year = 2013	−0.01	(−0.25)	−0.00	(−0.16)	−0.00	(−0.14)
Year = 2014	−0.02	(−0.73)	−0.02	(−0.67)	−0.02	(−0.66)
Year = 2015	−0.01	(−0.43)	−0.01	(−0.36)	−0.01	(−0.34)
Constant	0.76	(0.88)	0.78	(0.89)	0.80	(0.92)
N	3,901		3,901		3,901
R² within	.05		.05		.05
R² between	.63		.62		.62
R² overall	.64		.63		.63

Note. The sample size is smaller than 4,037 because observations with one paid employee resulting in zero when log transformed were removed from the analysis. The findings are not affected, but the models perform better. IGR = intergovernmental aid; NPV = net property taxable value; MRPV = mean residential property value.

†

p < .10. *p < .05. **p < .01. ***p < .001.

Table 5.

Volunteering (Lagged) and the Number of Paid Employees.

Variables	Model 10		Model 11		Model 12
Volunteering measure	Ratio (t – 1)		Ratio truncated (t – 1)		Dummy (t – 1)
Volunteering	−0.01	(−0.93)	0.03	(0.82)	0.02	(1.02)
Managerial functions provided	0.02	(0.78)	0.02	(0.83)	0.02	(0.83)
Services provided	−0.01	(−0.78)	−0.01	(−0.78)	−0.01	(−0.76)
Facilities provided	0.00	(0.26)	0.00	(0.25)	0.00	(0.25)
IGR	−0.12	(−1.62)	−0.12^†	(−1.67)	−0.12^†	(−1.67)
Charges for services	−0.16^†	(−1.71)	−0.16^†	(−1.75)	−0.16^†	(−1.76)
Number of seats on the board	0.07	(0.90)	0.06	(0.88)	0.06	(0.89)
CEO elected by commission	−0.04	(−0.55)	−0.05	(−0.57)	−0.05	(−0.58)
Board elected by district/at-large	0.12	(1.53)	0.12	(1.54)	0.12	(1.53)
Board elected by district	−0.01	(−0.11)	−0.01	(−0.12)	−0.01	(−0.13)
Board elections not applicable	0.06	(0.35)	0.05	(0.31)	0.05	(0.31)
Population total	0.30**	(2.68)	0.30**	(2.61)	0.30**	(2.62)
Population growth	−0.22	(−1.39)	−0.21	(−1.38)	−0.22	(−1.39)
Population density	0.00	(0.76)	0.00	(0.81)	0.00	(0.83)
Percent in labor force	−0.04	(−0.13)	−0.04	(−0.14)	−0.04	(−0.14)
Population in poverty	−0.00	(−0.55)	−0.00	(−0.47)	−0.00	(−0.48)
NPV	0.00	(1.29)	0.00	(1.28)	0.00	(1.28)
MRPV	−0.00	(−1.35)	−0.00	(−1.33)	−0.00	(−1.32)
Votes for Democrat	0.67*	(2.12)	0.66*	(2.07)	0.65*	(2.06)
Voter turnout	−0.00	(−1.26)	−0.00	(−1.25)	−0.00	(−1.25)
Year = 2008	−0.01	(−0.70)	−0.01	(−0.64)	−0.01	(−0.65)
Year = 2009	−0.02	(−1.02)	−0.02	(−1.04)	−0.02	(−1.04)
Year = 2010	−0.03	(−1.10)	−0.03	(−1.14)	−0.03	(−1.14)
Year = 2011	−0.05^†	(−1.85)	−0.05^†	(−1.81)	−0.05^†	(−1.81)
Year = 2012	−0.03	(−1.09)	−0.03	(−1.08)	−0.03	(−1.09)
Year = 2013	−0.04	(−1.27)	−0.04	(−1.26)	−0.04	(−1.26)
Year = 2014	−0.04	(−1.41)	−0.04	(−1.41)	−0.04	(−1.41)
Year = 2015	−0.03	(−1.08)	−0.03	(−1.06)	−0.03	(−1.05)
Constant	0.42	(0.40)	0.45	(0.43)	0.45	(0.43)
N	3,143		3,143		3,143
R² within	.04		.04		.04
R² between	.65		.65		.65
R² overall	.66		.66		.66

Note. IGR = intergovernmental aid; NPV = net property taxable value; MRPV = mean residential property value.

