Why did Melbourne 2030 fail?

(click) Heaps of parking in central Paris in 1976!

The Age editorialises (21/11/10) that Melbourne 2030 is effectively dead and I agree. The latest nail in the coffin in The Age’s opinion is the apparently burgeoning growth of housing in townships and hamlets located in the peri urban area outside the Urban Growth Boundary (UGB).

I’ve argued before that this sort of “decentralisation” is poor policy (e.g. here and here). But I also think The Age has tended to ‘catastrophise’ the scale of the problem, especially with its highly misleading contention that Melbourne has “sprawled 50% beyond the official growth boundary, spanning 150 kilometres from east to west”.

However what interests me at the moment is why Melbourne 2030 failed. The key reason in my view is that it blithely assumed that enough affordable dwellings – mostly town houses and apartments – could be provided within the established urban areas to avoid the need for the UGB to be extended.

This objective was never realistic for a number of reasons.

Residents living in established areas are not generally receptive to redevelopment. Given a choice, most people don’t want the house next door replaced with multiple dwellings. The planning system recognises this and imposes restrictions on what can be built within established areas and gives neighbouring residents the right of appeal.

The Government and planners reacted to resident opposition by deciding that larger redevelopments should be confined to a small area i.e. activity centres and major transport corridors. Planners and architects liked this approach because it emphasised proximity to public transport and created a sense of urbanity; Government liked it because it took the heat off the vast bulk of the established suburbs.

However this tactic didn’t do much to increase housing supply. Difficulties with land assembly combined with planning constraints in activity centres meant supply was restricted. Only up-market, higher density apartments were viable.

It costs more to build a two bedroom apartment within the established suburbs than a three bedroom house with a garden in the Growth Areas. According to property consultants, Oliver Hulme, the median price for a new entry-level two bedroom apartment in metropolitan Melbourne in the September quarter was $509,250. The median size of these apartments was 68m².

The median price of a lot in the Growth Areas, on the other hand, was $212,000 in the September quarter. A settler can build a 176m² home for a starting price of $150,000 or a 350m² behemoth for $250,000. The combined land and house price of either option is still less than the median apartment.

It was imagined that “empty nesters” in established suburbs would sell their large houses to young families and relocate to smaller dwellings. But they prefer to “age in place”. They still want space for their adult children to live at home, for visiting grandchildren and for home offices. They like the familiarity of their area and the garden they might have spent decades creating. Importantly, there is usually little financial benefit from “trading down”.

It was also imagined that “spare” infrastructure capacity in established suburbs would give government a financial incentive to tackle planning restrictions on redevelopment head-on and keep the cost of new dwellings down. Although definitive evidence is lacking, it seems likely that whatever spare capacity existed has now gone. Indeed, it seems probable that the cost of expanding capacity is higher in established areas than it is on the fringe.

Another assumption was that there would be sufficient disused industrial sites within established areas to absorb much of the growth that would otherwise go to the fringe. For example Monday’s editorial in The Age argues:

The assumption that new housing land is to be found only outside the UGB is simply false: large undeveloped or disused industrial sites can be found even close to the CBD, under the West Gate Bridge and in suburbs such as Richmond, Collingwood, North Melbourne and Footscray

The first flaw in this assertion is captured in Challenge Melbourne, the discussion paper released during the preparation of Melbourne 2030. It states that “capacity for an estimated 65,000 dwellings on large sites such as old factories has been identified in the established suburbs”. Disused sites are an important resource, but given that the number of households in Melbourne is projected in Victoria in Future to grow by 825,000 between 2006 and 2036 – more than half of whom are projected to settle in existing suburbs – the potential contribution from “brownfields” sites will be modest.

The second flaw is that the sorts of dwellings likely to be constructed in the inner city suburbs mentioned by The Age will, on average, be way too expensive and way too small for the kinds of households settling in fringe locations.

Given all these issues, it is not surprising that most home buyers headed for the fringe.

The additional transport costs they consequently face are much exaggerated. Workers settling in outer areas overwhelmingly work within the suburbs and relatively close to where they live. Petrol is not expensive relative to average incomes and new cars are remarkably cheap.

The Growth Areas are by no means perfect, but people who settle in them are prepared to make compromises. They consciously choose to trade-off higher transport costs to get the benefits of home ownership, principally security of tenure and capital gain. Many of these locations are in any event better in terms of services and facilities than the fringe was a generation ago. The new residents know they will improve over time.

I’m not claiming that the factors listed above cover every reason why Melbourne 2030 failed – I’ve barely mentioned transport, for example, and there’s a whole world of politics. But they do illustrate that Melbourne 2030 did not understand the implications of the policies it propounded.

It should’ve had more regard for the preferences of existing and new residents. It should’ve gathered data and tested assumptions more rigorously. Most importantly, it shouldn’t have advanced policies without a committment to the sorts of supporting actions needed to make them work in practice.