Are democratically elected local officials more responsive to citizens? Political science research has long presumed that responsiveness is both a reason for the preferability of democracy over other regime types, and a metric by which the quality of a democracy should be judged.Footnote 1 Scholars of Chinese politics have extended this line of argument to village elections in rural China, and to township and county people's congresses. Even in an authoritarian context, elections may give citizens the power to nominate officials whom they expect to be responsive to citizens and to vote out the officials who are least responsive to citizens’ needs.Footnote 2 However, sometimes democratic elections fail to hold officials fully accountable – and, by contrast, some non-electoral institutions may effectively motivate officials. Indeed, research on China has illustrated a variety of mechanisms, ranging from social pressure to the threat of protest, by which citizens can induce responsiveness even in the absence of fully democratic elections.Footnote 3
One way to adjudicate these competing claims is to focus on “apples to apples” comparisons of elected and appointed officials who perform the same duties. Valid comparison cases can be difficult to find, and the results of studies that focus on them have been mixed. Some suggest that elected officials are more responsive to citizens because of effective monitoring by voters or re-election incentives.Footnote 4 By contrast, other studies find that directly elected officials behave much like appointed ones.Footnote 5 Still others find that direct election may produce less responsive or effective local officials.Footnote 6
For both substantive and methodological reasons, Taiwan provides a valuable test case for assessing the relative performance of elected and appointed officials. Substantively, this is a live political issue in Taiwan. While Taiwan's democratic transition is a point of pride, factional politics, vote buying and other forms of corruption continue to mar local elections, although these problems are not as severe as they once were.Footnote 7 Indeed, problems with local-level democracy in Taiwan have led to proposals that would require directly elected township mayors to be replaced with bureaucratic appointees. The debate over these proposals reveals competing assumptions about the responsiveness of elected and appointed officials. DPP legislator Cheng Yun-Peng 鄭運鵬, who proposed amending the Local Government Act (difang zhidufa 地方制度法) to abolish township chief elections, argued that replacing elected chiefs with bureaucratically appointed ones would “increase government efficiency and reduce corruption.” Opponents of the proposed amendment, however, argued that bureaucratically appointed officials would be less responsive to citizens’ needs.Footnote 8
Methodologically, Taiwan provides a unique opportunity to compare elected and appointed officials, as the position of township (district) chief is directly elected by voters in some localities and bureaucratically appointed in others. Taiwan is divided into 368 district-level political units (townships, districts or township-level cities). The executives of 204 of these units are directly elected while the remainder are bureaucratically appointed by the city in which they are located. Figure 1 shows the current distribution of districts with elected and appointed chiefs across Taiwan. Crucially, a given district's method of executive selection is determined by factors that are exogenous to the district: districts within “special municipalities” (zhixiashi 直辖市) have appointed executives, while those within counties and smaller cities, as well as all indigenous districts, have elected ones. Whether a city is designated as a special municipality is also largely out of local politicians’ hands, as a city is automatically upgraded after its population reaches two million people. This institutional configuration allows for the direct comparison of elected and appointed officials who otherwise operate in similar positions and within the same national cultural and political context.
Figure 1: Districts with Elected (Dark) versus Appointed (Light) District Chiefs, 2019
This paper leverages this source of institutional variation to assess whether directly elected officials are more responsive than bureaucratic appointees to citizens’ needs. Using data from an experiment in which Sara Newland and John Chung-en Liu sent requests for help from putative citizens to 358 district chiefs at two points in time, I find that they are not.Footnote 9 This study measures responsiveness as the speed and efficacy with which officials provide help and information to citizens, in keeping with similar studies of local government responsiveness in the Chinese context.Footnote 10 Elected officials write shorter and lower-quality responses, and are less likely to respond at all, than appointed officials. Employing a multimethod design, I use 20 in-depth interviews and data from nonparticipant observation in district service centres to elucidate the mechanisms that produce these results. The qualitative data show that local elections, which remain quite personalistic in Taiwan, do not incentivize elected chiefs to be responsive to those outside of their networks. By contrast, frequent monitoring and evaluation by higher-level bureaucrats means that appointed chiefs must be highly responsive or risk losing their positions.
Quantitative Data and Results
This paper relies on a novel source of quantitative data to compare the responsiveness of local officials: requests for help emailed by putative citizens to 358 elected and appointed district chiefs. Taiwan's local government websites are robust, providing opportunities for citizens to give feedback to local officials, ask questions and, in some cases, apply for social welfare benefits and other government-run programmes entirely online. As early as 2002, Taiwan was regarded as a global leader in “e-governance,” and this reputation remains today.Footnote 11 Email and online platforms are thus an important site of local government responsiveness across Taiwan.
