Urban growth boundaries

Why does drawing a line to stop a city sprawling sometimes push building even further out?

A city draws a line around itself and says: inside, we build; outside, we do not. It is about as direct an instrument as planning offers. And yet the recurring complaint about such lines is not that they fail to hold, but that development leaps clean over them and lands further out than it would otherwise have gone.

That is a peculiar kind of failure. A wall that stops nothing is easy to understand. What we have here is closer to a wall that stops things and makes the situation worse anyway. So the question is not "does the line hold?" but something sharper: what is a developer actually deciding when they look at that line?

Put yourself in their position. You have money and a horizon of perhaps fifteen years. A field just outside the boundary is farmland today, priced as farmland. If the boundary were ever moved outward, that same field becomes building land and is worth many times more. What do you do?

You buy it and wait. Notice you did not need the boundary to move — you only needed it to be plausible that it might. And now notice what that does to the city: the land immediately outside the line is now held by people whose profit depends on the line moving, which is a constituency that did not exist before the line was drawn.

So the first thing the line's effect turns on is not geography but belief. Does anyone expect it to move? Ask what would make it credible. A boundary set by the same council that grants permissions, revisable by ordinary vote, is a statement of current intent. A boundary written into higher law, or defended by a body that gains nothing from releasing land, is closer to a promise. The physical line is identical in both cases; the behaviour it produces is not.

And the honest complication is that even well-designed schemes build a release valve in. Oregon's system, the best-studied example, requires each urban growth boundary to contain roughly a twenty-year supply of buildable land, so the boundary is periodically reviewed and expanded as that supply is drawn down. That is a defensible rule — it is what keeps the line from strangling housing supply — but it also tells every landholder that expansion is a matter of when, not whether. A commitment with a published procedure for being relaxed is a weaker commitment, and the land market prices that in.

Now take the other case, the one that produces the leapfrog. Suppose the boundary genuinely holds. Building inside must go up rather than out, land inside becomes dearer, and some households and firms are priced out. Where do they go? Not into the protected ring — they cannot. They go over it, to the towns and villages beyond, which typically sit under different councils that never agreed to the line and have their own reasons to welcome growth. England's green belts around London and other cities are the standard illustration: growth beyond them, and long commutes back across them, are a well-documented pattern, even if planners still argue about how much of it the belt caused versus merely shaped.

Do you see the structure now? The line constrains where you may build, but not whether people need somewhere to live, and not the jurisdictions past its far edge. Demand is conserved; only its geography is redirected. Whether that redirection is upward — denser city — or outward — a ring of commuter towns — depends on whether the city inside the line can actually absorb the growth, which is a question about zoning, height limits, and appeals, not about the boundary at all.

That is the pairing worth holding on to. A boundary that is not believed produces speculation at its edge. A boundary that is believed but is not matched by capacity inside produces settlement beyond it. Only the third case — believed, and paired with permission to build densely within — produces the compact city the line was drawn for.

That anticipatory quality is why this is a commitment problem rather than a design problem. The policy's effect is produced by expectations about its future, not by its present position, which means a boundary can be doing its work well or badly without moving an inch. It also means the usual instinct — redraw the line more carefully — addresses the wrong variable.

A misconception worth correcting gently: the leapfrog is often read as proof that boundaries do not work. It is better read as proof that a boundary is only half a policy. It states where growth may not go; it does not state where growth shall go, and it has no authority over the land past its outer edge. Pair it with genuine capacity inside — upzoning, permitted heights, fast consent — and the pressure has somewhere legitimate to release. Leave the inside as restrictive as before, and the line simply exports the growth.

And one more refinement, because the two failure modes are easy to conflate. Speculative land-banking happens at the boundary and depends on the line looking movable. Leapfrogging happens beyond the boundary and depends on the line looking firm. A city can suffer either, and diagnosing which one it has tells you whether the missing ingredient is credibility or capacity.

• Land value capture: taxing the uplift when a boundary does move, so the gain is public rather than speculative.
• Green belt versus growth boundary: the English and Oregonian designs differ in purpose, and the difference matters.
• Jurisdictional fragmentation, and why regional planning bodies exist at all.

Every term the collection defines is gathered in the glossary.