Research On Optimization Of Investment Portfolio Among Asian Stock Markets Based On Different Different Weighting Rules

With the rapid growth of financial markets and the continuous upgrading and upgrading of financial products,the relationship between cross-regional securities markets has become more and more complex,making their correlations more and more complex,and investment behaviors across countries and regions are increasing,and the volatility of securities markets has also increased.Intensifying.In this context,there are usually some special situations that cause people's attention.For example,the price fluctuations of the securities market in a certain region often spread quickly to other regions,sometimes even globally,resulting in significantly increased financial risks.According to this,the need to build an investment portfolio to prevent risks is even more urgent.How to construct an investment portfolio and how to optimize the investment portfolio to serve humanity efficiently to reduce financial risks and thereby build a more efficient securities market are subject to the theoretical and practical circles closely.How to make the constructed combination reliable is a question worth thinking about.Fortunately,Markowitz's portfolio theory has pointed out a direction for our thinking.The idea emphasizes that when the return is fixed,the risk requirement is the smallest.According to this,high-precision measurement of risk is of great significance to the construction of asset portfolios.However,using the Mean-Variance optimization model proposed by Markowitz,using variance as an indicator instead of risk,it is difficult to describe the actual risk of the asset portfolio,mainly because:(1)The size of future returns away from the expected return is to use variance to Representative,however,the actual scale of risk is ignored here.Therefore,it is difficult to describe the actual risk faced by the portfolio;(2)Variance does not have directionality,and the measurement of risk is vague.In other words,it may produce The gain from exceeding the mean value,that is,the first half of the variance,may also produce losses that are lower than the mean value,that is,the second half of the variance.From the perspective of investors,the second half of the variance is indeed the risk faced by the actual situation.But the variance in the first half is different,andthis part will generate the necessary returns for investors.Therefore,replacing risk with variance is flawed.Therefore,this paper uses Value at Risk(Va R)and Conditional Value at Risk(CVa R)to replace the variance in the mean variance model,make some adjustments to the model,and establish Mean-Va R and Mean-CVa R respectively.It is necessary to clarify that in the exploration of asset portfolio optimization,two core issues are involved,namely,the selection of the objective function with the least risk under a certain return of the asset collection and the correlation of each asset in the asset collection.For the second core problem,the past The research mainly uses qualitative,linear and quantitative methods to study asset correlation,but financial time series are often affected by some typical facts such as autocorrelation,and the financial field is a cumbersome dynamic system,between markets or between assets frequently show the relationship of non-linear interdependence,so the above method is weak in face of the actual situation.Therefore,this paper uses the Copula model that can depict the non-linear relationship between assets to capture the interdependence structure of assets in the portfolio to accurately weigh the risk of the collection,thereby promoting the improvement and optimization of the collection of assets.First of all,the logarithmic returns of the three major stock market indexes are obtained,and the ARMA-APARCH model is used to model the marginal distribution of the logarithmic return series of each stock market,and some typical factual characteristics included in the return series are described.Such as "biased",autocorrelation,heteroscedasticity,etc.Secondly,the maximum spanning tree algorithm is used to carry out research,the absolute values derived from each layer in the Copula method are summed,and the largest is selected to determine a multivariate Copula function sufficient to reasonably describe the correlation between the three markets,eliminating complex mutual The role of dependency structure in building asset portfolios.Then,after obtaining the interdependent structure of the assets,the obtained useful parameters and the conditional variance and mean of the risk factors are sorted and analyzed,and then,the Monte Carlo model is used to simulate the m types of possible future returns of the selected asset,And combine three weighting rules to build an asset portfolio optimization model.Finally,compare the risk backtest results of the three sets of weighted assets.After empirical analysis,the following conclusions are drawn:(1)The ARMA-APARCH method can efficiently capture the "autocorrelation" and "heteroskedasticity" characteristics of financial asset time series;(2)when describing the interdependent structure between multiple assets,The multivariate Copula model has a flexible and rich form,which makes the multivariate Copula method can improve the accuracy of describing the interdependent structure between multiple assets;(3)Compared with the CVa R weighting method and the GMV weighting method(Global Minimum Variance),The prediction value of the model under the equal weighting method is farther from the actual value,which shows that the prediction ability of the model under the equal weight is weaker than the prediction ability of the model under the GMV and CVa R weighting methods.The above conclusions provide a reliable technical reference for more accurate optimization of asset portfolios,and also provide theoretical support for financial risk management authorities to enhance the management of financial risks,and can also provide some help for investors to strengthen their risk understanding.