學術產出-Theses
Article View/Open
Publication Export
-
題名 以深度學習預測外匯超額報酬之研究
Deep Neural Network for Foreign Exchange Return Prediction作者 蔡玄中
Cai, Syuan-Jhong貢獻者 林建秀
Lin, Chien-Hsiu
蔡玄中
Cai, Syuan-Jhong關鍵詞 外匯預測
利差交易策略
動能交易策略
價值交易策略
深度學習
Exchange rate prediction
Carry trade
Momentum
Value
Deep learning日期 2023 上傳時間 2-Aug-2023 14:11:08 (UTC+8) 摘要 外匯走勢除市場供需影響走勢外，也深受該國政府政策，以及總體經濟的變化影響。因此本研究利用各國政策數據，例如利率決策、政府債務餘額佔GDP比例、經常帳餘額佔GDP比例、政府治理、外匯存底等五個因子作為各別因子，總體因子則以美國的數據為主，其中有VIX、TED利差、美國貨幣基數、AAA等級公司債殖利率、美國消費者物價指數、外國向美國購買證券、美國向外國購買證券等七個總體因子，另外也會納入過去文獻常探討的四個交易策略，市場投資組合、利差交易、動能交易、價值交易等四個傳統四因子。本研究將會把每個國家的各別因子與總體因子、傳統四因子進行交乘，將得到的60因子(5+(7+4)*5)分別放入DNN模型以及OLS模型進行匯率變化預測。接著再將兩個模型的預測結果加上各國利差得到的外匯超額報酬建構HML(High Minus Low)投資組合，並與市場投資組合、利差交易、動能交易、價值交易等四個交易策略績效進行比較。本研究最後會進行Permutation Importance，以判斷哪個因子對本研究的DNN模型重要性最高。實證結果發現，DNN模型的樣本外R^2顯著優於OLS模型，在各個交易策略中，DNN模型所建構的投資組合績效也是最好的。在因子重要性中，以政府債務餘額佔GDP比例、外匯存底、政府效能、外國向美國購買證券、美國向外國購買證券、美國消費者物價指數、美國貨幣基數、價值交易等幾個因子的重要度較高。上述的結果也顯示DNN模型在當下的經濟環境中，能夠有效應用因子所蘊含的資訊，這也是傳統線性模型所沒有的特性。
Foreign exchange trends are influenced not only by market supply and demand but also by government policies and macroeconomic changes in each country. In this study, DNN model uses various factors related to government policies as individual, and use US data as macroeconomic factors. Besides, this study also uses four strategies that has been discussed for long time, market portfolio, carry trade, momentum, and value. The individual factor will be multiplied and combined with macroeconomic factors and four strategy factor. Therefore, this study will use sixty (5+(7+4)*5) factors and separately input these factors into DNN and OLS models to predict the change of foreign exchange rate. Then, the predicted result will plus the interest rate differentials of each country, and get the predicted excess return. Next, use the predicted excess return to construct the High Minus Low portfolio. The performance of the HML portfolio will then be compared with the four trading strategies: market portfolio, carry trade, momentum, and value. Finally, the study will perform Permutation Importance to determine the most influential factor in the DNN model for this research.Empirical results reveal that the out-of-sample R^2 of the DNN model is higher than that of the OLS model. Furthermore, among the various strategies, the portfolio constructed by the DNN model exhibits the best performance. The factor importance analysis indicates that factors such as public debt, foreign exchange reserves, political stability, monetary base, CPI, gross purchases of domestic U.S security, gross purchases of foreign security from U.S, and value hold greater significance. These findings highlight the effectiveness of the DNN model in utilizing the information embedded in the factors within the current economic environment, a characteristic that traditional linear models lack.參考文獻 [1] Amat, C., Michalski, T., & Stoltz, G. (2018). Fundamentals and exchange rate forecastability with simple machine learning methods. Journal of International Money and Finance, 88, 1-24.[2] Ang, A., & Chen, J. (2010). Yield curve predictors of foreign exchange returns. AFA 2011 Denver Meetings Paper,[3] Asness, C. S., Moskowitz, T. J., & Pedersen, L. H. (2013). Value and momentum everywhere. The Journal of Finance, 68(3), 929-985.