Chapter 2: Forms of Fiscal Governance

From Hallerberg, Mark. The Treaty of Maastricht and the Making of Budgets in Europe 1980-2002. Book Manuscript, University of Pittsburgh. Abstract
This chapter defines three ideal forms of fiscal governance. Fiefdom exists when government ministers set their own budgets. Ministers do not consider the full tax implications of their decisions, and the result is a severe common pool resource problem. Both delegation to a strong finance minister and commitment reduce the severity of common pool resource problems. The chapter illustrates how the ideal types function in practice, and it explains why one-party and coalition majority governments generally choose delegation while coalition governments where parties run against each other in future elections choose commitment. It also indicates that minority governments lead to a mix of the ideal forms of government. Centralization around a finance minister occurs within the government, but fiscal discipline is tightened only when the government and selected opposition parties commit themselves to detailed budgetary contracts. The final part of the chapter presents evidence that the theoretical framework can explain the relative fiscal performance of European Union states 1980-2000.

Introduction To understand budgetary outcomes one must understand how budgets are made. This chapter provides definitions for four ideal types of decision-making, which I term forms of fiscal governance. To illustrate these ideal forms, I begin with a simple formal model of the budget-making process in parliamentary democracies where the principle decisions on the budget are made in the cabinet. The key premise is that every government faces a Common Pool Resource (CPR) Problem. This problem arises when policy-makers consider the full benefits of their spending decisions on their constituencies but only part of the tax burden. In practice, the problem permeates decision-making. Policy-makers for sound political reasons do not have an incentive in most cases to consider the implications of their spending decisions upon the full tax burden. If policy-makers do little to address this problem, then a fiefdom form of governance predominates. Policy-makers consider their domains their fiefdoms, and they policy decide spending levels more or less in isolation from one another. There are, however, two ideal forms of governance conducive to solving the CPR Problem, a mode of delegation and a mode of commitment. Delegation involves vesting the Finance Minister with significant decision-making powers over public monies. Under commitment a group of agents with similar decision making rights enters an agreement, or a fiscal contract, to commit themselves strictly to budgetary norms, i.e., targets for
1

budget aggregates set for one or several years. A third form of governance that solves the CPR problem, which is found in minority governments only and that combines elements of both ideal forms, I define to as mixed. Each of these forms of governance to solve the CPR problem has distinct implications for the adoption, and for the effectiveness, of fiscal institutions. In delegation states, the emphasis is on improving the discretion of the finance minister in the budgetary process. One expects formal or informal rules that enhance the position of the finance minister. Moreover, spending targets such as caps on personnel costs, explicit mechanisms to deal with open-ended expenditures, and the like that make the budget more manageable for the finance minister are likely. In contrast, in commitment states one expects a range of formal rules to maintain the fiscal contract among the political parties who make the initial agreements. This mode of governance is more rule-based, and a range of multi-annual targets and sub-targets as well as rules to deal with unexpected shocks so that the initial agreement is not broken should be common. The mixed form of governance has elements of delegation in the budget deliberations that take place within the cabinet and elements of commitment in the contracts the government signs with one or more opposition parties in parliament. This chapter is organized as follows. Section 1 begins with a discussion of the three stages of the budget process. The second section develops a formal model that operationalizes policy outcomes as decisions on expenditure and taxation levels. 1 The model indicates that each type of governance has clear implications for the design of fiscal institutions at each stage of the budget process. Section 3 provides preliminary evidence about differences across forms of governance. Section 4 concludes.

The model is based on similar models presented in von Hagen and Harden (1994), Hallerberg and von Hagen (1999), and Hallerberg (2000). Von Hagen and Harden (1994) is concerned with the common pool problem in cabinets. It does not have parties as actors, it assumes that ministers worry about ministry constituencies only, and it is a one-period game. Hallerberg and von Hagen (1999) is multi-period and considers the effects of the CPR problem on the size of deficits. The focus remains on the cabinet, not the party. It also links the type of electoral system with the institutional choice of a solution to the CPR problem. Hallerberg (2000) develops an extensive form game that stresses the monitoring and punishment aspects of the different ideal types of government. This paper considers explicitly parties as actors. While the substantive conclusions do not differ all that much from the earlier models, the consideration of parties allows a more explicit comparison with coalition theorists.

1.

The Budget Process

There are three principle stages of the budget process. It is possible that the stages

of the process overlap, but the different stages in an ideal sense allow us to consider how these stages may differ. The first is the planning stage. Governments make forecasts about revenues, spending needs, economic growth, and the like for future years. The planning can be detailed and go forward several years. It can also bind the government to certain levels of spending and/or certain levels of revenue unless new legislation is passed to change the budget plan, as the multi-annual plan does in Sweden today. The plan can also be vague, restricted to only the next budget year, and have little or no practical impact in the formation of the budget. This has been the case in Greece from (re)democratization in 1974 until quite recently. The second stage is the decision-making stage. In practice, there are two sequential parts. In the first, the government formulates its annual budget. In the second, the legislature passes a budget. In theory it is the legislature that ultimately possesses the power of the purse in all democracies. In practice, however, the legislatures input in the annual budget process varies widely. In some cases, the legislature makes significant changes to the executives budget proposal. Members of Congress in the US are fond of calling the presidents budget dead on arrival, implying that the president will get little of what he wants once Congress is through with the budget. In other cases, the legislature serves as little more than a rubber stamp of the governments original budget proposal. The third stage is the implementation of the budget. In some countries, the government is legally required to execute the budget exactly as the legislature passed it. In other countries, government ministers have either implicit or explicit permission to spend less or to spend more than what appears in the budget. For example, the Italian Treasury Minister in the mid-1990s, Carlo Ciampi, blocked the spending of money the Italian Chamber of Deputies had already approved (Chapter 7). On the other side of the coin, a Francophone Education Minister in mid-1980s Belgium spent millions of francs more than his ministry was allotted in the budget with no more justification than a vague claim that he had more Francophone students than he expected (see Chapter 5). An important institutional question concerns how much the government itself knows about what it is spending and what revenues it collects as well as when it knows these figures.
3

If the government (and, for that matter, the legislature) has little ability to monitor its spending and its revenue collection there may be fairly wide disparities between the budget that passes the legislature and the budget the government ultimately implements.2 The next section presents a model of the decision-making process within the cabinet. It then provides definitions for different packages of fiscal rules, and it explains how to identify the different forms of fiscal governance in practice at different stages of the budget process.

