Natural Gas Futures Research Articles

Extracting Signals from High-Frequency Trading with Digital Signal Processing Tools

As algorithms replace a growing number of tasks performed by humans in the markets, there have been growing concerns about an increased likelihood of cascading events, similar to the Flash Crash of May 6, 2010. To address these concerns, researchers have employed a number of scientific data analysis tools to monitor the risk of such cascading events. As an example, the authors of this article investigate the natural gas (NG) futures market in the frequency domain and the interaction between weather forecasts and NG price data. They observe that Fourier components with high frequencies have become more prominent in recent years and are much stronger than could be expected from an analytical model of the market. Additionally, a significant amount of trading activity occurs in the first few seconds of every minute, which is a tell-tale sign of time-based algorithmic trading. To illustrate the potential of cascading events, the authors further study how weather forecasts drive NG prices and show that, after separating the time series by season to account for the different mechanisms that relate temperature to NG price, the temperature forecast is indeed cointegrated with NG price. They also show that the variations in temperature forecasts contribute to a significant percentage of the average daily price fluctuations, which confirms the possibility that a forecast error could significantly affect the price of NG futures. TOPICS:Statistical methods, simulations, big data/machine learning Key Findings • High-frequency components in the trading data are stronger than expected from a model assuming uniform trading during market hours. • The dominance of the high-frequency components have been increasing over the years. • Relatively small changes in temperature could create a large price fluctuation in natural gas futures contracts.

A novel risk management framework for natural gas markets

AbstractThis paper examines dynamic hedges in the natural gas futures markets for different horizons and explores the gains from devising risk management strategies. Despite the substantial progress made in developing hedging models, forecast combinations have not been explored. We fill this gap by proposing a framework for combining hedge‐ratio predictions. Composite hedge ratios lead to significant reduction in portfolio risk, whether spot prices are partially predictable or not. We offer insights on hedging effectiveness across seasons, backwardation‐contango conditions and the asymmetric profiles of long‐short hedgers. We conclude that forecast combinations better reconcile realized performance with the hedging process, mitigating model instability.

Multiscale Market Integration and Nonlinear Granger Causality between Natural Gas Futures and Physical Markets

This paper comprehensively analyzed the price integration of the U.S. natural gas futures market and its physical markets. The analyses were conducted in the form of graphics using the ensemble empirical mode decomposition (EEMD) method and minimum spanning trees with various horizons. Our findings indicated that the network structures of the minimum spanning trees of the gas futures and physical markets are the same on different time scales. The citygate returns were always the core of the physical gas markets. In addition, the gas futures and physical markets were highly integrated on different time scales. Moreover, our findings showed that at the original data level, unidirectional linear and nonlinear causalities from gas futures to physical returns exist. Specifically, the relationships between futures and physical gas returns were not constant across various time scales. In the long term, futures gas returns had only a linear causality with the citygate, commercial, and industry gas returns, and a unidirectional, nonlinear causality with residential gas returns.

Stochastic volatility, jumps and leverage in energy and stock markets: Evidence from high frequency data

• Our empirical analysis makes use of high-frequency data from the futures markets. • We show that the introduction of both leverage and jumps in the futures return dynamics improves the ability to forecast volatility. • It also provides the best forecast for risk in both a VaR and a CVaR sense. • It benefits speculative firms who are short natural gas futures while minimizing tail risk in a VaR sense. • And it also benefits speculative firms who are long S&P500 futures and use either VaR or CVaR as financial risk management criteria. In this paper, we propose a model for futures returns that has the potential to provide both individual investors and firms who have positions in financial and energy commodity futures a valid tail risk management tool. In doing so, we also aim to explore the commonalities between these markets and the degree of financialization of energy commodities. While empirical studies in energy markets embed either leverage or jumps in the futures return dynamics, we show that the introduction of both features improves the ability to forecast volatility as an indicator for risk for both the S&P500 and natural gas futures markets.

Carbon Price Forecasting: Which Driven Factors Can Help

This paper proposes two dimension-reduction and forecasting quantile methods (i.e., the quantile group lasso and the quantile group SCAD models) to predict carbon futures returns and investigate the predictability of a comprehensive group of factors including market fundamental variables and technical variables. In terms of the predictive performance, the two proposed models outperform a series of popular competing models. In terms of robustness, the employed quantile method outperforms the mean shrinkage models, especially in the case of our empirical dataset with a non-normal distribution. Through the predictor selection process, the most powerful predictors of carbon futures returns are selected through the dimension-reduction mechanism of the two employed models, while the possible difference of the selected predictors for different quantiles of carbon returns are carefully considered. Moreover, the impacts of those selected predictors are eventually estimated on the quantiles of carbon returns through a quantile regression. We find that the Brent price is the key factor of carbon returns at different quantile levels. The crude oil closing stock and the natural gas futures in the UK significantly influence the carbon futures returns at lower and higher quantile levels. Our research shows that appropriate predictors for carbon futures returns and their impacts hinge on carbon market conditions.

