موجودی و پویایی قیمت بنزین بالادستی
|کد مقاله||سال انتشار||مقاله انگلیسی||ترجمه فارسی||تعداد کلمات|
|20699||2012||7 صفحه PDF||سفارش دهید||محاسبه نشده|
Publisher : Elsevier - Science Direct (الزویر - ساینس دایرکت)
Journal : Energy Economics, Volume 34, Issue 1, January 2012, Pages 208–214
This paper sheds new light on the asymmetric dynamics in upstream U.S. gasoline prices. The model is based on Pindyck's inventory model of commodity price dynamics. We show that asymmetry in gasoline price dynamics is caused by changes in the net marginal convenience yield: higher costs of marketing and storage lead to rising gasoline prices, whereas a drop in these costs lowers gasoline prices. The former effect is stronger. This indicates asymmetric dynamics. We also analyze the asymmetry across the sample by analyzing recursive and rolling regressions.
In June 14, 2001 when the crude oil price was 69 cents per gallon, John Cook (Director of the Petroleum Division in the U.S. Energy Information Administration's Office of Oil and Gas) testified before the Committee on Government Reform of the U.S. House of Representatives. He argued that “…Low stocks set the stage for gasoline price increases this spring. … This ongoing tightness has been a key factor in maintaining both low crude and product inventories since then…”. He expects prices to continue to decline in the summer of 2001 because production in the U.S. increased significantly. However, gasoline markets remain exposed to volatility because of low global oil inventories (Energy Information Administration, 2001). Crude oil prices indeed dropped and hit a low on November 15, 2001 of 41.7 cents per gallon. Crude oil prices started to rise again in 2002, and broke the 100 cents per gallon barrier for the first time on May 24, 2004. On May 16, 2008 one gallon of crude oil costs 301 cents. On June 25 that year Guy Caruso testified before the Committee on Appropriations of the U.S. House of Representatives that “…the current very tight oil market balances and the possibility of further oil supply disruptions are causing prices to rise to unprecedented highs…”. (Energy Information Administration, 2008a). The peak of 346 cents per gallon was reached on July 11, 2008, but only five months later the price of a gallon of crude oil dropped to only 72 cents. After that the crude oil price rose again steadily to 263 cents per gallon in May 2011. Gasoline prices followed more or less the same pattern, but the reaction of gasoline prices to changes in crude oil prices is not necessarily symmetric (see Frey and Manera, 2007 for a review). These facts show that fluctuations in crude oil prices and gasoline prices are large and volatile. These fluctuations are primarily driven by demand and supply shocks, and by stock levels. This paper examines the role of inventories in a framework that has been used extensively to addresses the question of a systematic tendency for gasoline prices to respond more rapidly to increases in crude oil prices than to decreases in crude oil prices (the so-called rockets-and-feathers hypothesis). Understanding gasoline price dynamics and the role of inventories in that process may help to improve to forecast prices of assets that are contingent on oil. Recent examples of this literature are Wu and McCallum, 2005, Knetsch, 2007 and Alquist and Kilian, 2010. In this paper we examine the role of inventories and volatility in crude oil markets using daily U.S. crude oil prices and regular upstream gasoline prices for the period November 1987 until March 2010. We use the standard two-step error correction estimation procedure and take account of volatility clustering and heavy tails (excess kurtosis) in the distribution of crude oil and gasoline price data. We find that marketing and storage costs affect gasoline prices asymmetrically. In the empirical literature so far this explanation is not considered. We also show that the asymmetry emerges after June 1998 when oil prices started to rise. Our paper differs from other papers by studying the role of inventories and volatility. Moreover, we test for possible structural breaks, and we carefully test our specification. We argue that it is important to account for volatility clustering. Finally, we analyze how asymmetry develops across the sample by carrying out recursive and rolling regressions. The outline of the paper is as follows. In Section 2 we discuss the related literature. Section 3 explains the inventory model that serves as the foundation of the empirical model presented in Section 4. This section emphasis the dynamics of gasoline prices focusing on persistence, volatility, and asymmetry. In Section 5 we present the data we use in the empirical analysis. We also analyze the data by addressing issues like stationarity, cointegration and breaks in the series. Section 6 describes the methodology and presents the estimation results. We conclude that asymmetric effects due to inventories do exist. We also present interval estimates for different sample sizes and sample periods by means of recursive and rolling regressions. Finally, we illustrate the results with an impulse response analysis. Section 7 summarizes the main conclusion of the paper.
نتیجه گیری انگلیسی
We analyze the pass-through of crude oil price to upstream gasoline prices emphasizing the role of inventories and volatility, both theoretically and empirically. The theoretical model implies a trade-off between selling out of inventory versus production. We analyze the implications of this literature for dynamics and volatility of upstream U.S. gasoline prices. The dynamics may include asymmetries of various types, but also allows us to study persistence of gasoline price dynamics. Volatility in our model is included through the net marginal convenience yield, and via the conditional variance in an ARCH model. As is standard practice in the rockets-and-feathers gasoline price literature we estimate an error correction model, but also model the conditional variance of the error term. The ARCH version we are using is GJR-model which allows for asymmetric volatility. Heavy tails in the daily data are accounted for by estimating the ECM–GJR model with t-distributed errors. We use daily data for U.S. oil and gasoline prices for the period 1986–2010. We show that asymmetry is driven by changes in the net marginal convenience yield. Lower marketing costs reduce gasoline prices, whereas higher marketing costs increase gasoline prices. The latter dominates and the difference is significant. This is illustrated by showing the results of impulse-response analysis. A drop in the net marginal convenience yield of 0.5 cents per gallon significantly affects the gasoline price more than a similar increase in the net marginal convenience yield does. The effects on gasoline prices are + 0.61 cents per gallon and − 0.42 cents per gallon, respectively. The effect is not permanent, but gradually fades away. This outcome is robust for samples that include the period after the mid-2000s. These results contribute to a better understanding of upstream gasoline dynamics, and may also have implications for pricing assets contingent on oil.