Sunday, April 14, 2013

Will land prices rise as population rises?

Just a quick post here. Bill McBride, like many others, thinks that rising population implies that housing prices should have a long-term upward trend. I'm not sure he's right. McBride writes:
A key reason for the upward slope in real house prices is because some areas are land constrained, and with an increasing population, the value of land increases faster than inflation...The bottom line is there is an upward slope to real house prices.
I'm not sure this is true. The reasoning is intuitive: More people + same amount of land = land is more scarce. But here's why I think that reasoning is not quite right: Most of the value of land is the value of its proximity to centers of economic activity. Any urban economic model that incorporates geometry, geography, etc. will tell you this.

In other words, New York City real estate is high-priced because New York City is an agglomeration of economic activity. It is not high-priced because an increasing number of people are being forced to live in New York City. That isn't even the case! No law makes people cram themselves into NYC (except in that Kurt Russell movie!); you are legally free to move out to North Texas and get a nice ranch. People choose to live in the heart of New York City because of the economic (and social) opportunities offered by proximity to all the other people living there. So they're willing to pay lots for land.

To use an old real estate cliche: Location, location, location.

So will land-use restrictions around big cities cause land prices there to increase? Maybe, maybe not. If those land-use restrictions hurt agglomeration, they will choke off the economic value of living in the city center, which might lower land prices. Also, a lot depends on what else is going on in the region, and in the rest of the country. Agglomeration effects can be weirdly non-linear and history-dependent. You can even have long-term de-agglomeration, simply due to shifting geographic patterns of trade and industry; as an example, just look at the Rust Belt. Even if land prices up-trended nationally due to population pressure, regional prices might experience long secular downtrends due to rearrangements of national economic activity.

Also, there's technology, which can change a lot faster than population. Urban economists like Masahisa Fujita will tell you that once transportation costs get really low - for example, the more we switch to a digital online economy - the less economically important it is for companies and people to be in close physical proximity. That will cause land prices to decline in some places, and rise in others, even as total economic value increases across the country. And if the imperatives of those technologies cause land-use restrictions to be altered, all bets really are off. Remember, the total amount of American land actually occupied by humanity is tiny. And (as a commented points out), there is plenty of unoccupied cheap land in the near vicinity of many cities.

So I say, don't fall for the easy logic of more people = more expensive land. It might work that way, but it very well might not.

Update: Paul Krugman has some data to back up my suspicion; it turns out that the average American actually lives in a less densely populated area than a decade ago. My suspicion is that this trend stretches back farther than the year 2000. But as Krugman notes:
McBride’s point that actual real housing prices do seem to have an upward trend remains important, and needs explaining.
My intuition says that it's because we've been occupying larger and larger living spaces. Now, that wouldn't always tend to raise land prices (maybe it happened because agglomeration effects decreased), but if it indicates a shift in consumer demand for dwelling space, it would. If the value of proximity stays the same and people at a given spot choose to spend their money on bigger dwellings, they'll crowd each other out, and land close to agglomeration centers will become more scarce. Also, it might be that agglomeration effects themselves have increased in some places like San Francisco and New York, where industrial clustering in tech and finance has raised the value of living close to other people in the same industry. Anyway, if I had more time, I'd look up data and try to solve the mystery; those are just guesses.

Update 2: A couple more issues I didn't raise before, but which are interesting:

* In the very long run, McBride is right; if population goes up enough, land prices will eventually have to rise. But there's good reason to think we'll never reach that very long run. American fertility rates are slightly below replacement levels, and most rich countries are far lower; developing countries aren't far behind (China's working-age population is already falling, India and Latin America and Southeast Asia are all at or near replacement levels). Our population is growing slowly and steadily because of immigration, but that will eventually trickle off as countries' incomes converge around the world (Mexican immigration is already at zero or negative). So even if population growth is positive right now, expectations of long-term future population growth may be negative or flat, which might affect land prices.

* If median incomes fall, people will have less to spend on land, which will exert downward pressure on land prices. This has been the norm in America for over a decade now; no one knows if it will continue.

* Public transit policy matters. If cities build a lot of light rail, a lot of people can give up their cars, freeing up lots of urban land that used to be garages and parking lots.

* For many investors, what matters for housing as an investment is not really whether land prices rise, but whether they rise faster than overall GDP (or some other benchmark like the risk-free rate or the stock market). Since population growth raises GDP, it's not clear whether any realistic amount of population growth will make land appreciate faster than - or anywhere near as fast as - the economy's stock of productive assets.

So these are all caveats to think about before concluding that population growth gives land a long-term upward price trend.


  1. "A key reason for the upward slope in real house prices is because some areas are land constrained, and with an increasing population, the value of land increases faster than inflation"

    The word "some" is important, and actually a very big qualifier, because while the supply of land is fixed, the supply of what really counts, developed, desirable, infrastructured, non-zoning restricted land is very very unfixed. I've read that something like 95% of the land in even the US is basically undeveloped, and can be bought dirt cheap.

    And here Shiller's advance in technology is very important. If it gets a lot cheaper to put infrastructure on undeveloped land, like with the help of our robot friends, to start a new city or expand an old one, or to build higher, suddenly it can get a lot cheaper to expand the supply of housing in response to a change in price. In other words, the supply curve for housing can shift out a lot with technological advance.

    Furthermore, advances in teleworking and computer driven cars -- which can become little computer driven offices -- can make it so people will be a lot more willing to move to the new developments on the vast vast tracks of undeveloped land that can be bought for very little per acre. And we may finally wise up on zoning, and not let selfish current owners make so many suffer so much with their supply restrictions to jack up prices -- "local rights" becomes personal fief rights.

    Finally, Shiller claims to be careful in his series -- but I haven't looked up the details -- to control for the fact that over time homes have gotten a lot bigger, better, more elaborate. Go back 100 years and many houses didn't even have indoor plumbing.

    1. An interesting somewhat-related paper is by Albert Saiz (2011 QJE) about geographic determinants of housing supply. From the abstract:

      "I process satellite-generated data on terrain elevation and presence of water bodies to precisely estimate the amount of developable land in U.S. metropolitan areas. The data show that residential development is effectively curtailed by the presence of steep-sloped terrain."

