Looking for an ideal legislative design
Exploring contributions to social choice theory and new institutional economics that inform legislative design
Ever since the debate between James Madison and Alexander Hamilton on how a Constitution needs to be structured and designed, many nation-States and democracies have struggled with drafting their Constitutions. We may recall that Madison and Hamilton took two different views which were highlighted in The Federalist Papers back in 1787. Madison preferred a strong Federal Government, with checks and imbalances. He had argued in Federalist No.10 that factions (either majority or minority) may derail the union. He, therefore, proposed proportional representation from different districts and more participation from the citizens. Hamilton, on the other hand, was not worried about factions. He also wanted a strong Federal Government but proposed that the US adopt a British style monarchy with a representative Government. This led to the first US Constitution which was essentially a compromise between the preferences of various founding fathers.
New Institutional Economics and Legislative Design
We will try to answer the following question: Do New Institutional Economics (NIE) (particularly its subset of New Economics of Organisation) and the closely related field of social choice theory, shed any light on what an ideal legislative design should be? In this context, we will also discuss important contributions in social choice theory, particularly on cyclicity of majority rule, possible solutions and implications for legislative design. We will cover the various theoretical contributions, particularly in Social choice theory in Part I and explore the 'way out' in Part II.
Recall that NIE relaxed the various assumptions of neo-classical economics viz., free entry and exit of firms, perfect information, zero transactions costs, etc. Various proponents in the literature of NIE had basically stated that hierarchy would be more efficient than the market mechanism to coordinate production within the firm (Coase, Williamson, Alchian-Demsetz).
While the legislature is nothing like a market, insights from the Economics of Organisation such as contracting, gains from trade and transaction costs, principal-agent problems and information asymmetry can be useful in studying legislative design.
As Barry Weingast and William Marshall have argued (1988), the US Congress is ideally designed to capture gains from trade. In a simplified model, they begin with three assumptions: Congressmen represent interests within their constituencies; political parties place no constraints on individual Congressmen; and proposed bills must be passed by a majority of the entire Congress.
Weingast and Marshall suggest that legislators must cooperate in order to pursue their re-election by attempting to provide benefits to the constituents. They suggest that legislators may "trade" their votes with each other to maximise benefits to their constituents as well as increasing the chance of their re-election. For example, if one group of legislators wants dams and bridges in their constituency and the other group wants a new regulatory agency to be set up in their area, both groups can exchange votes to benefit each other. This is also referred as Pork-Barrel politics.
It is necessary to ensure that both groups do not renege on their promises. Weingast and Marshall suggested that the Legislative Committee system overcomes this reneging problem as well as lowers the transaction cost of cooperation between various committees (which are manned by legislators). Hence the Committee ensures that ex-post reneging is not possible by invoking various rules such as agenda control – ensuring that all committees benefit from such an arrangement. The committee system is superior to a 'market exchange type mechanism' where legislators could only exchange votes. In this system, legislators exchange well-defined policies in various jurisdictions.
Majority Rule: Cyclicity and Instability
We may recall that the instability of majority rule was first demonstrated by Condorcet in 1785 and came to be known as the voting paradox. This can be illustrated as follows:
We can see from the above table, where three voters have to choose between three candidates A, B and C. If A is chosen as the winner, one can argue that C should win instead since Voter 2 and Voter 3 prefer C to A by a margin of 2 to 1. Similarly, A is preferred over B and B over C. We thus have a situation where voters' preferences have not given a clear winner even though each individual voter has a clear rank ordering of preference over a pair of candidates. In other words, rational individual preferences have led to a cyclical irrational collective preference.
Many economists and political scientists have worked on this paradox. Most notably this was rediscovered by Duncan Black who modified the Condorcet method, whereby each voter assigns points to alternatives, which are arranged in a single dimension, in order of his preference. Further, each voter's preference has a single peak (or a single maximum point). In such a situation, Black found that majority cycles would not occur. Instead, majority rule would yield an equilibrium at the ideal point of the Median Voter. This also came to be referred to as the 'median voter theorem'. However, Black realised that if the number of policy dimensions was increased from one to many, we would be back in the world of cyclicity and unstable majority rule.