[4] Avramov, D., Cheng, S., & Metzker, L. (2023). Machine learning vs. economic restrictions: Evidence from stock return predictability. Management Science, 69(5), 2587-2619.[5] Batista, G. E., & Monard, M. C. (2002). A study of K-nearest neighbour as an imputation method. His, 87(251-260), 48.[6] Brunnermeier, M. K., Nagel, S., & Pedersen, L. H. (2008). Carry trades and currency crashes. NBER macroeconomics annual, 23(1), 313-348.[7] Burnside, C., Eichenbaum, M., & Rebelo, S. (2011). Carry trade and momentum in currency markets. Annu. Rev. Financ. Econ., 3(1), 511-535.[8] Chaboud, A. P., & Wright, J. H. (2005). Uncovered interest parity: it works, but not for long. Journal of International Economics, 66(2), 349-362.[9] Cover, T., & Hart, P. (1967). Nearest neighbor pattern classification. IEEE TRANSACTIONS ON INFORMATION THEORY, 13(1), 21-27.[10] Della Corte, P., Sarno, L., & Tsiakas, I. (2009). An economic evaluation of empirical exchange rate models. The Review of Financial Studies, 22(9), 3491-3530.[11] Eichenbaum, M. S., Johannsen, B. K., & Rebelo, S. T. (2021). Monetary policy and the predictability of nominal exchange rates. The Review of Economic Studies, 88(1), 192-228.[12] Engel, C., & West, K. D. (2005). Exchange rates and fundamentals. Journal of Political Economy, 113(3), 485-517.[13] Feng, G., He, J., & Polson, N. G. (2018). Deep learning for predicting asset returns. arXiv preprint arXiv:1804.09314.[14] Filippou, I., Rapach, D., Taylor, M. P., & Zhou, G. (2022). Exchange rate prediction with machine learning and a smart carry portfolio. Available at SSRN 3455713.[15] Gu, S., Kelly, B., & Xiu, D. (2020). Empirical asset pricing via machine learning. The Review of Financial Studies, 33(5), 2223-2273.[16] Hoerl, A. E., & Kennard, R. W. (1970). Ridge regression: Biased estimation for nonorthogonal problems. Technometrics, 12(1), 55-67.[17] Hu, Z., Zhao, Y., & Khushi, M. (2021). A survey of forex and stock price prediction using deep learning. Applied System Innovation, 4(1), 9.[18] Jegadeesh, N., & Titman, S. (1993). Returns to buying winners and selling losers: Implications for stock market efficiency. The Journal of Finance, 48(1), 65-91.[19] Khalaf, G. (2012). A proposed ridge parameter to improve the least square estimator. Journal of Modern Applied Statistical Methods, 11(2), 15.[20] Kroencke, T. A., Schindler, F., & Schrimpf, A. (2014). International diversification benefits with foreign exchange investment styles. Review of Finance, 18(5), 1847-1883.[21] Lakonishok, J., Shleifer, A., & Vishny, R. W. (1994). Contrarian investment, extrapolation, and risk. The Journal of Finance, 49(5), 1541-1578.[22] Meese, R. A., & Rogoff, K. (1983). Empirical exchange rate models of the seventies: Do they fit out of sample? Journal of International Economics, 14(1-2), 3-24.[23] Menkhoff, L., Sarno, L., Schmeling, M., & Schrimpf, A. (2012a). Carry trades and global foreign exchange volatility. The Journal of Finance, 67(2), 681-718.[24] Menkhoff, L., Sarno, L., Schmeling, M., & Schrimpf, A. (2012b). Currency momentum strategies. Journal of Financial Economics, 106(3), 660-684.[25] Menkhoff, L., Sarno, L., Schmeling, M., & Schrimpf, A. (2017). Currency value. The Review of Financial Studies, 30(2), 416-441.