Modeling the commodity prices of base metals in Indian commodity market using a Higher Order Markovian Approach

TL;DR

This paper introduces a Higher Order Markovian model to better predict commodity prices in the Indian market, demonstrating improved accuracy over traditional models for most metals, especially during volatile periods.

Contribution

The paper proposes a Higher Order Markovian approach as an alternative to Markovian models for modeling commodity prices, capturing industry response delays.

Findings

HOM model improves forecasting accuracy for Copper Mini.

HOM model outperforms Markovian model for most metals.

Volatility during COVID-19 affected Aluminum predictions.

Abstract

A Higher Order Markovian (HOM) model to capture the dynamics of commodity prices is proposed as an alternative to a Markovian model. In particular, the order of the former model, is taken to be the delay, in the response of the industry, to the market information. This is then empirically analyzed for the prices of Copper Mini and four other bases metals, namely Aluminum, Lead, Nickel and Zinc, in the Indian commodities market. In case of Copper Mini, the usage of the HOM approach consistently offer improvement, over the Markovian approach, in terms of the errors in forecasting. Similar trends were observed for the other base metals considered, with the exception of Aluminum, which can be attributed the volatility in the Asian market during the COVID-19 outbreak.