For decades Thailand's bureaucracy dominated politics, a condition Riggs (Reference Riggs1966) labelled “bureaucratic polity,” wherein the main competitions for resources, authority, and influence occurred within and between bureaucratic cliques. Power slowly transitioned out of bureaucratic hands during the 1980s and 1990s, as business interests, the middle class, and regional bosses joined the political sphere (Anek Reference Anek1992; Ockey Reference Ockey1992), and following the 1992 crisis, the Thai state was believed to be firmly in the hands of civilian politicians (Bidhya Reference Bidhya2005; Chai-Anan Reference Chai-Anan and Hewison1997). Two decades later, in the wake of two coups and a military junta, observers have declared the resurrection of the bureaucratic polity (Porphant Reference Porphant2014; Puangthong Reference Puangthong2014; Supalak Reference Supalak2014; Surin Reference Surin2007).
The narrative of the bureaucracy's dominance, fall, and resurgence, though, presents an interesting challenge for political science theory. Standard political economy discussions of the relationship between politicians and bureaucrats rely heavily on principal–agent frameworks (Meier and Krause Reference Meier, Krause, Krause and Meier2003; Moe Reference Moe1984; Wood and Waterman Reference Wood and Waterman1991). The approach espouses a fundamental assumption of a hierarchical relationship between political principals and bureaucratic agents, and scholars in this vein seek to understand the mechanisms incentivizing, monitoring, and overseeing bureaucrats to ensure that they pursue the desires of their political principals (Moe Reference Moe1987; McCubbins, Noll, and Weingast Reference McCubbins, Noll and Weingast1987; Shapiro Reference Shapiro2005; Weingast Reference Weingast1984). Bureaucrats are treated as active during policy implementation but largely passive in the political process (Huber and Shipan Reference Huber and Shipan2002; Moe Reference Moe2006).
Melding these predominant theories of politician–bureaucrat relationships with the experience of Thailand appears somewhat problematic, as specialists frequently describe Thai bureaucrats as actively engaged in politics and independent of politician control (Chambers and Napisa Reference Chambers and Waitoolkiat2016; Merieau Reference Merieau2016; Unger Reference Unger, Schneider and Heredia2003). Such claims contradict the basic assumptions of agency theory, leaving us with the question: Who are the principals of the Thai state, politicians or bureaucrats? In this essay, I make two claims by examining methods of political control over the bureaucracy. First, I demonstrate that the assumptions underpinning the principal–agent model do have some applicability in describing the relationship between Thai civilian politicians and bureaucrats, especially during the period from 1992 through 2006; at the same time, though, the utility of agency theory is limited. Second, I demonstrate that the politician–bureaucrat relationship in Thailand is complicated by political power struggles between government agencies and elected politicians. Many officials within Thailand's bureaucracy actively hinder civilian control over the arms of the state. While positive signals emerged from 1992 through 2006, much of that progress has disappeared.
At the theoretical level, the Thai case suggests that we must seriously re-examine theories of bureaucratic politics, especially in the context of developing states. While the literature on politician–bureaucracy relations acknowledges the importance of the bureaucracy, its basic assumption of a hierarchical relationship overlooks the power of bureaucrats (Baekgaard, Blom-Hansen, and Serritzlew Reference Baekgaard, Blom-Hansen and Serritzlew2015; Fukuyama Reference Fukuyama2013). Such assumptions are dubious; in many countries bureaucracies are intimately involved in politics, and, as organizations and individuals, they are not passive recipients of the choices made by politicians (Moe Reference Moe2006; Carpenter Reference Carpenter2001). Indeed, in Thailand we see that the politician–bureaucracy relationship is not always hierarchical but rather transaction based. This confuses the distinction between agents and principals (Unger Reference Unger, Schneider and Heredia2003; Unger and Chandra Reference Unger and Mahakanjana2016, 80–92). We need better theoretical constructs to deal with such cases.
The remainder of this article is as follows. I begin by briefly reviewing agency theory, drawing a set of predictions for the Thai state. I then test these hypotheses. I do this first through examining the legislative productivity of parliament. Next, I turn to politician control over career trajectories of bureaucrats in two of Thailand's most important ministries, the Ministry of Interior and the Ministry of Defence. Finally, I close the essay with a discussion of the implications of these findings.
AGENCY THEORY AND BUREAUCRATIC CONTROL
The principal–agent framework, also known as agency theory, is the most prominent theoretical base for social science work on politician–bureaucrat relationships (Moe Reference Moe1984; Meier and Krause Reference Meier, Krause, Krause and Meier2003; Shapiro Reference Shapiro2005; Weingast Reference Weingast1984). Borrowed from economics, agency theory focuses on contractual relationships wherein a principal hires an agent to accomplish a task; the challenge is then to use incentives and monitoring to align the actions of the agent with the desires of the principal (Jensen and Meckling Reference Jensen and Meckling1976). Political scientists who adopted this perspective operate under the assumption that political authority is ultimately the domain of politicians, but their actions are circumscribed by institutions and contexts (Huber and Shipan Reference Huber and Shipan2002). Monitoring costs, bureaucratic expertise, and veto players in the political process largely determine the amount of slack that bureaucratic agents enjoy. The bureaucracy exercises influence over policy primarily through implementation, because of opportunities for shirking or sabotage (Brehm and Gates Reference Brehm and Gates1999). Bureaucrats, though, may ultimately recognize their relatively weak position and align their interests with politicians, thus creating a mirage of bureaucratic independence (Ramseyer and Rosenbluth Reference Ramseyer and Rosenbluth1993). At risk of oversimplification, politicians are the ultimate policy-makers; bureaucratic discretion exists primarily under conditions which hinder politicians’ ability to monitor and enforce their preferred policy.