A multivariate distance nonlinear causality test based on partial distance correlation: a machine learning application to energy futures

This paper proposes a multivariate distance nonlinear causality test (MDNC) using the partial distance correlation in a time series framework. Partial distance correlation as an extension of the Brownian distance correlation calculates the distance correlation between random vectors X and Y controlling for a random vector Z. Our test can detect nonlinear lagged relationships between time series, and when integrated with machine learning methods it can improve the forecasting power. We apply our method as a feature selection procedure and combine it with the support vector machine and random forests algorithms to study the forecast of the main energy financial time series (oil, coal, and natural gas futures). It shows substantial improvement in forecasting the fuel energy time series in comparison to the classical Granger causality method in time series.

Moment spreads in the energy market

In this paper, we first extend the variance risk premium (VRP) in Bollerslev, Tauchen and Zhou (2009) into the moment spreads, including the skewness spread (SKS) and the kurtosis spread (KTS), and then investigate the return predictability of the moment spreads in the energy market. There are three main findings. First, VRP can significantly predict energy market returns at daily and weekly frequencies. For example, the weekly in-sample and out-of-sample R2 for predicting the weekly Energy Select Sector SPDR ETF (XLE) excess returns are 1.13% and 1.32%, respectively. Second, VRP can significantly predict crude oil returns (i.e., returns of both crude oil futures and crude oil ETF) at a monthly frequency with both monthly in-sample and out-of-sample R2 of around 7%. Finally, the kurtosis spread (KTS) of the energy market is an important predictor for excess natural gas returns (i.e., returns of both natural gas futures and natural gas ETF) across different horizons.

Can limits‐to‐arbitrage from bounded storage improve commodity term‐structure modeling?

AbstractThis paper develops and estimates a two‐factor competitive storage model for the purpose of pricing commodity futures. The empirical relevance of the model is evaluated for US natural gas and crude oil futures by comparing the pricing performance to reduced form models. Results suggest jump models, both reduced form and economic, improve modeling due to incorporating pricing discontinuities. Furthermore, the economic model precludes carry arbitrage, which appears relevant for pricing natural gas futures. For crude oil, the reduced form models produce superior pricing under nonstationary market conditions, and the economic model produces superior long‐dated futures pricing under stationarity.

The Impact of Outliers on Computing Conditional Risk Measures for Crude Oil and Natural Gas Commodity Futures Prices

In this study, we aim to build better risk models for energy commodities by employing statistical procedures to identify outliers in the prices for all crude oil and natural gas futures contracts traded on the CME over the period of December 2003 through March 2017. Our results show that it is important to investigate and control for potential outlier effects when performing parametric estimation of risk parameters because outliers can have a large impact on the estimation of value at risk. We illustrate for actual crude oil and natural gas contract how the use of value at risk metrics based on raw data can lead to higher than expected actual losses. Our research demonstrates that it is crucial to include intervention parameters to address outlier impacts in order to obtain robust risk metrics.

What should a small load serving entity use to hedge its procurement cost risk?

This paper answers the question: what should a small load serving entity (LSE) use to hedge its procurement cost risk? In doing so, it proposes a transparent process to compute a hedging strategy’s procurement cost expectations and ceilings. Its key finding is that if the LSE is risk averse, it should use electricity forwards when the forward price premium is relatively small and cross hedging with natural gas futures is relatively ineffective.

Expectations for Statistical Arbitrage in Energy Futures Markets

Energy futures have become important as alternative investment assets to minimize the volatility of portfolio return, owing to their low links with traditional financial markets. In order to make energy futures markets grow further, it is necessary to expand expectations of returns from trading in energy futures markets. Therefore, this study examines whether profits can be earned by statistical arbitrage between wholesale electricity futures and natural gas futures listed on the New York Mercantile Exchange. On the assumption that power prices and natural gas prices have a cointegration relationship, as tested and supported by previous studies, the short-term deviation from the long-term equilibrium is regarded as an arbitrage opportunity. The results of the spark-spread trading simulations using historical data from 2 January 2014 to 29 December 2017 show about 30% yield at maximum. This study shows the possibility of generating earnings in energy futures market.