      Mian and Sufi use Saiz's data to instrument for household debt to income ratios in their 2012 paper.

    2. Hmm, steep-sloped terrain hasn't stopped San Francisco, has it? ;-)

    3. Anonymous4:41 AM


      Not your best work.

      1. Cities work due to network effects and other aspects of information such as location effects (suppliers want to be near customers). Best example of location effect: Sun Studios in Memphis in the 1950s/60s, which attracted extraordinary talent from throughout the Mid-South (from Elvis to Dolly and Porter)

      2. Against such gains must be offset the inefficiencies of a particular city (crime, cost of services, cost of transportation, etc.)

      3. Terrain = sanitary sewers and water become very expensive. Such is a very big deal in cities where terrain is rock. The study is accurate.

      4. You wholly fail to consider the artistic/aesthetic side of cities. A city which is visually screwed up is entirely different than a city with great views. Downtown Cleveland, for example, is now a mess because the view of the Lake for most of the year, especially at night, is a black hole. Same with the view of the river in STL. Both cities have thousands of acres where no one wants to live or work and thus both have major problems of sprawl.

    4. That no one wants to work in the "thousands" of acres you cite is more complicated than that, it's game theoretic.

      Let's take Central Park. And let's think of it as yet to be discovered land (so you don't have to buy it). For any one given person, perhaps a poor tenet sharing with five other people, it's every bit worthwhile to build on the park, just a tiny bit.

      It would be everyone's dominant strategy to move in, because personal gain from more space outweighs personal loss of slightly less uninhabited space. But when everyone does it the situation becomes nothing too much more complicated than your prisoner's dilemma.

    5. Not your best work.

      What do you think was my best work?

      Cities work due to network effects and other aspects of information such as location effects (suppliers want to be near customers).

      That's absolutely right...but as Fujita shows, location effects first become stronger, but then become weaker, as transport costs decrease. There's an inverse U-shaped curve. As transport costs tend to 0, location stops becoming more important, and becomes less and less important.

      Unintuitive, but it falls out of the math pretty easily.

      You wholly fail to consider the artistic/aesthetic side of cities.

      Yes, that's right; I also left out other natural amenities like weather. Those can change tool and their changes can outweigh any population pressure.

      See, I was even more right than I let on! ;-)

    6. Hmm.

      Yet Toronto's and Chicago's downtowns are vibrant, so the view of Lake Erie is a black hole yet the views of Lake Ontario & Lake Michigan are swell. Did I get that right?

    7. You people are all making really good observations. But no-one on this whole thread has mentioned regulatory distortions and zero-sum economic rent.

      It is absolutely correct that transport improvements allow the land supply in the urban economy to grow. So does the rise in urban incomes relative to the cost of farmland (or other non-rural land surrounding cities).

      We probably all understand that there is a "land rent curve" in all urban economies, that slopes gradually downward out into the surrounding countryside as increasing "cost of transport" reduces the value of the land to a participant in the urban economy.

      But regulatory interference, in the form of green belts or urban growth boundaries or "infrastructure plans" or "rural density zoning" or similar, cause a "discontinuity" in this rent curve. A sudden steepness in the slope will be encountered as the boundary is reached, and the value of the land inside the boundary will be typically tens of times higher than what it would have been absent the distortion. All the land within this boundary will also be inflated in value, due to the premiums commanded by "location".

      This is why the ratio of land value to GDP is many times higher in the UK, than in Europe, which in turn is many times higher than the USA. Theoretically, land values should be derived from the incomes earned in the use of the land. The difference between cities and nations, is "economic land rent" - straight-out wealth transfer to property owners.

      Most "Georgists" have long been obsessed with the uplift in land values that occurs due to normal economic growth and infrastructure investment. But the uplift that occurs due to regulatory distortions is an order of magnitude greater.

      There was a massive paradigm shift in the formation of economic land rents in the urban economy, due to "automobility". Robert Murray Haig's 1923 theory of transportation and economic land rent was correct. The greater the flexibility of the transport system, and the greater the supply of land brought into the urban economy per dollar of transport cost (including infrastructure), the more "dispersed" economic land rent is. Even rail based transport only brought ribbons of land into the urban economy, which still left the supply easily capturable by rentier interests.

      So the slope of the economic land rent curve was rendered continually shallower, out further and further into the surrounding countryside.

      What Haig did not anticipate, was the massive reduction in economic land rent made possible by genuinely competitive automobile based development. The supply of land accessible to the urban economy by automobile is so great that it ceases to be economic for "land bankers" to corner the available supply - because supplies further away can still be accessed and "new towns" or "splatter" developments built. So developers tend to be buying the land in genuinely rural land markets, and developing it in competition with each other, with all profits including uplift in land value, being minimised, just as in genuine free markets for goods and services.

      The trend in recent years, for anti-sprawl urban planning or even deliberately exclusionary policies, has undermined this paradigm and restored the pre-automobility era's levels of economic land rent in urban economies.

    8. Re Albert Saiz and "Geographic Determinants of Housing Supply"

      Wendell Cox critiqued this:

      Saiz committed a disappointing error of analytical approach. Drawing a radius the same size in all cases regardless of city size, and analysing the geographic constraints within that radius, means the conclusions will be meaningless.

      The Federal Reserve Bank of Atlanta also comments on the Saiz paper:

      “…….the local-regulation index is ITSELF strongly correlated with Saiz’s space-constraint index, as space-constrained cities tend to have stricter regulatory limits on new construction. This correlation provides compelling evidence for something that many housing economists have long suspected—local voters seek to protect the values of expensive homes by preventing new homes from being built. This finding may be puzzling to some, as it may be hard to imagine why land-constrained cities would need to implement further restrictions on new construction. However, some new development, perhaps via dense apartment buildings, is usually possible. Note that unlike a lot of correlations in economics, we can be reasonably sure that the direction of causality runs from space constraints to local regulations, not the other way around. After all, it is hard to create a new mountain, lake, or ocean through the political process……”

      A Korean paper with a completely different methodology to Saiz, finds that “geographic constraints” do NOT have any affect on housing prices in Korea.

