Home prices since 1870: No price like home

Katharina Knoll, Moritz Schularick, Thomas Steger 01 November 2014

a

A

For economists there is no price like home – at least not since the global financial crisis. Fluctuations in house prices, their impact on the balance sheets of consumers and banks, as well as the deleveraging pressures triggered by house price busts have been a major focus of macroeconomic research in recent years (Mian and Sufi 2014, Jordà et al. 2014, Shiller 2009). Houses are typically the largest component of household wealth and the key collateral for bank lending, and they play a central role in long-run trends in wealth-to-income ratios and the size of the financial sector (Piketty and Zucman 2014). Yet despite their importance for the macroeconomy, surprisingly little is known about long-run house price trends.

In our new paper (Knoll et al. 2014) we turn to economic history to fill this void. Based on extensive historical research, we present new historical house price indices for 14 advanced economies since 1870. Houses are heterogeneous assets, and great care is needed when combining data from a variety of sources in order to construct plausible long-run indices that account for quality improvements, shifts in the composition of the type of houses, and their location.

For the construction of the long-run database, we were able to build in part on the existing work of economic and financial historians such as Eichholtz (1994) for the Netherlands or Eitrheim and Erlandsen (2004) for Norway. In other cases we collected new information from regional and national statistical offices, central banks, tax authorities such as the UK Land Registry, and national real estate associations such as the Canadian Real Estate Association. We are indebted to many colleagues for their help and guidance in sorting through the historical sources.

Hockey stick pattern of house prices since 1870

The (unweighted) mean and median of the 14 house price indices are shown in Figure 1. Adjusted by the consumer price index, house prices in the early 21st century are well above their late 19th-century level, and increased in all advanced economies in the long run. Yet their trajectory is distinctive. Using the new dataset, we are able to show that they follow a hockey-stick pattern – real house prices remained broadly stable from the late 19th century to the mid-20th century, and increased strongly in the following decades. Real house prices have approximately tripled since 1900, with virtually all of the increase occurring in the second half of the 20th century, as Figure 1 shows.

We also find considerable cross-country heterogeneity in long-run house price trends. While Australia has seen the strongest, Germany has seen the weakest increase in real house prices. Moreover, in the paper we also demonstrate that urban and rural house prices as well as farmland prices display similar long-run trends.

Figure 1. Mean and median real house prices, 1870–2012

Note: CPI-deflated nominal house prices for 14 economies.

House prices and incomes

The trajectory of long-run house prices is all the more surprising since income growth was relatively stable in the long run, leading to pronounced fluctuations in the house price to income ratio. Figure 2 displays the relation between house prices and GDP per capita over the past 140 years. House prices declined until the mid-20th century relative to incomes. After World War II, the elasticity of house prices with respect to income growth was close to or even greater than 1. Finally, in the past two decades preceding the 2008 global financial crisis, real house price growth outpaced income growth by a substantial margin, as Figure 2 illustrates.

Figure 2. Mean real house prices and income per capita, 1870–2012

Booming land prices

Houses are bundles of the structure and the underlying land. In the paper, we study the driving forces of the hockey-stick pattern of house prices. Using an accounting decomposition of house price dynamics into replacement costs of the structure and land prices, we demonstrate that rising land prices hold the key to understanding the upward trend in global house prices.

While construction costs have flat-lined in the past four decades, sharp increases in residential land prices have driven up international house prices. Our decomposition suggests that up to 80% of the increase in house prices between 1950 and 2012 can be attributed to land price appreciation alone.

Figure 3. Construction costs and house prices

Note: Average construction costs and house prices for 14 economies.

The pronounced increase in residential land prices in recent decades contrasts with the earlier period. During the first half of the 20th century, residential land prices remained constant in advanced economies despite substantial population and income growth. We are not the first to note the upward trend in land prices in the second half of the 20th century (Glaeser and Ward 2009, Case 2007, Davis and Heathcote 2007, Gyourko et al. 2006). But to our knowledge, it has not been shown that this is a broad-based, cross-country phenomenon that marks a break with the previous era.

