Prosper’s Director of Research and Policy, Tim Helm, was interviewed by Cameron Murray for the Fresh Economic Thinking podcast on 9 November 2023.
Their conversation covered the relationship between the prices of land and housing, how to define and measure landbanking, incentives for landbanking, the relationship between landbanking and housing supply, and methodological issues in the leading study of upzoning in Auckland.
Audio is at fresheconomicthinking.com and the transcript of the interview is below.
Cameron Murray:
I’m Cameron Murray, founder of Fresh Economic Thinking, and I’m here today with a special guest, Tim Helm.
Like many of my guests, Tim found himself in a “barney” recently about some of the finer details of economic analysis, this time around zoning and house prices.
Tim and I have worked together on this topic and on a few other projects, and I thought it’d be interesting to get him on the show and pick his brain a little about where he’s come from intellectually, where he’s going, and how our paths crossed.
I don’t think we agree on everything, so there might be a good story here. Tim, welcome to the show.
Tim Helm:
Thanks, Cam, and good morning.
You’ll have to alert me to which barney it was you’re referring to. As I’m off Twitter I don’t engage in quite as many as you do – but there have still been a few.
Cameron Murray:
I’m talking about the Auckland paper… and one thing Tim has taught me is how to have barneys in a much more polite way.
But maybe we should start a little bit in the past.
Tim has a PhD in economics, and his PhD is called Essays on the Economics of Price Transmission, which sounds pretty… boring? What’s the right word? Pretty… straight?
Tim, how did you decide to do a PhD on that topic? Where should we start to understand how you ended up there, and how you got to where you are now?
Tim Helm:
I think it’s a great place to start.
It’ll take us right through to land and housing economics, because the title of that thesis is the title you put on a thesis when you’re wrapping a bow awkwardly around a package with a lot of lumps.
That’s because about halfway through my PhD I came across the economics of land tax, got particularly interested in that, and wrote a couple of chapters on it, particularly focused on the question of how a land tax is passed on – is it passed on to rents or passed back into land?
One of the reasons I happened upon this topic and puzzled so hard over it was that the first half of my PhD had been looking at abstract models on price transmission in supply chains. When the oil price goes up or down, for instance, how quickly does that flow through to the petrol price?
Through being engrossed in the models in that field, and in their particular application to modelling asymmetry in the rate of price transmission, I had adopted a mental worldview about how the world worked which became my lens on all markets. But that lens didn’t particularly fit well with many situations.
I had typically thought about costs as fixed, and prices as reflecting those costs in a “building blocks” manner – a linear causality if you like.
So, naturally, when somebody gave me a brochure about land tax, explaining that a land tax wouldn’t increase rents, I thought “well, they don’t know the first thing about economics”.
And that was a little gateway into the long history of thinking about land tax, and led me into my subsequent career working mostly on tax policy and on land and housing markets.
Cameron Murray:
I forgot to mention at the beginning that you are now the research director at Prosper Australia.
Is it my understanding that the flyer you’re talking about on land taxes was actually from Prosper?
Tim Helm:
You’ve got it exactly right.
My PhD is from Melbourne University, and I was living in Melbourne at the time and came across this think tank or NGO by chance, and their literature on land tax led me into getting interested in that question.
That was more than 10 years ago, and I’ve stayed loosely attached to Prosper ever since, while working in the state government of Victoria, and consulting in Victoria. I’ve been part of the informal group of advisors that help shape some of Prosper’s ideas and research directions, and as of this year I’ve taken on a part-time role as the research and policy director for Prosper.
I should introduce Prosper a little.
They’re a Georgist think tank, which means they’re from the tradition of spruiking the ideas of Henry George, which were focused on putting land tax at the centre of tax systems – focused, really, on the capture of rents in the economy, particularly those from land.
We at Prosper do a lot of policy advocacy particularly around increasing the share of property tax, as opposed to taxes on productive activity (on labour, etc). We also do a sideline in research, because you can’t just keep banging the drum yelling for a land tax every year before people tune out. We have capacity to do some original research, and that’s often focused around the nature of land markets and housing markets.
Cameron Murray:
How long ago was it that you did this PhD, and realised that you had this building blocks idea of prices that you maybe needed a different lens for?
And what is this building block idea of prices?
What changed, and what happened in the intervening years, to get you from there to here?
Tim Helm:
Great question. I see this “cost plus” idea – this assumption that many people have about prices in the economy – as a great big cancer of the mind which affects so much thinking around housing and tax policy. I was in that mindset, with those models of how fast a cost shock flows through. I took the cost side to be, as we say in economics, “exogenous”, or fixed, something determined in a market somewhere else, which we then see flowing through to prices.