[26] Molodtsova, T., & Papell, D. H. (2013). Taylor rule exchange rate forecasting during the financial crisis. NBER International Seminar on Macroeconomics,[27] Muthukrishnan, R., & Rohini, R. (2016). LASSO: A feature selection technique in predictive modeling for machine learning. 2016 IEEE international conference on advances in computer applications (ICACA),[28] Okunev, J., & White, D. (2003). Do momentum-based strategies still work in foreign currency markets? Journal of Financial and Quantitative Analysis, 38(2), 425-447.[29] Raza, A. (2015). Are Value Strategies Profitable in the Foreign Exchange Market? 28th Australasian Finance and Banking Conference,[30] Rossi, B. (2013). Exchange rate predictability. Journal of Economic Literature, 51(4), 1063-1119.[31] Taylor, J. B. (1993). Discretion versus policy rules in practice. Carnegie-Rochester Conference Series on Public Policy,[32] Tibshirani, R. (1996). Regression shrinkage and selection via the lasso. Journal of the Royal Statistical Society: Series B (Methodological), 58(1), 267-288. 描述 碩士
國立政治大學
金融學系
110352020資料來源 http://thesis.lib.nccu.edu.tw/record/#G0110352020 資料類型 thesis dc.contributor.advisor 林建秀 zh_TW dc.contributor.advisor Lin, Chien-Hsiu en_US dc.contributor.author (Authors) 蔡玄中 zh_TW dc.contributor.author (Authors) Cai, Syuan-Jhong en_US dc.creator (作者) 蔡玄中 zh_TW dc.creator (作者) Cai, Syuan-Jhong en_US dc.date (日期) 2023 en_US dc.date.accessioned 2-Aug-2023 14:11:08 (UTC+8) - dc.date.available 2-Aug-2023 14:11:08 (UTC+8) - dc.date.issued (上傳時間) 2-Aug-2023 14:11:08 (UTC+8) - dc.identifier (Other Identifiers) G0110352020 en_US dc.identifier.uri (URI) http://nccur.lib.nccu.edu.tw/handle/140.119/146599 - dc.description (描述) 碩士 zh_TW dc.description (描述) 國立政治大學 zh_TW dc.description (描述) 金融學系 zh_TW dc.description (描述) 110352020 zh_TW dc.description.abstract (摘要) 外匯走勢除市場供需影響走勢外，也深受該國政府政策，以及總體經濟的變化影響。因此本研究利用各國政策數據，例如利率決策、政府債務餘額佔GDP比例、經常帳餘額佔GDP比例、政府治理、外匯存底等五個因子作為各別因子，總體因子則以美國的數據為主，其中有VIX、TED利差、美國貨幣基數、AAA等級公司債殖利率、美國消費者物價指數、外國向美國購買證券、美國向外國購買證券等七個總體因子，另外也會納入過去文獻常探討的四個交易策略，市場投資組合、利差交易、動能交易、價值交易等四個傳統四因子。本研究將會把每個國家的各別因子與總體因子、傳統四因子進行交乘，將得到的60因子(5+(7+4)*5)分別放入DNN模型以及OLS模型進行匯率變化預測。接著再將兩個模型的預測結果加上各國利差得到的外匯超額報酬建構HML(High Minus Low)投資組合，並與市場投資組合、利差交易、動能交易、價值交易等四個交易策略績效進行比較。本研究最後會進行Permutation Importance，以判斷哪個因子對本研究的DNN模型重要性最高。實證結果發現，DNN模型的樣本外R^2顯著優於OLS模型，在各個交易策略中，DNN模型所建構的投資組合績效也是最好的。在因子重要性中，以政府債務餘額佔GDP比例、外匯存底、政府效能、外國向美國購買證券、美國向外國購買證券、美國消費者物價指數、美國貨幣基數、價值交易等幾個因子的重要度較高。上述的結果也顯示DNN模型在當下的經濟環境中，能夠有效應用因子所蘊含的資訊，這也是傳統線性模型所沒有的特性。 zh_TW dc.description.abstract (摘要) Foreign exchange trends are influenced not only by market supply and demand but also by government policies and macroeconomic changes in each country. In this study, DNN model uses various factors related to government policies as individual, and use US data as macroeconomic factors. Besides, this study also uses four strategies that has been discussed for long time, market portfolio, carry trade, momentum, and value. The individual factor will be multiplied and combined with macroeconomic factors and four strategy factor. Therefore, this study will use sixty (5+(7+4)*5) factors and separately input these factors into DNN and OLS models to predict the change of foreign exchange rate. Then, the predicted result will plus the interest rate differentials of each country, and get the predicted excess return. Next, use the predicted excess return to construct the High Minus Low portfolio. The performance of the HML