Enterprising government: the political and financial effects of fee-supported municipal services.

Jang, Hee Soun ; Kwon, Myungjung

Since the tax revolt of the 1970's, all levels of governments--federal, state, and local--have been pursuing various ways to overcome limited revenue sources. Cities often face more severe financial difficulties and lack of funding sources compared to other levels of government because they have fewer ways of generating revenue resources. Municipalities, whose general funds are mainly covered by property taxes, have a hard time balancing their budget since property tax has been capped and the creation of new taxes has been severely constrained in many states. Since typical city services are public safety and basic maintenance services--such as water, gas, electricity, sewage, and garbage collection--the funding decisions for those services immediately affect city residents and may turn into politically sensitive issues for the elected officials and the local voters.

In order to overcome these financial and political obstacles, municipalities often use the fee-supported service model for many of their public services outside of the tax-supported general fund services, because fee-supported services charge customers, not the community as a whole, and possibly generate profits. In other words, fee-supported services are voluntarily purchased public goods and the prices of these services are supposed to reflect the true cost of production including capital investment (Bierhanzl & Downing, 1998, 2004; Government Finance Officers Association, 2001; Mikesell, 2007). Since a user charge is a market-like financing which shifts general fund activities to fee-supported enterprise fund activities, the usage of user fee-supported goods and services has become a popular strategic effort for cities to achieve more responsive and efficient finance systems (Bailey 1994; Netzer 1992; Stumm and Kahn 1996; Stumm 2000; Wagner 1976; Zorn 1991).

This research argues that municipalities have shifted more towards reliance on user charges than general tax financing to generate their revenue and contextual factors, such as political institution, fiscal, economic, and demographic factors have influenced this shift. According to the U.S. Census, user charges made up 9% of total local revenue and property taxes generated 37% of the total in 1970. However, in 2006, 14% of total local revenue came from user charges and 24% from property taxes (U.S. Census). Previous studies also show that the increased reliance by local governments on user charge financing is due to its more efficient way of providing local services than tax revenue (Baily 1994; Sun & Jung 2012; Wagner 1976). However, despite the efficiency gains from the greater degree of reliance on user charge financing, cities' motivations for using fee-supported goods vary according to financial, managerial, social and political attributes affecting their decision making and some cities are actively involved in the provision of fee based enterprise services, but some cities are not.

In this study, we explore municipalities' strategic behaviors of introducing and increasing fee-supported enterprise services to respond to financial necessities and political pressures. Since enterprise funding is a required financing methodology for fee-supported service delivery,1 municipal revenue appropriated under enterprise funds have been used as an indicator to understand the usage of fee-supported service delivery in previous empirical studies (Bunch, 2000; Bunch & Ducker, 2003; DeHoog & Swanson, 1988; Gitajn, 1984; Molinari & Tyer, 2003; Rubin, 1985, 1988; Stumm, 1997, 2000; Stumm & Khan, 1996; Tyer, 1989). For empirical analysis, this research examines enterprise fund revenue to investigate introduction and reliance of fee-supported public services in municipalities and the terms fee-supported (based) services, user fee services, enterprise services and enterprise funds--are used interchangeably throughout this paper. The research questions that we examine are: Why do some cities use a fee-supported enterprise service model? Why do some of them rely more on fee-supported enterprise services than other cities? And descriptive analyses of Florida finance data elaborate on what types of local services are being chosen as fee-supported activities. These questions will be answered with testable hypotheses by employing the two stage Heckman model for empirical analysis with finance data from Florida municipalities. This research will examine the expected advantages and potential disadvantages of using fee supported services and examine how cities in Florida use the fee supported enterprise services model in diverse service areas.

Fee-Supported Enterprise Fund Services

Previous studies observed that an increasing number of municipal governments provided services through fee-supported financing systems. Most recently Sun and Jung (2012) examined longitudinal data of sewer, and park and recreation services and found fee-supported services led significant reduction in service expenditures. Bunch (2000), in her nation-wide survey of large cities, revealed the increased usage of fee-supported enterprise services and found services like solid waste and drainage were newly added to the enterprise fiscal system in the 1990s. Molinari and Tyer (2003) also reported that cities and counties in South Carolina showed a significant increase in fee-supported goods and services even when they were compared to the increase of tax-supported general fund activities.

It has been noted that in some cities, profits generated from fee-supported services transferred to the general tax revenue. Seminal research done by Colberg (1955), about enterprise fund use in Florida cities, found that cities provide electric utility services presented higher general total expenditures than elsewhere due to profits generated from electric utility services that were transferred to general fund. DiLorenzo (1982), in his analysis of a sample of fifty-two cities in the state of New York, examined electric utility services that increased per capita city spending and speculated that profits from the fee based service subsidized the general fund. DeHoog and Swanson (1988) tested the tax substitution effect of electric utilities by comparing cities that provided electric utility services and cities without their own electric services in Florida. They found that there was no significant difference in per capita total expenditures; however, it was observed that cities without electric services showed significantly higher property tax revenues and per capita total revenue. They concluded that cities with electric utility services relied less on general tax revenue since they used utility profit transfers as a substitute for tax revenue. Tyer (1989), in his study of cities in South Carolina, confirmed prior findings of profit-generating nature of fee-supported services. He reported that the cities with electric services transferred revenues to the general fund and had higher average per capita expenditures and lower revenue from property taxes.

Although earlier research efforts provided a solid foundation for the understanding of the financial effects of fee-supported services, much of the general effects of using and increasing fee-supported services remain unexplored. With a few exceptions (Stumm & Khan, 1996; Sun & Jung, 2012; Rubin, 1988), prior studies mainly relied on either comparison between cities with fee-supported services and those without, or descriptive reports of fee-supported service usages. It is because most of these studies focus on either a single utility fee-based service or a few selected services only; however, in order to understand the municipalities' reliance on fee-supported services, we argue it is reasonable to look at total revenue that comes from enterprise service as an aggregated measure of enterprise funds from the total revenue. If there is enough variation that cities have any level of fee based services or cities do not have any of them at all, we also need to explore why some of cities do not provide services via a fee-supported funding system. In order to fully examine the variety of motivations for using fee based services, we must account for multiple factors including financial, managerial, social and political attributes affecting municipal decision making. This research builds on the earlier studies mentioned above to develop a theory on the utilization of fee-supported services by building testable hypotheses with systematic empirical analyses while addressing the earlier studies' limitations.