I think many people hold this worldview. They instinctively think linearly – that if you add up a bunch of costs, and a profit margin reflecting the competitiveness of an industry, you end up at the price.
In many circumstances that’s a good enough heuristic. But the case of land is a special one, because land has no cost. It’s already there, and so the cost to use it in production, or to use it at all, is zero.
What do you sacrifice by using land in 2023? Well, you can’t store up those services for 2024, or bring them forward. Those services flow steadily with time. So it’s costless to use land.
Yet land has this high price, when we’re talking of urban land, and of course that price is causally determined backwards starting from the prices we pay to use land in ways such as housing. So that rule of thumb of pass through, where a cost in a supply chain causes a price, that way of seeing the world is wrong when it comes to land.
We should realise the reverse nature of it, and think about the price of land as reflecting the leftover – the residual – after we take the value of what we can do with land and net off the costs associated with the doing of it.
It’s a really old insight. It was second nature to economists for several generations, but somehow in the training that I’d gone through right to PhD level – and perhaps this is a reflection more on me than on the training – I’d missed the central point that land has some different attributes.
Certainly labour and capital prices are also determined in part by the value of what they produce, but land is particularly special in that regard, because it’s already there, and it’s fixed in supply.
That’s the directionality point that I’m talking about. That was an eye-opening change and I suppose that’s also how it led me into housing economics. Working in tax policy and government I would see this misunderstanding all the time amongst policymakers, and exploited by lobbyists.
Cameron Murray:
It’s quite interesting that you explain it and encountered it this way, because my background is in property valuation and property economics.
So when I studied economics, I was doing the reverse transition, starting with this residual idea of the value of land – that the value of land is determined by what people will pay for use at that location, minus the costs of investing on that land to get that use. The residual, or what’s left over, is the value of the land.
It’s interesting that you made the opposite transition.
I started with this idea, and have been banging my head against the wall when it comes to communicating it.
Tim Helm:
I think of this as learning it the hard way, in the order I learned it. Your mental model, that you started with as a practitioner, matches the economics better.
For me learning it the hard way was a humbling lesson. When you feel like you understand the world, there is no feeling of internal inconsistency. But then a better point of view comes along, and you are wrenched from one view to the other, and you realise that your prior feelings of internal consistency were no guide as to whether you were correct.
I think that lesson stuck with me through time.
Cameron Murray:
One of the puzzles I had with this reasonable view about where the value of land came from is that I could never explain why it might be optimal to own land but keep it undeveloped.
Because if the value of the land comes from using the land for something, subtracting the cost of constructing that use on the land, why would I not construct that use? Why would I own the land but never use it?
I remember during my PhD walking around the halls at UQ knocking on doors of economists and saying that I’ve got a new model about why people might not use land.
Everyone said “no, that doesn’t make sense – you would just instantly invest in land”. And I’d say “look out the window!”.
So I had a transition myself when observing the world and thinking “this model of the residual value is incomplete, because it doesn’t tell me why I would never use land”.
I looked at things like landbanking, and I used to work for property developers. We would own land and never seem to be in a hurry until we started paying the builders. There was always a patient period of thinking about what we should do. So I had a similar change in that way.
I want to talk about spatial equilibrium, which is an idea on which you influenced me quite heavily.
But maybe we’ll start with landbanking.
How did you come to this idea, and what do you make of it now? Because I think it’s super important for current policy debates. How would you explain this to someone on the street?
Tim Helm:
I think it’s really important in under-recognized ways for policy debates about housing.
So I’ll have a crack at selling the importance of landbanking.
What’s interesting is that there has been a parallel conversation where, over in the development sector, landbanking is an obvious thing. It’s a business strategy, it’s what everybody does, and you can measure it. There are all sorts of empirics, including some that you’ve done, that show land being held for a ridiculously long time compared to what it takes to get it ready for sale.
But then over in the economics profession, and in many of the fields of policy analysis that rely on that mainstream economics, it doesn’t really exist. There’s no place to slot it into the mental models.
The classic example in this regard, if you’re in Australia, is the Productivity Commission’s report from last year on the National Housing and Homelessness Agreement. It had a box about landbanking, reasoning that that it couldn’t exist. And it was really a master class in motivated reasoning. You know, “I don’t understand how this empirical phenomenon fits in my model, so I’ll work really hard to rationalise why it doesn’t exist”.
I think the advice for them is the same as you would have given to your professors: let’s just look out the window, because we can learn from empirics when doing science. We don’t need to limit ourselves to model-based exploration to find the truth.
To get to your question – why is landbanking important?