      SON and KIM (1998) “Analysis of Urban Land Shortages: The Case of Korean Cities”

      William Fischel, in “Do Growth Controls Matter”? (1990) suggested that they do not matter until a critical mass is reached with their imposition in MOST of the municipalities in a region; until then, the non-constrained municipalities act as a price “vent”. California tipped over this crucial point in the 1990′s, where there were no longer enough non-constrained municipalities.

      Geographic constraints are almost irrelevant to the PRICE OF LAND as long as “building” is permitted SOMEWHERE. The “cost of building” might be a little higher due to terrain, but the cost of LAND should not be, as long as there are not artificial land rationing regulations at work. The fact that urban LAND prices in the cases we are considering, are inflated by hundreds or thousands of percent, is evidence that the problem is regulations, not geography. The Korean paper’s methodology manages to make this distinction, while Saiz’ methodology does not.

    9. I'll have to give a good look at the Saez article when I have time, and the criticisms of it, but I will say this:

      I've driven tons all over our interstates. I see predominantly flat, or pretty flat, land that's empty or little developed. Water issues you definitely hear about, but again this gets into technology advance increasing the supply of infrastructured usable land for housing. With advances in robotics, it may get far cheaper to build pipelines, like the big one proposed for my home state, Arizona, from the Colorado River. Likewise, advances in desalination (and long pipelines form the desalination plants) may make tons more land developable. From a recent Bloomberg article:

      "And technology developed by San Leandro, California-based Energy Recovery Inc. and other companies that recirculates water in filtering plants has cut energy expenses by as much as 60 percent, says Energy Recovery CEO Tom Rooney. The technology lessens the cost of a cubic meter of desalinated water to about 50 cents, Felber says.

      Fresh ground supplies, by comparison, run less than 20 cents, according to the 2030 Water Resources Group."


      Now, note the 50 cents per gallon versus 20 cents per cubic meter comparison. If living on a new housing development means your water bill is twice as much, but your mortgage is half as much (or a quarter as much), you save vastly more than you spend. But desalination is very energy intensive. Imagine combining it with plummeting solar costs, and huge expected advances in solar. So, again, technological advance may make it far cheaper in the future to add greatly to our supply of infrastructured, developed land for new homes.

    10. Units are both price/m^3 not per gallon.

    11. Your point is absolutely valid that a city that was forced to stop growing by lack of water availability, would suffer the same consequences for its urban land market, as a city on an island that literally ran out of land.

      We would need to work out what to do in such cases as they arose, and cases where some other unalterable factor made the land market behave in this way.

      Of course a higher cost of water (desal plants or whatever) would increase the cost of living, but it would not need to force house prices up - an ongoing cost or higher local taxes, would not be an inflator of house prices at all, rather the opposite.

      But lumping the additional cost upfront into new house prices would be a different story. Lumping costs upfront into the price of new houses makes the prices of all houses go up, and this is a windfall gain to all property owners at the expense of future buyers and all renters.

      But what we need to understand is that urban growth boundaries or proxies for them, also make urban land markets behave this way, and it is little understood that this is a very costly course to adopt.

    12. @PhilBest
      The political dynamics of water seem rather different from the political dynamics of zoning.
      With zoning the (implicit or explicit) idea seems well-known to everyone who matters --- we'll reduce the stock of houses and that will boost, or at least maintain, the value of your house. Even with a fixed sized island and without zoning, everyone can see what's playing out, how the game is going to end.

      Water, for better or worse, is a lot more invisible, and no-one much seems poised to win big by restricting the number of water clients. Thus, while there will be a hard limit, it will be hit while traveling at 100mph, rather than the zoning/limited island case, where the rising prices will kick in before the ultimate limits are hit so as to reduce the speed of expansion, and to get everyone ready for the end point.

      I think we see this in the Phoenix area, where the annual hand wringing rhetoric over water is not accompanied by any sort of serious attempt to reduce growth. The same is true in LA, though there the water constraints are slightly less serious and the boosterist attempts to grow the population not so aggressive.

  2. the price of land is variable on many things.

    land tenancy is most critical thing we all need to secure, without it, we are reduced to living a hunter-gatherer quality of life.

    Purchases prices are driven by lending environment -- the 1974 Equal Credit Opportunity Act put the wives' salaries as the same standing as the husbands, and was a large driver of the 1970s home price bubble.

    After the Volcker Fed was done throttling credit, since 1983 we've seen rather monotonically falling interest rates:

    which have greatly boosted home values.

    What is more interesting to look at is what housing RENTS are going to do.

    if the past is any guide, they will move in lockstep with wages:

    because we all have to bid against ourselves for housing.

    Speaking as a neo-Georgist, modern-day economics does not understand the land element -- how it is a massive wealth siphon from working America to the predatory wealthy -- AT ALL.

    Just another component in their pretty graphs, with no cognizance of the centrality land has in everyone's lives.

    Try living without it, LOL.

  3. The income accruing to land rents, as a portion of national income, has declined steadily for two hundred years.

    1. The income accruing to land rents, as a portion of national income, has declined steadily for two hundred years.

      and yet housing rents are the dominant line item expense (along with health care) in all our budgets.

      Funny that.

      The economic rents in housing and medical care are of course immense. It is unclear to me what will happen with them if food and energy costs escalate rapidly this century as they well might.

      Our household budgets are the battleground of the rent-seekers.

      What we spend on housing is determined by our discretionary incomes -- every last penny of it, really, since buying a house is the obvious way to escape future inflation in housing.

    2. That graph is not a graph of real housing prices. It's an index of nominal housing inflation.

    3. The All-Devouring Rent is a thesis we Georgists have about that.

      Housing gladly eats every nominal wage gain the masses receive.

      How could it not?

      Demand is unbounded (we all want a newer house. better schools, bigger back yard, better view, or a 2nd house, plus some rental units . . .)

      And supply is so slow to respond, since real estate by definition is immobile, and new supply cannot be erected due to NIMBYism, and the basic physics that two things cannot occupy the same space at the same time.

      Looking at real housing purchase prices is deceptive given the interest rate environment of 1983-now (falling from 15%+ to 3%).

      You need to look at rents.

    4. I have no idea what you're talking about, Troy.

    5. ".....Our household budgets are the battleground of the rent-seekers....."