This theoretical work contrasts with a second body of scholarship on non-Western countries that has highlighted the political strength of autonomous bureaucracies in determining policy outcomes, a condition that may be very beneficial in the pursuit of growth strategies (Evans Reference Evans1995; Johnson Reference Johnson1982). Despite potential benefits, autonomy can also create dilemmas. When policy reforms come to the fore, bureaucrats are among the largest groups of “losers” who face the costs of downsizing, retraining, or restructuring (Grindle Reference Grindle2004). Thus, bureaucracies have strong vested interests in shaping government action, and they may leverage their political influence to shape or block policy outcomes that threaten their interests. This literature sees bureaucrats and politicians in a potential power struggle; only when politicians are strong are they able to exercise subjective control over their counterparts in the bureaucracy (Fukuyama Reference Fukuyama2013; Huntington Reference Huntington1957). Otherwise bureaucrats, with their norms, expertise, and goals, can control the policy arena.
Advocates of agency theory contest these claims, arguing that such observations are due to either agency slack or the alignment of bureaucrat and politician interests; principal–agent frameworks still apply (Ramseyer and Rosenbluth Reference Ramseyer and Rosenbluth1993). Politicians remain in command due to their ability to exert influence over bureaucrats through various mechanisms, including writing legislation (Huber and Shipan Reference Huber and Shipan2002; McCubbins et al. Reference McCubbins, Noll and Weingast1987), control over bureaucratic appointments and careers (Peters Reference Peters and Farazmand1997; Wood and Waterman Reference Wood and Waterman1991), budgetary decisions (Dunleavy Reference Dunleavy1991), bureaucratic reorganizations (Schwartz Reference Schwartz1994), and monitoring institutions such as courts or an ombudsman's office (Bennett Reference Bennett1997).
Scholarship on Thailand's politician–bureaucrat relationships tends to align with the second body of literature, treating bureaucrats as relatively autonomous and politically active (Bidhya and Ora-orn Reference Bidhya and Poocharoen2010; Ockey Reference Ockey2004; Unger Reference Unger, Schneider and Heredia2003; Unger and Chandra Reference Unger and Mahakanjana2016), and as such, it has largely side-stepped any challenge agency theory might pose to this interpretation of events. Can agency theory, then, apply to Thailand? To gauge this, I focus on two methods of political control over the bureaucracy: law-making and bureaucratic appointments.Footnote 1 First, legislators exert their policy authority by writing more laws and including greater detail in the laws they produce; specific and lengthy statutes tend to delineate the exact actions that bureaucracies are to take in their implementation of laws. Short and vague legislation, on the other hand, grants leeway to bureaucrats (Huber and Shipan Reference Huber and Shipan2002, 3–8). Second, politicians may also control bureaucrats through either rewarding those who please them or punishing those who do not via promotions and advancement, especially among high-ranking officials (Ramseyer and Rosenbluth Reference Ramseyer and Rosenbluth1993; Wood and Waterman Reference Wood and Waterman1991). According to agency theory, control of the legislative and appointment processes grants politicians dominance over bureaucrats.
Drawing from this literature, we can identify three observable predictions regarding the politician–bureaucrat relationship in Thailand that should be in evidence if politicians act as principals of the Thai state. Two of the predictions have to do with legislative productivity while the third focuses on politician control over bureaucratic career advancement. First, if politicians in Thailand are acting as principals of the state, we should see variation in the level of bureaucratic discretion written into legislation as the number of veto players in the legislature changes. The logic behind this claim is based on the increasing difficulty of passing legislation as coalition size grows. Each additional coalition partner brings a new policy preference to the table, constraining the acceptable range of policies (Tsebelis Reference Tsebelis2002). Due to the increasing probability of conflict in the coalition, politicians will reduce the amount of legislation and the detail of that legislation in order to reach a mutually agreeable compromise with coalition partners; this, in turn, grants greater discretion to the bureaucracy. The opposite should also hold. Bureaucratic discretion should decline as politicians are able to act collectively.
Second, we know that policy-making is costly both in terms of information and transactions. Politicians must pay these costs in their efforts to rein in bureaucratic slack. Certain politicians will have greater capacity to do so due to their education and talents, or what Huber and Shipan (Reference Huber and Shipan2002) refer to as legislative capacity. The most important skills in this realm are those related to understanding and writing legislation. Thus, we expect that if politicians are truly masters over the bureaucracy, we should see shifts in legislative output according to their individual talents. A more-skilled or better-trained politician should exercise greater control over the bureaucracy through legislation.
In other words, variation in legislative productivity resulting from coalition size or legislative capacity signals that politicians act as principals, directing the state via law-making, but their actions are constrained according to predictable patterns. A lack of variation along these hypotheses would indicate that some other mechanisms are at play that are not currently described within agency theory.
Third, beyond a focus on legislation, questions of power dynamics between politicians and bureaucrats play themselves out in the promotion and tenure paths of government officials (Peters Reference Peters and Farazmand1997). If a bureaucrat acts contrarily to politician preferences, presumably politicians have the authority to either dismiss the official or hinder his or her career advancement. Thus, policy that is developed and implemented by the bureaucracy would remain in line with politician preferences (Ramseyer and Rosenbluth Reference Ramseyer and Rosenbluth1993). Otherwise state officials would incur the wrath of politicians. The observable implication of this prediction is that politicians in control of the bureaucracy should regularly influence the tenure and promotion of bureaucrats (Wood and Waterman Reference Wood and Waterman1991). If bureaucratic appointments are made contrary to politician preferences, then politicians are not in command.
