تاثیر سیاست های پولی بر پارامترهای فرآیند نفت و انتظارات بازار

Following record low interest rates and a depreciating US dollar, crude oil prices came under increasing pressure during the period 2003M5–2007M10. Oil price process parameters changed dramatically toward consistently rising prices. Short-term forecasting implied the persistence of the observed trend, as market fundamentals and underlying monetary policy were supportive of these trends. Market expectations derived from option prices anticipated further surge in oil prices and indicated a significant probability of right tail events. A tightening of monetary policy may be essential for restoring stability to oil markets and in turn for sustained economic growth.

During the period 2003–2007, the path of crude oil prices appeared to be on an upward trend. All the more worrisome was the fact that oil price increases took place during a period of rising trends in general commodity prices, speculation in housing, equity, and credit markets, and a depreciating US dollar. The fast rise in commodity prices, including oil, could be seen as delayed effect of highly expansionary monetary policy during 2001–2004, as key interest rates were forced down to record post-WWII lows. Such monetary expansion led to high world economic growth and consequently higher world demand for oil and non-oil commodities. As the supply of crude oil and other commodities are rigid or increase at a slower pace than demand, the result was the highest level of commodity price inflation in the post-WWII era. With real interest rates turning low or negative, pressure on real aggregate demand, and therefore on oil markets, might not subside. Relaxation of monetary policy during August to December 2007 set off a new spiral in commodity price inflation and currency depreciation.1 In this paper we analyze oil price developments during 2000M1–2007M10. To analyze the effect of monetary policy, we examine two sub-periods with very different oil market fundamentals, 2000M1–2003M4 and 2003M5–2007M10. We assume oil prices are driven by Levy processes (LP) of generalized hyperbolic (GH) type and use daily data to estimate the parameters of these processes. Combining features of normal and stable distributions and offering more flexibility than Poisson-type processes, which were used to model large finite jumps, GH distributions have gained wide popularity for modeling stock market indices.2 We demonstrate that normal inverse Gaussian (NIG) process closely fits oil price returns during 2000M1–2003M4 and 2003M5–2007M10; parameters of the process have, however, changed: mean return increased due to persistence in an upward trend, and kurtosis has declined due to higher predictability in oil prices. Estimated parameters of the NIG process are in conformity with findings for the empirical distribution of oil price returns. To be applicable for pricing derivatives, statistical distributions had to be adjusted for market price risk and turned into martingale processes. This is done through applying the Esscher transform to the statistical process. Besides estimating the distribution of oil price returns from time-series data, the distribution of the returns was also estimated from cross-section data from call and put option prices (for options on November 2, 2007 for end-December 2007). Implied risk-neutral distribution based on NIG shows that traders were expecting further rise in oil prices and were assigning significant probabilities for right tail events. The paper is structured as follows: in Section 2, we review the literature on the relationship between monetary policy and oil prices. In Section 3, we examine daily oil prices during 2000M1–2007M10; in Section 4, we present the normal inverse Gaussian (NIG) distribution; in Section 5, we present the statistical estimation of the NIG distribution based on daily oil price data for 2000M1–2003M4 and 2003M5–2007M10; in Section 6, we derive the risk-neutral distributions from statistical LP by applying the Esscher transform to these processes; in Section 7, we derive density forecast from oil options prices; and our conclusions follow in Section 8.

Oil prices rose relentlessly during the period 2003–2007 because of growing demand for oil and a supply that did not increase commensurately. Using the normal inverse Gaussian (NIG) distribution, a subclass of the Levy processes of generalized hyperbolic (GH) type that have been found to closely fit high frequency financial data, we obtained the result that oil price parameters changed dramatically from 2000M1–2003M4 to 2003M5–2007M10. The changes were characterized by high mean returns that rose from −0.005 to 0.11, and a lower dispersion that fell from 2.22 to 1.70. NIG for both sub-periods had low kurtosis, implying flatter than normal distribution. Based on NIG parameters, oil prices would have been expected to rise by about 30% per year. Crude oil density forecast for the end of December 2007 was extracted from option prices data on November 2, 2007. Traders' expectations were in agreement with prevailing market fundamentals, characterized by widening oil demand-supply gap, the highest post-WWII commodity price inflation, falling real interest rates, a depreciating US dollar, and a meltdown of the sub-prime debt market. Relaxation of monetary policy during August to December 2007 jolted oil prices and apparently contributed to firming up of market price expectations. Derived risk-neutral density had low kurtosis, implying a flatter distribution and a significant probability for tail events. The persistence of these trends could culminate in explosive commodities prices and could turn out to be un-sustainable. Also worrisome is the fact that oil producers have become wary of rapidly falling value of their international financial reserves and investments, which could discourage higher oil supplies. Although it could be claimed that high commodity prices had not dented world economic growth and had not affected consumer price indices, recessionary and inflationary implication of oil price shocks should not be underestimated. Hamilton, 1983 and Hamilton and Herrera, 2004 have analyzed recessionary effects of oil prices and allowed for longer lag for these effects to be fully transmitted to output and prices. Bernanke et al., 1997, Jones et al., 2004 and Lee et al., 1995 have also studied the relationship between oil prices and GDP and inflation. These authors generally found significant recessionary and inflationary impact of oil prices on real GDP and consumer prices. In contrast to the period 2003M5–2007M10, we have shown that in 2000M1–2003M4 oil prices were evolving around slightly declining trend with moderate oil demand growth. Such scenario could be again restored if monetary policy were to be tightened. At the same time, restoring stability to oil markets is essential for sustainable economic growth and price stability. The brunt of this burden appears to be on monetary policy but monetary policy cannot be used as a general panacea. A well-targeted instrument best addresses each economic goal. For instance, balance of payments deficits could be best and most rapidly tackled by monetary approach to balance of payments, namely reducing public and private deficit financing through credit ceilings. Similarly, external competitiveness could be most durably achieved through gains in productivity, cost reductions, and technical innovations without necessarily depressing nominal exchange rates. Exchange rate depreciation may not restore external competitiveness in context of expansionary monetary policy; to be effective, the exchange rate may require the support of a restrictive monetary policy. In sum, oil prices are key economic variables. By juxtaposing two sub-periods, we have shown that oil price parameters could be highly sensitive to macroeconomic policies. Accordingly, prudent monetary policy may be a requirement for achieving longer-term oil price stability and sustained economic growth