2. A Model of Budget-Making in Parliamentary Cabinets

a. Ministry and Party Ideal Budgets
There are three relevant sets of actors in the model of decision-making within the cabinet. Spending ministers S1,2,,n make bids for the level of spending to be allocated to their ministries. They also execute the spending of budgets allocated to them once the budget is approved. All spending ministers together form the government G. A subset of the members of government belong to political parties P1,2,,l. Spending ministers seek full funding for those programs they consider important to reach their policy goals, and their proposals affect the spending side of the budget. I assume that members of the same political party share the same ideal budgets for each
* ministry.3 The variable xip represents the ideal spending level of party p for ministry i.

Ministers also seek to minimize the taxes that their constituencies must pay, and mi is the amount of the total tax burden that the minister expects her constituency to bear. This amount can be equal to or less than the total level of taxation on the ministers party constituency, so that mi m p . Finally, the minister may benefit simply from having larger budgets, and as a consequence she will request funds that are greater than the minimum needed to reach her policy goals (Niskanen 1971). The degree to which she values additional spending is represented by l. Assuming that the excess burden of taxation is quadratic, the ith minister will possess the utility function

Some authors treat the monitoring of the budget as a separate stage; see, for example, von Hagen (1992) and Hallerberg, von Hagen, and Strauch (2001). 3 This assumption can be relaxed by treating party factions as separate parties in the formal model.

Ui = lxi -

a m * 2 ( xi - xip ) - i T2 2 2

(1)

where xi is the amount of funding the ministry ultimately receives and

n

a the relative

weight that the minister places on spending concerns. If one assumes that there is a hard budget constraint, so that T = i=1 x i , then the spending minister j chooses a budget for his ministry that takes the form4

bsj =

l + ax *jp - m j a + mj

n i= 1, i j

xi

(2)

Now consider an alternative situation where the spending ministers reflect fully the preferences of their political party. In this case the ministers consider the effects of total spending and total taxation for party p, and they do not value higher spending for their ministry in its own right. The joint utility equation for the party is
mp 2 a 2 U p = - n ( x i -x * T , ip ) i= 1 2 2

(3)

and the budget for minister j would be

bpj =

ax *jp - m p

n i= 1, i j

xi

a + mp

(4)

Note that the size of the budget in both Equation 2 and Equation 4 is dependent upon what other ministers choose as their ideal budgets. To determine whether bsj > bpj in practice requires one further step. So that the argument is easy to follow, I assume that the ideal budgets for the political party are the same across ministries so that
4

Note that this hard budget constraint will be relaxed later in the chapter.

* * * x* p = x1p = x2 p = ... = xnp . I also assume for any ministries i and j that the budgets that the

ministers ultimately select are the same, such that b*j = bi* . Finally, I assume that ministry constituencies do not overlap such that m p = i=1 mi , and that mi = that is ideal for the party is
b* pj =
* p n

mp . The solution n

ax (5) a + nm p

while the ideal budget for the individual is

* bsj =

l + ax * l + ax * = a + nm j a + m p

(6)

Clearly, bsj > bpj when the minister does not consider the entire tax burden on the partys constituency and/or so long as l > 0, and this inequality indicates that the budget a given spending minister would like to propose is larger than the partys optimal budget. Given the importance of the tax burden to the outcome, how likely is it that ministers do not consider the partys full budget? Partial consideration of the tax burden is generally the rule for the following reasons. First, ministers are often judged by how well they protect the interests of the constituents of their particular ministry. An agriculture minister may care most about the effects of spending and taxation on farmers, while a labor minister has similar worries about the effects of spending and taxation on farmers. In Germany, for example, in Helmut Schmidts cabinet his colleague Hans Apel, a working class Social Democrat from the docks of Hamburg, became a voracious advocate of his ministry after his appointment as Defense Minister to such an extent that he was the only Bundestag member within his party besides his chancellor to support the stationing of American short-range missiles on German territory at the beginning of the 1980s. The general point is that where one stands on budget issues within ones party depends on where one sits at the cabinet table. If cabinet members cared only about the

tax burdens on their ministrys constituencies, then all mi would sum to the total tax burden on the cabinet.5 There is some empirical evidence supporting the contention that ministers care only about spending and taxation for their particular ministries. Kontopoulos and Perotti (1999) find that the more ministers in a given cabinet the higher the level of spending and the higher the budget deficit in a panel of OECD countries for the period 1970-1995. This is only one study, of course, but more empirical evidence will be presented in the case study chapters in this book. A second reason why ministers likely do not consider the full tax burden is that their parties are coalitions of different interests. Leaders within the party generally represent the spectrum of party supporters. In the March 2000 cabinet reshuffle in France, for example, Laurent Fabius as minister of finance represented the more monetarist wing of the socialist party while Martine Aubry remained as a more leftleaning minister of employment. In Germany, Norbert Blm came from a Christian Democratic Union post, and he was a steady advocate of pro-labor policies as Labor Minister within Helmut Kohls cabinet. Before moving on, one should note that the ideological distance among different factions within a political party is not a random phenomenon across countries. This variable is, to some degree, a function of the underlying electoral system. Under proportional representation systems with high district magnitudes, one would expect many political parties with potentially narrow constituencies. In contrast, under plurality one expects a two party system with big tent parties that bring together more diverse interests under one party label (Duverger 1954; Katz 1980). Moreover, plurality systems lead members of parties to care about geographically specific interests. As a consequence, Lancaster (1986) argues that legislators in plurality systems provide more pork barrel projects than legislators in PR systems. Weingast, Shepsle, and Johnsen (1981) similarly contend in an American system that legislators care only about the spending and the tax burden in their districts. Assuming that the tax burden is distributed equally across a given country, they assume that the amount of the tax burden a legislator considers in a legislature with m members is simply 1/m. Since cabinet members
5

This logic is the basis for the model in Hallerberg and von Hagen (1999).

generally come from the legislature in Europe, there is reason to believe that they consider only part of the total tax burden.6 In sum, if left alone there are good reasons to believe that ministers commonly do not consider the full tax burden when they decide the ideal spending level for their ministries. The shape the budget takes depends upon the decision-making process.