In the case that these three predictions follow the expectations of agency theory, we can more confidently argue that Thai politicians act as principals of the Thai state. On the other hand, if there is no clear preponderance of evidence in favor of such a politician–bureaucrat relationship, we will be forced to reconsider the applicability of agency theory's assumptions.
I evaluate the first two implications of agency theory by examining variation in legislative productivity across individual governments in Thailand. I consider governments formed during elected legislatures from 1992 through 2013, as it was only after 1992 that scholars widely recognized politicians as dominant over bureaucrats (Bidhya Reference Bidhya2005). This period, especially from 1992 to 2006, provides us the most likely scenario under which politicians would act as principals. In other words, if our predictions fail to describe the activities of these governments, then it is highly unlikely that the assumptions of agency theory hold in the Thai case. This leaves me with 10 governments formed from elected legislatures.Footnote 2 Despite constitutional changes, the legislative process remained constant, removing this as a potential confounding factor (Chanchai and Apirach Reference Chanchai, Phetsiri, Tanitcul and Imaizumi2011).
For the first two tests of agency theory detailed above, the outcome of interest is the degree of bureaucratic discretion gauged by legislative productivity. To measure this, I first look at the number of laws signed by each government per month. This signals how active each government was in promoting new policies. Without new laws, the bulk of policy-making is left to bureaucratic discretion. Two additional measures deal with the detail of legislation rather than the amount. Politicians exercise control over policy implementation by writing more detail into laws; the longer the law, the less discretion available to the bureaucracy.
These counts are based on the Yearly Summary of the Laws produced by the Secretariat of the House of Representatives. Calculations include both Royal Acts (Phraracha Banyat) and Emergency Decrees (Phraracaha Kamnod).Footnote 3 On average, between September 23, 1992 and December 31, 2013, elected legislatures in Thailand produced 2.74 laws per month.Footnote 4 The mean length of Thai laws in the same period is 7.64 pages with 23.47 articles, with median lengths of 3 pages and 8 articles. Table 1 presents legislative productivity numbers across the 10 governments in my analysis. Appendix 1 combines Christensen and Ammar's (Reference Christensen and Siamwalla1993) counts with my own to provide data for the 60-year period from 1953 through 2013.
* Counts include Acts (Phraracha Banyat), Emergency Decrees (Phraracha Kamnod), Organic Laws, and Constitutional Amendments drawn from the Yearly Summary of the Laws (multiple years), produced by the Secretariat of the House of Representatives.
** Calculated monthly average from government's tenure, including cabinet reshuffles.
The first observable implication that Thai politicians and bureaucrats are in a principal–agent style relationship would be that legislative productivity should shift according to coalition size. In other words, both the number of laws produced by a government and the detail in those laws should vary inversely with the number of veto players present; with fewer coalition partners, legislation should be more frequent and detailed (Tsebelis Reference Tsebelis2002). I use the effective number of coalition parties to gauge the number of vetoes present in the cabinet (see Blau Reference Blau2008).Footnote 5
Thai cabinets have exhibited a tendency toward broad coalitions, with a few important exceptions, most prominently the governments under Thaksin Shinawatra and his sister, Yingluck. If agency theory were truly describing the politician–bureaucrat relationship in Thailand, we should expect that these governments exhibit greater legislative productivity than their counterparts.Footnote 6 Yet this is not the case. There appears to be no significant increase in the number of laws produced by the legislature during periods in which a single party is dominant. Thaksin's first term hovered just above the average, despite his extremely convincing electoral mandate and ability to rein in troublesome coalition partners. Both Thaksin's second term and Yingluck's term were significantly below the average legislative activity.
One might explain these away by arguing that legislative activity was hindered by crisis situations during the both governments. In Thaksin's case, protests in the months leading up to the coup may have hampered his ability to pass legislation. If we constrain the analysis to only the months before the crisis, though, results are not substantially different. Accounting for a slight delay in the time legislation takes from parliamentary approval to receiving the official acceptance, from March 2005 through February 2006, Thaksin's second government only produced 24 laws, or two laws per month. This was still below average and during a period in which Thaksin's government enjoyed unprecedented dominance in the legislature.
Yingluck fared only slightly better. If we were to remove the months she and Bangkok were inundated by a flooding crisis (August–December 2011) as well as anything after the major protests began against her government in November 2013, we are left with the months from January 2012 through December 2013, again considering the slight delay between passage in the parliament to acceptance of the law. During this time, the government passed 53 pieces of legislation at a rate of 2.21 laws per month. Even accounting for crisis, both Thaksin and Yingluk performed worse than average on the amount of legislation passed.
In contrast, Thaksin's first government did exhibit a dramatic increase in the number of pages per law as well as the number of articles per law. These counts were the highest among all the governments under consideration, showing that, despite the lack of expansion in new laws, there was an improvement in the level of detail and thus constraints on the bureaucracy. If we look more closely at his government, we also see that legislative productivity jumped dramatically, from 2.64 laws per month to 4 laws per month, after Thaksin dropped the Chart Pattana Party from his coalition in November 2003, reducing the effective number of parties from 1.5 to 1.28. This is consistent with the theoretical expectations.
Alternatively, governments with large numbers of coalition partners do tend to produce fewer pieces of legislation and less detailed legislation. Banharn's government, with the highest number of coalition partners, was also among the least productive of the legislatures under consideration. Similarly, Chuan's first government was also unproductive.
