Thursday, June 25, 2015

Dodgy rezoning, a summary

If you click the news link above you will see that my research on land rezoning decisions and the relationship networks of land owners has had a fair bit of coverage.

One thing I have learnt is that this type of quite technical research requires some effort to translate into bite-sized pieces for the broad media-consuming audience.

This post will be a reference point for the media and interested people that summarises the key findings and provides a couple of simple graphs and visuals that are not specifically included in the original research paper, but that can communicate the basic findings well.

What I did

I took a sample of landowners inside and outside rezoned areas in 6 locations in Queensland, where the statuary body, the Urban Land Development Authority (ULDA), took planning controls away from councils with the intent of increasing density, land values, and the speed of housing development.

In the maps below the blue disks are the landowners in my sample inside the rezoned area (black outline), sized by their land area, while the red disks are outside landowners in my sample. 

The logic of doing this is that these outside landowners could have been rezoned had the boundary of the areas been decided differently. From interviews with former public officials, and many others involved with planning decisions, it came to my attention that there is quite a bit of discretion about boundary decisions in zoning and that well-connected landowners often use their political clout to make sure the boundaries encompass their properties. The odd shapes of these areas are quite suggestive of this type of favouritism, as they have no apparent economic justification.

The sample is filtered to ensure I capture only undeveloped parcels that are at least as large as the minimum size rezoned parcel. There are 1,137 landowners owning 1,192 parcels in the sample.

Land characteristics or owner characteristics?

To see whether the characteristics of landowners likely to reflect political influence were a key determinant of these boundary decisions I collected data from both inside and outside landowners on the following:
  • Political donation activity 
  • Professional lobbying activity 
  • Membership of property industry lobby groups 
  • Corporate relationships through cross-directorships and ownership 
I also created a network of relationships that included the landowners using lobbyist-to-client connections, industry group connections, corporate connections, and
  • Connections from ULDA staff to their former employers 
  • Connections to politicians from their former employers 
The network had 13,740 entities and 272,810 edges.

I then model the effect of these characteristic on rezoning success, finding that being connected in the network increased chances of rezoning by around 19%, and getting into the most favourable part of the network gained an additional 25% chances. Employing a lobbyist improves your chances 37%, even after controlling for all other factors. Political donations don’t show and significant prediction on land rezoning when controlling for these other factors.

Connected landowners owned 75% of the rezoned area, and only 12% of the land outside. 

In the graph above I have the proportion of connected and not-connected landowners that were rezoned, the proportion of landowners who don’t employ lobbyists rezoned versus the proportion of those who do, and the same with political donors. 

The size of rezoning value gains

Using 822 historical sales of development sites inside and outside the areas in the study I estimate the price deviation attributable to rezoning. Essentially the rezoning increase prices across all areas by 81% relative the the neighbouring sites outside the rezoned areas. The below graph shows the price deviation through time, and the big change that comes the year of rezoning.

In all the value to rezoned land owners was $710million, of which connected landowners gained $410million. In terms of the marginal gains to becoming connected in our sample, this was $190million. While on a per hectare basis the mean gains are just $56,000, which isn't much, the sheer scale of the the gains to a narrow group of people represent a problematic political transfer from the unconnected to the connected. It also suggests that many billions in value are regularly transferred to connected landowners through routine rezoning decisions.

What to do

There are two main ways to stop this political favouritism - disrupt the favour exchange, and remove the honeypot.

Disrupting the favour exchange means requiring cooling-off periods for former politicians and bureaucrats, but also a policy of electing independent people to boards and decision-making positions. Why does the ULDA need to have its board stacked with local developers? Why not have a planning expert from Europe? There is a myth that local expertise is somehow required in the regulatory positions, but most of the time you don't get expertise, just loyalty to mates.

Greater transparency from all our public registers would also go a long way. Why aren't land titles a freely available database? Why isn't ASIC register of companies available for free? Why can trusts conceal who owns what so easily?