To collect data, Newland and Liu sent requests for a basic citizen service (instructions for applying for a public subsidy) to all district chiefs in Taiwan, excluding Kinmen 金門縣 and Matsu 連江縣. We then collected data on four measures of responsiveness: whether a given district chief responded to the email they received as well as the length, timeliness and quality of their responses.Footnote 12 We sent two rounds of emails, so that each chief received two different requests for assistance.Footnote 13
The quantitative results are presented in Table 1 in which I include a short list of pre-treatment control variables. I avoid including individual-level covariates associated with particular officeholders; because a district's status – whether it is led by an elected or appointed chief – is the “treatment” in this study, (s)election of individuals into these positions occurs after treatment is assigned and controlling for individual-level factors would thus run the risk of introducing post-treatment bias.Footnote 14
Table 1: Results
Note: *p < 0.1; **p < 0.05; ***p < 0.01
In all cases, the relationship between status as an elected district and our measure of responsiveness is in the expected direction, although in some cases these results are not statistically significant. Elected officials respond less frequently to constituent emails, and the responses they write are shorter, lower in quality and slower to arrive. The results are most striking for our measure of response quality. Responses by elected officials scored .44 lower than appointed ones on average in the first round, and .82 lower in the second, on a 5-point scale. These differences are statistically significant at p < .05 and p < .001, respectively.Footnote 15
The results are quite consistent across both rounds of the experiment and across these various specifications. Bureaucratically appointed officials are more responsive than elected ones. In both rounds of the experiment, appointed officials were more likely to respond to putative citizens’ requests for help than elected ones. In round one, 85 per cent of appointed officials responded to requests for help, compared to 78 per cent of elected officials. In round two, the gap grew to 15 percentage points (83 per cent of appointed officials responded versus 68 per cent of elected officials). By international standards, these numbers suggest that local officials in Taiwan are quite responsive overall. Nonetheless, the difference in behaviour between elected and appointed officials is a striking pattern. To put these results in a comparative context, the response rates by appointed officials are higher than any reported in Mia Costa's meta-analysis of elite responsiveness studies.Footnote 16 By contrast, the response rates by elected officials are still high, but much less exceptional (the 68 per cent response rate would put Taiwan in the 83rd percentile of the studies Costa analyses).
Exploring Mechanisms: Personalistic Elections and Bureaucratic Monitoring
What explains these persistent differences between elected and appointed officials? This section relies on non-participant observation in district service centres led by elected and appointed chiefs, as well as 20 in-depth interviews with citizens and local political elites, to describe the ways in which the different performance incentives of elected and appointed officials shape their behaviour. Taiwan's local elections remain quite personalistic, as voters are linked to elected township chiefs via patronage networks that encourage responsiveness to particular subsets of voters at the expense of the community as a whole. By contrast, bureaucratically appointed district chiefs are frequently monitored by their superiors in city government. While they sometimes seem distant and impersonal to voters, constant performance evaluation (and the possibility of immediate removal from office for poor performance) incentivizes them to provide effective citizen services even to citizens with whom they lack personal ties.
Several reasons for the relatively poor performance of elected district chiefs emerge from the qualitative data. First, re-election incentives encourage elected chiefs to emphasize campaigning over governance. As candidates for district chief are not always affiliated with a political party, they need high name recognition to succeed on election day and so put substantial effort and expense into campaign posters and events that bring them into contact with voters. Some see this as a useful mechanism for ensuring accountability to citizens, as “citizens may not recognize [appointed] district chiefs, and district chiefs may not care about citizens.”Footnote 17 Others, however, find that the work of campaigning gets in the way of service. As one employee of a district service centre with an elected head put it, “Bureaucrats who get their position by passing the civil service exam develop deep knowledge of a particular area. Elected officials are always just pursuing votes. They have to always pay attention to winning votes over everything else.”Footnote 18
Second, local elections remain personalistic in much of Taiwan, encouraging local elected officials to be responsive to citizens within their network of influence but creating limited incentives for broader responsiveness. One appointed district chief (quzhang 区长) from Taoyuan 桃園市, who replaced an elected chief when Taoyuan was upgraded to a special municipality, offered this example:
One newly elected village representative sent me at least ten messages thanking me when I agreed to put 5,000 NTD [about $150] towards a new water system. That is not a lot of money, but he kept thanking me over and over! He said that he had asked for four years for this from the former mayor and was ignored because he wasn't part of the old mayor's patronage network. I try to provide services fairly to everyone.Footnote 19
As another appointed district chief described, “My [elected] predecessor only served those who voted for him. He didn't meet with others.”Footnote 20 Personalistic ties between candidates and citizens, often facilitated via vote brokers and party factions, played an important role in local elections during the authoritarian period.Footnote 21 Although voters now have substantial autonomy and Taiwan's local elections are widely regarded as free and fair, both existing research and my interviews suggest that these institutions continue to shape relationships between local elected officials and citizens today.Footnote 22