We do have two anomalies, which can be explained by considering the context in which they occurred. The second Chuan Leekpai government, with an initial count of 2.89 effective and 6 actual parties in the coalition, exhibited relatively high legislative productivity. This would be contrary to our theoretical expectations. Even so, the Chuan government's productivity came in the aftermath of the 1997 financial crisis, which created a special situation. Politicians who would regularly fight over resources began to cooperate due to the severe external shock they had recently faced, and this decreased the number of effective veto players in the system. Thus, the number of parties with seats in the second Chuan cabinet belies the fact that vetoes had actually declined during this time, which would be in line with the hypothesis. Furthermore, if we look more closely at the Chuan government's legislative productivity, we can see that it increased substantially from 4 laws per month to 5.93 laws per month after the Social Action Party withdrew from the governing coalition in July 1999, dropping the effective number of parties to 2.66. This shift in productivity aligns with theoretical predictions.
Another anomaly is the relative detail found in the legislation produced by the Abhisit Vejjajjiva government, which had 2.33 effective and 7 actual coalition parties. Again, context can explain this increase in legislative detail. Abhisit's cabinet enjoyed behind-the-scenes support from the military faction that had conducted a coup in 2006. Military interference had assisted in the coalition formation and acted as a relatively invisible force reducing alternate parties’ veto threats (Chambers Reference Chambers2010). With reduced veto threats, our hypothesis would predict this outcome regarding detailed legislation. This pattern is further confirmed through statistical analysis of Thai laws from 1992 through 2013. I assembled a dataset of legislative productivity of each month during this period, including data on the effective number of coalition parties, which varied within government due to frequent cabinet reshuffles.
The dataset includes two control variables for crises that have affected Thai legislative productivity. The first is a dummy variable highlighting the Tom Yang Kung Crisis, starting in May 1997, when speculative attacks began on the Thai Baht, and continuing until December 1999, when economic reports showed that the Thai economy was again experiencing positive growth. All other times scored a zero. The economic crisis presumably reduced the number of veto players and allowed for increased legislative productivity. The second crisis variable accounts for mass protests, such as the protests against the Thaksin government (February–September 2006), yellow-shirt activities against the Samak and Somchai governments (May–November 2008), or the red-shirt protests during Abhisit's administration (March–May 2010); these are scored as a 1 during the presence of demonstrations. The turmoil should have reduced legislative productivity. The results of these analyses are presented in Table 2.
Robust Standard Errors in Parentheses
* p < 0.1
** p < 0.05
Regression analysis indicates that monthly legislative productivity was negatively correlated with the effective number of coalition parties. These numbers were statistically significant in the amount of detail in legislation measured by law length. As the post-2006 governments have been subject to military influence, I repeated the analysis including only the months from September 1992 to October 2006. The correlation held in both cases, displaying that there is an observable relationship between an increase in coalition size and a decrease in the amount of detail that is included in Thai laws. These findings suggest that Thai politicians did experience veto player constraints on their ability to produce legislation, consistent with the predictions of agency theory.
In sum, there is some support for the hypothesis that broader coalitions result in less law-making. Qualitatively, it does appear that in situations of fewer veto players, politicians were legislatively more productive, especially during the second Chuan and Abhisit periods, wherein outside forces reduced the number of veto players in the system. Quantitatively, we also see that the amount of detail written into legislation is negatively correlated with the effective number of coalition parties in government; this holds for both the full dataset as well as the constrained (1992–2006) dataset. Thai politicians, then, had the potential to control bureaucrats via legislation, but their actions varied in line with the predictions of agency theory. Even so, this cannot explain why both the second Thaksin and Yingluck governments failed to produce much legislation despite their political clout.
The second implication from agency theory is that as the legislative capacity of politicians increases, so should their ability to rein in the bureaucracy through legislation. Huber and Shipan (Reference Huber and Shipan2002) gauged capacity through measuring compensation across legislators in the USA; this approach would not work in Thailand, though. Instead, we can proxy legislative capacity of the executive through measuring educational attainment of the prime minister. First, education is vitally important in the composition and understanding of legislation. Certain types of education lend themselves to greater ability to convey policy preference into written laws. Also, higher levels of education would decrease the information costs a politician must pay in order to understand legislation, its production, and implementation. Second, the capacity of the prime minister is key to the development of laws, as the prime minister sets the legislative agenda and potentially has strong influence over the legislation emerging from parliament (Chanchai and Apirach Reference Chanchai, Phetsiri, Tanitcul and Imaizumi2011).
Again, we should expect that prime ministers who presided over periods of particularly detailed legislation should have either higher educational attainments or educations more pertinent to writing laws than their counterparts. Here we have three prime ministers to consider: Chuan Leekpai, Thaksin Shinawatra, and Abhisit Vejjajjiva. In 1962 Chuan obtained a law degree at the prestigious Thammasat University. He was admitted to the bar association two years later and practiced law before joining politics. Thaksin graduated from the Royal Thai Police Cadet Academy in 1973 and then pursued graduate degrees in criminal justice in the United States. He holds an MA from Eastern Kentucky University and a PhD from Sam Houston State University. Abhisit holds both an undergraduate degree (Philosophy, Politics, and Economics) and a master's degree in economics from Oxford University. All three would be considered highly educated in areas that have some relevance to law making. As such, I evaluate their legislative capacity as high.
Of course, Chuan and Thaksin also presided over less productive periods. The reasons for Thaksin's less productive second term and Chuan's opportunity to engage in legislative production during his second term have already been discussed above. It is worth noting, though, that when given the opportunity to engage in law-making, they presided over some of the most detailed legislation in Thai history. Abhisit's tenure is also instructive. Despite passing relatively few laws, his government did make those laws more detailed. The level of specificity could have been due to his greater legislative capacity.
