Conditional spending grants

Why does a national government buy a policy with grant money when it could simply have legislated it?

A national government wants every state to set the same minimum drinking age. It does not legislate one. Instead it offers highway money with the age attached as a condition, and within a few years every state has complied.

The obvious reading is that money is simply more persuasive than law. But that cannot be right on its face — a law does not need to persuade anyone, it just applies. So the roundabout route is not chosen because it works better. It is chosen because the direct route is closed. The question is what closes it, and what the detour costs.

Begin with the premise. In a federation, the centre's legislative power is enumerated: it may act on the subjects assigned to it and not on others. Road safety, schooling, policing, licensing — the everyday texture of government — are commonly on the other list. So there are policies the centre wants and has no authority to command.

Now ask what it does have. It has revenue, usually far more than its own functions require, because the taxes that raise the most money are the ones only a large jurisdiction can collect efficiently. Meanwhile the units below spend heavily on services and cannot easily raise what those cost. The two facts together produce a standing gap, and money flows down to close it.

That is the raw material. The move is to attach a string. And notice the legal form this takes: the centre is not commanding, it is offering. The recipient may refuse. So no enumerated power over the subject is required, because on this framing nothing is being regulated at all — a contract is being proposed. That is the whole trick, and it is why the technique appears wherever competence is divided and revenue is not.

Is the framing honest? Sometimes plainly yes: a small grant for an optional programme is a genuine offer, and refusals happen. But run the logic outward. If the centre may condition one dollar, may it condition every dollar the recipient receives? At some ratio of grant to budget, "you may refuse" stops describing anything real — a state told it will lose the funding of an entire existing service if it declines is not choosing.

Courts have had to draw that line and found it hard. The United States Supreme Court upheld conditioning a small percentage of highway funds on the drinking age as mild encouragement, and later struck down a condition putting a state's entire existing health-programme funding at stake, calling it a gun to the head. Between those poles it has not said where the boundary lies. I do not think anyone has a principled formula; the honest description is that the doctrine names both ends and leaves the middle to judgement.

What would falsify the account? If conditional grants are a workaround for missing competence, they should cluster in subjects the centre cannot legislate directly, and thin out where it can. That is testable across federations and over time: when a centre's direct powers expand, grant conditionality in the newly covered field should fall away as redundant. The refuting observation would be conditions spread evenly across all fields including those the centre already regulates outright — which would mean conditionality is about something else, most plausibly administrative control over how money is spent rather than about competence at all.

There is a second reading worth keeping alongside the first, because in some systems it dominates: even where the centre could legislate, grants let it enlist an administration it does not own. Germany's federal structure runs much national law through the Länder, and the money follows the administrative duty rather than substituting for a missing power. So "buying a policy" and "paying an agent to deliver one" are different mechanisms that use the same instrument.

Three refinements.

First, the constraint being evaded is legal competence, not political difficulty. A centre that could legislate but expects a hard fight will not usually reach for a grant, because grants must be funded every year and are therefore more politically exposed, not less. This is why the mechanism sits in federations rather than in unitary states with strong local government — the neighbouring question of why decisions get devolved at all is about matching decisions to information and spillovers; this one is about what happens after the constitutional line is drawn and the money sits on the wrong side of it.

Second, conditionality has a ratchet. Once a recipient builds staff, buildings and clients around a funded programme, the cost of later refusing a new condition includes unwinding all of that. The centre's leverage therefore grows over time without any new grant being offered, which is a quiet and unlegislated shift in the balance of a federation.

Third, this arrangement has a distinctive accountability defect, running in both directions. The centre sets the policy but does not administer it, so it is not blamed when delivery is poor. The recipient administers it but did not choose it, and can say so. Voters facing a failing programme find two governments each pointing at the other, both accurately.

The cost side, stated plainly: conditional grants let a federation act coherently on problems that cross its internal lines without amending anything, and that flexibility is real. What it buys with is the clarity of the original division of powers, the recipients' room to try different answers, and a legible chain from a policy to whoever chose it. And the arrangement is unusually stable against reform, because a centre asked to give up conditionality is being asked to surrender influence it already holds, while recipients are being asked to risk money they already receive. Nobody with standing to change it gains by doing so.

• How vertical fiscal imbalance — the gap between who taxes and who spends — is created deliberately in some federations and reluctantly in others.
• Why unconditional block grants keep being proposed as the remedy, and what recipients actually do with them.

Every term the collection defines is gathered in the glossary.