      Agree very much. There is a noticeable difference in the "housing affordability" data from Demographia and others - cities with median multiples of "3" tend to actually have much larger, newer houses and much more land per household. In contrast, the cities with median multiples of "6" and over tend to have much smaller houses and even smaller amounts of land per household, and housing tends to be much older. The difference is this.

      A median multiple of around "3" represents an amount of expenditure that households are comfortable with, so that as incomes rise relative to the cost of raw land, and relative to the cost of construction, and technology improves, households tend to consume more land and attributes of housing. The value of the land tends to be about 30% of the value of the "housing", even as the size of the lot grows, typically as large as 1/2 acre on average now.

      But the median multiple of "6" represents "the most people can stretch to paying". It will be found that economic land rent is rising and households are paying more and more for a square foot of land; and households are sacrificing not just land consumption but other attributes of housing, so as to house themselves at all within "what they can afford". Typically the value of the land represents 70% of the value of the "housing", even as the section sizes shrink, typically to 1/16 of an acre on average in the UK now.

      Of course the increase in economic LAND rent is exponential, when you "net" this out. The difference is always due to regulatory distortions (there might be an island nation-state that has literally run out of land, but this is certainly not the explanation in, say, Australia).

    6. Developers in these "planned" growth-constrained markets are the meat in the sandwich – it is not them who are the beneficiaries of the quota system in urban land; it is the original vendors of the land.

      The developers are placed in a very high risk situation. They have to out-bid each other for the quota’d quantities of land, and still sell finished housing at a price the market can stand.

      Developers working under these conditions do work backwards from “what they think they can get for the finished product”, and this is reflected in “what they are prepared to bid for the quota’d land supply”. However, we would not regard it as acceptable for bread or milk, to run a quota scheme that allows the prices to settle at “what people can pay” rather than “what suppliers freely in competition with each other can bring the product to market at”.

      What keeps the price of bread and milk and everything else, being set not by “what people are prepared to pay”, but by “what suppliers freely in competition with each other can bring the product to market at”, is the lack of anti-competitive barriers to anyone converting land from other uses, to dairy farming or wheat growing; and to supplying raw product to supply chain intermediaries; and to processing and producing finished product.

      This market freedom does NOT result in a massive “oversupply” of bread or milk or cars or TV’s; nor does “freedom to convert land from other uses to urban uses” (just as easily as converting it from growing one crop to another) result in massive over-supply of new houses.

  4. There is also the minor point that global population growth is decelarating, having already gone negative in several high income nations, starting with Japan. US population growth continues to be positive, but the birth rate is reaching new lows, so our time is coming, and land prices in Japan have been falling for more than two decades now.

  5. Noah, I'm very tempted to agree with this argument, and do for the most part. I would note though you seem to make the very classically-economist mistake of assuming people can "just move" when they deem their utility derived from a North Texan ranch outweighs the total benefit of NYC (salary, nightlife, prestige, whatever). Even ignoring all the economic frictions of moving, I also think there is a huge anchor to staying in big cities that precedes agglomeration.

    People want to stay with their historical familial network, both immediate and extended. While this effect is particularly weak in America (definitely stronger in Europe) I still think it's important.

    If in the future population density growth in cities is more correlated with land price in Europe than America, this might well be the reason.

    I think a lot of what you're arguing is the exception, not the norm. But I'm not convinced of that, and am really curious to see how this plays out. (Indeed we'll both – hopefully! – be alive when world population is set to peak).

    1. starbird3:05 PM

      This is backwards. In fact, people can "just *not* move" from North Texas to NYC. The cities' growth is not driven by high fertility among the people who already live there, it's driven by people moving in. So as long as a city is growing at all, I think the various frictions associated with moving will fall on the other side of the ledger.

    2. The USA is actually outstanding for the rate of migration from high-economic-land-rent cities to low-economic land rent ones. Of course in most other nations this option does not exist at all, universal urban planning practices making all their cities high-economic-land rent.

      British and Australians and others trapped in paying gouging housing costs (and business startup land costs) might really envy Americans their "relocation to affordability" option if they were better informed.

  6. If one assumes rational expectations and a predictable human growth rate. If we assume technology is constant, and people preferences do not change, then house prices will rise with the time preference discounting and inflation not more right?

    Now clearly not all the above assumptions are realistic, question is which one(s)?

    1. Rational expectations don't (can't) exist, technology isn't constant, and preferences change.

      Human growth rate is probably fair predictable.

      It's probably just a weaker assumption to assert that land prices rise in proportion with population!

    2. Rational expectations don't exist -> How far off are we from it? Is it really the cause of the failures here? Legitimate and interesting question here, the answer of which could probably have wider implications for the economic analysis.

      Preferences change? Do you really think they change that frequently? Especially aren't they continuous, and therefore in a span of let's say 10Y we can assume they are constant?

      What I meant to say is that if we observe that land prices rise in proportion with population is saying that one of the above hypothesis is wrong, and the question is which one?

      Are the land prices increasing just because technology is not constant? I don't think it's down to preferences changing, is it an irrational bubble? Or what is "market failure" that is causing this?

    3. You can't even define rational expectations. W Brian Arthur is much smarter than I am, so I'll let him to the heavy-lifting:

      "This observation of course is not new. Other economists, Shackle in particular (1955, 1992), have written much about this. But it has an important consequence for theorizing. To the degree that outcomes are unknowable, the decision problems they pose are not well- defined. It follows that rationality—pure deductive rationality—is not well-defined either, for the simple reason that there cannot be a logical solution to a problem that is not logically defined. It follows that in such situations deductive rationality is not just a bad assumption; it cannot exist. There might be intelligent behavior, there might be sensible behavior, there might be farsighted behavior, but rigorously speaking there can not be deductively rational behavior. Therefore we cannot assume it."

    4. OK then, if people are forward looking, and you tell them population growth is increasing in a certain rate, while the land supply is fixed (and the technology to use that land is fixed, and the preference of people for housing space is constant), they can compute the equation of the land price.

      They don't even have to do it, some smart economist does it and publish it. Then given some discounting this gives you the current and all future prices of housing.