†

p < .10. *p < .05. **p < .01. ***p < .001.

Because we observe employee reduction in Year 1 with no budget reductions, and budget reductions in Year 2 without employee reductions, this picture, although generally within our expectations, is still somewhat perplexing. Why would there be no budget reduction in Year 1 if the level of employment is significantly lower? It is possible that while interchangeability between paid and unpaid employees is in process, substantial administrative costs are incurred to manage the change, which eats up any cost saving from volunteer engagement. Once the transformation is complete, we begin to see tangible cost reductions. This would suggest we might expect an increase in administrative costs associated with volunteering programs (Handy & Mook, 2011).

To check the plausibility of this explanation, I estimate the relationship between volunteer use and the total level of city general government expenditures, which includes human resources and financial administration. The results summarized in Table 6 indicate that volunteering is indeed associated with significant increases in general government expenditures. A 1% increase in the volunteers to employees ratio may be associated with a 0.07% increase in the total level of general government expenditures. On average, this category of expenditures in the cities that engage volunteers is 0.05% higher than in the cities that do not. There is no lagged relationship between volunteers and general government expenditures.³

Table 6.

Volunteering and General Government Expenditures.

Variables	Model 13		Model 14		Model 15
Volunteering measure	Ratio		Ratio truncated		Dummy
Volunteering	0.01	(0.75)	0.07**	(2.80)	0.05**	(2.77)
Paid employees per capita	2.61**	(2.76)	2.67**	(2.79)	2.60**	(2.76)
Managerial functions provided	0.01	(0.81)	0.01	(0.78)	0.01	(0.78)
Services provided	−0.00	(−0.01)	−0.00	(−0.02)	−0.00	(−0.00)
Facilities provided	−0.01^†	(−1.71)	−0.01^†	(−1.74)	−0.01†	(−1.73)
IGR	0.05	(0.78)	0.05	(0.79)	0.05	(0.79)
Charges for services	−0.03	(−0.32)	−0.03	(−0.34)	−0.03	(−0.31)
Number of seats on the board	0.07	(1.11)	0.07	(1.09)	0.07	(1.09)
CEO elected by commission	−0.07	(−0.81)	−0.07	(−0.86)	−0.07	(−0.88)
Board elected by district/at-large	0.07	(0.49)	0.07	(0.50)	0.07	(0.50)
Board elected by district	0.03	(0.63)	0.03	(0.66)	0.03	(0.63)
Board elections not applicable	−0.17	(−1.58)	−0.18^†	(−1.88)	−0.18^†	(−1.82)
Population total	0.18	(1.51)	0.18	(1.51)	0.18	(1.52)
Population growth	−0.03	(−0.30)	−0.04	(−0.38)	−0.04	(−0.39)
Population density	−0.00	(−0.37)	−0.00	(−0.37)	−0.00	(−0.39)
Percent in labor force	−0.25	(−1.08)	−0.25	(−1.09)	−0.25	(−1.08)
Population in poverty	0.00	(1.08)	0.00	(1.11)	0.00	(1.08)
NPV	0.00	(1.10)	0.00	(1.06)	0.00	(1.06)
MRPV	−0.01*	(−2.22)	−0.01*	(−2.17)	−0.01*	(−2.18)
Votes for Democrat	0.68*	(2.57)	0.69**	(2.60)	0.68*	(2.55)
Voter turnout	−0.00	(−0.10)	−0.00	(−0.10)	−0.00	(−0.10)
Year = 2007	0.08***	(4.20)	0.08***	(4.23)	0.08***	(4.20)
Year = 2008	0.11***	(5.97)	0.11***	(5.88)	0.11***	(5.88)
Year = 2009	0.10***	(4.68)	0.09***	(4.62)	0.09***	(4.61)
Year = 2010	0.11***	(4.75)	0.11***	(4.77)	0.11***	(4.75)
Year = 2011	0.13***	(5.09)	0.13***	(5.07)	0.13***	(5.07)
Year = 2012	0.15***	(5.85)	0.14***	(5.80)	0.14***	(5.80)
Year = 2013	0.16***	(6.20)	0.16***	(6.21)	0.16***	(6.22)
Year = 2014	0.21***	(7.99)	0.20***	(7.95)	0.20***	(7.99)
Year = 2015	0.20***	(6.54)	0.20***	(6.54)	0.20***	(6.54)
Constant	10.57***		10.59***		10.58***
N	4,037		4,037		4,037
R² within	.10		.10		.10
R² between	.59		.59		.60
R² overall	.56		.55		.56

