Farmland as Portfolio Diversification?

In 2013 Martijn Cremers published a paper called The Performance of Direct Investments in Real Assets: Natural Resources, Infrastructure and Commercial Real Estate. The paper analysed the benefits of owning a number of different investments and compared them to the common stock/bond portfolios that many investors are familiar with. The result is that a historical investment in farmland looked like a big winner.

Sharpe ratio is return divided by the volatility. Many investors consider this to be a much better measure than simply measuring the return because if the want a higher return than simply investing in the asset provides, then they can use leverage.

Besides the risk and return of an investment, in order to see how if fits in a portfolio it is important to consider correlation with other assets. Even if we find a good investment, if it is highly correlated with the other assets in the portfolio it could add much more risk than if we find an investment with a lower expected return but which has diversifying properties. Cremers shows the following correlation matrix in his paper. We chose to use annual returns rather than quarterly returns correlation matrix (which are also in the paper) due to the issue with farmland’s lumpy revenue sharing payments.

From the paper’s results, not only does farmland have the highest sharpe ratio, but it also is negatively correlated with stocks (represented by the S&P 500) and bonds.

Risk and Correlation Issues

As mentioned before, there are only a couple of publicly traded farmland REITs, and they have not been around for very long. However, there have been publicly traded Property REITs for decades. Though, in the paper Cremers uses the NCREIF Property Index. In order to assess how well the NCREIF Property Index tracks publicly traded real estate companies we created a portfolio using all the real estate companies in the Russell 3000 (we used a bigger universe to get a better sample size) and weighted the portfolio using each company’s market capitalization. Below, we track the investment returns to this portfolio and compare it to the NCREIF Property Index and the S&P 500.

The NCREIF Property index, which is based on appraised values, is not at all reflective of the publicly traded Real Estate companies, even though they have the same return over the last 30 years. The volatility of publicly traded companies is often not reflective of the true underlying volatility in expected future cash flows, which is one of the reasons why we think a publicly traded farmland REIT is not the best way to invest in farmland (here and here). The correlation between the Russell 300 Real Estate stocks and the NCREIF Property Index is actually slightly negative and has almost 5 times the risk (as measured by standard deviation of quarterly returns)! This makes any risk comparison or correlation calculation meaningless. The paper is comparing apples to oranges.

Because the paper also uses the NCREIF Farmland Index to represent an investment in farmland, one can also assume that the low risk and correlation as compared to publicly traded stocks and bonds is also likely overstated.