Wednesday, 28 May 2014

House prices, rents, and supply

In a previous post I argued that high house prices, in the UK and perhaps elsewhere, could simply reflect lower expected long term interest rates. This had some people puzzled, because it appeared to ignore supply and demand. Surely it is higher demand combined with inelastic supply (supply that is insensitive to changes in prices) which is to blame for high house prices? This short post clarifies how both views can be correct at the same time. I also raise a question about housing wealth and inequality at the end.

The easiest way to think about this, at least for me, is to imagine no one owned their own home. Everyone rents, and houses are owned by landlords. Rents represent the price of ‘housing services’, which is the flow of benefits we get by having a roof over our head. If we calculate these rents in real terms, we have the relative price of housing services compared to other goods. Real rents will reflect the supply and demand for these housing services. If we all suddenly decided we wanted to rent a house in the countryside as well as our house in the city (or vice versa), and if the supply of houses did not increase, rents would rise dramatically until enough of us thought that maybe this wasn’t such a good idea after all.

Suppose real interest rates fall, but the supply of housing is fixed. There is no particular reason why lower real interest rates should change the demand for housing services relative to other goods. Then with no change in demand or supply, rents do not change in real terms. But house prices will, because they are - to use a bit of jargon - the present discounted value of future rents, where the discount rate is the real interest rate. This is a shorthand way of saying that lower interest rates mean that investments in financial assets yield lower returns than the same amount invested in housing, so investors will want to own houses rather than financial assets. This pushes up house prices until the rate of return on both types of asset are equalised.

That assumes housing supply is inelastic, which is a reasonable assumption in the short term. However suppose by some means (economic or political) these higher house prices generated an increase in the supply of houses to rent. If the demand for housing services is unchanged, greater supply will begin to push down rents. Falling rents push down the yield from owning housing assets. As a result, housing becomes less attractive as an asset, and prices begin to fall.

So if housing supply was very elastic, permanently lower real interest rates need not lead to higher house prices in the long run. Instead, they could produce much lower rents, because a lot more houses get built. Such an outcome seems unlikely in a country like the UK, but it could happen in a country like the US where there is plenty of land available to build on. This is one possible reason for the different trends in house prices in different countries that I commented on in my previous post.

So my original post was certainly not suggesting that increasing housing supply would have no impact on prices. What it was suggesting was that in evaluating whether current prices represent a bubble, we need to allow for the possibility that high prices today reflect a view that real interest rates may stay low for some time.

If interest rates do stay low for some time, and this does keep house prices high, an interesting question is why this matters. Take the case where everyone rents. Rents are unchanged, so those renting are no worse off. Landlords appear richer, but their future income in real terms has not changed. If people own their own houses, their houses have not suddenly got bigger or better. This is related to, but is not quite the same as, a question recently raised by Chris Dillow. It is different because it potentially applies to anyone who owns assets that get more valuable simply because interest rates fall, and not because the future incomes they generate increase. It is the same issue that is raised when some complain that Quantitative Easing, by - say - raising share prices, is benefiting the rich. I think this change in wealth does matter, as this evidence suggests, and I hope to explore the reasons why in a future post.


13 comments:

  1. Very useful. In Canada with relatively anchored expectations about inflation and interest rates and poor growth prospects in the populous areas, it is a very appealing idea. The issue of changing expectations worries me. What happens when the Fed moves?

    ReplyDelete
  2. There are two points here, as always brought to mind by my having read as much Shiller as I can.

    First, as it is difficult to get quickly in and out of the housing market unlike other assets, then who are the investors pushing up these prices as they were not finance-types in the last housing bubble - they provided the loans but not the impetus for housing as an asset class?

    Second, Shiller's point about pyramid schemes and housing bubbles is that there was a switch in the mid-1970s California and early 1980s UK that saw global and national explanations be used to justify home price rises, not local factors (new railway or factory, say) as was the case before. People in the UK know it was the US housing bubble that burst first, not the UK one, and that could explain the UK's quicker pick up rate, especially with immigration stories being big in the press and justifying the idea that an overcrowded nation should be used to justify the idea that someone will always want to buy the house I have just bought. This is the London glamour city pyramid scheme effect.


    ReplyDelete
  3. Could this explain changes in oil prices?

    ReplyDelete
  4. "So if housing supply was very elastic, permanently lower real interest rates need not lead to higher house prices in the long run. Instead, they could produce much lower rents, because a lot more houses get built."

    What do you mean here when you say elastic housing supply? Isn't the correct framework for understanding housing supply real options? If I recall, the price LEVEL is irrelevant for supply decisions, it is merely the rate of change of expected value developed vs undeveloped that enters the decision.

    Moreover, why would the price level of the asset affect supply decisions? It's like saying Exxon share price determines whether they build an oil rig. Doesn't the price of oil and the cost of the rig determine whether they invest?

    ReplyDelete
    Replies
    1. The way investments are compared is by calculating their respective NPV's. A more or less permanently lower interest rate will greatly increase the NPV from building houses while giving most other investments only slightly higher NPV's. Clearly, this raises the value of the house. The reason for this is that house owners receive an implicit cash flow that lasts long into the future. The higher the interest rate is, the less such a cash flow will be worth and vice versa.

      That is why the elasticity of housing supply is important. If it were possible to build houses fairly quickly in attractive locations, the higher NPV would lead to a lot of houses being built which would push down the NPV again since the implicit cash flow becomes smaller.