Note. IGR = intergovernmental aid; NPV = net property taxable value; MRPV = mean residential property value.

†

p < .10. *p < .05. **p < .01. ***p < .001.

Thus, these findings support the notion expressed in the literature that the initiation of a volunteer program may require time and money investment, which, in our case, is evident in significant increases of the general government expenditures. This may explain why we do not see significant net savings in the first year despite the reductions in the number of paid employees. Once the transformation process is complete and the volunteers are properly trained, their involvement may bring net benefits, which we can observe in the second year. These finding provide empirical support to our main hypothesis that predicts significant negative association of volunteering with expenditures and paid employees.

It remains unclear, however, whether volunteers are truly associated with costs savings. Namely, given the high chance of paid employee reduction, there is a possibility that cities undergo downsizing and engage volunteers temporarily simply to smooth off the transition. The smaller budgets observed afterward may thus result from fewer services provided rather than volunteering efforts. If this scenario is true, it suggests that volunteering may be associated with fewer services provided by the city. To test this alternative, I explore the relationship between volunteering and the number of services, facilities, and managerial functions provided by each city.⁴ These three outcome variables capture the main functional responsibilities performed by the cities in Georgia. The full list of services, functions, and facilities is provided in the appendix.

The results summarized in Table 7 suggest that volunteering is not correlated with the number of services, functions, or facilities the cities provide. There is no lagged effect either.⁵ It appears that these outcomes are mostly driven by the size of city population and time trends captured by the year dummies.

Table 7.

Volunteering and the Number of Functions, Services, and Facilities.

Variables	Model 16		Model 17		Model 18
Outcome	Managerial functions		Services		Facilities
Volunteering	0.00	(−1.64)	0.00	(0.12)	0.00	(0.09)
Paid employees per capita	−0.00*	(−2.05)	−0.00***	(−3.63)	0.00	(−0.35)
Financial director employed	−0.01	(−0.56)	−0.01	(−0.83)	0.00	0.00
Regulatory fees used	0.01	(0.71)	−0.01	(−0.85)	0.00	(−0.06)
User fees used	0.02*	(2.03)	0.01	(1.32)	0.02	(0.86)
IGR	0.01	(0.83)	−0.01	(−0.50)	−0.03	(−1.09)
Charges for services	0.04*	(2.46)	0.03*	(2.19)	0.03	(0.89)
Number of seats on the board	−0.01	(−0.65)	0.00	(0.49)	0.01	(0.49)
CEO elected by commission	−0.03	(−1.47)	0.01	(0.23)	−0.04	(−0.40)
Board elected by district/at-large	0.02	(1.12)	0.03	(0.96)	0.04	(0.60)
Board elected by district	0.02	(1.40)	0.02	(1.22)	0.03	(0.62)
Board elections not applicable	−0.08	(−1.27)	−0.04	(−0.72)	−0.14	(−0.98)
Population total	0.06***	(7.89)	0.07***	(8.98)	0.14***	(7.61)
Population growth	−0.03	(−1.41)	−0.02	(−1.28)	−0.01	(−0.36)
Population density	−0.00^†	(−1.72)	0.00	(0.02)	−0.00^†	(−1.82)
Percent in labor force	−0.01	(−0.15)	0.02	(0.27)	−0.16	(−1.02)
NPV	0.00	(−0.79)	0.00	(−0.21)	0.00	(−0.61)
MRPV	0.00	(0.07)	0.00	(1.60)	0.00	(−0.82)
Population in poverty	0.00	(−1.52)	0.00	(−0.54)	0.00	(−1.61)
Votes for Democrat	0.04	(0.98)	−0.04	(−0.98)	−0.18	(−1.61)
Voter turnout	0.00	(−0.46)	0.00	(1.02)	0.00	(1.19)
Year = 2007	−0.01**	(−2.78)	0.00	(−0.62)	0.02*	(2.51)
Year = 2008	−0.01*	(−2.06)	0.01	(1.46)	0.04***	(4.08)
Year = 2009	0.00	(−0.70)	0.01	(1.21)	0.04***	(3.72)
Year = 2010	0.00	(0.44)	0.01	(1.59)	0.05***	(3.47)
Year = 2011	0.01	(0.85)	0.02***	(3.41)	0.05***	(3.85)
Year = 2012	0.01	(0.91)	0.02**	(3.01)	0.06***	(3.78)
Year = 2013	0.01	(0.94)	0.02***	(3.32)	0.07***	(4.23)
Year = 2014	0.01	(1.41)	0.02**	(3.26)	0.06***	(3.35)
Year = 2015	0.01	(1.21)	0.02*	(2.42)	0.05**	(2.65)
Constant	2.04***	(31.45)	2.20***	(32.16)	1.26***	(6.59)
N	3,976		3,976		3,976

