You Tell Me It’s the Institution … To Cut Spending, Focus on the Rules of the Game, Not the Players

ABSENT SIGNIFICANT POLICY CHANGE, America’s governments at all levels—and the people who depend on them—face a bleak future. Spending growth threatens to push both federal and state debt-to-gross domestic product ratios past 90 percent in a matter of years. That is the level at which the largest and most comprehensive studies suggest debt begins to dramatically reduce economic growth.

It might seem that the natural solution is to elect politicians committed to reining in spending, especially on the entitlement programs and pensions at the heart of state and federal overspending. The problem, however, is that even fiscally conservative politicians face significant perverse incentives to spend beyond their constituents’ means. And even if they do manage to trim the budget, today’s cuts can be reversed by tomorrow’s leaders.

Luckily, there is hope. Political incentives are shaped, in part, by institutions, i.e., the rules that govern budgeting, electioneering, and legislating. These rules influence the decisions of legislators, governors, presidents, bureaucrats, voters, and even lobbyists. So if we can improve the institutions, we can enduringly diminish the incentive to overspend.

Major institutional reform does, of course, pose a risk. If done poorly, it risks institutionalizing bad incentives rather than good ones. But the institutional reformer need not fly blind. Much can be learned from the U.S. states. Over the years, each state has experimented with different institutional designs and scores of researchers have spent the past quarter century studying these institutions and their effects on spending. Those studies have made the best of cutting-edge econometric techniques to control carefully for other factors and, to the best of their ability, sort out cause and effect. My colleague Nicholas Tuszynski and I recently reviewed dozens of these studies and found that a number of institutions offer hope to those who wish to put government on a more sustainable course.

Consider, for example, one of the most studied institutions: a strict balanced-budget requirement. While nearly every state has a balanced-budget requirement, the stringency of these requirements varies widely. Studies find that, other factors being equal, per-capita spending in states with stricter requirements is about $180 less compared with other states. Studies also find that tighter requirements tend to yield better-funded “rainy day” funds and larger surpluses. States with stricter requirements are also less likely to suffer from a political business cycle in which spending spikes just prior to an election only to be cut back shortly thereafter.

Another well-studied institution is the supermajority requirement for tax increases. Fifteen states currently require a supermajority vote to raise taxes. Studies find that, after controlling for other factors, states with this requirement tend to spend between $100 and $150 less per capita.

Or consider the line-item veto. Researchers find that, in cases where the legislature and the executive are controlled by different parties, this institution reduces per capita spending by about $100. Better yet, consider a special variety of line-item veto known as the “item-reduction veto.” In states with this institution, the governor need not completely defund items, but instead may write in a lower amount. This changes the negotiating power between the governor and the legislature by effectively denying the legislature the ability to make the governor a take-it-or-leave-it offer. One study finds that, other factors being equal, this institution can limit per capita spending by as much as $471.

Committee structure seems to have a profound effect on spending. Theoretically, when only one legislative committee has jurisdiction over spending, it should be relatively easy for voters to assign responsibility for excessive spending. But when multiple committees have jurisdiction, no one has the incentive to restrain its appetite, and the entire structure is biased to overspend. The evidence seems to support this hypothesis. One study of state spending in the 1980s finds that on a per capita basis, those states with multiple spending committees tend to spend about $200 more than those with just one.

The same study also examines the effect of separate taxing and spending committees. In this case, if one committee has jurisdiction over taxation but not spending, its members—unable to lavish spending on their constituents—have an interest in constraining the budget. They should, therefore, be a check on the interests of the spending committees. But when one committee has jurisdiction over both spending and taxing, its members will be less interested in reining in the budget. Indeed, it appears that states with unified spending and taxing committees spend over $1,000 more per capita compared with states that separate the two functions.

Some institutions affect spending in counterintuitive ways. For example, while many fiscal conservatives champion biennial budgeting, studies suggest that states with biennial budgets actually spend about $120 more per capita than states with annual budgets. Similarly, in some states, if legislators fail to reach an agreement on the budget in time, the government automatically shuts down. Though one might think this would lead to less spending, the very threat of a government shutdown seems to be enough to keep it open. Moreover, it seems to tilt the balance of political power toward those who favor more spending not less: in states with automatic shutdown procedures, per capita spending is about $80 more.

A number of states have experimented with formal rules that are specifically designed to arrest excessive government growth. These so-called “tax and expenditure limits” (TELs) bind state spending, taxation, or both through formulas. These formulas include factors such as the personal income of state residents, population growth, the inflation rate, or some combination thereof. The evidence here is mixed. Some studies find that in high-income states the most popular variety of TEL—tying spending growth to growth in residents’ income—actually seems to be associated with more spending. Other studies find that TELs can be effective restraints on spending, but only if properly designed. It helps, for example, if the formula references inflation and population growth rather than resident income.

As the last example illustrates, not all of the research on government budgetary institutions is clear. Some topics such as term limits and direct democracy are still hotly debated by academics. And in many cases, subsequent studies of a particular institution have yielded more nuanced results. For example, a number of the institutions we reviewed are only effective in certain circumstances or when they are designed in a certain way (such as TELs). It is likely that we have more to learn about some institutions, such as the item-reduction veto, that have received comparatively less academic attention. But in other cases, such as the effect of strict balanced-budget requirements or supermajority requirements for tax increases, the literature is relatively clear: These institutions save money.

Justice Brandeis famously described the states as laboratories of democracy. Policy makers interested in arresting the unsustainable growth of government can learn much from the results of these experiments.

Mr. Mitchell is a senior research fellow at the Mercatus Center at George Mason University. This article is based on research he conducted with Nick Tuszynski and published as a Mercatus Center Working Paper, “Institutions and State Spending,” October 2011.