THEORY AND TESTABLE HYPOTHESES

Fee-supported services are expected to create a quasi-market mechanism in the local finance since customers who use the services are expected to pay a share of the costs of service production and the costs of the original capital investments (Mikesell, 2007; Zorn, 1991). Fee-supported services which rely on user charges, have a separate enterprise finance system that measures flow of economic resources, and expected to have different financial impact to the municipalities. It has been also noted that fee-supported services would be a more efficient way to charge the use of those source based services rather than services charged based on residence (Bailey, 1994; Bunch & Ducker, 2003; Inman, 2005; Rubin, 1988). While users of fee based services pay all or part of the cost of the goods or services and they don't have to be voting taxpayers, residents of the community who pay local taxes, cover non-specified general service functions of the city.

As Tiebout (1956) proposed, fragmented local governments are municipal corporations controlled by homeowners whose preferences are revealed by their choices of the communities where they live. The Tiebout model depicts property tax payers as the consumer voters who "vote with their feet" and regard their property taxes as benefit taxes contributed to the services which keep up the values of their properties. Along this urban economic perspectives, Peterson (1981) in City Limits argued that local governments are sensitive to the developmental needs of their community and political leaders are pressured by local tax payers to spend tax-supported general fund for the direct and indirect services that only benefit city residents. Furthermore, in the local decision-making process, homeowners are shareholders of municipalities and push city governments to utilize collected local taxes for services that maximize their property values; this is why property taxes have been conventionally spent in general fund services. Oates (1969) argued that local voters will agree to pay property taxes if they believe those taxes will be capitalized into the added value of owner-occupied homes. Ideally this benefit tax logic of property taxation allows cities to avoid free-rider problems of geographically limited public goods (Fischel, 2001a, 2001b; Tiebout, 1956).

By the same token, since costs of fee based enterprise goods would not be subsidized by property taxes paid by city residents, services funded though user charges promote a sense of fairness in that customers pay their share of the costs for the amount of use of the services (Bierhanzl & Downing, 1998, 2004;; Bunch, 2000; Molinari & Tyer, 2003; Tyer, 1989). However, it is important to note that the local tax and expenditure limitations make it difficult to realize optimal efficiency of benefit tax from property taxation. This is one of reason why local governments push to diversify their finance mechanisms and increase more fee based enterprise services to charge service users rather than city residents who pay property taxes in the district.

Another major benefit of fee-supported services comes from the ability to lower the tax burden on property tax payers by substituting for property tax-based general fund services (Bunch, 2000; DeHoog & Swanson, 1988; McGuire, 2001; Molinari & Tyer 2003; Vogt, 1978). Since enterprise finances are self-supporting, the use of enterprise services helps cities to expand services and keeps the quality of services high without substantial tax increases. Some cities use profits from fee-supported services to lower the tax burden of citizens even more by transferring revenue from an enterprise fund account to the general fund account. DiLorenzo (1982) reported that profits, created by fee-supported services, were transferred to general funds in cities in New York. A study done by the State of Minnesota (2004) reported that cities in Minnesota make policy decisions regarding their use of transferred profits to lower the tax burden of local citizens because fee-supported enterprise services are profitable. Tyer (1989) also examined profit transfers from enterprise fund in cities in South Carolina by comparing locally raised revenue per capita from cities with enterprise services and cities without. He concluded that cities with enterprise utility services subsidized the property tax generated general fund. (2) Tyer (1989) found that South Carolina cities with electricity enterprise services subsidize the property tax by substituting enterprise service profits from residents as well as, and more importantly, non-residents of the city. Research done by DeHoog and Swanson (1988) also reported that cities in Florida justified inter-fund transfers from enterprise fund based city utilities to the general fund.

If we can assume that city residents are sensitive to tax-supported general fund services which funded mainly through property tax, and they are aware of the differences of service levels among neighboring communities, it is logical to expect that city residents welcome fee-supported enterprise services, which enhance the sense of fairness over general fund services which come mostly from property taxes (Fischel, 2001a; 2001b). Thus we propose that when cities have only a small percentage of revenue coming from property taxes, they are more likely to provide services through fee-supported enterprise funds and increase fee-supported services since the capacity to provide general fund based services are very limited. Thus fee-supported enterprise services are more preferred for public service arrangement of communities which present low property tax proportion from total revenue since enterprise service charges may subsidize general funds (Bunch, 2000; Vogt 1978). We also consider the per capita total tax as an indicator of the tax burden to the community residents and anticipate that cities with a higher tax burden would want to avoid tax increases by using more fee-supported services (Molinari & Tyer, 2003). The reluctance of taxpayers to pay higher taxes requires municipal decision makers to increase non-tax revenue to avoid tax increases and expand local public services.

Financial independence provides a city with autonomy regarding how it operates its revenue. Cities receive revenues from various channels including their own sources, such as property taxes, sales taxes, and service charges, and external intergovernmental sources like state and federal aids. The varying ratio between a city's own source of revenue and intergovernmental revenue affect its financial system differently, especially in terms of a city's discretion to use those sources of revenues. If a city heavily relies on its intergovernmental aid, it may lack the ability to control its own finances since intergovernmental revenue from higher governments is distributed with specific use only category and strict strings attached from the jurisdictions which provided the funds. High reliance on intergovernmental revenue could be a sign of a city in poor fiscal health and it may limit the city to participate in autonomous planning activities (Hanna, 2005). Also, heavy reliance on intergovernmental revenue would limit the city's discretionary authority to design its own finance. However, if a city maintains its revenue from mainly its own sources, it would independently decide how to use its revenue, maintain its finance system, and be more likely to be innovative and autonomous by increasing market-like systems of enterprise funds to please its local tax paying residents (Oates, 2001). This research proposes that cities with a high percentage of revenue from intergovernmental sources are less likely to increase enterprise fund services and have fewer services provided under an enterprise finance system since heavy dependence upon other government sources would not allow cities financial discretionary room to create their own service systems. It is also expected the user fees are welcomed by local taxpaying residents as an alternative way of increasing a city's revenue. Financial independence was measured as the percentage of total revenue that comes from a city's own sources. The hypothesized associations of the financial conditions and the use of fee-supported enterprise services are presented as follows.

