Have food and financial markets integrated?

Abstract

This article analyses comovements and discusses possibly greater market integration between aggregate food commodity and stock prices in the period 1990 to 2012. Return correlations, price return distributions, cointegration and Granger-causalities are tested in subsamples on monthly FAO Food Price Index and MSCI World Stock Market Index. Empirical results suggest that while there is only weak indication of greater comovements concurrent with structural changes such as changed agricultural policies, new demand due to growth in emerging markets and energy mandates and the financialization of food markets since the early 2000s, they did start to increase substantially in particular during the financial stress of the Lehman crisis and the Great Recession. While structural changes may have amplified price linkages across markets, results do not suggest that they are the key factors for greater price comovements. Instead, the effects of the late-2000s recession as a time of great economic weakness and uncertainty may have changed concurrently the behaviour of both food and financial market participants, such that different market prices exhibit large comovements.

Keywords:

• comovements
• food commodity market
• stock market
• financialization
• market integration

JEL Classification:

• Q11
• Q13
• G11
• G12

Acknowledgements

The research for this article was started while the author was visiting the Department of Agricultural and Consumer Economics at the University of Illinois at Urbana-Champaign, and he gratefully acknowledges their hospitality. He also thanks Michel Robe for very useful discussions. An earlier version of this article was presented at the 2013 GEWISOLA conference, and comments of conference attendants have very much contributed in improving this article. Useful suggestions of an anonymous reviewer are also gratefully acknowledged. They are, of course, not responsible for errors that may remain.

Notes

¹ Another approach would be to choose either the VAR or VECM for the causality test, conditional on test results for cointegration (e.g. used in Röthig, 2011). However, this is an example of ‘preliminary test testing’, and significance levels and the power of the causality test will be distorted. Results obtained by Clarke and Mirza (2006) suggest that an approach such as proposed by Toda and Yamamoto (1995) is preferred to the practice of pretesting for cointegration. In a recent article, Bauer and Maynard (2012) indicate that the lag-augmented approach can provide robust Granger-causality tests not only in the case of nonstationarity, but also for problems such as long-memory and certain (unmodelled) structural breaks.

² The analysis was also additionally conducted on the subsample before the financial crisis (January 1990–September 2008). Over that period, results show insignificant return correlation, neither cointegration nor Granger-causalities; however, Kolmogorov–Smirnov tests cannot reject the null of common return distributions. This adds to the evidence that comovements increased concurrently with the financial crisis. These alternative results are available from the author upon request.