Do their counterparts exhibit less legislative capacity? Banharn Silpa-Archa left secondary school during World War Two and pursued a business career. He returned to education only after becoming an influential politician, completing both a bachelor's and master's degree in law at Ramkhamhaeng University, where he was accused of plagiarizing his thesis.Footnote 7 Chavalit Yongchaiyudh, a retired general, completed a degree at the Royal Military Academy. Samak Sundaravej completed a bachelor of law degree at Thammasat, although he did not practice. Somchai Wongsawat also completed a bachelor of law degree from Thammasat in 1970. He was admitted to the bar association in 1973. Nearly 30 years later he completed an MPA degree at the National Institute of Development Administration. Yingluck Shinawatra graduated in 1988 with a degree in politics from Chiang Mai University before pursuing an MPA at Kentucky State University.
These degrees provide some indication as to the legislative capacity of these prime ministers. Banharn exhibits lower educational attainments than his counterparts, only obtaining a degree long after he was an active politician. We can judge his legislative capacity as likely lower than the three high-performing executives. Chavalit's degree in military studies was unlikely to have granted him expertise in law-making, which also bodes poorly for his legislative capacity. As such, I evaluate both Banharn and Chavalit as having low legislative capacity. Samak appears to never have joined the bar association, a potential indicator that his performance in law school was not on par with that attained by Chuan. On the surface, Yingluck's degrees appear relevant to writing legislation, but her graduate specialization was in management information systems. Her work after graduation was also business oriented rather than linked to legislative capacity. I rank both Samak and Yingluck as having moderate legislative capacity. Only Somchai's background seems to approach those of the three high-performing legislators. Even so, he was in office for only two and a half tumultuous months and presided over the passing of only two laws, one of which was the annual budget. Had he been in office longer, he may have had the opportunity to preside over more detailed legislation.
It appears, then, that there is some support for the claim that politicians with greater legislative capacity will engage in writing more detailed legislation and thus reduce the discretionary authority of the bureaucracy. This is supported by statistical analysis (Table 2), which showed that the legislative capacity of the prime minister, measured as a three-point ordinal variable, was positively correlated with both the number of laws produced by a government per month as well as the number of pages in those laws. Even so, this relationship is somewhat weaker than that seen between coalition composition and legislative productivity.
To summarize, these findings should prove heartening for advocates of agency theory, as the observed variation implies that Thai politicians have the capacity to exercise authority over their bureaucratic agents via legislation. Also, there does seem to be some support that legislative capacity does shape the ability of executives to produce more detailed legislation and rein in the bureaucracy. While high numbers of veto players in cabinets might explain low levels of legislative productivity, we don't see the expected result with the second Thaksin and Yingluck governments when a single party dominated.Footnote 8
CONTROLLING BUREAUCRATIC CAREER PATHS
Now I turn to the third test of agency theory's implications for bureaucratic tenure and promotion decisions. If the assumptions of agency theory are correct, we should see that politicians are able to readily influence bureaucratic career paths. Thai civil service reshuffles can be very disputatious, especially among resource-rich ministries such as Interior, Transportation, and Commerce (Bidhya Reference Bidhya2010). Control of these portfolios determines the distribution of resources throughout the country. Political parties first seek to command the cabinet seat of the portfolio, which is done through coalition bargaining during government formation. Once in office, they may also promote their supporters within the bureaucracy to higher positions.
The causal chain is complicated, as bureaucrats pursue alliances with political figures. Officials also strive to reduce political interference within their agencies, as this may upset their own path to promotion or diminish their ability to provide adequate services. By and large, promotion within the Thai civil service is based on tenure and personal networks rather than on capacity or skill (Akira Reference Akira2014, 321–324).
Because of these complexities, it is difficult to identify exactly who controls the appointment process. Even so, we can observe the effect of political influence by considering the role of politicians in the appointment and promotion of officials. Here I choose to focus on two of the most important and powerful ministries in the Thai bureaucracy: The Ministry of Interior and the Ministry of Defence (Bidhya Reference Bidhya, Burns and Bowornwathana2001). These two ministries present the most stringent test of political control over the bureaucracy as they have a history of autonomy from civilian politicians. Indeed, if predictions drawn from agency theory fit the experience of the fiercely independent Defence and Interior ministries, we would have a strong case for the application of principal–agent frames in the Thai context.
First, the Ministry of Interior (MOI) is important in terms of both budget and political influence. Monetarily, MOI receives one of the largest allowances of any ministry, accounting for over 11 percent of the total budget in 2011, second only to the Ministry of Education. Unlike Education, where most money is committed to staff salaries, most of the MOI budget (approximately 78 percent in 2011) is available for lucrative investments and subsidies. Politically, MOI has a strong and extensive network of civil servants throughout the country. Officials in this ministry include provincial governors, the police force, and local administrators. MOI bureaucrats are at least tangentially involved in almost all aspects of domestic governance (Achakorn and Tatchalerm Reference Achakorn and Sudhipongpracha2014). Indeed, MOI had a long history as a “super ministry” with broad-reaching authority and independence (Bidhya and Ora-orn Reference Bidhya and Poocharoen2010, 311; Achakorn and Chandra Reference Achakorn, Mahakanjana and Berman2011, 56–63).
