Chapter 3 Coalitional Politics of Non-Ideological Groups: LDP Factions and Budget Outcomes in Japan, 1956-2003∗ Kuniaki Nemoto Gradually from the 1960s, the fiscal situation in Japan began worsening. In 1966, the Japanese government had to resort bonds to finance the deficits for the first time since the wartime period. The fiscal situation improved a little in the end of the 1960s and the early 1970s, but the problem became worse again in the mid-1970s, when the everlasting Japanese economic miracle finally slowed down. The Japanese government had to rely on the issuance of much more bonds to finance its expenditures, and for the first time in 1976, the outstanding bonds of the central government overcame the net revenues from tax (see Figure 1). The growth rate of tax revenues recovered in the 1980s as a result of the economic boom (Noguchi 1991), but the problem was further deteriorated in the 1990s as the bubble economy collapsed. The general governments' gross liabilities exceeded the total amount of the GDP and have been around the OECD's worst level since 2000 (see Figure 2). What is puzzling is that Japan was once seen as a unique nation where the government could manage its macroeconomy with the balanced budget norm, the efficient bureaucracy, and the relatively small size of the public sector. First, the government always tried to keep the norm of balanced budget. With its high-level policy expertise and fiscal discipline, the finance bureaucracy in Japan has made every effort to intervene in the budget process, in an attempt to reduce expenditures and increase revenues (Campbell 1977; Grimes 2000). Second, the Japanese bureaucracy was once said to be highly autonomous and efficient. Free from inefficient allocation of monetary resources, it became an idealized, stylized bureaucratic model required for the rise of a developmental state (Johnson 1982). Third, Japan's public sector was the smallest among the industrialized democracies. Before the 1970s, government consumption was kept under 10 percent of GDP, while after the 1980s and especially in the late 1990s its level has reached the OECD average (Figure 3). It is true that macroeconomic conditions largely explain a government's fiscal conditions (e.g. Barro 1979), but scholars have increasingly pointed to the role played by ∗
The present essay is prepared for one of the empirical chapters to appear in my dissertation. I have benefited from helpful comments by Stephan Haggard, Ellis S. Krauss, Mathew D. McCubbins, Matthew Soberg Shugart, and especially Megumi Naoi. The usual caveat applies. 1
political and institutional conditions that can still explain substantial variations. Which political and/or institutional factors lead to increase and/or decrease in expenditures on public policies and budget deficits still lack a definite answer, and this casts empirical and theoretical questions. In explicating the size, scope, and orientations of public spending, one group of authors, mainly based on the finding by Roubini and Sachs (1989), focuses on the aggregation of different interests. They argue the varying economic and political institutions help to account for the differences in patterns of public spending and deficits.
Especially, they find that large
spending and therefore the large deficits are the result of political weakness, measured by a short tenure in office and a dispersion of political power – both are the features of coalition governments. Logic behind this finding is that governing requires a consensus among a large number of veto players and thus the need to forge such a consensus should result in more generous spending. Others have confirmed this finding in various ways (Crepaz and Moser 2004; Kontopoulos and Perotti 1999; Milesi-Ferretti et al 2001; Perotti and Kontopoulos 2002; Persson and Tabellini 1999, 2003). Used throughout these models is the implicit or explicit assumption that each political party in a coalition can be a potential veto player that has a unique policy preference over public policy. However, this assumption ignores two subtle but important considerations. First, a party is not a unitary actor. A party's policy goals can be best viewed as an aggregation of different backbenchers' preferences, but most often researchers simplify this aggregation mechanism, when they incorporate the number of parties as an independent variable to the right side of their equations, or when they identify a party's preference with the median legislator inside the party. But, the actual mechanism of how backbenchers' interests are translated into the formal policy goals of any parties should be conditional upon party rules that stipulate the balance of power between rank-and-files and leaders, and electoral rules that might make backbenchers' interests divergent from party leaders' public policy orientations. Second, even if a coalition government contains several veto players, the existence of veto players may not automatically result in expansion of the budget. This is the case if, as Tsebelis (2002) notes, each veto player is flexible enough to make policy adjustments easily. To put it differently, if veto players inside a coalition are non-ideological, the automatic relationship between a number of veto players and fiscal expansion would not hold – especially when parties may be best viewed as a group of legislators who only want to institutionalize focal points for exchanging their own unique interests (e.g. Aldrich 1995).
2
Thus adding a new member to a coalition or changing coalitional partners may not necessarily result in the change of aggregate fiscal outcomes. Another group focuses on the formation of the different interests to explain fiscal policy patterns. Certain types of electoral institutions, for example, are said to encourage legislators to have local and parochial interests and therefore engage in fiscal competition (Carey and Shugart 1995; Cox and McCubbins 2001). This implies the common pool problem of budget: Backbenchers may intensely compete each other in a zero-sum conflict for pork to their own districts, possibly exhausting the available resources (Weingast, Shepsle, and Johnsen 1981). Expansion of the entire pie might be pursued as the second-best solution on which all the relevant actors can agree to moderate this fiscal competition game. First, the problem in this local bias approach is its static nature. How to reach this structural equilibrium is relatively unknown. Even under the electoral systems that cultivate personal vote, politicians may take time to learn how the system works and how they can optimize their reelection possibilities. This suggests that one might uncover the historical process of how the structural equilibrium is achieved. Second, even if all the politicians have distinctly parochial interests, this might not result in the tragedy of the commons. This is because we cannot assume that all the politicians necessarily have equal opportunities for policymaking, even if they share the incentives to cultivate local interests. Under certain party rules, only a few politicians may monopolize the policymaking process to shut off relatively young, inexperienced backbenchers. This may partly because the existing analysis of personal vote does not differentiate the incentives for constituency services and the capacity to bring pork to districts – if access to policymaking is relatively limited to veterans, then the electoral rules that encourage personal vote may not necessarily result in fierce fiscal competition (Nemoto, Krauss, and Pekkanen 2006). Competition in the latter case may involve contests over limited posts that are tied to governmental privileges. Thus these theoretical models are insufficient to explain the link between the tragedy of the commons and political/institutional variables in Japan. First, the Liberal Democratic Party (LDP) has been the dominant ruling party in the country since 1955 with only a minor interruption in 1993-94. Coalition governments have been the norm since 1993, but the LDP has continued to be the largest and thus the most important party in every coalition government formed after 1994, so the political variables that are useful in explaining cross-country variations in fiscal outcomes – the number of veto players and partisan
3
ideologies – are not appropriate in this one-party-dominant context.1 Second, the government's spending showed different patterns throughout the observed period, even though the country had used for a fairly long time the same electoral rule, which is known as so personalistic and particularistic that politicians developed a local bias. My argument is that the decision to spend more is conditional upon how the power of party leaders is constrained internally by backbenchers. When the decision making process is centralized enough to exclude the demands from rank-and-file members for both private and public goods, spending is under monopolized control of the central authority. If this is the case, then party leaders can limit expansion of spending. By looking at the changing degrees of centralization, one can differentiate various governments, even in a regime continually dominated by one single party. This dynamics inside the party may involve changing coalitions as a matter of factions, but not necessarily because factions have unique policy preferences as argued by McCubbins and Thies (1997). Still, another empirical question about non-ideological veto players needs to be answered, as the above prediction would hold whether groups in a coalition (in the Japanese context, factions in a party) are viewed as ideological or non-ideological (cf. Heller 2001). Therefore I will test several hypotheses that can be derived from the two different assumptions – one is that factions are ideological units with unique policy preferences, and the other is a claim commonly made in the literature of the Japanese politics; that is, "factions of interest" work only as institutions to allocate privileges and resources to members and are more inclined toward provision of private goods, rather than public-regarded policies pursued by ideological entities (Bettcher 2005). We will see that, although some of our results are still mixed and we cannot be decisive at this point, evidence is confirmatory to the factions-of-interest view: the demand for public-good spending arises more saliently from the central leadership than from the factional competition mechanism; and factions compete only for pork-related policy issues and the result is that spending for pork-related items expands when factions become threatening to the leadership. The focus in this article is on the Japanese government's fiscal policy. Specifically, 1
Some recent analysts show factions inside the LDP have their own policy preferences (e.g. McCubbins and Thies 1997; Thies 2001), challenging the traditional view that factions are only groups to exchange parochial interests and they do not have ideological policy preferences (e.g. Leiserson 1968; Ramseyer and Rosenbluth 1993). The former view suggests that still the veto-player theory is usable in the Japanese context. These competing assumptions have theoretically predictable implications that can be tested, as I will do below. 4
we replicate the Japanese fiscal data by McCubbins and Thies (1997) – 63 budget items under authority of 13 different ministries – and expand its observations up to 2003. The items to be covered in this study are appropriate in this theoretical setting because they have a clear division between highly politicized and therefore pork-related items (e.g. educational subsidies exclusively for Okinawa) on the one hand, and public goods items (e.g. the maintenance and management fees for labor protection offices across all the prefectures) on the other hand. First, this present study will describe the Japanese budget-making system, and review the existing literature on each of the budget systems in Japan to show the increasing politicization of the process especially after the 1970s. Second, I will show the politicization of the process was in fact party leaders' response to personal-vote externalities arising from Japan's unique electoral system and increasing threats from factions to defect. Third, I will introduce my hypotheses and data, followed by a data analysis section. The concluding section will follow. I.
