Time dynamics of connectedness between commodity convenience yields and zero-coupon inflation swap rates

ABSTRACT

Globalisation and financial liberalisation have made financial markets more correlated and connected. In this context, it has become extremely important to understand the connectedness and correlation among different financial markets and commodities. This paper attempts to extend applicable empirical studies by examining the connectedness between volatilities of commodity convenience yields and zero-coupon inflation swap rates. We conduct our study by using both the spillover index methodology provided by Diebold and Yilmaz (2009, 2012) as well as Barunik and Krehlik’s (2018) methodology to decompose the index to its frequencies for short-, medium and long-term dynamics. Although, empirical results based on Diebold and Yilmaz’s (2012) methodology show that high total connectedness exists between the variables for the whole time period, our results based on Barunik and Krehlik’s (2018) approach shows that this connectedness exists only in the long-term. The results also indicate that the connectedness dynamics change when the effect of cross-correlations is considered.

KEYWORDS:

• frequency connectedness
• spillover
• convenience yield
• volatility

JEL CODES:

• Q02
• Q11
• Q41
• E31
• G13
• G14

Notes

1 This approach assumes that convenience yields have single dimensional relationship with its independent variables such as interest rates, carry costs and spot and futures prices and thus rules out the stochastic of convenience yields, which in fact receive influences from its independent variables. Casassus and Collin-Dufresne (2005) show that this is the case. This paper studies the stochastic process of convenience yields of individual commodities to identify the influences that convenience yields receive from interest rates, spot prices and risk premia, using maximum likelihood tests, maximal model of historical parameters and unconditional first and second moments with the Maximal Model Estimates. This approach allowed to find the optimal representation of convenience yield distribution under normalised measure. Following this, the unconditional first and second moments with the Maximal Model Estimates have been studied to find the persistence parameters. This allows to identify the statistical significance of exogeneity (time-varying) and endogeneity (consistent patterns), thus, mean reversion properties of convenience yield of a particular commodity. Nevertheless, since carry costs are ruled out due to the fact that there is no centralised reliable data to be employed, one cannot claim that findings are unbiased. That being said, although paying attention to convenience yield stochastic could yield more accurate results, for the sake of simplicity, this study assumed convenience yields as single dimensional data.