Note. IGR = intergovernmental aid; NPV = net property taxable value; MRPV = mean residential property value.

†

p < .10. *p < .05. **p < .01. ***p < .001.

These results corroborate the notion that volunteers may help reduce the costs of public service provision, most likely by reducing the need for paid employees, without significant reductions in the service outputs. In short, public sector volunteers may allow the delivery of the same level of services but at a lower cost.

Conclusion

It has been argued in the literature that volunteer involvement in the co-delivering of public services, among other things, has the potential to generate tangible costs savings. This work represents one of the first attempts to test these predictions in a longitudinal study of general-purpose local governments with different levels of volunteer engagement. The main conclusion of the study is that volunteer programs may have a significant and prolonged impact on government fiscal behavior. It appears that higher levels of volunteer involvement, as measured by a volunteers-to-employees ratio, are unlikely to generate cost savings immediately. Indeed, initiation of volunteer programs may significantly increase administrative costs necessary to recruit, train, and manage citizen volunteers. Because volunteering engagement is associated with significant reductions in paid staff, this temporary rise in administrative costs does not result in net cost increases. Once properly trained, however, volunteers may help provide similar levels of public services with fewer paid employees and therefore less spending. These findings provide empirical support to one of the main arguments in the literature that public sector volunteering may help generate cost savings. By exploring the relationship between the use of volunteers and public budgets and employees over time, this study improves our understanding of short- and long-term implications of citizen volunteering in the public sector.

This study has some important policy implications for public administrators. First, the common perception that volunteers represent “free” labor force may be inaccurate in practice. Public managers should plan to allocate funds for an investment in volunteering programs, especially for recruitment and training. Second, volunteers may not be able to undertake important responsibilities and save resources in early stages of the program. However, with more training and experience, the initial investment can be expected to generate some returns. This implies that along with recruitment and proper training, efforts should be made to ensure volunteer retention for the benefits to occur. Without such retention strategies, public organizations may end up being worse off with volunteering programs due to volunteer turnover. Third, most of the savings in this study of the city governments in Georgia likely came from reductions in paid employees and interchangeability with volunteers. Volunteers thus helped city administration to deliver the same level of services with fewer paid employees. This displacement does not need to be the case, and government officials may be able to save resources by relying on volunteers for government transition through service expansion and during peak loads.

Although the results of this study are based on cities within a single state, these findings may be generalizable beyond the state of Georgia. The extent of generalizability depends on whether the cities in other states are similar to those included in this study. The status and powers of local governments are determined by state institutions, which jointly establish state systems of local governments. These systems differ from state to state (Krane, Rigos, & Hill, 2001), which, in practice, means that city officials in two different states possess different levels of authority and autonomy. However, there are commonalities among states in the way they organize their local units (Stephens & Wikstrom, 2000), which implies that these findings may be generalizable beyond the state of Georgia.