Similarly, Anthony Downs borrowed from spatial theory whereby voters will choose parties on the 'left-right' axis of economic ideology. His basic finding was that if the distribution of ideologies in society remains constant, the system will move towards an equilibrium where the number of parties and their ideologies will be the same over time. This will typically happen when the distribution of voters resembles a bell curve with a single peak. Downs also postulated that a voter would weigh expected benefits and costs of voting. The benefit for a voter would arise only if his vote were pivotal (or the decider). Since this has a very low probability, Downs suggested that almost no one will vote.
A number of other works in social choice theory invoked the cyclicity of the voting paradox. Perhaps the most famous is that of Kenneth Arrow, who proposed his 'Impossibility Theorem'. According to this theorem, a clear rank ordering of preferences cannot be determined which meets certain mandatory principles of fair voting procedures. These principles include: non-dictatorship; Pareto efficiency (meaning that if every voter prefers candidate A over candidate B, candidate A should win); independence of irrelevant alternatives (if the candidate A is preferred over B, then A would still be ahead of B even if a third candidate C is removed from the competition); unrestricted domain (meaning that all individual preference must be accounted for); and social ordering (each individual should be able to rank order his preference in any way and even indicate ties).
In short, it states that if the above five conditions are met, it is impossible to formulate a social ordering where individuals have to rank preferences over three or more options.
The Mckelvey chaos theorem (later extended by Schofield) attempted to rescue majority rule from cyclicity and instability. He found that under perfect information and rational foresight, there is no policy combination that cannot be beaten by another similar combination. This has an implication for decision making in political institutions. Anyone who controls the agenda can guide the voting trajectory to an outcome desired by the agenda setter. In other words, the majority vote can lead to a stable outcome under certain conditions. This is compatible with the findings of Krehbiel, discussed below, which underlined the gatekeeping powers of Legislative Committees.
William Riker, in his book Liberalism against Populism, also analysed cyclical majorities and applied them to real political situations. Extending Down's analysis, William Riker and Peter Ordeshook argue that voting is, after all, a rational act. The voter will consider the benefits and costs of voting and since benefits are rather small, it is the 'duty' element that propels the voting decision. This was an extension of Downs "Paradox of Voting".
Another influential work having implications for the legislative design was the book, The Calculus of Consent by Buchanan and Tullock (1962). They laid out two levels of collective decision making – 'ordinary politics' (e.g. decision made by majority voting in legislatures) and 'constitutional politics' (decision on the framework of rules within which ordinary politics unfolds). They suggested that majority rule could operate over decisions made through 'ordinary politics'. However, unanimity should be the basis when the more critical 'constitutional politics' are considered.
Distributive and informational perspective
Keith Krehbiel, the political theorist from Stanford, has proposed two perspectives of legislative organisation viz., distributive and informational perspective. The distributive perspective pertains to the question of "who gets what and at whose expense". The main characteristic of the distributive perspective is that there is no difference between policies and outcome, these are one and the same. The studies referred to above by Weingast, Marshall and Shepsle belong to the distributive perspective. In these studies, the collective action problem facing the legislature is how to ensure cooperation among legislators that lead to stable outcomes. The solution to the problem as discussed above was institutional, a set of rules which facilitates gains from trade by bargaining on policy jurisdiction. In other words, the Committee system in the US Congress is one such set of rules.
The informational perspective, on the other hand, refers to gains from specialisation in addition to gains from trade. Legislators need to become policy experts and specialise themselves rather than rely on the executive branch. In this approach, policy expertise is good and the solution is again institutional. However, here the solution is to provide resources to legislators to become specialists and accordingly assign them to appropriate committees. Another important aspect of the informational approach is that legislators are initially uncertain about the relationship between policies and outcomes (which can only be forecast). On the other hand, in the distributive approach, policy and outcomes are one and the same. Hence, the key concepts in Informational perspective are:
Incomplete information with legislators which determines the choice of their policies and outcomes.
Asymmetric Information where some legislators have more and better knowledge than others about the relationship between policies and outcomes.
Krehbiel characterises the legislature and committees as players in the above scenario of incomplete and asymmetric information. Committees and the legislature are involved in Legislative Signalling Games (LSG) where the committee sends a signal and legislature receives it. The interaction between the committees and legislature ultimately determines the number and nature of the bills that will be passed, which in turn will decide the equilibrium of the game.
...to be continued
The writer is an IAS officer, working as Principal Resident Commissioner, Government of West Bengal. Views expressed are strictly personal