A monthly crude oil spot price forecasting model using relative inventories

doi:10.1016/j.ijforecast.2005.01.001

International Journal of Forecasting

Volume 21, Issue 3, July–September 2005, Pages 491-501

https://doi.org/10.1016/j.ijforecast.2005.01.001 Get rights and content

Abstract

This paper presents a short-term forecasting model of monthly West Texas Intermediate crude oil spot prices using readily available OECD industrial petroleum inventory levels. The model provides good in-sample and out-of-sample dynamic forecasts for the post-Gulf War time period. In-sample and out-of-sample forecasts from the model are compared with those derived from other models. The model is intended for the practicing forecaster and designed to be simple enough to implement easily in a spreadsheet or other software package, with the variables easy to update. The simplicity and ease of updating make this model attractive for investigating various scenarios to see the impacts that market changes can have on monthly crude oil spot prices, if inventories, production, imports, or demand change. Finally, the model structure can easily be updated periodically should there be a fundamental market change or a shift in the normal level of inventories.

Introduction

In this paper, we develop a simple and practical model for forecasting monthly crude oil spot prices under normal market circumstances¹ using readily available data. The objective of this model is to provide a dynamic forecast of monthly West Texas Intermediate (WTI) prices for the post Gulf War I time period using readily available data (WTI spot prices and OECD petroleum inventories) readily available and that can be implemented in spreadsheet applications. Thus, the model is intuitively appealing and useful to industry and government decision-makers in forecasting prices and in investigating the impacts on price of changing market fundamentals, such as petroleum inventories, production, imports, and demand. Because of the few variables involved, the results from the forecast model can be easily interpreted. The model is also practical from a maintenance standpoint in that it is based on two readily available data series.

We focus on the concepts of normal and relative levels of petroleum market variables such as demand, field production, net imports, and inventory. This is similar to the natural rate theory widely applied in macroeconomic modeling and empirical studies. We decompose the observed level of a petroleum market variable into two components: the normal level, determined by historical seasonal movements and trends, which reflects the normal market demand and operational requirements; and the relative level, the difference between the observed and normal levels, which reflects short-run market fluctuations. Seasonality exists in petroleum market variables such as demand, field production, and net imports. Since change in inventory equals the difference between demand and supply (the sum of field production and net imports), petroleum inventory demonstrates seasonal movements as well. Long-term inventory trends exist due to the trends in government inventories, increases in long-term product demand, and the larger distribution and storage infrastructure needed to meet demand growth and product differentiation. It is the relative levels of demand, field production, net imports, and inventory that respond to the unexpected short-run variations in the market.

Among these relative level variables, it is the relative inventory that matters most in the short run for the determination of crude oil price.² On the demand side, many consumers have few options and little discretion as to how much volume they can use in the short term. People must heat their homes in the winter and drive to work regardless of price. This implies that short-term demand elasticity is small. On the supply side, supply chains that move crude oil from its source of production to refineries, and then move product from refineries to consumers, are long and cannot be adjusted quickly when unexpected shifts in supply or demand occur. Empirically, we have found that the short-run demand elasticity is indeed much less than the short-run inventory elasticity in the United States and the production elasticity is virtually zero. Since total petroleum inventory levels are a measure of the balance, or imbalance, between petroleum production and demand, they reflect changing market pressures on crude oil prices, and thus provide a good market barometer for crude oil price change in the short run. Intuitively, relative inventory levels must be negatively related to prices, with distributed lagged effects.

The observed behavior of WTI spot prices and total OECD industrial inventories since Gulf War I supports this argument. The large swings in WTI spot prices during the late 1990s are coupled with counter-swings in OECD total inventories, particularly when measured in relative terms. For example, in 1998 production consistently exceeded demand; as a result, inventories grew to unusually high levels. Demand growth slowed during this period in part due to warm winters in the Northern Hemisphere and the Asian financial crisis, while supply increased substantially as Iraqi crude oil came back into the export market in 1997 through the “Oil-for-Food” Program. WTI spot prices fell to near $10 per barrel by the end of 1998 due to the excess of production over demand and the resulting inventory build. The supply/demand balance reversed in 1999 when OPEC cut back production to a level well below demand, and the demand for crude oil simultaneously increased as the Asian economies recovered. With demand exceeding production during 1999, inventories were drawn down to help meet demand. The excess inventories that had been built up fell rapidly to well below normal levels, and WTI spot prices correspondingly rose to over $30 per barrel by early March 2000. More recently, we have observed that the loss of Venezuelan crude oil production during the oil workers' strike in that country resulted in a large drop in U.S. crude oil inventories, while demand changed very little, even though crude oil prices climbed about $5 per barrel.³

Section snippets

A forecast model

The basic idea underlying the forecast model is the natural rate theory. Theoretically, the normal inventory level is determined by normal demand and normal operating requirements, while the relative inventory level responds to market fluctuations. For practical purposes, we calculate the normal inventory level by appropriately de-seasonalizing and de-trending historical data.⁴