      If this were to be true we would not be seeing what we see in the housing prices (or maybe we do if we believe Schiller). Since we do not see that, what is wrong here? Is it the fact that people follow animal spirits and just switch preferences all the time (so a potential bubble), or is it that we cannot know this deterministic function above (incomplete information), or is it that technology is changing and some market friction is causing the prices not to adjust immediately? Or maybe some other reason?

      Same reasoning of the fixed land applies to Gold or to oil (if we can roughly equate the technonological changes in land use to the unknown supply of gold and oil), what is driving their price movements as they do not exhibit the same behavior as housing.

    5. Well, the rate of population increases changes, land supply is not fixed, technology is not constant, preferences change: every single condition you name does not hold!

    6. I know that it's a very complex topic, I'm just wondering if there was any research done along those lines. Plus you would assume some of the conditions above would have more impact than the others.

      Plus why would land prices behave so differently than gold prices, if we are saying that land price is increasing because of the fixed supply of land but increasing population, well this applies to gold as well (gold mining is quite limited).

      So clearly something more than just saying:

      more people = more expensive land

      is driving the land price (for instance the safe asset issue, people see land as a safe asset or maybe the MBS market has created a bubble in the housing market, etc ...)

      I don't view the fact the population is increasing as a real predictor of the house prices, as if the issue is unpredictability of the population growth is to blame, then we should see flucutations, and not bubbly like growth.

    7. Over the very long-term, appreciation of real estate minus the costs that you have to put in to maintaining a house is about zero when you aggregate across everywhere. For instance, while real estate in NYC has risen a lot, there are a ton of towns in western Nebraska and Kansas where real estate, in real terms, was worth more 100 years ago. Real estate in most neighborhoods in Detroit and Flint in real terms almost certainly were worth more 50 years ago.

      RE is very, very localized. Saying more people in the world (and even that isn't terribly true; look at changes in fertility rates almost everywhere outside of Africa) -> higher RE prices in the world is one of the more idiotic facile arguments I've heard in my life.

      If you're betting in RE, you're betting on a small region. People will migrate to certain places from certain places for different reasons (employment, weather, change-in-wealth, crime, culture, change-in-mode-of-employment, etc.) RE in certain metropolitan areas could also go on massive decades-long bubbles (mostly due to land-use & zoning restrictions and geography). All those factors matter much more than whether there are more people or not.

    8. What everyone on this discussion thread is missing, bright as they are, is the role of regulatory distortions and economic land rent.

      The original “Ricardian” assumption made by so many economists who lack specialist land economics training, is that "the supply of land is fixed"; therefore the only thing that matters for the price of land, is "demand", and the income generated from the use of each piece of land. This assumption falls over in the case of a "planned" urban land market because Ricardo was only ever talking about the “total supply of land for all uses”, and assuming that land could be freely converted from one use to another. Obviously a quota system for land for building houses, violates this assumption.

      Yes, improvements in transport and technology SHOULD lower the ZERO SUM economic rent captured by land owners, due to the increased competition provided by the increased supply of land made accessible to the participants in the urban economy. "Positive sum" economic rent is different, this is the rent that accrues to land owners as a result of growth in incomes and in the urban economy all around them. It is not extraction of a higher share of status quo income levels, which is what regulatory rationing of land for a particular use, allows to happen.

      See my other comments on this thread.

    9. PhilBest, good point.

      Dohsan, not sure you understood my point, with RE, prices are stale, they don't move (well not more than discounting). If 100 years ago you predicted this human growth, then land prices would have went up already 100 years ago and not moved, so your first paragraph is actually showing the RE point.

      My point was one cannot say:

      land is fixed / population is rising => house prices are rising

      since this would mean either people are not forward looking or population growth is not predictable.

      Now obviously this actually almost only applies globally, as locally one could have as PhilBest pointed out regulatory issues or some other issue that may cause a local to outprice the other one. But what I'm saying is this issue cannot just be population growth as people are expecting this.

    10. This comment has been removed by the author.

    11. Georges:

      Except that human growth in different regions (and probably more importantly, whether artificial restrictions about land-use will be put up) are not predictable. Developers try to predict growth (get caught in in euphoria, use easy credit, etc.) So RE prices move.

      You logic applies to equities as well, yet stock prices fluctuate a lot.

      In any case, I don't disagree that a rising population does not mean rising RE prices. Did I not say that?

    12. Dohsan, then we agree :)

      My point was we have all those determinants above, if all the hypothesis were true, prices would be stale. Clearly they are not, so something must break.
      I'm saying to conclude that rising population means land prices should go up in time means it's the forward looking assumption that is breaking (it's not the unpredictability as this would lead to flucutating prices, people cannot be underestimating the population growth all the time), which I find unlikely.

      So I think one cannot use the argument rising population => land prices should be going up.

  7. The Housing Bubble of 2003-2006 in the four Sand States of CA, AZ, NV, and FL, where about 7/8ths of home price appreciation in the U.S. took place, was driven in large part by the assumption that rising populations dues to immigration had to lead to higher home prices. It turned out, however, that marginal homebuyers couldn't earn enough money to pay for those mortgages.

    1. Love your work, Steve, and quote it often.

      Interestingly, some of the cities with the highest in-migration of all, did not have a house PRICE bubble at all. Eg Houston, Atlanta, Austin, Charlotte, Raleigh NC.....

      Stable urban land costs, due to high elasticity of supply, was actually an attractor of migrants.

      In the unusual cases of housing bubbles that involved both price and overbuilding, the mechanism always involved the active participation of government chasing "planning gain" or similar. "Supply" can be abundant but quota'd, leading to bidding wars for it. Urban planners "releasing 10 years supply of land"" is just throwing fuel on a fire. The only way to put the fire out completely is to abolish the barriers to conversion of rural land to urban anywhere.

      When Portland established its "20 year UGB", the price of land began to inflate just 4 years later. Here is the basic flaw with the idea that “X years supply of land has been zoned” on greenfields beyond the existing urban fringe.

      The zoned amount is NOT the amount that NORMALLY WOULD COME TO MARKET ANYWAY on the normal rural land market, within that zone, IN THE NEXT “X” YEARS. It is the TOTAL quantity of land within the zone. To REALLY have “X years of supply”, it would be necessary to include in the zone, enough land that the quantity that IS “X years supply” WOULD come up for sale anyway in a normal rural market, at normal rural prices. This would mean a zone with about ten to twenty times as much land as “X years supply for urban growth”.