portfolio will then be compared with the four trading strategies: market portfolio, carry trade, momentum, and value. Finally, the study will perform Permutation Importance to determine the most influential factor in the DNN model for this research.Empirical results reveal that the out-of-sample R^2 of the DNN model is higher than that of the OLS model. Furthermore, among the various strategies, the portfolio constructed by the DNN model exhibits the best performance. The factor importance analysis indicates that factors such as public debt, foreign exchange reserves, political stability, monetary base, CPI, gross purchases of domestic U.S security, gross purchases of foreign security from U.S, and value hold greater significance. These findings highlight the effectiveness of the DNN model in utilizing the information embedded in the factors within the current economic environment, a characteristic that traditional linear models lack. en_US dc.description.tableofcontents 第一章 緒論 1第一節 研究背景與動機 1第二節 研究目的 2第三節 研究架構 3第二章 文獻回顧 3第一節 外匯預測 3第二節 利差交易 4第三節 價值交易 5第四節 動能交易 6第五節 K-Nearest Neighbor演算法 6第六節 機器學習模型 7第三章 資料 8第一節 資料來源 8第二節 因子建構 10第四章 研究方法 16第一節 研究流程 16第二節 研究方法 17第五章 實證結果與分析 23第一節 DNN與線性迴歸之預測能力比較 23第二節 策略投資組合績效比較 29第三節 因子重要度 31第六章 結論與建議 35第一節 結論 35第二節 未來建議 35參考文獻 37 zh_TW dc.format.extent 2264938 bytes - dc.format.mimetype application/pdf - dc.source.uri (資料來源) http://thesis.lib.nccu.edu.tw/record/#G0110352020 en_US dc.subject (關鍵詞) 外匯預測 zh_TW dc.subject (關鍵詞) 利差交易策略 zh_TW dc.subject (關鍵詞) 動能交易策略 zh_TW dc.subject (關鍵詞) 價值交易策略 zh_TW dc.subject (關鍵詞) 深度學習 zh_TW dc.subject (關鍵詞) Exchange rate prediction en_US dc.subject (關鍵詞) Carry trade en_US dc.subject (關鍵詞) Momentum en_US dc.subject (關鍵詞) Value en_US dc.subject (關鍵詞) Deep learning en_US dc.title (題名) 以深度學習預測外匯超額報酬之研究 zh_TW dc.title (題名) Deep Neural Network for Foreign Exchange Return Prediction en_US dc.type (資料類型) thesis en_US dc.relation.reference (參考文獻) [1] Amat, C., Michalski, T., & Stoltz, G. (2018). Fundamentals and exchange rate forecastability with simple machine learning methods. Journal of International Money and Finance, 88, 1-24.[2] Ang, A., & Chen, J. (2010). Yield curve predictors of foreign exchange returns. AFA 2011 Denver Meetings Paper,[3] Asness, C. S., Moskowitz, T. J., & Pedersen, L. H. (2013). Value and momentum everywhere. The Journal of Finance, 68(3), 929-985.[4] Avramov, D., Cheng, S., & Metzker, L. (2023). Machine learning vs. economic restrictions: Evidence from stock return predictability. Management Science, 69(5), 2587-2619.[5] Batista, G. E., & Monard, M. C. (2002). A study of K-nearest neighbour as an imputation method. His, 87(251-260), 48.[6] Brunnermeier, M. K., Nagel, S., & Pedersen, L. H. (2008). Carry trades and currency crashes. NBER macroeconomics annual, 23(1), 313-348.[7] Burnside, C., Eichenbaum, M., & Rebelo, S. (2011). Carry trade and momentum in currency markets. Annu. Rev. Financ. Econ., 3(1), 511-535.[8] Chaboud, A. P., & Wright, J. H. (2005). Uncovered interest parity: it works, but not for long. Journal of International Economics, 66(2), 349-362.[9] Cover, T., & Hart, P. (1967). Nearest neighbor pattern classification. IEEE TRANSACTIONS ON INFORMATION THEORY, 13(1), 21-27.[10] Della Corte, P., Sarno, L., & Tsiakas, I. (2009). An economic evaluation of empirical exchange rate models. The Review of Financial Studies, 22(9), 3491-3530.