Budget Making Process in Japan
Regular Budget. The fiscal year in Japan starts on April 1. From then on, the preparation of requests starts in spending ministries. In July, the cabinet issues guidelines to all spending ministries for ceiling. By the end of August, further discussions in spending ministries result in a consolidated request, which is submitted to the Budget Bureau of the Ministry of Finance (MOF). The Bureau reviews the requests from ministries and recommendations from various advisory councils inside the MOF from September. The MOF drafts a preliminary budget around December 20. Any amendments and consolidation of these intense requests are finalized by the MOF and the cabinet as the final government draft budget proposal by the end of December. The final draft is then sent to the MOF, which prepares an official budget bill by the end of January. After the bill is sent to the Diet, the Budget Committees of the Lower and Upper Houses discuss the budget. After approved by the committees, the bill is finally sent to the floor voting and legislated by the end of March. Although the cabinet was a primary agenda-setter up until the 1960s (e.g. Makihara 2003), in this process of regular budget making, rank-and-file politicians make every effort to intervene through a wide range of channels especially after the 1970s (Campbell 1977; Noguchi 1991). First, requests for budget items are closely discussed between politicians and
5
spending ministries. Most often, bureaucrats' interests in maximizing their budget and extending authority over policy issues coincide with politicians' interests in spending more for public and private goods to buy the voter's support (Niskanen 1975). Bureaucrats from spending ministries intensively participate in the meetings of the ruling party's most authoritative policymaking organ – the Policy Affairs Research Council (PARC) –, where bureaucrats try to get political backing from politicians for their policy issues, while listening to politicians' demands for their local districts.
Second, even after a final request is consolidated
and sent to the MOF, until December, MOF bureaucrats face intense lobbying by spending ministries and pressure groups backed by some LDP legislators for expansion of their favorite budget items. Requests for further revisions are reviewed at various levels inside and outside the MOF, as well as at the PARC. Third, although it is relatively rare to find amendments to a final government budget bill, still politicians have opportunities to make a voice on it, for example by trying to delay the submission to the Diet (McCubbins and Thies 1997, fn.14). In addition, as argued below, the amendments to the budget were further made possible by the politicization of the supplementary budget. Supplementary Budget.
Even after the original budget is submitted to and approved by
the Diet, it can still legislate "Supplementary Budgets" to modify the original budget. The rationale behind these budgets is that the government has to respond to any sudden economic fluctuations due to downturns or unexpected fiscal demands from natural disasters. The cabinet creates budget proposals for approval by the Diet, and modification of the original budget may simply involve transfers among items without any increase of the entire budget, or increases in some items with additional funding, which usually come from issuance of bonds. Usually before the 1990s, the net additional spending in this supplementary budget was only about 3-4% to the main regular budget. As such, the MOF has tended to hold Supplementary Budgets in low regard, as they tend to be only counter-cyclical (Campbell 1977). However, even this slight increase was politicized, especially after 1972, when "Prime Minister Tanaka used the supplement as the opening wedge in an attempt to break down Finance Ministry defenses" (Campbell 1977, p.206). As Meyer and Naka (1998) show, the LDP had a significant impact on the formation of supplementary budgets: in particular when the LDP had a larger share in the Diet and more legislators in the LDP were from rural areas, the LDP could change priorities in supplementary budget items.
6
With increasing frequency since the 1990s, supplementary budgets have been enacted to fund economic stimulus efforts. These counter-cyclical economic packages in the 1990s have become more and more ineffective: as Ihori et al (2003) show, increasing public investment in the 1990s crowded out private investment to some extent and did not increase private consumption much. In addition, they find the Keynesian effect was not observed in the 1990s: the adverse effect was often observed in recent years, with expansion of the Keynesian fiscal policy having a negative impact on Japanese economic performance.
As
Ihori et al (2003) suggest, this is partly because of the further politicization of supplementary budgets, of course in addition to the Japanese economy's structural malfunctioning. The government's expenditures are increasingly targeted as unproductive rents before elections as legislators in the Diet, pressured by local interest groups as principals, cannot form an effective collective action to reduce spending (cf. Grimes 2000; Patterson and Beason 2001). Special Accounts. Special Accounts are those accounts for special public programs, such as national forestry projects and public pension schemes, that are administered and managed by the government. They are "special" in the sense that, according to an official description, the government has to separately manage them from the general revenue and expenditures, in order to clarify the relationship between the beneficiary and costs-bearers. As of 2005, there are 31 special accounts, which total 387 trillion yen. Some of them are more or less financially independent: for example, the Motor Vehicle Inspection program is mainly financed by inspection fees and registration taxes from car owners. Others are much more dependent upon the government, as the national pension plan is financially supported partly by the regular budget. Several of these special accounts are notoriously politicized. One typical example is the coal program, in which the government subsidized coal production and passed the costs of adjustment onto the petroleum industry and ultimately the taxpayer. The program, which was launched in 1962, just after the political turmoil in the 1960 Mitui Miike labor strike, was essentially taxing imported heavy crude oil to help subsidize the domestic coal industry, which was no longer internationally competitive by the late 1950s. The petroleum refining industry also had to pay the costs of spurring economic demands in the local industries based on coal-mining companies; for example, the government gave subsidies for firms employing job-less coal-miners. The program continued through 2001 to delay adjustments for 40 years (Lesbirel 1991; Samuels 1987, pp.108-32; Uriu 1996, pp.95-102).
7
Lesbirel (1991) argues that imposing heavy taxes on the consumer to save uncompetitive firms was possible because some members from the LDP intervened in the process for electoral reasons. The Special Coal Policy Committee in the PARC, established in the 1960s, was a response to the political need to protect the interests' of the coal sector. The committee was dominated by Diet members from Kyushu, where most of Japan's coal production is located. FILP. In addition to regular and supplementary budgets, the government has the Fiscal and Investment Loan Program (FILP), or what is called the second budget. Launched in 1953, the FILP uses the government's capital stock – mainly from postal savings accounts – to loan financial resources to a range of public institutions. As of 1953, the number of recipients was only 15, while into the late 1990s, the number was increased to 70. Recipients range from highly public-regarded institutions like Pension Welfare Service Public Corporation and Japan Scholarship Foundation to the roads construction project, which is known as an example of the inefficient and sometimes corrupt iron triangle (e.g. Woodall 1996). Of 3.1 trillion yen (as of 2005) of loans poured into this road-making project, Japan Highway Public Corporation and its local affiliates receive most. Because of the nature of the program – i.e., recipients have to return loans and MOF bureaucrats at Trust Funds Bureau had relative autonomy in deciding how to allocate resources – some argue that the FILP is exempt from parochial political pressure.2 But, two things are of importance to note. First, as Patterson (1994) points out, public financial institutions receiving the FILP loans had to have their annual budgets and plans approved by the Diet. The Diet's monitoring was further strengthened in 1973, when a law was passed that required the FILP get Diet approval for investing five-year or longer assets managed.3 After that, each element of the FILP has been made part of the budget plan subject to legislators' demands. Second, loans from the FILP through public banks to private sectors were not productive (Beason and Weinstein 1996). Showing that loans from the Japan Development Bank, one of the largest recipients of FILP loans, were not diverted to the high growth industries, Beason and Weinstein (1996) conclude that this is partly because the government could not target 2
See the review by Patterson (1994), p.438. The law was titled "Law Concerning Special Measures for Long-Term Investment of the Trust Fund Bureau Funds and Funds Accumulated from Postal Annuities and Postal Life Insurance Annuity," or Shikin Unyōbu Shikin narabini Kan'i Seimei Hoken oyobi Yūbin Nenkin no Tsumitate Kin no Chōki Un'yō ni Taisuru Tokubetsu Sochi ni Kansuru Hōritsu.
3
8
properly, and partly because industrial policy considerations were dominated by the desire to aid declining sectors or protect the interests of large unproductive industries. The debate over the efficiency of industrial policy being aside,4 it is clear that a certain portion of the FILP was directed toward sectors that are known as corrupt or have been not necessarily productive throughout the postwar period.
A typical example of the
latter is public financial institutions specializing in loaning money to small and mid-size enterprises (SMEs) (Calder 1988), including: the People's Finance Corporation (Kokumin Kin'yū Kōko), the Small Business Finance Corporation (Chūshō Kigyō Kin'yū Kōko), and the Small Business Credit Insurance Corporation (Chūshō Kigyō Shin'yō Hoken Kōko). Patterson (1994) shows that the LDP's support rate among SMEs has a negative impact on the amount of loans from these SME-specialized institutions, arguing that the ruling party responded to losses of support from members of its small-business constituency with increased amounts of financial help. This clear politicization of a portion of the FILP is due to the fact that SMEs constitute the core support group of the LDP.