      The land comprising “X years supply” by the planners standards, is being used for something already and no owner of it has any inducement to sell it at the value that it is worth IN ITS PRESENT USE. But having become aware that they are part of a newly created oligopoly holding of the next “X years supply of land for urban growth”, they cease to think in terms of rural land values at all, and start thinking like investors/speculators. None of them will be satisfied with a 10% capital gain, or even 20%. Expectations of gains become a reason to HOLD land and NOT sell it. So the planners “X years supply” becomes a LOT LESS than “X years supply” purely because of typical investor psychology. The planners have REDUCED the likelihood that any one land owner within the zone will in fact sell the land within the “X” years at all.

      Then developers not only have to door-knock and persuade "attached" owners of land to part with it, they need to offer greater than the present value of what the owner thinks the price of the land MIGHT yet inflate to......!

      This is why there is no middle ground, "moderately unaffordable housing" cities. Containing urban growth and rationing the land supply is not like a "flow control valve" over house prices as planners like to think it is, it is like an "on" switch for a nuclear chain reaction.

  8. Yep, when price gets high enough, people find substitutes. Meaning that they live move up (skyscrapers) or they move out (suburbs, exurbs, elsewhere).

    We tend to remember the communities that are successful and how those land values have risen, but we forget about the cities with largely obsolete economic foundations like manufacturing centers around the Great Lakes. It doesn't have to be a zero sum, but they certainly offset each other.

    Some land is valuable because of the proximity to natural beauty or water traffic ways. But most land is valuable because of investment in transportation, education, security, and entertainment. For example ports, subways, roads, colleges and good public schools, police, and parks all increase the value of land. Without those, most land isn't terribly useful or desirable.

    Notice that most of these are public improvements paid for with taxes, and if done correctly they are a good return because the value of the land increases with greater demand, and thereby increases the tax base (along with taxes on higher commerce). Which is a good reason for government to tax land at a higher rate than improvements, since it should create an incentive for wise public investment while not discouraging the largely private investments that are made in improvements to the land itself.

    If the public decides that they don't want to invest in these public goods anymore, then the land value declines along with the decline in transportation, education, entertainment, and safety because people don't want to live or do business in such a place.

    So the value of land will tend to follow the trend of public investment and the economy. It is the same idea as a company's value increasing if they make good investments in their technology, production, distribution, and marketing. No asset's long-term appreciation is certain, not even land.

  9. No asset's long-term appreciation is certain, not even land.

    Although there is an interesting tradeoff between development/urbanization and the limited supply of high quality arable farmland.

    1. But the value of farmland has been decreasing relative to urban incomes, for decades. Urban economies have tended to increase in size 16-fold over 7 decades as economies "develop". Greater efficiency in farming means that the gains in farmland production exceed the loss of farmland by a significant factor. The balance between rural land and urban land worldwide, is about 30 to 1. If urban areas doubled, the ratio could change to 15 to 1 - but there is much more land not yet converted to farmland.

      Transport improvements have of course brought the farming production of much larger areas of land, into supply for urban economies. Spoilage and wastage still occurs on a considerable scale in developed nations due to the fact that we have such excess.

      Furthermore, the amount of farmland that WAS used to grow feed for horses and draft animals, exceeds the amount of land lost to urban expansion. So we are ahead anyway.....!

  10. Most of the large cities in the world are built on natural harbors allowing low transport costs. Even Mexico City was built on an island in a lake.

  11. Anonymous11:50 AM

    In any given location, land price is going to increase with population. The question you are asking is not a counterpoint to that, you are instead questioning the assumption that population is monotonically increasing in any given region. To determine that, there are a bunch of factors involved: base birth rate, relative value of the jobs and people that are accessible to a location (and any changes in the locations that are accessible, based on transportation improvements), and also things like base cost of maintaining life in a region (in other words, how hard is it to get base necessities, how much does it cost to keep a comfortable home in good repair, including A/C and heat, running water, etc).

    A lot of these factors in population change, though, are very discrete - they change only when certain thresholds are reached, like a new train line is constructed, or a new technology becomes widespread. But there is, I think, something to the idea that as more people occupy a region, people get further out from the center of it, but still commute to that center. This provides a population of workers that would be available to new businesses that moved to the periphery, which gradually redistributes the population, as businesses draw off workers, then workers move further out to have cheaper housing while commuting to the new businesses, etc. But it's a really sticky process - people have ties to the community even if their job moves to a more convenient place - if their friends and family are still in the city, they will stay near the city.

    It would be interesting to see an attempt to actually measure these effects, but my guess would be that the stickiness of relationships causes the spreading-out effect to be much more gradual than the steady birth-rate-based increase. In fact, my expectation would be that it is a fixed percentage of the birth rate (though maybe very close to 100%) - each new generation redistributes according to the spread of businesses, with only some fraction of them being frozen into the place their parents live. So prices increase gradually, punctuated by big changes (up or down) from various events (like transportation improvements, huge economic successes - a single company growing like crazy and spawning a new sector in its home city - and big new technologies)

    1. Here's part of your answer:

      Dispersion of economic activity is a natural free market response to congestion, economic land rent, and workforce cost pressures; as the real cost of transport decreases and communications technology improves. It is not dispersion that needs explaining, it is sustained centralisation in a few exceptional outlier urban economies. Often this is due to heavy-handed urban planning, and results in massive zero-sum transfers of economic rent from the bottom of society to the top.

    2. Anonymous10:48 AM

      I don't think either need explaining, really. People stay because they don't want to change jobs, and they have friends and family and homes near where they have been, so even when they do change jobs, they look for one near their home and social connections, when they change homes, they look for one near their job and social connections.

      I don't know if theory can really prove which influence is stronger (the economic incentive to spread out vs. the social incentive to stay put). The best it could probably do is tell you which factors contribute and predict which influence will be stronger, given certain assumptions about the future progress of the factors involved. Technology might make it easier to spread out, for example, if it made it much easier to stay connected to friends after doing so (so for example, facebook and twitter might be major influences on future suburban job expansion).

