Time series analysis of volatility in the petroleum pricing markets: the persistence, asymmetry and jumps in the returns series

• Date: 01 September 2016
• Author: Olusanya E. Olubusoye,OlaOluwa S. Yaya
• DOI: http://doi.org/10.1111/opec.12077
• Published date: 01 September 2016

Time series analysis of volatility in the

petroleum pricing markets: the persistence,

asymmetry and jumps in the returns series

Olusanya E. Olubusoye and OlaOluwa S. Yaya

Department of Statistics University of Ibadan, Ibadan, 23402, Nigeria. Email: oe.olubusoye@ui.edu.ng;

os.yaya@ui.edu.ng

Abstract

The petroleum energy market is becoming more volatile owing to recent fluctuations in oil price,

which in the long run affects the pricing and volatility persistence levels of other petroleum

products. Apart from the symmetry and asymmetry that are known with volatility series, jumps

have recently been identified, while the symmetric and asymmetric models failed in predicting the

jump components in the financial series. The historical prices of crude oil and its distilled

constituents possess occasional jumps as a result of global political or economic constraints. We

applied both fractional persistence and volatility modelling frameworks in studying the volatility

persistence in crude oil and petroleum products prices. We chose among symmetric, asymmetric

and jumps volatility models. Results indicated that prices of crude oil and gasoline were less

persistent when compared with volatility series of other petroleum products. The newly proposed

jump volatility model variants outperformed other existing volatility models in predicting the

volatility in the prices of crude oil, heating oil and diesel. The exception was the Asymmetric

Power ARCH (APARCH) model, which emerged best in predicting the prices of gasoline,

kerosene and propane prices; but GAS variants were still ranked second and third competing

models in predicting the volatility in gasoline and kerosene prices. Using wrongly specified model

for predicting the volatility in petroleum pricing can misinform oil markets, thereby generating

intense conditional oil market volatility that is capable of distorting the price of oil and

macroeconomic stability of the entire globe.

1. Introduction

The energy issue has received considerable attention in the global sustainable

development agenda in the last few decades. This is due to the fact that energy pricing

significantly impacts on economic growth, and this consequently affects the poverty

level (UNDP, 2005). Crude oil and petroleum products are the major global energy

sources. Apart from crude oil, there are gasoline, heating oil, diesel, kerosene and

JEL classification: C22, C50, Q40, Q43.

©2016 Organization of the Petroleum Exporting Countries. Published by John Wiley & Sons Ltd, 9600 Garsington

Road, Oxford OX4 2DQ, UK and 350 Main Street, Malden, MA 02148, USA.

235

propane, which are used domestically and industrially. These products are affected by

market forces, which make their prices volatile. The demands for the products are

seasonal as well. For example, there is higher demand for gasoline during the summer,

while there is higher demand for heating oil during winter. It is, therefore, necessary to

study the volatility levels and suggest a more robust volatility models for predicting the

conditional volatility in the prices of crude oil and its products at the international

market.

Information on the volatility levels of oil price and other petroleum products is

necessary since it allows policy makers and individuals to know market reactions on

each of these petroleum products with regard to oil price. Forces militating against the

pricing of oil include the prevailing exchange rates and international stocks, particularly

those in the United States and the United Kingdom (Gil-Alana and Yaya, 2014). Recent

innovations by financial time series experts and economists have, therefore, concentrated

on studying the variability in the prices, rates and values of these economic and financial

time data, at both national and international markets. This variability is termed volatility.

Quite a number of models to support this stylised fact have been proposed in the

literature.

Since the seminal work by Bollerslev (1986) on the Generalized Autoregressive

Conditional Heteroscedasticity (GARCH) model, hundreds of other models variants

have been proposed, and volatility has been a major topic of discussion in empirical

financial time series analysis. These models are proposed based on properties of the time

series, such as symmetry, asymmetry, non-linearity, stationarity, persistence and

structural breaks. Actually, persistence is investigated directly using the log-returns,

absolute and squared returns series. The persistency level of volatility in these

transformed series is determined using the semi-parametric approach.

1

This approach is

widely used owing to its robustness in estimating the level of persistence for both

stationary and non-stationary time series. With a model-based approach, persistence of

volatility can be investigated using Fractionally Integrated (FI) variants of the GARCH

model.

2

It is very common to consider both symmetric and asymmetric volatility models

in time series data, but recent innovations have shown that jump is another inherent

property in financial data, for example, in prices of oil and other petroleum products. As

a result of these occasional jumps, Generalized Autoregressive Score (GAS) class of

models were introduced in Harvey and Chakravarty (2008) and Harvey (2013).

The classical GARCH variants are not robust enough to capture these occasional

jumps in financial time series, and therefore, they underestimate the magnitude effect of

the returns. The persistence effect of volatilities across markets can be measured using

either fractional integration or GARCH-based approaches. In this article, we consider

estimating the persistence in the spot prices of crude oil and petroleum products using

the fractional persistence approach and determining the appropriate volatility models

OPEC Energy Review September 2016 ©2016 Organization of the Petroleum Exporting Countries

236 Olusanya E. Olubusoye and OlaOluwa S. Yaya