Hypothesis 1. Property tax revenue of municipalities is inversely associated with the use of fee-supported services.

Hypothesis 2. Higher per capita tax burden of municipalities is positively associated with use of fee-supported services.

Hypothesis 3. Own source revenue of municipalities is positively associated with use of fee-supported services.

The creation of a more customer-sensitive managerial environment is another benefit that may affect local government when using fee-supported services (Bunch, 2000; Osborne & Gaebler, 1992). Since the nature of fee-supported services is self-sustaining through fees and charges, users of those services are expected to pay a price which is closer to the true production cost and their fair share of capital investment costs. The revelation of true cost brings efficiency to municipal government and forces public managers and service providers to behave much more like market actors in order to achieve optimum service prices and quality (Feiock & Clingermayor, 2001). From a managerial perspective, fee-supported services are financially conservative and stable choices since service costs incurred from the service investment and production are covered by customer charges. One of the recommendations from Osborne and Gaebler (1993) in Reinventing Government is to be more conscious about spending and more innovative about earning revenues. Thus, the efficiency-seeking nature of fee-supported enterprise services may motivate professional city managers to pursue more fee based services as they work for efficient service management and financial accountability of their cities (Rubin, 1988). Fee-supported services may also provide residents with more choices regarding which services to consume and how much of each service to consume (Bunch & Ducker, 2003).

We should also note that there are divergent theories in terms of explaining the bureaucratic behavior of using fee-supported services. Elected politicians may enjoy the low political visibility of enterprise services by transferring profits generated from fee-supported enterprise funds to the tax-supported general funds (Buchnan, 1967; Wagner, 1976;). Politicians expect to enjoy surplus transfers since the public is neither aware of the complicated city budget process nor the true price of enterprise services. Thus elected mayors in strong mayor forms of government may prefer to generate profits from enterprise funds to reduce the tax burden of tax-paying voters and avoid financial accountability from balancing budget. They may make inter-fund transfers to balance budget and avoid public scrutiny of surplus transfers from fee-supported enterprise funds to tax-supported general funds. As presented above this study found theoretically, both the strong mayors and professional managers are interested in providing fee-supported enterprise services based on their distinctive needs. The form of government is coded as 1 for the mayor-council form and as 0 for the council-manager form.

Hypothesis 4. Form of government of municipalities is associated with use of fee-supported services.

Control Variables

Demographic characteristics measured by race, poverty and population, are considered important factors explain the use of fee-supported services (Bunch, 2000; Molinari & Tyer, 2003). Earlier researchers have raised concerns regarding the regressive nature of fee-supported services since they impose a heavier burden, as a percentage of income, on low-income citizens than on high-income citizens (Bunch, 2000; Gitajn, 1984; Mikesell, 2007; Molinari & Tyer, 2003). If decision makers find that the use of fee-supported services are politically unpopular and economically ineffective when the city shows a high poverty rate, they would not consider enterprise fund financed services as an alternative way of providing city services. However, the effect of using fee based services on the poor has neither been described systematically nor tested analytically. Also, a heterogeneous service population may reflect more diverse service demands. It has been reported that cities create diverse service provision mechanisms to satisfy diverse preferences of the majority of heterogeneous constituents (Feiock & Jang, 2009). Population would measure the demand of fee based services by the size of the local market. Cities with a large population would have a bigger local market for fee based activities because of the large number of customers. This research accounts for poverty by measuring the percentage of the population under poverty level as defined by the U.S. Census Bureau and uses the log of the total population, and the percentage of the dominant race from the U.S. Census data. Since we use the percentage of the dominant race as a measurement of the heterogeneity we expect the direction of the coefficient will be negative. These demographic indicators are mostly included as control variables since our primary focus is on the impact of fiscal, and political factors on the likelihood of using and increasing fee-supported services.

FEE-SUPPORTED SERVICES IN FLORIDA CITIES

Currently 43 states have some forms of limitations on revenue and expenditure of local governments and, among those 43 states, 34 states have property tax rate limits. Florida cities are regulated by state regulations for raising local taxes and have very limited discretion to create new taxation. Florida also limits the increase in the taxable value of residential properties (Anderson, 2006). (3) Thus cities need to be more innovative and create alternative ways of increasing revenue to overcome these current tax restrictions, or saving by tightening up expenditures. (4) One of the most popular alternatives in maintaining local revenue is to create and increase fee-supported services because of its private nature of charging the production cost of services to the users and the possibility of transferring revenue to general funds (Freeman, Shoulders, & Lynn; 1988; Stumm, 1997).

One hundred and ten cities (81.5 percent) in our Florida city data reported that they provided some services through an enterprise financing system and 25 cities (18.5 percent) did not (see Table 1). The degree to which fee based services are used, in terms of the percentage of total enterprise fund revenue divided by total revenue, varies from 0 to more than 80 percent. Ten cities from our sample reported they appropriated more than 70 percent of their revenues from enterprise fund revenue.

This research looks into some of the popular types of fee based services from Florida cities, such as water, gas, electricity, sewer, and garbage/solid waste. (5) We also include public safety as well as recreation and cultural services to examine the differential usage of fee-supported enterprise services and tax-supported general fund services. Table 2 reports how individual services are financed either by general funds, enterprise funds, or by a mixture of general and enterprise funds. There are also many cities that do not appropriate revenue for specific services since they rely on the county, special districts, or private utility companies to provide certain services for these communities. As it is shown, water, garbage/solid waste, and sewer are the most common services provided through enterprise financing. Forty-five percent of cities (60 cities) in our sample provide fee-supported water utility services and three cities use a mix of general funds and enterprise funds for water service. About 43 percent of the 135 cities report that they provide sewer services through fee-supported enterprise financing, while only 3.7 percent of the cities provide sewer services through tax-supported general funds. Cities with public garbage/solid waste services use fee based enterprise financing to provide the service and there are only two cities that provide garbage/solid waste services exclusively through tax-supported general funds. Since human and social services have been traditionally provided by general funds, it is noteworthy to see that about 18 percent of Florida cities provide recreation and cultural services through a mix of both enterprise and general funds.