b. Fiefdom Governance
How are individual budget bids are translated into the total budget? The first case I consider is the fiefdom approach. Ministers decide the spending levels for their respective ministries, and the budget process involves simply aggregating the individual budget bids into a total budget. In practice, this model applies when the general cabinet sets the budget and spending ministers can coordinate log rolls with one another that allow them to get actual spending levels that correspond to their spending bids.7 What the budget ultimately looks like depends crucially upon the number, and the ideological distance among, parties in government because of differences in ideal budgets across parties. I begin with a one party government. For simplicity and without loss of generality, I again assume that the ideal budgets for the political party are the same across
* * * ministries so that x * p = x1 p = x 2 p = ... = x np . I also assume that the budgets that the

ministers ultimately select are the same, such that b*j = bi* . The aggregate level of spending if the ministers receive their ideal budgets is then
nl + nax * p Bi = a + mp (7)

which is larger than the collectively optimal budget of

6

For a discussion of the relationship among electoral systems, party discipline, and the choice of open or closed rules in European parliaments see Hallerberg (2002), for Latin America Hallerberg and Marier (2002). 7 Baron and Ferejohn 1989 remains the most succinct model. In their terminology the cabinets discussed here are merely small legislatures where the number of legislators is equal to the number of cabinet ministers. Their formal model indicates that votes with open rules in small legislatures and a costly delay can lead to universalistic solutions. The legislators unanimously agree on the outcome with each legislator receiving her preferred budget.

nax * p Bp = a + nm p

(8).

The situation under a multi-party government is somewhat more complicated but nevertheless tractable. The model assumes that parties hold different ideal spending preferences for each ministry, which in the model means that the parties have different x*i. Many coalition theorists assume, in fact, that negotiations concern the distribution of portfolios among parties with different policy preferences (Laver and Shepsle 1996). To simplify the discussion, consider the case with only two parties in government, party A and party B. Assume as well that party constituencies do not overlap so that
n m = i=1 mi , and that mi = m , where m is the total tax burden on the coalition. If one

simply adds up the spending requests of the different ministries under the two parties the total budget is
nA * i =1 nB i= 1

BD =

nl + a x iA + a x iB *

a +m

, nA+nB = n.

(7)

If the parties consider the full tax burden on the coalition instead, and if they set the spending levels that maximize their joint utilities (that is, they set the budget figures and not individual ministers so that l=0), then the budget takes the form
nA nB

Bd =

a x iA* + a x iB*
i =1 i =1

a + nm

(8)

Once again, spending under the fiefdom approach is higher than if the parties had jointly set the budgets. The general problem is therefore the same under both one-party and multi-party coalition governments. When decision-making is decentralized the players suffer from a common pool resource (CPR) problem. The greater the fragmentation of the cabinet into different cabinet posts, the smaller the tax burden a given minister considers and the larger the CPR problem. All of the players suffer from additional spending by other ministers that comes from the same tax revenue pot. The utility equations for both individual ministers (equation 1) and political parties (equation 3) indicate that both types

of actors suffer utility losses because of higher spending by ministers who do not consider the full tax burden. This CPR Problem is therefore also a collective action problem. All actors would be better off if they would consider the entire tax burden than if they all considered the tax burden on only their ministry.8 One should note that, unlike in the forms of governance that follow, there is no internal logical that connects the institutions present at the three stages of the budget process to the fiefdom model. Instead, it is a matter of identifying whether fiefdom is present at one or more stages of the process. Concerning the decision-making stage, one looks for situations where the party that receives a given portfolio determines the policies of the ministry without interference from other coalition partners. If the minister does exactly what the party wants, then the budget decision reflects consideration of the tax burden on the given ministers party only, not on the entire coalition. If the ministers determine their budgets more or less autonomously, then the budget outcome for a given ministry matches that given in Equation 7. One way to facilitate this outcome is to have the full cabinet make decisions on the budget. This allows the ministers effectively to log roll their preferred ministerial budgets. One also expects little or no centralization of the budget process in the cabinet. The budget outcome is simply an aggregation of individual budget bids. The finance minister for her part serves as a glorified accountant who may audit books and keep track of the budget figures. She has no say on the final budget outcome.
Table 1: Indicators of a Fiefdom Form of Fiscal Governance

Rule 1. Parties that receive certain portfolios autonomously set policy in those portfolios (check coalition agreement) 2. The full cabinet votes on items in the budget 3. The budget is the sum of individual budget bids (bottom-up budgeting) 4. Finance Minister serves an accounting function 5. No multi-annual planning of consequence
8

This model is potentially applicable to parliaments as well if parliaments have the ability to change the budget and parliamentarians can set the budget figures for areas of their expertise as they choose. The question here concerns the relevant actors. Even in fiefdom states there is usually some party discipline, and majority governments usually pass their cabinets budget proposals. Parliament does play a more critical role in minority governments, and the fiefdom model appears in cases where the minority government must buy votes from opposition parties.

10

A likely symptom of fiefdom is that budget planning into the future has little or no practical importance, and indeed it may be totally absent. There is no functional need for fiscal planning because ministers are likely to ignore what planning does occur. This form of fiscal governance can exist at the implementation stage if ministers can easily adjust their spending levels without regard to the effects of their decisions upon the total tax burden. For example, it was routine in Italy under the pentapartito for parliament simply to approve any spending above what the minister originally received in the budget in supplemental budgets mid-year. Of course, it is possible that states manage to avoid a fiefdom outcome at one stage of the process but not at others. In Belgium in the mid-1980s, for example, the prime minister often negotiated fiscal contracts among his ministers that led to budgets that, on paper, might have approximated the budget in equation 8, which was the solution the parties themselves favored. Yet in practice ministers could spend pretty much what they wanted during the implementation stage, which led to a serious CPR problem and budgets that approximated equation 7.

c. Delegation Governance
How to solve collective action problems is the concern of a significant body of literature (Olson 1965; Ostrom 1990; Ostrom, Gardner, and Walker 1994). One possible solution is to appoint a central player serve as an entrepreneur, whose function is to assure that all actors choose to cooperate. To be effective, the entrepreneur must have the ability to monitor the other players, possess selective incentives that he can use to punish defectors and/or reward those who cooperate, and have some motivation to bear the costs of monitoring himself (Olson 1965; Cox and McCubbins 1993). The spending ministers themselves have reason to delegate power to such a fiscal entrepreneur. While most players have an individual incentive to "defect" or overspend, they usually prefer the solution where all players cooperate to the solution where all defect.9

It is true that, if a spending minister ignored completely the tax effects of spending, the minister would always want more spending and would not support delegation to a strong finance minister, but such ministers should be rare.