Figure 4. Land prices and house prices

Note: Land prices derived from an accounting decomposition of house price trends into replacement value of the structure and land values.

Why didn’t land prices increase before?

What explains the fact that residential land prices remained stable until the mid-20th century and increased strongly thereafter? Our explanation focuses on the different dynamics in land supply in these two periods. From the 19th to the early 20th century, the transport revolution – mostly the construction of the railway network, but also the introduction of steam shipping – led to a massive and well-documented drop in transport costs (Jacks and Pendakur 2010). An important side effect of the transport revolution was to substantially augment the supply of economically usable land.

In the paper, we develop a model with land heterogeneity to demonstrate how a sustained decline in transport costs endogenously triggers an expansion of land, such that the land price may remain low despite continuous growth of income and population. We also show that this land-augmenting decline in transport costs subsides in the second half of the 20th century, so that land increasingly became a fixed factor.

At the same time, zoning regulations and other restrictions on land use also inhibited the utilisation of additional land in recent decades (Glaeser et al. 2005, Glaeser and Gyourko 2003), while rising expenditure shares for housing services added further to rising demand for land. Yet our stylised facts are also compatible with other explanations that help explain surging land prices in the past few decades, such as growing subsidies for home ownership or easy borrowing conditions (Mian and Sufi 2014, Jordà et al. 2014).

Ricardo might have been right

Our results have potentially important implications for the much-debated long-run trends in the distributions of income and wealth (Piketty and Zucman 2014).1 We offer a lens for reinterpreting Ricardo’s famous principle of scarcity. Ricardo (1817) argued that in the long run, economic growth disproportionally profits landlords as the owners of the fixed factor. Since land is highly unequally distributed across the population, market economies produce ever-rising levels of inequality.

Writing in the 19th century, Ricardo was mainly concerned with the price of agricultural land, and reasoned that as population growth pushes up the price of corn, the land rent and the land price would continuously increase. In the 21st century, we may be more concerned with the price of housing services and residential land, but the mechanism is similar. The decline in transport costs kept the price of residential land constant until the mid-20th century. The price surge in the past half-century could be an indication that Ricardo might have been right after all.

References

Case, K E (2007), “The Value of Land in the United States”, in G K Ingram and Y-H Hong (eds.), Land Policies and their Outcomes, Cambridge: Lincoln Institute of Land Policy.

Davis, M A and J Heathcote (2007), “The Price and Quantity of Residential Land in the United States”, Journal of Monetary Economics 54(8): 2595–2620.

Eichholtz, P M (1994), “A Long-Run House Price Index: The Herengracht Index, 1628– 1973”, Real Estate Economics 25(2): 175–192.

Eitrheim, Ø and S K Erlandsen (2004), “House Price Indices for Norway, 1819–2003”, in Ø Eitrheim, J T Klovland, and J F Ovigstad (eds.), Historical Monetary Statistics for Norway 1819–2003, Oslo: Norges Bank, Norges Bank Occasional Paper 35.

Glaeser, E L, J Gyourko, and R Saks (2005), “Why Have Housing Prices Gone Up?”, American Economic Review 95(2): 329–333.

Glaeser, E L and J Gyourko (2003), “The Impact of Building Restrictions on Housing Affordability”, Federal Reserve Bank of New York Economic Policy Review 9(2): 21–39.

Glaeser, E L and B A Ward (2009), “The Causes and Consequences of Land Use Regulation: Evidence from Greater Boston”, Journal of Urban Economics 65(3): 265–278.

Gyourko, J, C Mayer, and T Sinai (2006), “Superstar Cities”, American Economic Journal 5(4): 167–199.

Jacks, D S and K Pendakur (2010), “Global Trade and the Maritime Transport Revolution”, Review of Economics and Statistics 92(4): 745–755.

