| As the world economy has become more integrated,financial markets in various global economies are closely linked together and the Butterfly Effects are intensified.When we look back at every major fluctuation in capital market,economic policy has played an important role in it.Therefore,the research of economic policy uncertainty is becoming more and more important for both the academic and industrial circles.From a global perspective,the frequency and interval of financial crises are increasing and shortening.Resolving and preventing systemic financial risks has become a current top priority of China's financial regulatory authorities,and in the context of the current restructuring of the global financial system,Studying the inherent law of economic policy uncertainty is not only the need of theoretical progress,but also beneficial for China to take the lead in formulating correct economic policies that are conducive to determining competitive advantage.In academic field,researches on economic policy uncertainty have started very early.Stanford University Baker et al.compiled an economic policy uncertainty index in 2009,and published an article about measurement economic policy uncertainty in 2013(Measuring economic policy).From then on,various researches have been based on this index,and researchers have studied the impact of economic policy uncertainty in both macro and micro fields.In recent years,with the application of quantitative technology and big data,researches on the impact of economic policy uncertainty on capital markets have taken another step forward.This paper selects the stock market and the bond market for research,clarifies the transmission mechanism of economic policy uncertainty to the capital market,and empirically analyzes the impact of economic policy uncertainty on the two types of markets and the correlation between them.This paper takes the interaction of policy uncertainties in eight major economies,which includes the United States and the United Kingdom,into consideration.At the same time,the volatility data of macroeconomic variables are added,and reconstructs the EPUnew index for Chinese market by using the neural network model.Compared with the original China Economic Policy Uncertainty(EPU)Index,EPUnew uses Feature Extraction Method to take China Economic Policy Uncertainty Index(CEPU)as the benchmark,and adds the impact of policy changes of major foreign economies on Chinese financial market,which should be much more extensive and comprehensive in scope.The advantages of the EPUnew are mainly shown in the following aspects: the first is the globalization characteristic of the new index.The new index uses the Baker's original domestic EPU indexes of 8 major economies,which includes the United States,Canada,Britain,France,Germany,Italy,Japan and Russia,and combines the China domestic EPU index with the China EPU index according to Hong Kong news media.These 8 countries mentioned above account for more than 70% of China's foreign trade.Given that China is a major exporter,the policy changes of each country will significantly affect Chinese foreign trade and even cause domestic economic fluctuations.The new index combines the policy changes of the above-mentioned countries and weights them to maximize the degree of reflecting the uncertainty indicators in the whole world economy that can affect the Chinese domestic economy.The second is the use of neural network model.Compared with the traditional model,neural network model has excellent nonlinear feature,and less effective information is missing in feature extraction,which lead to the fact that the neural network model has been widely used in academic circles.Compared with principal component analysis,as a typical representative of unsupervised model,variational autoencoder model has many advantages such as high signal-to-noise ratio,strong expansibility and others.The third one is its significant impact on Chinese macro economy.Compared with the original index,the new index shows a stronger impact on domestic macroeconomic indicators,such as industrial added value,market fluctuations and business climate expectations.Those indicators show more significant impulse shock response.This paper studies the impact of economic policy uncertainty on Chinese capital market by using the new index.Chapter 4 is the analysis of stock market.Compared with those in the existing literature,we pay more attention to the impact of policy uncertainty on small and medium-sized markets as well as the time interval of subsamples.Garch-Midas and derivative models are used in this paper to model and analyze the long-term and short-term components of stock market fluctuations.In addition,the horizontal value and fluctuation value of economic policy uncertainty index are used to analyze the impact of policy uncertainty on the stock market in the modeling of long-term fluctuations.We find that the policy uncertainty has a long-term positive impact on the stock market volatility.The greater the policy uncertainty,the more unstable of the market sentiment and the greater the volatility in the stock market,which shows the obvious timeliness.When comparing the impacts on three types of markets,we find that policy uncertainty has a consistent positive correlation effect on the market.The higher the policy uncertainty causes the greater