Posted on by Colin Twiggs

Investing in Real Estate

In Monday’s update, we compared investing in stocks to investing in financial securities and concluded that stocks offer better long-term performance. Today, we use long-term data series for US and Australian real estate to evaluate their comparative performance.

US real estate data was sourced from Prof Robert Shiller, who created the Case-Shiller Index series. The chart below shows that US home prices from 1933 to 2023, a period of ninety years, appreciated to 69.5 times their original value.

Home Price Index

Adjusting for inflation, we get real appreciation of 2.9 times.

Again, CPI seems to understate inflation. Comparing the home price index to Gold, rather than CPI, provides a more accurate measure of appreciation in real terms. Gold appreciated 94.5 times over the same period, so the home price index actually lost value.

Calculation: 69.5/94.5 = 0.735 (i.e. a 26.5% loss of value)

Comparing data sources

A second source of home price data compares median prices, based on sales of existing homes, from 1953 to 2023. That shows growth of 22 times over the past seventy years, which is close to the Home Price Index appreciation of 21.25 over the same period.

Median Home Prices, Existing Homes

Australia

Australian housing data is harder to come by but we found an excellent source of long-term median house price data in the 2007 UNSW thesis of Dr Nigel David Stapledon. Using data from within the thesis, we were able to adjust nominal house prices to reflect constant quality (house values with no improvements) below.

Median Home Prices, Existing Homes, Constant Quality

Australian house prices appreciated 191.4 times between 1933 and 2006.

Unfortunately the data ends there, so we had to calculate a weighted average of median houses for 2007 to the present.

CoreLogic kindly provided us with values for their hedonic Home Value Index (for 5 Capital Cities) which also adjusts for quality:

Property Type	31/01/2007	31/07/2023
Houses		$399,182	$891,747

The gain of 2.23 is slightly higher than the 2.14 calculated from weighted average data for the 8 capital cities provided by the ABS.

Median Home Prices, Weighted Average of 8 Capital Cities

We opted for the higher figure from CoreLogic as likely to be more in line with the earlier Stapledon data. That gives a total nominal gain, adjusted for quality, of 426.8 for the ninety years from 1933 to 2023.

The price of Gold fines was fixed at £6-3/9 per troy ounce fine according to the Sydney Morning Herald on 2 January 1933, that converts to 12.375 Australian Dollars. Total gain for Gold in Australian Dollars over the past ninety years is therefore 232.9 times (A$2881.90/12.375).

We calculate the real gain for Australian house prices as 1.83 times over the past ninety years (426.8/232.9).

Conclusion

The US Home Price Index lost 26.5% in real terms, over the past ninety years (1933 – 2023), when compared to Gold.

The S&P 500 appreciated 6.8 times over the past ninety years when measured against Gold, and 9.7 times compared to US real estate (Home Price Index).

Using Gold as the benchmark, we conclude that Australian real estate prices appreciated faster than US real estate over the past ninety years, growing 1.83 times in real terms, whereas the US depreciated to 0.735 of its original real value.

We suspect the difference is largely due to the substantial fall in US real estate values after the 2008 sub-prime crisis, whereas Australian home prices continued to grow. We expect that performance of the two will converge in the long-term.

Lastly, when measured against Gold, US stocks outperformed Australian real estate. The S&P 500 grew 3.7 times against Australian home prices, in real terms, over the past ninety years.

This does not mean that we should ignore real estate as an investment medium. But a portfolio concentrated in real estate, without diversification into stocks and precious metals, could underperform in the long-term.

Acknowledgements

7 Replies to “Investing in Real Estate”

  2. It would be interesting to see how US major coastal city house prices compared with Australian major coastal city house prices. I’ll bet they fared comparably, and the US cities may have done better. Overall real decline in US house prices reflects in part the long term stagnation of housing in the heartland thanks to the deindustrialization trade policies of the 1980s and 90s. For many years you couldn’t give away houses in Detroit. Those same polices may have changed what jobs Aussies did, but did not have such a disparate geographic impact.

    1. The Case-Shiller 10-City Composite Index (focused on major metropolitan areas) is 6% higher than the National Home Price Index over the period 1987 – 2023. Misguided trade policies are no doubt a factor in some areas but do not have a marked influence on the national average.

      https://fred.stlouisfed.org/graph/?g=17KtI

      1. The scariest chart to me is Australian household net wealth relative to disposable income. Dwellings and net wealth are growing at far faster rates than income. These should be flat lines (i.e. a constant ratio). At some point we are likely to witness reversion to the mean.

        ABS: Household Net Wealth relative to Disposable Income

    1. Yields are not included. The S&P 500 dividend yield has fallen from about 4% to 2.0% in the last few decades (since the growth of buybacks). Gross residential yields for dwellings in the 5 combined capitals, in the CoreLogic index, were 3.6% in June 2023. I have no data for net yields but think 75% would be reasonable, giving a net yield of 2.7%.

Comments are closed.