Monday, November 22, 2021

Alice in housing economics wonderland

“When I use a word,” Humpty Dumpty said in rather a scornful tone, “it means just what I choose it to mean — neither more nor less.”
’The question is,” said Alice, “whether you can make words mean so many different things.”
― Lewis Carroll, Through the Looking Glass

An increasing academic and policy focus on housing supply has unfortunately not brought with it an increase in clarity over the meaning of words. Words like housing supply, demand, zoning, price and affordability, have come to mean whatever the authors want them to mean. 

This is unhelpful. 

I've put together in the below table various potential meanings of these catch-all economic terms, and the words I think we should use instead to increase clarity. 

For example, economists should be crystal clear that the economic price of housing is the market rental price (i.e. the price that  measures the amount of good and services given up to get that good). The sale price of  a dwelling is merely the market’s judgement of the asset value of buying a rental income stream in perpetuity and is heavily affected by prevailing interest rates (capitalisation rates), land and property taxes, and expectations of changes to the asset value (capital gains expectations).

With such a variety of definitions, what can the phrase “an increase in the supply of dwellings relative to demand will reduce dwelling prices” actually mean?

The obvious and strictly true definition is when the terms mean

supply → market willingness to sell housing in this period,
demand → market willingness to pay for housing in this period, and
price → the sale price of a dwelling in this period.

This merely describes asset market bid and offer schedules. When there is a "supply shift" in these schedules that massively reduces prices, we call that a market crash (a topic that is also poorly understood). But the economics of housing stock change is more sophisticated than this.

Supply and demand might also mean number of dwellings and total population—completely different ideas from bid and offer schedules in asset markets, instead focussed on material quantities that are expected to move together. 

Sometimes supply is used to mean zoned capacity (how much can build built within current zoning rules) and is often assumed to be synonymous with the absorption rate (how quickly the market will develop new housing subject to asset market conditions). Changing zoned capacity does not necessary change the absorption rate, yet they are often described as one and the same thing.

Perhaps if we could describe what we mean more specifically we can start to allow the evidence to support or disprove theories about how the housing and property market operate. 

I'm open to improving and expanding the table with better, or more widely used, terminology and will update and refine it over time. 

Tuesday, November 16, 2021

Opening remarks to housing inquiry

A video of my testimony to the inquiry is here (from 11:00:00)

My full written submission is here and my follow-up response to questions raised is here.

I believe I am one of the few witnesses who has worked for residential and industrial property developers, in government departments dealing with infrastructure charges and regulation, and now as a housing researcher. 

To be clear, Australia has more, bigger, better dwellings per capita than any point in history. We are also building new dwellings at a near record pace in a period where population growth is the lowest in decades.

More housing is better than less housing. Absolutely. I agree.

The argument I disagree with is that private landowners want to build faster, but only pesky council and state government red tape is slowing things down.

While I certainly have many ideas for improving and simplifying the planning system I see this as a separate topic to housing affordability.

We’ve heard from previous witnesses that housing developers have a lot of trouble building on unzoned land. No doubt. The whole point of unzoned land is to not have development at that location and get it located in the zoned land. It’s hardly evidence of anything except that these developers are bad at their jobs, always buying the wrong land for what they want to build.

Indeed, they certainly appear to be terrible lobbyists. If what they claim is true about zoning keeping prices up, then lobbying for mass rezoning is financial suicide. It would vastly increase the number of competitors in their market and reduce prices, wiping billions in value from their balance sheets. What sort of industry lobbies for that?

Perhaps this story is a lie.

In 2003, the AFR ran the headline “Brisbane running out of land for housing”, with land for housing expected to run out by 2015 according to the same expert lobbyists who have attended this inquiry. Yet detached housing lot production was 30% higher in the 6 years since 2015 than in the 6 years prior.

Are they terrible at their jobs? Or just telling stories that conceal the true nature of property markets?

Remember, only landowners can choose to make planning applications. Only landowners can choose when to build homes. Councils don’t do this. There is no speed limit to building new housing in the planning system. Planning regulates the location of different uses and densities, like road lanes regulate locations on the road. Density (dwellings per unit of land) and supply (new dwellings per period of time) are completely different concepts. More density does not equal faster supply.

The key issue at stake in this debate is that land is an asset. It is therefore priced like one. This is why when it is a good time to buy it is also a good time not to sell.

This is true for developed and undeveloped land. Stocks of undeveloped land sit on the balance sheet of developers, earning a return by growing in value while undeveloped.

The trade-off between the return from delaying developing land, versus developing now, creates a built-in market speed limit on the rate at which private landowners develop. As a previous witness mentioned, “… if you are a property owner or developer that had land that was consented and you hadn’t sold it a year ago then you are in a very strong position.” Builders might like to build faster, but landowners prefer to maximise returns on their assets. 