Pervasive corruption in electoral politics also increases politicians’ incentives for personal responsiveness to small groups of voters and financial supporters.Footnote 23 Vote buying has long been part of Taiwan's political culture. A central component of local elections during the authoritarian period, local corruption became a wedge issue that the DPP used to attract voters during Taiwan's first competitive elections, and a crackdown on widespread election-related corruption followed the January 1994 local elections.Footnote 24 Despite regular investigation and prosecution of election-related corruption in the democratic period, vote buying and other forms of corruption remain endemic in Taiwan's district-level electoral politics. Collusion with local property developers and embezzlement of public funds are common problems – in one extreme case, the last three township chiefs were removed from office for corruption.Footnote 25 Although corruption is not limited to elected district chiefs, high campaign costs may lead these problems to be especially severe among elected officials. Local campaigns in Taiwan have always been expensive.Footnote 26 Today, Taiwan's enormous roof- and wall-mounted campaign ads can be prohibitively costly and sometimes can only be secured via agreements between candidates and property developers.Footnote 27
Bureaucratically appointed district chiefs face a different set of incentives. They are typically experienced public servants with long careers in a highly professionalized bureaucracy, elite educational backgrounds and deep knowledge of local governance. These bureaucrats feel a strong sense of professional purpose and public obligation. This often manifests as a sense of obligation to higher-level officials rather than to citizens, and citizens sometimes perceive these officials as distant and unresponsive. Nonetheless, their training, their professional ethos and the strong city-level oversight to which they are subjected all lead appointed bureaucrats to be relatively responsive to a broad range of citizens.
That Taiwan's bureaucratically appointed district chiefs are generally competent and responsive should not come as a surprise given the training, history and ethos of the civil service. Civil service posts are not especially well compensated; nonetheless, the positions are prestigious and highly competitive. In 2018, the pass rate for normal administrative posts (yiban xingzheng 一般行政) in the “common exam” (for applicants with at least a high school diploma) was 3.5 per cent, and the pass rate for the elementary exam (open to all adults regardless of educational background) has hovered around 1–2 per cent in recent years.Footnote 28 In addition, bureaucratic appointees often have elite educational backgrounds; elected district chiefs tend to be less educated and enjoy strong family ties to incumbents in local government.Footnote 29
District chiefs typically reach their posts after relatively long careers, and successive promotions, within this elite system. In Taoyuan, for example, most appointed district chiefs reach their positions through one of two routes. Some have already served in vice-leader (fushouzhang 副首长) positions in the city government and request to serve in district chief positions late in their careers.Footnote 30 A second route involves promotion to district chief after at least 10–15 years in lower district-level or city-level posts. Bureaucratic appointment incorporates district chiefs into social and professional networks that run between the city and its various districts, and good performance as a district chief is likely to lead to promotion to a higher position within the city government.Footnote 31
This close relationship with the city government means that bureaucratically appointed chiefs are subject to frequent and consequential evaluation by higher-level officials. As one district chief put it, “evaluation is every day.”Footnote 32 Taoyuan's research and development council partnered with a local college to evaluate the telephone service to citizens; a team from the college called each branch office of the district government every month to ask for help, and the staff were evaluated on how helpful and polite they were.Footnote 33 In addition, city governments designate several high-priority policies each year and evaluate each district's implementation.Footnote 34 In New Taipei, districts are ranked according to their scores, and these ranks and the justifications for them are made public to encourage lower-performing districts to emulate the top performers. In Taoyuan, all district chiefs attend a weekly meeting with the mayor. There is also a monthly district governance meeting (quzheng huiyi 区政会议) led by the secretary-general (mishuzhang 秘书长) of the Taoyuan city government, to which every office within the city government must send a representative.Footnote 35 In general, this allows for smooth coordination between districts and the city.Footnote 36 Finally, the Taoyuan civil affairs bureau conducts an evaluation of the district chiefs and sends the results to the mayor, who ultimately assesses each district chief's performance.Footnote 37 These forms of monitoring can have immediate consequences for underperforming district chiefs: “Before, if the township chief did a bad job, it took four years before they could be voted out of office. Now, they can be removed from office the next day by the mayor.”Footnote 38
This combination of elite professional identities with strong and consequential higher-level monitoring has produced an effective, responsive corps of appointed district chiefs. For these appointed chiefs, responsiveness is not primarily to citizens. Instead, appointed district chiefs feel a strong sense of obligation to their superiors in the city government. Asked about his typical work routine, one appointed district chief replied: “I was appointed by the mayor of Taoyuan almost two years ago. My every day daily life is to serve our local village and city councilmen and execute what the city government asks me to do … If I have a good relationship with city councilmen, I will have a good time.”Footnote 39 This interviewee came across as a model public servant. He saw his post as a “24-hour job” and had received a national award for exemplary local governance. However, his sense of obligation to citizens was indirect; satisfying his directly elected superiors was his primary task.
