The battle is joined. Should commodities figure in a portfolio? It is to be expected that there be a big division and difference of opinion between investors with an active strategy trying to beat the market and those with a passive, market index allocation strategy. But commodities seem to be controversial even amongst the passive index asset allocation crowd, in which I count myself.
Main proponent: Larry Swedroe, best-selling author and principal of Buckingham Asset Management
Arguments: collateralized commodity future funds provide positive return, that is negatively correlated with both stocks and bonds (i.e. a strong diversification benefit), as well as protection for unexpected inflation, which together are excellent portfolio insurance
Main proponents: Rick Ferri, author of several "All About ..." books, and head of Portfolio Solutions LLC investment advisors and managers; William Bernstein, author of The Intelligent Asset Allocator and The Four Pillars of Investing amongst others; Kenneth French, renowned finance prof and researcher.
Arguments: commodities have no real expected return net of inflation and the backtested returns on various commodities indexes are essentially trading strategies, not passive market investments in an index, whose future success in uncertain
Rick Ferri and Larry Swedroe go head to head in a discussion hosted by Hard Assets - The Great Commodities Debate part 1 and part 2
Kenneth French video interview on Yahoo Finance Why Investors Shouldn't Own Commodities
William Bernstein writes on his Efficient Frontier website "On Stuff"
Dimensional Fund Advisors quotes paper by Truman Clark on their website
Hard Assets Investor lists a half dozen seminal papers on the Hard Assets University page - see the Grad School section which links to free downloads from SSRN. Must read: The Tactical and Strategic Value of Commodity Futures by Claude Erb and Campbell, 2006, which everyone seems to quote. It's almost a book at 61 pages but a very worthwhile couple of hours of slow reading. Around pages 39-40 there is a great general explanation of the beneficial effects of negative correlation, variance and number of securities interacting with regular rebalancing in a portfolio. HAI also has useful primers, interviews, current news.
Despite his being outnumbered, my reading of Erb and Campbell tells me that I believe more Swedroe's conclusion. Commodity ETFs/ETNs do depart from the purist passive asset allocator model, being essentially active portfolio strategies based on futures (i.e. derivatives) but there is good reason to believe that the diversification return of around 3% will continue and the powerful portfolio risk reduction effect from (most of the time) negative correlation makes them worthwhile. I don't believe in the inflation protection benefit. An allocation of about 5% of my portfolio seems reasonable. So I'm sticking with my holding of DJP, the iPath Dow Jones-UBS Commodity Index Total Return ETN (recently renamed, replacing the AIG reference in the name with UBS).
The Truman Clark paper cited by Dimensional, which seems to have no visibility (or credibility?) amongst the other researchers on commodities, being cited by none of them, seems too impossible to believe - how could there be no reduction in portfolio standard deviation, as it concludes, when negatively correlated commodity assets are added? It seems to be a mathematical impossibility the way covariance works.