Table of Contents

Long-term market overview

The Case-Shiller Housing Index http://en.wikipedia.org/wiki/Case-Shiller_index measures price appreciation. It is not inflation-adjusted but it is calculated in a way that makes it one of the best summaries of how equivalent houses have appreciated over time.


1987 is as far back as the index goes, 2006 was the peak of the US housing market, and 2009 is when I wrote this article.

Here's some inflation data measured by CPI (note: CPI is heavily weighted toward housing, so I'm not sure about implications of using CPI here. Suggestions for improved methodology welcome)


From 1987-2009 (22 years)

If we look at at average US real estate housing appreciation for equivalent houses over 22 years we get a real return of 0.4%. Here's how you calculate it:

Compound annual growth rate of 3.3%:

((128.81/62.03)(1/(2009-1987))-1)*100 = 3.3% (note: @ here means ^)

Inflation of 2.9%:

((212.174/111.400)(1/(2009-1987))-1)*100 = 2.9%

So the real return was 3.3-2.9 = 0.4%

Note: This is for the entire US and is an average.

From 1987-2006 (19 years)

Of course the market is not great in 2009. If you look at the peak, 2006, then the return is better, but still less than many people expect.

Compound annual growth rate of 6.0%:

((188.66/62.03)@(1/(2006-1987))-1)*100 = 6.0%

Inflation of 3.1%:

((199.400/111.400)@(1/(2006-1987))-1)*100 = 3.1%

So real return was 6.0-3.1 = 2.9%


In http://www.youtube.com/watch?v=8IR5LefXVPY Sal Khan looks at factors influencing demand and supply of housing. He claims that for the price of something to go up demand has to increase faster than supply. He identifies population growth, income growth, and housing growth as being the key factors influencing supply and demand. I've collected long-term data for each factor.

Population growth can be measured by new households.

I found data from http://www.economagic.com/em-cgi/data.exe/cenHVS/table13c01

YearQuarterHouseholds (in thousands)

Long-term population growth (as measured by households, which are generally the unit that occupies houses) from 1987 to 2009 looks to be 0.8%.

It can be calculated like this:

((111079/92699)(1/(2009-1987))-1)*100 = 0.8%

Income growth

Income growth from 2000 to 2004 was -3%. The following data is from census.gov but I lost the link.

YearIncome (2007 dollars)

Income growth from 1987 to 2006 was 1% and can be calculated like this:

((68459/56587)(1/(2006-1987))-1)*100 = 1.0%


Housing units completed

I found from http://www.census.gov/ftp/pub/const/co_cust.xls that from 1987 through 2006 there were 30,146,000 new housing units built in the US.

In 1993 there were 106,611,000 total housing units (see http://www.census.gov/prod/1/constr/h121952c.pdf). In 1987 there were 8,177,000, meaning from 1987 to 1993 there were 98,434,000 (106611k-8177k) units built. So in 2006 there were 98,434k + 30,146k = 128,580k (huh? I calculated this a while ago and it doesn't make sense to me now so I'm not sure it's right).

((128580/98434)@(1/(2006-1987))-1)*100 = 1.4%

Combining Supply and Demand

Roughly over the last 20 years households increased 0.8% per year and income increased 1.0% while housing supply increased 1.4%.

I don't know if these factors can just be added or what the correct interpretation of the result is, and I welcome clarification from readers, but I'll note that 1.0+0.8-1.4=0.4% which was the real compound annual growth rate from Case-Shiller on a house from 1987-2009. Like I said, I don't know if this supports these being the main predictors of housing prices or if it is just coincidence.

An effect of easy credit

Sal Khan (http://www.youtube.com/watch?v=wYAhlTHIBT4) points out that relaxed mortgage requirements have recently allowed millions of people without enough for a down payment to buy homes, increasing house prices.

Down but not necessarily bottom

I think it's important to remember that just because housing prices are down over the last couple years does not mean they are at the bottom and set to increase in the short or even long-term.

Functions of population, income, housing supply, credit policies, etc., not history, will determine future housing prices.

Keep in mind:

  • Mortgage readjustments in upcoming years and their effect on foreclosures
  • Job outlook

Costs of buying

Some costs that people often overlook:

  • Closing costs
  • Property tax
  • Insurance
  • Interest
  • HOA
  • Maintenance

None of these expenses increase the value of the home.


With a long-term real return of 1% or less, residential real estate on average doesn't look so hot.

I think what possibly makes residential real estate more attractive is leverage. With leverage, meaning you can go into debt with someone else's money to buy the home, if the house appreciates and you hold onto it long enough, you can make that 1% on a principal significantly greater than your outlay.

While other investments offer much higher average returns than residential real estate, it is much more difficult to leverage. I wonder if there are any institutions that make long-term loans that are at an interest rate that makes buy-and-hold investing in an S&P 500 index financially viable.


The sales pitch for rentals sounds almost like an infomercial. It goes something like this : "Stop throwing your money away on rent. Buy a house and start making money. You don't need a down payment or any special knowledge, it's easy. There are tons of renters and you won't have to do anything. They'll pay your mortgage for you. When you sell you'll have built equity and the house will have appreciated, making you tons of money."

For there to be a high return, there generally is one or more of the following:

  • high risk
  • high amounts of labor
  • high amounts of capital
  • high amounts of knowledge/experience/innovation
  • high amounts of luck

Assuming an efficient market and a potential rental purchaser with no special knowledge, little capital, and average luck, in order for a landlord to experience high returns they're either bearing a high amount of risk or they're doing a lot of work.

I think the market has been somewhat inefficient, but also labor and risk have been underappreciated. Landlords have understated the amount of time and energy they put into maintenance and finding tenants. Landlords also fail to recognize many risks that they bear, including that a housing price correction or bear market can leave them with an underwater mortgage, illiquidity, inability to rent, or credit impacts.

Date: 2009-08-31