Pricing dynamics of natural gas futures

Growth in the natural gas market is pronounced since the shale gas boom. Natural gas has become increasingly important in international trade, especially after the recent financialization in commodity markets. Motivated by the high volatility and time-varying nature in natural gas futures prices, understanding the pricing dynamics of natural gas is essential for risk management. In this paper, we adopt a class of computationally efficient discrete-time pricing models and construct futures dynamics by differentiating three subsamples which represent time-varying market conditions. We find strong evidence of positive jumps in the natural gas market and higher jump intensity in a more volatile period. The dynamic jump intensity model has a better model fit both in-sample and out-of-sample, suggesting time-varying jumps are necessary for pricing natural gas derivatives.

The relationship between the Nasdaq Composite Index and energy futures markets

This paper sheds light on the relationship between the Nasdaq Composite Index and a newly proposed Energy Futures Conditions Index (EFCI). While various financial conditions indices provide information about the financial stability of a country, the existence of an energy condition index, using futures markets, is scarce. Using weekly data over the period 1992–2017, this paper introduces an energy futures index using principal component analysis and test its predictability over the Nasdaq Composite Index. The EFCI captures 95% of the variability inherent in crude oil, heating oil and natural gas futures’ total reportable positions. Stability in forecast errors over different lags suggests a one week lag is sufficient to forecast weekly Nasdaq Composite Index. 95% prediction levels support that the estimated model captures actual equity market index values, except for the 2000 technology bubble. Distributions of level data were non-normal, not serially correlated and homoscedastic under the whole sample period, with diagnostics on pre and post technology bubble crisis showing mixed results. While differencing ensured homoscedastic errors in the forecasting model, Granger causality supported non-causality from both energy futures and equity markets, suggesting no evidence of cross market information flows.

A novel market efficiency index for energy futures and their term structure risk premiums

This paper proposes a novel futures market efficiency index which aggregates the efficiencies of futures contracts across their term structure spanning from one month to 12 months. The index measures the ability for price discovery of a long-term futures contract on its nearby short-term contract consistently across its terms. The index uses a recently developed futures market efficiency test which accounts for heteroscedastic prices and time-varying risk premiums. It simultaneously estimates the term premiums of futures, providing valuable information for the investor. The proposed index is employed to investigate the efficiencies of four major energy commodities, namely, crude oil, natural gas, heating oil, and gasoil during the sample period 1990–2016. Numerical results indicate that the market efficiencies vary across terms and across energy commodities. Gasoil futures traded on the Intercontinental Exchange (ICE) are the most efficient and natural gas futures are the least efficient in prices. The spillover dynamics of market risk information across futures markets are investigated using conditionally heteroscedastic common factors (CHCFs) extracted for each commodity using the estimated term premiums. There exist significant delayed, contemporaneous, and potential information spillovers among term premiums of the energy commodities in our sample. Our findings are robust and they can help investors to optimally diversify their portfolios.

Analysis of risk premium in UK natural gas futures

In many futures markets, trading is concentrated on the front contract and positions are rolled-over until the strategy horizon is attained. In this paper, a pair-wise comparison between the conventional risk premium and the accrued risk premium in rolled-over positions on the front contract is carried out for UK natural gas futures. Several novel results are obtained. Firstly, and most importantly, the accrued risk premium in rollover strategies is significatively larger than conventional risk premiums and increases with the time to delivery. Specifically, for strategy horizons between three and six months, this difference increases from 1% to 10% (or from 4% to 20% in annualized returns). Secondly, it is the first time that risk premium in day-ahead forwards has been measured in this market. The average value of the day-ahead risk premium is 0.5% per day and it is statistically significant. Thirdly, all risk premiums are significantly larger and more volatile in winter. Finally, risk premium time-variation is analysed using a regression model. It is shown that reservoir shocks, demand shocks and spot price volatility are significant explicative variables and its sign reflects equilibrium models implications for storable commodities.

Anatomy of Risk Premium in UK Natural Gas Futures

In many futures markets, trading is concentrated in the front contract and positions are rolled-over until the strategy horizon is attained. In this paper, a pair-wise comparison between the conventional risk premium and the accrued risk premium in rolled-over positions in the front contract is carried out for UK natural gas futures. Several novel results are obtained. Firstly, and most importantly, the accrued risk premium in rollover strategies is significatively larger than conventional risk premiums and increases with the time to delivery. Specifically, for strategy horizons between three and six months, this difference increases from 1% to 10%. Secondly, it is the first time that risk premium in day-ahead futures has been measured in this market. The average value of the day-ahead risk premium is 0.5% per day and it is statistically significant. Thirdly, all risk premiums are significantly larger and more volatile in winter. Finally, risk premium time-variation is analyzed using a regression model. It is shown that reservoirs, weather, liquidity, volatility, skewness, and seasons are able in all cases to explain between 21% and 59% of the risk premium time-variation (depending on the futures maturity and sub-period).