II.
Politicization of the Process Reviewing the existing literature on each of the budget categories – regular budgets, supplementary budgets, special accounts, and FILP – we have seen that politicization can be found and it became especially visible after the 1970s. The politicization of the budget making process was one of the factors that worsened the situation of the governmental budget as a common pool. What accounts for the different budget making patterns is still subject to ongoing debate. The bureaucrat-dominant view that was once commonly held (e.g. Johnson 1982) would see that Japanese bureaucrats, with their national, public and long-term goals, utilize every measure to allocate limited fiscal resources to only needed areas.5 Inefficient rents would be eliminated and thus there should be little space for political competition over fiscal resources. The consequence should be that Japan's budget, if 4
The best summary of the debate is available in Noble (1989). To be more precise, Johnson (1982) sees that Japanese bureaucrats could be autonomous as long as "the legislative and judicial branches of government must be restricted to 'safety valve' functions" (p.315). In his view, politicians only fended off the numerous interest groups, which if catered to would distort the priorities of the developmental state, to maintain bureaucrats' rule. Note his analysis ends in the early 1970s, when the politicization of the budget making process just started.
5
9
temporarily countercyclical to unexpected economic fluctuations, would be balanced at best (Campbell 1977). Although this view has not been systematically tested with fiscal data (e.g. McCubbins and Noble 1995), this bureaucratic-dominant argument has been challenged by several theoretical and empirical studies. First of all, as we have seen above, politicians are said to have encroached on the policymaking process and once autonomous MOF bureaucrats began to face increasing pressure from politicians. This is in part because throughout the long-term stable rule, LDP politicians could learn policy expertise (Murakawa 1989) and this expertise of politicians resulted in conflicts and cooperation between politicians and bureaucrats in various issues (Muramatsu and Krauss 1987). A slightly strong view even argues that politicians as principals dominated bureaucrats as agents with personnel management power (Ramseyer and Rosenbluth 1993). Second, bureaucracy cannot be assumed to be monolithic.
As Downs (1967)
noticed, there should be several types of bureaucrats even in the same agency; and different agencies might develop different institutions, rules, and preferences that are very likely to go against each other. MOF bureaucrats are known to be highly conscious about reducing expenditures (Grimes 2000), but bureaucrats in policy agencies usually have preferences to enlarge their policy areas that might involve budget expansion and thus compete with the MOF's balanced-budget preferences (cf. Niskanen 1975).6 This inter-ministry struggle gets fierce especially when spending ministries can receive backing from politicians. Third, an empirical study by McCubbins and Noble (1995) shows that changes in spending are not always incremental, that each agency and ministry does not necessarily receive a stable or fair share of the budget each year, and that spending on a program can be retrenched under certain circumstances. These empirical findings, contrary to the commonly held view that the Japanese budget is so balanced and tightly managed by the MOF (cf. Campbell 1977), are consistent with the fact that the politicization of the process occurred and budget items were increasingly under pressure from politicians. These contending views suggest that what is important here is to explore why the politicization occurred and how one can measure the changing degrees of the politicization. There are three types of explanations: the veto player theory; the approach looking at the local bias of politicians, and the principal-agent relationship. Another possibility is the changing relationship between politicians and bureaucrats, but I temporarily assume that the 6
The best empirical example of an inter-ministry bureaucratic battle is Johnson (1989). 10
politician-bureaucrat relationship can be endogenous to other political variables. In other words, the executive-legislature relationship, or the balance of power between the balanced-budged coalition (the cabinet and the MOF) and the budget-expansion coalition (backbenchers and spending ministries), can shape and define the ultimate political decisions to spend more or less. Veto Players Roubini and Sachs's (1989) study is one of the first attempts to uncover the relationship between political weakness and budget deficits. They find that the executive's political constraints, measured by a short tenure in office and a dispersion of political power across coalition partners, result in large budget deficits. In other words, when the executive is constrained by a large number of political parties with distinct policy preferences inside the coalition, then the need to buy off their support to maintain the office should result in generous spending. This finding has been confirmed in various ways. Kontopoulos and Perotti (1999) and Perotti and Kontopoulos (2001), for example, show that fragmentation both at the legislative level (i.e. the number of parties) and at the administrative level (i.e. the number of spending ministries) matters in explaining expansion of expenditures. This intuition is consistent with the veto player argument by Tsebelis (2002), who argues that reaching a consensus gets more difficult as the number of veto players increases. In the Japanese context, however, there are two limits in this theory. First, Japan has been dominated by one single party, so it is difficult to explain the fiscal-policy variations over time only by using the logic of the veto player theory. Second, even if we might assume that groups inside one party are so fragmented that they can be viewed as different policy groups,7 the veto player theory is indecisive at best in predicting policy directions, as Tsebelis (2002) himself admits. This is because the policy preferences of veto players can be quite complex. For example, one group's indifference curve may be flexible enough to include all the other parties' preferences in a coalition, meaning that increase in the number of veto players may not automatically result in policy change. Electoral Institutions When one sees an assembly as a group of legislators elected from local districts, such 7
This point will be addressed in detail later. 11
legislators should have coordination problems in streamlining the legislation process. Each legislator with distinct constituents, pursuing his/her parochial interests for local districts, competes with each other to appropriate limited fiscal resources. The result may be the tragedy of commons (Weingast et al 1981). Certain electoral institutions are known to cultivate the incentives to focus more on local goods (Cain et al 1987; Carey and Shugart 1995; Cox and McCubbins 2001). The Japanese electoral system before 1994 was unique in that voters cast one single non-transferable vote (SNTV) to one of the candidates in multimember districts (MMDs). This system forced a large party like the LDP to run multiple candidates to win a majority, and thus candidates from the same party had to appeal as individuals to their constituency. The result was that legislators had to differentiate themselves from other candidates in the same district. This called for them to specialize in unique policy areas where they could bring private policy benefits to targeted constituents as "policy tribes" (zoku giin) (Inoguchi and Iwai 1987; Ramseyer and Rosenbluth 1993). Scheiner (2005) sees this institutional basis for clientelism, combined with the fiscal centralization, as the main source for the opposition failure in Japan. When the central government grasps most of the fiscal resources, a ruling party can discretionarily and politically allocate money to local districts to exclude the opposition's access to this clienteslistic structure. Furthermore, this structure is further strengthened by the personal vote system, wherein legislators in local assemblies are tightly tied together with national legislators to exchange local benefits and votes. The problems in this local bias approach are: (i) how to reach this structural equilibrium is relatively unknown and (ii) politicians' power is implicitly assumed to be equal. First, even under the electoral systems that cultivate personal vote, politicians may take time to learn how the system works and how they can optimize their reelection possibilities. For example, Japanese legislators in the very early periods of postwar democracy could not coordinate their strategic entries into districts, contrary to the view that there should be M+1 candidates (Reed 1990). This suggests that one might uncover the historical process of how the structural equilibrium was achieved. Second, even if all the politicians have distinctly parochial interests, this might not result in the tragedy of the commons. This is because only a few politicians may monopolize the policymaking process to shut off relatively young, inexperienced backbenchers. Furthermore, when there is a rational reason to preclude politicians even in the same party from the policymaking process, the party leadership may not
12
want to allocate money to these rivals. In fact, intense factional competition and party presidential elections discouraged factional bosses in the LDP to allocate policy privileges equally. Thus institutions other than electoral systems (e.g. the degrees of democratic consolidation, legislators' capacities, and party rules) might deserve attention. The Party Leadership and Party Rules Thus far I have argued that, in order to look at the political dynamics surrounding fiscal policymaking decisions, one has to relax the party-as-a-unitary-actor assumption and pay attention to intra-party politics. Therefore I focus on those party rules that stipulate how backbenchers can have the lawmaking powers, participate in the policymaking process, and allocate policy benefits to their constituents. In other words, such rules should work as a structural bond that ties the leadership with otherwise would-be defectors in the party (Cox and McCubbins 1994). A study on Japan using this approach can be found in Ramseyer and Rosenbluth (1993), who see the intra-party dynamics as the principal-agent relationship between leaders and backbenchers. They argue that the relationship was efficiently managed, as backbenchers and factions without any policy preferences only delegate the authority for drafting national policy issues. Meanwhile, leaders could effectively control backbenchers, by utilizing the tightly structured career promotion ladder. This view correctly points to how leaders and backbenchers behaved in the equilibrium, but the problem inherent in it is that it is very static. The history of the intra-party politics of the LDP suggests that, in the 1950s and 1960s, backbenchers most often complained about leaders' tyrannical nature of policymaking and personnel management, while leaders most often ignored the career ladder and picked up their favorites to the highest posts (Watanabe 1958; 1967). Kawato (1996a; 1996b) points out the career ladder only emerged gradually after the late 1960s and was finally stabilized in the late 1970s. Therefore the more correct story is that the monopolized control by party leaders over policy and personnel issues in the 1950s was gradually replaced by the stable career ladder. The introduction of the fair and proportional rule was a response to backbenchers' internal constraints on party leaders, and these constraints represent the changing degrees of the power balance between leaders and backbenchers. When the decision making process is centralized enough to exclude the demands from rank-and-file members for both private and public goods, spending is under monopolized control of the central authority. If this is the
13
case, then party leaders can limit expansion of spending (Hahm et al 1996; Hallerberg 2004; Hallerberg and Marier 2004; Hallerberg and von Hagen 1999). Overall, the demand for spending more arises when the decision making process is decentralized. Seen this way, the politicization of the budget making process was actually party leaders' conscious attempts to decentralize the policymaking process to buy off would-be dissidents inside the party and expand support from the general voter. Party leaders inside the LDP had to decentralize the policymaking process to allow rank-and-file backbenchers to exert parochial interests for two reasons. First, there was the need to internalize negative externalities arising from the Japanese electoral system. As argued above, the electoral system encouraged legislators to develop local and parochial interests for the reelection purpose, but the party leadership in the early periods never allowed backbenchers to have the influential voice in the process. The cabinet, which should be read as the leadership of the LDP, monopolized the process. Although the 1955 LDP party rule already included a provision requiring that any party bills should be approved by the PARC before they would be submitted to the Diet (Article 30, Item 2) (Jiyū Minshutō 1987, p.32), the membership at some of the important decision-making bodies – such as the General Council and the PARC – was virtually monopolized by members from the party president's faction and allied factions, or "mainstream" (shuryūha) factions (Watanabe 1958). Legislators from the anti-mainstream factions kept challenging the leadership through a variety of filibustering measures, including sending many legislator-initiated bills to the floor and boycotting deliberations in standing committees. The party leadership demanded in 1960 that rank-and-file members should be prohibited from both approving and rejecting any bills if they were not based on the party's decision, and, if they would submit their own bills to the Diet, those bills should be approved by the General Council, Secretary-General, General Council Chair, PARC Chair, and Diet Affairs Committee Chair (Kokutai Iinchō) beforehand (Murakawa 1989, p.162). As a compromise, the government had to promise in 1962 that it should refer to the party before submitting government-initiated bills and budget plans to the Diet (Hoshi 2004). After this incident, any cabinet-initiated bills had to pass subdivisions in the PARC, which thus became virtual veto points for legislation (Satō and Matsuzaki 1986). PARC subdivisions became effective subgovernments where backbenchers could exchange their local interests to streamline the policymaking process in the Diet. Only after this practice appeared gradually after the end-1960s, policy tribes could exert their influence over policymaking.
