Gold has not served very well as a hedge against bad macroeconomic and stock market outcomes. That is the central conclusion of research by Professors Robert Barro and Sanjay Misra, which is forthcoming in the Economic Journal. Their study draws on evidence from long-term US data on gold returns, as well as gold returns during some of the worst macroeconomic disasters experienced across the world.

Gold has historically played a prominent role in transactions among financial institutions even in modern systems that rely on paper money. What''s more, many observers think that gold provides a hedge against major macroeconomic declines. But after assessing long-term US data on gold returns, the new research finds that gold has not served consistently as a hedge against large declines in real GDP or real stock prices.

From 1836 to 2011, gold delivered low average real price appreciation and experienced high average volatility. The mean real rate of price change was 1.1% per year, close to the 1% average real rate of return on three-month US Treasury Bills and comparable assets. The standard deviation of annual gold returns was 13.7%, almost as high as the 16.7% on the US stock market.

Some analysts justify holding gold as a hedge that delivers relatively high real returns during recessions. A true hedge would display significantly negative correlation with growth rates of real per capita GDP and real stock returns. But this study finds that these variables have a correlation of essentially zero with US real gold returns.

To explore gold''s hedging potential further, the researchers study gold returns during the worst macroeconomic disasters experienced globally, defined as periods when a country''s real per capita GDP fell by 10% or more. They identify 56 macroeconomic disasters since 1880 across 19 countries for which it is possible to calculate a country''s inflation-adjusted gold returns from data on exchange rates and price indexes.

Across these 19 countries, gold''s average real rate of price appreciation was 1.5% per year from 1880 to 2011. By comparison, during the 56 disaster periods, gold''s real price appreciation averaged 2.1% per year – negligibly above its overall mean – with an annual standard deviation of 22%.

Moreover, the real price appreciation of gold was negative in over half of the macroeconomic disasters (30 of 56). Thus, the relatively low average real returns and large volatility during macroeconomic disasters further demonstrate that gold has not been a great hedge against bad economic times.

The research finds that historical changes in the volatility of gold''s real rate of price appreciation derive mainly from changes in the monetary role of gold. The lowest volatility applied to the period of the classical gold standard from 1880 to the start of the First World War.

Since then, the important changes involve the transitions among different forms of gold standards, leading up to the full termination of convertibility of the US dollar into gold in 1971. Because of the near disconnect between gold and monetary systems since the 1970s, the volatility of real gold returns has become particularly high since that time.

The data reveal changes in the real price of gold but do not directly indicate ''dividends'', in the sense of the non-pecuniary services that gold provides to holders. But the analysis suggests that the average of this ''dividend yield'' would not be much greater than 1% per year.

Moreover, the researchers'' measures of the volatility of real returns on gold can be accurately gauged by the observable data on fluctuations in real gold prices. Therefore, they are confident in their conclusion that gold has not served very well as a hedge against bad macroeconomic and stock market outcomes.

''Gold Returns'' by Robert Barro and Sanjay Misra is forthcoming in the Economic Journal. The authors are at Harvard University.