*Adithya Raajkumar reviews Paul Bochtler’s work on legislative scrutiny in presidential systems. *

Legislative scrutiny refers to the process of debate, investigation and query that bills undergo in the legislature. While it is a well-documented feature of parliamentary systems, the concept has seldom been applied to legislatures in presidential systems. Paul Bochtler of Freieuniversität Berlin in “Legislative Scrutiny in Presidential Systems” therefore seeks to investigate the mechanisms by which bills in the presidential systems are subjected to legislative scrutiny. Choosing to specifically focus on coalition governments, Bochtler focuses on the Congreso Nacional de Argentina as a case study.

Bochtler starts by reviewing the theories of legislative scrutiny as applied to coalition governance in presidential systems. In governments composed of more than one party, each party has diverging interests regarding bills, and so any potential bill will generally be a compromise of these interests. This compromise must be checked and enforced through subjecting bills to scrutiny; if it is clear that a bill is violating the interests of one party too strongly, it is unlikely that it will make it out of the chamber. By using the time it takes for bills to make it out of Congress as proxy for the degree of legislative scrutiny to which they are subjected, Bochtler is able to hypothesize about various factors impacting the degree of scrutiny. For example, he predicts that bills which are “more divisive”, or which end up in committees chaired by the opposition, will take more time to make it from one stage of the legislative process to another.

Since legislative scrutiny is proxied by time spent in the legislative process, Bochtler uses survival analysis to model the expected time until a bill passes from one stage of the process to another. Thus, by using data on which bills made it to each stage and how long they took to do so, Bochter is able to estimate how long a particular bill will take to move on to the next stage given various characteristics of the bill. More technically, he uses statistical methods to estimate the bill’s “hazard rate function”, which models the expected time *until* a bill moves on. The “hazard rate”, derived from this unction, gives the probability that a bill will take a particular amount of time before moving on. In other words, for a given length of time, the hazard function returns the probability that the bill will take exactly this amount of time to move from one stage of the legislative process to the next (for example, a specific hazard function may say that for the average bill, the probability that it will take three months to be passed by Congress before being signed into law is 39%; the probability of 39% is the hazard rate). The hazard function is assumed to be determined by factors such as the ones Bochtler hypothesizes about. In this way, the amount of legislative scrutiny to which a bill is subjected can be estimated from its characteristics, which determine how long it spends in each stage of the legislative process.

Specifically, Bochtler proposes a Cox proportional hazards model, in which the covariates (factors) affect the *proportional* hazard rate at which these legislative events occur. For example, being passed in a special session may be expected to halve the time it takes for a bill to be passed, or doubling the number of committees to which a bill is sent for debate may triple it. Thus, for the first scenario, the hazard rate would be doubled, whereas in the second it would be cut in three. Mathematically, the covariates are used to estimate the parameters of an exponential distribution modeling how long the bill takes to pass throughs the stages of the legislative process in the Argentine Congress. In other words, the covariates are used to estimate the hazard function of these bills, which can then be used to estimate the time they spend in each stage.

While technically appropriate, Bochtler’s analysis is complicated by abstruse specification of the independent variables. At times, it is unclear how he measures certain variables or why he chooses a particular method. Moreover, it is unclear whether his results are robust to alternative specification of these variables. For example, it is not clear how he measures a bill’s divisiveness or how measuring it in a different manner would affect the results. On a larger scale, Bochtler’s results are complicated by the fact that he only considers bills which *passed*, whereas legislative scrutiny also necessitates that many bills will *fail* in the chamber. It is unclear why he chose to exclude failed bills from consideration, which would not only be an interesting alternative but perhaps useful as well. Finally, Bochtler chooses to further manipulate his results by *varying* the hazard rate over the course of the legislative process, which may not be entirely appropriate from a statistical standpoint.

Nevertheless, his work constitutes an important preliminary investigation into the behaviour of coalition parties in presidential systems, and makes important contributions to the understanding of legislative scrutiny in these systems. In this sense, it is an excellent springboard for further research into the matter.