Lots of people think government can do good things, even if its always in fashion to wink about the state’s legendary incompetence. It can do lots of things, but the only way it can do them is by exercising its power to coerce. It’s simply impossible to form an effective government without granting it that power. We must hope it will do only good things, and there is reasonable consensus that its basic functions are good, at least in kind: national defense, law enforcement and protection of basic rights, and a judicial system. The mere performance of those functions requires coercive power, and funding them requires the coercive power of taxation.

To make things simple, for now let’s stipulate that all agree on both the necessary functions of government, some minimal scale and scope of those functions, and the taxes necessary to pay for them. We may all feel that we are better off. Anything in excess of that minimal portfolio as might be desired by an individual or group would necessarily make some feel better off and some feel worse off. Additional taxes would have to be collected to pay for it, and the activities themselves might be seen in some quarters as inappropriate, wasteful, or intrusive. Now, the coercion of the state becomes more binding on some individuals and groups. We no longer have a win-win proposition, and that is what  distinguishes marginal government activity from marginal private exchange. The latter is always predicated on mutual benefits for the transacting parties. In the jargon of economics, these voluntary, private trades are Pareto-improving moves, meaning that some individuals are made better off and no one is made worse off. In general, if all mutually beneficial trades are exploited, the final result is Pareto optimal (after Vilfredo Pareto), because no further activity can make anyone better off without making someone else worse off.

The limited government described in the hypothetical sounds as if it might be Pareto optimal, but let’s add a little more realism. Are there additional government functions that would improve well being without doing harm to anyone? There is general agreement that government should provide for other “public goods”, which would otherwise be under-demanded in the market, and under-provided, due the nonexclusive nature of their benefits (think public parks). Once those are provisioned, the outcome may be Pareto optimal. There may be unanimous agreement, as well, that government should take actions to mitigate certain external costs arising from private activity. (If some of the costs of private activity are not internalized, then those market transactions fail the test of Pareto improvement). These additional government functions require coercive power, of course. Now we are into more complex issues of public choice. The provision of goods with at least some public benefits requires judgement as to degree, and judgement is necessary as to the appropriate degree of mitigation of external costs when they are an issue. In other words, Pareto-improving moves get scarce once government assumes responsibilities beyond those described in our original hypothetical.

As the scale and scope of government grow, its coercive force must advance as well. Therefore, unanimous consent for this growth, and even widespread consensus, will be impossible to achieve. Its size will reach a level at which a substantial share of the population will assume the roles of “public servants”, all having a vested interest in the state’s continued growth, if only to boost their own pay. The potential conflict of those personal interests with the public interest could not be clearer. That’s a good reason to support strict limits on the size and power of government, not to mention restrictions on the power of public employees to unionize.

Those who wish for government to play a dominant role in society might think it’s all for the good. They might support changes in the rules of governance that facilitate the dominance and coercive power it confers upon them. That might include, for example, pushing the use of executive authority to extreme levels based on interpretations of complex, but often vague, legislation. It might include changes in parliamentary rules that make it easier for thin majorities of legislators to work their will. No doubt these rule  changes will lead to Pareto-degrading actions, though the ruling faction will be quite happy with their new powers.

But what happens when a shift in the balance of public opinion brings new leaders to power? Those leaders will inherit rules that facilitate their agenda and authority to exercise coercive power. No one at any point along the ideological spectrum should dismiss this sort of risk. That’s the spirit of a recent Bryan Caplan post, “Limited Government as Insurance“. Stretching powers in the service of particular policy goals may well backfire when those powers become available to an opposing faction:

“Imagine going back in time to January 20, 2009. Obama’s Inauguration Day. You’re a cheering fan. On that day, an angel appears and makes you this offer: If you give up on Obama’s best ideas, none of Trump’s worst ideas will happen either. Obamacare will never happen – but neither will Trump’s immigration policies. Would you take that deal?

I know, it’s a galling hypothetical. You want the good stuff without the bad stuff.“

Caplan characterizes strict limits on government as a form of insurance against the risk of swings in the balance of power. He also considers plausible reasons for rejecting such a deal: “the arc of the moral universe”, or, you think your side will ultimately win, and will win for all time; and “asymmetric hyperbole”, or, the greatness of your policies outweighs any damage the other side can do with the same powers. If you really believe those things, it might seem reasonable to take your chances on an expansive state with expansive powers! A preference for limited government, however, does not require the contorted logic required to reject this insurance.

The U.S. Constitution includes many provisions originally intended to limit government and the exercise of coercive power. Those protections from the state have eroded over time, a process hastened by increasingly flexible judicial interpretations of the founding document. Caplan notes that there are a number of mechanisms by which limited government can be made durable:

“Supermajority rules require more than a majority to act. Division of powers makes it hard for government bodies to accomplish anything on their own. Judicial review allows judges to invalidate acts of government. Federalism greatly reduces the cost of “voting with your feet.” If you think these institutions aren’t working, the obvious solution is to strengthen them. Impose more supermajority requirements. Divide more powers. Overturn legislation that fails to get support from six, seven, eight, or all nine Supreme Court Justices. Make states pay for their own spending with their own taxes, not federal grants.“

Then comes the most insightful, but most disheartening, part of Caplan’s post: real steps to limit government will never be taken:

“Limited government helps everyone in the long-run, but immediately hurts the ruling party. They fought hard to win power; now that they have it, they yearn to flex their muscles.“

We might see a federal department abolished here or there, and we might see certain regulations rolled back, but those steps will be selective. The powers that put them there in the first place can be reapplied in the future. We might see more “business-friendly” actions, but those will be selective as well. In other words, corporatism will persist. And we might see tax cuts, but that won’t reduce the government’s absorption of resources, which is driven by the spending side. While this sounds discouraging, I nevertheless admire Caplan’s characterization of limited government as insurance against the other side’s bad policies. If only we could pull it off!