Regime switching in the reactions of stock markets in Saudi Arabia to oil price variations

• DOI: http://doi.org/10.1111/twec.12785
• Author: Jamel Jouini,Wajih Khallouli
• Date: 01 August 2019
• Published date: 01 August 2019

ORIGINAL ARTICLE

Regime switching in the reactions of stock markets

in Saudi Arabia to oil price variations

Jamel Jouini

1

|

Wajih Khallouli

2

1

Department of Economics, College of Business Administration, King Saud University, Riyadh, Saudi Arabia

2

Department of Banking and Financial Markets, College of Islamic Economics and Finance, Umm Al-Qura University,

Mecca, Saudi Arabia

1

|

INTRODUCTION

The focus on the extent to which international stock markets are potentially affected by oil price

fluctuations received great interest in the literature. The linkages between macroeconomic aggre-

gates, notably real output, and world oil price influence such extent through the effects on factor

input costs and corporate gains. Increases in oil price affect negatively stock prices due to lower

corporate gains and cash flows that result from the negative responses of real output to such

increases.

1

Within this context, Jones and Kaul (1996) outline that Canadian and US equit y mar-

kets are negatively influenced by oil price shocks. This negative relationship between stock and oil

markets is also supported by Sadorsky (1999), Papapetrou (2001), Miller and Ratti (2009) and

Jammazi and Aloui (2010). Moreover, some empirical studies argue that stock markets are posi-

tively sensitive to oil price shocks (see Lescaroux & Mignon, 2008; Narayan & Narayan, 2010;

Ono, 2011), while others do not conclude in favour of a significant relationship between equity

and oil markets (see Al Janabi, Hatemi, & Irandoust, 2010; Cong, Wei, Jiao, & Fan, 2008; and

Huang, Masulis, & Stoll, 1996).

The country's dependence on oil can also play a decisive role in determining the reactions of equity

markets to oil price shocks, which may explain the mixed results of the impacts of oil price movements

on stock market returns. Park and Ratti (2008) document that the effect of oil price on equity markets

and its sign differ greatly between oil‐importer cou ntries and oil‐exporter countries. Aloui, Nguyen, and

Njeh (2012) provide evidence to support Park and Ratti's (2008) conclusions and find that for stock mar-

kets that are positively associated with oil price, the oil risk is predominant. Bhar and Nikolova (2009)

find that the country's dependence on oil affects the stock–oil nexus in the BRIC (Brazil, Russia, India

and China) economies. As seen, there is no evidence of a clear consensus about the responses of equity

markets to oil price changes, whichincentivises us to revisitthe stock–oil nexus.

In addition to the dissimilar reactions of equity markets to oil price movements, the related lit-

erature deals with the stock–oil nexus in a time‐varying framework, provided many influential

events, such as the Great Depression, the 1973 oil price shock, the financial crises, the September

1

Hamilton (1983) points out that increases in oil price contribute to US recession cycles over the post‐World War II period. Mork

(1989) supports this conclusion and reveals, in addition, that decreases in oil price do not enhance economic growth during the same

period. Estimates in this direction are provided by International Energy Agency (2004) indicating that an increase from 25 to 35 dol-

lars in oil price generates a decline in GDP of 0.3% for the United States, 0.4% for Japan and 0.5% for the Eurozone.

Received: 8 December 2017

|

Revised: 30 October 2018

|

Accepted: 15 January 2019

DOI: 10.1111/twec.12785

World Econ. 2019;42:2467–2506. wileyonlinelibrary.com/journal/twec © 2019 John Wiley & Sons Ltd

|

2467

11 Attacks, the oil boom (2002 to mid‐2008), the 2011 global economy unrests and the oil price

falls since late 2014. These events incentivise the use of regime switching approach to study the

instability of the interactions between equity and oil price markets over time and to support their

sensitivity to the market situation (bear or bull). Within this context, Hammoudeh and Choi (2007)

find that oil price plays a pivotal role in explaining the behaviour of Gulf Cooperation Council