Jordà, Ò, M Schularick, and A M Taylor (2014), “The Great Mortgaging: Housing Finance, Crises, and Business Cycles”, NBER Working Paper 20501.

Knoll, K, M Schularick and T Steger (2014), "No Price Like Home: Global House Prices, 1870-2012", CEPR Discussion Paper No. 10166.

Mian, A and A Sufi (2014), House of Debt: How They (and You) Caused the Great Recession, and How We Can Prevent It from Happening Again, Chicago: University of Chicago Press.

Piketty, T and G Zucman (2014), “Capital is Back: Wealth-Income Ratios in Rich Countries, 1700–2010”, Quarterly Journal of Economics, forthcoming.

Ricardo, D (1817), Principles of Political Economy and Taxation, London: John Murray.

Schularick, M and A M Taylor (2012), “Credit Booms Gone Bust: Monetary Policy, Leverage Cycles, and Financial Crises, 1870–2008”, American Economic Review 102(2): 1029–1061.

Shiller, R (2009), Irrational Exuberance, New York: Broadway Books, 2nd revised and updated ed.

Footnote

[1] We are grateful to Thomas Piketty for pointing this out.

a

A

Topics:  Economic history Financial markets

Tags:  housing, house prices, global crisis, land prices, transport costs, transport revolution, land-use restrictions, zoning laws, Inequality

Comments

I have been working on a hypothesis from a different angle to these authors, and it is encouraging to see others taking an interest in these very important questions.

I agree completely that transport system effectiveness at providing increased access to land, is responsible for reducing land rent. This is surely self-evident; not only is there is established literature about this: Robert Murray Haig (1926) “Towards an Understanding of the Metropolis” is an early example of a work that includes an intellectual approach to urban land rent and transport system flexibility; but one encounters a certain familiarity with the principle in the writings of non-specialists like Frank Lloyd Wright, Henry Ford and Charles Booth.

There is an extremely helpful overview paper that discusses this and later theoretical advances by Alonso, Wingo, and others – Michael A. Goldberg (1970) “Transportation and Urban Land Rents: A Synthesis”.

In an essay published in Quadrant Magazine (Australia) December 2013, I criticised the economics profession for its failure to recognise the realities observable in real estate markets and develop theory that fitted them. My argument is based on the observation that there are two forms of land rent – differential rent and extractive rent. I owe this insight to Alan W. Evans (Emeritus, University of Reading) – it is included in his 2004 book “Economics, Real Estate and the Supply of Land”. But I recently was amazed to discover an unjustly-obscure theoretical and modelling paper that approached the question from almost the same way as I have been working. This is Dimitri Emmanuel (1985) “Urban Land Prices and Housing Distribution: Monopolistic Competition and the Myth of the ‘Law' of Differential Rent”.

Emmanuel describes what I call “extractive rent” in his paper, but calls it “monopolistically derived minimum land price”. His approach is based on the theory of “monopolistic competition”, which is not something I had tried to apply, but it seems to be appropriate to urban land supply. In fact his insight that advanced my conceptual grasp of the problem considerably, is that “differential rent” itself becomes monopolistically derived, or extractive. This perfectly harmonises with my observations of housing market reality around the world. The successful “bids” for every attribute of housing (according to location advantage and so on) are progressively derived from the incomes of bidders higher and higher up the income distribution, as the overall urban land market is increasingly dominated by “monopolistic” effects. And the bottom of the income distribution is marked by people excluded altogether from the formal housing market.