11

Among the relevant cabinet members, the minister of finance (or minister of economy in some political systems) is most likely to play the role of a 'fiscal' entrepreneur. He usually has the responsibility to coordinate the formation of the budget, and, fair or not, the overall conduct of fiscal policy, including the size of budget deficits, is the principal indicator that others use to judge his effectiveness. He may also have a trivial budget when compared with other ministers so that he cannot defect in the prisoner's dilemma game being played in the cabinet. His interests therefore generally coincide with the general interests. Indeed, as former Swiss Finance Minister Otto Stich noted after his resignation in 1995,"every federal ministry has its lobbyI have only the taxpayers. They are the real majority, but their interests are unfortunately not well represented in parliament.10 A second reason why the minister of finance (or economy) is the most likely entrepreneur is that he may have the institutional capability to monitor effectively the other ministries. The finance minister usually has an extensive staff trained in budgeting and accounting, and he is most likely to find tricks spending ministers use to justify additional spending. Because his prestige and hence his personal benefits depend on the effectiveness of his ministry, he has a private incentive to guarantee that the monitoring occurs. The only question is whether the finance minister has a privileged position in the budget process so that he can address the CPR problem. As the case study chapters will show, the power a minister of finance has varies from country to country. Yet it is important at this stage to explain what rules (formal as well as informal) strengthen the hand of the finance minister vis--vis other ministers. There are generally two stages of the budget process when the CPR problem is relevant, and when, as a consequence, the power of the minister of finance minister matterswhen the budget is being formulated (ex ante) and when the budget is implemented (ex post). Ex ante ministers submit budget bids that correspond to equation 2. Importantly, ministers would still like to receive budgets that correspond to equation 2 even if they do not receive this budget allocation in the budget law. This means that ex post ministers benefit from spending over their initial budget allocations. An effective finance minister

10

The Financial Times, September 1, 1995

12

most therefore hold a privileged position at both stages of the budget process to be effective. Table 2 displays a series of rules at the budget formation stage that strengthen the finance minister. In the first stage, the minister of finance can serve as an agenda setter in the formulation of the budget by possessing the right to make the first proposal for the budget. This power is undermined if spending ministers can ignore the finance ministers budget and ask for cabinet votes on the spending ministers budget bid, which allows the spending minister effectively to ignore the finance ministers proposal. This procedure is the one commonly found in France. The minister of finance, together with the prime minister, issues a framework letter for every spending ministry that indicates the level of spending that ministry should expect to receive in the next budget year (Chapter 4).

Another method ex ante is to have the minister of finance negotiate one on one with the spending ministers, as is done in the United Kingdom. The Chancellor of the Exchequer is generally regarded as second in power only to the prime minister, and she usually negotiates

Table 2: Indicators of Finance Minister Strength under Delegation during the Governmental Stage of the Budget
Rule 1. Finance Minister (FM) proposes the budget and serves as an agenda-setter 2. Spending Ministers cannot ask for individual votes on their budget bids 3. There are one-on-one negotiations between the FM and Spending Ministers 4. FM has veto right over budget issues 5. Cabinet cannot overrule a decision of the FM 6. FM must approve any changes to previous spending allocations
This table is based on the presentation in Hallerberg, Strauch, and von Hagen 2001.

one-on-one with spending ministers about their budget allocations. If there is a dispute between the finance minister and a given spending minister, it goes to a committee composed of senior ministers without portfolio for consideration and not to the full cabinet for resolution. These ministers do not have budgets of their own, and a logrolling

13

situation in favor of the spending minister is not possible. Since the senior ministers are appointed to consider the general interests of the cabinet as a well, they usually support the Chancellor (Chapter 4). Further rules that strengthen the power of the finance minister include a right to veto unilaterally any spending proposal, as is the case in Germany. One can strengthen this rule with a requirement that cabinet cannot overrule the decision of the finance minister, which is the case in Austria but not the case in Germany, where the Chancellor plus a majority of the cabinet can overrule a finance ministers veto. A final rule requires the finance minister approve any changes to multi-annual spending allocations. An effective delegation to a fiscal entrepreneur also requires that the actors ex post do not defect from their initial budget allocations. The minister of finance therefore needs to have adequate monitoring power. There also has to be a punishment mechanism available to the minister to punish intransigent spending ministers. As long as he has the ability to modify a spending minister's budget proposal, he can punish defectors in future years. If immediate action is required, he can appeal to the prime minister to take action, and, in the most extreme case, he can insist that the prime minister relieve the spending minister of his position. For her part, the prime minister generally enjoys a privileged position in any such battles. She can call a vote of confidence on a given issue that puts the very existence of the government at issue if a given minister does not support her position (Huber 1996). Table 3 presents different powers that the finance minister may have in practice. They include the power to block spending in-year, the requirement that the ministry approve all
14

Table 3: Indicators of Finance Minister Strength under Delegation during the Implementation Stage of the Budget
Rule 1. The Finance Minister can unilaterally block spending in-year 2. The Finance Minister must approve all cash disbursements 3. The Finance Minister can impose cash limits 4. The Finance Minister must approve the transfer of funds across budget chapters
This table is based on the presentation in Hallerberg, Strauch, and von Hagen 2001.

cash disbursements as well as the power to limit disbursements when the ministry thinks appropriate, and the requirement that the ministry approve transfers across budget chapters. This discussion also has implications for the parliamentary stage of the budget. Parliaments that can easily change the governments budget proposal undermine the effectiveness of the finance minister. Disgruntled spending ministers can potentially go around the backs of finance ministers and request that parliament pass their ideal budgets. An effective form of delegation therefore requires that parliament also delegate its power in the budget process to the finance minister. The expectation is that parliaments cannot change to any significant degree the governments budget without the governments approval, which, in practice, means the finance ministers approval. If delegation would be equally effective across all forms of government, one could stop here and suggest that all countries strengthen their finance ministers appreciably and weaken the role of parliaments in the budget process.11 Yet it is clear that the effectiveness of this form of governance depends upon the party structure of the government. To see this, return to the model presented in the previous section and consider two parties in government, with Party A controlling the Finance Ministry and some spending ministries while Party B controls some spending ministries. The ideal budget for Party A remains the budget in equation 6; that is, it is a budget that reflects the ideal spending preferences of Party A only. The ideal budget for the coalition, however, is Equation 8, which includes the ideal budgets of Party B for those ministries that Party B controls.12 What this suggests is that the greater the ideological distance between Party A and Party B the less likely Party B will be willing to delegate any power to a finance minister on the budget; in game-theoretic parlance, the principal-agent problem becomes more severe. Coalition members simply will not trust a central player who must inevitably come from only one party to monitor and punish spending ministers in a
11