(GCC) equity markets during shifts in the fads and the fundamentals. Miller and Ratti (2009) show

evidence of an unstable relationship between oil price and six OECD equity returns during the

post‐2000 period, thereby supporting the presence of unrests in stock and oil markets since the

beginning of the century. Jammazi and Aloui (2010) find similar conclusions for three international

equity markets over the period 1989–2007. Aloui and Jammazi (2009), and Jammazi and Nguyen

(2015) investigate the sensitivity of international stock markets to oil price shocks over the period

1989–2007 and find that equity index returns experience a regime switching behaviour and

respond differently to oil price surges.

2

Balcilar, Gupta, and Miller (2015) reveal that the stock–oil

nexus is regime‐dependent from 1859 to 2013.

The current study explores the empirical evidence of the reactions of stock market returns to oil

price shocks by contributing to the existing literature in various ways. First, in view of the absence of

prior empirical studies on the stock–oil nexus in the framework of regime switching models at a sec-

tor level, we examine the sensitivity of equity sectors to changes in oil price.

3

The analysis of stock–

oil nexus at a sector level may be more useful than that at a global level or national level, as the latter

might mask the specificities intrinsic to each equity sector. In addition, the stock sector reactions to

oil price shocks may be heterogeneous depending on whether oil is utile for the industry related to

the considered sector, and the importance of these industries for the economy. A noticeable feature is

that considering stock sector prices for an economy is very useful for investors to understand well the

sensitivity of stock sectors to oil price changes, thereby enabling them to be informed of opportuni-

ties to diversify their stock sector/oil portfolios or to make the portfolio with the equity secto r index

that satisfies the risk‐adjusted return performance. It is also important for policymakers to compre-

hend well the influence of oil price on the efficiency of cross‐market hedging policies.

Second, owing to the existing various international economic and financial events , as documented

above, the analysis of the relationship between stock markets and oil price should be clearly con-

ducted in a time‐varying framework. Within this context, several econometric specifications deal with

multiple structural change forms whose the most attractive is the Markov switching a pproach

4

that

2

Several research studies examine the behaviour of either stock markets or oil price. For equity markets, Chu, Santoni, and

Liu (1996) find that volatility of the New York Stock Exchange (NYSE) index is regime‐dependent. Maheu and McCurdy

(2000) identify recession and expansion phases in NYSE returns. Diamandis (2008) reveals regime‐shifts in the volatility of

Latin American equity markets during the 1990s and early 2000s. Wang and Theobald (2008) provide mixed evidence as

regards the effect of financial liberalisation on stock market volatility in each regime for East Asian countries. Liu, Margari-

tis, and Wang (2012) reveal that the S&P 500 stock index switches between two regimes (a low‐return volatile regime and

a high‐return stable regime), and that information on price range allows market agents to obtain better volatility forecasts.

Regarding oil price, Fong and See (2002), Vo (2009) and Arouri, Lahiani, Lévy, and Nguyen (2012) find that the regime

switching process performs well in forecasting oil price volatility. Chang (2012) outlines that forecasting performance

depends greatly on regime switching and asymmetric volatility.

3

The linkages between stock sectors and oil price have been addressed in the literature using models without structural

change (see Boyer & Filion, 2007; Jouini, 2013; Malik & Ewing, 2009; Nandha & Brooks, 2009; Nandha & Faff, 2008).

Similar outcomes indicate that the responses of stock markets to oil price shocks differ across sectors.

4

See Hamilton (1990) for univariate autoregressive models, Krolzig (1997, 1999) for vector autoregressive and vector error

correction models and Cai (1994), Hamilton and Susmel (1994) and Henry (2009) for GARCH‐type models. The existing

literature on the application of the Markov switching approach to the stock–oil nexus is quite limited, as documented above.

