Which People?

Think about two problems. A streetlight out on your road, and a disease spreading through the country. Nobody would send the broken light to Canberra. Nobody would leave the epidemic to your local council. We already know, without being taught, that some decisions belong close to home and some belong a long way up.

The hard part is everything in between. And most of the public argument about it gets run as though we had to choose, once and for all, between big and small. That is what I want to look at here. Because underneath all the questions about how a decision gets made, who shapes it, who votes on it, there is a plainer one that comes first. Where should the decision sit at all.

Walker Briefing is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber.

I should tell you where I start from, because it matters. I am drawn to the local answer. I have spent a working life persuaded that decisions made far from the people who live with them tend to be worse decisions, and that closeness to a problem is a discipline, not just a comfort. That conviction is mine, and I will own it. But a conviction is not evidence. So I have tried to do here what I have tried to do with every repair in this series. Take the thing I am fond of, and put it through the same test as the rest. Find out what it actually does, and to whom, once it leaves the page.

Let me state the case for keeping things local at its strongest, because it deserves that before anyone tests it.

The idea has a name. Subsidiarity. The word is ugly and the principle is simple. A decision should sit as close to the people who live with it as it can competently sit, and move up when it gets too big for that level. The version I find most honest comes from John Macgregor, who takes it from the Swiss. A system should be as centralised as necessary and as decentralised as possible. Notice what that does. It builds in its own brake. It does not say push everything down. It says put a decision at the lowest level that can actually carry it, and send it up the moment that level is overwhelmed.

The clearest way I know to picture this comes from my own trade. In medicine, care belongs as close to the patient as the patient’s condition allows. With the GP who knows the history. In the home where the person actually lives. At the level where someone can see the whole of the case and use their judgement. But the same system sends a problem up the instant it outgrows that level. To the specialist. To the hospital. To the national service no single clinic could ever build. Closeness and escalation are not enemies there. They are two halves of one rule. And it is the whole rule that matters, not the cosy half about staying small that people tend to remember.

Underneath the principle sits a real argument about knowledge. A lot of what makes a decision good is knowledge that does not travel well. The particular street. The way the water sits in that ground. The history of who was promised what, and by whom. A distant office does not just lack this knowledge. Often it cannot get hold of it without flattening the very thing it was trying to measure. The anthropologist James Scott gave this kind of on-the-ground knowledge a name, metis, and it is the backbone of the local case. Decisions made close to the ground can draw on things that decisions made at a distance never reach.

There is an accountability argument too, and it is plainer. A councillor who has to do her shopping in the town she governs is answerable in a way a minister three steps removed is not. Macgregor points to Switzerland, where power is spread unusually thin and where, by most measures of prosperity and wellbeing, things go well. That is worth taking seriously. Whether the spreading of power is what produces the result is a harder question, and I will come back to it. For now it is enough to say the case for keeping things local, put at full strength, is neither silly nor sentimental. It rests on knowledge, on accountability, and on a principle that disciplines itself. Which is exactly why it is worth testing rather than cheering.

A principle can discipline itself on paper and still go wrong in practice. And the troubles with localism are not small.

The first is capture. Push a decision down to a small enough body and you do not necessarily hand it to the people. You hand it to whoever is organised at that size. And that is rarely the whole community. It is usually a determined few.

The second is exclusion. The smaller the patch, the easier it is to draw the line around people like yourself, and to push the cost of that choice onto everyone outside the line.

The third is capacity. A small council may simply not have the money, the expertise, or the depth to do the job handed to it. When that happens, closeness buys you nothing and costs you plenty.

And some problems are just too big for any one place. A river. An outbreak. A jobs market. A carbon budget. None of them respects a council boundary, and a system that insists on handling them locally handles them badly.

So what about going the other way. What about merging small councils into big ones. Here the evidence is more interesting than either side likes to admit.

The promise of mergers has always been the same. Bigger units, lower costs per head, savings from scale. Two decades of Australian research, led by the local-government economists at the University of New England in Armidale, has tested that promise against the real mergers of recent years. The broad promise does not hold. The savings mostly do not show up. Where they do show up, they are small, and confined to back-office administration. Services barely change.

When New South Wales pushed its merger programme from 2016, the modelling used to sell it was judged misleading on several counts by the economists Dollery and Drew, and the state’s own budget office put the cost of the exercise at around $445 million. When Queensland forced mergers in 2008, communities later got the chance to vote, and a wave of them chose to split apart again, at an average of about $8 million each to unwind. And there is a cost on the other side of the ledger that is not just sentiment. When you consolidate, the quality of local democracy tends to fall. There is a real trade-off between running things efficiently and people being heard.

