Thoughts on Federalism

Should We Abolish the Australian States?

This is an attempt to put on paper my own thoughts on Australia's federal structure and some proposed modifications. It's very much an early draft.

Is it Practical to Abolish a Tier?

A commonly expressed desire, amongst those campaigning for the abolition of the states, is to reduce the number of tiers of government in Australia from three (federal, state and local) to two. Assuming Australia is to continue to exist as a nation state, this implies one of three things: abolition of local government; abolition of state government; or abolition of both and their replacement by a new level, presumably in between.

The third option is the one most commonly supported in what I understand is to be known as the Abolish States Collective. All state, territory and local governments (check) are replaced by around thirty to fifty regional governments. Each regional government would have a population of something like a quarter to half a million people.

Applying this to, for instance, country New South Wales shows that a typical region would contain several largish country towns and a sprinkling of small ones. It would be difficult for the democratically elected representatives in such a body to make good decisions about, say, zoning (probably the classic example of local government responsibility), since many of them would have at best occasionally visited the town in question. It might not be impossible: electoral laws could be structured so that at least one member of council had a responsibility to a smaller area; there could be appointed officials whose job it was to advise on local issues; those inclined to the futuristic may anticipate virtual reality walkthroughs and their ilk in the longer term. But this is a long way from an ideal of representative democracy. There are some decisions best made by bodies responsible for small areas.

The Woldring variant of this is (check) to have the regional governments indirectly elected through local governments. I take this to imply the continued existence of local governments below the regional level, and hence that we abandon the hope of eliminating a tier. My own suspicion is that this is inevitable, though not necessarily through the Woldring mechanism.


All this leaves open the issue of whether it is desirable to eliminate a tier. One motive given, for instance in the Drummond calculation, is to eliminate duplication between governments.

Drummond comes up with an estimate of the population-independent cost of government: I'm concerned whether there aren't alternative interpretations to the graph. For instance that governments simply spend as much as they are given, and that it costs the Federal government $46M to bribe each senator into acquiesence, and $666M for each member of the house of representatives.

We might hope to achieve a reduction in duplication costs by choosing better demarcation lines between the responsibilities of the various tiers.

Demarcation Between Tiers

We need to decide exactly which state responsibilities and powers would be transferred up to the national level, and which down to some sort of regional or local level. My own feeling, subject to revision without notice, is that we could transfer almost all a state's powers up to the national level. At least, I can't see any that couldn't be.

Some people would be concerned about concentrating all the police power in one body. Basically it's scary that any government control police. There is, though, some feeling of security, perhaps spurious, in thinking that the AFP keep an eye on state police forces and might kick up a stink if they got too repressive.

State governments may have a special responsibility over critical parts of their capital city (check). There are arguments for that to go up or down. The Federal government would presumably retain its special authority over parts of Canberra.

Drawing the Boundaries

A lot's been written on where the boundaries between regions should be drawn, and I don't have much to add. Whatever lines we draw, however, will at best be appropriate to the early twenty-first century, and will sooner or later need to be revised.

If the regions have some sovereignty then that's going to be very difficult. This is one reason I'm anxious about the Woldring proposal.

Mechanisms for Transfer of Responsibility

Something that interests me more than the demarcation line itself is the nature of the mechanism that exists to determine it. This is obviously unimportant in the short run, but in the long run I believe it's what made Australian Federalism such a poor system of government. Some of the mechanisms used include:

At present our system has a great deal of inertia in it: that is, a great deal of popular vote (plus the approval of the federal parliament) is required to make a change. This makes it very hard to transfer responsibilites between states. When railroads were first introduced to Australia, for instance, they generally ran for a short distance between a gold field, mine or sheep area and the nearest port. Compatibility with other railroads half a continent away wasn't much of an issue, so it made perfect sense for them to be a state responsibility. The reason they can't be made anyone else's responsibility now is that it would require either a constitutional change or the agreement of the states. They both happen occasionally (the latter occurred after Port Arthur, more or less) but it's hard. I suppose it may be too late for railroads anyway.

The difficulty of shifting powers and responsibilities is one reason why I'm presently opposed to giving sovereignty to the regional governments. Perhaps we could borrow some of the principles above in determining the level of sovereignty the regions deserve. For instance, we could allow voters to say at an election whether they approved of their regional government having sovereignty. A region that got enough support would be immune to changes to its powers or existence by the national government, or perhaps partially immune (say, three-quarters vote of the national parliament required). Perhaps the votes are made at the federal election, so that the voters get to approve/disapprove their regions mid-term. Perhaps a region must win a certain level of support two elections in a row, or at least has more immunity if it can. There's plenty of fine-tuning possible here. Some might be concerned that this will lead to a lot of Canberra-bashing at the regional level. I'm not sure how we avoid that.

Tradition is another issue, but I have no idea how to create a good tradition. Any proposal that hopes to harness the traditions that exist in Europe is going to need a convincing plausibility argument for a similar tradition arising in Australia, or it needs a good transition structure that will function adequately even in the absence of a tradition.

The Best Size for a Region

There's a principle in the military that a weapon is best assigned at the level covering a frontage comparable with the weapon's range. Similarly, we could argue that a power is best assigned at a level covering an area comparable with the externality range of the power. Certainly you wouldn't want the power assigned any lower than this if you could help it: you don't want regions of 100 km in radius handling thousand-kilometre rail links, or rivers. Use of the range principle might help to determine the size of regions if we know their powers, or vice versa.

Another approach is the time principle. The ramifications of a decision by a government whould affect its voters, so we don't want the unit so small that people have generally moved out of the region by the time the decision takes effect. With increasing mobility this is going to be more and more of an issue. This is another reason why we need to keep the definitions of regions flexible: a hundred years from now whatever structure we establish is going to look quaint at best and stupid at worst.

One concern with local government is that it doesn't get enough attention. I suggest this is partly because local governments are much smaller than the main units of media reporting, which are regional TV coverage footprints. I'm assuming here that nobody reads country town newspapers. We could choose to make the region about the same size as the TV footprint.

At local government the representatives are part-time. At state level they are professional politicians. This has to be considered in choosing our region size: what sort of representatives do we want?

Should regions even be geographical? Not that I have an alternative to propose: increasing geographical mobility won't be making regions ridiculous for a few years yet. At the local government level lots of people commute across boundaries. There might be something to be said for August's idea of giving people a vote where they live and another where they work (is this done today?).

Confidence in Local Government

At present I have very little faith in local government, and see no way to be confident regions would be efficient or honest. It would be nice if someone with real data on the subject could provide a more objective measure of the relative levels of efficiency and corruption at the national and local levels.

Assuming my prejudices are correct, how do we fix the problem? One argument is that with increased responsibility they will become ... well, more responsible. But that's a hell of a risk to take, especially if we give them enough sovereignty they can in practice evade national sanction.

Summary and Conclusions

One point I'd like to make is the importance of the mechanism used for making changes in the number and responsibility of governments, rather than the particular arrangement and number we might adopt when some new system is introduced.

Sub-national sovereignty is evil. Its evilness is in proportion to the number of sub-national units possessing it. Except, of course, that human rights are good: I'll have to work out if I'm being consistent there. I'm exaggerating my position here, and I reserve the right to deny I ever said it.

Still thinking. More later.

I welcome feedback at

This page is hosted by GeoCities, in return for carrying their advertising they will give you a free home page much like mine. Everything on this site varies without notice, especially after I get feedback. 1

Hosting by WebRing.