    3. I don't think the proportion of people who stay put for most of their lives is all that high. And my point is not about whether an individual decides to stay or relocate, it is about where the growth occurs spatially in the urban economy.

      Put simply, my point is, "out" is cheaper than "up" for most sectors of the urban economy most of the time. To make a very crude illustration, no-one would dream of forcing the RURAL sector to expand "upwards" to allegedly achieve greater efficiencies. And there is a lot of sectors in the economy that fall somewhere between "rural" and "global finance" in their level of income per unit of space required (and hence their ability to pay higher levels of economic rent).

      Sure, there might be subsidies to "sprawl", but it is possible to end these subsidies, rather than adopt blunt-instrument anti-growth regulations that do far more damage to the economy and society (via economic land rent maximisation). If we want to reduce energy use and emissions, we should tax these things and let people work out the most efficient response. Many people may adopt sustainable solutions that are a better fit with LOW density living.

  12. Anonymous11:59 AM

    Noah, many of your fellow economist freq assert that one of the reasons for high prices in the NYC/Boston corridor is that (partly due to zoning) their is a lack of land

    Doesn't this contradict your message ?

    In any event, a more interesting question is the psychology here - either, why do economist waste their time asserting idiotic ideas, or, equally plausible, why is it so difficult for non economist to accept the odd ideas of economists ?

  13. New York is an example of speculation and desire, NOT population. Prices are up but human density is way down...the actual people living (not working in Manhattan is a fraction for what it was 90 years ago. It is also, like London, a direct beneficiary of the 30-year financialization of assets, often extracted from other cities or regions. Less people with more assets competing over rising prices in a space-constrained environment. Nothing at all to do with population.

  14. You are correct. But note this economic rent captured by Manhattan property owners is nowhere near as great as the economic rent captured by property owners in the CBD’s of not just London, but nearly every UK city. Regulatory distortions to urban land markets deliver massive economic land rent even in the case of far weaker local economies.

    Cheshire and Hilber of the LSE estimate that in most US cities, there is negligible land rent; in Manhattan, the “regulatory tax” on office space is 50%; in European cities, it is more like 300%; and in the UK, it is higher still….!

    The Cheshire and Hilber paper is HERE:

    In general this discussion thread is an illustration of the blindness of even the world's brightest experts, to the connections between regulatory distortions in land markets, economic land rent, and economic (and socio-economics) outcomes.

  15. I think we licked the housing cost, food and fuel thing in the 20th century. What is unaffordable now is any form of assurance, whether of continued employment, medical care, retirement support and so on.

    1. Kaleberg, you get at a crucial point, and not just eat more kale. Like with assets, it's not just expected return it's risk -- It's really risk! The same with income. Risk, or variability, of income is hugely important, not just expected income, or expected lifetime income.

      A job that you can never lose -- tenure, schwiing! -- can be much more valuable, and utility creating, than one which has far higher expected pay, but you can easily lose it, and long term, or have a massive and lasting, or permanent, paycut, especially after you and your family have grown accustomed to that high pay.

      And a big reason for this is that positional externalities, the pink elephant of economics, are ginormous.

    2. Yes, the trouble with assurances is that they result in someone else bearing even more risk.
      Rare risks can be shared across a community at minimal cost, eg fire insurance. But very common risks, such as retiring and then going on to live a lot of years, or the risk of losing your job, are much more costly to share.

      For example, if a company promises to pay a pension to all of its retired employees and doesn't set aside money ahead of time to do so, and then suffers a fall in profits, then to keep paying the pension it has to squeeze either currently-working employees or shareholders, and shareholders are often pension funds of employees at different companies.

      If the government subsidies an industry to keep jobs going, it must either raise taxes, cut spending on something else, or run deficits. The rise in taxes causes job losses in whatever taxpayers' money was going on before the tax rise, the cut in government spending cost the jobs of whomever was employed by that government spending (and perhaps carry on jobs), and the deficits mean that eventually more and more taxpayer money goes on debt repayments rather than government services. Furthermore, the government becomes less able to respond to emergencies, like a tsunami wiping out thousands of miles of coastline settlements.

    3. Tracy W. you really underestimate just how uniquely great government often is at risk pooling and insuring. This is one of the biggest advantages that government often has over the private market, and why our government is often described as an insurance company with an army.

      With the federal government you can pool risk with enormous economies of scale and simplicity that private insurance companies often cannot even begin to match. And you can fund it with highly progressive taxes, and that just adds a lot more to the risk decreasing, as you only pay more in good or very good states, and in bad states your payment plummets, even to zero, or negative, like with the EITC.

      The federal government can also enormously decrease risk with free universal health care, free high quality universal preschool and bachelors degree (How much longer will the Republicans keep us at only the free K-12 we've had for 100 years; do they want us to be as unadvanced as 100 years ago, or take us back further to before the enlightenment?), higher social security taxes but higher pensions (uncapping both the taxes and the size of the pension you get), and so on.

      All of this can easily be funded with progressive taxation without increasing unemployment. The taxes may lower some jobs, but the spending will increase others -- less people building mansions, more building clinics and schools. And smart monetary and fiscal stimulus, and regulating rapacious finance, can basically always keep unemployment low, or at least bring it down quickly if a recession sets in.

      No the wealthy won't stop working hard if you raise their taxes -- income and substitution effects and empirical evidence, like 100 years ago people worked more hours for a fraction of the after tax pay per hour.

    4. Oh, and I left out one of the biggest advantages, no private company can insure your payout like the government, which can be legally required to pay and can always get the money through taxation or new money creation. And without the profit motive, only the civic good motive, and getting reelected, for both parties and officials, there is not the incentive to deceive you out of your claim with fineprint, an often serious risk with private insurance companies.

  16. House prices also depend on ease of credit; on the financial hedge factor.

    If a bank is willing to lend me five times my yearly income for mortgage, house prices will high up compared to a lending factor of two yearly incomes.

    We have recently been through a period of uneasy feelings towards easy credit; and house prices responded.

    On average, over the last 80 years or so, lending conditions have eased, I would think. That easing made possible by better monetary management, starting in the thirties (of the gold standard etc).

    And so house prices measure our financial peoples mindset.

    1. Actually, it is "supply" of housing that determines how many people will be able or unable to own a house; the ease of credit merely determines the price and debt level at which everyone will either get to buy, or miss out.