The Inter-American Bank (1997) suggested as much in their annual report, stating that more hierarchical budget institutions that grant more power and responsibility to the finance minister (vis--vis other ministers)can contain deficit bias and lead to permanent and meaningful improvements in fiscal discipline (p. 143).
12

Laver and Shepsle (1994, 9-10), for instance, in summarizing the findings of the case studies in their edited volume, note that the distribution of portfolios among members of the same political party has little effect on the policies which the government adopts; much more important is the distribution of portfolios among different parties.

15

manner that does not benefit the minister of finances party. In coalitions where the parties will run against each other in future elections, there may also be a zero-sum game played between the parties. Any excessive spending by one party potentially helps that party in the next election and hurts its present coalition partners. In contrast, in one-party governments, or in coalition governments where ideological differences are minor, conflicts of interest among cabinet members arise primarily because of the CPR the problem. The different ministers can be fairly sure that the finance minister holds more or less the same spending preferences as they do, and delegation of power to the finance minister is not problematic. There is a second reason why delegation is most effective in governments with one party governments. The punishment mechanism depends upon the power of the prime minister and, by extension, the finance minister to reprimand and to dismiss intransigent spending ministers. In one-party governments the dismissal of one minister from the same political party as the prime minister can be heavy for the minister but light for the government. If the minister comes from a coalition partner, however, the partner may rally around the minister and force a showdown that can lead to a collapse of the government. Delegation is therefore an option in one-party majority governments, such as in the United Kingdom, where ideal budgets of the players are close and where all cabinet members will be on the same side in the next election. Delegation is also possible in countries where political parties are close to each other ideologically and where they usually run together as blocks in elections. In France, for example, the RPR and the UDF usually coordinate their electoral strategies. In Germany the last 20 years the FDP, CDU, and CSU have in one block opposed the SPD and the Greens. In such countries differences in ideal budgets among block partners are not large. Parties that expect to run together in the future do not have the worry that their partners will intentionally defect to increase their chances of winning the next election. Excessive spending hurts the coalition more generally.

16

d. Commitment
The alternative to delegation that is available to coalition governments is that the players commit to fiscal contracts. Parties that negotiate the budgets for every ministry with each other consider the total tax burden on their coalition rather than the tax burden just on their parties, or, potentially even more damaging, than the burden that individual ministers consider. They therefore internalize the tax externality. Because the parties negotiate the budgets between themselves and set the contract, the benefit that individual ministers get simply from larger budgets for their ministries is also not included in the utility equation. The natural place to negotiate such contracts is during the coalition negotiations. There are therefore two alternative ideal types of coalition agreements. The first type is the fiefdom model, where the parties simply negotiate the distribution of portfolios. The second type is commitment to fiscal contracts, and it involves detailed negotiations for every ministry. The prototypical example comes from the Netherlands. Future coalition partners negotiate detailed budgets for every ministry before the negotiations are concluded.13 This discussion suggests that the planning stage of the budget is important to commitment states. Coalition partners negotiate multi-annual plans, and they expect annual budgets to remain consistent with these plans. In practice, in fact, the plans amount to fiscal contracts, and this analogy is useful here. Questions to ask include what provisions exist to make sure that the parties stick by the contract, what happens if one or more parties break the contract, and under what conditions are contracts renegotiable. First, negotiations among parties that result commitment to a fiscal contract do not solve the CPR problem if the parties and/or ministers can easily violate the agreement ex post. Such violations can take place during either the decision-making phase of the budget or during the implementation phase. If they take place during the decisionmaking phase, one or more parties decide to break the agreement and to demand more spending in one or more areas that matter to it. If the violation occurs during the

13

This case, as well as the development of fiscal contracts in Belgium and in Finland, is discussed in detail in Chapter 5.

17

implementation phase, one or more parties chooses to spend more than it was awarded in both the coalition agreement and in the annual budget. In both cases, it is important for coalition members to be able to detect possible defections. One issue concerns the contract design itself. Detailed provisions make it easier to determine whether a given action violates the letter of the contract. Wide circulation of the provisions of the contract also make it more likely that people outside of government can identify defections. With the government, the minister of finance can again play a role, but that role is necessarily more limited. Unlike in delegation states, she will not have agenda-setting power and other strategic powers ex ante because the CPR problem is addressed through the fiscal contract. Coalition parties have reason to suspect that the minister is biased towards her party. While the minister can still assist in monitoring ministers ex post, the other parties will still want to monitor each other in other forums. The obvious institution that has some staff and potentially some expertise to monitor cabinet members is parliament. This argument suggests an important role that parliament can play in the budget process. One would expect committees to have the ability to monitor government ministries in commitment states. Committees may be designed so that they are responsible for keeping watch over one government ministry. The committee chair may be expected to come from a different party than the Minister, which occurs in practice most of the time in countries like the Netherlands. They committees may have other rights that reinforce their ability to collect information on the ministry in their jurisdiction, such as the right to call a minister before the committee for testimony and to subpoena ministerial documents. The second question to ask is what happens if a party defects. Indeed, the punishment mechanism clearly differentiates commitment from delegation. It is unlikely that a prime minister in a multi-party coalition can easily dismiss individual ministers who violate the contract if they come from a different party. Moreover, this solution is ineffective anyway so long as the root cause of the problem is the party and not the individual minister. The likely punishment is therefore the dissolution of the coalition government itself, but this form of brinksmanship is a blunt, and often unsuccessful, tool. The effectiveness of such threats depends upon the ability of current coalition members
18

