Why Your Supply Elasticity Number Is Probably Wrong
The same cities can look like over-regulated one decade and free-market miracles the next, simply because the housing market moves in cycles
The global housing debate is dominated by the idea that planning regulations affect the way a city’s housing market responds to changes in demand, leading to the wrong number of homes being built.
Sydney. Vancouver. Los Angeles. You name it.
Want to dive deep into the economics of land and housing?
On the 9th and 10th of June 2026, I will present a two-day workshop in Brisbane—Ground Rules: Land & Housing Economics for Professionals.
Get your Early Bird tickets and peruse the workshop details at the Eventbrite booking page. Make that professional development budget work for you!
Discounts are available for annual paid FET subscribers and groups of three or more.
All enquiries to cameron@fresheconomicthinking.com.
Those regulations might include zoning that limits density in high value areas, development charges on new builds, NIMBYism and its legal challenges, or delays in permitting or approvals. Whatever they are, they are widely thought to cause a slower rate of housing production and higher rents and prices.
In the economic jargon, these regulations are thought to make housing supply less elastic, meaning that for any change in demand leading to a certain price increase, fewer new homes are built.
As has been pointed out many times here at Fresh Economic Thinking, this particular story is full of common misunderstandings about how both how town planning operates and the fundamental economic forces in property markets.
The point I want to make clear today is that if you try and measure supply elasticity at different times and over different periods, you can get very different estimates under the same regulatory conditions.
Instead, what you are measuring are predominantly the trends within a longer-term housing market cycle.
These figures are therefore not very informative and potentially highly misleading.
Using short time windows to measure supply elasticity is a common error. This is why the more detailed academic studies uses periods as long as possible for their analysis, with many papers using all the comparable data from the first reliable collections in the 1970s and the subsequent four or five decades, with some even separating our boom and bust cycles because of their large effects on measurement.
To get some clarity before we dive in, a confusing reality is that a variety of different things are called supply elasticity. Two main ones are flow elasticity, and stock elasticity.
Flow elasticity is the rate of change in the quantity per period of new homes built in a region divided by the rate of change of price. So if the rate of new homes built per year rises from 1,000 to 1,500 when the prices increase from $100,000 to $120,000 then the flow elasticity (using the midpoint formula) is (500/1250)/(20,000/110,000) = 2.2.
Stock elasticity is the rate of chance in the stock of dwellings divided by the rate of change of price. So if the number of homes in an area rises from 1,000,000 to 1,100,000 over a ten year period when the price has increased from $100,000 to $120,000, then the stock elasticity is (100,000/1,050,000)/(20,000/110,000) = 0.52.
These are not comparable with each other.
Indeed, it is also not clear why the stock of homes should be related in the long run to the asset value of housing, and not the rental price, as asset prices are heavily affected by interest rates and tax settings.
When measuring flow elasticity, lags and time period choices matter a great deal, as does the choice of controls (such as for population or income).
Another problem is that the flow of new housing might be low in one city, but the variation of that flow in response to price might be very high, leading to a large measured flow elasticity in a city with hardly any new homes being built and hence a low stock elasticity.
Consider two cities. City A builds 1% of the stock per year on average, but 2% in a boom a none in a bust. City B builds 2% of the stock per year on average, but with hardly any variation in this flow rate (say, 2.2% in a boom and 1.8% in a bust). If they both have identical price patterns, City A will have a high flow elasticity and low stock elasticity, and vice-versa for City B.
The point is that the elasticity measurement game can reveal your assumptions as much as any real underlying economic relationship.
Here I demonstrate the effect of the market cycle on supply elasticity measurement. To do this I, calculate the stock elasticity based on the price of homes being the inflation-adjusted asset price over two different recent time periods at different points in the housing cycle in cities in Australia, Canada, and the United States.
Using stock elasticity should be less sensitive to short term variations and issues about lags and so forth, but is more open to issues about endpoint selection (which is very much the point of this exercise!).
This exercise demonstrates how problematic it can be to use short term estimates (or changes in them) to guide policy advice.
Australia
The chart below shows our simple stock elasticity of supply estimate for the three biggest Australian cities over the 2011-18 period in blue and the 2018-25 in yellow.
From the perspective of 2018, looking back over seven years, you would see that Brisbane had the highest supply elasticity. You might advise other states to copy Brisbane’s zoning regulations to get higher supply elasticity.
