In the last post I noted that real estate has not been a particularly good long-term investment, except in the hottest markets, and even there has lagged the S&P500 over the last 25 years. In his research, Shiller has constructed an index based on repeated sales of the same house. The index shows that over very long periods of time (like 50-100 years) housing barely outperforms inflation - house prices do not even keep pace with average income growth (i.e., GDP growth), let alone share prices of companies. Although this is contrary to conventional wisdom, it is what one might expect for an inert asset in a place like the US where population densities are still low. Otherwise, families would be forced over time to spend larger and larger fractions of their incomes on shelter.
But Mr. Shiller has unearthed some rare historical housing data for other countries. Using old classified advertisements, he was then able to fashion a chart for the United States that goes back to the 19th century.
It all points to an unavoidable truth, he says. Every housing boom of the last few centuries has been followed by decades in which home values fell relative to inflation. Over the long term, the portion of income that families spend on their shelter stays about the same.
Builders become more efficient, as they are doing today. Places that were once sleepy hinterlands, like the counties south of San Francisco or a patch of desert in southern Nevada, turn into bustling centers that take pressure off prices elsewhere. Even now, the United States remains a mostly empty nation.
"This is the biggest boom we've ever had," said Mr. Shiller, who bought into the boom himself in 2002, with a vacation home near one of Connecticut's Thimble Islands. "So a very plausible scenario is that home-price increases continue for a couple more years, and then we might have a recession and they continue down into negative territory and languish for a decade.
"It doesn't even attract that much attention," he continued. "There will be many people thinking it was a soft landing even though prices may have gone down in real terms by 40 percent."
steve wrote, The index shows that over very long periods of time (like 50-100 years) housing barely outperforms inflation - house prices do not even keep pace with average income growth (i.e., GDP growth), let alone share prices of companies.
ReplyDeleteI saw this awhile ago, and I flat-out don't believe it.
If it were true, the price of housing as a fraction of personal income would tend to zero in the long run, which hasn't happened.
Does Shiller even consider the "dividends" thrown off by owner-occupied housing?
--SJF
I don't know the details of Shiller's methodology, but his result seems believeable to me, especially if one accounts for the expenses of maintaining and upgrading a home over that length of time.
ReplyDeleteWell, there are many commodities with fixed total supply (e.g., copper or iron) whose prices don't keep pace with GDP growth unless scarcity is encountered. Their prices merely increase (on average) with the commodities inflation rate, which is typically not so different from the overall inflation rate.
ReplyDeleteIn most of the US, land is not yet scarce. In the previous post (NYTimes shows OFHEO data), you can find cases of real estate markets that appreciated even faster than GDP, but they are exceptional. When you average over the whole country you probably come out below GDP growth rates. (Note the OFHEO numbers are in nominal, not real, dollars. Most of the country had zero or only slightly positive real appreciation over the last 25 years.)
I was a bit surprised to see zero real appreciation in Shiller's data - I was expecting something slightly positive. There is actual scarcity in certain areas (Manhattan), and that will boost the average appreciation rate. Of course, at the level of 1% error there are a lot of factors to disentangle, like improvements to the properties.
steve wrote, Well, there are many commodities with fixed total supply (e.g., copper or iron) whose prices don't keep pace with GDP growth unless scarcity is encountered. Their prices merely increase (on average) with the commodities inflation rate, which is typically not so different from the overall inflation rate.
ReplyDeleteBut in the long run, the market can find substitutes for these commodities.
Not so land. Land can be used more efficiently, of course, but that's hard to do right now because of zoning regulations.
While land isn't scarce, the supply of land is still quite fixed---particularly land near desirable jobs and schools.
--SJF
anonymous said, especially if one accounts for the expenses of maintaining and upgrading a home over that length of time.
ReplyDeleteRoughly speaking, the value of improvements on a parcel of land will depreciate to zero in a few decades' time.
However, I claim that the value of the parcel will increase roughly with GDP, at least for a good chunk of urban land.
Note also that the value of the capital stock in productive industry also depreciates with time, probably much faster than residential buildings.
Again, I haven't seen an answer (anywhere, not just on this blog) to the question of whether Shiller's numbers take into account the "dividends" thrown off by housing. I can't believe they do.
--SJF
steve said, Well, there are many commodities with fixed total supply (e.g., copper or iron) whose prices don't keep pace with GDP growth unless scarcity is encountered. Their prices merely increase (on average) with the commodities inflation rate, which is typically not so different from the overall inflation rate.
ReplyDeleteFurthermore, you can track that these commodities have aggregate input costs that decline as a fraction of GDP. Again, I don't buy that in the case of housing. The cost of housing seems to do a good job, over the long haul, of remaining above a fixed percentage of personal consumption. Since personal consumption keeps up with GDP, housing costs must also.
Note this is definitely not the case for basic food staples.
--SJF
SJF,
ReplyDeleteI'm not sure all your comments are making it through the system. Do you notice any missing comments?
Regarding dividends, I have a feeling those might be in the S&P data but not in Shiller's index. If you take a 3-5% return on rent (net of property taxes, expenses, etc. and using a 15-20 price to rent ratio), then that would give a total return on real estate which is closer to GDP (perhaps even higher than GDP), but still lagging the equity markets.
Since equities are more volatile than real estate, it would be surprising if their total returns were not higher. The risk-adjusted returns might be similar...