I have been pointing out for some years that there is (and has been in the modern history of housing markets) numerous cities with a house price median multiple of 3, and that these cities are all low density and with generously sized housing – while cities that are not in this category all tend to have median multiples that are a lot more volatile, and trend around 6 and upwards, along with the housing generally being smaller and “lots” even more so (and density being higher). And in reality, the correlation between urban density/average housing space per household, and median/average housing cost, “by city”, runs in the direction of higher density/lower average housing space = higher median/average housing cost. This is the opposite of the shallow assumptions made by advocates of “compact city” planning. This is because economic land rent falls faster than additional space is consumed by households, as long as the superabundant supply of lower-cost land in non-urban uses is freely being added to the urban economy. If that superabundant land supply is denied to the urban economy, by regulations, geographic realities, or lack of automobility in still-developing economies, the opposite occurs: land rent rises faster than households average consumption of space falls, as incomes rise. Paul Cheshire at the LSE, and various co-authors, have noted that this effect underlies the long-developing crisis in housing in the UK since the 1947 Town and Country Planning Act. (I hold that this is also the reason why developing-nation cities increasingly struggle to include latecomers in the urbanisation process, in the formal housing market in cities, in tighter and tighter accommodation such as the infamous tenements of the first world in the Victorian era. Urban land rent rises with the incomes of those already there, while the rural incomes of as-yet-migrants has been left behind).

I had noted already that density zonings in the cities with growth boundaries (or proxies for them – this can be “rural” zoning, not an explicitly enacted UGB) affected house and lot sizes but not the “price”; and concluded that the evidence is (as Emmanuel says) that site rent is elastic to density – only under the conditions of growth containment. For example, lots and house sizes in Boston and Santa Clara are very large on average yet their median multiples are no higher than cities in, say, the UK where permitted densities of development are several times higher and housing unit size several times smaller.

So I have been criticising advocacy that blames low density zoning as a cause of affordability problems, because this is contrary to the evidence, as is the assumption that upzoning and building “up” will provide “affordability” as fringe growth is explicitly constrained in new “Plans”. I have corresponded with a lot of academics about this; I believe Bertaud, Cheshire and Levinson are among those who are clearer on this reality, but no-one has been aware that there is already a theory as thorough as Emmanuel’s one.

But Emmanuel does not emphasise as I have, that there is a data set of cities where the process of automobile-based growth and the dispersion of employment and amenities has been so uninhibited that land rent has ended up being almost entirely of the “differential” type undistorted by extractive effects. The urban land rent curve tends to have steadily lowered and flattened in cities that have evolved along these lines. William Wheaton, in “Commuting, Ricardian Rent and House Price Appreciation in Cities with Dispersed Employment and Mixed Land Use” (2002) observes that these effects act in the direction of reducing land rent and reducing commute lengths, even if rebound effects obscure these forces.

In my Dec 2013 Quadrant essay I suggest that in the automobile-based free-spreading cities, housing is marked by “consumer surplus” – that is, as is the case with most goods in a free market, they tend to incorporate more and more attributes for a lower and lower real price or share of income. I believe that this is the case for any city with a house price median multiple of around 3 – which Demographia and indeed the UN and other global institutions regards as the gold standard for “affordability”. I hold that the reason for this is that urban fringe and “splatter” development is taking place so competitively that the resulting derivation of the urban land rent curve in those developments, is anchored in a fringe land “rural price plus cost of development plus a moderate profit”. This is in stark contrast to the “planning gain” noted to exist on the fringes of UK cities, of a factor of 100 to 900….!!

“Option values” mean that the “differential rent” relative to the value of fringe urban land, is what determines the urban land prices everywhere in the city, including the centre. There is a recent paper from New Zealand (Productivity Commission, 2013) which includes a time series of graphs of urban land rent in the city of Auckland during the recent years of historically unprecedented “house price inflation”, and this rent curve rose along its entire length, with the size of the “discontinuity” at the growth boundary increased – and this episode followed the enactment of growth containment policies.

Consumer surplus and extractive rent are the opposing manifestations of the same phenomenon. Where extractive or monopolistically derived rent in urban land exists, it is elastic to not just density, but all other attributes of housing. Increased density, cheaper construction, sacrificing of attributes of housing, will merely create more site rent, as will the enduring of un-renewed and un-maintained housing conditions. Extractive economic rent rises to take out any slack in “housing cost” that has occurred in any other attribute of it.

