Systemic Risk for Financial Institutions in the Major Petroleum-based Economies: The Role of Oil.

• Author: Khalifa, Ahmed

1. INTRODUCTION

   The oil rich countries, Bahrain, Kuwait, Oman, Qatar, Saudi Arabia and UAE, are heavily petroleum-dependent economies that are underpinned by huge foreign assets and powered by foreign labour. More specifically, oil accounted for 26.4% of the total GDP in the whole GCC area, 25% of Saudi Arabia's, 22.3% of the UAE's, 42% of Kuwait's, 32.3% of Qatar's, and 29% of Oman's. Bahrain stands out among those oil rich countries because oil accounts for only 12% of its GDP due to the depletion of its oil reserves over the years. (1) The oil dominance in these countries underscores that a marked change in the level of oil prices will significantly affect all the sectors of their economies and may exacerbate existing financial systemic risks, thereby harming the stability and the functioning of their financial sectors. In turn, this would have further consequences on the cyclical sectors.

   Notably, these countries attempt to coordinate their policies to achieve their common goal of realizing full economic integration through the Gulf Cooperation Council (GCC), an international organization that embraces all GCC countries. Furthermore, the financial institutions in those countries are highly connected and characterized by economies of scale (Al-Jarrah et al., 2019). Additionally, they carry the systemic risks usually associated with large financial firms (Al-Jarrah et al., 2016). Within such a business environment of heavy oil dependence, high financial intercon-nectedness and strong propagation of risk, the examination of the risk tolerance of GCC financial institutions to oil price movements presents itself as an interesting case study, particularly in the wake of the recent global financial crisis and the recent reoccurrence of collapses in oil prices. For this reason, this paper attempts to address two major questions related to the financial sectors of those petroleum-based economies, which possess large foreign assets but are still vulnerable to oil risk. First, do oil shocks cause stress to petroleum-based financial institutions? (2) Second and more relevant, what is the impact of the movement of the level of oil prices on the systemic risk indicators for those financial institutions?

   We may postulate that the empirical evidence should indicate a relevant impact of oil price movements on the (systemic) financial risk of GCC countries. Despite this reasonable and expected result, this study is the first that attempts to deal with such important questions by focusing on a large panel of GCC financial institutions. Furthermore, our approach is innovative, because it accounts for the impact of oil price variations on financial risk over different horizons, as inspired by the heterogeneity of market players of Corsi (2009) and includes it into one of the most common systemic risk measures which is the change in the Conditional Value-at-Risk (or [DELTA]CoVaR) of Adrian and Brunnermeier (2016). The introduction of a direct impact of oil on the evaluation of systemic risk in GCC financial institutions will facilitate the detection of the presence of the oil impact, and, thus, will allow assessing the potential effect of oil price swings on the GCC financial sector.

   The interest in our analyses is not limited to GCC financial institutions and GCC regulators. Indeed, the study provides relevant insights into the systemic risk in financial institutions at the global level. In fact, we cannot exclude the possibility that a very high risk in a major financial institution could cascade further risks in the highly vulnerable GCC economies, with grave consequences for the global economy. Thus, our findings will be of interest to global financial institutions and market regulators, as they will provide an approach to monitoring the impact of oil price variations on systemic risk measures. To investigate the impact of oil price variations on a GCC financial institution's systemic risk, we collect data on stock prices and balance sheets for many financial companies as well as on national market indexes for the GCC region for the period from March 2004 to November 2018.

   Building on these data, we address the first question and attempt to detect if oil shocks cause a stress on petroleum-based financial institutions. Following the approach of Jeong et al. (2012), we initially run a quantile causality test from oil returns to financial institutions' returns. This test sheds light on the possible impact of oil movements on the quantiles, as proxies of risk measures, of the financial institutions. As well as can be expected, the findings show that oil returns have a significant and diffused impact on the quantiles of the GCC financial institutions' stocks returns.

