Citizen Dissatisfaction Leads to Budget Cuts, or Not: A Case Study of a Local Taiwanese Government (低满意度必然导致预算删减？未必：台湾地方政府的个案研究)

Beginning in the 1990s, performance measurements have frequently been adopted by governments around the world to help make budgetary decisions more efficient (King, Feltey and Susel 1998; Poister and Streib 1994; Schachter 1997). Two types of performance measurements have been adopted: internal and external measurements. The internal performance measurement is associated with target-driven indicators, which facilitates administrative oversight. The external performance refers to citizen evaluations of service quality (Kelly and Swindell 2002). Although concerns about using citizen evaluations as performance measurements remain (Ebdon and Franklin 2004; Hatry and Blair 1976; Thomas 1995), they are now widely used as performance indicators (Wilson 1983). This would suggest that governments need more citizen input in the budgetary process to help allocate limited resources among competing policies (Ebdon and Franklin 2006). Traditional views on budget allocations stress bureaucrats’ response to the institutional constraints in a closed system (Rubin 1990). But in the age of new public management (NPM), public interests should be given an equal emphasis (James 2003). Theoretically, both the elected official and citizens can hold agencies accountable. However, there remains a debate about how bureaucrats should act when there is a discrepancy between political priority and citizen preference.

The preferences of those directly participating in budget allocation decisions are important determinants of budget results (Bowling, Cho and Wright 2004:489). The traditional public choice school has focused on bureaucrats’ budgetary preferences. Niskanen (1971) argued that bureaucrats only pursue their self-interest, which he defined as power, income, prestige, and reputation. The public service sector often lacks a pecuniary incentive, leaving such analysts to explain budget results on the basis of other self-interested elements. This budget-maximising model has provided one theoretical explanation of the expansion of United States of America federal agency budgets during the 1970s, but later it was criticised because it did not seem to apply to non-American bureaucracies (Dollery and Hamburger 1995). Dunleavy (1985, 1991) revised Niskanen's model and elaborated a so-called ‘bureau-shaping’ framework to explain why sometimes budget-maximising is not the result. Retaining the assumption of the role of self-interest, Dunleavy argued that an agency head compares the political benefit he or she may claim with the political blame that may be incurred for each policy decision. This leads to agency heads preferring to be policy-makers rather than operators who only undertake delegated tasks with limited discretion. This model also suggests that programs that enhance the policy-making role of the agency head are the ones most likely to be expanded.

There have been criticisms of the bureau-shaping model. Lowndes (1996) argued that the self-interested elements used have not been empirically tested, and that the public choice models rely heavily on assumptions, which make it unreliable. Marsh, Smith and Richards (2000) noted that the agency heads may not be aware of the difference between policy-function tasks and prescribed routines. Another criticism has been that the model focuses only on the influence of institutions on bureaucratic budgetary behaviour: it does not deal with the relationship between the bureaucrats and the elected official (James 2003; Marsh et al. 2000). Perhaps even more importantly, the bureau-shaping model explores bureaucratic budgetary behaviour within a closed system, but budget content is commonly influenced by external or exogenous factors, including macroeconomics and citizen opinions (Rubin 1990; Ryu, Bowling, Cho and Wright 2008).

The strategy that bureaucrats employ requires more comprehensive analysis to account for the complexity of the budgetary environment. In order to explore this problem, this article looks at three unpopular programs in New Taipei City, Taiwan. It examines why survey outcomes led to different program budget results. The analysis is based primarily on the bureau-shaping framework which assumes that agency heads would compare political benefits with political blame, and prefer programs that give the agency greater primacy in policy making.

Three Unpopular Programs

Beginning in the mid-1990s, Taiwan initiated a series of budgetary reforms, which included new communication mechanisms to obtain information from citizens on policy issues and their level of satisfaction about the public sector (Hwang 2005). The aim was for a more effective and efficient budgetary system (Su 2007). Beginning in September 2009, the Department of Budget, Accounting, and Statistics (DBAS) of New Taipei City initiated a biannual citizen survey to evaluate the effectiveness of the city's major policies. All local DBAS come under the Directorate-General of Budget, Accounting, and Statistics of the Executive Yuan. In the beginning of the budget preparation phase, the DBAS communicates with the mayor on policy for the budgeted year, and then sets the budget allowances for the agencies based on available resources. To help decide resource allocation among competing projects, the survey canvasses views on critical policies which the city mayor may be concerned about, as well as on newly initiated projects that deal with city issues. The director of DBAS chooses which policies and projects to include in the survey with assistance from DBAS staff and directors of agencies whose programs may be covered.