to find alternative partners to replace any defectors. This requires some dynamism in party system structure. Coalition theorists tell us about the likelihood of a new coalition forming. In particular, Laver and Shepsle (1996) argue that some parties cannot be excluded from possible coalition because of their electoral strength combined with their position in the middle of a given countrys political spectrum.14 Other coalition partners cannot punish such strong parties in the perspective presented here. Fiscal contracts in the presence of a strong party are unlikely to work. Countries with a history of strong parties are likely to have fiefdom governments.15 The final question concerns renegotiation of the contract. One can imagine situations where one or more party is no longer satisfied with the original contract. If a coalition partner expects to do much better after new elections, the contractand for that matter the coalitionare probably at their end and renegotiation is not possible. Early elections are the result and there is no bargaining among the players. A more interesting case is one where an unforeseen event places a burden on the budget. How are the consequences of the burden distributed among the coalition partners? Grilli et. al 1991 concentrate on the effects of economic downturns, and they argue that all coalition governments (not just commitment governments) face gridlock during economic downturns because they cannot negotiate the distribution of costs (both actual in the budget as well as political) across coalition partners. The result is policy drift and, during an economic downturn, large budget deficits. The Grilli et. al 1991 scenario probably describes well the situation facing fiefdom governments, but it is not satisfactory for explaining the functioning of commitment governments for two reasons. First, parties can write detailed provisions into the contracts to assure that there are in practice very few unforeseen effects. Provisions in the original contract for across-the-board spending cuts during negative shocks or provisions that require that all additional revenue during a positive shock go to reducing the overall debt level (what todays commitment government in Belgium refers to as the
14

In particular, Laver and Shepsle (1996) differentiate between two types of strong parties. A very strong party is one where no majority in parliament prefers an alternative coalition that excludes it. A merely strong party, in contrast, can threaten to veto certain alternative cabinets. See Laver and Shepsle (1996), pp. 69-71. 15 I will return to the point more formally later in the chapter when I consider a multi-period game that allows borrowing from the future.

19

golden hamster rule) take the decision out of the hands of the coalition. Second, governments can make problematic negative shocks less likely by writing contracts based on intentionally conservative estimates of future economic growth and the like. If the expected annual real growth rate of the economy over a four-term is 3% per year the government may write fiscal contracts that estimate a growth rate of only 2.5%. As Hallerberg, et. al 2001 indicate in their survey of fiscal rules in EU countries during the period 1998-2000, commitment states regularly employ both tactics to try to avoid contract renegotiations in the middle of a coalitions term. A summary of institutional characteristics common in commitment states appears in Table 4.

Table 4: Indicators of Finance Minister Strength under Delegation during the Implementation Stage of the Budget
Rule 1. Multi-Annual budget plan written into the coalition agreement as a fiscal contract 2. Rules for what to do in case of unforeseen economic shocks written into the coalition agreement 3. Finance minister is one guarantor of the fiscal contract, but she does not have the same powers as a finance minister under delegation
This table is based on the presentation in Hallerberg, Strauch, and von Hagen 2001.

Mixed Governance
The previous sections introduced three ideal types of governance in parliamentary democraciesfiefdom, delegation, and commitment. They also indicated that the underlying party system structure plays a critical role in determining which structure is likely. One party governments or governments where the parties run in electoral blocks have either fiefdom or delegation as options, while multi-party governments are either fiefdom or commitment governments. Both delegation and commitment assume that governments can pass their budgets through parliament with little difficulty and relatively unchanged.
20

This assumption may be realistic in West European countries with majority governments. It does not describe, however, the passage of legislation in countries with minority governments. By definition minority governments do not have the votes in parliament to pass legislation without the help of the opposition. Since just under a third of all postwar European governments have taken a minority form, this point is not trivial (Tsebelis 1995). Moreover, although minority governments have received less attention than majority governments, there is little consensus in the literature about how such governments affect the making of budgets. Some authors consider them to be highly unstable and incapable of passing significant legislation. Edin and Ohlsson (1991), in their re-analysis of Roubini and Sachs' (1989) dataset, find that minority governments are more likely than any type of majority government to run large budget deficits. In contrast, Strm (1990) insists that such governments are relatively stable and that they do at least as well as majority parties on the effectiveness of their policies. Tsebelis (1995, 1999, 2001) even argues that there is no functional difference between one-party minority and one-party majority governments. For the purposes of this book, which deals with budget policy, minority governments can take one of two forms of governance. Like majority governments they can lapse into fiefdom modes. To the extent that minority governments consider a smaller share of the overall tax burden than majority governments, one can anticipate that the CPR problem may be even worse in minority governments than in majority governments. Yet there is also a way that minority governments can avoid this fate, which in practice is a combination of the ideal types of delegation and commitment. A strong finance minister solution to the common pool problem is possible under minority situations at the cabinet level in one-party governments and in minority coalition governments where the coalitions partners are close to one another ideologically. While a finance minister may coordinate the budget-making process within the government if there is one dominant party in the government as in delegation, this step is not enough. Centralization at the governmental stage can come to naught if the budget unravels in last-minute deals with the opposition that simply buy support from this or that party on key dimensions. As Chapter 7 indicates, in fact, Swedish Finance Minister Feldt learned

21

this lesson in the late 1980s when it proved difficult for the minority Social Democratic government to pass its budget. A more effective method is to include selected opposition parties in budget negotiations early in the process and, as in Denmark after 1982, reach agreement with them on all dimensions of the budget. This type of agreement, of course, resembles very much the agreements reached under commitment. The key difference is that the agreement is not negotiated with coalition partners. The government's negotiations with the opposition are instead functionally equivalent to the coalition negotiations necessary under multi-party majority governments. The government agrees to fiscal contracts with certain opposition parties. In Sweden in 2001, for example, the minority Social Democrats negotiated a spending agreement with the Left Party as well as with the Green Party across 27 different spending categories that, together, constituted the entire budget. Consistent with the data presented in Strm (1990), one would expect parliamentary committees to be especially strong in countries with frequent minority governments. Similar to the commitment case, these committees would monitor the governments implementation of the budget to assure that it keeps its part of the fiscal contract.

