Machine Learning Regularization Methods in High-Dimensional Monetary and Financial VARs
Vector autoregressions (VARs) and their multiple variants are standard models in economic and financial research due to their power for forecasting, data analysis and inference. These properties are a consequence of their capabilities to include multiple variables and lags which, however, turns into...
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2022-03-01
|
| Series: | Mathematics |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2227-7390/10/6/877 |
| _version_ | 1797445452634783744 |
|---|---|
| author | Javier Sánchez García Salvador Cruz Rambaud |
| author_facet | Javier Sánchez García Salvador Cruz Rambaud |
| author_sort | Javier Sánchez García |
| collection | DOAJ |
| description | Vector autoregressions (VARs) and their multiple variants are standard models in economic and financial research due to their power for forecasting, data analysis and inference. These properties are a consequence of their capabilities to include multiple variables and lags which, however, turns into an exponential growth of the parameters to be estimated. This means that high-dimensional models with multiple variables and lags are difficult to estimate, leading to omitted variables, information biases and a loss of potential forecasting power. Traditionally, the existing literature has resorted to factor analysis, and specially, to Bayesian methods to overcome this situation. This paper explores the so-called machine learning regularization methods as an alternative to traditional methods of forecasting and impulse response analysis. We find that regularization structures, which allow for high dimensional models, perform better than standard Bayesian methods in nowcasting and forecasting. Moreover, impulse response analysis is robust and consistent with economic theory and evidence, and with the different regularization structures. Specifically, regarding the best regularization structure, an elementwise machine learning structure performs better in nowcasting and in computational efficiency, whilst a componentwise structure performs better in forecasting and cross-validation methods. |
| first_indexed | 2024-03-09T13:26:03Z |
| format | Article |
| id | doaj.art-238f90ca91fa4c90ac8bfe52347dc893 |
| institution | Directory Open Access Journal |
| issn | 2227-7390 |
| language | English |
| last_indexed | 2024-03-09T13:26:03Z |
| publishDate | 2022-03-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Mathematics |
| spelling | doaj.art-238f90ca91fa4c90ac8bfe52347dc8932023-11-30T21:23:26ZengMDPI AGMathematics2227-73902022-03-0110687710.3390/math10060877Machine Learning Regularization Methods in High-Dimensional Monetary and Financial VARsJavier Sánchez García0Salvador Cruz Rambaud1Mediterranean Research Center for Economics and Sustainable Development (CIMEDES), Universidad de Almería, 04120 Almería, SpainDepartamento de Economía y Empresa, Universidad de Almería, 04120 Almería, SpainVector autoregressions (VARs) and their multiple variants are standard models in economic and financial research due to their power for forecasting, data analysis and inference. These properties are a consequence of their capabilities to include multiple variables and lags which, however, turns into an exponential growth of the parameters to be estimated. This means that high-dimensional models with multiple variables and lags are difficult to estimate, leading to omitted variables, information biases and a loss of potential forecasting power. Traditionally, the existing literature has resorted to factor analysis, and specially, to Bayesian methods to overcome this situation. This paper explores the so-called machine learning regularization methods as an alternative to traditional methods of forecasting and impulse response analysis. We find that regularization structures, which allow for high dimensional models, perform better than standard Bayesian methods in nowcasting and forecasting. Moreover, impulse response analysis is robust and consistent with economic theory and evidence, and with the different regularization structures. Specifically, regarding the best regularization structure, an elementwise machine learning structure performs better in nowcasting and in computational efficiency, whilst a componentwise structure performs better in forecasting and cross-validation methods.https://www.mdpi.com/2227-7390/10/6/877VARmachine learningLASSO (Least Absolute Shrinkage and Selection Operator)regularization methodssparsitymonetary economics |
| spellingShingle | Javier Sánchez García Salvador Cruz Rambaud Machine Learning Regularization Methods in High-Dimensional Monetary and Financial VARs Mathematics VAR machine learning LASSO (Least Absolute Shrinkage and Selection Operator) regularization methods sparsity monetary economics |
| title | Machine Learning Regularization Methods in High-Dimensional Monetary and Financial VARs |
| title_full | Machine Learning Regularization Methods in High-Dimensional Monetary and Financial VARs |
| title_fullStr | Machine Learning Regularization Methods in High-Dimensional Monetary and Financial VARs |
| title_full_unstemmed | Machine Learning Regularization Methods in High-Dimensional Monetary and Financial VARs |
| title_short | Machine Learning Regularization Methods in High-Dimensional Monetary and Financial VARs |
| title_sort | machine learning regularization methods in high dimensional monetary and financial vars |
| topic | VAR machine learning LASSO (Least Absolute Shrinkage and Selection Operator) regularization methods sparsity monetary economics |
| url | https://www.mdpi.com/2227-7390/10/6/877 |
| work_keys_str_mv | AT javiersanchezgarcia machinelearningregularizationmethodsinhighdimensionalmonetaryandfinancialvars AT salvadorcruzrambaud machinelearningregularizationmethodsinhighdimensionalmonetaryandfinancialvars |