The Global Financial Crisis, starting in 2007, served as a reminder of the serious impact that imbalances originating in financial markets can have on economic growth. The aftermath of this economic shock with the ensuing recession continues to concern policymakers to this day. The subsequent period characterized by subdued growth and few but severe recessions gave rise to the importance of linkages between economic policy and risk management. The connection between this idea and the relevance of financial variables for analyzing growth risks is established by Adrian et al. (2019). They employ quantile regressions to examine the conditional distribution of future GDP growth and find that its left tail is exposed to substantially more volatility over time than the right tail. Moreover, they find that financial conditions for the US measured by the National Financial Conditions Index (NFCI) can serve as a relevant predictor of downside risk to conditional future economic growth.
This thesis examines some machine-learning based variable selection methods that have been largely unexplored in the GaR context. The focus is on generating higher predictive power compared to the model by Adrian et al. (2019) rather than on analyzing economic relationships. The approaches described here are easy to apply and can help to automate the selection of variables for GaR estimation instead of having to manually choose relevant indicators. In detail, the LASSO method is used in the quantile regression context (Belloni and Chernozhukov 2011; Li and Zhu 2008), as well as the Adaptive (Wu and Liu 2009) and Relaxed LASSO (Meinshausen 2007), two of its modifications. In addition, the Elastic Net method is investigated as a compromise between Ridge and LASSO regression.
To test the performance of these models, a backtesting exercise is conducted based on US data ranging from 1986 to 2019. The out-of-sample analysis is performed under the expanding and rolling window approach. For evaluation of the models, some of the backtesting tools used by Brownlees and Souza (2019) to perform a similar analysis for volatility models in the GaR context are utilized. In this regard, the following research question is formulated: Can the machine learning-based models improve the predictive power measured by the introduced backtesting tools for the investigated period compared to the quantile regression base model?
Inhaltsverzeichnis (Table of Contents)
- Introduction
- The Concept of Growth-at-Risk
- Literature Overview
- Quantile Regression
- LASSO Quantile Methods
- LASSO Quantile Regression
- Relaxed LASSO Quantile Regression
- Adaptive LASSO Quantile Regression
- Elastic Net Quantile Regression
- Backtesting
- Empirical Analysis
- Data
- In-Sample Analysis
- Out-of-Sample Analysis
- Expanding Window Results
- Rolling Window Results
- Variable Selection in Out-of-Sample Analysis
- Discussion
- Conclusion
Zielsetzung und Themenschwerpunkte (Objectives and Key Themes)
This master thesis examines the effectiveness of various machine learning-based variable selection methods in the Growth-at-Risk (GaR) context. It seeks to determine whether utilizing these methods for variable selection can improve the accuracy of GaR predictions compared to traditional methods that rely on aggregated indices.
- Comparison of Machine Learning Methods for Variable Selection in GaR
- Evaluation of the Performance of Machine Learning Models in Out-of-Sample Analysis
- Analysis of the Impact of Variable Selection on Prediction Accuracy
- Investigating the Stability of Variable Selection Across Different Quantiles, Prediction Horizons, and Time Periods
- Exploration of the Potential for Improved Prediction Results in Times of Economic Crisis
Zusammenfassung der Kapitel (Chapter Summaries)
- Introduction: This chapter provides a brief overview of the research question and the methods used in the thesis. It introduces the concept of Growth-at-Risk (GaR) and its relevance in economic forecasting.
- The Concept of Growth-at-Risk: This chapter delves into the theoretical background of GaR, discussing its definition, relevance, and existing literature. It introduces quantile regression as a fundamental tool for estimating GaR and explores various LASSO quantile methods used for variable selection. The chapter concludes with a discussion of backtesting methodologies.
- Empirical Analysis: This chapter presents the empirical study conducted in the thesis. It describes the dataset used, the in-sample analysis results, and the out-of-sample analysis results. The out-of-sample analysis is conducted using both expanding and rolling windows to assess the predictive power of the models over time. The chapter concludes with a discussion of the variable selection process and its impact on the models' performance.
Schlüsselwörter (Keywords)
The primary focus of this thesis revolves around machine learning applications in the context of Growth-at-Risk (GaR). Key concepts include variable selection methods, LASSO quantile regression, Elastic Net, quantile regression, backtesting, expanding window, rolling window, and out-of-sample analysis. The research examines the performance of these methods in predicting GaR, with a particular interest in their effectiveness during periods of economic crisis.
- Quote paper
- Franz Lennart Wunderlich (Author), 2022, Machine Learning in the Growth-at-Risk Context. A Comparison of Predictors, Munich, GRIN Verlag, https://www.grin.com/document/1274858