   Then, we proceed to the estimation of the systemic risk measure proposed by Adrian and Brun-nermeier (2016), the [DELTA]CoVaR. The main idea behind the [DELTA]CoVaR is that the risk of a financial system depends on the financial health of individual institutions. When a financial institution faces stress, this will change the distribution of asset values within the system. Therefore, by measuring the relationship between a financial company and a financial market index, we can infer the systemic impact of a single financial institution. The [DELTA]CoVaR monitors the changes in the asset values of the financial system, conditioning on the stress situation in a single financial company and contrasting the obtained values with those observed in a normal state of the same company. This measure provides insights that help answer the second research question related to the potential impact of changes in oil prices on the systemic risk indicators for those financial institutions. As a first step, we set our benchmark by ignoring oil as a potential systemic risk factor, and hence excluding it from [DELTA]CoVaR estimation. The results show that pronounced increases in the [DELTA]CoVaR levels correspond to the global financial crisis, which is an exogenous shock to the financial sectors of these petroleum-based GCC economies.

   Then, we proceed and evaluate the changes in systemic risk measurement obtained by introducing oil returns as a potential risk driver. Inspired by the work of Corsi (2009), we deviate from the Adrian and Brunnermeier (2016) approach and introduce the cumulated lagged oil returns in the CoVaR equations to capture both the short-term (one week) impact of oil price movements and the more pronounced movements that can be detected over longer periods (one month). This is coherent with the recent contribution of Khalifa et al. (2017), who in a different framework find that oil price movements may influence the oil production process with a delay up to a quarter. The empirical results suggest that the impact of oil price movements on the extreme quantiles of the financial companies' returns is relevant and is associated with both weekly and monthly impacts. In this regard, by means of the dynamic quantile test proposed by Engle and Manganelli (2004), we show that there is an improvement in the systemic risk measurement through CoVaR with the inclusion of oil.

   Interestingly, the difference between the CoVaR with and without oil returns seems to correlate with the occurrence of the shock that stroke oil prices in correspondence with the global financial crisis but with a longer time length. Indeed, by using a Markov switching model, we show that the stress regime of the difference between CoVaR with and without oil returns for the GCC region is longer than the stress regime of oil returns. This implies that the recent financial crisis has a real effect on oil prices. In turn, this leads to a further worsening of the financial institutions' risk levels, and an increase in the time needed to recover from the effects of financial crises.

   From a policy maker's or a regulator's perspective, the results of our study suggest that the conditioning on real control variables is fundamental to monitor the real effects of financial crises and the possible feedbacks between the real and the financial sides of an economy. In the case of the GCC markets, the role of oil, as expected, is crucial and allows for a more appropriate estimation of the systemic impact of financial companies, in addition to potentially facilitating the determination of the financial impact of shocks striking oil prices.

   The remainder of the paper proceeds as follows. Section 2 provides a review of the literature, while Section 3 discusses the empirical strategy by presenting the data, the methodology and the results. Section 4 provides the conclusions and recommendations. A complementary document available on-line accompanies the manuscript.

2. LITERATURE REVIEW

   The present paper relates to two strands of the financial economics literature. The first focuses on the estimation of systemic risk for financial institutions, while the second deals with the consequences of oil price variations on financial markets.

   Within the first strand, the literature has proposed several Systemic Risk Measures (SRMs) by defining and modelling systemic events using different approaches. Acharya et al. (2017) present an economic model of systemic risk and show that the Marginal Expected Shortfall (MES) can measure each financial institution's contribution to the systemic risk, while Systemic Expected Shortfall (SES) evaluates the expected contribution of a financial institution to a systemic crisis. Brownlees and Engle (2016) propose SRISK, a systemic risk measure that is a function of a firm's size, leverage, volatility, and dependence on the market. The SRISK measures the capital shortfall of a financial institution, conditional on a severe market decline. Billio et al. (2012) propose Granger-causality tests to measure the interconnectedness among financial institutions such as hedge funds, banks, brokers, and insurances. Their findings show that interconnectedness represents a reliable indicator of the...