Many thanks to Alex for introducing me yesterday. Having written several papers on term limits, I will use my first post of the week to raise a new question that has emerged from this aging policy intervention: Why does government spending increase under term limits?
Back in the 1990s, when about half the states’ voters slapped term limits on their state legislators, the idea was to rein in government spending and decrease the growth of government. Instead, spending per capita increased in those states relative to states without term limits. See this empirical paper, this survey article, or this book this book for details.
These results are counterintuitive insofar as we put stock in the intended mechanism, which was simple: As legislators spend more time in office, they tend to vote for more government spending – so if legislators are required by law to spend less time in office, they’ll spend less money.
There are two problems with this. First, the premise that tenure and spending positively correlate has not held up to empirical scrutiny. Most papers found no positive link between tenure and spending, although a few reported small effects.
Besides, even if there were a strong tenure-spending correlation without term limits, that correlation is not likely to hold up once term limits are imposed. This is due to a version of the Lucas Critique (or Goodhart’s Law), which in general argues that observed behavioral patterns are not invariant to policy interventions. In this case, term limits will change the dynamics between voters and politicians in ways that lead to greater spending. More specifically, three explanations seem plausible.
- Term limitation exacerbates fiscal commons problems within the legislature. Because term limits decrease the variance of tenure within a legislature, the relative power of party leaders and ranking committee members will decrease. As the distribution of power flattens, this increases the proportion of legislators who possess access rights to budget items, thus decreasing the control rights that a relatively few leaders and committee chairs would otherwise have. When everyone can get their pet project through, more projects get through.
- Term limits shorten legislators’ time horizons. If legislators use their time in office to advance their careers, and if the career-value of being in the statehouse increases with the support of more spending, then term limits can impart an incentive to spend more and sooner. For example, rank-and-file legislators support more spending to secure leadership positions, and leaders let more projects through in order to quickly build durable coalitions.
- Term limits might lure legislators into very wasteful forms of pork spending, according to this paper by Michael Herron and Kenneth Shotts:
Term limits can, in some cases, inhibit voters from selecting representatives who deliver particularistic benefits, and, in these cases, term limits reduce pork spending. On the other hand, when pork is extremely socially inefficient, representatives who want to deliver pork to their districts have incentives to refrain from doing so to reduce future pork in other districts. In this scenario, term limits actually prevent legislators from promoting future spending moderation and thus paradoxically increase pork spending.
These explanations can, of course, be mutually inclusive. I suspect there is more to #1 and #2, if only because they are more salient.
In general, term limits increase spending because voters and legislators rationally respond to changes in their institutional environment. As this question invites further study, good papers will unpack the specific mechanisms that drive those responses.
— Notes: Since most people seem surprised by the actual effects of term limits, here are pointers to similar findings: Gubernatorial term limits worsen fiscal volatility — this paper (co-authored by my co-author Pete Calcagno) and this paper (by my dissertation advisor Bob Tollison) — because governors invest less in reputation (this paper). States with legislative term limits might also have worse bond ratings (here). Here on MR, neither Alex nor Tyler have put much stock in term limits, though Alex is less skeptical.