The first survey was conducted in September of 2009 and the second in May of 2010. Both used the Computer-Assisted Telephone Interview (CATI) System. There were two stages in the sample selection. First, a sample size was assigned to each of the twenty-nine districts in relation to its share of the city's total population aged twenty and older. Next, a Random Digit Dialling (RDD) sampling method generated dialling lists, and an adult resident aged twenty and older was selected for questioning. Total valid surveys were 1,808 with a response rate of 19.72 per cent in the first Survey, and 1,812 (17.8 per cent) in the second Survey.

This article looks at three low-satisfaction projects to explore the influence the survey had on resource arrangements. These are, the ‘New Hope Centre’ (NHC), which provides assistances to the unemployed and their families, the Pay-As-You-Throw (PAYT) program, which charges citizens for trash collection fee by the sizes and quantities of the trash bags they use, and the Traffic Signal Maintenance (TSM) system, which maintain the working order of traffic lights. They were all indentified as not very popular in the surveys (see Table 1 below). They differ from each other in the strength of political support and agency preference, and they demonstrate different kinds of functions an agency head may have to take into account when making budget decisions.

The Budget Results

According to the director of DBAS of New Taipei City, the survey results are expected be a useful tool for agencies to set budget priorities, as well as for the mayor and the DBAS to make final budget decisions. The goal of the survey, in the context of budget allocation, is to help with the allocation of resources among agencies rather than help any particular agency enlarge its budget allowance. Agency heads also claim that the survey result influences the success of their budget recommendations, as Interviewee III³ said: ‘the survey results influence whether or not our budget proposal can be safe not only in the stage of budget preparation but also in budget review… your funds cut less when citizens think it is good and useful.’

The three New Taipei City programs all gained low support and low satisfaction ratings from residents. In response, the corresponding departments in charge took different strategies. The Social Welfare Department planned to retire the NHC program, the Environmental Protection Department made efforts to promote the PAYT system, and the Transportation Department increased funding to improve the for TSM.

The Budget Result of NHC

One year after the introduction of the NHC program, the number of applications dropped by half (Ho 2010). The former mayor said that the city government would put more effort into promoting the program to raise the rate of applications. This was not reflected in the budget of his successor's administration the next year. The budget total for NHC was NT$215 million in 2010, including NT$15 million for outsourced functions and NT$200 million for financial aid to eligible families. In 2011, the budget of NHC programs decreased to NT$105 million, including NT$5 million for outsourced function and NT$100 million for financial aid. According to the 2012 budget proposal, funding for NHC program will shrink to NT$20 million, all of which will be allocated to financial aid. This implies that the program is being phased out, and an annual plan for NHC is no longer included in the Social Welfare Department's budget proposal. In this case, the department responded to citizen preference immediately by removing resources from the project which it also did not prefer.

Comparison of the Features of the Programs and the Budget Results

Many researchers and practitioners have advocated greater use of information obtained from citizens following New Public Management (NPM). Citizen evaluations of government performance speak to the effectiveness of policy, and can be used to help allocate resources among competing policies. A program that gains high level of satisfaction or support may be given priority over other programs, but when the satisfaction level is low, how does the department-in-charge respond? There are two assertions (Behn 2003): one is to cut its budget as a punishment for poor performance, and the other is to increase funding on the basis that poor performance is due to underfunding. From the position that public officials pursue self-interest rather than public interest, we might expect that the bureaucratic impact on the budget is contingent on the strength of political priority and the possible benefit to the bureaucrats from the program.

First, a program that is of less political priority and allows the department in charge to exert less policy impact may be more easily cut. The NHC program experience is in line with this situation. The program has long been claimed as one of the former mayor's policy achievements (Chen 2009), although some have criticised the decreased number of applications since the second year the NHC was introduced to the city. The current mayor does not see any political benefit from continuing this program, and he might be blamed for wasting money on an underused program. So the political support for the program is weak. In addition, the Social Welfare Department is simply a fund distributer, and has no discretion over the program. In a bureau-shaping framework, this program generates very little utility to the agency head. As a result, using low-satisfaction as an excuse to shrink the budget of the program seems to create a win-win situation.