Cameron Murray:
Maybe we should define what it is, because I think some people call landbanking two things.
One is buying land, knowing it can be developed, for the return prior to development.
And one is the fact that all land is owned by someone, whether it has development potential or not, and whether their intention is to develop or use it for its current use.
So how would you define this clearly as a phenomenon? Versus the fact that development opportunities exist, and all property is owned by someone; therefore any land owned by someone that could be developed and isn’t is being land-banked.
What do you think?
Tim Helm:
Yeah, it’s a great question.
I think we’re working it out on the fly, because there isn’t really a textbook definition of landbanking. So I think it’s incumbent on us to create definitions that are useful and meaningful for analysis.
I’m also going to throw in the word speculation, which has fuzzy meanings as well, and also gets used in everyday terms. Sometimes speculation is defined in terms of the intent of the speculator.
So I’m thinking about these two words as needing or benefitting from tighter definitions.
What I’m interested in, and where I think the right definition of this lies, is that I want to know about land that is already feasible to redevelop.
The classic example is farmland on the edge of the city that you can sell and someone can build a house on, and where all you’re losing is the value of running a few sheep on it. It’s clear on a development feasibility calculation that there’s a new use that is immediately more profitable than the old use, and that is to build a house today.
That’s the feasibility test. So I define landbanking as being the holding of land in undeveloped use or less developed use that is already feasible – that is, profitable – to develop to higher intensity use.
That’s a definition that doesn’t rely on the intent of the operator, and I think that’s useful if we’re going to make sense of it.
Why I think that’s a useful definition is that we’re engaged in a lot of these policy conversations around the contribution of zoning and other policy levers towards housing development and towards house prices. And a really key question is “how much land is already sitting there ready to go, profitable to develop but not being developed?”. Because there is this strategic delay in developing – private sector behaviour which is nothing to do with public sector rules – which is delaying development.
We want to be able to quantify that, so I think the test of how much land is already feasible to redevelop is a good one.
Whether we call that landbanking, or feasible capacity, or something else, is maybe a secondary point. But I think it’s just simplest to define landbanking this way.
But I’m curious also about your thoughts. Is this how you conceive of landbanking – i.e. that a feasibility test passes yet it’s not being developed yet?
Cameron Murray:
I’ll be honest: you just explained it how it is in my head, but I’ve never been able to articulate it as well you just did. For listeners, this is a common theme: I have an idea, I explain it badly to Tim five different ways, and then he explains it back to me very clearly and concisely.
So if something like that is valuable for your organisation please give Tim a call!
I think you’re exactly right. For example, a lot of people might have a large house lot that they could split and subdivide. At a market price today, the value of that house lot is higher as a development project than as a single home. Another simple case of landbanking is that they’re not doing this.
And I don’t think mind-reading the intent of property owners is a good way to understand a phenomenon. People don’t really know what they think quite often. So I think if it’s feasible to develop, and the property owner is choosing not to develop today for any reason – whether they don’t have the money, whether it’s never crossed their mind, whether they don’t realise, whether they’re waiting with the intent to make more money in the future, whatever they are doing – then they are waiting, and they are banking that land, and this time dimension to development is being exposed.
Another point you made back to me is about the incentive. I think a lot of people don’t understand the incentive. Can you articulate why it might be better, even for the person who understands the payoff today, to not develop today, and instead hold that sheep farm or big housing lot and wait for the future?
Why is it better to not have cash today? What’s the trade-off?
Tim Helm:
I’ll acknowledge that your thinking has informed mine heavily on this, even if I express it slightly differently.
Although I’ll touch on one more point first, which is this: why do we care about landbanking?
I earlier said it’s important for knowing how much the private sector is contributing to slower development. But I also want to join this conversation about landbanking to policy conversations about housing supply and affordability.
Because what it is to supply a house is to end landbanking on a site.
Suppose you’ve got a site. It’s feasible to develop. You’re not developing it, you’re not developing it, you’re not developing it… and then one day you develop it. That means you end landbanking – and it also means you supply a house.
We hear incessantly about the importance of housing supply, and yet people don’t understand the importance of landbanking.
I just want to say: this speculation, this landbanking, is the opposite of housing supply.
So if you want to understand where housing supply comes from, you’ve got to understand the incentive to landbank.
Cameron Murray:
The line I think I’ve heard you use is “supply starts when speculation ends”.
You are taking this timeline of feasible sites, and at one point in that timeline you build a house. Before that, you’re speculating – you have a feasible site that you’re choosing not to develop yet.
I thought that really captured the idea pretty well.
Tim Helm:
I hope so. That’s how I see it.