[11] Eichenbaum, M. S., Johannsen, B. K., & Rebelo, S. T. (2021). Monetary policy and the predictability of nominal exchange rates. The Review of Economic Studies, 88(1), 192-228.[12] Engel, C., & West, K. D. (2005). Exchange rates and fundamentals. Journal of Political Economy, 113(3), 485-517.[13] Feng, G., He, J., & Polson, N. G. (2018). Deep learning for predicting asset returns. arXiv preprint arXiv:1804.09314.[14] Filippou, I., Rapach, D., Taylor, M. P., & Zhou, G. (2022). Exchange rate prediction with machine learning and a smart carry portfolio. Available at SSRN 3455713.[15] Gu, S., Kelly, B., & Xiu, D. (2020). Empirical asset pricing via machine learning. The Review of Financial Studies, 33(5), 2223-2273.[16] Hoerl, A. E., & Kennard, R. W. (1970). Ridge regression: Biased estimation for nonorthogonal problems. Technometrics, 12(1), 55-67.[17] Hu, Z., Zhao, Y., & Khushi, M. (2021). A survey of forex and stock price prediction using deep learning. Applied System Innovation, 4(1), 9.[18] Jegadeesh, N., & Titman, S. (1993). Returns to buying winners and selling losers: Implications for stock market efficiency. The Journal of Finance, 48(1), 65-91.[19] Khalaf, G. (2012). A proposed ridge parameter to improve the least square estimator. Journal of Modern Applied Statistical Methods, 11(2), 15.[20] Kroencke, T. A., Schindler, F., & Schrimpf, A. (2014). International diversification benefits with foreign exchange investment styles. Review of Finance, 18(5), 1847-1883.[21] Lakonishok, J., Shleifer, A., & Vishny, R. W. (1994). Contrarian investment, extrapolation, and risk. The Journal of Finance, 49(5), 1541-1578.[22] Meese, R. A., & Rogoff, K. (1983). Empirical exchange rate models of the seventies: Do they fit out of sample? Journal of International Economics, 14(1-2), 3-24.[23] Menkhoff, L., Sarno, L., Schmeling, M., & Schrimpf, A. (2012a). Carry trades and global foreign exchange volatility. The Journal of Finance, 67(2), 681-718.[24] Menkhoff, L., Sarno, L., Schmeling, M., & Schrimpf, A. (2012b). Currency momentum strategies. Journal of Financial Economics, 106(3), 660-684.[25] Menkhoff, L., Sarno, L., Schmeling, M., & Schrimpf, A. (2017). Currency value. The Review of Financial Studies, 30(2), 416-441.[26] Molodtsova, T., & Papell, D. H. (2013). Taylor rule exchange rate forecasting during the financial crisis. NBER International Seminar on Macroeconomics,[27] Muthukrishnan, R., & Rohini, R. (2016). LASSO: A feature selection technique in predictive modeling for machine learning. 2016 IEEE international conference on advances in computer applications (ICACA),[28] Okunev, J., & White, D. (2003). Do momentum-based strategies still work in foreign currency markets? Journal of Financial and Quantitative Analysis, 38(2), 425-447.[29] Raza, A. (2015). Are Value Strategies Profitable in the Foreign Exchange Market? 28th Australasian Finance and Banking Conference,[30] Rossi, B. (2013). Exchange rate predictability. Journal of Economic Literature, 51(4), 1063-1119.[31] Taylor, J. B. (1993). Discretion versus policy rules in practice. Carnegie-Rochester Conference Series on Public Policy,[32] Tibshirani, R. (1996). Regression shrinkage and selection via the lasso. Journal of the Royal Statistical Society: Series B (Methodological), 58(1), 267-288. zh_TW
