In the preceding chapter we made a determined attempt to derive testable empirical implications from our model of government formation and to evaluate these using data on actual government formations in 12 postwar European democracies. We found considerable empirical support for our approach. Many of the implications we derived – in particular those dealing with the cabinet membership of very strong and merely strong parties – give the analyst a massive improvement over chance in making forecasts about the allocation of cabinet portfolios. In this chapter we extend our empirical analysis in three substantive directions, using more powerful statistical techniques.

The substantive extensions of our empirical analysis are concerned with two important features of the strategic position of any legislative party in the government formation process – its size and its position in the policy space. In Chapter 5 we investigated the relationship between size, policy, and strong-party status in large numbers of simulated legislatures. In the present chapter, we investigate this relationship in real legislatures. We have three key objectives in this phase of the analysis. First, we know from Chapter 5 that, in simulated legislatures, strength is strongly related both to the weight of parties and to their position in the policy space. Larger parties, and parties at a median position on at least one policy dimension, are far more likely to be strong or very strong in our simulations. We noted when reporting these simulations that party positions in real party systems are not randomly located; in this chapter we assess the extent to which a party's strength is related to its weight and policy position in the real world.

INTRODUCTION

Thus far we have elaborated our model, set out some of its implications for coalition formation in the real world, explored these implications in a couple of particular settings, and discussed how these implications can be more generally evaluated using data that are actually available to us. We are now in a position to look in a more systematic manner at how well the model seems to capture important features of the making and breaking of real governments. We do this by determining whether each of the cabinets in our dataset conforms to each of the theoretical implications of our model set out in Chapter 7.

Using our data for each government formation situation that we consider, we use Winset to calculate which of the possible cabinets that could have formed were consistent with the theoretical implications of our model. Since, out of very many cabinets that could possibly have formed, our model typically identified rather few as potential equilibriums, if the actual cabinet that formed was one of these equilibrium cabinets identified by our model, we regard this as an empirical “success.”

Furthermore, since every day in politics is one in which the incumbent government must survive in the face of the possibility of being brought down, every day is in a sense a new government formation situation. If the parameters of the situation change in such a way that the incumbent government is no longer equilibrium, then our model implies that the government should change. We thus consider the duration over which the theoretical implications of our model are fulfilled in each case.

As is conventional, we bring this book to a close in the next two chapters by pointing out some directions for future research. The usual way to do this is to throw out a few general ideas in a relatively haphazard manner, some tough and not necessarily very tasty bones for graduate students to get their teeth into. However, we regard one of the great virtues of our approach – indeed its greatest virtue – to be the rich and varied menu of potential applications that it opens up for future research. We do not see these as the boring bits we could not be bothered to chew over ourselves. We see them as the exciting rewards that can be reaped as a result of developing an explicit and we hope realistic model of the government formation process. Accordingly, in this chapter we limit ourselves to three of the most promising potential applications of our approach, and take time to develop them in some detail.

The first application that we consider concerns the interaction of intraparty politics and government formation. We consider the general area of intraparty politics to be one of the most exciting and underdeveloped in the entire literature, since it generates the potential to provide some motivation for the actions of political parties, hitherto unrealistically seen by most theorists as anthropomorphic unitary actors. The second application concerns minority and surplus majority governments. The third and final application of this chapter focuses on administrative reform. Here we raise the prospect of endogenizing the assignment of issues to ministerial portfolios, thereby affecting the very way in which governments go about their business once installed in power.

One of the key conclusions of the previous chapter – and indeed of our entire analysis – had to do with the role of strong parties in the making and breaking of governments. In a nutshell, if a party is very strong, then the cabinet giving that party all portfolios is preferred by some legislative majority to any alternative cabinet. A party may also be merely strong, however, if it participates in every cabinet that is preferred by a legislative majority to the cabinet in which it controls all portfolios. In the previous chapter we show that there can be at most one strong party. Our central result, stated in Proposition 4.2, is that a strong party can dominate the business of government formation and guarantee itself a place at the cabinet table. This result has fundamental implications, both for the analysis of particular cases and for our general understanding of the making and breaking of governments in parliamentary democracies.

If we want to analyze a particular government formation situation using our model, one of the first things we need to know is which party (if any) is the strong party. If we want to understand the process of government formation more generally, then we need to know about the conditions under which a configuration of legislative parties generates a strong party. Thus in this chapter we take a closer look at strong parties and the party configurations that sustain them.

We do this in several ways. First, we briefly discuss how to squeeze as much juice as we can out of our model by focusing on necessary and sufficient conditions for the existence of a strong party.

The Cambridge series on the Political Economy of Institutions and Decisions is built around attempts to answer two central questions: How do institutions evolve in response to individual incentives, strategies, and choices, and how do institutions affect the performance of political and economic systems? The scope of the series is comparative and historical rather than international or specifically American, and the focus is positive rather than normative.

Laver and Shepsle's theoretically innovative book pushes the study of government formation a big step forward. Rather than concentrating on the equilibrium collective policy that emerges from the balance and distribution of party strengths in the parliament, they treat governments as collections of ministers with individual jurisdictions and policy as a bundle of individual party-preferred policies, depending on which party receives which portfolio. Jurisdiction is the key concept, in their view. It makes a party's promises credible to its electors, for it can carry them out (only) if it receives the relevant ministry. At the same time portfolio allocation monitors party behavior, for the party receiving a portfolio has no excuse for not carrying out its preferred policy.

Their portfolio-based model of government formation is based on constitutional features carefully documented in a massive 14-country study, published in the companion volume Cabinet Ministers and Parliamentary Government. The model bases its predictions on a sequence of proposals from an historically determined status quo. Bargaining takes place among rationally foresighted policy-motivated parties, each with an explicit veto over every cabinet in which it could participate, in a lattice of feasible governments. Laver and Shepsle's predicted importance for “strong” parties opens up significant possibilities for comparative empirical evaluations of theirs and other contending models.

