本研究旨在建立台灣的通膨風險評估架構，透過結合機器學習與分量迴歸方法，分析未來不同期別下消費者物價指數年增率的極端變動風險。本研究以2000 年1月至2024年10月之台灣總體經濟與金融變數月資料為樣本，使用LASSO、Elastic Net、Adaptive LASSO、PCA、隨機森林、XGBoost、LSTM 七種特徵選取法，從大量變數中篩選重要因子，並透過分量迴歸估計各分位數下的通膨風險，結合t-skew分布進行尾部風險調整，建立ES、EL兩指標來選擇模型，以其篩選出來的變數做各分位係數分析，最後以Q95與Q5的差距為衡量，探討重大政策調整是否有助於穩定風險。

實證結果顯示，不同模型在上行與下行風險的預測上具異質性，Elastic Net在短期EL評估上表現較優，PCA與LSTM在中長期EL評估上均表現優異。而不同變數對通膨風險的影響在各分位間呈現分位數不對稱性，部分變數甚至在低分位與高分位方向相反，顯示傳統線性預測方法或許低估極端價格變動風險。本研究建議政策制定者應採用多模型架構與分量風險評估方法，不同觀察期採以不同模型分析，以強化通膨風險管理之全面性。

This study aims to establish an inflation risk assessment framework for Taiwan by integrating machine learning techniques with quantile regression to analyze the extreme fluctuation risks of the Consumer Price Index (CPI) annual growth rate across different forecast horizons. Using macroeconomic and financial data from Taiwan spanning January 2000 to October 2024, the study applies seven feature selection methods including LASSO, Elastic Net, Adaptive LASSO, PCA, Random Forest, XGBoost, and LSTM to extract key predictors from a high-dimensional dataset. These selected features are then incorporated into quantile regression models to estimate Inflation-at-Risk (IaR) at various quantiles. To better capture tail risk dynamics, the t-skew distribution is employed for smoothing, enabling the computation of Expected Shortfall (ES) and Expected Longrise (EL). Models with superior performance are identified, and the corresponding features are visualized using quantile coefficient plots. Finally, the study uses the gap between Q95 and Q5 as a measure to evaluate whether major policy adjustments help stabilize inflation risk.

Empirical results reveal heterogeneity in model performance regarding upward and downward risk prediction. Elastic Net performs better in short-term EL estimation, while PCA and LSTM outperform in medium to long-term EL forecasting. The impact of individual variables on inflation risk demonstrates nonlinearity and quantile-specific asymmetry, with some variables exhibiting opposite effects across low and high quantiles. These findings suggest that traditional linear forecasting models may underestimate extreme inflation risks. The study recommends that policymakers adopt a multi-model architecture and quantile-based risk assessment approach, applying different models across different forecast horizons to enhance the accuracy and robustness of inflation risk management.

