Discussion Paper

No. 2018-55 | July 11, 2018
What drives food price volatility? Evidence based on a generalized VAR approach applied to the food, financial and energy markets


The aim of this study is to investigate sources of food prices volatility. The analysis uses daily series for volatility of corn, soybean, wheat, rice, US dollar, crude oil, and SP500 futures spanning the period January 4, 2000 to April 1, 2017. The authors employ the generalized vector autoregressive framework in rolling sample approach in order to capture the time-varying nature of volatility spillovers. The results reveal that: volatility spillovers measures change over time; most of the volatility spillovers are observed within the two groups of markets: food markets and “non-food” markets; corn market is net volatility transmitter.

Data Set

JEL Classification:

Q17, G15, C58


  • Downloads: 1142


Cite As

Sławomir Śmiech, Monika Papież, Marek A. Dąbrowski, and Kamil Fijorek (2018). What drives food price volatility? Evidence based on a generalized VAR approach applied to the food, financial and energy markets. Economics Discussion Papers, No 2018-55, Kiel Institute for the World Economy. http://www.economics-ejournal.org/economics/discussionpapers/2018-55

Comments and Questions

Anonymous - Invited Reader Comment
July 31, 2018 - 09:47

1. The article analyses the volatility of food prices from 2000 to 2017. I think it’s important to consider that this includes the so-called 2000’s commodity boom from 2004 to 2014, approximately. I don’t know if the boom affected the volatility as well as the level. Can their results be ...[more]

... extended to regular times? At least, more on this should be clarified. Or robustness checks can be obtained adding periods of normal food prices.

2. What is the variable US dollar? Is it the exchange rate vis a vis the euro?

3. The authors present four results but they only provide a plausible explanation for one of them: i.e, the corn as source of volatility because of biofuels. I think it will be interesting to provide plausible explanations for the rest of the evidence as well.

4. In equation 1, the authors give their proxy for volatility which is the log-diff of the logs of max and min values. Why not just the log-diff of the original max and min values?

5. In the results presented in Table 2, wouldn’t it be relevant to show standard deviations of the spillovers?

6. In Figure 7, the authors present volatility spillovers within the food markets. Does this mean that the rest of variables are excluded from the GFEVD estimation? If yes, isn’t there any danger of omitted variable bias?

7. I find kind of hard to follow the results presented in Figure 9. Why not just plot the IRFs as usual?

8. The authors claim in their conclusion that financial and energy markets have a limited role in food prices volatility, and that the volatility comes mostly from the corn market, a food commodity itself. While this result might have an statistic relevance, I believe that there is not much of an economic contribution here. I think that a more important finding would be to know where does the volatility come from, besides the food market itself. E.g, is it from demand or supply driven shocks? But maybe this is beyond the scope of the article.

Sławomir Śmiech - responses to the comment
August 07, 2018 - 17:19

the attached pdf file contains detailed responses to all eight commentaries

Anonymous - Referee Report 1
August 08, 2018 - 08:31

see attached file

Sławomir Śmiech - reply to referee 1 report
September 10, 2018 - 11:09

see attached file

Anonymous - Referee Report 2
October 22, 2018 - 10:19

To the authors

The objective of the paper is to analyze volatility spillovers between different markets, namely commodities, US stock, energy and foreign exchange markets. A special emphasis is given to the understanding of food price volatility within the last decades which includes the food crises of 2008 and ...[more]

... 2011. Thus, the work contributes to the literature on the causes of increasing volatility in international food markets, in particular to the strain of literature on the financialization of agricultural commodity markets. The contribution of the work is twofold. First, the authors apply a Generalized Vector Autoregressive Model (GVAR) allowing for joint modeling a large number of markets. Second, the authors use a different measure to compute volatility, specifically a range-based measures which is increases the efficiency of the estimation. In addition to that, the authors test the robustness of the results and present impulse response functions to illustrate the empirical results.

The results indicate that volatility spillovers are observed mostly within each category of markets, e.g. within food markets. However, the volatility spillovers vary over time and are larger during crises periods. Among the food markets, the corn market transmits the largest amount of volatility to the other markets. The authors relate this to the fact that corn is used to produce biofuels in the U.S.

The paper is nicely written, data and methodology are well explained. However, I have some major comments.

Major comments:

•The authors apply a new method (GVAR methodology; range-based volatility) to a thoroughly investigated research question. However, the authors do not critically review their methods over the others (in particular mgarch). I suggest to list advantages/disadvantages and hypothesize how the results may differ.
•The authors focus on three strands of literature in their literature review. First, studies examining the relationship between energy and food markets; second research on the financialization of agricultural commodity markets; third, the transmission between different agricultural commodity markets. Certainly, it is not possible to cite the complete literature on the topic, but it would be beneficial to look at studies with a similar research objective, which is, to my understanding, the identification of the contributors (and their importance) to food price volatility. In my opinion, most relevant to this research objective are studies that look at the fundamental factors, the financialization of commodity markets, and the energy-food nexus jointly, e.g. Tadesse et al. (2014) in Food policy. Moreover, the methods of the other studies are neither discussed nor critically reviewed.

-Related to the last point, there is no discussion on how the findings of this study differ or coincide with the other works presented. For this comparison, it is important to critically review the methods of the other studies and to explain the advantage/implication of the GVAR methods over the others.

•The last comment relates to the discussion of the results. I appreciate the detailed analysis and robustness checks by the authors. However, the discussion appears a bit lengthy as compared to the rest of the paper. More importantly, the results are not convincing. The volatility transmitted is always very similar to the volatility received. One could argue that this just reflects the level of correlation between the categories of markets. Moreover, this is different from the literature presented in the review section. Last, the discussion of results and the conclusion is merely descriptive. What drives the results, apart from the heterogeneity of the markets? For instance, the sentence “the most general conclusion of the paper is that the role of the financial and energy markets in creating the food markets volatility is limited” is not explained. The tension between and the theoretical arguments of advocates and opponents of the financialization hypothesis is not mentioned in the paper. The respective literature Irwin et al. is also not cited.

Sławomir Śmiech - reply to referee report 2
November 03, 2018 - 19:59

Reply to referee report 2 in an attached file