We can see how this pays off with a case study of Jordan Springs, a Lendlease subdivision of around 2,000 housing lots in Sydney that took a decade to sell. I looked at the sales rates over time and saw that the average rate was only 45% of the peak rate (3-month average), though some periods had sales just 12% of the peak rate. The speed of developing new housing lots was far below the capability of developers and the capacity of zoned land. By selling at this slower rate, and capturing overall market gains in the form of higher prices, they made an additional $137 million on the project. Prices were 31% higher at the end than the start for land lots on a per sqm basis.

It would have been financially irresponsible of them to develop faster than they did.

I’m not saying that this behaviour is wrong, or a conspiracy, or even that it has major price effects. The stock of housing only changes a couple of percent a year at the best and small changes to those small changes make tiny price differences. This is just normal market behaviour. This is why for the century prior to the existence of zoning we had the same issues of unequal access to land and housing ownership, only much worse.

The current housing asset price boom is a global one. Average prices are up 20% in the last year in the US, the same as Australia, and many places that were previously lauded as having flexible zoning, like Germany and cities in Texas, have had the highest price growth. What we are seeing is a period of global asset re-pricing, as intended by monetary policy.

If you really want more homes build them. Flood the market with a public housing developer—you know, just in case the private developers don’t do what they said they would. It might be a sensible insurance policy. No doubt the property lobbyists will find something wrong with this, even though it is exactly the outcome they pretend to be lobbying for—more competitors and lower prices.

Wednesday, November 3, 2021

Public housing is way cheaper than rental subsidies

A discussion about the best way to provide below-market-priced housing popped up on Twitter recently. Peter Tulip noted many of the limitations of such systems—queuing, quotas, qualifying criteria, etc—concluding that a cash payment to help pay market rents is an economically-efficient way to get the policy outcome of reducing housing costs to low-income households.
I am not against providing such cash payments. They are clearly better than nothing. But the reason I believe governments should build and own some housing is that it provides a better bang for your housing subsidy buck.

Consider the two alternatives over a “tenant life” of say 30 years.

With cash rental assistance, the government pays, say for the sake of argument, $13,000 per year the first year. But to have a meaningful effect this must grow over time to reflect growth in rents and incomes. At a 2% growth rate, by year 30, the subsidy is $23,000 and over 30 years the total subsidy paid is $527,000. The present value of this 30-year flow of subsidy payments at a 2% discount rate is $374,000.

With public housing ownership, the government builds or buys a dwelling worth $500,000 today to supply that dwelling at a rent that is currently $13,000 below market rent per year (i.e. the same rental subsidy to the resident). The remaining rent paid by the tenant covers ongoing costs only. Like the cash rental subsidy, the gap grows over time to be $23,000 in the 30th year. Instead of $374,000 in present value terms, this option costs $500,000 today to build or buy the dwelling (much less if built on under-utilised publicly-owned land).

However, with public housing ownership, a government agency owns the property at the end of the 30 years. Over this period, the asset value grows. Even if it grows in line with the 2% growth of local incomes it means that the property is worth $890,000. In reality, because incomes at a location rise faster than the average (because cities expand), it is likely to be more. For reference, this is only a 76% rise in three decades; a conservative figure when compared to the 143% price rise seen in Australia’s capital cities in the past 18 years. 

The table below shows a comparison of the two alternative ways of providing the same value of housing subsidy to a resident over 30 years. Although the public housing ownership option costs $500,000 upfront, today's value of the final sale price is $490,000, leaving a net economic cost of just $10,000. This approach gets 40x better value for the budgetary spend. If capital growth is closer to historical norms then public housing can more than pay for itself.

What we learn from this is that 
  • The cost of rental assistance over the long term is not much different than simply buying a dwelling and giving it to the household ($374k vs $500k). 
  • The cost of rental assistance over the long term is much more than providing the same rental subsidy via owning the property ($374k vs $10k) 
  • Getting out of the housing ownership game over the past three decades and shifting towards rental subsidies has cost government budgets billions.
[UPDATE] I've updated the figures to reflect 2% growth of incomes and rents and 2% interest and made the spreadsheet available here. Play around with the numbers. 

[UPDATE] People seem to think that interest payments need to be taken into account somewhere. They do not. Prevailing interest rates are incorporated via discounting. 

[UPDATE] Thanks to Jago Dodson for letting me know that a 1993 review by the Industry Commission (now Productivity Commission) ranked public housing first in terms of efficiency and cost-effectiveness out of a variety of alternative housing subsidy approaches they assessed. 

[UPDATE] Thanks to Vivienne Milligen for letting me know that the 1989 National Housing Policy Review found similarly—that public ownership of housing is the lowest-cost strategy for housing poverty relief.