market volatility,and the impacts on different markets are basically the same.Combined with the empirical results,we conclude that the economic policy uncertainty can indeed explain some fluctuations for the stock market.Compared with the large market(CSI 300)and the small and medium-sized market(Small and medium-sized board and entrepreneurial board),the impacts of economic policy uncertainty on fluctuations are not significantly different.We get the conclusion as follows: firstly,the economic policy uncertainty has a consistent effect on long-term fluctuations.Compared with individual stocks,economic policy uncertainty as a macro factor has a major impact on the overall capital market.We consider that systemic risk is the main transmission channel of economic policy uncertainty;secondly,although economic policy uncertainty and market volatility cycle have basically remained the consistent,but in some periods,especially after the “stock market crush” occurred in 2016,there has been an inconsistency.After “stock market crush”,the market has been in a downturn for a long time.During this period,the government introduced a number of economic policies such as financial policy and monetary policy,but the result has been contrary to common sense.Furthermore,this paper studies the forecasting ability of the economic policy uncertainty towards stock market volatility and builds the economic policy uncertainty prediction model which contains macro and micro control variables.The forecasting ability of economic policy uncertainty in the Chinese A-share market in the full sample period(2002-2017)and sub-sample period(2008-2017)is analyzed.In this case,the significant forecasting results(within 5% and 10%,respectively)are obtained.We have made an empirical forecast of the primary industry for the overall market segmentations.The results show that except for telecom industry,EPUnew's forecasting ability is significant,and the financial industry is most affected by economic policy uncertainty.In the out of sample robustness test,MSFE-adjust is used to test its significance and we find that the economic policy uncertainty has significant out of sample forecasting ability(as compared with the benchmark model)for all industries,which include materials,industry,consumption,finance and real estate.To sum up,economic policy uncertainty can predict the volatility of A-share market in China,which is also significant under the out of sample robustness test.Then,the paper makes an analogy research on another part of the capital market-bond market.Compared with the overall fluctuation of the stock market,the bond market is less volatile.Economic policy uncertainty is mainly transmitted through the direct impact on market sentiment and the impact on real enterprises.Firstly,this paper studies the impact of Chinese economic policy uncertainty on Chinese bond market,especially the long-term fluctuations of Treasury and corporate bonds.Similarly,Garch-Midas is used to introduce the EPUnew index with the corresponding frequency to analyze the influence mechanism of the long-term components of the volatility in China's bond market.The results show that the China EPUnew is positively correlated with the long-term fluctuation of the bond market.Compared with Treasury bonds,the volatility of economic policy uncertainty has a more obvious impact on the volatility of corporate bonds,which is due to the fact that private enterprises bonds account for larger proportion among corporate bonds,and they are more affected by the overall macro environment.The increase of economic policy uncertainty will increase the volatility of the bond market.We use Markov switching model to find that when the market is in a state of great fluctuation,economic policy uncertainty has a great impact on the market risk and the credit spreads widen.However,when the market experiences low levels of volatility,the uncertainty of economic policies still has a positive impact on the market,which leads to a higher credit spread,while the changing range is not as great as that of state 1.When economic policy uncertainty increases,Chinese Treasury bonds in the bond market are barely affected due to government endorsement,while corporate bond yields will decrease due to worries that industry will be influenced by policies and investors will lower expectations about future corporate earnings.The credit spread reflects the differences between corporate bond and Treasury bond.Thus,bond market fluctuations mainly follow the more affected corporate bonds.Further,we analyze the impact of economic policy uncertainty on excess returns in the bond market.Similar to the market fluctuation,the excess return of bonds is often affected by macro variables.On the basis of the existing research,we control the traditional macro variables such as one-year deposit interest rate(RF),inflation index(CPI),market liquidity index(LIQ)and SLOPE and we find that economic policy uncertainty has a significant impact on the excess return of bond market.With the increase of economic policy uncertainty,the excess return of long and short term bonds in the Treasury bond market is decreasing.Yields on long-term Treasury bonds fall more than those on short-term ones.We take the new comprehensive wealth index(WI)of China bonds as the representative of the bond market.Different from Treasury