Estimating oil price ‘Value at Risk’ using the historical simulation approach
Introduction
Since the end of the nineteenth century, oil has been the main source of energy for mankind, and is the most widely used natural resource in energy production. Furthermore, movements in oil markets affect supply and demand of the other energetic commodities and incentive or disincentive research on alternative sources of energy. The strategic role of oil in our society became evident in the crisis of the seventies (1973 and 1979), and once again in the year 2000. Crude oil and refined products prices are not only relevant to oil market agents but also to governments and society. As Sadorsky (1999) suggests, changes in oil prices have an impact on economic activity, but changes in economic activity have little impact on oil prices. He states that oil price volatility shocks have asymmetric effects on the economy and finds evidence of the importance of oil price movements when explaining movements in stock returns. In the same way, Faff and Brailsford (1999) also find significant evidence of oil price sensitivity across some equity returns of Australian industry, using an augmented market model.
The agreements that OPEC members reached in the seventies and the reaction that those agreements caused in the oil market agents changed the controlled environment that had characterised the oil market up to then. The new market is relatively free and characterised by high price shifts. An unpredictable, volatile and risky environment has arisen and protection against market risk has become an essential issue.1
This volatile oil price environment requires a risk quantification. Value at Risk (VaR) has become an essential tool within financial markets when quantifying portfolio market risk (the risk associated with price movements). VaR determines the maximum loss a portfolio can generate over a certain holding period, with a pre-determined likelihood level. Therefore, VaR can be used, for instance, to evaluate portfolio managers’ performance by providing a risk quantification, expressed in monetary units (m.u.), which can be used, together with portfolio return (also expressed in m.u.), to this end. Moreover, it helps portfolio managers to determine the most suitable risk management policy for any given situation.
Within oil markets, VaR can be used to quantify the maximum oil price change associated with a likelihood level. This quantification is fundamental when designing risk management strategies.
This study proposes to quantify oil price VaR through the historical simulation (HS) approach. We develop a HS methodology (the historical simulation with ARMA forecasts approach, or HSAF), which, as we demonstrate, improves the VaR estimations provided by the HS standard approach. Furthermore, HSAF VaR estimations are more efficient than those provided by an autoregressive conditional heteroskedasticity (ARCH) model.
The rest of the paper is set out as follows: Section 2 introduces the VaR calculation procedure, and analyses the main methodologies and models that can be used to determine VaR. In Section 3 we put forward the fundamentals of the proposed methodology. Section 4 presents the empirical analysis: we estimate oil price VaR using the HS standard scheme, the proposed HSAF methodology and the ARCH models approach. The final section of the paper summarises the main conclusions of our study.
Section snippets
VaR calculation procedure
Within the area of finance, VaR can be defined as ‘an estimate of the largest loss that a portfolio is likely to suffer during all but truly exceptional periods. More precisely, the VaR is the maximum loss that an institution can be confident it would lose (in) a certain fraction of time over a particular period’.2 Thus, VaR sums up the risk which a portfolio or, in general, an entity is exposed to in just one figure: the amount which could be lost during an established holding
The historical simulation with ARMA forecasts methodology
The HSAF methodology, introduced and developed in this paper, belongs to the HS approach. However, HSAF methodology does not directly use the distribution of past returns, but rather the distribution of forecasting errors, derived from an estimated ARMA model. HSAF methodology implementation requires a four-stage procedure.
Stage 1: Past portfolio returns: HSAF methodology uses the autocorrelation property observed in most financial and economic data series when they are taken in absolute value
Data
We used daily spot Brent oil prices, from January 1992 to December 1999, and divided this time period into two smaller time periods:
- •
The first, from 1992 to 1998, was used to estimate the model coefficients; this is the ‘in the sample period’.
- •
The second, the year 1999, was used for forecasting purposes; this is the ‘out of the sample period’.
Fig. 1 shows oil price trends corresponding to the analysed period.
Historical simulation standard approach
Within the oil price environment, the HS standard approach assumes that price change
Concluding remarks
The volatile oil price environment requires a risk quantification. VaR has become an essential tool for this end when quantifying market risk (risk associated with price changes). Within oil markets, VaR can be used to quantify the maximum oil price change associated with a likelihood level. This quantification constitutes a fundamental point when designing risk management strategies.
There are several methods used to estimate oil price VaR. In this paper we analyse three of these methods: the
References (14)
Generalized autoregressive conditional heteroskedasticity
J. Economet.
(1986)- et al.
ARCH modelling in finance: a review of the theory and empirical evidence
J. Economet.
(1992) - et al.
Oil price risk and the Australian stock market
J. Energ. Finance Dev.
(1999) - et al.
The impact of energy derivates on the crude oil market
Energ. Econ.
(1999) - et al.
Unbiasedness and time varying risk premia in the crude oil futures market
Energ. Econ.
(1994) - et al.
Are oil markets chaotic? A non-linear dynamic analysis
Energ. Econ.
(2000) Oil price shocks and stock market activity
Energ. Econ.
(1999)
Cited by (157)
A high-frequency approach to VaR measures and forecasts based on the HAR-QREG model with jumps
2022, Physica A: Statistical Mechanics and its ApplicationsSpillover effects between commodity and stock markets: A SDSES approach
2022, Resources PolicyImplications of clean energy, oil and emissions pricing for the GCC energy sector stock
2022, Energy EconomicsSystemic risk of commodity markets: A dynamic factor copula approach
2022, International Review of Financial Analysis