Second, the Ministry of Defence, including the military, has been the most influential branch of the bureaucracy throughout Thailand's history. From the date of the coup that overthrew the absolute monarchy, on June 28, 1932, until June 28, 2017, the office of prime minister has been filled by either an active or retired military official for 21,010 days—or approximately 68 percent of the time. Bidhya Bowornwathana, one of the most prominent experts on Thai bureaucracy, argued that it would have been best to categorize Thailand's “bureaucratic polity” as a “military polity” (interview with author, Bangkok, February 13, 2012). Civilian control over the military is one of the major challenges facing the Thai state (Chambers Reference Chambers2010; Punchada and Ricks Reference Punchada and Ricks2016).
If the assumptions of agency theory hold, then we should see that politicians exert regular influence over promotions. If the contrary is true, the promotion and tenure of bureaucrats should be handled internal to the agency, despite political preferences to the contrary.
Ministry of Interior
In the MOI, the position of provincial governor is among the most coveted, reserved for those who reach C-10 status, the second highest rank in the Thai civil service. Provincial governors not only hold the highest seat in the ministry at the provincial level, they also supervise the efforts of other ministries and departments within their administrative boundaries. MOI officials at this level have been able to dominate the distribution of public finances to lower levels (Achakorn and Tatchalerm Reference Achakorn and Sudhipongpracha2014). The power of the position makes the annual reshuffle of governors among the most important regular events in the MOI, and the reshuffle regularly features in national politics.
Historically, management of governors has been dominated by career bureaucrats. Early in the 1990s, a reform proposal emerged to democratize the provinces, changing governors from MOI appointments to elected offices. The first Chuan government was forced to abandon this idea under fierce resistance from the MOI, indicating the ministry's strength in the policy arena (Nagai, Ozaki, and Kimata Reference Nagai, Ozaki and Kimata2007, 5–7). Beyond this, during the Chuan and Banharn governments, appointments of provincial governors generally followed the desires of ministry officials. Indeed, politicians complained that the October 1995 reshuffle list was compiled by civil servants without allowing politicians to provide input (Bangkok Post 1995b). The ministry's permanent secretary confirmed that the prime minister “made no change to the [reshuffle] list” the agency had provided to the government (Bangkok Post 1995a).
Late in the 1990s, though, civilian politicians gained greater control over the ministry. Chavalit moved some governors known to be close to Banharn into inactive posts (Temsak Reference Temsak1997). When Chuan took office after the Asian Financial Crisis, he was soon criticized for orchestrating a major shakeup of the MOI outside of the annual schedule; 15 governors were transferred in April rather than waiting for the usual October reshuffle. Chuan and MOI officials denied that the reshuffles were politically motivated (Bangkok Post 1998a; Bangkok Post 1998b). No matter the cause, the reshuffle signaled that Chuan's government was able to implement changes outside of the regular promotion schedule.
When Thaksin came to power in 2001, he set his sights on reforming the bureaucracy, including governorships, adopting what would become known as the “CEO” governor policy. The reform was meant to improve lines of accountability by making governors responsible for all local governments in their province. They would then report directly to the prime ministers’ office (Mutebi Reference Mutebi2004; Painter Reference Painter2006). This centralization of power gave Thaksin added incentive to become more involved in the annual MOI reshuffle. Reshuffle lists now passed directly through the prime minister's office for review and approval (Yuwadee and Temsak Reference Yuwadee and Traisophon2003). Thaksin used control over gubernatorial reshuffles to reward supporters and punish defectors, a fact recognized in US diplomatic cables (US Embassy 2005). One retired deputy governor explained, “Prior to Thaksin there may have been a little bit [of political influence in annual reshuffles], but it wasn't overt; it became very obvious during Thaksin's time” (interview with author, Khon Kaen, October 28, 2015).
After elections returned in 2008, ensuing governments continued to exert influence over gubernatorial appointments. Political sparks flew between Abhisit's Democrat Party and its coalition partner, Bhumjai Thai, over control of the annual reshuffle (Bangkok Post 2009; Veera Reference Veera2010). Through directing the placement of governors and the forced retirement of Thaksin appointees, the new coalition government sought to assert itself in MOI (Pradit, Aekarach, and Surasak Reference Pradit, Sattaburuth and Glahan2009). This approach was repeated by the Yingluck administration (Pradit Reference Pradit2012; Bangkok Post 2012).
The experience of the MOI provides evidence for the claim that politicians have come to serve as a hierarchical authority over bureaucracies. While during the 1990s the bureaucracy could resist politician preferences, this autonomy ended after the 1997 constitution. Thaksin's concentrated political power allowed him to gain control over the reshuffle and politicize it further. A former member of the Civil Service Commission argued that bureaucrats, “can't [resist] if there is a strong government. The 1997 Constitution gave this power to the government” (interview with author, Bangkok, September 15, 2014). The civil servants in the MOI became much more proactive in aligning their actions with the desires of politicians, including pre-emptively vetting the annual reshuffle list with the prime minister and members of the cabinet (Retired deputy governor, interview with author, Khon Kaen, October 28, 2015). Post-Thaksin civilian leaders have continued to intervene in the gubernatorial reshuffle list, indicating that politicians were acting as principals, exercising much greater control over MOI.
Ministry of Defence
Civilian politicians in Thailand have maintained an uneasy rapport with military leaders. Prime ministers seek to maintain good relations with military officers, often by filling the post of Minister of Defence with a retired military officer loyal to their cause. When Chuan Leekpai appointed himself as Defence Minister in 1997, it was the first time a civilian had done so since 1976. Since Chuan, though, Samak, Somchai, and Yingluck all filled the portfolio in hopes of cultivating closer ties with the military (Patsara and Wassana Reference Patsara and Nanuam2013). Taking these positions, though, did not signal that politicians were in control of the annual military reshuffles. Instead they sought to develop close ties with the armed forces to forestall possible coup attempts.