It is also important to note that Georgia is a so-called “employment-at-will” state, which in this state means that “in the absence of a written contract of employment for a defined duration, either party may terminate the employment relationship for good cause, bad cause, or no cause at all, as long as it is not for an illegal cause” (Butler, 2017, p. 38). The employment-at-will doctrine varies from state to state (Muhl, 2001), but the implication for generalizability of this study is that the extent of expected interchangeability between volunteers and paid employees may vary depending on how the doctrine is construed. Government employees in the states with fewer exception to the doctrine, like Georgia, can be replaced more easily than public employees in the states with stronger civil service protections. A public official in the latter group, however, may still find volunteering helpful as an enhancement for paid employees during peak loads and expansion. On a related note, future studies could also contribute to the literature by exploring the effects of volunteering on the level of paid employees and public budgets in the cities from states with different levels of unionization.

Footnotes

Appendix

Acknowledgements

The author would like to thank Seong Cheol Kang and the anonymous reviewers for their help with this manuscript.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Notes

ORCID iD

Mikhail Ivonchyk

Author Biography

Mikhail Ivonchyk obtained his doctoral degree from the Department of Public Administration and Policy at University of Georgia. His current research focuses on public management, financial administration, policy adoption and implementation.

References

Brudney

J. L.

(1984). Local coproduction of services and the analysis of municipal productivity. Urban Affairs Review, 19, 465-484.

Brudney

J. L.

(1990a). The availability of volunteers implications for local governments. Administration & Society, 21, 413-424.

Brudney

J. L.

(1990b). Expanding the government-by-proxy construct: Volunteers in the delivery of public services. Nonprofit and Voluntary Sector Quarterly, 19, 315-328.

Brudney

J. L.

(1990c). Fostering volunteer programs in the public sector. San Francisco, CA: Jossey-Bass.

Brudney

J. L.

(1993). Volunteer involvement in the delivery of public services: Advantages and disadvantages. Public Productivity & Management Review, 16, 283-297.

Brudney

J. L.

(1999). The effective use of volunteers: Best practices for the public sector. Law and Contemporary Problems, 62, 219-255.

Brudney

J. L.

(2005). Designing and managing volunteer programs. In Renz

D. O.

Herman

R. D.

(Eds.), The Jossey-Bass handbook of nonprofit leadership and management (pp. 310-344). San Francisco, CA: Jossey-Bass.

Brudney

J. L.

Duncombe

W. D.

(1992). An economic evaluation of paid, volunteer, and mixed staffing options for public services. Public Administration Review, 52, 474-481.

Brudney

J. L.

Gazley

(2002). Testing the conventional wisdom regarding volunteer programs: A longitudinal analysis of the service corps of retired executives and the U.S. Small Business Administration. Nonprofit and Voluntary Sector Quarterly, 31, 525-548.

10.

Brudney

J. L.

, et al. (2017). Volunteers in the public sector. In Farazmand

(Ed.), The global encyclopedia of public administration, public policy, and governance. New York, NY: Springer.

11.

Brudney

J. L.

Kellough

J. E.

(2000). Volunteers in state government: Involvement, management, and benefits. Nonprofit and Voluntary Sector Quarterly, 29, 111-130.

12.

Brudney

J. L.

Warren

(1990). Multiple forms of volunteer activity in the public sector: Functional, structural, and policy dimensions. Nonprofit and Voluntary Sector Quarterly, 19, 47-58.

13.

Brunet

DeBoer

McNamara

K. T.

(2001). Community choice between volunteer and professional fire departments. Nonprofit and Voluntary Sector Quarterly, 30, 26-50. doi:10.1177/0899764001301002

14.

Butler

(2017). Employer handbook. Georgia Department of Labor. Retrieved from https://dol.georgia.gov/sites/dol.georgia.gov/files/related_files/document/dol224.pdf

15.