Another scenario is that strong political support may overcome citizen disapproval. This situation is represented by the PAYT program. With respect to political priority, although the former mayor claimed some credit for himself in trash reduction by introducing the PAYT system (Lin 2010), the current city mayor also shares political benefits by saying that he has invested effort to increase recycling. The Environment Protection Department has been delegated to design the PAYT system, and the role the department takes is more like an administrator who engages in micro-policy-making. In a bureau-shaping framework, the agency head's attitude towards the program depends on whether the political benefits claimed are in excess of, equal to, or less than the potential blame that might be incurred. A low level of public support of this program shows that there has been political blame. However, the agency head can still benefit from promoting the PAYT program in two ways: (1) by showing loyalty to the mayor by introducing a better design of the program; and (2) by increasing citizen support of the program through education and communication. As a result, the department reallocates funds to policy advocacy so that the political benefits are in excess of the blame indicated in the survey.

The third situation is that when tension between political and citizen attitudes is not apparent, the bureaucrat's preference becomes the most important determinant in budget allocation. This situation is represented in the TSM program. This program is essentially different from the above two cases. Traffic signal operations and maintenance is a prescribed task of the Transportation Department, and it receives ongoing funding. The department has full discretion over the design of the TSM system. According to the bureau-shaping model, this program may shape the agency head's policy and professional role. Therefore it is likely to receive strong agency preference. On the other hand, a poorly performing agency may attract public discontent, and the mayor might seek citizen support through other programs. In the extreme case, the mayor might shift resources from unpopular programs to programs with higher level citizen satisfaction. In the TSM case, the political attitude on expansion or diminution of the budget was neutral. The agency head expanded the budget of the TSM system based on the survey results. According to the director of the Transportation Department, the survey result

This information was used to persuade the budget bureau, the mayor, and even the legislators to allocate more funds to improve the quality of TSM services.

The following table summarises the different dimensions of support for these three programs, and the budget results.

Conclusion

Current literature suggests that citizen surveys are often used to prioritise agency budgets-a program with a high level of citizen satisfaction or support attracts higher priority. But few studies focus on programs that are unpopular. A result-oriented budget system may suggest a budget cut for the low efficiency programs, but an input-oriented proponent may argue that low satisfaction results from insufficient funds (Behn 2003). Therefore, there is no consensus about how information about dissatisfaction is transformed into policy changes. This article proposed an analysis by examining how agencies deal with citizen dissatisfaction by observing the subsequent resource allocations. In particular, the analysis is based on the framework of bureau-shaping model which assumes that bureaucratic behaviour is motivated by self-interest, and that activities which enhance the agency's policy function will receive higher priority (Dunleavy 1985, 1991).

Three unpopular programs in New Taipei City, Taiwan were selected to explain why agency heads use different budgetary strategy responses to dissatisfied citizen feedback. The finding indicates that the program budget result was influenced by the strength of political support as well as the agency preference. Tension between political priority and citizen preference is apparent in the PAYT case, uncertain in the TSM case, and non existent in the NHC case. Adding the variable of agency preference into the three cases, the budget results tell interesting stories. When a tension existed between political and citizen preferences, and when the agency was not particularly in favour of the program, political interest defeated citizen disapproval (the PAYT program). However, citizen dissatisfaction also influences the agency directly. To reduce blame on the agency, resources were allocated to policy advocacy and communication. In addition, when the tension between political and citizen priorities was not intensive, and the agency had full policy-making discretion, the agency persuaded the mayor to consent its fund increases to improve service quality. Citizen dissatisfaction, in this case (the TSM program), was used as an excuse to enlarge the program budget. Finally, when there was no conflict among citizen, mayor, and agency preferences in an unpopular program, citizen dissatisfaction was an excuse to retire the program (the NHC program).

This article argues that the agency heads choose to use different strategies to respond to citizen dissatisfaction. The main strategy is to pursue self-interest through a series of comparisons between the likely gains and losses for each option. Citizen surveys are claimed to be a useful tool for policy decisions and resources allocation, yet they may be manipulated strategically, especially when the survey result is politically unpleasant. First, not all unpopular programs are punished by poor performance-unless the elected politician and the agency in charge also show low preferences to it. Second, although Chan and Gao (2009) indicated that public officials simply comply with the supervisor's preferences regardless of public interest, this article argued that agency heads may respond differently. In the case of PAYT, political priority defeated public resentment in the short run, but in the long run the agency head mitigated the tension between the elected politician and the citizens. Finally, when a program can enhance the agency's policy function, there is no punishment of a budget cut despite evidence of unpopularity: on the contrary, extra funding to improve service quality. This result corresponds to the fundamental argument of the bureau-shaping theory.