14
Second, factions were increasingly becoming threats for party leaders. Access to posts important for policymaking was limited to the president's faction and its allies in the very early periods of the LDP. In fact, this limited access was a natural consequence of intense factional competition in the 1950s and 1960s over the party presidency: as the presidential elections required a majority to win, bosses had to reward their supporters with posts. Some favorite protégés could gain fast access to important posts; legislators with only one or two terms in the Diet could get cabinet posts if they made great contributions to their factional bosses (Watanabe 1958). The party president and secretary-general could monopolize personnel management as long as the anti-mainstream factions were too small to be threatening. After the 1970s, when the LDP's entire seat share declined and factions' size enlarged, each faction could have the power to threaten the stability of the party leadership (see Figure 5; also see Kohno 1992). After the 1976 election, each of the big five factions – Fukuda, Miki, Nakasone, Ōhira, and Tanaka – could be a threatening force to the cabinet; every faction except tiny ones became large enough to defect the LDP and make a non-LDP coalition government with opposition parties excluding the communists. As signaling to defect was now credible, party leaders now had to institutionalize fair personnel systems even for the antimainstream factions. Thus factional balance in allocating posts to rank-and-file members was designed to maintain members' loyalty to the party and moderate factional conflicts (Kawatō 1996a, 1996b; Satō and Matsuzaki 1986). Therefore by the mid-1970s, a set of institutions had been established to allow policy tribes and dissidents from the antimainstream factions to have voice in policymaking bodies inside the party and fair access to posts directly related to policy privileges. This came with large costs: with the increasing power of legislation, politicians now intensively sought for private goods that would improve their electoral fortunes. Many items in various budgets became the target for allocating benefits to local districts. Factions of Interest or Factions of Principle? Thus far we have focused on decentralization, but the argument may hold whether factions are viewed as non-ideological groups of interest (Bettcher 2005) or ideological groups of principle (McCubbins and Thies 1997; Thies 2001). The party leadership may want to exclude the antimainstream factions from the core of the coalition simply because it wants to allocate posts and privileges to its supporters or because it cannot allow ideologically distant
15
factions to get in to the policymaking process.8 I will test the implications resulting from these two competing assumptions in three ways. First, the assumption that factions are more like ideological entities may result in the same consequence as Roubini and Sachs (1989) and others hold: the more veto players, the more spending. Thus I will include the number of factions as an independent variable. If factions are only non-ideological groups, then increasing or decreasing the number of factions may not necessarily change fiscal performance. An alternative way may be looking at the effects of "which factions were in a coalition in a given year" on spending as McCubbins and Thies (1997) do, but this research design may not work for two reasons: first, if factions can flexibly and frequently change their policy preferences, then jumping into a coalition can cause a shift in their policy positioning; and second, this problem could be solved if we have an ex ante objective measurement of the policy positioning of each faction, but simply we lack such a systematic measurement. All McCubbins and Thies (1997) could test is to reject the null hypothesis – that factions do not matter – without any ex ante predictions. Second, if factions are groups only designed to enhance exchange of parochial interests and privileges, then they should be more sensitive to pork-related budget items. The politicization and therefore decentralization of the budget making process should have larger upward impact on pork spending than public goods. On the other hand, if factions are groups of ideological principle, then factions would show greater sensitivity to public policy items, or, at least, the difference between pork preferences and public policy orientations would be significantly small. Third, I will look at the dynamics of the party leadership. If party leaders are mere delegates from policy-seeking factions preoccupied with dividing a limited pie and expanding their policies, then the party leadership may engage in fiscal competition before the current winning coalition will be reshuffled – policy-seeking incentives outweigh office-seeking ones. Therefore policy-seeking leaders should pursue their interests more eagerly. But, if factions can more flexibly change their policy positions, then leaders may care more about the sustainability of the party's reputation, as their being in office ultimately depends on the capacity to maintain a majority in parliament – office-seeking incentives outweigh policy-seeking ones. Spending more is one of their viable options for survival, as it can make 8
As McCubbins and Thies (1997) raise almost the same point (fn.23), that this differentiation of factions of interest and factions of principle may not result in different conclusions makes research design much difficult. 16
an appeal to the general voter as a provider of public goods. Therefore office-seeking leaders should be interested more in protecting the size of a pie when their survival is shaky.
III.