I am not surprised that there are wide disparities in the different markets analysed by Schularick and colleagues, but I am surprised that so many of the markets demonstrate a rising price of houses in the second half of the 20th century. My understanding was that home ownership increased considerably in most first world countries because the real price of housing fell; and furthermore, that the reversal of this trend is related to the more recent global mania for “constraining” urban growth. There were several decades where in fact government positively enabled and promoted urban growth with infrastructure investment and planning. This was understood to be serving the objective of home ownership, and improved living conditions relative to pre-automobile dense urban living. Certainly the most volatile period of house price inflation is the IPCC-era anti-automobility phase of urban planning fashion.

There is another complicating factor, though, which is discussed by Goldberg (1970) – which is that consumption of the attributes of housing is in fact price elastic to such an extent that households willingly spend a higher proportion of their income than previously, on “housing”, when they are getting value for money in those attributes. I would suggest that there is a kind of s-shaped curve relationship in housing markets where this is the case.

So there is a mixture of two completely different “economic rent” effects at work in the apparent increase in housing prices in the second half of the 20th century. If we were to focus completely on land rent per unit of land space I would expect to see that this was falling in all markets during the era in which automobile based suburban development was a major phenomenon. I know the UK is the exception that proves the rule.

But it gets worse. From the McKinsey Institute's latest Global Report on Affordable Housing: in London, 45 percent of land with permission to be developed remains idle. The UK's waiting list for social housing has 1.8 million people on it; a pitiful 98,000 new units were constructed in 2012; and 400,000 sites with development permission remain undeveloped!

The reason for this is that in these “created scarcity” urban land markets, site owners are thinking like speculators, not like “producers”. Why should they even bother to develop their site to maximum potential, or sell it, when its value already embodies the “rights of development” and that value is going up? (And when the crash comes and the value is down, no-one is interested in doing development).

Where building “up” did lower floor rents for some decades and still does – is in the presence of competitive fringe development that brings housing onto the market at the lowest cost that developers in competition with each other can do so on rural land acquired at minimal uplift over rural values. “Option values” from the fringe to the centre takes care of the rest of the urban land market. Houston, Dallas and Indianapolis are examples of cities today where impressive levels of intensification in the right places is occurring, and this is all occurring for good sound functional reasons, plus the capital available for building “up” is higher due to the very low site acquisition costs. Manhattan’s famous skyscraper boom and economic rebalancing from manufacturing to “financial services” occurred under similar market conditions – i.e. rapid, transport-improvements-based urban area spread. Manhattan floor rents are still significantly lower than those of any other “global” city (with the possible exception of Tokyo, due to factors that require another essay in its own right). The value for money in housing options at any given level of travel time to central Manhattan, is far higher than in any other global city and even than that in many secondary cities where very stringent growth containment policies apply, such as in the UK and now Australia.

Automobile based development is more effective than rail based, for reduction of urban land rent, because rails only bring long ribbons of land into supply; the distance/land price trade-off is steeper. Haig (1926) used the term “friction” regarding the relationship between the transport system and land use. Obviously a door-to-door fast transport method is much lower-friction than one requiring a walk at both ends, waiting times, and possibly transfers as well – especially if we are considering access to “the entire urban economy” from any given point.

Glaeser makes some interesting comments in “Nation of Gamblers” (2013) regarding this.

“……Almost everywhere, prices in 1970 were below 1950 prices plus this construction cost related price increase. Even after the most stupendous change in America’s mortgage history, and a post-war economic boom, housing prices had gone up less than construction costs would warrant.

The natural explanation for the missing boom in prices after World War II is that there was an enormous increase in housing supply over the same time period. During the 1950s, America permitted 11.84 million housing units, which is roughly the same as America permitted during the twenty-six years from 1920 to 1945. The construction was disproportionately on the urban fringe (Jackson, 1979) and disproportionately in the Sunbelt.