In terms of removing the honeypot, or reducing the value of discretionary political decisions, the obvious point here is to enact a process of selling additional development rights rather than giving them to selected landowners for free. This requires a pretty radical change in the way planning is viewed, but it makes perfect sense. Planning rules, including zoning, are part of the definition of property rights. You wouldn't give away a land area for free, so why give away land right of a different kind for free?

Alternatives to selling development rights are betterment taxes, and land taxes, both of which are effective administrative alternatives.

We can also think outside the box and look at development timing fees. If the rezoning is intended to increase density and increase housing supply, why not provide incentives to build sooner rather than delay? A per-dwelling fee that increases every year in the rezoned area would encourage developers to bring forward construction and sales.

Lastly, we can have local referenda on town planning changes to allow for representation of the interests of the politically unconnected. 


  1. Great work. It would be even better if you could extend a similar analysis to Victoria and NSW and see if the ban on political donations from developers has had any effect in NSW.

  2. Land Titles are not free because it costs an enormous amount of money to provide these certificates to every single law firm in the country on a daily basis. Same with ASIC certificates. It's not about concealment, it's about the cost of upkeep.

    1. I'm not really sure what you are trying to say. It seems you are saying that the automated electronic delivery system from the Asic and land titles databases is costly and needs to be recovered through fees.

      I doubt it.

      Sure, having a whole titles and company register is a costly exercise, but why should costs be recovered from those who want to see the public information rather than those who are registered and hence benefit from being in the database?

      Also, these are core government functions and can be funded from any general revenue source. There is absolutely no reason once you have such a database to not make it freely available.

  3. Cameron, did you consider the value uplift/capture charges in your analysis?

    1. I believe this is mentioned in the paper. However you will note that the additional charges under the label 'value uplift' are not in fact calculated from a valuation but are an additional fee for development that exceeds that allowed for in prior council level plans.

      The price estimates are net of any of these fees and show that being in a declared area still adds considerable value to the land.

    2. Cameron, I did a quick search of the report and it does not seem to be mentioned. That seems a massive oversight to me, as your key claim seems to be that private owners are getting a windfall gain that the government does not share in. That is patently untrue in most (but not all) circumstances in the ULDA/EDQ declared areas.

      The value capture charge for Flagstone/etc is $8,462/lot. Developers in this areas average around 15 lots per hectare or $126,930/ha. You summarize that the marginal gain of "being connected" is $56,000/ha - clearly the government is also doing well out of this arrangement and I feel that is most worthy of mention in your report.

    3. "be that private owners are getting a windfall gain that the government does not share in"

      Yes. That is true with or without the charges.

      The net effect (after value uplift charge) is a windfall gain.

      I am comparing to the case where governments sell additional development rights and capture 100% of the potential value gains. You are comparing to the case where government takes 0% of the potential value gains.

      So in your view the government is doing very well. In my view, since they actually could have sold the total value, they are not doing very well. What do you think should be the appropriate counterfactual and why?

      Just remember that in my estimate the value gains are net of the additional charges. That is, I estimate the value gains to private land owners after paying additional value uplift charges and other infrastructure charges etc. I don't estimate it before the charges and ignore the fact that landowners then need to pay charges.

      If you want to estimate the potential value before the uplift charge then simply add on to my estimate an estimate of total value uplift charges to get that total potential value gain amount.

      I did have comments about the existence of these charges and that my calculations are net of charges in an earlier draft, but after reorganising the paper it is no longer in it. I will certainly add a comment about this in future versions.

    4. can you point to any relevant examples of government capturing 100% of the windfall gain?? it seems like a very socialist approach to me.

    5. Every time a government sells property right - radio and tv spectrum, surplus land, fishing and water rights etc - they capture 100% of the value.

      I assume you would think it corrupt if a instead of the state government selling public land to private owners, they simply gave it to their mates for free. This is exactly the same thing, but in terms of incremental additional property right associated with a location (i.e. the landowner already has one set of property right, and are being given a new, more valuable set for free).