Forecast evaluation

To evaluate the relative inventory (RSTK) model developed in the last section, we compare in-sample and out-of-sample forecast results with two alternative models. The first alternative model is a naïve autoregressive forecast (NAIV) model ${WTI}_{t} = a + b AR (1) + \sum_{i = 0}^{5} c_{j} D_{j} 911 + dLAPR99 + e AR(12) + ε_{t}$ in which AR(1) and AR(12) are the 1st-order and 12th-order autoregressive terms, respectively. Other variables are as defined in Eq. (3) to ensure comparability. Table 3 shows the estimates for the naïve model's

Conclusion

This paper presents and evaluates a simple and practical dynamic forecast model using petroleum inventories as the only independent variable.¹⁹ This practical model is needed because we found that existing simple models lack a theoretical basis and more complex models (in the general equilibrium

Acknowledgements

This work is partially sponsored by the Office of Petroleum Reserves, U.S. Department of Energy. The authors thank two anonymous referees for many helpful comments.

Michael YE teaches economics at St. Mary's College of Maryland. He has experience as a consultant to a wide variety of private and government organizations, including the U.S. Department of Defense and U.S. Department of Energy. He has published many papers on modeling in microeconomics, operations research, and statistics.

References (15)

R. Lucas
Expectations and the neutrality of money
Journal of Economic Theory
(1972)
A.C. Arize
U.S. petroleum consumption behavior and oil price uncertainty: Tests of cointegration and parameter instability
Atlantic Economic Journal
(2000)
J. Bentzen
The impact of petroleum taxes on residential oil demand in Denmark
Journal of Energy Finance and Development
(1997)
F.X. Diebold
Elements of forecasting
(2001)
M. Friedman
The role of monetary policy
American Economic Review
(1968)
D.F. Gordon et al.
On the theory of price dynamics
S. Husted et al.
Linear rational expectations equilibrium laws of motion for selected U.S. raw material imports
International Economics Review
(1987)

There are more references available in the full text version of this article.

Cited by (108)

