本文探討高維總體經濟資料下之預測問題，並提出一套兩階段方法，以處理變數選擇問題與模型不確定性。我們提出一套結合單一共變數逐一多重檢定（One Covariate at a Time Multiple Testing, OCMT）與 BMA 的兩階段預測方法。此方法之第一階段透過多重邊際顯著性檢定，有效縮減因共線性與維度災難所衍生之龐大模型空間；第二階段則運用 BMA 演算法，於縮減後之模型空間中計算後驗機率並進行預測加權。

本文將此兩階段架構應用於美國 109 個季頻總體經濟指標，進行 GDP 成長率與 CPI 通膨變動率之樣本外預測。實證結果顯示，結合 OCMT 預先篩選能有效克服維度詛咒，大幅改善純 BMA 的預測表現。本架構的核心貢獻在於，在維持預測精準度的同時，能直接針對原始變數計算後驗包含機率（PIP），突破了傳統因子模型缺乏經濟意涵的限制。透過 PIP 之分析，我們發現 GDP 成長率的預測動能主要來自房市實體指標與服務業消費；而 CPI 通膨變動率的表現則高度反映通膨變動率的自我迴歸慣性。最後，先驗穩健性檢定亦支持上述實證發現不受特定參數設定所主導。

This thesis proposes a two-stage framework combining One Covariate at a Time Multiple Testing (OCMT) and Bayesian Model Averaging (BMA) to address variable selection and model uncertainty in high-dimensional macroeconomic forecasting. In the first stage, OCMT pre-screens potential predictors via marginal significance testing to mitigate the curse of dimensionality. Subsequently, BMA constructs weighted forecasts within the reduced model space by accounting for the joint distribution of candidate variables.Applying this approach to 109 U.S. macroeconomic indicators for out-of-sample forecasting of real GDP growth and changes in CPI inflation, the OCMT-BMA framework offers enhanced variable interpretability while maintaining competitive predictive accuracy compared to Lasso and factor-augmented autoregressive (FA-AR) models. Specifically, by computing the posterior inclusion probabilities (PIP) for the original variables, this framework overcomes the interpretability limitations of standard factor models, which often struggle to map extracted factors back to specific economic indicators.The PIP analysis further clarifies the relative importance of the selected variables: real GDP growth forecasts are primarily driven by real residential investment and service consumption, whereas predictions of changes in CPI inflation rely heavily on their autoregressive persistence. These empirical findings carry clear economic intuition and remain robust across different prior specifications. Overall, the OCMT-BMA framework provides a structured approach to evaluate the relative importance of individual economic variables under model uncertainty, serving as a valuable complement to existing high-dimensional predictive methods.

[1] Alexandre Belloni, Victor Chernozhukov, and Christian Hansen. High-dimensional methods and inference on structural and treatment effects. Journal of Economic Perspectives, 28(2):29–50, 2014.
[2] José M. Bernardo and Adrian F. M. Smith. Bayesian Theory. John Wiley & Sons, Chichester, 1994.
[3] Patrick Breheny and Jian Huang. Coordinate descent algorithms for nonconvex penalized regression, with applications to biological feature selection. The Annals of Applied Statistics, 5(1):232–253, 2011.
[4] Alexander Chudik, George Kapetanios, and M. Hashem Pesaran. A one covariate at a time, multiple testing approach to variable selection in high-dimensional linear regression models. Econometrica, 86(4):1479–1512, 2018.
[5] David Draper. Assessment and propagation of model uncertainty. Journal of the Royal Statistical Society: Series B (Methodological), 57(1):45–70, 1995.
[6] Carmen Fernández, Eduardo Ley, and Mark F. J. Steel. Benchmark priors for bayesian model averaging. Journal of Econometrics, 100(2):381–427, 2001.
[7] Peter L. Flom and David L. Cassell. Stopping stepwise: Why stepwise and similar selection methods are bad, and what you should use. In NorthEast SAS Users Group Inc 20th Annual Conference, pages 1–5, 2007.
[8] Frank E. Harrell. Regression Modeling Strategies: With Applications to Linear Models, Logistic Regression, and Survival Analysis. Springer, 2001.
[9] James S. Hodges. Uncertainty, policy analysis and statistics. Statistical Science, 2(3):259–275, 1987.
[10] Jian Huang, Shuangge Ma, and Cun-Hui Zhang. Adaptive lasso for sparse high-dimensional regression models. Statistica Sinica, 18(4):1603–1618, 2008.
[11] Robert E. Kass and Larry Wasserman. A reference bayesian test for nested hypotheses and its relationship to the schwarz criterion. Journal of the American Statistical Association, 90(431):928–934, 1995.
[12] Gary Koop and Simon M. Potter. Forecasting in dynamic factor models using bayesian model averaging. The Econometrics Journal, 7(2):550–565, 2004.
[13] Edward E. Leamer. Housing is the business cycle. In Proceedings - Economic Policy Symposium - Jackson Hole, pages 149–233. Federal Reserve Bank of Kansas City, 2007.
[14] Eduardo Ley and Mark F. J. Steel. On the effect of prior assumptions in bayesian model averaging with applications to growth regression. Journal of Applied Econometrics, 24(4):651–674, 2009.
[15] Feng Liang, Rui Paulo, German Molina, Merlise A. Clyde, and Jim O. Berger. Mixtures of g priors for bayesian variable selection. Journal of the American Statistical Association, 103(481):110–123, 2008.
[16] Serena Ng and Jonathan H. Wright. Facts and challenges from the great recession for forecasting and macroeconomic modeling. Journal of Economic Literature, 51(4):1120–1154, 2013.
[17] Mark F. J. Steel. Model averaging and its use in economics. Journal of Economic Literature, 58(3):644–719, 2020.
[18] James H. Stock and Mark W. Watson. Forecasting inflation. Journal of Monetary Economics, 44(2):293–335, 1999.
[19] James H. Stock and Mark W. Watson. Macroeconomic forecasting using diffusion indexes. Journal of Business & Economic Statistics, 20(2):147–162, 2002.
[20] James H Stock and Mark W Watson. Has the business cycle changed and why? NBER Macroeconomics Annual, 17:159–218, 2003.
[21] James H. Stock and Mark W. Watson. Why has u.s. inflation become harder to forecast? Journal of Money, Credit and Banking, 39(s1):3–33, 2007.
[22] James H. Stock and Mark W. Watson. Generalized shrinkage methods for forecasting using many predictors. Journal of Business & Economic Statistics, 30(4):481–493, 2012.
[23] Robert Tibshirani. Regression shrinkage and selection via the lasso. Journal of the Royal Statistical Society: Series B (Methodological), 58(1):267–288, 1996.
[24] Jonathan H. Wright. Forecasting US inflation by bayesian model averaging. Journal of Forecasting, 28(2):131–144, 2009.
[25] Arnold Zellner. On assessing prior distributions and bayesian regression analysis with g-prior distributions. In Bayesian Inference and Decision Techniques: Essays in Honor of Bruno de Finetti, pages 233–243. Elsevier, 1986.
[26] Peng Zhao and Bin Yu. On model selection consistency of lasso. The Journal of Machine Learning Research, 7:2541–2563, 2006.
[27] Hui Zou. The adaptive lasso and its oracle properties. Journal of the American Statistical Association, 101(476):1418–1429, 2006.
[28] Hui Zou and Trevor Hastie. Regularization and variable selection via the elastic net. Journal of the Royal Statistical Society: Series B (Statistical Methodology), 67(2):301–320, 2005.