Best Fitting Fat Tail Distribution for the Volatilities of Energy Futures: Gev, Gat and Stable Distributions in GARCH and APARCH Models

Precise modeling and forecasting of the volatility of energy futures is vital to structuring trading strategies in spot markets for risk managers. Capturing conditional distribution, fat tails and price spikes properly is crucial to the correct measurement of risk. This paper is an attempt to model volatility of energy futures under different distributions. In empirical analysis, we estimate the volatility of Natural Gas Futures, Brent Oil Futures and Heating Oil Futures through GARCH and APARCH models under gev, gat and alpha-stable distributions. We also applied various VaR analyses, Gaussian, Historical and Modified (Cornish-Fisher) VaR, for each variable. Results suggest that the APARCH model largely outperforms the GARCH model, and gat distribution performs better in modeling fat tails in returns. Our results also indicate that the correct volatility level, in gat distribution, is higher than those suggested under normal distribution with rates of 56%, 45% and 67% for Natural Gas Futures, Brent Oil Futures and Heating Oil Futures, respectively. Implemented VaR analyses also support this conclusion. Additionally, VaR test results demonstrate that energy futures display riskier behavior than S&P 500 returns. Our findings suggest that for optimum risk management and trading strategies, risk managers should consider alternative distributions in their models. According to our results, prices in energy markets are wilder than the perception of normal distribution. In this regard, regulators and policy makers should enhance transparency and competitiveness in the energy markets to protect consumers.

What drives informed trading before public releases? Evidence from natural gas inventory announcements

This paper shows evidence of informed trading in the natural gas futures market before gas inventory announcements. We examine whether traders can predict the upcoming announcement by processing public information. The results show that the difference between the median forecast of analysts with high historical forecasting accuracy and the consensus forecast can be used to predict inventory surprises. This predictor explains some of the pre‐announcement price drift, suggesting that informed trading before the announcement is likely to be driven by superior forecasting rather than by information leakage. A simple trading strategy conditioned on the predictor would have generated an annualized Sharpe ratio of 1.26.

Risk premia in commodity price forecasts and their impact on valuation

Commodity price driven valuation models require a stochastic price input if the value of managerial flexibility, such as the option to defer investment until the optimal time and the option to abandon a project, is to be estimated. The risk-neutral version of the stochastic price model is typically used in academic work; however, risk-adjusted models of the expected spot price are often used in practice. These two approaches are connected by a risk premium which is unfortunately often difficult to estimate. In this work, we use natural gas futures prices in a Kalman filter approach with maximum likelihood estimation to parameterize the Schwartz and Smith (2000) stochastic price model, and then apply an asset pricing model to address the large uncertainty of the risk premia parameter estimates. To evaluate the impact of the risk premia and other parameters in the two-factor price model on project valuation, we apply the price model to a prototypical shale gas investment, both for a base reference case as well as for cases where there are real options to optimally time decisions to invest or to abandon the project. Using this approach, we are able to determine the implied risk-adjusted discount rate that would be used with the spot price forecast, given the two-factor model risk premia, and we also discuss the impact of the risk premia on project value relative to other model parameters.

Testing Co-Volatility spillovers for natural gas spot, futures and ETF spot using dynamic conditional covariances

There is substantial empirical evidence that energy and financial markets are closely connected. As one of the most widely-used energy resources worldwide, natural gas has a large daily trading volume. In order to hedge the risk of natural gas spot markets, a large number of hedging strategies can be used, especially with the rapid development of natural gas derivatives markets. These hedging instruments include natural gas futures and options, as well as Exchange Traded Fund (ETF) prices that are related to natural gas stock prices. The volatility spillover effect is the delayed effect of a returns shock in one physical, biological or financial asset on the subsequent volatility or co-volatility of another physical, biological or financial asset. Investigating volatility spillovers within and across energy and financial markets is a crucial aspect of constructing optimal dynamic hedging strategies. The paper tests and calculates spillover effects among natural gas spot, futures and ETF markets using the multivariate conditional volatility diagonal BEKK model. The data used include natural gas spot and futures returns data from two major international natural gas derivatives markets, namely NYMEX (USA) and ICE (UK), as well as ETF data of natural gas companies from the stock markets in the USA and UK. The data used for the empirical analysis is from 14 May 2007 to 15 April 2016, incorporating 2330 observations. The empirical results show that there are significant spillover effects in natural gas spot, futures and ETF markets for both USA and UK. Such a result suggests that both natural gas futures and ETF products within and beyond the country might be considered when constructing optimal dynamic hedging strategies for natural gas spot prices.