An Analysis of Temporal and Spectral Connectedness and Spillover in Commodity Markets

Abstract

This thesis is concerned with evaluating the temporal and spectral connectedness and spillover dynamics of commodity prices. The industries of interest are crude oil, agricultural commodities, aquaculture, and Norwegian salmon as the primary datasets. World agricultural and energy commodity indexes as well as the aquaculture sector and salmon price index have experienced exceptionally volatile periods throughout the last decade. Therefore, the objective of this thesis is to detect and quantify the temporal and spectral connectedness and spillover dynamics in the prices of these assets. This thesis falls in line with a large collection of research papers evaluating the dynamics of commodity markets. More specifically, the first two papers examine the connectedness structure between crude oil and agricultural commodities and between various aquaculture species by utilizing wavelet-based copula approach. By combining the methodologies from physics and econometrics, we evaluate how the dependence structures among the underlying assets varies across di˙erent frequencies and in the tails of the distributions. The third paper evaluates the static and temporal return and volatility spillover dynamics between crude oil and agricultural commodities. The last paper examines the firm-level cointegration relation and return spillover dynamics between Fish Pool Index (FPI) and major salmon producers. Incorporating methodologies from physics, economics, and finance is relevant when examining spectral relationship and providing an alternative angle to examine the commodity markets. The findings of this thesis indicate that the connectedness between oil and agricultural commodities increased during post-2006 across all considered frequencies of return movements. Specifically, the wavelet decomposition reveal that the interconnectedness structure is negative during the pre-2006, but it turns positive over the post-2006 subsample. Furthermore, the findings indicate persistence in dependence variation is higher over the long-run return movements. In terms of spillover analysis, the findings indicate minuscule information transmission between crude oil and agricultural commodities over the pre-2006 subsample, but crude oil tends to be a net receiver of volatility over the post-2006 subsample. Furthermore, we report asymmetric and bidirectional information transmission between crude oil and agriculture during periods of financial and economic turmoil. In terms of connectedness in di˙erent aquaculture species, the findings indicate limited dependence in the short-run horizon, however, the price linkage among various species significantly increased over the medium- and long-run horizon, suggesting market integration over the long-run. In regard to cointegration and spillover among FPI and major salmon producers, we report that the prices of exchange traded salmon stocks reflect the flow of salmon market information earlier than the price index. Furthermore, our findings indicate that the FPI and small producers are net receiver of spillover from major salmon producers.