EFF/KRF: The answer to your question is "yes" — if you judge that the past average return on gold is the best estimate of its future expected return. We don't buy this, for two reasons. First, much of the stock of gold is in the form of jewelry and other goods that pay a "consumption dividend." This dividend increases the current price of gold and lowers its expected capital gain return. But an investor who holds gold bullion as a portfolio asset only expects to get the expected capital gain, which does not suffice to compensate for the risk of the asset. Second, if the capital gain return on gold is uncorrelated or even negatively correlated with the returns on other assets like stocks, then portfolio theory and its asset pricing offspring imply that the expected return on gold is low because of its role in reducing portfolio risk. Both arguments point to the conclusion that the past average return on gold probably overestimates its expected return.

To summarize, gold makes sense as a portfolio asset only for investors who also get the consumption dividend from gold, since this "dividend" lowers gold's expected capital gain. Thus, for investors who do not value the consumption dividend, the expected return on gold does not cover its risk as a portfolio asset, (which takes account of its value for portfolio diversification).