      Delete
    2. "A more or less permanently lower interest rate will greatly increase the NPV from building houses "

      Why? Didn't Simon say that lower interest rate increase value, but that would mean it increases the land values as well in proportion. So a house that can be built for $100k costs $200k because of the land value is $100k. Interest rates go down, and now that house is worth $300k. But can still be built fro $100k. So the land value goes up to $200k. What changes? Nothing except that the current owners of land make money for nothing. They still don't have incentives to sell or build.

      That's why the real options framework is critical here.

      Delete
    3. I think I see what you mean now. However the high NPV comes from people actually living on the land. The person who builds an apartment complex and rents out the apartments will receive monthly rents. These rents are worth when the interest is low because the discount we apply to future rents becomes smaller.

      Someone who simply owns land will not be getting these rents. He will therefore have a much stronger incentive (than before the interest decrease) to either build something on it or sell it.

      Delete
  5. Yes, by a simple argument from arbitrage, lower rates increase the prices of all assets which can generate a future cash flow. They translate the demand curve. The supply curve staying where it was, the new point of intersection will be at a higher price. As Mr. Wren-Lewis points out, this does affect wealth inequality by increasing it. A priori a bad thing.

    ReplyDelete
  6. I've seen data suggesting that, at least in the US and in places that are not space-constrained (e.g., Manhattan), the (long-run) supply curve of owner-occupied housing is, essentially, horizontal. And the Case-Shiller data sets more-or-less confirm this--the long-run trend of inflation-adjusted housing prices is, essentially, flat. (http://en.wikipedia.org/wiki/Case%E2%80%93Shiller_index)

    ReplyDelete
  7. There are other things like eg borrowing capacity (collateral value) that are important.
    Imho UK banks should reduce mortgage multipliers when prices rise much faster than incomes.
    Otherwise when the RRE market turns they would have big problems.

    As I see it the UK has a permanent supply shortage of RRE, which gives a permanent push upwards (everything else remaining the same which it doesnot of course).
    And in general housing has a high priority.
    Plus now lack of yield on almost everything else and QE money seeking yield.
    Lot of unbalances at the moment.

    It is not so much a bubble that is dangerous. It is a considerable drop in value of an important assetclass (or more than one) while these are debt financed. Of course bubbles are more likely to see a drop invalue.
    Things that kick the collateral value out from under loans and at a large scale are imho the main danger.
    The higher prices get the riskier it gets. There is the main danger. Chance of downward correction increases and potential drop invalue/price increases as well.
    Usually these things happen with other negative stuff like economic problems.

    Cash paid investments not required for spending now or in the future (pensions), drop in value isnot very important.
    Heavily debt financed assets at the max of people's borrowingcapacity and/or yield required for spending now or in the future. Drop in value of the asset class is extremely dangerous for the economy as a whole (unless the assetsclass is small, say stamps, wine ot things like that). But RRE is huge.

    An economy largely depending on debt like the UKs should be very careful with this, if not you have all the ingredients for a new financial crisis in house.

    NB Now we see as well a safehaven value playing. Basically an insurance premium to get your investment back.

    ReplyDelete
  8. Storewars News29 May 2014 22:44

    Nice read! Very informative. Something interesting: Tesco and China Resources Enterprise reach retail deal. Read it here: http://bit.ly/1hDOqkT.

    ReplyDelete
  9. The point is that people borrow to buy a house. From a bank. They are able to borrow more when interest rates get down. At least: when deregulation allows them to borrow more. And they do want to borrow more, as they have to pay back less cash, every month. Most people do not calculate discounted value. They look at cash flows. That's a fact of life. That's what happened in the Netherlands: interst only mortgages (very, very popular for some time) decreased cash flows and increased borrowing - and therewith house prices. Banks did not have any problem with this, as they did not expect prices to be lower in the future - economic models explaining houseprices by (a) interest rates, (b) income and (c) lagged houseprices of one and two years before explained developments very good (as household income hardly changed this variable was sginifican in a statistical but not in an economical way). Everything looked really fine. Until mr. Minsky came along, in October 2008. And banks did not want to lend anymore. And people did not want to borrow anymore. And house prices went down (wow. it took economists almost three years to acknowledge that fact, really!). The point: money matters. Banking matters. Discounted imputed rents don't. Market prices in France are much higher than discounted assumed future rents. http://spire.sciencespo.fr/hdl:/2441/30nstiku669glbr66l6n7mc2oq Look here for the disbelieve among Dutch economists when, partly due to a 'sudden reversal' of mortgages, house prices crashed http://www.paecon.net/PAEReview/issue55/Knibbe55.pdf

    On a meta level, this of course shows why we can't estimate the value of capital in a general sense as the discounted value of expected future earnings.

    Merijn Knibbe

    ReplyDelete
  10. Simon,
    Your comparison with the US seems to suggest that housing prices in the UK are high because there is not enough land to build on. What do you say to Paul Cheshire's argument that the short supply of housing is "caused primarily by policies that intentionally constrain the supply of housing land" (http://blogs.lse.ac.uk/politicsandpolicy/archives/41540)? Sounds very plausible to me given how much land in the UK is still undeveloped. It's hard to believe that the fact that British homes are among the smallest and most expensive in Europe is really justified by population density. Cheshire's story also makes sense politically: Most voters own homes and therefore have a financial interest in high house prices. Hence, presumably, the conservative media's vociferous opposition to changes in planning laws (e.g. http://www.telegraph.co.uk/earth/hands-off-our-land/).

    ReplyDelete