Hypotheses and Data Based on the theoretical framework discussed above, I will hypothesize that: Hypothesis 1. If the personnel rule guarantees fair and proportional access even to the antimainstream faction members, then spending will increase in budget items. Hypothesis 2. The upward pressure of such fair and proportional access to the antimainstream faction members on spending will be more salient in pork categories. Hypothesis 3. If the party leadership is externally constrained by decline of the seat share of the party, then it would expand spending in public categories. I will test some alternative hypotheses based on the factions-as-ideological-groups
assumption: Null Hypothesis 4: If the number of factions increases, then spending will increase both in pork and public categories. Null Hypothesis 5: Any change in mainstream coalitions can result in change in spending patterns. My sample covers 63 regular-budget items authorized by 13 different ministries, from 1956 to 2003. The sample coincides those items used by McCubbins and Thies (1997). Their strategy is to collect data on various budget items to include a broad range of substantive types – small and large ones and public and pork types – from all the major spending ministries. The items have to cover no fewer than 10 consecutive years (see Table 1). I modified their data in two ways. First, to maintain the consistency of those budget items that changed only their titles, splintered some functions to newly created items, or integrated different items, I closely looked at Ōkurashō Shukeikyoku (various years) to check the subcomponents of items and added functionally equivalent items to the sample. For example, there was a legal change in 1982 to splinter an item newly created, "National
17
Supplement to Employees' Pension Fund," from the item "National Supplement to Social Insurance Fund." To maintain the consistency before and after 1982, I added these two items after 1982. Second, I expanded the observation periods. Their observation starts at 1961 and ends at 1984, but I added available information on the items before and after their periods. Now my data covers the period between 1956 and 2003 for some of the items. The panel is unbalanced, as some items have observations throughout the period, while others are not. My dependent variable is the share of the absolute money spent each year for each item in the GDP. As the panel is unbalanced, the panel-corrected-standard-error procedure is not usable here, but fixed-effect models with lagged dependent variables would be appropriate. Thus my structural model for an individual budget item i at t is: Itemi,t = b0 + b1*Itemi,t-1 + b2*Xt + b3*Ct + ai + ei,t where i represents each budget item (i = 1, ..., 63), t is the year (t = 1957, ..., 2003),9 X is a vector of political variables of interest, C is a vector of economic controls, ai is a group-specific constant term for i, and ei,t is an error term. Political independent variables include the cabinet balance, the LDP seat share in the Lower House, the effective number of factions in the LDP, and the number of the mainstream factions. A cautious note is required here about how to code these political variables. As the budget is legislated in the end of March every year, these variables have to represent the values as of March every year (see Table 2). I primarily used Satō and Matsuzaki (1986) to calculate these variables, but as it covers only before 1986, I used Asahi Nenkan (various years). In addition, as factional affiliations were highly fluctuated before 1963, I also used Leiserson (1968) and Watanabe (1958) to check if there is any inconsistency in the data. The variable the cabinet balance (cabbalance) measures how fair and proportional allocation of posts in the cabinet is. This is measured as: cabbalance = ∑N|EPj - Pj|, N=1, ..., j where EPj is an expected share in the cabinet of faction j, that is, total available posts in the cabinet times j's percentage size in the LDP; and Pj is an actual share in the cabinet of faction j. If expected and actual shares of all the factions closely coincide, the variable cabbalance 9
As models include a lagged dependent variable, the sample for analysis starts from 1957. 18
gets close to 0, while if there is a skewed balance among factions, then the variable becomes distant from 0. The history of LDP cabinets starts with the Hatoyama cabinet launched in late 1955, which was highly exclusive in that he did not allocate any posts to ex-Liberals except Ohno faction members – especially, Yoshida's affiliates (members from Ikeda and Sato factions) got no post in the cabinet.10 After the Hatoyama cabinet, the general trend is that party leaders gradually introduced fair proportion to all the factions, while Koizumi is apparently trying to destroy this practice. The variables the LDP seat share (ldpshare) and the effective number of factions (enf) are intuitive as the names suggest. I will also include mainchange, a dummy representing whether or not there was any change in the mainstream coalition inside the LDP from the previous year. Economic controls include the growth rate (growth), the deflator (deflator), the unemployment rate (unemp), and the government's budget surplus (gov_sur), all taken from Nihon Tōkei Nenkan (various years). As the LDP was ousted from the government temporarily in 1993-94 and during this period the Hosokawa and Hata cabinets drafted the budget, I include a dummy for the year 1994 (y1994). In addition, to check if there are any effects of the 1994 electoral reform, a dummy for the years after 1994 (postreform) is also included. Furthermore, the effects of coalition cabinets (1984-6, 1994-6, and 1999-) will be checked with the number of parties in the government (coalition). The variable coalition is coded 1 if the LDP is a single party in the government. Summary statistics is given in Table 3, with predicted signs.
IV.
Analysis and Discussion The results of fixed-effect cross-section time-series estimation models are presented in Table 4. First of all, the lagged dependent variable shows very strong significance throughout the models, suggesting that there is a strong path-dependent effect in budget items. This is consistent with the literature on incrementalism for example found in Widlavsky
10
Ohno was treated really well because he, together with Taketora Ogata, ousted Yoshida from the Liberal president, closely contacted with Nobusuke Kishi and Bukichi Miki for a Democrats-Liberals merger, and consistently supported Hatoyama as the first president of the LDP (and therefore as the Prime Minister). 19
(1975). Second, the economic controls show that the Japanese fiscal policy patterns are countercyclical. Economic growth encourages the government to moderate expansion of the budget, while the economic hardships such as unemployment lead to fiscal measures to boost the economy. The sign of the variable gov_sur is positive and significant, suggesting that the government tends to expand budget items when it has the fiscal capacity to do so. Overall, combined with the above finding about the path dependent effects in budget items, the Japanese budget items tend to be incremental and functional to some degree. But the third and more important finding is that the political independent variables of interest show statistical significance, indicating that still political dynamics matters significantly in explaining variations in this incremental and functional process. The variable cabbalance shows the negative sign only in the pork model, as expected. In Models 1 and 2 – the models for all the budget items and public-policy items – the variable is not significant, suggesting that competing factions were not concerned with budget items related to public goods, whose benefits by definition are non-exclusive and non-rival to different factions. Rather, the significant and negative sign of cabbalance in Model 3 suggests that the goods that are exclusive in nature – private goods – are targeted as the goals of factional competition. This finding overall suggests that as the party leadership opened up the once monopolized personnel system to the antimainstream factions, budget items increasingly faced the upward pressures; this corresponds with the argument that policy tribes could exert more influential power to expand fiscal policy especially after the 1970s. On the other hand, as the party leadership tightened up the system again especially after Koizumi cabinets, budget allocation to private goods was significantly cut – this finally reached the revolt from a coalition of policy tribes and antagonistic factional bosses in the 2005 elections (Nemoto, Krauss, and Pekkanen 2006). The variable ldpshare is significant in the public goods model (Model 2), as expected. Statistically significant at the 5% level, this variable shows that the party leadership tried to buy off support from the general voter by expanding public policy budget items in the times of the political hardships. Pork-related items are not related to the seat share of the LDP, indicating that improving the seat share through distributing pork was not an option to the party; this suggests distributing pork was a matter of intra-party dynamics (as indicated by the significantly negative sign of cabbalance) and legislators' individual reelection goals, more than the party leadership's decisions – consistent with Hypothesis 3, as well as the literature on
20
personal vote and clientelism in Japan (Scheiner 2005). The effective number of factions is significant throughout the models. These suggest that the degree of factional fragmentation inside the LDP is a significant determinant for decisions to expand expenditures. To put it differently, buying support from veto players inside the LDP with expansion of budget items was required for streamlining the budget making process and, ultimately, the stability of the government. This is confirmatory to Null Hypothesis 4 that we wished to refute; we cannot negate the assumption that factions have their own policy preferences and the logical corollary that there are upward pressures for the costs of coordinating the differences in these preferences. Refinement of the theory is still necessary on this point. On the other hand, the dummy mainchange is not significant throughout the models, suggesting that the change of the winning coalition inside the LDP does not matter for spending patterns. Our Null Hypothesis 5 is refuted and this is encouraging for the argument that factions are more like non-ideological entities. Overall, the Japanese budget system is highly path-dependent and functional in that the budget size in the past and economic variables are significant throughout the models. However, even in this process, political variables are important in explaining the variations. Specifically, if the factional balance of cabinets – a measure for the party leadership's monopolization of important decision-making posts – was skewed to leaders, then the leadership could reduce expansion of spending in pork-related items. Meanwhile, factions were relatively unconcerned about public policy items that are non-exclusive and non-rival in nature. The LDP leadership, concerned about its nation-wide reputation, tended to expand public-regarded items to buy support from the general voter, but its bare majority status was not significant for pork spending, which was not necessarily effective in improving the party's national image.
V.
Conclusion The existing theory on government expenditures has difficulties in explaining fiscal policy patterns of a one-party democracy like Japan. I suggest in this essay that looking at the party structure – especially some party rules that stipulate the power balance between party leaders and backbenchers and therefore determine the degree of constraints on the party leadership – can help to understand political logic behind the decisions to increase/decrease
21
spending. I argue that the decision making process, proxied by how fair and proportional allocation of posts is, influences the degree of the internal constraints on party leaders: if the system is centralized enough to exclude the demands from rank-and-file members, then leaders can limit spending; on the other hand, if party leaders have to allow for would-be dissidents inside the party to have louder voice, then fiscal consequence will become more encompassing and expansionary. I have shown this argument with some empirical evidence. I also tried to shed a new light on the debate over whether or not factions inside the LDP are really ideological and have their own unique policy preferences. Evidence is mixed with both positive and negative results. First, there is the sign showing that factional competition only matters in pork-related items, suggesting that factions may be in fact only factions of interest. Second, on the other hand, as I have shown in the empirical section, factional fragmentation inside the party did result in expansion of the budget, conducive to the existing theory on the relationship between coalition governments and fiscal performance. One way to show whether or not factions are ideological should require more refined predictions on directions in different budget items. The case in this study is budget items under authority of 13 different ministries. The finding in this study is more consistent with the recent literature on the politicization of policymaking process than the traditional bureaucrat-dominant view that bureaucrats were highly autonomous in deciding how to allocate limited financial resources to productive and promising sectors. Substantial amounts of money went to some sectors that were not necessarily efficient and productive; rather, the decision to allocate money to these sectors was at least to some degree influenced by the internal dynamics of the LDP. The results of my models also suggest that, as the party leadership started to regain monopoly over personnel issues after Koizumi cabinets, allocation of now extremely limited financial resources to unproductive rents is being cut down.
