標題: 資產配置全球化之投資策略實證研究
An Empirical Study on Global Asset Allocation Strategies
作者: 郭洪生
Hung-sheng Kuo
陳安斌
An-Pin Chen
資訊管理研究所
關鍵字: 全球資產配置;均異模式;固定組合策略;固定比例策略;時間不變性投資組合保護策略;Global Asset Allocation;MV Model;Constant-Mix;CPPI;TIPP
公開日期: 1999
摘要: 摘要 在所有投資決策中,都少不了資產的選擇與權重配置,對日後的績效表現影響極為顯著。本研究的目的就在於提出一資產配置全球化的模式,期望能達到較低風險與較高報酬。 目前的投資工具往往侷限於國內市場,容易受到景氣等總體經濟因素影響而較難規避風險,許多研究指出將海外資產納入資產配置中,即能建立較好的效率前緣,預期可以有較好的表現。因此本研究以Markowitz(1952)所提出的均異(Mean-Variance)模式出發點建立一台港日區域型基金,並以Benchmark、Equal、Optimized三種權重配合B&H、Constant-Mix、CPPI、TIPP四種策略,實證全球資產配置最佳化及動態調整操作策略之有效性。 由本研究的實證,可得以下的結論。權重設定方面,標竿或市值權重是被動式資產配置的最佳參考,但是要考慮權重資料取得成本較高;等權重容易操作,能規避系統化風險,但是表現績效較最佳化權重差;最佳化權重往往能有最好的績效,但是在進行最佳化時必須注意樣本資料的選取及各項參數設定。 動態策略方面:買入持有的交易成本、操作成本低且容易,效果也不錯,但是很難有最好的表現;Constant-Mix在市場橫向整理時表現較好,但全球資產配置波動性較小,所以並不適用;CPPI能提供市場下跌時的保護,但是風險性較高;TIPP原理近似CPPI,但是在市場上揚時報酬及風險都較CPPI小。
Abstract Asset allocation is essential in all investment decisions and has great influence on the performance of portfolios, and according to many previous studies, global asset allocation (GAA) is more effective in risk diversification since local investment is too dependent on native macro situations. Therefore, the objective of this research is to propose an asset allocation model that expands the investment targets globally to gain lower risk and higher return. Based on the Mean-Variance model (Markowitz, 1952), this paper present three empirical studies that include various initial weights and rebalance strategies to study their effectiveness on global investing. The initial weight setting methodology includes “market capitalization”, “equal”, and “optimized”, while the rebalance strategies include “buy and hold”, “constant-mix”, “CPPI”, and “TIPP.” The following are several conclusions summarized from the empirical result. Market capitalization weights or benchmarks are good references of passive management, but the cost of weighting information is very high; equal weight is simple and risk-diversified, but the performance is not as good as optimized weighting. At last, the optimized weights are expected to reach best returns, while the performance is too sensitive to parameters and in-sample data. Optimization analysts must take care about the sensitivity problem. In the comparison of rebalancing strategies, although “buy and hold” got less transaction and manipulation costs, it has few chances to outperform other strategies; Constant-Mix is good for high volatility markets, however it is not suitable to global investing which is well diversified. CPPI provides downward protection, upward return, and high volatility. TIPP, modified from CPPI, is more conservative and provide fewer return and risk than CPPI.
URI: http://140.113.39.130/cdrfb3/record/nctu/#NT880396008
http://hdl.handle.net/11536/65587
顯示於類別:畢業論文