[1] Adams, P. A., Adrian, T., Boyarchenko, N., & Giannone, D. (2020). Forecasting macroeconomic risks (Staff Report No. 914). Federal Reserve Bank of New York.
[2] Adrian, T., Boyarchenko, N., & Giannone, D. (2019). Vulnerable growth. American Economic Review, 109(4), 1263-1289.
[3] Adrian, T., Boyarchenko, N., & Giannone, D. (2022). Financial conditions and their asymmetric effects on economic activity. Journal of Monetary Economics, 125, 1-17.
[4] Almosova, A., & Andresen, N. (2019). Nonlinear inflation forecasting with recurrent neural networks. Working paper, Humboldt University of Berlin and Technical University of Berlin.
[5] Araujo, G. S., & Gaglianone, W. P. (2023). Machine learning methods for inflation forecasting in Brazil: New contenders versus classical models. Latin American Journal of Central Banking, 4, 100087.
[6] Azzalini, A., & Capitanio, A. (2003). Distributions generated by perturbation of symmetry with emphasis on a multivariate skew t-distribution. Journal of the Royal Statistical Society: Series B (Statistical Methodology), 65(2), 367-389.
[7] Ball, L., & Mankiw, N. G. (1995). Relative-price changes as aggregate supply shocks. The Quarterly Journal of Economics, 110(1), 161–193.
[8] Berger, D., & Vavra, J. (2019). Shocks vs. responsiveness: What drives time-varying dispersion? Journal of Political Economy, 127(5), 2104–2142.
[9] Blanchard, O., & Galí, J. (2007). Real wage rigidities and the New Keynesian model. Journal of Money, Credit and Banking, 39(s1), 35–65.
[10] Breiman, L. (2001). Random forests. Machine Learning, 45(1), 5-32.
[11] Busch, M. O., Sánchez-Martínez, J. M., Rodríguez-Martínez, A., Montañez-Enríquez, R., & Martínez-Jaramillo, S. (2022). Growth at risk: Methodology and applications in an open-source platform. Latin American Journal of Central Banking, 3, 100068.
[12] Chen, T., & Guestrin, C. (2016). XGBoost: A scalable tree boosting system. Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (KDD’16), 785-794.
[13] Friedman, M. (1970). The counter-revolution in monetary theory. Institute of Economic Affairs.
[14] Fuhrer, J. C., & Moore, G. R. (1995). Inflation persistence. Quarterly Journal of Economics, 110(1), 127–159.
[15] Giannone, D., Lenza, M., & Primiceri, G. E. (2022). Economic predictions with big data: The illusion of sparsity. Econometrica, 90(1), 315-345.
[16] Goldani, M. (n.d.). Evaluating feature selection methods for macro-economic forecasting, applied for Iran’s macro-economic variables. Unpublished manuscript, Hakim Sabzevari University.
[17] Hamilton, J. D. (2012). Import prices and inflation. International Journal of Central Banking, 8(1), 271–279.
[18] Hochreiter, S., & Schmidhuber, J. (1997). Long short-term memory. Neural Computation, 9(8), 1735–1780.
[19] Hooker, M. A. (2002). Are oil shocks inflationary? Asymmetric and nonlinear specifications versus changes in regime. Journal of Money, Credit and Banking, 34(2), 540–561.
[20] Iseringhausen, M. (2021). A time-varying skewness model for growth-at-risk (Working Paper Series No. 49). European Stability Mechanism.
[21] Koenker, R., & Bassett, G. (1978). Regression quantiles. Econometrica, 46(1), 33-50.
[22] Makabe, Y., & Norimasa, Y. (2022). The term structure of inflation at risk: A panel quantile regression approach. Bank of Japan Working Paper Series, No. 22-E-4.
[23] Martin, L. (2019). Machine Learning vs Traditional Forecasting Methods: An Application to South African GDP. Stellenbosch Economic Working Papers WP12/2019.
[24] Medeiros, M. C., Vasconcelos, G. F., Veiga, A., & Zilberman, E. (2021). Forecasting inflation in a data-rich environment: The benefits of machine learning methods. Journal of Business & Economic Statistics, 39(1), 98-119.
[25] Naghi, A. A., O'Neill, E., & Danielova Zaharieva, M. (2024). The benefits of forecasting inflation with machine learning: New evidence. Journal of Applied Econometrics, 39(7), 1321–1331.
[26] Narmandakh, B. (2021). Inflation forecasting with machine learning methods: A case of Mongolia. Monetary Policy Department, Bank of Mongolia.
[27] Nguyen, T. T. T., Pham, S. D., Li, X.-M., & Do, H. X. (2024). Does the U.S. export inflation? Evidence from the dynamic inflation spillover between the U.S. and EAGLEs. International Review of Economics and Finance, 94, 103427.
[28] Paranhos, J. (2023). Real-time inflation forecasting using nonlinear dimension reduction techniques. International Journal of Forecasting, 39(2), 901–921.
[29] López-Salido, D., & Loria, F. (2020). Inflation at risk. Finance and Economics Discussion Series, 2020-013. Board of Governors of the Federal Reserve System.
[30] Stock, J. H., & Watson, M. W. (2002a). Macroeconomic forecasting using diffusion indexes. Journal of Business & Economic Statistics, 20(2), 147-162.
[31] Stock, J. H., & Watson, M. W. (2002b). Forecasting using principal components from a large number of predictors. Journal of the American Statistical Association, 97(460), 1167-1179.
[32] Theoharidis, C., Guillen, R., & Lopes, S. (2023). Predicting Inflation with Recurrent Neural Networks. Journal of Economic Forecasting, 42(1), 337–405.
[33] Tibshirani, R. (1996). Regression shrinkage and selection via the lasso. Journal of the Royal Statistical Society: Series B (Methodological), 58(1), 267-288.
[34] Ho, T. K. (1995). Random decision forests. Proceedings of the 3rd International Conference on Document Analysis and Recognition (ICDAR’95), 278-282.
[35] Yang, Y., Xu, X., Ge, J., & Xu, Y. (2024). Machine learning for economic forecasting: An application to China’s GDP growth. arXiv preprint arXiv:2407.03595.
[36] Zou, H. (2006). The adaptive lasso and its oracle properties. Journal of the American Statistical Association, 101(476), 1418-1429.
[37] Zou, H., & Hastie, T. (2005). Regularization and variable selection via the elastic net. Journal of the Royal Statistical Society: Series B (Statistical Methodology), 67(2), 301-320.
[38] 林馨怡、顏佑銘、葉錦徽，2024，「台灣通膨率預測：運用大數據資料分析」，中央銀行委外研究。
[39] 陳博志，2024，「風險機率很小就可以忽略嗎？」，台灣經濟研究月刊。