      Government "housing policy" that does not address "supply" when this is a problem, is not "housing policy" at all - it is mortgage finance sector profit increasing policy, and land-rentier profit increasing policy.

    2. Note that in every city with a high elasticity of housing supply, the ease of credit does not affect house prices any more than the ease of credit affects the price of cars and TV's.

  17. I think I'm in the McBride camp, and here's why. I live in the LA area so I'll use that as an example. For most people the primary consideration for location is proximity to work, with community considerations like quality of life and schools being of secondary importance. And for LA, weather has a huge impact. There are lots of people who live here who don't even have jobs just because they can.

    Now suppose you work in aerospace(aerospace is centered near LAX and in northern Orange County), you're early in your career, and don't earn enough to live in Santa Monica. You have basically two choices. You can live in an affordable neighborhood nearby, but you might not like the choices in your price range. Or for the same money you can live in a much nicer house with good schools in the Inland Empire and have a tough commute to Orange County. Thirty years ago the commute might have been manageable, because the infrastructure-to-population ratio was higher. But since then, infrastructure investment has not kept pace with population growth, and commuting costs are higher. That causes the neighborhoods closer in to be more appealing, and thus increases prices.

    The higher cost of living has had an impact on employer location decisions. While it's not likely that highly educated and compensated engineers are willing to move from So. California to Huntsville, Alabama, much of the production has moved to lower cost locations, so that tends to moderate housing prices.

  18. Bill Ellis3:11 PM

    Noah, YOU WIN !

    Paul Krugman says..."I think I call this one for Smith — although McBride’s point that actual real housing prices do seem to have an upward trend remains important, and needs explaining."

    If PK says you are right is it good enough for me. ;-)

  19. This is all reasonable but I think you've got the transit sign wrong.

    1. Actually I think it's indeterminate.

    2. Well, maybe. But I think the overwhelming historical evidence would be on the side that more transit = higher land values. The fact that it frees up some space that would otherwise be used for parking is totally swamped by the higher levels of density it allows.

      Anyway, it's just a quibble, everything else you say seems right.

    3. The effect of different transport systems on economic land rent was theorised by Robert Murray Haig back in 1923, and has been expanded on by Alonso, Wingo, and Goldberg.

      The increased value of real estate near transit stops is merely concentration of economic rent - zero sum wealth transfer.

      From one of Colin Clark’s seminal books, “Regional and Urban Location” (1982).

      (Page 231)”….If rail and subway services to the centres of large cities were charged at full cost, including interest and depreciation, two consequences would follow. The employers of the lower paid service workers in the city would have to raise their wages, in some cases reduce the services offered, or move to suburban locations (for example, some of the retail businesses still carried on in city centres). Meanwhile the higher paid salaried and businessmen, who in most cases could not change their workplace, would have an incentive to move their residences closer to the centre (while at the same time having less incentive to reside close to railway or subway stations).

      These movements would have their reflections in the price of urban land. They would reduce the demand for and the price of business land in the city centre, and of residential land in the outer susburbs, particularly land now high priced because of its proximity to railway stations. There would be countervailing movements raising the price of business land in the suburbs and of residential land nearer the city centre; but these would probably be of a lesser order of magnitude. In net effect, the subsidies on rail and subway suburban transport are subsidies to the owners of certain types of land – for which there is no social justification….”

      In case you are wondering who Colin Clark was:

      The transfer of wealth from taxpayers to property owners is surprisingly quantifiable in terms of capital gains reaped by the conveniently placed property owners. These people, of course, become a natural lobby group in favour of transit subsidies, and the whole thing becomes another example of the old, old story where the great majority of people have a little to lose “each”, but not enough to turn any one of them into a lobbyist.

      "Automobility" as a system, however, disperses rather than concentrates economic land rent. See my other comments on this thread.

  20. Alan Goldhammer4:53 PM

    Well Bill McBride and I graduated from the same high school albeit about 10 years apart so we are both brilliant(I'm the older and it's questionable if wiser). I've been tracking Calculated Risk for over six years and have seldom found him wrong either intuitively or in data driven posts. Of course this whole thing is dependent on location which is not only based on the geography but the economy as well.

    From my vantage in the greater Washington DC area there is no question that the closer in one lives the higher valued the land is (in our case our property is assessed at almost four times the value of the dwelling and most of the houses such as ours are knocked down shortly after sale so that a larger house can be built on the same property). Sure one can find lots of land in some US locations that's cheap and land in other areas that's expensive. It's not simply explained and I'm in agreement with Bill on this one and one can see this reflected in the price of stocks related to building.

  21. The value of existing homes used grow at an inflation adjusted rate of 0.75% annually. The increase was in the value of permitted Housing lots. But as housing has become a less attractive investment, people are buying smaller less elaborate homes. One formulation of housing bubbles is that the value of the house would fall at the end of the correction to the annual gain in housing plus inflation. At the end of the current housing bubble their is a 3% growth over the 15 years of the bubble. All of this growth is in a few expensive cities. Out side these cities the real growth is negative.

  22. I don't really see what Krugman's data has to do with anything. If less populated suburbs grow faster than dense inner cities, that could show up as a decrease in population-weighted density. Moreover, while it would imply that prices would probably be rising in those suburbs, we would not expect prices to fall in inner cities (which are also growing). So Krugman's data doesn't really tell us anything one way or another.

    1. Disagree. It tells us that agglomeration is not linear in time. Reduced agglomeration reduces land values. Remember that what is being discussed is land value, not land + structure value (see McBride's original post). This means that rising population is no guarantee of riding land values. In reality land values may go up, or they may go down.

  23. David S4:18 AM

    Not seeing enough comments on how local zoning constrains density. That describes much of the Boston metro region. Efforts to reduce housing costs by increasing higher density supply in the city core face significant permitting hurdles which extend project timelines.

  24. I think we licked the housing cost, food and fuel thing in the 20th century. What is unaffordable now is any form of assurance, whether of continued employment, medical care, retirement support and so on.

  25. Awesome blog over here! Thanks for sharing this very usefull information. I will visit your blog again into a couple off days to check if you have some new articles
    moving services California

  26. Tell that to the family farmer...