City or bush a question of history

An illustration by Peter Nicholson.

POLICIES to move people out of cities into the regions have been tried and failed before, says Henry Ergas.

IN its recently signed agreement with the two independents, the government commits to promoting "an ex-urban generation in Australia". What this means is unclear but the underlying premise is clear enough: we would be better off shifting Australia's population from the cities to the regions.

That belief is hardly new. Closer settlement, as it was called, was an aspiration even before Federation. Japan's victory over Russia in 1904-05, combined with US president Theodore Roosevelt's warning that "an unmanned nation invites disaster", then gave renewed impetus to efforts at settling our sparsely populated areas. Combined with the rise of rural populism, the result was a succession of poorly conceived initiatives, stretching from soldier settlement to the Ord River scheme and the fantasies of Gough Whitlam's regional centres.

None of these made Australia as a whole better off. Nor did they even do enduring good to the recipients of the government's largesse. And if the objective was to shift Australia's population distribution, the projects were miserable failures. That record of failure is unsurprising. The geographical distribution of our population reflects powerful underlying economic forces. Policies that ignore those forces have failed repeatedly and will continue to do so.

The crucial fact is that dispersion is expensive. Infrastructure, health and education costs soar as population densities decline. And dispersion causes the loss of agglomeration economies, which are the efficiencies that come from the knowledge spillovers, greater choice and intense competition that characterise large centres.

As a result, only highly productive but completely site-specific activities, such as mining and some forms of pastoral agriculture, can be economically viable in remote locations.

However, cramming people into cities causes costs of its own, notably congestion.

And the high productivity of resources in large centres increases competition for urban land, raising land prices and driving less productive uses to places where land is more abundant.

Market forces balance between these costs and benefits of urbanisation, concentrating activities for which the agglomeration economies are greatest, while dispersing those that benefit less from being co-located.

The resulting balance is not perfect. But most attempts at changing it cause more harm than good. This is not because people cannot be convinced to move. On the contrary, in an integrated labour market, populations will shift to areas that become more attractive. But it is one thing to make particular places bigger; it is another to make people better off.

To see why, consider the Pilbara. It is an awful place to work but mining can pay the high wages needed to compensate. Now, assume government spends billions converting the Pilbara into an earthly paradise. This will have three effects.

First, as people flock there, the value of local land will rise, but with a better living environment and labour more abundant, relative wages will fall: the hardship premium disappears. Second, with people leaving the cities, some economies of agglomeration are lost, though some may be gained in the Pilbara itself. And, third, the billions spent need to be paid for, with the resulting taxes causing distortions of their own.

Overall, the Pilbara's beautification will make local landowners richer. But if the project's benefits are less than its costs, the gain in land values will be less than the wealth that could be obtained by putting the resources diverted to the project to other uses.

As a result, Australian real incomes must fall, so local rentiers will have gained at the expense of workers and of the nation. For politicians with close ties to local landowners, that may be fine, but it is hardly good public policy.

The lesson is simple: to durably raise incomes, projects aimed at increasing the attractiveness of regional areas must have benefits that exceed their costs, including any loss in agglomeration economies from population movement. Meeting that condition will be no easier in the future than it has been in the past. This is because the efficiency advantages cities obtain from agglomeration seem to be getting stronger rather than weaker.

A recent study by the Organisation for Economic Co-operation and Development finds that within countries growth is increasingly concentrated in a small number of regions. Typically, these regions have long had high levels of human capital, while attempts at boosting performance in lagging regions have had scant success.

Nor is broadband likely to change that pattern. Rather, broadband increases output most in urban areas, in part because it boosts growth in services, which are best produced in cities. Its main consequence, therefore, is to flatten the distribution of city sizes, as has recently happened in Australia, but not to shift population out of cities themselves.

Little wonder a recent survey by Harvard professor Ed Glaeser concludes that policies aimed at promoting regional places are not worthwhile. Better outcomes, he suggests, may be achieved by removing the obstacles zoning creates to further expanding urban centres, thus making it easier to shift population from less to more productive areas, while using congestion and developer charges to better manage the disamenities of city life. Glaeser's bleak assessment of regional policies is controversial. But what is not controversial is the long history of failure. That should inspire approaches that are cautious, well-targeted and thoroughly tested.

More the pity, then, that the government's agreement with the independents promises vast outlays without any cost-benefit assessment. No less worrying is the promise to track and publicise the delivery of services in regional areas, but the absence of any commitment to do the same for costs.

This is a recipe for projects that merely shuffle wealth around while losing some along the way.

Ultimately, the trouble lies in the confusion that underpins policies such as these.

It makes no more sense to have a preference for Dubbo to be a certain size relative to Sydney than to have a preference for the textile industry to be a certain size relative to the software industry. And policies that benefit one place at another's expense waste resources in exactly the same way as does industry protection. Until that is understood, we will merely repeat the errors of the past.