Monday, February 5, 2018

New paper: Developers pay developer charges

I have a new paper out — Developers pay developer charges in Cities: The International Journal of Urban Policy and Planning.

In this paper, I estimate the economic incidence of developer charges (taxes paid upon approval to use land for a higher value purpose) using a natural experiment in Queensland, Australia, where a surprise political announcement varied the charges. Using data on developer charges and dwelling prices during this ‘natural experiment’ period I estimate their economic incidence. The data clearly shows that the administrative incidence on the landowner (developer) happens to also be the economic incidence. An increase in the charge comes at no cost to the buyer of a new dwelling but instead decreases the land value by an equal amount.

The motivation for doing this analysis was an article in The Conversation that suggested the opposite — that the economic incidence was on the buyers of new dwellings, against all logic and reason. In fact, this research showed a significant correlation between developer charges and home prices at a ratio of 1:4. Erroneously interpreting this relationship as causal would mean that increasing charges by $1 would increase home prices by $4.

Can you see the nonsense here? If there really is a causal link property developers would be lobbying to massively increase charges in order to earn a 400% markup on them! In reality, the development industry has been lobbying hard to remove them.

In my paper, I demonstrate the problem with this causal interpretation, which arises because the variation in the developer charges is due to the way they are set by regulations. The regulations state that the charge per new dwelling of 2 bedrooms or less can be a maximum of $20,000. The charge for a 3 bedroom or larger dwelling can be a maximum of $28,000. Because councils have no incentive to charge less than this maximum this was the size of the charges in their data. The regression analysis merely showed that the average 3-bedroom or larger dwelling is 4 x $8,000, or $32,000, more than the average 2-bedroom or smaller dwelling (controlling for other quality and location factors).

Because my study covered a period where surprise political decisions varied the charges themselves for each dwelling type, my analysis shows no relationship. In fact, if you take out this surprise variation in my data and leave the charge at the fixed price for each size dwelling I can replicate the earlier results of a 1:4 correlation.

Why is this important? 
This result is significant because the economics of property is almost the exact opposite of the economics taught in most modern university degrees, and bad economics is being used to justify bad policy. All too often I see the following implicit assumption about causality:

Cost of capital ⇒ Rental price of capital.

If you increase the cost of investing in capital, you increase the rental price of capital. That is the logic behind the idea that developer charges, or a land tax, can be passed on to users.

But this clearly makes no sense in the case of land. Land is costless to produce. It is obviously not costless to buy it from someone else, but ultimately, there is no prior investment that provides its value. It is merely a legal right to claim certain incomes associated with that location. So for land (and ownership rights in general), the direction of causality must be:

Rental price of capital ⇒ Cost of capital.

This is not a secret. It has been widely known for hundreds of years in the property valuation profession, which uses variations of the ‘residual value’ method to isolate the pure rental price of land and use it to determine the cost (price) of land.

So what? 
Vested interests in the property industry continue to argue that shifting the tax base to land will increase the cost of housing — after all, they argue, the rental price is caused by the cost of land plus other costs, including taxes and charges.

We know this argument is bogus because it simply begs the question that if prices come from input costs, why does land have any value at all? All land rents should be zero.

And again, if the rental price of capital was the result of a summation of costs, the property industry would have nothing to fear from increasing developer charges, as they could pass on those costs in the price of new dwellings.

One step further
We can take this logic another step and see that because the economic incidence of land taxes (or development charges) is on the landowner, increasing these taxes can encourage more development sooner since it reduces the payoff from delaying investment in new housing.

Consider the table below. It comes from my paper. I use it to demonstrate the changed incentives to delay or bring forward new housing development from increasing land taxes (which effectively decreases the net rental price of land).

The table shows three scenarios where the discount rate is 5%. In each scenario, the price in time one (t=1) reflects the expected rate of growth. The present value (PV) is the price at t=1 discounted at the 5% rate. Where that present value is higher than the current price, there is an incentive to delay sales, which feeds back into delayed construction [1].

If the rate of price growth is higher than the discount rate (the rate of return on the sale price available from investing it elsewhere) it makes sense to delay the sale to get the higher price (Scenario A). If the rate of price growth is low, there is an incentive to bring forward sales to get your money out of this property to put it somewhere else an get a higher return (Scenario C).



The property industry likes to promote the myth that they would never delay selling. Yet, when I worked for a major property developer during a price boom period, we did exactly that. The decision was made to close the sales office one Saturday because there were too many sales. These rapid sales meant that the price was too low and that delaying the sales would fetch a higher price (and a higher PV of that future price). So instead of selling the whole building in one day and starting construction, the prices were raised, and it took years afterwards to sell the whole building and massively delayed construction.

The absolutely crucial lesson in from the Scenarios in this table that the imposition of a developer charge can turn Scenario A into Scenario C by reducing the net revenue from each future dwelling sale to a developer due to the charge. For example, if a charge of $10,000 is announced to be imposed in the next financial year in Scenario A, it becomes Scenario C in net terms, and the developer will prefer to bring forward planning applications to get a lower charge and incur sales in the current period.

Increase taxes on land to get more construction, not less!

To be clear, this is not some crazy idea I just invented. This is the standard result of real options theory, and it applies equally to increasing costs to landowners and decreasing their future development options. Here’s a 1985 paper from the AER making the point.
… the initiation of height restrictions, perhaps for the purpose of limiting growth in an area, may lead to an increase in building activity in the area because of the consequent decrease in uncertainty… 
Imposing height restrictions can turn Scenario A, where future revenues (price x number of dwellings) are higher because of the option for increased density, to Scenario C, where future revenues are lower because the number of dwellings able to be built on the site is fixed.  This brings forward sales and construction.

In sum
My new paper is a small contribution that demonstrates the well-established economics of property markets, but which flies in the face of conventional theory. Understanding land and property markets helps to understand how backwards the standard economic understanding of ‘capital’ really is.

fn [1]. Another thing many economists get wrong about the property market is they ignore the fact that most sales come before construction, not after. This means that when people just say “increase supply” they don’t realise that market incentives mean this will never happen — supply only responds to demand. Only a housing developer without a profit motive would increase supply at a rate that would depress local prices, and yet we hear nothing from the ‘supply-siders’ about the creation of a public housing company that could do just that.

2 comments:

  1. Sorry Cameron, I don't agree. An increase in taxes may temporarily reduce the land value, but not in the long term. It will make developing new apartments or subdivided land more expensive and raise the market price of existing units/land as they compete in the same market.

    ReplyDelete
  2. This makes some sense. In big cities like NYC, land owners often build a small building, much smaller than could be built by right, called a "taxpayer". The idea is to invest as little as possible to produce just enough rental income to cover the cost of taxes, maintenance and so on. The landlord expects the value of the land to rise as a general trend over the years and decades. When the land is valuable enough, the taxpayer is torn down and replaced with a more substantial structure. This works because land in cities is usually taxed based on what is currently there, not on what could be there.

    On the other hand, investors do often build on spec. I live near Seattle, so I've seen an entire neighborhood go up with thousands of apartment units. Many were sold to investors, not necessarily tenants, before construction was started, but the sheer number of units has slowed the rise of housing prices in the face of a booming economy. I have friends who have rented though they can afford to buy since the small investors who bought the units need someone to cover the taxes, fees and maintenance. The idea is that as supply is absorbed and prices rise over the years, they will be able to raise the rent or sell at profit. In the interim, they need a "taxpayer" to cover their costs.

    ReplyDelete