Put the two sides next to each other and a pattern shows up that neither one sees on its own. The decisive thing turns out not to be how far down you push a decision. It is two other things entirely.

The first is money.

Hand a small body a job without handing it the money to do the job, and you have not empowered anyone. You have created what is politely called an unfunded mandate. A small purse, a large task, and the blame when it fails. This is where the Swiss example, Macgregor’s own best case, turns in the hand. And it has to be turned carefully, because the Swiss result is over-determined. They are wealthy. The population is small and settled. They have a long direct-democratic tradition. Any single-cause story about Switzerland, including this one, should be held loosely. But one difference survives the comparison. The Swiss commune sets and raises its own taxes. The Australian council cannot. That is the better candidate for the work we keep crediting to the mere direction of devolution. Fiscal capacity is the thing the romantic version of the local idea leaves out.

The second is fit.

The gain from putting a decision at one level rather than another comes from matching the level to the size of the problem. And the matching runs both ways. Some things genuinely belong lower than they sit now. Others, the river and the outbreak again, belong higher. That is the escalation the principle built in, and the part its keenest supporters forget first. The research holds a finding here that sounds at first like a contradiction. It reports no evidence of a best size for a council. That is not a problem for matching the level to the problem. It is the reason for it. There is no best size because the right level is not a property of councils in general. It is a property of problems in particular. The useful question is never how big a council should be, which has no general answer. It is at what level a given job gets done well. And the answer is different for water, for planning, for public health, and for waste.

Which leaves the argument that organises almost every public fight on this subject, the choice between centralising and localising, looking less like the real question than like a substitute for it. There is no best size, so neither direction is better in itself. Western Australia, as it happens, resisted both a split and a merger inside a single generation. Capture turns up at the small scale and the large one, changing only in who does the capturing. The centralise-or-localise argument is not where the real work happens. It points past itself, to money and to fit, and on to a harder question. Who actually gains when a decision is moved.

Every repair in this series has been put to the same test, and this one must take its turn. The test asks not whether a proposal means well, but who actually gains when it is adopted. The phrase that sells localism to almost everyone is a warm one. Returning the decision to the people. The test’s first question is the one that phrase is built to stop you asking.

Which people.

The honest answer is uncomfortable for an idea I support. The people is not one settled body. Push a decision to a small place, and the winner is whoever is strongest in that place. On the questions that matter most to most households, that is a very particular group.

Most people’s biggest asset is the home they live in. When the local council controls what can be built nearby, the homeowner who already owns has every reason to use that control to protect what their home is worth. The economist William Fischel called this person the homevoter, and a long body of work on exclusionary zoning backs it up. Who exactly is dominant depends on the place. In a mining shire it might be a single big proponent. In a thinly settled country council, something else again. The point is not that the homeowner rules everywhere. It is that wherever owner-occupiers are the organised majority, which is most of urban Australia, the local vote hands them the decision. And the results are not really in dispute. Zoning that keeps people out. Housing supply held down. The cost of both pushed outward, onto people who do not live inside the boundary and so have no say within it.

This is the finding that does the real damage, and it does it at the root, not the edges. The whole premise of the local idea is that a decision should sit closest to the people most affected by it. But on housing, the people most affected by a refusal to build are exactly the ones the local vote shuts out. The young woman who cannot afford to live in the suburb she grew up in. The worker priced out of the town where the work is. The renter who has no vote in the council deciding whether anything new will be allowed. The most affected are, by the design of the thing, the ones who do not live there. At the precise point where closeness is meant to be the virtue, the affected person is not in the room. The principle does not just underperform here. Its founding promise turns over completely. The rule that a decision belongs closest to those most affected delivers, on housing, a decision controlled by those least affected of all.

Honesty needs the other half, and it cuts the other way. The same research shows why homeowners often fight being merged into bigger city-wide bodies. At the larger scale the capture does not disappear. It changes hands. Business and development interests can organise across a big area better than scattered residents can, and they tend to run the consolidated body. Macgregor supplies a fact from that side himself. Merging councils in the growth fringe around our cities raised the incentive for zoning corruption, because each rezoning decision was worth more once the unit got bigger. So the test does not end up preferring big to small. It ends somewhere more demanding. Capture is not a feature of one scale that the other escapes. It lives at both. It changes who, not whether. The incumbent homeowner below. The developer above. Returning the decision to the people names no one in particular, because at every scale it returns the decision to whoever was already best placed to take it.