DATA AND METHODS

The data for the empirical analyses were drawn from several sources. The data for cities' financial operations including use of fee-supported enterprise funds were obtained from the local government Comprehensive Annual Financial Reports (CAFR) in 2005. Municipal CAFR are compiled by the Office of Economic and Demographic Research (EDR) at Florida State Government and EDR provides a full list of individual municipal government's revenue by account codes and fund types for local fiscal years. (7) Financial data from Municipal Revenues were matched with 135 responding cities from a mail survey of Public Management: Florida Municipal Governments which conducted in 2005. The respondents to this Public Management survey were city managers and chief administrative officers and they were asked various public management and service provision questions. The survey was mailed to 410 city managers and mayors in Florida city governments using names and addresses provided by the Florida League of Cities. With two mailings, we received a total of 135 completed surveys for an overall response rate of 33 percent, which is considered adequate for this type of survey. (8)

Two-Stage Heckman Model

This study employs a unique Two-stage Heckman modeling process to capture the structure of data truncation and correct possibly biased sample selection. A Two-stage Heckman modeling process tests the common structure of models of censoring, sample selection, and truncation. As it is presented in Table 1, there are 25 cities that do not have any enterprise services. However, we neither simply truncate those cities nor include them as zero values in regression analysis since these non-users are not a random subset of cities using enterprise services and selection bias is possible. Hence, error term correlation of factors affecting the use of fee-supported services and the degree of use of fee-supported services should be tested by a Two-stage Heckman model (Heckman, 1976; Jang, 2006). When we have a dependent variable y which is completely observed depending upon another variable zi, we have a sample selection problem. So we can model the selection stage (the probability that the ith observation is included in the selected sample) and the outcome stage (the expected value of yi conditional on having been included in the selected sample), and test error term correlation of these two stages (Breen, 1996; Heckman, 1979). Thus the empirical analyses will be conducted in two parts: 1) selection model tests whether cities have any fee-supported enterprise services, 2) outcome model tests magnitude of usage of fee-supported enterprise services. The Heckman model analyzes the systematic choices of the use of fee supported services (non-user vs fee-supported service user) with the Probit test for the first selection stage and the second stage analyzes estimates for the percentage of total revenue that comes from enterprise funds revenue with the OLS regression model. Table 3 presents the Two-stage Heckman model specifications.

Dependent and Independent Variables

We have two dependent variables for Heckman selection model estimates: 1) selection model- a dichotomous measure of the use of fee-supported services and 2) outcome model- a continuous measure of the magnitude of use of fee-supported services. First we employ the presence of a fee-supported service as a dummy variable in the empirical estimation, which was coded as 1 if there was any revenue appropriate in enterprise fund account obtained from Comprehensive Annual Financial Reports (CAFR) in 2005, and 0 otherwise. To estimate the magnitude of the use of fee-supported services, we use percentages of enterprise funds revenue from total revenue which is a second dependent variable.

To estimate the effects of financial variables of municipalities to use fee-supported services, we use three financial indicators, including a percentage of property tax revenue from total revenue, per capita total tax and financial independence, the ratio of own revenue to total revenue of a city. Political institution was measured by the presence of an elected strong mayor in a city. If a city has the elected mayor form, coded as 1 and 0 otherwise. Poverty, measured as a percentage of the population under poverty level as defined by the U.S. Census Bureau, was employed to understand the effect of low-income citizens on the use of fee-supported services. The percentage of the dominant race from the U.S. Census data was used to capture heterogeneous service needs of cities. To capture the demand of enterprise activities, the log of population was employed to understand the size of the local market. Correlation analysis confirmed that multicollinearity did not exist among the independent variables.

FINDINGS

Table 4 and 5 present Two-stage Heckman model estimates for the cities' utilization of fee-supported services and Heckman outcome model estimates for the percentage of the total revenue that comes from enterprise fund revenue in cities. The Two-stage Heckman model estimation failed to reject the null hypothesis that "the use of enterprise funds and the percentage of the enterprise funds are independent of each other." This result shows that the regression model with truncated data would be unbiased. Thus Heckman model allows us to use the estimates of the parameters in the regression model. The Likelihood-Ratio (LR) chi-square showed that the selection model was statistically supported at less than .001. The regression outcome model fits the data well ([R.sup.2]=.58). Overall, the supports for the two models are strong and there were interesting findings based on the strength of each of the individual variables.

As presented table 4, the percentage of property tax from total revenue, per capita tax, and total population were statistically associated with the selection model which tested the association of the use of enterprise funds and those individual variables. High property tax based municipal revenue may provide a city with the financial capacity for more tax-supported general fund services and may decrease the need for any alternative ways of service provision, such as fee-supported services. The per capita tax and population were significantly and positively associated with the use of fee based services. As we hypothesized too much tax burden would lead a city to seek another revenue source, such as fee-supported services. Cities with a large population are more likely to use fee based services due to its size of customers for their local market. Yet, the fiscal independence measured by revenue from own sources divided by total revenue, had no significance in the selection model and might not influence cities to introduce the fee based services.

Most of the independent variables were statistically associated with the outcome model in table 5. (9) The property tax percentage from total revenue and per capita tax were the most consistent factors associated with both selection and outcome models. The property tax percentage was negatively associated with the revenue of enterprise services and the per capita tax is significantly and positively associated with the revenue of enterprise fund services. Although the fiscal independence had no significance in the selection model, it is significantly and positively associated with the percentage of enterprise fund revenue from total revenue in outcome model. We interpret this to mean that financially independent cities that are more autonomous, are more likely to increase fee-supported services than cities that heavily rely on revenue from higher governments.

The mayor form of government was statistically supported in the outcome model. According to the significant positive coefficient sign of mayor variable, strong mayor-council cities tended to increase fee-supported services more than council-manager cities. It is our interpretation that strong mayors may try to please local voters by expanding services provided by user fees rather than tax-supported general fund services to increase their chances to get re-elected.

Interestingly, even though the poverty rate showed statistical significance, the direction of its association was opposite of what it had been expected. The poverty rate was positively associated with the use of enterprise services. Cities with high poverty rates may provide more fee-supported services to cover the shortage of city revenue from the high poverty level. The heterogeneity, which was measured by the percentage of a dominant race, was negatively associated with the revenue of the services as expected. This finding showed that cities with heterogeneity had more demand for fee-supported services than cities with a racially homogeneous population.