22
References Aldrich, John (1995). Why Parties? Chicago: The University of Chicago Press. Barro, Robert J. (1979). "On the Determination of the Public Debt." Journal of Political Economy 87(5): pp.946-71. Beason, Richard, and David E. Weinstein (1996). "Growth, Economies of Scale, and Targeting in Japan (1955-1990)." Review of Economics and Statistics 78(2): 286-95. Bettcher, Kim Eric (2005). "Factions of Interest in Japan and Italy: The Organizational and Motivational Dimensions of Factionalism." Party Politics 11(3): 339-58. Cain, Bruce, John Ferejohn, and Morris Fiorina (1987). The Personal Vote: Constituency Service and Electoral Independence. Cambridge, MA: Harvard University Press. Calder, Kent E. (1988). Crisis and Compensation. Princeton: Princeton University Press. Campbell, John C. (1977). Contemporary Japanese Budget Politics. Berkeley: University of California Press. Carey, John M., and Matthew Soberg Shugart (1995). "Incentives to Cultivate a Personal Vote: A Rank Ordering of Electoral Formulas." Electoral Studies 14(4): 417-39. Cox, Gary W., and Mathew D. McCubbins (1994). "Bonding, Structure, and the Stability of Political Parties: Party Government in the House." Legislative Studies Quarterly 19(2): 215-31. Cox, Gary W., and Mathew D. McCubbins (2001). The Institutional Determinants of Economic Policy Outcomes. Presidents, Parliaments, and Policy. Stephan Haggard and Mathew D. McCubbins. Cambridge, Cambridge University Press: 21-63. Crepaz, Markus M. L., and Ann W. Moser (2004). "The Impact of Collective and Competitive Veto Points on Public Expenditures in the Global Age." Comparative Political Studies 37(3): 259-85. Downs, Anthony (1967). Inside Bureaucracy. Boston: Little Brown. Ferejohn, John (1986). "Incumbent Performance and Electoral Control." Public Choice 50: 5-26. Fiorina, Morris P. (1981). Retrospective Voting in American National Elections. New Haven: Yale University Press. Grimes, William W. (2000). Unmaking the Japanese Miracle. Ithaca: Cornell University Press. Hahm, Sung Deuk, Mark S. Kamlet, and David C. Mowery (1996). "The Political Economy of Deficit Spending in Nine Industrialized Parliamentary Democracies: The Role of Fiscal Institutions." Comparative Political Studies 29(1): 52-77. Hallerberg, Mark (2004). Domestic Budgets in a United Europe: Fiscal Governance from the End of
23
Bretton Woods to EMU. Ithaca: Cornell University Press. Hallerberg, Mark, and Jurgen von Hagen (1999). Electoral Institutions, Cabinet Negotiations, and Budget Deficits in the European Union. Fiscal Institutions and Fiscal Performance. James M. Poterba and Jurgen von Hagen. Chicago, The University of Chicago Press: 209-32. Hallerberg, Mark, and Patrik Marier (2004). "Executive Authority, the Personal Vote, and Budget Discipline in Latin American and Caribbean Countries." American Journal of Political Science 48(3): 571-87. Heller, William B. (2001). "Political Denials: The Policy Effect of Intercameral Partisan Differences in Bicameral Parliamentary Systems." Journal of Law, Economics and Organization 17(1): 34-61. Hoshi, Hiroshi. 2004. "Jimintō Seichōkai-to Seisaku Kettei Katei." in Kimihiko Kitamura et al. eds. Gendai Nihon Seitō Shiroku. Vol.5: '55-Nen Taisei Ikō-no Seitō Seiji. Tokyo: Daiichi Hōki. 401-38. [in Japanese] Ihori, Toshihiro, Toru Nakazato, and Masumi Kawade (2003). "Japan's Fiscal Policies in the 1990s." World Economy 26(3): 325-38. Inoguchi, Takashi, and Tomoaki Iwai (1987). Zoku Giin no Kenkyū. [The Study of Policy Tribes.] Tokyo: Nihon Keizai Shimbun Sha. [in Japanese] Jiyū Minshutō (1987). Jiyū Minshutō Tōshi: Shiryō Hen. Tokyo: Jiyū Minshutō. [in Japanese] Johnson, Chalmers (1982). MITI and the Japanese Miracle. Stanford: Stanford University Press. Johnson, Chalmers (1989). "MITI, MPT, and the Telecom Wars: How Japan Makes Policy for High Technology." in Chalmers Johnson, Laura D'Andrea Tyson, and John Zysman. eds. Politics and Productivity: The Real Story of Why Japan Works. Cambridge, MA: Ballinger. pp177-239. Kawato, Sadafumi (1996a). "Jimintō ni Okeru Yakushoku Jinji no Seidoka." [The Institutionalization of Rules Governing Promotion in the LDP.] Hōgaku 59(6). 933-57. [in Japanese] Kawato, Sadafumi (1996b). "Sinioriti Rūru to Habatsu: Jimintō ni Okeru Jinji Haibun no Henka." [The Development of Seniority and Interfactional Balancing Rules in the LDP.] Revaiasan Extra Issue. 111-45. [in Japanese] Kohno, Masaru (1992). "Rational Foundations for the Organization of the Liberal-Democratic Party in Japan." World Politics 44(3): 369-97. Kontopoulos, Yianos, and Roberto Perotti (1999). Government Fragmentation and Fiscal Policy Outcomes: Evidence from OECD Countries. Fiscal Institutions and Fiscal Performance.
24
James M. Poterba and Jurgen von Hagen. Chicago, The University of Chicago Press: 81-102. Leiserson, Michael (1968). "Factions and Coalitions in One-Party Japan: An Interpretation Based on the Theory of Games." American Political Science Review 62(3): 770-87. Lesbirel, S. Hayden (1991). "Structural Adjustment in Japan: Terminating 'Old King Coal'." Asian Survey 31(11): 1079-94. Makihara, Izuru (2003). Naikaku Seiji to "Ōkurashyō Shihai". [The Cabinet Politics and the "Rule of the MOF".] Tokyo: Chūō Kōron Sha. [in Japanese] McCubbins, Mathew D., and Gregory W. Noble (1995). Perceptions and Realities of Japanese Budgeting. Structure and Policy in Japan and the United States. Peter F. Cowhey and Mathew D. McCubbins. Cambridge, Cambridge University Press: 81-115. McCubbins, Mathew D., and Michael F. Thies (1997). "As a Matter of Factions: The Budgetary Implications of Shifting Factional Control in Japan's LDP." Legislative Studies Quarterly 22(3): 297-328. Meyer, Steven A., and Shigeto Naka (1998). "Legislative Influences in Japanese Budgetary Politics." Public Choice 94(3/4): 267-88. Milesi-Ferretti, Gian Maria, Roberto Perotti, and Massimo Rostagno (2002). "Electoral Systems and Public Spending." Quarterly Journal of Economics 117(2): 609-57. Murakawa, Ichirō. 1989. Jimintō-no Seisaku Kettei Sisutemu. Tokyo: Kyōikusha. [in Japanese] Muramatsu, Michio, and Ellis S. Krauss (1987). The Conservative Policy Line and the Development of Patterned Pluralism. The Political Economy of Japan, Vol. 1: The Domestic Transformation. Kozo Yamamura and Yasukizi Yasuba. Stanford, Stanford University Press: 516-54. Nemoto, Kuniaki, Ellis S. Krauss, and Robert Pekkanen (2006). "Party Discipline and Policy Defection." manuscript. Nihon Tōkei Nenkan (various years). Tokyo: Sōrifu Tōkeikyoku. Niskanen, Willam A. (1975). "Bureaucrats and Politicians." Journal of Law and Economics 18(3): 617-43. Noble, Gregory W. (1989). The Japanese Industrial Policy Debate. Pacific Dynamics: The International Politics of Industrial Change. Stephan Haggard and Chung-in Moon. Boulder, Westview: 53-95. Noguchi, Yukio (1991). Budget Policymaking in Japan. Parallel Politics: Economic Policymaking in the United States and Japan. Samuel Kernell. Washington, Brookings Institution: 119-41.
