Elster (2000) on institutional solutions for procrastination-prone political institutions

Elster does some interesting speculating in Ulysses Unbound (p. 143) about how delay in the implementation of rules may help political institutions to overcome time inconsistency. In this post I flesh out and extend his ideas a bit.

Suppose that a legislature’s preferences were such that it wanted balanced budgets in the future but not now. From the perspective of year 0, the legislature would like to be able to run a deficit in year 0 and year 1 but not after that; from the perspective of year 1, however, it would like to run a deficit in year 1 and year 2 but not after that. Like Augustine, the legislature wants chastity — but not yet. (One reason for this tendency to procrastinate might be that legislators benefit while in office from kickbacks on government contracts, and the median legislator expects to be in office for two more years.)

The solution Elster suggests (based apparently on work by Tabellini and Alesina from 1994) is to have a rule-making process by which any balanced-budget rule — or modification to that rule — can only take effect after two years. Under that process, the current legislature would pass a balanced budget rule, knowing that it will not constrain the legislature in the near future; a year later, the legislature would want to pass a new law delaying the original law by one year (because legislators anticipate the rule to hurt them when it goes into effect in the next year), but given that any such change would take effect after _two_ years, the legislature chooses to stick with the balanced budget rule.

Elster mentions two kinds of institutional delay that might make this possible: one is delay required to pass an amendment after it is first proposed (delay between proposal and adoption), which is a feature of a number of constitutions (Elster mentions Norway, Sweden, France, Bulgaria, and Finland); the other is delay to implement an amendment once passed. He does not make very clear how these interact, but after thinking about it for a few minutes here’s how it seems to me. In the case of a legislature that can’t get around to balancing its budget (like the one above), two years of the second kind of delay (implementation delay) is necessary to induce the legislature to pass the balanced budget rule; two years of any combination of the first and second kinds of delay (adoption and implementation delay) is sufficient to prevent the legislature from revoking the rule. For example, if the constitution requires two time periods between proposal and passage, with no delay for implementation, then the legislature would never propose an amendment revoking the balanced budget rule, knowing that by doing so it would only be making deficits possible in the future, when it actually wants a balanced budget. But without at least two years of delay for implementation, the rule would not be passed in the first place: without implementation delay, the legislature would choose not to pass the balanced budget rule no matter when it finally comes up to a vote, no matter how long the adoption delay. The most straightforward solution is thus to have two years of delay between passage of any rule (or amendment) and implementation.