Application of GARCH type model is a key for modeling and forecasting volatility for high frequency data such as daily commodity price. Following the same framework, the objective of the present study is to apply the multiplicative GARCH-MIDAS model for daily exported coffee price as proxy of daily total coffee price of Ethiopia over the period of 1-1-2008 to 7-17-2018 with the purpose of fitting and forecasting coffee price returns volatility.
The GARCH-MIDAS model decomposes the conditional variance as short-term component of GARCH (1,1) process, and long-term component, with monthly frequencies of macroeconomic variables. In this study exchange rate (nominal exchange rate), inflation rate (general inflation), interest rate (lending interest rate), fuel oil price (price of imported petroleum and petroleum production), total consumption and money supply (broad money) macroeconomic variables were employed through MIDAS specification using beta-weighting scheme to analyze impact of the variables on the long-term volatility component. For fitted ARMA (1,1) of coffee price return ARCH effect test on the residual from the mean model revealed the existence of time varying conditional variance for the selected mean model. A conditional variance model GARCH (1,1) was selected and used to model the conditional variance of coffee price return with Quasi Maximum Likelihood along with Bayesian estimation methods and both estimation procedures indicated the persistence of conditional variance observed even for small sample under Bayesian estimation framework. Asymmetry test show the insignificance of the asymmetric term, while Lundbergh and Terasvirta Lagrange Multiplier and the Li-Mak portmanteau test for the residual of GARCH model show the existence of time varying unconditional variance and made call for GARCH-MIDAS model. From the result of estimated GARCH-MIDAS model exchange rate and inflation rate were found to be the best drivers of coffee price volatility in Ethiopia and used for in and out of sample forecast.
Finally, the Mean Absolute Error Root Mean Square Error and Diebold Mariano test were used for evaluating and comparing the forecasting ability of GARCH-MIDAS component model against standard GARCH (1,1) model which indicated that, including exchange rate and inflation rate make efficient forecasting of coffee price volatility in Ethiopia.
Table of Contents
1. INTRODUCTION
1.1. Background
1.2. Statement of the Problem
1.3. Objectives of the Study
1.4. Significance of the Study
1.5. Scope and Limitations of the Study
1.6. Organization of the Thesis
2. LITERATURE REVIEW
2.1. Theoretical Review
2.1.1. Volatility
2.1.2. Price Volatility and Coffee Consumption in Ethiopia
2.1.3. Price Volatility and Coffee Export Status in Ethiopia
2.1.4. Economic Importance of Price Volatility
2.2. Related Theories of Price Volatility
2.2.1. Arbitrage Pricing Theory
2.2.2. Future Cash Flow Discounted Model Theory
2.3. Theoretical Framework of Volatility Models
2.3.1. Volatility Measurement and Theoretical Review of Volatility Models
2.3.2. Model of Mixed Data Sampling (MIDAS)
2.3.3. GARCH-MIDAS Component Model
2.4. Empirical Literature Review
3. RESEARCH METHODOLOGY
3.1. Data Source and Variables
3.2. Measurement and Definitions of Variables
3.3. Financial Time series and their Characteristics
3.4. Test for Stationarity
3.4.1. The Augmented Dickey Fuller (ADF) Test
3.4.2. The Phillips and Perron (PP) Test
3.5. Model Specification
3.5.1. Autoregressive Moving Average (ARMA)
3.5.2. Autoregressive Integrated Moving Average (ARIMA)
3.6. Conditional Volatility Model
3.6.1. ARCH model
3.6.2. GARCH model
3.7. Residual Diagnostics for Building GARCH Type Model
3.7.1. Lagrange Multiplier Test for ARCH Effect
3.7.2. Ljung Box Test
3.7.3. Test for Asymmetry in Volatility
3.7.4. Test of Normality of Residual
3.8. Estimation of GARCH (p, q)
3.8.1. Maximum Likelihood Estimation
3.8.2. Distribution Assumptions of Error
3.8.3. Bayesian Estimation of GARCH (p, q)
3.8.4. Model Priors and MCMC Schemes
3.9. GARCH-MIDAS Model
3.9.1. Short-term Volatility Component
3.9.2. Long-run Volatility Component (MIDAS)
3.9.3. MIDAS-Realized Volatility
3.9.4. MIDAS with Macroeconomic Variables
3.10. Assumptions of GARCH-MIDAS Component Models
3.11. Test for Building GARCH-MIDAS Model
3.11.1. Test for Time Varying Unconditional Volatility
3.11.2. Lagrange Multiplier (LM) Test
3.11.3. Portmanteau Test
3.12. Estimation of GARCH-MIDAS Component Model
3.12.1. Quasi Log-Likelihood Estimation
3.12.2. Quasi Log-Likelihood Function
3.13. Model Adequacy Checking
3.14. Forecasting Using GARCH-MIDAS Model
3.15. Evaluation of Forecasting Accuracy
3.15.1. Statistical Loss Function
3.15.2. Modified Diebold and Mariano Test
4. RESULTS AND DISCUSSION
4.1. Data Description
4.2. Descriptive Analysis Results
4.2.1. Graphical Analysis Results
4.2.2. Summary Statistics
4.3. Unit Root Test of Stationarity for Study Variables
4.4. Mean Model Specification
4.5. Residual Diagnostics for Mean Model
4.5.1. ARCH-LM Test for Estimated ARMA Model of Coffee Price Return
4.5.2. Ljung-Box Test from Estimated ARMA Model of Coffee Price Return
4.5.3. Result for Normality Test
4.6. Specification and Estimation of GARCH Model
4.6.1. Estimation of coffee price return using GARCH (1,1) under student-t distribution
4.6. 2. Normality test for residual from GARCH (1,1)
4.6. 3. Asymmetry Test
4.6.4. Time Varying Unconditional Volatility Test
4.6.5. Result of Li and Mak Portmanteau test
4.7. Bayesian Estimation Result of GARCH (1,1)
4.8. GARCH-MIDAS Model Estimation Result
4.9. In-sample Forecasting Using GARCH-MIDAS Model
4.10. Out- sample Forecasting Using GARCH-MIDAS
4.10.1. Evaluation of Forecasting Accuracy
4.10.2. Modified Diebold and Marino Test
5. CONCLUSION AND RECOMMENDATIONS
5.1. Conclusion
5.2. Recommendations
Objectives and Topics
The primary objective of this study is to model and forecast the volatility of daily exported coffee prices in Ethiopia using the multiplicative GARCH-MIDAS model, incorporating various macroeconomic variables such as exchange rates and inflation to better understand their impact on both short-term and long-term volatility components.
- Application of the GARCH-MIDAS model to high-frequency coffee price data.
- Evaluation of macroeconomic determinants (exchange rate, inflation, interest rate, etc.) on price volatility.
- Comparison of GARCH-MIDAS forecasting performance against standard GARCH(1,1) models.
- Utilization of Bayesian estimation methods for small-sample robustness.
- Assessment of the informational gain from using mixed-frequency data sampling.
Excerpt from the Book
2.1.1. Volatility
Volatility is one of the most important concepts in finance. It tells us about the uncertainty or risk of financial assets. Financial series bear a widely accepted feature called volatility clustering. Volatility clustering is the phenomenon, which a large volatility is followed by a large volatility and vice versa. In other words, there is a positive correlation of returns in different periods. This phenomenon violates the assumption of homoscedasticity (Bolerslev, 1986).
Summary of Chapters
1. INTRODUCTION: Outlines the background of the coffee industry in Ethiopia, identifies the problem of price volatility, and sets the research objectives and scope.
2. LITERATURE REVIEW: Reviews theoretical concepts of volatility, relevant economic theories such as Arbitrage Pricing Theory, and existing empirical studies on GARCH-type models.
3. RESEARCH METHODOLOGY: Details the data sources, variable definitions, and the mathematical framework for ARMA, GARCH, and GARCH-MIDAS models, including estimation techniques.
4. RESULTS AND DISCUSSION: Presents the empirical findings, including descriptive statistics, unit root tests, estimation results of GARCH and GARCH-MIDAS models, and performance evaluation.
5. CONCLUSION AND RECOMMENDATIONS: Summarizes the study's findings and provides policy recommendations for the Ethiopian government regarding fiscal and monetary measures.
Keywords
Daily coffee price, GARCH-MIDAS model, Bayesian estimation, Short-run volatility component, Long-run volatility component, Ethiopia, Exchange rate, Inflation rate, Price volatility, Forecasting accuracy, Macroeconomic determinants, Econometrics, Volatility clustering, Financial markets, Commodity exports.
Frequently Asked Questions
What is the core focus of this research?
The research focuses on modeling and forecasting the volatility of daily exported coffee prices in Ethiopia by identifying the macroeconomic determinants that influence its fluctuations.
Which primary economic variables are analyzed?
The study examines exchange rates, inflation, interest rates, fuel oil prices, total consumption, and money supply as exogenous variables affecting coffee price volatility.
What is the central research goal?
The goal is to apply the GARCH-MIDAS model to overcome frequency mismatches in data and to accurately identify which macroeconomic factors serve as the most effective drivers of long-term coffee price volatility.
What methodology is employed in the study?
The study uses the GARCH-MIDAS (Mixed Data Sampling) approach alongside traditional GARCH(1,1) and ARMA models, utilizing both Maximum Likelihood and Bayesian estimation techniques.
What topics are covered in the main body of the work?
The work covers theoretical reviews of volatility, the development of GARCH-type models, empirical literature on commodity price behavior, methodology for stationary testing and model specification, and an extensive discussion of empirical results.
Which keywords best characterize this work?
Key terms include GARCH-MIDAS, Ethiopia, coffee price volatility, macroeconomic determinants, Bayesian estimation, and financial forecasting.
How does the GARCH-MIDAS model differ from standard models?
Standard GARCH models fail to capture time-varying unconditional variance when dealing with variables of different frequencies; the GARCH-MIDAS model specifically addresses this by decomposing volatility into short-term (GARCH) and long-term (MIDAS) components.
What does the study conclude regarding the drivers of volatility?
The study finds that the exchange rate and the inflation rate are the most significant drivers influencing coffee price volatility in Ethiopia, both having a positive impact.
Why is the Bayesian estimation approach used?
Bayesian estimation is employed to ensure robust results, particularly when dealing with small sample sizes, and to provide probabilistic statements about model parameters.
What practical recommendations are offered?
The author recommends that the government implement specific fiscal and monetary policies to stabilize exchange rates and mitigate the inflationary pressures that exacerbate coffee price instability.
- Arbeit zitieren
- Tekle Bobo (Autor:in), Tesfaye Abera (Autor:in), Jema Haji (Autor:in), 2020, Macroeconomic Determinants of the Coffee Price Volatility in Ethiopia. Application of the Garch-Midas Model, München, GRIN Verlag, https://www.grin.com/document/936705