Citizens and officials perceive local state–society linkages to be weaker under appointed district chiefs than under elected ones. Both widely agree that citizens often do not even know the name of their appointed district chief.Footnote 40 Neither officials nor citizens saw this kind of distance as necessarily problematic, however. As one city government employee put it:
The old system had its advantages – [elected] chiefs were willing to work hard for citizens, even if what they did wasn't always legal (hefa 合法). Now, they work hard but their work is just on solving problems, not getting name recognition. Instead, they collect opinions from many citizens, and on the basis of that make policy decisions.Footnote 41
Citizens also perceive appointed chiefs as relatively removed from citizens, but at the same time see them as more effective than they were under the old, election-based system:
The workers at the district service centre do a really good job. If you go to them for help they help you very quickly and efficiently. Twenty years ago, it was not like this … Before Taoyuan was elevated to a special municipality, we all knew who the [elected] district chief was. Now, the district chief has no power … After the elevation, everything is controlled by the city.Footnote 42
The interviews described in this section thus confirm the central finding of the quantitative results – that elected officials are less responsive to citizens than appointed ones. They also highlight key reasons for this responsiveness gap: whereas frequent monitoring and close ties to city government incentivize appointed bureaucrats to respond quickly and effectively to citizens’ needs, the personalism and expense of local elections induces elected officials to be responsive to voters and economic elites with whom they share close ties – potentially at a cost to citizens excluded from these networks.
Conclusion
On the whole, district-level officials in Taiwan are a model of local government responsiveness. In large cities, district service centres often provide a “one-stop shop” where citizens can apply for social welfare benefits, check out library books, seek medical attention, charge electronic devices, drink tea provided by a volunteer and even get a massage. While government offices in smaller towns do not offer the same range of amenities, they nonetheless provide polite and efficient service from specialized public servants who staff walk-up desks, phone lines and online communication portals where citizens can request assistance.Footnote 43 The overall response rate of district governments to online requests for constituent service is about 80 per cent, one of the highest rates reported in cross-national research on local government responsiveness.Footnote 44
Nonetheless, there is substantial variation across localities in the level of responsiveness that local officials display towards citizens. This paper has argued that one important source of this variation is the method by which local township or district chiefs are (s)elected. I find that chiefs who are directly elected are less responsive to citizens than chiefs who are civil servants appointed by the city government. In districts with elected chiefs, the personalistic nature of local elections means that chiefs are often highly responsive to small groups of voters at the expense of broader attention to the needs of other voters or the district as a whole. By contrast, appointed chiefs’ elite backgrounds and frequent monitoring by – and strong sense of obligation to – the city government combine to make them serve local residents more effectively.
These findings should not be taken as an indictment of Taiwan's democratic development. Indeed, that Taiwan's bureaucratically appointed district chiefs are highly responsive is partly a function of relatively effective city-level democracy. Regular monitoring by two bodies – the elected city council and the elected city mayor and their staff – is central to appointed chiefs’ sense of their mission. Put differently, Taiwan's elite bureaucracy remains so in part because elected “principals” exert meaningful control over bureaucratically appointed “agents.” What would enable voters to play a similar role in local politics – in other words, to more effectively manage the principal-agent relationship with elected township chiefs – remains an open question.
Acknowledgements
I am grateful to John Chung-en Liu for allowing me to use our data from another project for the quantitative component of this paper. I would also like to thank Villanova University and Smith College for financial support, John Chung-en Liu, Ben Read, Steve Goldstein, Dai-Lin Hsi, Ching-Ping Tang and two anonymous reviewers for helpful comments, and student researchers in the East Asian Politics Lab at Smith College for superb research assistance.
Conflicts of interest
None.
Sara A. Newland is assistant professor of government at Smith College. She researches local governance in China and Taiwan and subnational diplomacy between the US, China and Taiwan.
