Last week I discussed why bicameralism is not necessarily a status-quo preserving institution, at least in the sense that conventional wisdom suggests bicameral legislatures produce less legislation relative to analogously situated (however defined) unicameral legislatures.
Commentators often ignore that in “strong” bicameral systems, as exist in the U.S., “second” legislative chambers can initiate legislation itself as well as kill legislation approved by the other chamber. Depending on how much legislation each chamber initiates, and on cross-chamber kill rates, it’s entirely possible that a bicameral legislature will enact more legislation than a similarly-situated unicameral legislature.
To be sure, it is a bit of a bait-and-switch to purport to consider the impact of veto players on legislative production and then initially discuss an institution that can initiate legislation as well as stop legislation. So let’s now face the original question fairly: What about institutions that can only veto legislation without also having the power to initiate legislation? Think of judicial review. Or perhaps the executive veto. And, of course, in some countries second legislative chambers can only stop or delay legislation, they do not have the power formally to initiate legislation.
Even in these cases the impact of the addition of a veto player – a player with only the power to stop legislation and not to initiate it – is indeterminate on the amount of legislation that is enacted and implemented.
How can this be?
Static analysis of separation-of-power systems has led many commentators to conclude the addition of a veto player, such as a judge with the power of judicial review, necessarily reduces the amount of legislation implemented. As rehearsed last week, the argument seems obvious: If in a system of legislative supremacy the legislature enacts N laws, then adding a veto player who can only veto laws, say n ≥ 0, means the overall number of laws implemented when a veto player is added to the system must be (weakly) fewer than the number implemented by the legislature without added the veto player, N – n ≤ N.
Simple and straightforward. But also misleading. The static nature of the traditional analysis leads us to the misleading result. In a dynamic conception of separation-of-power systems, adding a veto player who acts subsequently to the legislature can nonetheless change the incentives legislators face earlier in time when producing legislation. Expectations of future outcomes can affect what legislators (indeed, what any of us) do today.
To make it more concrete, recall from last week, drawing on James Bradley Thayer’s argument, I sketched how non-deferential judicial review might induce legislative moral hazard. Because judicial review, as it were, can “insure” against bad legislative outcomes, judicial review can induce legislatures to legislate more recklessly than they would without the institution.
But what does legislative “recklessness” mean here? Start with the more-standard case of moral hazard resulting from insurance for auto accidents. Because drivers are able to insure against paying the full cost of an accident, they drive more miles and drive with less care than they do without auto insurance.
So consider how judicial “insurance” against bad legislation might affect legislative incentives: By insuring against (some of) the consequences of bad legislation, non-deferential judicial can increase overall legislative production (“more miles”) and decrease the average quality of legislation (“less care”).
Importantly, this is not an argument that the addition of judicial review (or a veto player of any sort) necessarily increases legislative production, at least in the sense of increasing the overall number of laws ultimately implemented. Whether legislative production increases or decreases this number depends on the particular forms of a legislature’s “production function,” how the addition of the veto player changes legislative incentives to produce legislation, and the number of laws the judiciary actually vetoes.
Nonetheless, when we consider separation-of-power systems as dynamic rather than static, the impact of the added veto players on the number of implemented statutes is indeterminate. That is, there exists what seems to be a reasonable possibility that separation-of-power systems can be more statutorily disruptive of the status quo than systems of legislative supremacy.