Now the question you raised is “what is that incentive?”. Why land bank? Why not just build everything immediately?
The answer is that producing housing services is not like producing haircuts or lettuces – something you just bring to market with no long-term investment involved.
The nature of land development is that you put structures on a site. You commit the site to those structures for the economic life of those structures. And that’s effectively irreversible – or it’s economically irreversible, in that it doesn’t make sense to destroy capital once it’s on the site.
Those two things give rise – as you’ve expressed very well numerous times – to an option. You have a real option, and that has value. The option means you’ve got the right to develop the site but not the obligation.
That has value, because one thing you could do is develop the site today, but you may be able to do even better than that by not developing it today.
So the answer I give to people who say “why would landbanking arise?” is that even when development is profitable, not developing might be even more profitable. Obviously that means developing tomorrow.
Why would that be more profitable?
One reason is that prices rise through time, very often, in many circumstances. As incomes rise, what people can pay for rentals rise, and so what it’s worth an investor to buy a rental property rises. In various ways the price you can sell a built house for rises through time.
It may be that if the price is rising faster than the return on (or borrowing cost for) the capital you’ve used to finance the land then it makes more sense to build and sell the house later.
The other point that you’ve raised is that if you want to sell more today – if you already have some sort of rate at which you’re supplying housing today – then you might have to discount the price below what you had anticipated or hoped to get. And that might mean you’re selling it for less today than you could tomorrow.
By the same token, we can look to the construction side, the cost side, as another reason why people might landbank.
In certain markets, like the present construction market, or particularly in 2021, if you want to build more today you’re going to have to bid some construction resources away from someone else, which means paying over the odds for those resources. You’re paying a bit extra today, which is also an incentive for more resources to move into the construction sector tomorrow.
So you’re changing the relative cost of building today versus tomorrow, and you’re making that “slope down” a bit – you’re choosing to build in a way that lowers future costs, but you’re exposing yourself to the high present costs of construction. It’s the flip side to selling a bit more today and therefpre causing the gradient of price over time to slope up.
So there are two factors: you have to discount your prices if you want to sell a bit more today, and you have to hike up your construction costs if you want to construct a bit more today. That leads you to a balance where it makes sense to delay developing your site (or developing all your sites). And I think in essence that’s how I understand the incentive for landbanking: you have this balance that reflects the trade-off between profits from landbanking and profits from developing, and there’s an equilibrium rate at which you want to supply housing
This is how I see that incentive.
Cameron Murray:
I think from different directions we’ve come to similar views on that, and informed each other’s thinking, and I think you articulated that well. I want to bring that together to this barney that I flagged.
Before we get on to Auckland, I want to say that you also alerted me to an article by Eric Crampton, a New Zealand economist, who articulated the same intertemporal trade-off that you just explained for landbanking when it came to carbon credits in New Zealand.
There was a system of carbon credits, where if farmers wanted to plant trees they could get carbon credits for them. And there was a concern that if you just let any old property owner plant trees and get carbon credits, won’t they flood the market today, and plant heaps and heaps of trees until the price of carbon credits goes to zero?
Eric Crampton really nicely articulated in an article that it’s not logical to flood the market and undermine future profits by planting lots of trees today.
And so on my Substack I changed all his wording from carbon credits to property, and trees to houses, to use his explanation to show why the same logic applies to housing.
He responded and for some reason didn’t think it applied when the thing you put on your property is houses, but does apply when the thing you put on your property is trees.
I think you got me into that barney… but I want to move on so we can wrap up.
Tim Helm:
I think Crampton’s expression amended by you is one of the clearest expressions I have seen of this intertemporal trader, so I highly recommend everybody has a look at your column.
Cameron Murray:
He explained it so well, and that’s a funny thing in economics, right?
People can see this logic clearly when it’s politically convenient for them, or when it suits their preferences or their tribe or their social group.
And when it doesn’t? It’s like “oh, I don’t know what you’re saying – it’s impossible that that could be true, because all my mates are saying the opposite”. Which as you know is a recurring theme in this podcast.
Now, Auckland.
Auckland upzoned because they thought property owners would panic and flood the market with new homes – because it wasn’t the intertemporal trade-off that made property owners wait, it was regulations that forced everybody to wait.
A couple of economists in New Zealand looked at dwelling consents, which are the equivalent of building approvals, in Auckland. Their analysis found that there were an extra 22,000 dwellings in Auckland in the four or five years after this upzoning.
I initially had some concerns about the data simply because the numbers were so huge. I had a ballpark idea that with 22,000 extra the graph should look different.
Tim actually read all the appendices, and replicated the analysis, and explained to me some of the conceptual errors.