Carr

J. B.

Karuppusamy

(2010). Reassessing the link between city structure and fiscal policy: Is the problem poor measures of governmental structure? The American Review of Public Administration, 40, 209-228. doi:10.1177/0275074009334641

16.

Coursey

D. H.

Perry

J. L.

Brudney

J. L.

Littlepage

(2008). Psychometric verification of Perry’s public service motivation instrument: Results for volunteer exemplars. Review of Public Personnel Administration, 28, 79-90.

17.

Dobrin

Wolf

(2016). What is known and not known about volunteer policing in the United States. International Journal of Police Science & Management, 18, 220-227.

18.

Donahue

A. K.

(2004). The influence of management on the cost of fire protection. Journal of Policy Analysis and Management, 23, 71-92.

19.

Dover

G. J.

(2010). Public sector volunteering: Committed staff, multiple logics, and contradictory strategies. Review of Public Personnel Administration, 30, 235-256.

20.

Duncombe

Brudney

J. L.

(1995). The optimal mix of volunteer and paid staff in local governments: An application to municipal fire departments. Public Finance Quarterly, 23, 356-384.

21.

Duncombe

Yinger

(1997). Why is it so hard to help central city schools? Journal of Policy Analysis and Management, 16, 85-113.

22.

Ellis

S. J.

Noyes

K. H.

(1990). By the people: A history of Americans as volunteers. San Francisco, CA: Jossey-Bass.

23.

Emanuele

(1996). Is there a (downward sloping) demand curve for volunteer labour? Annals of Public Cooperative Economics, 67, 193-208.

24.

Eom

T. H.

Rubenstein

(2006). Do state-funded property tax exemptions increase local government inefficiency? An analysis of New York State’s STAR program. Public Budgeting & Finance, 26, 66-87.

25.

Feiock

R. C.

Jeong

M.-G.

Kim

(2003). Credible commitment and council-manager government: Implications for policy instrument choices. Public Administration Review, 63, 616-625. doi:10.1111/1540-6210.00324

26.

Fischer

L. R.

Schaffer

K. B.

(1993). Older volunteers: A guide to research and practice. Thousand Oaks, CA: Sage.

27.

Frederickson

H. G.

Johnson

G. A.

(2001). The adapted American city: A study of institutional dynamics. Urban Affairs Review, 36, 872-884. doi:10.1177/10780870122185127

28.

Gazley

Brudney

J. L.

(2005). Volunteer involvement in local government after September 11: The continuing question of capacity. Public Administration Review, 65, 131-142. doi:10.1111/j.1540-6210.2005.00439.x

29.

Georgia State Government. (2017). Cities & counties in the State of Georgia. Retrieved from https://georgia.gov/municipality-list

30.

Grosskopf

Hayes

K. J.

Taylor

L. L.

Weber

W. L.

(2001). On the determinants of school district efficiency: Competition and monitoring. Journal of Urban Economics, 49, 453-478. doi:10.1006/juec.2000.2201

31.

Gujarati

(2007). Basic econometrics (4th ed.). New York, NY: McGraw-Hill Education.

32.

Halaby

C. N.

(2004). Panel models in sociological research: Theory into practice. Annual Review of Sociology, 30, 507-544.

33.

Handy

Mook

(2011). Volunteering and volunteers: Benefit-cost analyses. Research on Social Work Practice, 21, 412-420.

34.

Handy

Mook

Quarter

(2008). The interchangeability of paid staff and volunteers in nonprofit organizations, 37, 76-92. doi:10.1177/0899764007303528

35.

Handy

Srinivasan

(2004). Valuing volunteers: An economic evaluation of the net benefits of hospital volunteers. Nonprofit and Voluntary Sector Quarterly, 33, 28-54.

36.

Hatry

H. P.

(1983). Review of private approaches for delivery of public services. Washington, D.C. : Urban Institute Press

37.

Hayes

Razzolini

Ross

(1998). Bureaucratic choice and nonoptimal provision of public goods: Theory and evidence. Public Choice, 94(1-2), 1-20.

38.