In this country all of that sits on ground that makes the money question even sharper. Local government here is not even in the Constitution as a thing in its own right. Section 107 leaves it to the states, who create, change, and abolish councils by ordinary law. Twice we have been asked to change that, and twice we have said no. A third attempt was pulled before it reached a vote. In 1974 a proposal to let the Commonwealth fund and borrow for local government failed, carried by not a single state, backed by about 47 per cent of voters across the country and just over 40 per cent here in Western Australia. In 1988 a proposal merely to guarantee that local government would exist and be elected failed in every state. In 2013 a proposal to let the Commonwealth fund councils directly passed both houses of Parliament and was then abandoned before any referendum, because everyone expected it to fail as well. The local tier exists, in law, at the pleasure of the states.

The money position follows from that, and its sharp edge is not what people assume. It is not that councils live on handouts. They actually raise most of their own money, through rates and charges. The constraint is somewhere else. They cannot set the base their rates are drawn from. Several states cap the rate in the dollar they are allowed to charge. And for anything beyond what rates will cover, they have to look upward, to grants whose constitutional footing is itself shaky. This is the exact opposite of what makes Macgregor’s best case work. A Swiss commune commands its own revenue. An Australian council collects from a base it is not free to grow, and reaches up for the rest to a tier the Constitution does not firmly require to provide it. Any serious plan to push decisions downward in this country has to face the fact that the bodies it would hand power to do not control the money that power would need, and that the constitutional structure stands in the way of giving it to them. The direction of devolution is the easy part. The money to make it mean anything is the hard part. And it is the part the country refused at two referendums and ducked at a third.

Western Australia gives the cleanest picture of the deeper point, which is that people resist change being done to them, not the direction it travels in. In the mid-1990s the old City of Perth was broken into four smaller councils, on the argument that smaller would be closer and more responsive. It was unpopular. Two decades later the Barnett government proposed the reverse, cutting the metropolitan councils from thirty down to fourteen. After about six years of effort it collapsed, in the face of ratepayer polls, a revolt on the government’s own benches, and local government withdrawing its support. The state tried devolution by splitting, and consolidation by merging, inside a single generation. Residents resisted both. Whatever the public was defending, it was not a direction. It was the integrity of arrangements they had not been genuinely consulted about. Two smaller details sit beside this as colour rather than argument. The City of Perth council was suspended in 2018 for a governance inquiry that ran about two years and cost close to $8 million, which is a useful reminder that small and local is not the same as clean. And the smallest council in Perth, and one of its wealthiest, was left out of the merger proposals altogether. A very small and very well-off council has little to gain from a merger and the means to resist one, and reform of this kind tends to leave such places exactly where they are. The point is the incentive the structure creates. Not anyone’s motive.

I set out drawn to this principle, and I have to report that the test has left it standing, though not in the shape its keenest supporters, or Macgregor at his most expansive, would choose. Subsidiarity properly stated, the lowest competent level, with escalation built in, is sound. What does not survive is the simpler and more seductive idea that pushing power downward is itself the cure. The evidence does not put the cure in the direction. It puts it in two conditions the direction leaves untouched. Whether the body receiving a decision has the money to carry it. And whether the level fits the size of the problem. Where those two hold, closeness is a genuine good. Where they do not, devolution hands a community responsibility without power, and calls the result self-government.

The test leaves a harder residue still. Returning the decision to the people is a phrase that holds only as long as nobody asks which people. Once the question is asked, the answer is whoever was already strongest at the chosen scale, and the people most affected by a decision are often precisely the ones the local vote leaves outside the door. I owe one honest admission about this test. It pressed hard on the move downward, and it largely took the move upward on trust. Who decides that a problem belongs higher, and whether that judgement can itself be captured, is the same question owed in the other direction, and I have not paid it here. None of this is an argument for centralising. It is an argument against believing that the question of level is the same as the question of power, when the evidence is that they are not.

There is one thing this whole chapter has quietly assumed, and it is the thing the next one article looks at. The local idea assumes a community actually equipped to govern at the level power is handed back to. Informed. Paying attention. Able to hold a local decision to account. Closeness without that is not democracy brought nearer. It is capture made more convenient. Whether that capacity can be built, and over what kind of timescale, is the question of civic education. And that is where this turns next.

Walker Briefing is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber.