Code task G3C. Maritime Analytics individual assignment SPM 502. Class 2023 (Statistics Project Sample)

TABLE OF CONTENTS LIST OF TABLES …………………………………………………………………………….3 LIST OF FIGURES ……………………………………………………………………………3 LIST OF ABBREVIATIONS………………………………………………………………….3 3. FINDINGS 3.1 Introduction ………………………………………………………………………………….4 3.2 Findings………………………………………………………………………………………5 3.2.1 Descriptive Statistics…………….........................................................................................5 3.2.2 Natural Log. ………………………………………………………………………..............6 3.2.3 Stationary Test (Unit Root Test)............................................................................................6 3.2.4 Correlation Test between the Independent Variables ……………………………………....9 3.2.5 Coefficient Diagnostics Test (T-Test) ……………………………………………………...9 3.2.6 Eagle Granger-Cointegration Test ………………………………………………………...10 3.2.7 Arma Model ……………………………………………………………………………….10 3.2.8 Residual Test ……………………………………………………………………………....12 3.2.9 Chow Test (Breakpoint) …………………………………………………………………...12 3.2.10 Forecasting Test ………………………………………………………………………….13 3.2.11 Model before the Shock ………………………………………………………………….14 3.2.12 Model after the Shock ……………………………………………………………………19 3.3 Conclusion …………………………………………………………………………………..23 References ……………………………………………………………………………………….25 LIST OF TABLES Table 1 Descriptive statistics Table 2 Correlation Table 3 Correlation Table 4 T-Test (main regression) Table 5 Cointegration (main regression) Table 6 Autoregressive Process (AR) and Moving Average (MA) (ARMA Model) Table 7 Autoregressive Process (AR) and Moving Average (MA) (ARMA Model) Table 8 Residual Test (main regression) LIST OF FIGURES Figure 1 Chow test (main regression) Figure 2 Forecasting test (main regression) Figure 3 Break Point test (before the shock) Figure 4 Forecasting test (before the shock) Figure 5 Break Point test (after the shock) Figure 6 Forecasting test (after the shock) LIST OF ABBREVIATIONS OPEC ADF WTI The Organization of the Petroleum Exporting Countries Augmented Dickey-Fuller West Texas Intermediate ARMA Autoregressive Moving Average CPI Consumer Price Index MA Moving Average CLRM Classical Linear Regression Model GLS Generalized Least Square HAC Heteroscedasticity and Autocorrelation Consistent KPSS Kwartuwski Philips Schmidt Shin OLS Ordinary Least Squares PP Philips Perron RESET Regression Equation Specification Error Test US United States VLCC Very Large Crude Carrier 3. FINDINGS 3.1 Introduction: Oil prices and oil price volatility both have a significant impact on the world economy, albeit the impacts vary based on the time of year, the area, the industry, the cause of the oil shock, and other factors. Various points of view have been put out regarding how changes in oil prices would affect the world economy. Give a good explanation of these many viewpoints, for instance. Numerous research revealed via this discussion that increasing oil prices have a negative effect on the world economy. Furthermore, it is discovered that the negative economic effects of increasing oil prices for oil-importing nations like South Korea are considerably more severe. It is crucial to precisely predict future oil prices with effective models in order to make informed judgements regarding the direction of economic policy. Oil prices throughout the world, which had been rising since 2003, soared to $134/Bbl (for West Texas Intermediate, WTI) in June 2008. After the worldwide economic slump of 2008, oil prices declined, but they soon began to increase in early 2009. Up to 2040, a general decreasing trend in the growth rate of the world's oil demand is anticipated. Studies have proposed a number of potential explanations for this anticipated slowing in the increase of the global oil demand, including shale gas extraction in the United States, consumer reactions, and governmental initiatives. The price of crude oil fell to less than $50/Bbl in 2014 when the Organisation of the Petroleum Exporting Countries (OPEC) resolved to sustain oil output. Due to the persistent weakness in oil demand and the robust shale production in 2015 and 2016, the price has remained at mid-$40/Bbl. As a result, concerns over fluctuating oil prices and a new oil crisis have grown. Understanding the long-term trend in the structural changes in the factors affecting oil prices in this context is important. Up to the oil price crash in the mid-1980s, supply-side factors dominated the determination of crude oil price. As a result, since the late 1980s, there has been an oil pricing system linked to the oil market, and the price of crude oil is influenced by both supply and demand. Particularly in the 1990s, rising oil costs were influenced by developing nations like China and India. Since 2000, financial variables have drawn attention as potential predictors of global oil prices. These reasons include the penetration of speculative forces, a declining dollar, and the financial crisis. For instance, Morana discovered that financial shocks significantly, and to a much greater extent, since the middle of the 2000s, contributed to the rise in oil prices. Speculative anticipation has been identified as one of the financial elements that significantly influences commodity price. In addition, several micro and macroeconomic variables impact crude oil prices. These characteristics are qualitative as well as quantitative in nature. However, for the sake of this study, we have limited the modelling and analysis to quantitative factors. The Brent Crude Oil Price has been defined as the Dependent Variable, with 37 independent factors that could affect the Brent Crude Oil price. Our selection of independent variables is based on the interaction of crude oil demand and supply, which sets prices at the equilibrium point assuming all other parameters remain constant. This project will uncover elements that influence oil prices and will provide a realistic forecast model of oil prices. A lot of things must be completed in order to reach this goal: To investigate the elements that influence the price of oil; consider the auto-regressive integrated moving average (ARIMA) forecasting technique. Collect and analyse descriptive data; create a regression model and find significant variables; Obtain forecasts using the methods given above, select the best one, and base a future forecast on it. 3.2 Findings: Factors affecting the price of Brent crude oil are estimated with OLS regression, using MATLAB statistical software. The estimation results are also obtained using one model. Lastly, this study will continue and show all of the OLS regression phases' findings. If there is a substantial influence of the significant independent variable on the dependent variable, it can be told from the data shown here. To get these findings, three regressions were run as follows: * Global regression * Regression before the shock * Regression After the shock. Also, the outcomes will demonstrate the model's goodness of fit and illustrate whether or not the models are reliable and suitable for use as inputs for the ARIMAX model. 3.2.1 Descriptive Statistics: The table below shows the result of the descriptive statistics (character of the data) Table 1: 3.2.2 Natural Log Following the descriptive statistics, a MATLAB program was used to convert the data into a natural logarithm. This is because, for this study, we're interested in seeing how the data we've gathered on Brent crude oil prices have evolved over time. In order to continue with the stationarity test, the data was transformed into returns following the descriptive statistics. 3.2.3 Stationary Test (Unit Root Test) The outcomes of the stationarity test utilizing the ADF, KPSS, and PP techniques are shown in the table below. All non-stationary variables were changed to stationary using the MATLAB program. The Unit Root Test showed that of the 38 variables, only two were originally stationary. I (0) process while the rest were non stationary but all became stationary at first difference. That means I (1) process. The two I (0) are: * VLCC order book * China’s Interest Rate Unit Root Test: Brent crude oil price Even if the variables are cointegrated, it is still vital that they be close to one another and not traveling in opposite directions, hence it is crucial that all 36 variables in I (1) process are checked for the Engle-Granger technique later on. After the coefficient diagnostic, the study's findings using cointegration will be given. Table 2: Table 3: 3.2.4 Correlation Test between the Independent Variables Testing for multicollinearity came after the non-stationary variables had been transformed. It was necessary to find in...