2468

|

JOUINI AND KHALLOULI

fits well economic and financial relationships with influential regime‐shifts and business cycles,

5

thereby alleviating the shortcomings raised by linear approaches (ARIMA process, GARCH‐type

models, unobservable components model, etc.) which do not allow us to account for such features. In

this study, we use the two‐regime Markov switching EGARCH model, proposed by Henry (2009),

which has the advantage to capture asymmetry and persistence, thereby describing well the underl y-

ing data of time series. This is well documented by Henry (2009) for the connection between UK

stock returns and short‐term interest rates, Khallouli and Sandretto (2012) for the contagion in Middle

East and North African (MENA) equity markets caused by the US subprime crisis, Chang (2012) for

the behaviour of crude oil futures and Aloui and Jammazi (2009) and Jammazi and Nguyen (2015)

for the interactions between international equity markets and oil price.

6

By doing so, we extend the

study of Jouini (2013), who examines the linkages between stock sectors in Saudi Arabia and world

oil price in the context of linear GARCH‐type models, for a longer period. Accordingly, the current

study allows us to explain well the regime‐shift behaviour of the stock–oil nexus and to scout about

the state (expansion or recession) on which depend the time‐varying reactions of stock sectors to oil

price fluctuations, in order to obtain accurate, meaningful and useful results for investors and policy-

makers. In addition, our research attempts to highlight whether the reactions of stock sectors to oil

price movements differ across sectors and regimes, and to what extent. This allows investors to build

portfolios with stock sectors that are less sensitive to oil price variations and financial authorities to

intervene effectively in stock sector markets.

Third, two pertinent indicators of oil price shocks are used to identify the responses of equity

markets to oil price changes. Indeed, we distinguish between negative and positive oil price shocks

to highlight the asymmetric effects of oil price on stock index returns, which allow policymakers

to make appropriate measures depending on the nature of oil price variations, downward or

upward. The incentive is that many prior studies provide evidence of non‐linear linkages between

oil price and economic activity; that is, negative and positive oil price variations have different

impacts on the variables (see Arouri, Jouini, Le, & Nguyen, 2012; Basher & Sadorsky, 2006;

Cologni & Manera, 2009; Cong et al., 2008; Hamilton, 2003; Lardic & Mignon, 2008; Mork,

1989; Nandha & Faff, 2008).

Lastly, we focus on estimating the effect of oil price movements on stock sectors for Saudi

Arabia, a large OPEC oil‐exporting country that is too sensitive to oil price fluctuations. It should

be noted that the Saudi equity market is the largest Arab market in terms of market capitalisation

(US$448.831 bn in 2016). Table 1 displays the evolution of four indicators for the Saudi stock

market from 2009 to 2016.

7

The ratio of market capitalisation of listed domestic companies to

GDP, the share of stocks traded in the GDP and the turnover ratio of domestic shares can be con-

sidered as sufficiently high, thereby highlighting the importance of the Saudi stock market. As

expected, the Saudi equity market becomes attractive, as the number of listed domestic companies

experiences upward movements from 2009 to 2016, thereby increasing trading activity. This

increase may be explained by the equity market liberalisation and structural reforms established in

the GCC countries in the mid‐2000s. Overall, the Arab spring triggered in some Arab countries

5

See Hamilton (1989), Diebold, Lee, and Weinbach (1994), Filardo and Gordon (1998), Kim, Nelson, and Startz (1998),

Kim and Nelson (1998), Psaradakis, Sola, and Spagnolo (2004), Brunetti, Scotti, Mariano, and Tan (2008), Aloui and Jam-

mazi (2009), Henry (2009), Jammazi and Aloui (2010), Balcilar et al. (2015) and Jammazi and Nguyen (2015).

6

To the best of our knowledge, these authors are the only ones who have applied the Henry's (2009) empirical approach.

Therefore, our study is the first to opt for such approach to examine the sensitivity of stock market returns to oil price

movements at a sector‐level.

7

This is the only common period for all indicators considered.

JOUINI AND KHALLOULI

|

2469