Next, we conducted predictability analyses to show the relative contribution of key factors on the size of fee-supported services. Table 6 and Figure 1 show the predicted impacts of property tax and financial independence underlying the percentage of municipal total revenue that comes from enterprise funds revenue. The predicted value of the percentage of enterprise funds would be, on average, 10.3 percent lower in the highest percentile of percentage of property tax than in the lowest percentile, holding all other variables at their means. The predicted value of the percentage of total revenue that comes from enterprise funds would be, on average, 10.1 percent higher in the highest percentile (43.2%) of percentage of total revenue that comes from own sources than in the lowest percentile (33.1%), holding all other variables at their means.

[FIGURE 1 OMITTED]

Table 6 and Figure 2 illustrate the predicted impacts of form of government underlying the percentage of total revenue that comes from enterprise funds revenue. The predicted value of the percentage enterprise funds revenue in strong mayor cities, on average, would be 1.3 percent higher than council-manager cities or other forms, holding all other variables at their means. It shows that having a strong mayor carries a stronger effect on the predictability of a city increasing fee based services than manager council form of government. In a nutshell, these predicted probabilities imply that the lack of property tax revenue and financial autonomy have strong impact on the magnitude of usage of fee-supported services than impact from the strong mayor government. Although political institution affects cities' motivations for utilizing user charges, financial conditions in municipalities, especially property tax revenue, are more influential to explain the variation of enterprise financing utilization than any of the other factors.

DISCUSSION

Why do some cities use the fee-supported service model? Why do cities shift their revenue more to fee-supported services than general tax-supported services? This study indicates that the use of fee-supported service model has significant implications in the financial, political, and managerial aspects of city administration. One finding that stands out from the Heckman model is the strong negative relationship between the total revenue from property tax and the magnitude of reliance on fee-supported services. This association between property tax and revenue from fee-supported services reflects that, when cities have limited property tax revenue to provide tax-supported general fund based services, they may choose to increase fee-based services since those services are funded directly by consumers of services and can possibly generate profits. A previous study of Florida cities by DeHoog and Swanson (1988) provided a descriptive analysis of cities with electricity enterprise services and reported that cities without electricity utility services showed higher property tax rates.

Financial independence is strongly associated with the use of the fee-supported service model which was measured as percentage of enterprise funds from the total revenue. Our results indicate that financially autonomous cities which maintain large portions of their finances through their own sources are better positioned to design private-like public service packages including fee-based services. All cities show some revenue variation in the mixture of intergovernmental funding sources in the form of grants from higher governments and their own sources in the form of taxes and fees. Oates (1972) noted that this variation indicates significant differences in the overall local finance system. However previous research has not analyzed the systematic, direct relationships between local finance system and financial structures of municipalities. It is our contribution to confirm that financial independence can give cities some degree of freedom in their financial management and the ability to try new ways, such as increasing fee-supported public services, to increase their revenue.

Another important discovery is the influence that the form of government has on the use of fee based services. The results in our Heckman outcome model suggest that, when possible, elected mayors prefer to increase fee-supported services rather than rely on property taxes, possibly to please tax-paying city residents. Since revenue from enterprise funds would be used to make inter-fund transfers in order to balance the budget, the fee-supported services could possibly allow politicians to avoid public scrutiny of transferring surpluses from enterprise funds to general fund services (Bunch, 2000; DiLorenzo, 1982; Tyer, 1989). Bunch (2000) found from her case study, that either the creation of fee supported services or the increase of rates of user charges are more politically feasible choices than increasing tax-supported general fund services or cutting essential public services. Thus politically-sensitive elected mayors in the elected mayor form of government may prefer to generate and increase user charge services to achieve allocative efficiency by providing desired levels of service and avoiding financial accountability issues from heavy reliance on property tax (MacDonald, 2008).

To understand the political behavior of local politicians, we must consider the fact that local tax payers do have a political voice and can impact the local finance system by presenting their preferences to politicians. Fischel (2001a, 2001b) modified the Tiebout model by adding "politics" because residents make rational decisions to choose the community where they live and, more importantly, local residents proactively control the decisions of local governments. However, this research found divergent theories regarding the influence of the form of government on the use of fee-supported services. The efficiency-seeking and possibly profitable nature of enterprise financing of fee supported services makes the standard applications of the Reinventing Government reform movement (Obsborne & Gaebler; 1993) more plausible. Professional city managers who pursue more conscious spending and innovative earning may increase the use of fee-supported services to boost reform efforts in the city's financial system. From a managerial perspective, one of the benefits we expect from using fee-supported services is the customer-sensitive management of services when the true cost of public service provision would be revealed and the needs of customer would be met. Thus we empirically tested the political aspect of using fee-supported services in different political forms of local government and found that cities with strong mayors are positively associated with the use and increase of fee based services. Although we found that cities with strong mayors are more likely to create and increase fee-supported services than cities with professional manager council forms, we do not disregard professional city managers' efficiency and innovative management efforts through fee-supported enterprise financing systems. A city manager's pursuit of financial stability and accountable management of service provision, no matter their political leadership form, are bound by the political institution and other factors (Zhang & Feiock, 2009).

In addition to the above findings, we found that there is a statistical association between cities with a high poverty rate and their likelihood of having a high percentage of enterprise fund revenue, which contradicts our hypothesis. Since fee-supported services are regressive in nature due to the bigger burden on low-income households to pay user charges (Bunch, 2000; Bunch & Ducker, 2002; Gitajn, 1984), a low-income community was expected to be less likely to use fee-supported public services. However, the results from the Heckman outcome model lead us to at least tentatively conclude that, if a city has a high poverty rate, it is more likely to provide services via fee based enterprise services since its property tax base is thinner than cities with lower poverty rate. This finding is related to the argument on "cutback management" (Levine, 1980; Rubin, 1985, 1990) which states that cities with economic hardship from a high poverty rate would prefer to use more market driven financial mechanisms to overcome the hardship. However, it is not concrete enough to conclude that this correlation of poverty and use of fee-supported services is confirmed or further explored with this data set.

Community attributes would also be influential in the use of fee-supported services. Local markets with a large size population would allow cities to have fee-supported services, while cities with high heterogeneity may face a backlash from diverse citizens' demands. The high heterogeneity of the political and ideological context increases communication and information costs, which may hinder local governments from creating and continuing to implement stable and consistent management innovation.