From 1992 through 2001, we see evidence that political control of the military was growing through efforts made by civilian politicians to influence military appointments.Footnote 9 Chuan's first government took several actions contrary to the interests of the military, including limiting budget growth and arms purchases. Beyond this, Banharn's cabinet blocked some military appointments desired by top brass (Tasker Reference Tasker1995; Wassana Reference Wassana1995). The military's diminished political role played out in the senate as well where, in 1996, only 18.4 percent of senators were drawn from the ranks of retired soldiers, a sharp drop from the 55.2 percent previously (Chambers Reference Chambers2013a).
In the aftermath of the 1997 financial crisis, Chuan chose Surayud Chulanont as the new military chief; Surayud claimed to be firmly committed to keeping the military out of politics (Pasuk and Baker Reference Pasuk and Baker2009; Wassana Reference Wassana2001). Reform efforts, though, were slow. Chuan admitted that earlier attempts to rehabilitate the system of military promotions had failed, as the ranks of generals continued to increase (Chuan Reference Chuan, Gearing and Mitton1999). Ockey (Reference Ockey and Alagappa2001, 208) wrote that civilians were still “unable to change the basic structure of the military.” Thus, political control over promotions and tenure from 1992 through 2001 was growing, but it was also limited.
As the new constitution took effect, and the popular Thaksin Shinawatra government was formed, it seemed that politicians were prepared and powerful enough to influence military reshuffles. The senate also became an elected body, with retired military officials almost completely disappearing from its ranks.
Thaksin saw the military as a possible threat to his political dominance, and he moved to bring it under his personal control, if not under the institutional control of politicians (McCargo and Ukrist Reference McCargo and Pathmanand2005). By appointing former Prime Minister Chavalit to his cabinet, he hoped to take advantage of the general-turned-politician's skills and influence in directing military reshuffles (Chambers Reference Chambers2013a). While Thaksin did re-politicize the military (Pasuk and Baker Reference Pasuk and Baker2009), he also brought the annual military reshuffles more directly under politician control. As a graduate of the Armed Forces Preparatory School and a former police officer, Thaksin had many contacts in the military, and he began promoting officers friendly to him. From 2001 to 2004, the prime minister's office determined many important appointments in the annual military reshuffles, with some notable exceptions wherein current and retired military officers intervened to promote anti-Thaksin officers (Chambers Reference Chambers2013a, 258–260). While this may not have signaled objective control of the military, at least the soldiers were largely under subjective control (Huntington Reference Huntington1957; Ockey Reference Ockey2007; Punchada and Ricks Reference Punchada and Ricks2016).
As Thaksin's second term moved forward, though, factions within the military as well as the Privy Council were increasingly unhappy with the prime minister's meddling. With political protests on the streets of Bangkok and the threat of violence, the unassailable Thaksin juggernaut weakened. Retired general and former Prime Minister Prem Tinsulanonda publicly spoke to military officials, openly arguing that the military owed little allegiance to an elected government (Pasuk and Baker Reference Pasuk and Baker2009). The 2005 and 2006 military reshuffles were both highly contentious (Chambers Reference Chambers2013a). While not the sole cause, the increasing subjective control over the military was a strong contributing factor for the 2006 coup (Ockey Reference Ockey2007).
With the coup, the military reasserted its dominance in the political arena, demonstrating its independent political authority. In 2008, the Defence Administration Act institutionalized military dominance in promotions, removing the capacity of elected officials to independently control promotions above the rank of brigadier general. Post-coup governments have never been able to re-establish control over the military reshuffles, which have been dominated by supporters of the 2006 coup. Instead governments bowed to military preferences (Chambers Reference Chambers2010, Reference Chambers2013b). Military expenditures have also increased, reversing a trend in reduced military spending since the 1990s.
We see, then, that Thaksin and, to a lesser extent, Chuan could exert influence over promotions in the military. In both cases, the power balance had tipped toward civilian rule. For Chuan, military leadership was significantly weakened in the aftermath of the 1992 protests. This allowed civilian politicians to exercise greater control over the armed forces. Thaksin also enjoyed greater political power than the military, at least during his initial years as prime minister, as the military was dominated by officers who openly proclaimed their apolitical status (Wassana Reference Wassana2001). Thaksin thus faced relatively little resistance as he leveraged his electoral victories to dominate bureaucratic appointments, within and without the military. This, though, seeded resistance within certain factions of the military, contributing to the coup (Ockey Reference Ockey2007).
The relationship between civilian leaders and the military can be seen as a power struggle. From 1992 through 2006, civilians gained some control over the military apparatus, but this did not mean that the military bureaucracy had adopted Weberian characteristics or submitted itself to politician control. As one of the retired top brass explained in an interview, military officials believe themselves independent of civilian control and politicians should be forbidden from involvement in the promotion and appointment process (Retired Air Chief Marshal, interview with author, Bangkok, 28 January 2014). Indeed, if politicians interfere in military decisions, generals feel justified in overturning the political system.
To summarize the discussion of political control over bureaucratic appointments, the fundamental assumption that politicians hold a dominant position over bureaucrats has limited applicability in the Thai case. Elected politicians, as they accumulated political power and influence from 1992 through 2006, were able to exercise increasing control over the bureaucracy, which indicates that Thailand was moving toward a situation that fits the predictions of agency theory. From the example of the Ministry of Interior, we see that politicians had made massive gains as principals of the Thai state. At the same time, though, there was bureaucratic resistance spearheaded by the military and, finally, in 2006, a coup that turned back the clock, demonstrating the weakness of political control over the bureaucracy. The Thai military bureaucracy maintains a separate source of political power and influence, placing it in competition with elected officials rather than under their control.