25
OECD (2005). Economic Outlook. Vol. 78, Release 2. Paris: OECD. Ōkurashō Shukeikyoku (various years). Ippan Kaikei Saishutsu Sainyū Kessan. Tokyo: Ōkurashō Shukeikyoku. Ōkurashō Shukeikyoku Chōsaka (various years). Zaisei Tōkei. Tokyo: Ōkurashō Shukeikyoku Chōsaka. Patterson, Dennis (1994). "Electoral Influence and Economic Policy: Political Origins of Financial Aid to Small Business in Japan." Comparative Political Studies 27(3): 425-47. Patterson, Dennis, and Dick Beason (2001). "Politics, Pressure, and Economic Policy: Explaining Japan's Use of Economic Stimulus Policies." World Politics 53(4): 499-523. Perotti, Roberto, and Yianos Kontopoulos (2002). "Fragmented Fiscal Policy." Journal of Public Economics 86(2): 191-222. Persson, Torsten, and Guido Tabellini (1999). "The Size and Scope of Government: Comparative Politics with Rational Politicians." European Economic Review 43(4-6): 699-735. Persson, Torsten, and Guido Tabellini (2003). The Economic Effects of Constitutions. Cambridge: MIT Press. Ramseyer, J. Mark, and Frances McCall Rosenbluth (1993). Japan's Political Marketplace Cambridge, MA: Harvard University Press. Reed, Steven R. (1990). "Structure and Behavior: Extending Duverger's Law to the Japanese Case." British Journal of Political Science 20(3): 335-56. Roubini, Nouriel, and Jeffrey Sachs (1989). "Government Spending and Budget Deficits in the Industrial Countries." Economic Policy 4(8): 99-132. Samuels, Richard J. (1987). The Business of the Japanese State. Ithaca: Cornell University Press. Satō, Seisaburō, and Tetsuhisa Matsuzaki. 1986. Jimintō Seiken. Tokyo: Chūō Kōron. [in Japanese] Savage, James D. (2000). "A Decade of Deficits and Debt: Japanese Fiscal Policy and the Rise and Fall of the Fiscal Structural Reform Act of 1997." Public Budgeting and Finance 20(1): 55-84. Scheiner, Ethan (2005). Democracy Without Competition in Japan. Cambridge: Cambridge University Press. Thies, Michael F. (2001). "Keeping Tabs on Partners: The Logic of Delegation in Coalition Governments." American Journal of Political Science 45(3): 580-98. Uriu, Robert M. (1996). Troubled Industries. Ithaca: Cornell University Press. Volkerink, Bjorn, and Jakob de Haan (2001). "Fragmented Government Effects on Fiscal
26
Policy: New Evidence." Public Choice 109(3/4): 221-42. Watanabe, Tsuneo (1958). Habatsu: Hoshutō-no Kaibō. Tokyo: Kōbundo. [in Japanese] Watanabe, Tsuneo (1967). Habatsu to Tatōka Jidai. Tokyo: Sekkasha. [in Japanese] Weingast, Barry R. (1979). "A Rational Choice Perspective on Congressional Norms." American Journal of Political Science 23(2): 245-62. Weingast, Barry, Kenneth Shepsle, and Christopher Johnsen (1981). "The Political Economy of Benefits and Costs: A Neoclassical Approach to Distributive Politics." Journal of Political Economy 89: 642-664. Wildavsky, Aaron B (1975). Budgeting: A Comparative Theory of Budgetary Processes. Boston: Little Brown. Woodall, Brian (1996). Japan under Construction. Berkeley: University of California Press. Wright, Maurice (1999). Coping with Fiscal Stress: Illusion and Reality in Central Government Budgeting in Japan, 1975-1997. Fiscal Institutions and Fiscal Performance. James M. Poterba and Jurgen von Hagen. Chicago, The University of Chicago Press: 349-76. Zaisei Kin'yū Tōkei Geppō (various issues).
27
Figure 1: Outstanding Bonds and Tax Revenues 600000
500000
Million Yen
400000
300000
200000
100000
20 03
20 01
19 99
19 97
19 95
19 93
19 91
19 89
19 87
19 85
19 83
19 81
19 79
19 77
19 75
19 73
19 71
19 69
19 67
19 65
19 63
19 61
19 59
19 57
19 55
0
Year Outstanding Bonds
Tax Revenues
GDP
Source: Ōkurashō Shukeikyoku Chōsaka (various issues); Nihon Tōkei Nenkan (various issues).
28
Figure 2: General Government Gross Financial Liabilities 180.0 160.0
140.0 France
120.0
Germany Japan
100.0 %
UK 80.0
US Canada
60.0
Italy
40.0
20.0 0.0 1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
Year
Source: OECD (2005).
29
Figure 3: Government Consumption to the GDP a) 1952-79 0.25
0.2 Canada France
0.15
Germany Italy Japan
0.1
UK US
0.05
19 7
8
6 19 7
19 7
4
2 19 7
0 19 7
8 19 6
6 19 6
4 19 6
2 19 6
0 19 6
8 19 5
6 19 5
4 19 5
19 5
2
0
b) 1980-2000 0.3
0.25
0.2
Canada France Germany
0.15
Italy Japan UK US
0.1
0.05
0 1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
Source: IMF. International Financial Statistics.
30
Figure 4: The Number of Factions and the Effective Number of Factions in the LDP 16
14
12
10 Number of Factions Effective Number of Factions
8
Mainstream Factions 6
4
2
20 01
19 99
19 97
19 93 19 95
19 91
19 87 19 89
19 85
19 81 19 83
19 79
19 77
19 73 19 75
19 71
19 67 19 69
19 65
19 61 19 63
19 59
19 57
19 55
0
Source: Leiserson (1968) and Watanabe (1958) for 1957; Sato and Matsuzaki (1986) for every election year from 1955 to 1983; Asahi Nenkan (various years)
31
0.7
5
0.6 0.5
4
0.4
3
0.3
2
0.2
1
0.1
0
0
The Share of the LDP
6
Threatening Factions LDP Share
1956 1960 1964 1968 1972 1976 1980 1984 1988 1992 1996 2000 2004
The Number of Threatening Factions
Figure 5: The Number of Threatening Factions and the Share of the LDP in the Lower House
Year
Source: See Figure 4. Note: The "Threatening" factions are defined as those factions whose size is larger than a [bare majority minus non-communist opposition parties]; that is, such factions are large enough to defect the LDP and make a non-LDP coalition government with opposition parties excluding the communists.
32
20 04
20 02
20 00
19 98
19 96
19 94
19 92
19 90
19 88
19 86
19 84
19 82
19 80
19 78
19 76
19 74
19 72
19 70
19 68
19 66
19 64
19 62
19 60
19 58
19 56
Figure 6: Cabinet Balance of Factions 1.4
1.2
1
0.8
0.6
0.4
0.2
0
Source: See Figure 4.
33
Table 1: Expenditure Items Agency General Administrative Agency of the Cabinet Ministry of Education / Okinawa Dev't Agency National Land Agency Hokkaido Development Agency Hokkaido Development Agency Hokkaido Development Agency Defense Agency Defense Agency Defense Agency Defense Agency Defense Agency Ministry of Education Ministry of Education Ministry of Ministry of Ministry of Ministry of
Education Education Education Health and Welfare
Ministry of Health and Welfare Ministry of Health and Welfare Ministry of Health and Welfare Ministry of Health and Welfare Ministry of Health and Welfare Ministry of Ministry of Ministry of Ministry of Ministry of
Health and Welfare Health and Welfare Health and Welfare Health and Welfare Health and Welfare
Ministry of Health and Welfare Ministry of Agriculture, Forestry, and Fisheries Ministry of Agriculture, Forestry, and Fisheries Ministry of Agriculture, Forestry, and Fisheries
Available Item Years 1956-2003 Pension Expenditures 1972-2003 Okinawa Education Promotion National Integrated Land 1974-2003 Development Coordination Hokkaido Road Construction 1956-2003 Expenditure 1969-2003 Hokkaido Fishing Port Facilities Hokkaido Housing Construction 1956-2003 Industry 1969-2003 Weapons and Vehicles 1961-2003 Aircraft 1957-2003 Shipbuilding 1957-2003 Maintenance of Facilities 1961-2003 Research and Development Subsidies for Educational Expenses 1957-2000 for the Handicapped 1956-2003 Educational Promotion Maintenance of Public Educational 1956-2003 Facilities 1956-2003 Scholarship Expenses 1956-2000 Science Promotion 1956-2003 Research and Development Subsidy to National Health Insurance 1956-2003 Plan Aid to Bereaved Families and Families 1956-2003 of Internees Maintenance of Environmental Health 1956-1999 and Sanitation Facilities Expenditure for Atomic Bomb 1957-2003 Victims 1956-2003 Sanitation Expenses Public Assistance (Lifestyle Protection) 1956-2003 Expenditure 1956-2003 Support for the Disabled 1963-2000 Welfare to the Elderly 1957-2003 Support for Women 1956-2003 Support for Children National Supplement to Social 1956-2003 Insurance Fund 1967-1999 Agricultural Promotion Strategic Restructuring of Agriculture 1961-1999 Agricultural Pensions and Related 1970-1999 Expenses