CONCLUSION

Cities either rely on local taxes mainly from property taxes or user fees to support delivery of public goods and services. However, relying heavily on tax-supported services must be done with caution because doing so can cause resentment from tax paying citizens. The impact of a higher tax burden on city residents also leads to cities increasing fee-supported services to avoid tax increases. Hence, if there was a shortage of revenue and a high tax burden, cities would consider increasing fee-supported services rather than relying heavily on tax supported general funds. Only 18 percent of the cities from our Florida city data are non-user cities which do not have any fee-supported services and we reasonably assume that these nonuser cities are more likely dependent upon property taxes as revenue source than cities using the fee-supported enterprise services model. (10)

Municipalities should analyze the consequences of shifting their revenue away from property tax to other sources. Earlier, we pointed out that there are important values, a sense of efficiency and a sense of equity, involved when municipalities rely more on fee-supported services. Fee-based services can promote fairness and efficiency by giving citizens a market choice about the use of public services. However, if the profit from provision of fee-based services is not spent to improve the services adequately, the individuals who use the fee-based services may no longer feel a sense of fairness and may become dissatisfied with the poor quality of the services. Their dissatisfaction may in turn affect the demand curve for the fee-supported services (Duff, 2004). Thus, municipal governments should carefully understand that the willingness of individuals to pay for fee-supported services is closely associated with satisfaction with the quality of the services and expenditures should be adequately allocated to finance the services.

Also, fee-supported services are politically popular according to our analyses, however these services can violate equity issues by creating a regressive financial burden on low-income citizens, especially when the services must be purchased. We find that fee-supported services can help cities with a higher poverty rate increase their revenue through cutback management, but we recommend that cities need to analyze whether their mandatory fee-based services are affordable to low-income families. Municipalities need to ensure that individuals who cannot afford to pay are guaranteed access through special arrangements. Some municipalities give fee discounts or vouchers to low-income families for fee-based services to increase a sense of equity. For instance, California cities have electricity bill discount programs for low-income families whose annual income is less than $25,000. Also, civic engagement, such as citizen surveys or interviews, may help municipalities set price rates for fee-supported services for different segments of income.

This research attempted to make local finance systems more transparent by analyzing fee-supported service activities. Nevertheless, the analysis was still limited because the intertransferable characteristics of various enterprise funds were only assumed and not measured empirically. Future studies should use empirical investigation to expand our understanding of both the political and financial impact of using fee-supported services in specific service areas to verify whether fee-supported services are used primarily to enhance efficiency and effectiveness of finance systems or to appease tax payers. The main purpose of this research was to identify and describe which factors encourage or discourage cities from using the fee-supported services model and we concluded that the local political institution, local finance, and community attributes are associated with cities' various uses of the fee-supported services model and the property tax percentage relatively have a higher impact on the use of enterprise services.

REFERENCES

Anderson, N. B. (2006). Property tax limitations: An interactive review. National Tax Journal LIX, 3, 685-694.

Baily, S. J. (1994). User charges for urban services. Urban Studies, 31, 745- .

Bierhanzl, E. J. & Downing, P. B. (1998). User charges and bureaucratic inefficiency. Atlantic Economic Journal. 26, 175-.

Bierhanzl, E. J. & Downing, P. B. (2004). User charges and special districts. In J. R. Aronson & E. Schwartz (Eds.), Management policies in local government finance. (5th ed.). Washington, D.C.: International City/County Management Association.

Breene, R. (1996). Regression models: Censored, sample selected or truncated data. California: Sage Publications.

Buchnan, J. M. (1967). Public finance in democratic process: Fiscal institutions and individual choice. Chapel Hill, NC: University of North Carolina Press.

Bunch, S. B. (2000). Changes in the usage of enterprise funds by large city governments. Public Budgeting and Finance, 20(2), 15-29.

Bunch, S. B., & Ducker, R. R. (2003). Implications of using enterprise funds to account for public works services. Public Works Management & Policy, 7(3), 216-225.

Colbeg, M. (1955). Utility profits: A substitute for property taxes. National Tax Journal, 8(4), 382-387.

DeHoog, R., & Swanson, B. E. (1988). Tax and spending effects of municipal enterprises. Public Budgeting and Finance,8, 48-57.

DiLorenzo, T. J. (1982). Utility profits, fiscal illusion and local public expenditures. Public Choice, 38, 243-252.

Duff, D. G. (2004). Benefit Taxes and user fees in theory and practice. The University Of Toronto Law Journal, 54(4), 391-447.

Erie, S. P. (2004). Globalizing L.A. Stanford, CA: Stanford University Press.

Feiock, R. C., & Clingermayor, J. C. (2001). Institutional Constraints and Local Policy Choices: An Exploration of Local Governance; SUNY Press: Albany.

Feiock, R. C., & Jang, H. (2009). Nonprofits as local government service contractors. Public Administration Review,69(4). 668-680.

Fischel, W. A. (2001a). Municipal corporations, homeowners and the benefit view of the property tax. In W. E. Oates (Ed.), Property taxation and local government finance. Cambridge: Lincoln Institute of Land Policy.

Fischel, W. A. (2001b). The homevoter hypothesis: How home values influence local government property taxation, school finance, and land-use policies. MA: Harvard University Press.

Florida Focus. (2007). Economic implications of Florida's proposed property tax amendment. Florida Focus, 4(1). Bureau of Economic and Business Research: University of Florida

Freeman, R. J., Shoulders, C. D., & Lynn, E. S. (1988). Governmental and nonprofit accounting: Theory and practice. Eaglewood Cliffs, NJ: Prentice Hall.

Gitajn, A. (1984). Creating and financing public enterprises. Washington, D.C.: Government Finance Officers Association.

Government Finance Officers Association. (2001).

Governmental Accounting, Auditing and Financial Reporting. Chicago.

Hanna, K. S. (2005). Planning for sustainability: Experiences in two contrasting communities. Journal of the American Planning Association, 71(1), 27-40.

Heckman, J. J. (1976). The common structure of statistical models of truncation, sample selection and limited dependent variables and a simple estimator for such models. Annals of Economic and Social Measurement, 5, 475-492.

Inman, R. P., & National Bureau of Economic Research. (2005). Financing cities. Cambridge, Mass: National Bureau of Economic Research.