The evidence above suggests that elected politicians in Thailand increasingly acted as principals of the Thai state from 1992 through 2006. Higher numbers of veto players resulted in reduced legislative productivity, and higher levels of legislative capacity tend to align with greater legislative productivity. We also see that the Ministry of Interior had been brought to heel according to politician wishes through political control over bureaucratic promotions.
These relationships were somewhat attenuated by negative examples, however, such as Thaksin's second government, wherein we would predict above-average legislative productivity but the opposite occurred. Additionally, military officials from the Defence Ministry have resisted politician control; even overthrowing their political principals in 2006 and 2014. In essence, the assumption that politicians and bureaucrats are in a hierarchical relationship has some explanatory power, but it is also incomplete. There is a more complicated relationship in play, one which includes a delicate balancing act wherein civilian politicians cannot always trust in their control the bureaucracy. At times they may command, but at others they must cajole their bureaucratic counterparts.
These findings hold a set of implications for the study of politician–bureaucrat relationships in Thailand and other developing countries. First, agency theory, although holding some tenable predictions, faces major challenges as a tool for researching the bureaucracy in developing states. The Thai state, among others in the developing world, is far from the Weberian ideal, and the underlying monitoring and oversight functions that are central to agency theory do not necessarily hold. In cases where politicians do not have clear and coherent mechanisms to exert control over the bureaucracy, theories based on a hierarchical politician–bureaucrat relationship may highlight the growing power of civilian politicians, but they fail to capture the nuances of power struggles taking place in the process. Instead, we need a more pragmatic approach to describe the role of bureaucrats in the Thai polity. Rather than agents of the Thai state, they are also actors in the political sphere themselves, and they have repeatedly inserted themselves into politics to protect their interests.
Theoretically, then, we should consider an interactive model of the politician–bureaucrat relationship, one which does not assume that politicians enjoy a hierarchical advantage over bureaucrats (Carpenter and Krause Reference Carpenter and Krause2015). Power relationships in these situations are more fluid and are linked to patronage, and they have not yet crystalized as in a Weberian world. Bureaucrats have additional tools whereby they can shirk or sabotage efforts to exert political control over their actions. Recognizing this, we should take the political influence of the bureaucracy, especially powerful agencies, into consideration when researching both politics and policy decisions. I therefore echo Moe (Reference Moe2006, 25) who called for “more research on how public sector bureaucrats … attempt to exercise power in gaining control over the political authorities that govern them.”
In studying the Thai state, rather than merely declaring the return of the bureaucratic polity, we need to recognize the methods by which bureaucrats actively pursue political goals, both individually and as agencies, even when civilian politicians seem to dominate (Ockey Reference Ockey2004; Bidhya Reference Bidhya2010). For instance, electoral lists obtained from Thailand's Election Commission for parliamentary elections in 2005, 2007, and 2011 all contained a strong showing of both retired and active bureaucrats, with 16.4, 16.8, and 14.3 percent of the respective party-list candidates coming from the bureaucracy. Civil servants were also well-represented in district campaigns, with 12.9, 11.9, and 8.6 percent of district candidates hailing from the bureaucracy. Among those elected in 2011, 78 parliamentarians (15.6 percent of the total) were former bureaucrats. Additionally, former bureaucrats make strong showings in political leadership. Theera Wongsamut, a Royal Irrigation Department official, has presided over the Chat Thai Pattana party since 2013; Yongyuth Wichaidit, a former Permanent Secretary of the Ministry of Interior, headed the Pheu Thai Party from 2008 to 2012. By the end of her administration, over half of Yingluck's cabinet was composed of former bureaucrats. Also, lest we forget, Thaksin Shinawatra began his career as a bureaucrat in the police force for 14 years. where he laid foundations for his later business and political success (Pasuk and Baker Reference Pasuk and Baker2009, 36–40).
Taking a longer view of Thai history (see Appendix 1) we see an interesting pattern in Thai legislation, wherein the bureaucrats who populate legislatures during juntas and appointed governments tend to write and adopt many more statutes than their civilian counterparts. From 1958 through 2013, juntas produced approximately 10 times as many laws, while appointed governments were about three times a fruitful. Appointed legislatures also compose the most detailed legislation found in the Thai legal canon. Bureaucrats have thus authored laws to regulate the behavior of politicians and reduce the amount of discretion they enjoy, turning agency theory upon its head. The new 2017 Constitution furthers this effort, as conservative military officials and other actors have tailored the document to weaken the potential power of civilian politicians. The charter, in its present form, enshrines a leading role for bureaucrats in the coming years (Khemthong Reference Khemthong2017). It remains to be seen, though, whether such institutional engineering will endure when social and political pressures that have been held in check by the junta are released. Civilian politicians will likely challenge these constraints, and Thai constitutions are notorious for their short lifespans.
In sum, labelling Thai bureaucrats and politicians as agents and principals is something of a misnomer. As demonstrated above, Thai politicians have played the part of principals, but their command of the bureaucracy is not absolute. Even Thaksin, who enjoyed unprecedented sway, was constrained in his efforts to control the Thai state. Thai bureaucrats, on the other hand, often act as agents, but some venture into the realm of principals. The politician–bureaucrat relationship embodies tension, interaction, cooperation, and transactions that leave politicians as something less than absolute principals and bureaucrats as more than agents.