Empirical Evaluation, Forms of Governance, and Budget Deficits

This section briefly reviews empirical evidence illustrating differences among the four models. Based upon the case study material presented in the remainder of the book, it first classifies the fifteen current members of the European Union according to these four types of governance for the period 1980-1997. It then presents a regression analysis that provides evidence that the forms of governance have a real effect on budget outcomes, and in particular in levels of fiscal discipline. Table 5 classifies the 15 European Union states according to the different possible forms of fiscal governance for the time period 1980-2000. One of the first things to notice is that the forms of fiscal governance do change. Eleven of fifteen countries have changed their form of fiscal governance at least once. The timing of changes is also relevant. Several states, including Austria, Denmark, Ireland, and the Netherlands moved away from fiefdom before the Maastricht Treaty was agreed to in December 1991. The
22

EMU process did not cause the domestic changes in the budget process in these states. At the same time, there are a number of states that either introduced a new form of fiscal governance, or that strengthened a pre-existing one, in the mid-1990s. One question the case study chapters will ask is whether the changes in the budget process are due to EMU or due to other reasons. Finally, the table reveals that delegation states tend to be more stable. Three large countriesFrance, Germany, and the United Kingdomhave had delegation forms of fiscal governments for the entire period studied here.

Table 5: Categorization of States, 1980-2000

Country Austria Belgium Denmark Finland France Germany Greece Ireland Italy Luxembourg Netherlands Portugal Spain Sweden United Kingdom Form of Fiscal Governance Fiefdom 1980-84, Commitment 1985-91, Fiefdom 1993-95, Commitment 9697, Fiefdom 98-99, Delegation 00 Fiefdom 1980-92, Commitment 1993-2000 Fiefdom 1980-81, Mixed 1982-00 Commitment; Strengthened mid-1990s Delegation Delegation Fiefdom 1980-96, Delegation 1997-2000 Fiefdom 1980-87, Mixed 1988-91, Commitment 1992-00 Fiefdom 1980-95, Delegation 1996-2000 Commitment Fiefdom 1980-82, Commitment 1983-00, strengthened 1994 Fiefdom 1980-00 Fiefdom 1980-93, Mixed 1994-2000, Delegation (2000-) Fiefdom 1980-96, Mixed 1997-2000 Delegation

The next question to ask is, do differences in forms of fiscal governance have a tangible impact on policy outcomes? It would be most desirable to test the formal directly, but it is not possible to measure ideal spending preferences in practice. What we can observe are strategies that individual parties and ministers pursue, but the strategies are not the same thing as preferences. A way around this difficulty is to consider a multi-period game instead of a one-period game that allows government borrowing. As Hallerberg and von Hagen (1999) and Velasco (1999; 2000) indicate, governments with serious common pool resource problems should run larger budget

23

deficits all else equal.16 In terms of this books framework, fiefdom governments should run larger budget deficits than the forms of government that address the CPR problem. Table 6 provides the results from a regression analysis involving the current European Union states for the time period 1980-1997. The dependent variable is the change in the gross debt burden. This measure is more consistent over longer periods of time than changes in the budget balance. States did not necessarily use the same accounting standards to determine their budget balances, but there are few differences across countries in measuring gross debt. There are two groups of independent variables. To test the effects of forms of fiscal governance, I include dummy variables for each form that addresses the common pool resource problem. Fiefdom exists when the dummy variables all equal zero. One might anticipate that a change in government would lead to paralysis (or drift) in budget policy, so the fourth political variable codes changes in government. The fifth political variable measures the percent of cabinet positions that left parties occupy to test for possible partisan effects. Tsebelis (2002) argues that countries with larger ideological distances among veto players are likely to have fixed fiscal policies while parties with smaller ideological distances are likely to have more flexible fiscal policies. While not using the term veto player, Roubini and Sachs (1989) and Grilli et als (1991) contention that coalition governments perform worse than one-party governments would imply that increases in the ideological distance among veto players should lead to increases in the gross debt burden over time. I also include a set of economic variables that one would expect would affect the gross debt burden, including changes in economic growth and unemployment. Some scholars argue that European Union states made only temporary changes to their budgets to get into EMU. This implies that there should be a negative coefficient on year dummy variables for 1996 and 1997. I therefore report the results for those years in the Table. For robustness reasons, I include year dummy variables for the remaining years in the sample except for 1980 (not reported year), a lagged dependent variable to address auto-correlation, and panel-corrected standard errors (Beck and Katz 1995; Beck 2001).

16

If Editor will allow it, provide the mathematical proof in an Appendix that presents the multi-period game and that allows deficits.

24

Table 6: Fiscal Performance of European Union States, 1980-97

Dependent variable: change in the gross debt burden
Coefficient (Standard Error) Political Variables Delegation Commitment Mixed Change in Government Percent of Cabinet Positions Occupied by Left Parties Ideological Distance, Veto Players (Laver-Hunt Coding) Control Variables Intercept Change in Debtt-1 Debtt-1 Growth Change in Unemployment Rate 1996 1997 N R-Squared -1.74** (0.61) -1.34** (0.57) -1.63* (1.00) 0.48 (0.48) -0.72 (0.54) 1.22 (1.47) 2.32** (0.67) 0.46** (0.08) -0.016* (0.008) -0.70** (0.15) 0.47** (0.24) 1.28 (1.54) 2.30 (2.32) 222 .63

The results indicate that the forms of fiscal governance have real effects on budget outcomes. The intercept term indicates that debt levels are expected to grow at little over two percentage of GDP per year in fiefdom countries. Introducing any of the forms of governance to address the common pool resource problem reduces the growth of the debt level almost to zero. This means that swings in the economy in the form of output growth or unemployment are passed along directly to debt burden. An increase in the veto player ideological distance has the correct sign but is not statistically significant. The remaining political variables are also not significant.