Jang, H. (2006). Contracting out park and recreation services: A Heckman selection approach. International Journal of Public Administration, 29(10&11), 799 -818.

Long, J. S. (1997). Regression models for categorical and limited dependent variables. Thousand Oaks, London: Sage Publications.

Levine, C. (1980). Organizational decline and cutback management. In C. Levine (Ed.), Managing fiscal stress: The crisis in the public sector (pp.-). Chatham, NJ: Chatham House.

MacDonald, L. (2008). The impact of government structure on local public expenditures. Public Choice, 136, 457-473.

McGuire, T. J. (2001). Alternatives to property taxation for local government. In W. E. Oates (Ed.), Property taxation and local government finance (pp.-). Cambridge: Lincoln Institute of Land Policy.

Mikesell, J. L. (2007). Fiscal administration: Analysis and applications in the public sector. 7th edition. Belmont, CA: Thomson.

Molinari, J., & Tyer, C. (2003). Local government enterprise fund activity: Trends and implications. Public Administration Quarterly, 27 (3&4), 369-391.

Netzer, D. (1992). Differences in reliance on user charges by American state and local governments. Public Finance Review, 20. 499-511.

Oates, W. E. (1969). The effects of property taxes and local public spending on property values: An empirical study of tax capitalization and the Tiebout hypothesis. The Journal of Political Economy, 77, 957-971.

Oates, W. E. (1972). Fiscal Federalism. New York: Harcourt Brace Jovanovich.

Oates, W. E. (2001). Property taxation and local government finance: An overview and some reflections. In W. E. Oates (Ed.), Property taxation and local government finance (pp.-). Cambridge: Lincoln Institute of Land Policy.

Osborne, D., & Gaebler, T. (1992). Reinventing government: How the entrepreneurial spirit is transforming the public sector. MA: Addison-Wesley.

Peterson, A. J. (1984). Preface: Creating and financing public enterprises. In A. Gitajn, Creating and financing public enterprises. Washington, DC: Government Finance Officers Association.

Peterson, P. E. (1981). City limits. Chicago: The University of Chicago Press.

Rubin, I. S. (1985). Shrinking the federal government: The effects of cutbacks on five federal agencies. New York: Longman.

Rubin, I. S. (1988). Municipal enterprises: Exploring budgetary and political implications. Public Administration Review, 48(1), 542-550.

Rubin, I. S. (1990). Managing cycles of growth and decline. In J. L. Perry (Ed.), Handbook of Public Administration. San Francisco: Jossey Bass.

Sun, R. & Jung, C. (2012). Does user-charge financing reduce expenditure levels for the charge-financed services? The American Review of Public Administration. 42 (2). 170-185.

State of Minnesota, Office of the State Auditor. (2004). Special Study: Municipal Enterprise Activity. State of Minnesota.

Stumm, T. J., & Khan, A. (1996). Effects of utility enterprise fund subsidization on municipal taxes and expenditures. State and Local Government Review, 28(2), 103-113.

Stumm, T. J. (1997). Comparing alternative service delivery modes: Municipal enterprises require special consideration. Journal of Urban Affairs, 19(3), 275-289.

Stumm, T. J. (2000). Fiscal illusions and municipal enterprises. Municipal Finance Journal, 20(4), 61-70.

Stumm, T. J., & Corrigan, M. T. (1998). City managers: Do they promote fiscal efficiency? Journal of Urban Affairs, 20(3), 343-351.

Svara, J. H. (1985). Dichotomy and duality: Reconceptualizing the relationship between policy and administration in council-manager cities. Public Administration Review, 45(1), 221-232.

Svara, J. H. (2006). The search for meaning in political administrative relations in local government. International Journal of Public Administration, 29(12), 953-976.

Tiebout, C. M. (1956). A pure theory of local expenditures. The Journal of Political Economy, 64, 416-424.

Tyer, C. B. (1989). Municipal enterprises and taxing and spending policies: Public avoidance and fiscal illusions. Public Administration Review, 49(3), 249-256.

Vogt, A. J. (1978). Operating revenue of North Carolina municipal governments: Overview and comparison of electric and nonelectric cities. Popular Government, 44, 11-20.

Wagner, R. E. (1976). Revenue structure, fiscal illusion, and budgetary choice. Public Choice, 25 (1), 45-61.

Wang, X., Dennis, L., & Tu, Y. S. (2007). Measuring financial condition: A study of U.S. states. Public Budgeting & Finance, 27(2), 1-21.

Zafra-Gomez, J. L., Lopez-Hernandez, A. M., & Hernandez-Bastida, A. (2007). Developing a model to measure financial condition in local government: Evaluating service quality and minimizing the effects of the socioeconomic environment: An application to Spanish municipalities. The American Review of Public Administration, 39(4), 425-449.

Zhang, Y., & Feiock, R. C. (2009). City manager's policy leadership in council-manager cities. Journal of Public Administration Research and Theory, 20, 461-476.

Zorn, C. K. (1991). User charges and fees. In J. E. Perterson & D. R. Strachota (Eds.) Local government finance: Concepts and practices .135. Chicago, IL: Government Finance Officers Association.

HEE SOUN JANG

University of North Texas

MYUNGJUNG KWON

California State University Fullerton

NOTES

(1) Generally Accepted Accounting Principles (GAAP) requires state and local governments to use enterprise funds to account for fee-supported services which funded primarily through user charges and fees (Government Finance Officers Association, 1999). GAAP also states that enterprise funds are used in "situations where a given fund provides services primarily to customers other than citizens--outside the funding reporting entity" (Government Finance Officers Association, 2001, p.87). An enterprise fund is a separate finance system to measure the flow of economic resources for financing business type activities in the public sector similar to private business.

(2) On the other hand, it has been noted that excessive revenue transfers from enterprise fund services to the general fund would cause severe lack of funding for the maintenance of the particular enterprise service itself and lead to a serious debt increase to cover infrastructure improvement of service functions. Erie (2004) examined the Los Angeles International Airport (LAX) development plan during 1993 to 2001 and reported that the LAX expansion plan was largely stalled by excessive transfers from airport revenues to municipal general fund.

(3) In 1992, the Florida Constitution was amended to limit the annual increase in property tax. This amendment was implemented in 1995. The Save Our Homes (SOH) provision limits its growth rate in subsequent tax years to 3% or the rate of inflation, whichever is lowest (Florida Focus, 2007).

