Search results
Jan 13, 2021 · Commercial inventories declined. WTI oil prices were mostly at or near $40 per barrel, but closed out 2020 at $48.52. Indicators suggest a recovering oil market with potential for higher, yet moderate, prices in the short term. However, demand, supply, and uncertainties in U.S. policy toward Iran could change 2021 market and price conditions.
- Realised Volatility
- Traditional Predicting Models
- Machine Learning Models
The monthly RV of oil futures is calculated in this study using daily returns: where \({r}_{t,j}\) represents the jth daily return of month t and \(M\) denotes the number of observations of month t.
3.2.1 AR model, benchmark
Following Paye (2012), we apply the AR model as follows: where \({\text{L}\text{n}\left(\text{R}\text{V}\right)}_{t+1}\) represents the natural logarithm of \({\text{R}\text{V}}_{t+1}\), \({\alpha }_{k}\)represents the coefficient of k-lags Ln(RV), and \({\epsilon }_{t+1}\)represents the disturbance term. Three lags is a relatively good choice based on the AIC and SC information criteria, and we employ the AR(3) model to perform the analysis.
3.2.2 Kitchen sink model
Furthermore, according to Rapach et al. (2010) and Ma et al. (2019), we construct the kitchen sink model (AR-all model) by adding all macroeconomic and economic predictors to the AR(3) model. where \({X}_{n,t}\) represents the n-th predictor and Nis the number of total macroeconomic predictors. To avoid overfitting, we use several commonly used dimensionality reduction methods, including a principal component analysis (PCA), scaled PCA (SPCA), and partial least squares (PLS), to obtain new in...
3.2.3 Principal component analysis model
The AR-PCA method can be expressed as follows: where \({F}_{p,t}^{\text{P}\text{C}\text{A}}\) represents the principal component extracted from all the macroeconomic and economic predictors. Following Neely et al. (2014), P is chosen based on the adjusted\({R}^{2}.\)
3.3.1 Support vector regression model
The support vector regression (SVR) model is a prominent machine learning algorithm for regression. SVR’s objective is to cover as many sample points as feasible with a fixed width band (the width is regulated by the parameter ϵ) while keeping the overall error as small as possible. The SVR model function is a linear function written as follows: where \(\text{y}\) = \({\text{L}\text{n}\left(\text{R}\text{V}\right)}_{t+1}\) represents the natural logarithm of \({\text{R}\text{V}}_{t+1}\), \(x\...
3.3.2 K-nearest neighbour model
The K-nearest neighbour (KNN) model (Altman, 1992) is a popular supervised learning method that determines the k training samples in the training set that are closest to a given test sample based on a distance measure and subsequently predicts the target value of the test sample using the information of these k “neighbours.” The distance, Lp, between the samples is measured using the following formula: where \({x}_{i}=({x}_{i}^{1},{x}_{i}^{2},\ldots,{x}_{i}^{N})\) and \({x}_{j}=({x}_{j}^{1},{...
3.3.3 Decision trees model
Yeh (1991) defined decision trees (DT) as a nonparametric supervised learning approach. A DT is generative by a recursive construction of a binary tree. To create the binary tree in the regression task, feature selection is accomplished using the squared error minimisation criterion. The decision tree generation algorithm is as follows. First, the optimal split variable \(v\) and the split point \(s\) are selected. This is accomplished by traversing the variable \(v\), scanning the split poin...
- GDP Growth Rebounds Following a Steep Decline in 2020. Description. Description: This chart illustrates the short-term macroeconomic impact of COVID-19 through real GDP and GDP growth trends from 2018 to 2025.
- Economic Indicators, Evolving and Current Policies Scenarios (2019-2050) Description. Description: This chart shows average annual growth rates from 2019 to 2050 of several economic indicators for both the Evolving Policies and Current Policies scenarios.
- End-use Demand Declines in All Sectors in the Evolving Policies Scenario.
- End-use Energy Consumption Peaks in 2019 and Declines over the Long Term in the Evolving Policies Scenario. Figure R.3 Description.
Mar 17, 2021 · Oil demand in 2025 is set to be 2.5 mb/d lower than was forecast a year ago in our Oil 2020 report. All of this demand growth relative to 2019 is expected to come from emerging and developing economies, underpinned by rising populations and incomes. Asian oil demand will continue to rise strongly, albeit at a slower pace than in the recent past.
Oct 1, 2024 · The study utilizes external factors to improve the precision of predicting fluctuations in oil prices. Based on the pertinent research, the present research gathers 62 external factors from 2000 to 2021, that indicate the changes in oil need, oil supplies, oil stock, economic basics, monetary metrics, and estimations of unpredictability.
Table 1 displays the data provided by the Bloomberg terminal on January 31, 2014, which includes (1) a list of the different analyst firms in the first column labeled as Firm, (2) the exact date on which the price forecast is released under the second column designated as As Of, (3) the predictions for the average crude oil price in 2014 under the third column labeled 2014, and (4) subsequent ...
People also ask
What uncertainties affect oil and petroleum price forecasts for 2021?
Does oil price uncertainty affect growth rate?
Does oil price uncertainty affect climate policies?
What factors determine the fluctuation of crude oil prices?
Why does crude oil production increase in 2032?
What are the key factors affecting oil demand in 2021?
Dec 27, 2022 · We investigate whether the United States economy responds negatively to oil price uncertainty and whether oil price shocks exert asymmetric effects on economic activity. In doing so, we use monthly data and modify the (Elder and Serletis J Money Credit Bank 42, 1138–1159, 2010) bivariate structural GARCH-in-Mean VAR to accommodate the interaction between oil price uncertainty and the climate ...