The post-World War II era demonstrated exactly what textbook economics predicts should happen when robust demand meets relatively elastic supply. Quantities rose and prices stayed relatively flat. The relatively elastic supply owed much to the rise of automobile-based living on the urban fringe, which can be seen as either a shift in housing supply or a change in supply elasticity. For example, in an open-city formulation of the Alonso-Muth-Mills model, with supply costs that increase with density, lower transportation costs will increase supply but not change supply elasticity. Yet it is possible that the automobile made supply more elastic as well. On the urban fringe, lower cost, low density housing can be built in massive quantities, essentially using a constant returns-to-scale technology……

“……..The missing post-war price boom is not a problem for conventional economics, but it does present a challenge to those who seek to explain bubbles as the outcomes of a stable process where readily observable exogenous variables translate into the presence of a bubble. The 1950s had easier credit for homeowners than the 1920s and economic conditions were at least as good. Any model that suggests that there is a stable relationship between either of those variables and price bubbles has difficulties with this epoch……”

Another extremely interesting discussion is by Nicholas Crafts (University of Warwick) on this site:

http://www.voxeu.org/article/escaping-liquidity-traps-lessons-uk-s-1930s...

“Escaping liquidity traps: Lessons from the UK’s 1930s escape”

Nicholas Crafts 12 May 2013

Professor Crafts points out that monetary easing in the UK in the 1930′s, WORKED because it had somewhere productive to GO:

“…….Obviously, for the cheap-money policy to work it needed to stimulate demand – a transmission mechanism into the real economy was needed. One specific aspect of this is worth exploring, namely, the impact that cheap money had on house-building. The number of houses built by the private sector rose from 133,000 in 1931/2 to 293,000 in 1934/5 and 279,000 in 1935/6 – many of these dwellings being the famous 1930s semi-detached houses which proliferated around London and more generally across southern England. The construction of these houses directly contributed an additional £55 million to economic activity by 1934 and multiplier effects from increased employment probably raised the total impact to £80 million or about a third of the increase in GDP between 1932 and 1934. House building reacted to the reduction in interest rates and also to the recognition by developers that construction costs had bottomed out; both of these stimuli resulted from the cheap-money policy (Howson 1975).

Why was house-building so responsive in the 1930s? Two factors stand out. First, the supply of mortgage finance grew rapidly and became more affordable in an economy in which there had been no financial crisis that curtailed lending.

Building society mortgage debt rose from £316 million with 720,000 borrowers in 1930 to £636 million with 1,392,000 borrowers in 1937 when about 18% of non-agricultural working-class households were buying or owned their own homes. In these years, deposits fell in some cases to 5% and repayment terms were extended from around 20 to 25 or even 30 years reducing weekly outgoings by 15% (Scott 2008).

Second, houses were affordable to an increasing number of potential buyers.

85% of new houses sold for less than £750 (£45,000 in today’s money). Terraced houses in the London area could be bought for £395 in the mid-1930s when average earnings were about £165 per year. Houses were cheap because the supply of land for housing was very elastic which in turn meant that there was no incentive for developers to sit on large land banks. Underpinning the availability of land for house-building was an almost complete absence of land-use planning restrictions which applied to only about 75,000 acres in 1932 – the draconian provisions of the 1947 Town and Country Planning Act were still to come……”

It has been noted by the late Sir Peter Hall and others, that the houses built in Britain in the 1930’s tend to be highly sought after today for their qualities, which have been increasingly absent in housing developed since.

The importance of these issues justifies a reversal of the current inverse level of theoretical clarity regarding them.

Doctoral Candidate in Economics/Economic History, Free University of Berlin

Professor of Economics, University of Bonn and CEPR Research Fellow

Professor of Economics, Leipzig University

Vox Talks