(4) McGuire (2001), in his national study of "Alternatives to Property Taxation for Local Government," showed how much cities relied on enterprise funds in 1977, one year before the passage of Proposition 13, and compare them with later in 1992 (303). Cities in California demonstrated the biggest changes in their reliance on enterprise funds accounted for 37 percent of total revenues in 1992, compared to 21 percent in 1977. Likewise, cities in Florida relied on service charges for 41 percent of total revenues in 1992 compared to 30 percent in 1977. In other words, municipal reliance on enterprise funds substantially increased over the period and enterprise funds are a becoming a popular local alternative to the limited property taxation after the tax revolt of 1978.

(5) Municipal Revenues data collected by the Florida Legislative Committee reported the use of enterprise funds under the Charges for Services category. The category is divided into twelve sections: 1) general government, 2) public safety, 3) court related, 4) electricity, 5) gas, 6) water, 7) garbage/ solid waste, 8) sewer, 9) transportation, 10) economic environment, 11) Human services, and 12) recreation /culture. Among those service areas reported on the Municipal Revenues data, we found that only seven services were usually provided by enterprise finance system with rare exceptions and constructed a descriptive review of the use of those seven services on Table 2. However, it should be noted that the measurement of "fee-supported service use" in our analyses includes all of the possible enterprise financed service items, not only those seven services listed on Table 2.

(6) Since the cities that provide services through a mixture of enterprise and general funds are double counted, the total number of cities exceeds N of 135 cities in our analyses.

(7) Florida Statue, section 218. 32, F. S., requires each local government that is determined to be reporting entity, as defined by generally accepted accounting principles, to submit to the Florida Department of Financial Services (DFS) a copy of its comprehensive annual financial report (i.e., CAFR) for the previous fiscal year in a format prescribed by the department.

(8) Currently local government's mail surveys range from 25% to 35% according to the local surveys done by International City/County Management Association (ICMA).

(9) We test for the endogeneity issue of financial indicators of 1) percentage of property tax from total revenue, and 2) tax per capita variables using Durbin Wu Hausman tests. [H.sub.0]: the regressor is exogenous; [H.sub.1]: it is endogenous. Values of test statistics allow us to accept the [H.sub.0] and we confirmed that the use of enterprise funds and measurements of percentage of property tax and tax per capita are not endogenous.

Durbin Wu Hausman Test for the measurements of percentage of property tax and tax per capita variables

Independent    Wu-Hausman      Durbin-Wu-Hausman   Accept [H.sub.0];
variables      F test          chi-square test     the regressor
                                                   is exogenous

Property tax   11.94305 ****   11.22786 ****       Yes
Per tax        9.63174 ***     9 23914 ***         Yes

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

(10) Florida city finance data indicates that the mean value of the percentage of property tax of cities without enterprise funds was 32% compared to the 17 % mean value of the percentage of property tax of cities with enterprise funds.

Table 1
Use of Fee-supported Services in Florida Cities

Percentage of Enterprise   Number of cities (%)
Funds from Total Revenue

None                       25 (18.5%)
10%<                       2 (1.5%)
10-30%                     43 (31.9%)
31-50%                     42 (31%)
51-70%                     13 (9.6%)
70-80%                     8 (6%)
80%>                       2 (1.5%)
100%                       135 cities (100%)

Table 2
Services and Enterprise Funds in Florida cities (6)

                        Cities            Cities
                        provide           provide
                        service           service
Local Service           through an        through a
                        enterprise        general fund
                        fund

Water                   60/135            4
Electricity             14                2
Gas                     11                1
Garbage/Solid Waste     52                5
Sewer                   58                4
Public Safety           8                 87
Recreation/Culture      25                92

                        Cities            Cities do
                        provide           not
                        service           appropriate
Local Service           through           revenue
                        mixture of
                        enterprise
                        and general
                        funds

Water                   3                 74
Electricity             2                 121
Gas                     1                 124
Garbage/Solid Waste     3                 76
Sewer                   3                 76
Public Safety           7                 47
Recreation/Culture      24                42

Table 3
Two-stage Heckman Model Specifications

Selection Model

Dependent Variable       Independent Variable

Use of Enterprise        Property tax, Tax per capita, Financial
Services (Dichotomous    independence, Population
Variable)

Outcome Model

Dependent Variable       Independent Variable

Percentage of            Property tax, Tax per capital, Financial
Enterprise Funds from    independence, Strong mayor, Poverty,
Total Revenue            Population, Single race
(Continuous Variable)

Table 4
Heckman Selection Model Estimates on Use of Fee-supported
Enterprise Services

Use of
Enterprise
Services (n=135)         Coefficient   Z

Property Tax             -.1643 ***    -5.16
Financial Independence   .0072         0.36
Tax per capita           .0098 ***     4.59
Population               .4627 **      3.00
Cons                     -3.682 *      -2.29
LR Chi2                  81.03 ***

*** p < .001; ** p < .01; * p < .05

Table 5
Heckman Outcome Model Estimates on
Percent of Enterprise Funds from Total Revenue

Percent of Enterprise
Funds from Total         Coefficient    t
Revenue (n=109)

Property tax             -1.045 ***     -7.74
Tax per capita           .0118 **       2.60
Financial                7196 ***       6.10
Independence
Strong Mayor             6.77 *         -2.00
Poverty                  .6572 *        2.00
Population               .3252          0.39
Single Race              -.2004         -1.73
Cons                     -33.32 **      -2.66
[R.sup.2]                0.58 ***

*** p < .001; ** p < .01; * p < .05

Table 6
Predicted Probability of Fee-supported Enterprise Services

                Predicted           Difference
                Probability of
                Percentage of
                Enterprise Funds

Property Tax    44.1% (25 p)        10.3%
                /33.8% (75 p)

Financial       33.1% (25 p)/       10.1%
Independence    43.2% (75 p)

Form of         39.4% (Mayor)/      1.3%
Government      38.1 (Manager)

Figure 2
Histogram of Forms of Government by Average Predicted
Values of Enterprise Funds Percentage

Average Predicted

Average Predicted Value,
Mayor-Council Form,           39.4

Average Predicted Value,
Other Forms,                  38.1

Note: Table made from bar graph.