25

Conclusion This chapter defined four forms of fiscal governance. Fiefdom exists when ministers set their own budgets and when they do not consider the full tax implications of their decisions. Delegation exists when minister delegate strategic powers to a finance minister to centralize the budget process. The mode of governance relies upon the discretion of the finance minister. Commitment exists when political parties write what amount to fiscal contracts into their coalition agreements. A mixed form of fiscal governance integrates elements of delegation and commitment. This chapter also discussed reasons for why one finds variation in forms of fiscal governance in parliamentary democracies. Delegation is found in countries where there are one-party majority governments. They also exist if the ideological distances among coalition partners are small and if the parties expect to run together in elections. Commitment can address the CPR problem under multi-party coalition governments. Mixed systems are found exclusively under minority governments. Finally, fiefdom is common where party systems are uncompetitive. It is not possible to punish a defecting minister or party if the offender cannot be removed from office. The party system is therefore a crucial variable, and the following case study chapters are organized according to the type of party system. Four countries have systems either where one party consistently wins a majority of seats (Greece, United Kingdom) or where groups of parties with minor ideological differences run in elections as blocks. One would expect delegation forms of fiscal governance. Four further countries (Belgium, Finland, Luxembourg, and the Netherlands) have regular majority coalition governments where partners routinely run against each other, and commitment would be expected to develop in these countries. Denmark and Sweden have minority governments, and they should have mixed forms. Austria, Ireland, Portugal, and Spain do not have stable party systems that yield one form of government (one-party/coalition, minority/majority) in the twenty covered in this study. The forms of fiscal governance, while often taking root, are not as well institutionalized as in countries with more stable party systems. Finally, Italy is a wonderful test case to apply the theory. It changed its electoral system in 1994, which had notable effects on the countrys party system. The prediction about the type of fiscal
26

governance that would be appropriate for the country shifted from commitment to delegation. The chapters together indicate that the competitiveness of the party system is crucialstates remain stuck with fiefdom governments when they have uncompetitive elections. Moreover, the EMU framework can be interpreted as a set of rules as well as a set of requirements to monitor budgetary performance. EMU is consistent with a rulesbased form of fiscal governance, and it has enhanced the fiscal performance of commitment states like Belgium, Finland, and the Netherlands. It is not, however, consistent with a discretion-based form of fiscal governance. It is no surprise that delegation states like France, Germany, and (today) Italy are having difficulties abiding by the provisions of the Stability and Growth Pact.

Appendix Present the multi-period results of the model here.

27

Bibliography Baron, David P, and John A. Ferejohn. 1989. "Bargaining in Legislatures. American Political Science Review 83, no. 4 (December): 1181-1206. Duverger, Maurice. 1954. Political Parties. Their Organization and Activity in the Modern State. New York: Wiley. Edin, Per-Anders, and Henry Ohlsson. 1991. "Political Determinants of Budget Deficits: Coalition Effects Versus Minority Effects." European Economic Review 35: 1597-1603. Hallerberg, Mark. 2000. The Role of Parliamentary Committees in the Budgetary Process within Europe. In Strauch, Rolf, and Jrgen von Hagen, Eds. Institutions, Politics and Fiscal Policy. Kluwer: 87-106. Hallerberg, Mark, Rolf Strauch, and Jrgen von Hagen. 2001. The Use and Effectiveness of Budgetary Rules and Norms in EU Member States. Report Prepared for the Dutch Ministry of Finance, July 2001. Available for download at http://www.minfin.nl . Hallerberg, Mark, and Jrgen von Hagen. 1999. "Electoral Institutions, Cabinet Negotiations, and Budget Deficits within the European Union." in Poterba, James, and Jrgen von Hagen, Eds. Fiscal Institutions and Fiscal Performance. Chicago: University of Chicago Press: 209-232. Hallerberg, Mark, and Jrgen von Hagen. 1998. Electoral Institutions and the Budget Process. In Fukasaku, Kiichiro, and Ricardo Hausmann, Eds., Democracy, Decentralisation and Deficits in Latin America. Paris: Organisation for Economic Cooperation and Development. Hardin, Russell. 1982. Collective Action. Baltimore: The John Hopkins University Press. Huber, John. 1996. The Vote of No Confidence in Parliamentary Democracies. American Political Science Review 90, 2: 269-282. Inter-American Development Bank. 1997. Latin America After a Decade of Reforms. Washington, D.C.: Johns Hopkins University Press. Krehbiel, Keith. 1991. Information and Legislative Organization. Ann Arbor: The University of Michigan Press. Lancaster, Thomas D. 1986. Electoral Structures and Pork Barrel Politics. International Political Science Review 7, 1 (January): 67-81.

28

Laver, Michael, and Kenneth A. Shepsle. 1996. Making and Breaking Governments. Cambridge: Cambridge University Press. Laver, Michael, and Kenneth A. Shepsle. 1994. "Cabinet Ministers and Government Formation in Parliamentary Democracies," in Michael Laver and Kenneth A. Shepsle, eds. Cabinet Ministers and Parliamentary Government. Cambridge: Cambridge University Press: 3-14. Lupia, Arthur, and Mathew D. McCubbins. 1994. "Who Controls? Information and the Structure of Legislative Decision Making." Legislative Studies Quarterly 19, 3: 361-384. Mattson, Ingvar, and Kaare Strm. 1995. Parliamentary Committees. in Herbert Dring, ed. Parliaments and Majority Rule in Western Europe. New York: St. Martins Press: 249-307. Olson, Mancur. 1965. The Logic of Collective Action. Cambridge: Harvard University Press. Ostrom, Elinor. 1990. Governing the Commons. Cambridge: Cambridge University Press. Ostrom, Elinor, Roy Gardner, and James Walker. 1994. Rules, Games, and Common Pool Resources. Ann Arbor: The University of Michigan Press. Roubini, Nouriel, and Jeffrey D. Sachs. 1989. "Political and Economic Determinants of Budget Deficits in the Industrial Democracies." European Economic Review. 33: 903-938. Strm, Kaare. 1990. Minority Governments and Majority Rule. Cambridge: Cambridge University Press. Tsebelis, George. 2001. Veto Players: How Political Institutions Work and Why. Unpublished Book Manuscript, UCLA. Tsebelis, George. 1999. Veto Players and Law Production in Parliamentary Democracie: An Empirical Analysis. American Political Science Review. 93: 591-608. Tsebelis, George. 1995. Decision Making in Political Systems: Veto Players in Presidentialism, Parliamentarism, Multicameralism and Multipartyism. British Journal of Political Science 25 (3): 289-325. Velasco, Andrs. 2000. Debts and Deficits with Fragmented Fiscal Policymaking. Journal of Public Economics. 76 (2000): 105-125.

29

Von Hagen, Jrgen, and Ian Harden. 1995. Budget Processes and Commitment to Fiscal Discipline. European Economic Review. 39: 771-779. Von Hagen, Jrgen, and Ian Harden. 1994. "National Budget Processes and Fiscal Performance." European Economy. 3: 315-418. Von Hagen, Jrgen. 1992. "Budgeting Procedures and Fiscal Performance in the European Communities. Economic Papers 96.

30