Strategic alignment of pricing oil products from Brent conditionally on market regimes using Markov and cluster‐based switching dynamics

• Published date: 01 March 2023
• Author: Petros Theodorou,Nikos Apokoritis
• Date: 01 March 2023
• DOI: http://doi.org/10.1111/opec.12261

34

|

OPEC Energy Rev. 2023;47:34–54.wileyonlinelibrary.com/journal/opec

1

|

INTRODUCTION

Refiners face a key economic trade off among flexibility and cost strategy in order to attain strategic alignment

(Theodorou,1996; Theodorou & Karyampas,2008). The tradeoff is whether to pay a price premium for higher grades

(procurement rigidity) or accept to pay higher capex and refining costs in order to take advantage of lower grades (pro-

curement flexibility). Costs and payback periods must be weighed against anticipated prices and the projected differential

or price spread among oil products. Flexibility strategy refers to the alignment of operations to the continuous changes of

crude oil prices and the prices of refined products. Thus, crude slate and refinery's complexity must be aligned with the

states, regimes reflected in the prices of both crude oil and the products of distillation. Crude oil prices can determine the

states or regimes of the market and play a very significant role in pricing and energy policy. Refiners have limited range of

flexibility in setting the gasoline to distillate production ratio. According to refinery economics and shifting yields, diesel

along with gasoline is of primary focus as high value lights versus residuals (lowest- value by- product).

Moreover, the significance of the price states, regimes is derived by the oil's market share and its mark- to- market

use in energy contracts, as it is applied in the pricing formulas of most fossil fuels and energy products (Jammazi &

Aloui,2012; Kuck & Schweikert,2017; Yu et al.,2008). Their latter role facilitates analysts to assess market behaviour and

DOI: 10.1111/opec.12261

ORIGINAL ARTICLE

Strategic alignment of pricing oil products from Brent

conditionally on market regimes using Markov and

cluster- based switching dynamics

PetrosTheodorou1,2

|

NikosApokoritis3

© 2022 Organization of the Petroleum Exporting Countries.

1Athens University of Economics and

Business, Athens, Greece

2Public Power Corporation, Athens,

Greece

3Information Resources Inc., Analytics

Center of Excellence, Athens, Greece

Correspondence

Petros Theodorou, Athens University of

Economics and Business, 76 Patission

Str., Athens, Greece.

Emails: theodorou@aueb.gr and

p.theodorou@dei.gr

Abstract

Asymmetries and non- linearities in oil prices are of high importance for energy

policy and strategy. The effectiveness of energy policy depends heavily on the

‘conceptualization’ of states, regimes and less on the absolute price level. Based

on the strategic alignment theory and the matching perspective, we claim that

recurring oil market dynamics showcase the existence of various regimes across

which pricing relationships differ. Regimes are examined by combining two

methodologies: Cluster- based Regression (CBR) and Markov- Switching Dynamic

Regression (MSDR). CBR identifies spatial regimes by clustering Brent as a proxy

variable, while MSDR identifies temporal regimes by presuming a probabilistic

superposition of states which transition as a discrete time Markov chain. Results

have both policy and methodology contributions. A reversion and a cyclicality

phenomenon in pricing, resembles the Joseph and Noah effects in terms of per-

sistence, expected duration and probability of transition. CBR describes a serial

process in pricing from ‘bull’ to ‘bear’ and then to ‘transitory’ state. Differences

in asymmetry and high price premiums, conditionally on market states, are de-

tected among products. Matching probabilities show robustness of results.

|

35

ALIGNMENT OF PRICING OIL PRODUCTS ON MARKET STATES

monitor the market, both of which are important factors for decision making, for analysing the investment behaviour

