东吴大学经济系.PDF

鏉 鍚 澶 瀛 缍 婵 绯 纰╁＋璜栨枃 鎸囧皫鏁欐巿锛 Christos Michalopoulos 鍗氬＋ 澶氬厓鍒嗛厤涔嬪缓妲婜GARCH-Copula 涔嬫噳鐢 Construction of Multivariate Distributions : GARCH-Copula Application 鐮旂┒鐢燂細琚佸唬鏄 鎾 涓 鑿皯 鍦 105 骞 10 鏈 鎽樿 鏈枃鐐哄缓绔嬭偂绁ㄦ寚鏁稿牨閰 鐜囦箣闁 涔嬮棞鑱祼妲嬶紝閬哥敤妯欑殑 鐐哄叾鍚堢ū鐐烘柊 閲戠鍥涘湅鐨 鍙扮仯 銆佷腑鍦嬨€佸嵃搴︺€ 鍗楅煋 銆傚皪鏂兼柊 閲戠 鍥涘湅鑲＄エ鎸囨暩鍫遍叕 鐜囦箣閭婇殯 鍒嗛厤锛 鏈枃浣 鐢 ARMA-GARCH妯″瀷閫茶妯℃摤锛屽啀閬 浠ラ仼鐣朵箣 Copula 鍑芥暩灏囧叾閫ｇ祼鐐鸿伅鍚 鍒 閰嶃€傛渶寰 鏈枃鍒╃敤寤烘涔嬪鍏冨垎閰嶆ā鎿 鍑哄皪鎳変箣妯ｆ湰鏁告鏌 鍏跺鍏冨垎閰嶄箣閰 閬╂€э紝 鐢 浠ヨ鏄 灏嶆柤璩 鐢㈠牨閰 鐜囷紝GARCH-Copula妯″瀷寤烘鑱悎 鍒嗛厤 涔嬭兘 鍔涖€ Abstract In this paper, we construct the dependence structure between stock markets of four countries in Asia, they also be called TICKs. The first step consists of the application of the Autoregressive Moving Average and the Generalize Autoregressive Conditional Heteroscedastic process, then we extracting the standardized residuals for each return series through the ARMA-GARCH model which is a useful method to fit margins in finance. In the second step, the work of liking the fitted margins is completed by the copula function. Finally, we generate some simulations from the multivariate distribution we constructed and examine whether the GARCH-copula is an appropriate combination model in finance. Contents 1 Introduction 鈭 1 2 Modeling Volatility 3 2.1 Stylized facts for financial returns 鈭欌垯鈭欌垯鈭欌垯鈭欌垯鈭欌垯鈭欌垯鈭欌垯鈭欌垯 3 2.2 ARCH process 鈭欌垯鈭欌垯鈭欌垯鈭欌垯鈭欌垯鈭欌垯鈭欌垯鈭欌垯鈭欌垯鈭欌垯鈭欌垯鈭欌垯鈭欌垯鈭欌垯鈭 4 2.2.1 Stationarity of ARCH process 鈭欌垯鈭欌垯鈭欌垯鈭欌垯鈭欌垯鈭欌垯鈭欌垯 6 2.3 GARCH process and GJR extention鈭欌垯鈭欌垯鈭欌垯鈭欌垯鈭欌垯鈭欌垯鈭欌垯鈭 8 2.3.1 Stationa