Hilke

(1986). The impact of volunteer firefighters on local government spending and taxation. Municipal Finance Journal, 7, 33-44.

39.

Karuppusamy

Carr

J. B.

(2012). Interjurisdictional competition and local public finance: Assessing the modifying effects of institutional incentives and fiscal constraints. Urban Studies, 49, 1549-1569. doi:10.1177/0042098011415435

40.

Kiser

L. L.

Percy

S. L.

(1980). The concept of coproduction and its implications for public service delivery. Paper presented at the Workshop in Political Theory and Policy Analysis. Retrieved from https://dlc.dlib.indiana.edu/dlc/handle/10535/1466

41.

Krane

Rigos

P. N.

Hill

M. B.

(2001). Home rule in America: A fifty-state handbook. Washington, DC: CQ Press.

42.

Leibenstein

(1966). Allocative efficiency vs. “x-efficiency.” The American Economic Review, 56, 392-415. doi:10.2307/1823775

43.

Leibenstein

(1978). On the basic proposition of x-efficiency theory. The American Economic Review, 68, 328-332.

44.

Levine

C. H.

Fisher

(1984). Citizenship and service delivery: The promise of coproduction. Public Administration Review, 44, 178-189.

45.

Muhl

C. J. J. M. L. R

. (2001). The employment-at-will doctrine: Three major exceptions. Retrieved from https://www.bls.gov/opub/mlr/2001/01/art1full.pdf

46.

Nabatchi

Leighninger

(2015). Public participation for 21st century democracy. Hoboken, NJ: John Wiley.

47.

Nesbit

Brudney

J. L.

(2010). At your service? Volunteering and national service in 2020. Public Administration Review, 70(Suppl. 1), S107-S113.

48.

Nesbit

Brudney

J. L.

(2013). Projections and policies for volunteer programs: The implications of the serve America act for volunteer diversity and management. Nonprofit Management and Leadership, 24, 3-21.

49.

Neshkova

M. I.

Guo

(2012). Public participation and organizational performance: Evidence from state agencies. Journal of Public Administration Research and Theory, 22, 267-288. doi:10.1093/jopart/mur038

50.

Parks

R. B.

Baker

P. C.

Kiser

Oakerson

Ostrom

Wilson

(1981). Consumers as coproducers of public services: Some economic and institutional considerations. Policy Studies Journal, 9, 1001-1011.

51.

Pynes

J. E.

(2008). Human resources management for public and nonprofit organizations: A strategic approach (Vol. 30). Hoboken, NJ: John Wiley.

52.

Reeves

T. Z.

(1988). The politics of the Peace Corps & VISTA. Tuscaloosa: University of Alabama Press.

53.

Russel

A. R.

Mook

Handy

(2017). Interchangeability of labor: Managing a mixed paid and volunteer workforce. In Sowa

J. E.

Word

J. K. A.

(Eds.), The nonprofit human resource management handbook: From theory to practice (pp. 271-284). New York, NY: Taylor & Francis.

54.

Sowa

J. E.

Word

J. K. A.

(2017). The nonprofit human resource management handbook: From theory to practice. New York, NY: Taylor & Francis.

55.

Stebbins

R. A.

(2009). Would you volunteer? Society, 46, 155-159.

56.

Stephens

Wikstrom

(2000). Metropolitan government and governance: Theoretical perspectives, empirical analysis, and the future. New York, NY: Oxford University Press.

57.

Stukas

A. A.

Hoye

Nicholson

Brown

K. M.

Aisbett

(2016). Motivations to volunteer and their associations with volunteers’ well-being. Nonprofit and Voluntary Sector Quarterly, 45, 112-132.

58.

Svara

J. H.

(2005). Exploring structures and institutions in city government. Public Administration Review, 65, 500-506. doi:10.1111/j.1540-6210.2005.00475.x

59.

Wooldridge

(1999). Distribution-free estimation of some nonlinear panel data models. Journal of Econometrics, 90, 77-97. doi:10.1016/S0304-4076(98)00033-5

60.

Wooldridge

(2013). Introductory econometrics: A modern approach (5th ed.). Mason, OH: Cengage Learning.