Predicting electricity demand in Cameroon with enhanced modified exponential models: A focus on the southern integrated grid (S.I.G)
2024, e-Prime - Advances in Electrical Engineering, Electronics and Energy
This paper is centered on the prediction of electricity demand within the Southern Integrated Grid (S.I.G) for the forthcoming six years (2022–2027) through the utilization of advanced forecasting techniques. These methodologies encompass modified exponential, probabilistic or stochastic extrapolation and linear regression models. The foundation of these forecasts is established on a comprehensive examination of load curve data spanning 12 years (2010–2021) and an analysis of the factors influencing electricity demand in Cameroon. The models are implemented using Matlab and Excel, enabling the projection of results. The results of this analysis reveal a consistent growth trend in Cameroon's electricity demand between 2022 and 2027. the modified exponential model foresees an increase from 991.16MW to 1189.25MW, the probabilistic model from 969.68MW to 1109.20MW, and the linear regression model from 853.46MW to 975.89MW. Despite the inherent limitations, the outcomes prove to be satisfactory, marked by a high level of accuracy and reliability, as evidenced by mean absolute percentage error (MAPE) values of 0.2% for the modified exponential and probabilistic methods and 0.55% for the linear regression method. These scores fall well below the industry-accepted benchmark of MAPE < 10% outlined by the Lewis benchmark. In a comparative analysis, it is evident that the modified exponential and probabilistic models exhibit superior accuracy.
A new hybrid deep learning model for monthly oil prices forecasting
2023, Energy Economics
The forecast of crude oil prices has always been important for investors and scholars and has drawn more attention to applying deep learning techniques in recent years. Under this circumstance, firstly, this paper proposes a novel hybrid deep learning forecasting model named Mod-VMD-BiLSTM based on the variational mode decomposition (VMD) and bidirectional long short-term memory (BiLSTM) algorithms. Next, several empirical studies and statistical evaluations are conducted to evaluate its forecasting performance. Our empirical results show that the preprocessing of decomposed series is beneficial to capture temporal general feature patterns hidden in sub-series, thereby helping to produce more accurate and robust forecasting results than the competing benchmark models among all scenarios. And all the evaluation metric values can pass the corresponding statistical tests, making the conclusions more convincing and comprehensive. Finally, the robustness tests confirm that the proposed forecasting framework is robust and superior for modeling and forecasting monthly oil prices time series.
Transformer-based forecasting for intraday trading in the Shanghai crude oil market: Analyzing open-high-low-close prices
2023, Energy Economics
The Shanghai crude oil futures market exudes distinct speculative attributes, underscoring the pivotal significance of precise price forecasts. Accurate forecasting of Shanghai crude oil futures prices assumes vital importance for investors to optimize their portfolios profitably, for producers to mitigate production risks in the crude oil spot market, and for providing cogent decision-making support to government entities. This study implements a groundbreaking unbiased structural forecasting for Shanghai crude oil futures' open-high-low-close (OHLC) prices leveraging the Transformer framework coupled with the model-driven and penalty term-based loss function designs. Based on OHLC forecasts, this study devises three intraday trading strategies. Notably, our results evince that the forecasting accuracy of the Transformer outperforms the Naïve method, vector autoregression (VAR) and vector error correction model (VECM), multiple linear regression (MLR), support vector regression (SVR), and long short-term memory (LSTM) neural network when applied to different temporal granularities of Shanghai crude oil futures OHLC data. Furthermore, the three proposed intraday trading strategies exhibit higher annualized return rates and Sharpe ratios, alongside lower maximum drawdowns and standard deviations of returns in comparison to the conventional close-to-close strategy relying solely on the close price. Remarkably, the forecasting process and intraday trading strategies explicated in this study can equally apply to other futures products in the energy sector, including electricity, coal, and natural gas.
Price dynamics in the oil market: a bond-graph modeling approach
2023, IFAC-PapersOnLine
Current oil modeling techniques lack a comprehensive approach, as long-term oil prices are qualitatively modeled based on first principles, while short-term price transients are modeled using econometric methods. In this paper we propose a comprehensive bond-graph modeling approach in which price dynamics follow from first principles. The first principles that we use are derived from the recently developed economic-engineering theory in which price dynamics are modeled using Newtonian mechanics and price drivers are identified as forces. We reformulate a qualitative first-principles model developed by the Energy Information Administration (EIA) as a bond graph by modeling six identified price-driving factors as port-elements. The constitutive laws of these port-elements generate the price drivers, which through the interconnection structure of the bond graph yield the price dynamics. We demonstrate the bond-graph model by identifying its parameters and letting it estimate the oil price given historic oil supply data. Compared to a benchmark black-box model, we find that the bond graph has two advantages: (i) it achieves a better performance, and (ii) we know what its parameters and variables represent. The latter advantage allows us to validate the bond-graph model by reconstructing the oil inventory stocks and to manually adjust parameters by expert input.
Have the predictability of oil changed during the COVID-19 pandemic: Evidence from international stock markets
2023, International Review of Financial Analysis
The COVID-19 has undoubtfully brought fierce shocks to the real economic activities, financial market and public lives. Under this special condition, this study explores whether the predictability of crude oil futures information has changed before and during the COVID-19 pandemic for 19 international stock markets. From an in-sample perspective, we find that the crude oil futures RV can significantly affect future stock volatility for each equity index except SSEC. Moreover, the out-of-sample results from statistic and economic perspective reveal that crude oil futures RV is a more efficient predictor during the COVID-19 pandemic compared with the pre-crisis period. Furthermore, we find that the predictability of crude oil futures information is stronger from March to May 2020, when the epidemic is seriously prevailing. The empirical results from alternative evaluation method, recursive window method, alternative realized measures, controlling VIX and the seasonal effect, asymmetric forecasting window and different testing windows are robust and consistent. Our findings could offer novel and significant policy and practical implications.
Forecasting crude oil prices with shrinkage methods: Can nonconvex penalty and Huber loss help?
2022, Energy Economics
This study explores different specifications of shrinkage methods to forecast crude oil prices, and examines the forecasting performances from both statistical and economic perspectives. In addition to the widely-used LASSO, elastic net and ridge, we consider two popular nonvonvex penalties and Huber loss function in penalized regressions. The results indicate that, first, the out-of-sample performance of different shrinkage methods depend on the forecasting horizons. The shrinkage forecasts with nonvonvex penalty and Huber loss outperform the benchmark and competing models within one year; the popular shrinkage methods of LASSO, elastic net and ridge perform relatively well more than one year, whereas, no better than the no-change benchmark. Second, in terms of net-of-transaction-costs portfolio performance for different horizon forecasts, the portfolios based on nonconvex penalties achieve the majority of the largest economic gains, followed by the popular version of LASSO and elastic net. When crude oil prices decline sharply, shrinkage forecasts outperform the benchmark forecast remarkably, both in statistical and economic perspectives. Finally, extended analysis indicates that imposing statistical and economic constraints on coefficient estimation of shrinkage models can further improve forecasting performance in most cases.

View all citing articles on Scopus

John ZYREN joined the Energy Information Agency at the U.S. Department of Energy as a senior industry economist after working as a private consultant in the energy industry. His principal areas of research are in the quantitative modeling of crude oil and petroleum product price behavior.

Joanne SHORE has spent over 25 years analyzing and modeling energy markets. She began her career in the nuclear industry, but later joined Gulf Oil's corporate planning department where she performed studies in areas such as downstream profitability and the implications of economic, political, and social changes on the oil business. Ms. Shore then spent 15 years in private consulting, focusing on analyses of crude oil, gasoline, distillate, and propane markets to explain price movements and to determine the impacts of changing regulations. In 2000, she joined the Energy Information Administration's Petroleum Division as lead operations research analysis team leader.

View full text

A monthly crude oil spot price forecasting model using relative inventories

Abstract

Introduction

Section snippets

A forecast model

Forecast evaluation

Conclusion

Acknowledgements

Journal of Economic Theory

U.S. petroleum consumption behavior and oil price uncertainty: Tests of cointegration and parameter instability

Atlantic Economic Journal

The impact of petroleum taxes on residential oil demand in Denmark

Journal of Energy Finance and Development

Elements of forecasting

The role of monetary policy

American Economic Review

On the theory of price dynamics

Linear rational expectations equilibrium laws of motion for selected U.S. raw material imports

International Economics Review