Tim, why did you read it and replicate the analysis in such detail, and what did you find?
Tim Helm:
As with you, for me it didn’t quite pass the sniff test. The magnitude of the number being claimed, 22,000, pointed to something like half the houses built in Auckland since the Unitary Plan only being built due to the Unitary Plan.
I’ve read through the guts of this paper and tried to explain the methodological errors made by the authors, but I think those explanations are really complex. So I’ll start by just posing a litmus test for listeners, and if it doesn’t pass this sniff test, we can go through why the methods have led to this wrong conclusion.
The litmus test was this. When you’re evaluating the impact on a certain variable – permits – of a policy intervention, you need a counterfactual. You need a view of the world about what would have happened without the intervention. And your counterfactual, if you claim there’s a certain number of permits that were due to the policy, is what actually happened minus the number of permits that you claim are attributed to the policy.
In the counterfactual created by these authors, even though it wasn’t transparent, up to the point of the policy in the real data you had permits growing at about 12% a year for four or five years. So permits were on a growth curve. It’s a cyclical data series, and it was growing after the post GFC lull.
And the counterfactual by the authors said that without the policy that growth of 12% a year would actually just be 6% a year in future.
In our Substack posts you can just see how dramatically off the past trend the counterfactual imposed by these authors actually was. And, maybe unsurprisingly, the authors found that the Unitary Plan doubled construction from what it otherwise would have been – which is to say, things continued growing at about 12% a year for a few years.
So what did they in essence do wrong?
I should note there were two elements to this paper.
One was a quantification of the number of permits attributed to the upzoning, and another was a very complex econometric exercise to prove that there was a statistically significant increase in these overall permits. We ignored this econometrics in our columns, because it didn’t seem particularly relevant, but it rests on the same fundamental logical error, which is linear extrapolation on a cyclical series.
The econometrics proved the statistical significance of permit growth beyond a linear extrapolation from the policy date. They use a linear extrapolation as their control group because their actual chosen control group isn’t a useful control group. It’s not a useful control group because it’s part of the same city-wide housing market so, of course, a lot of building activity moved from these non-upzoned areas into the upzoned areas, making the non-upzoned areas not a useful control.
So they invent a control, which is the counterfactual for the non-upzoned areas, which they just create by linear extrapolation from the past trend. And in essence the econometrics then proves that total permits exceeded that control by a statistically significant margin. That makes the econometrics basically a complicated way not finding out much, and the overall number a complicated method of not finding out anything particularly meaningful.
There were a couple of issues.
There was one major issue aside from the inappropriateness of linear extrapolation, and that was that they only used a partial data set. Many of the permits missing from their data set, that they carved out of their sample, were a special type of permit that grew rapidly before the Plan and then died away rapidly after the Plan kicked in.
By excluding it they turned this Auckland permit series, which had effectively steady growth, from something that didn’t have a structural break into a sub-sample with a structural break. They unsurprisingly found a structural break in their sub-sample – a contrived structural break – by carving out a subset of permits that had its own structural break.
So those were the issues with Auckland. And I think to do Auckland properly, to understand the effect, I just don’t think you can take a single housing market and try to carve up that market into an intervention and control group. You’re going to have to compare Auckland to elsewhere, as best you can.
Cameron Murray:
And what has the fallout been? What have you learned about academia, and this publishing game, and the approach to knowledge creation from this exercise?
My feeling is that for you it has been a bit of a lesson. Is that right?
Tim Helm:
Yeah, no doubt.
I’m not an academic. You’ve been in the business and seen this. Academics have their publication incentives, right? I’d just be publishing what I could as well.
The shocking thing for me is that in the year after this working paper was released, and got heavily cited and used in policy advice in New Zealand, no economist in New Zealand apparently took the Stats NZ data, and did an Excel time series, and found that this apparent structural break didn’t appear in the overall series, and raised some questions.
I just happened to get into it a year after this working paper was first released. So I think the New Zealand economics fraternity just dropped the ball on that. There aren’t enough people scrutinizing these things.
I suppose it passed peer review as well, right?
It took a little bit of curiosity. I, like you, was surprised at this big number.
Cameron Murray:
I was already skeptical of peer review. It’s not obvious to me how much you can review a paper. You’re just squeezing it in, having a read. Does it fit with the stuff I like? Does it fit with the way journals like papers written?
It’s not a replication, peer review.
The other thing for me in this example is how this result can circulate around the world before our critique does and people will still ignore the critique.
So they’re the lessons.
But anyway, Tim, it’s been great having you on.
I had more questions for you, but I think what we might do is wrap it up, and I’ll have to get you back on another time.