Type Public Pork Semi-Public Pork Pork Pork Pork Pork Pork Pork Pork Semi-Public Public Semi-Public Semi-Public Public Public Public Semi-Public Pork Semi-Public Public Public Semi-Public Semi-Public Semi-Public Semi-Public Public Semi-Public Pork Semi-Public
34
Ministry of Ministry of Ministry of Ministry of Ministry of
Agriculture, Forestry, and Fisheries Agriculture, Forestry, and Fisheries Agriculture, Forestry, and Fisheries Agriculture, Forestry, and Fisheries Agriculture, Forestry, and Fisheries
Ministry of Ministry of Ministry of Ministry of Ministry of Ministry of Ministry of
Agriculture, Forestry, and Fisheries Agriculture, Forestry, and Fisheries Agriculture, Forestry, and Fisheries Agriculture, Forestry, and Fisheries International Trade and Industry International Trade and Industry International Trade and Industry
Ministry of International Trade and Industry Ministry of International Trade and Industry Ministry of International Trade and Industry Ministry of International Trade and Industry Ministry of International Trade and Industry Ministry of Transportation Ministry of Ministry of Ministry of Ministry of
Transportation Transportation Transportation Transportation
Ministry of Post and Telecommunications Ministry of Ministry of Ministry of Ministry of Ministry of
Post and Telecommunications Labor Labor Labor Labor
Ministry of Ministry of Ministry of Ministry of
Construction Construction Construction Construction
Silk Cultivation and Horticulture 1956-1999 Promotion Diffusion of Agriculture Improvement 1956-2001 Strengthening of Wet Rice Agriculture 1969-1999 1956-1999 Promotion of Animal Husbandry 1965-2003 Sugar Price Stabilization Strategic Planning of Food Product 1974-1999 Distribution 1956-2003 Promotion of Forestry 1956-2003 Promotion of Fisheries 1956-2000 Fishing Harbor Facilities 1972-2000 Promotion of Industrial Relocation Promotion of the Computer Industry 1972-1998 1956-2003 Industrial Plumbing Facilities Promotion of Mining and 1956-2003 Manufacturing Technology Large-Scale Industrial Research and 1966-1992 Development 1969-2000 Agency of Industrial Science and Technology Experimental Research Lab 1962-2003 Underground Resources Strategy Small and Meduim-Sized Enterprise 1956-2003 Strategy Subsidies to Accounts of Japan 1968-1986 National Railways Corporation Subsidies to Japan National Railways 1965-1991 Construction Corporation 1956-1991 Railroad Track Maintenance 1960-2004 Coastline Enterprises Subsidies for Maritime Transportation 1965-2004 Integrated Telecommunications 1956-2003 Research Lab Local Telecommunications 1956-1995 Administration Bureau 1956-1995 Unemployment Expenditure 1966-2003 Career Change Planning 1956-1999 Labor Protection Office 1956-1999 Employment Security Office 1967-2003 Strategic Planning Regarding Crumbing of Steep Inclines and Related Expenses 1960-2003 Coastline Enterprises 1956-2003 Housing Construction 1956-2003 City Planning
Semi-Public Semi-Public Semi-Public Semi-Public Semi-Public Public Semi-Public Semi-Public Semi-Public Semi-Public Semi-Public Pork Semi-Public Pork Public Semi-Public Semi-Public Semi-Public Semi-Public Pork Pork Pork Semi-Public Public Semi-Public Semi-Public Public Public Semi-Public Semi-Public Semi-Public Semi-Public
Source: Ōkurashō Shukeikyoku (various years). 35
Table 2: Cabinets Year Cabinet 1956 Hatoyama C 1957 Ishibashi 1958 Kishi A 1959 Kishi B 1960 Kishi C 1961 Ikeda B 1962 Ikeda C 1963 Ikeda D 1964 Ikeda E 1965 Ikeda F 1966 Sato A 1967 Sato C 1968 Sato D 1969 Sato E 1970 Sato F 1971 Sato F 1972 Sato G 1973 Tanaka B 1974 Tanaka C 1975 Miki A 1976 Miki A 1977 Fukuda A 1978 Fukuda B 1979 Ohira A 1980 Ohira B 1981 Suzuki A 1982 Suzuki B 1983 Nakasone A 1984 Nakasone B 1985 Nakasone C 1986 Nakasone D 1987 Nakasone E 1988 Takeshita A 1989 Takeshita B 1990 Kaifu B 1991 Kaifu C 1992 Miyazawa A 1993 Miyazawa B 1994 Hosokawa 1995 Murayama A (to be continued)
Mainstream Hatoyama; Kishi; Miki-Matsumura; Ohno; Ishii Ishibashi; Miki-Matsumura; Ono; Ishii; Ikeda Kishi; Sato; Ohno; Kohno Kishi; Sato; Ohno; Kohno Kishi; Sato; Ohno; Kohno Ikeda; Sato; Kishi Ikeda; Sato; Fukuda; Kawashima Ikeda; Kohno Ikeda; Kohno; Ohno Ikeda; Ohno; Kohno; Miki; Kawashima Sato; Fukuda; Funada; Murakami; Miki; Mori; Kawashima; Ishii Sato; Fukuda; Kawashima; Ishii; Funada; Sonoda; Miki Sato; Fukuda; Kawashima; Ishii; Funada; Sonoda; Miki Sato; Fukuda; Kawashima; Ishii; Funada; Sonoda Sato; Fukuda; Kawashima; Ishii; Funada; Sonoda Sato; Fukuda; Kawashima; Ishii; Funada; Sonoda Sato; Fukuda; Ohira; Shiina; Ishii; Funada; Mizuta; Nakasone; Sonoda Tanaka; Ohira; Nakasone; Miki; Shiina; Ishii; Funada; Mizuta Tanaka; Ohira; Nakasone; Miki; Shiina; Ishii; Funada; Mizuta Miki; Fukuda; Nakasone; Shiina; Ishii; Funada; Mizuta Miki; Fukuda; Nakasone; Shiina; Ishii; Funada; Mizuta Fukuda; Ohira; Tanaka; Shiina; Ishii; Funada; Mizuta Fukuda; Ohira; Tanaka; Shiina; Ishii; Funada; Mizuta Ohira; Tanaka Ohira; Tanaka Suzuki; Tanaka; Nakasone Suzuki; Tanaka; Nakasone Suzuki; Tanaka; Nakasone Suzuki; Tanaka; Nakasone Suzuki; Tanaka; Nakasone; Komoto; Fukuda* Suzuki; Tanaka; Nakasone; Komoto; Fukuda* Miyazawa; Tanaka; Nakasone; Komoto; Fukuda* Miyazawa; Takeshita; Nakasone; Komoto; Abe* Miyazawa; Takeshita; Nakasone; Komoto; Abe* Takeshita; Nakasone; Komoto; Abe Takeshita; Watanabe; Komoto; Abe Takeshita; Miyazawa; Komoto Obuchi; Miyazawa; Komoto Miyazawa; Obuchi; Mitsuduka
36
(continued from the last page) Year Cabinet Mainstream 1996 Hashimoto A Obuchi; Watanabe; Komoto; Miyazawa (Kato) 1997 Hashimoto B Obuchi; Watanabe; Komoto; Miyazawa (Kato) 1998 Hashimoto C Obuchi; Watanabe; Komoto; Miyazawa (Kato) 1999 Obuchi B Obuchi; Komoto; Miyazawa (Kato); Watanabe (Yamazaki) 2000 Obuchi C Mori; Kamei; Komoto; Kono; Obuchi 2001 Mori C Mori; Kamei; Hashimoto; Komoto; Kono; Horiuchi 2002 Koizumi A Mori† 2003 Koizumi B Mori† 2004 Koizumi C Mori† Source: Satō and Matsuzaki (1986) for cabinets before 1986; Asahi Nenkan (various years) for cabinets after 1986. Note: * It is difficult to define the mainstream factions from Nakasone's second reshuffled cabinet (1985) to Takeshita's reshuffled cabinet (1989) as grand coalitions of all factions (Sō Shuryūha) occurred. † Koizumi cabinets (2002-2004) are exceptional in that usual factional coalitions up to the presidential elections cannot be observed. I tentatively assigned the mainstream-ship only to the Mori faction.
37
Table 3: Summary Statistics IV Name Mean Std. Dev.
Min
Max
cabbalance ldpshare enf growth deflator unemp gov_sur
0 0.39 4.86 -0.012 -0.024 0.011 -0.14
1.19 0.64 13.63 0.12 0.19 0.055 0.27
0.32 0.55 7.11 0.051 0.035 0.024 0.052
0.21 0.064 1.66 0.038 0.042 0.011 0.093
Predicted Sign Public 0 0 + + +
Pork 0 0 + + +
38
Table 4: Estimation Results Model 1 Sample: All lagged DV Primary IVs
Econ Controls
Pol Controls
Model 2 Sample: Public 0.98*** 0.98*** (0.0045) (0.0052) cabbalance -0.000050 -0.000048 (0.000032) (0.000043) ldpshare -0.00024** -0.00033** (0.00011) (0.00015) enf 0.000011*** 0.000012** (0.0000038) (0.0000050) mainchange 0.0000063 0.0000087 (0.0000066) (0.0000088) growth -0.00064*** -0.00073*** (0.00013) (0.00017) deflator 0.00034*** 0.00043*** (0.00012) (0.00016) unemp 0.0015* 0.0017 (0.00086) (0.0012) gov_sur 0.00023*** 0.00027*** (0.000054) (0.000072) postreform 0.00000050 0.0000011 (0.000015) (0.000020) coalition -0.000031*** -0.000040*** (0.0000087) (0.0000088) y1994 0.00010** 0.00014** (0.000046) (0.000061)
Model 3 Sample: Pork 0.97*** (0.0093) -0.000053** (0.000026) -0.000017 (0.000085) 0.0000084*** (0.0000030) 0.000000041 (0.0000051) -0.00040*** (0.00010) 0.00011 (0.000097) 0.00082 (0.00068) 0.00011** (0.000043) 0.0000043 (0.000012) -0.0000080 (0.0000068) 0.0000073 (0.000036)
Constant
0.00010 0.00016* -0.000017 (0.000068) (0.000092) (0.000053) obs 2646 1932 714 R2 0.9828 0.9822 0.9891 Note: *** p<0.01, ** p<0.05, * p<0.1. Standard Errors are in the parantheses.
39