(Kuck & Schweikert,2017) and for the accurate prediction of REN and energy policy. Specifically, gas market uses a

pricing scheme based among other factors on certain crude oil price states, bands. In Europe, the price of pipeline gas is

contractually defined by weighting the prices of crudes and distillate products using specific oil price states, regimes in

an averaging and time lagging formula. Coal, as well, uses contractual provisions for conditional progressive pricing on

crude's price states. All the above, give a good reasoning, why research on oil price bands (states, regimes) is very import-

ant for the determination of pricing alignment. The thesis of this research is influenced by the matching perspective of

the Strategic Alignment Theory (Kumar et al.,2018; Manthou et al.,1996; Theodorou,2005; Theodorou & Florou,2008)

and applied to the relationship of crude's price with the price of distillates, as this relationship changes significantly

among the different states, clusters or regimes. The method developed in this research to extract these price states, re-

gimes is derived from the movement of the benchmark price (Brent) through which the price of distillates products is

estimated. In this way, all the information included in the benchmark can be decoded, decompiled in a way as to explain

the price movement of oil products. Within those clusters, potential outcomes and likelihood of occurrences are known,

thus persistence is increased, while uncertainty is increased when moving or changing clusters.

Different states denote different price clusters and neutrality denotes a certain level of fluctuation and risk within each

state. These states are estimated in two different ways, one by using certain deterministic clustering algorithms and the

other stochastically by Markov switching models. As will be presented, our models determine the bull, bear and transi-

tory state relative to the optimal number of clusters formed by the price levels. Bear state prevails for 4 years starting from

2015 in the Brent's level of (28– 62)$/bbl while the bull state mainly for 5 years starting from 2011 within the price band of

(92– 142)$/bbl. Transitory state appears around the levels of (62– 92)$/bbl in the periods from 2006– 2011 and 2018– 2019

interrupted by the 2 years period of global financial crisis. For each time period and each state, the reaction of distillates

prices will be analysed as well as the movement of prices which form the cyclical phenomenon: a specific move from

bear to bull state without the intervention of transitory in the middle. The strategic alignment of prices in bull, bear and

transitory regime will also be estimated. Stochastically, two temporal regimes will be identified and explained below as

moderate (temporal regime 1) and turbulent (temporal regime 2). The mathematical models developed in this study:

Cluster- based Regression (CBR) and Markov Switching Dynamic Regression (MSDR) are implemented in the relation-

ship of the prices of distillates fuels and of the crude oil, under the below generic expression:

In the subsequent paragraphs, a literature review section will follow, after which details of the data sources and study period

will be given. Next, the methodology used in this research will be presented and described in detail, followed by a section

explaining and discussing the results and the model's estimations. Finally, conclusions and policy implications will be given

as well as extensions for future research.

2

|

LITERATURE REVIEW

The relationship and the effect of Brent's price on various distillate products is an essential issue for the budgeting, cash

flow, and risk management of many energy companies. The effect of Brent is conditional on the market's state, regimes.

However, even if this effect has been examined with various economic variables, research falls short on the effect on

pricing various oil products, while nothing can be found on the conditional effect regarding the market states, regimes.

Specifically, the relationship of oil price with other economic variables (GDP, stock indexes, interest and exchange rates)

has long been debated. Discussion mainly started after great depression and continues up to day in a high momentum

(Hamilton,1983; Hamilton,2008). The impact of oil price on equities has been examined by Basher et al.(2018), Jones

and Kaul(1996), Tiwari, Nasreen, et al.(2020), Tiwari, Raheem, et al.(2020), Ye et al.(2016) and Zhu et al.(2016), while

the relevant impact on the price of money (interest rate) by Akram(2009) and Wang & Chueh(2013). Regarding the rela-

tionship of oil with foreign exchange it is worth to mention the papers of Anjum and Malik(2019); Golub(1983); Jawadi

et al.(2016); Lizardo and Mollick(2010); Roubaud and Arouri(2018). For readers who need an extended literature re-

view on the research of the relationship among oil and exchange rate Kim et al.(2019) provide an extended taxonomy

per country, period, method and results. Finally, as regards to the macroeconomic effects we can refer to the work of

Kim et al.(2019); and Mork(1989). The impact of oil price on growth rate has been studied by Jiménez- Rodríguez and

Sánchez(2005), and on inflation by Bloomberg and Ethan(1995).

Pdistillate

=f

(

P

crude|

State of crude

)