But then, from the perspective of 2025, looking back over seven years, you would see that Brisbane now has the lowest supply elasticity. You might at this point advise Brisbane to copy the zoning regulations of Sydney or Melbourne to get a higher supply elasticity.
Canada
It might not surprise you but similar cyclical variations arise in Canada too.
From 2010-16, our results are similar to an influential CMHC report that looked at a long expansion period from 1992-2016. Vancouver and Toronto appear to be supply laggards, with a low measured stock elasticity of supply.
But what about in the years since? These years differ in that they have see two periods of substantial real price declines (2016-20 and 2022-25).
We see complete reversals in measured flow elasticity of supply.
Toronto was inelastic and should have learnt from Montreal in the first period, but then if you trust these measurements, they are now supply leaders that Montreal should learn from.
The negative result for Vancouver in the second period is due to the fact that real prices fell 6% while the stock of housing grew 17%. Wow!
Of course, economic theory has trouble dealing with this unless we invoke a supply shock, which moves out the entire supply curve. So I don’t think it is wise to look for an interpretation beyond the data capturing cyclical changes and asset prices being a poor metric for the price of housing services.1
A study from New Zealand found an unexpected negative relationship between population growth and dwelling stock growth using data from 2012-23, then simply threw the data out because it was inconvenient to the economic story they were trying to tell. Read more about that here:
United States
Lastly, a quick look at the United States.
We see a large cyclical gain in measured stock elasticity of supply in Los Angeles and most other cities in the second period, and we see the famous Texas “supply leaders” with higher overall elasticity.
But again, we need to recognise the cycle.
Prices have fallen dramatically in Austin since they peaked at the end of 2022. Had we only looked at the 2018-22 period, then its measured supply elasticity would be back at 0.6, and not far out of line with many other cities.
I know many readers might want to bail out these measure now by thinking “Yeah, but Houston and Austin seem always to be higher than the other cities, so the measure must be telling us something!”
And I would caution that you are seeing only the patterns that confirm your beliefs, as the data also requires you to believe that in 2016 Vancouver suddenly became even better at building homes than Austin.
So what?
Taking measures of supply elasticity as a proxy for regulation without recognising the cyclical nature of property markets is a bad idea. Cycles are long, price corrections can be swift, and different cities are often in different phases of their cycle at the same time. Interpreting these numbers as key indicators of the effect of planning regulations is a big leap.
Your choice of time period, start and end dates, as well as data sources, all matter. You might use asset prices, of which there are a variety of indexes, market averages, all of which don’t quite match up. Usually you would adjust them for inflation. Or you can take rents as the price of home, and deal with the variety of different rental price metrics in each city (I explain the difference between advertised rents, paid rents, etc in the below downloadable Explainer).
Each data source will also have a different slightly different cyclical timing.
What we know is that when we zoom out over time and space and control of demand variation, all cities follow the same basic economics of housing production, including the emergence of price cycles.
A recent paper by Federal Reserve Bank of San Francisco researchers documents this comprehensively, finding that there is essentially no difference in the rate of new housing production in response to similar changes in demand over a 40-year period across major United States metro areas.
Here’s a detailed explanation of that paper and its methods.
Ad hoc approaches measuring supply elasticity over short periods within a longer cycle (that can last two decades) have been used to justify a raft of planning policy changes—some good, some bad, and some involving billions in value created and given away to connected political mates.
We should argue about planning regulations on their own local merits. But when we do, we need avoid jumping straight to elasticity-type reasoning based on arbitrary measures.
As always, please like, share, comment, and subscribe. Thanks for your support. You can find Fresh Economic Thinking on YouTube, Spotify, and Apple Podcasts.
Interested in learning more? Fresh Economic Thinking runs in-person and online workshops to help your organisation dig into the economic issues you face and learn powerful insights.
I agree that exact elasticity numbers are meaningless but it's hard to deny that cities without zoning rules, with large unincorporated areas within urban area, easy county permits and low state regulation, keep long term housing prices not only lower but also less volatile (not without volatility because most volatility is driven by demand side - easy credit). Elasticity is not a silver ballet to stop volatility or high prices but the very idea in peoples heads that if they need or want they can build a house nearby helps to some degree to reduce FOMO which is one of main drivers of boom bust cycles (in addition to easy credit which is an enabling factor).
Sorry, replied to the wrong comment...