Government Formation and Parliamentary Democracy

This book is about the making and breaking of governments in parliamentary democracies. The essence of parliamentary democracy is the accountability of the government (or cabinet or executive or administration) to the legislature. In most working parliamentary democracies, this relationship is enshrined in a constitutional provision that the government must retain the support of a majority of legislators, tested in a legislative motion of confidence (or no confidence) in the executive. A government that loses such a vote is defeated and obliged to resign. Although formal mechanisms for installing new governments differ considerably between countries, a government, however installed, is immediately exposed in the legislature to a potential vote of no confidence. It is the ability of the executive to win this vote and thus maintain the confidence of the legislature that is the universal acid test of government viability in a parliamentary democracy.

Even though the executive is responsible in this sense to the legislature, the cabinet and its ministers nevertheless retain wide-ranging power to govern the country as long as they remain in office. This power comprises both effective political control over the administrative departments of state and a firm grip on the day-to-day activities of the legislature. Apart from the opposition's ability to seek dismissal of an administration by proposing a motion of no confidence, it is difficult in most parliamentary democracies for anyone outside the executive to have a significant impact on the process of legislation.

This chapter and the next use the transactions approach to explain the institutional choices the enacting legislature makes when it turns to tax-financed bureaus to supply goods and services or distribute resources. This chapter focuses on the budget. It examines the nature of expenditure control and why so much expenditure is mandated and therefore cannot be changed without a change in the law. Chapter 5 focuses on the employment arrangements in bureaus. It examines the decline of patronage and the emergence and persistence of civil service rules that effectively constrain legislative influence over bureaucrats. It also examines the characteristic features of this merit-based system and explains those features in terms of the approach described in Chapter 2.

Tax-financed bureau production and distribution has a number of distinctive characteristics that help shape the transaction problems legislators face and the potential institutional solutions available. The taxpayers who fund the bureau are a very large group with a relatively small per capita stake in the operation of any particular bureau. The beneficiaries of bureau activity are typically in a similar position. The whole citizenry benefits from the provision of public goods like justice, foreign relations, and defense, from the provision of policy advice and from the maintenance of the social security system. Because private per capita costs and benefits of this production are often very small, there is less incentive for private interests to monitor provision or to participate in agency decision making.

Considerable academic and popular criticism is directed at the way public administration is organized and how it functions. Many claim that the public sector lacks the incentives for effective performance and that there is a disturbing lack of accountability to the elected representatives whom administrators are supposed to serve. Given these common perceptions, it is a real puzzle to find that the administrative arrangements that shape administrators' incentives and determine their accountabilities are so common and so persistent. It is even more of a puzzle given that the lack of any neat separation between policy and administration means that administrative decisions influence policy outcomes. If these arrangements are so bad, why do we find them in so many jurisdictions and why have they persisted for so long? This book develops a theory to explain key institutional characteristics of the modern administrative machinery of government. It is part of a growing appreciation of the role of institutions in both political science and economics.

Just as the private sector includes corporations, partnerships, and non-profit organizations, the public sector is made up of different forms of organization, each with its distinctive characteristics. In the regulatory arena, laws can be administered by the courts, independent regulatory commissions, or executive agencies. When it comes to the production of goods and services, legislators typically turn to tax-funded bureaus with personnel policies determined by civil service legislation. Sometimes, however, they use state-owned enterprises (SOEs), which are primarily funded by sales revenues and organized along more commercial lines.

The previous chapter sought to explain why legislators preferred public to private enterprise in certain, well-defined situations. Public enterprise is not, however, the only alternative. Legislators could have formed a government bureau to provide electricity, gas, rail and air transportation, or telephone and postal services. Although some countries have used bureaus to provide these services during some periods of their history, they are typically provided by state-owned enterprises (SOEs). A complete explanation for the distribution of SOEs must answer this question: If the government does want to own the producer of these goods and services, why does it choose the SOE form of organization rather than the bureau form?

DISTINGUISHING PUBLIC ENTERPRISE FROM BUREAUS

Aharoni suggests that SOEs have three distinguishing characteristics: “First… they must be owned by government. Second,… [they] must be engaged in the production of goods and services for sale… Third, sales revenues of SOEs should bear some relationship to cost” (1986, p. 6). By comparison, Niskanen defines bureaus as “nonprofit organizations which are financed, at least in part, by a periodic appropriation or grant” (1971, p. 15). Nonprofit organizations are defined as those in which neither managers nor owners can appropriate the difference between costs and revenues as personal income. A public bureau is one that is owned by government and is primarily tax-financed.

The Cambridge series on the Political Economy of Institutions and Decisions is built around attempts to answer two central questions: How do institutions evolve in response to individual incentives, strategies, and choices, and How do institutions affect the performance of political and economic systems? The scope of the series is comparative and historical rather than international or specifically American, and the focus is positive rather than normative.

Murray Horn has combined his academic insight and practical experience to fashion a major advance in transactions cost theories of institutional design. He starts with a frequently described situation in which issue complexity and the consequent desire to take advantage of specialization lead representative assemblies to rely on the opinions and actions of bureaucratic agents. Since the exchanges between legislators and their constituents are typically not simultaneous (flows of benefits to legislators are often more immediate than to constituents), he argues that legislators when creating agency missions need to add durability to their deals with forward-looking constituents by protecting the agency against subsequent legislative modification, while protecting themselves against agency actions detrimental to legislative interests.

Horn theorizes that enacting coalitions in this situation design agencies by solving an instrument assignment problem to minimize the transaction costs they face.