bonds,the prediction coefficient of economic policy uncertainty on WI is positive,which indicates that with the increase of economic policy uncertainty,WI's excess return relative to benchmark bond(one-year Treasury bond yield)is also increasing.Finally,we reconsider the forecast of the yield difference between the composite wealth index and Treasury bonds of the same maturity due to economic policy uncertainty,and the results show that the prediction coefficient of excess return of bonds with different maturities under economic policy uncertainty is basically unchanged.There is no obvious monotonic change,which indicates that EPUnew's influence on bonds credit risk of different maturities is close.Finally,this paper analyzes the influence of the correlation between stock market and bond market from the perspective of single market.As two major components of Chinese capital market,the stock market and the bond market reflect the trends of China's economy,and the correlation between the two often reflects economic situation and market sentiment.We use the daily data of Shanghai Composite Index and Shanghai Treasury bonds from January 2003 to May 2018 and Dcc-Garch model to obtain the Dynamic Conditional Correlation between the two Markets(ADF test shows that it is a stationary process and J-B test rejects the assumption that it is normally distributed).When we use Garch-Midas model to analyze the impact of economic policy uncertainty on correlation coefficients,we find that the model estimation results based on the horizontal value and fluctuating value of economic policy uncertainty show a positive relationship between economic policy uncertainty and correlation coefficient.When the horizontal value or fluctuation value of economic policy uncertainty increases,the volatility of market correlation increases.We adopt Markov switching model to analyze the impact of policy uncertainty on the correlation level.We classify the market into two state,under the overall “active” state of the market,we find that the increased uncertainty of economic policy will lead to the decrease of market correlation(from positive to zero to negative),which reflects that risk aversion towards assets increases when the macro environment is unstable and capital flows from one market to the other.In a period of inactivity,economic policy uncertainty is negatively correlated with the long-term correlation coefficient of the two Markets as well,which indicates that the impact of policy uncertainty on the market turns weak.Both states above show the sensitivity and risk aversion of capital to the external environment.Further analysis of the impact mechanism shows that due to the fluctuations caused by the newly issued policies on the industry,the financing costs of enterprises rise and the capital flows into the relatively stable bond market instead of the stock market with higher risks,which leads to stocks and bonds move in opposite directions(stock market declines and bond market rises),and reduce the correlation between two markets.Therefore,it can be said that the decrease of Chinese economic policy uncertainty will lead to the increase of the correlation between Chinese stock market and Chinese bond market.The innovation of this paper is mainly reflected in the following aspects: 1.it takes the lead in using machine learning method in economic policy uncertainty index analysis.This paper proposes the construction of new economic policy uncertainty index for the first time based on the neural network model for Chinese capital market.Compared with the original quantitative index,the new index can better reflect the disturbance and impact of economic uncertainty on the Chinese capital market,and it is more effective in terms of industrial growth,market sentiment and market volatility,which is verified from multiple perspectives in the paper.In the follow-up studies,different results of the new index and the original index in the empirical study are also compared and the validity as well as practical application value of the new index are further pointed out.2.This paper progressively illustrates the impact of economic policy uncertainty on Chinese capital market from multiple levels.From the perspective of classic returns and market fluctuations,we select several important stock and bond indexes as the sample cross-sectional data,which basically cover the main part of China's capital market,the results of the analysis are significant and effective,and have universal significance.At the same time,traditional macro factors are introduced for each independent market,including macro factors such as inflation,market sentiment index,and bond market liquidity.After controlling the above variables,the new economic policy uncertainty index analysis results are still significant.This paper also conducts a robustness test,for example,in the empirical analysis section(Chapters 4-6),Baker's original China Economic Policy Uncertainty Index is compared with the EPUnew index,which enhances the persuasiveness of the research.3.Some significant research results are found in the study,which can be used as policy improvement and suggestions,for instance,economic policy uncertainty has different degrees of influence on the fluctuations of different industry portfolios.When implementing economic policies,regulators can take full account of the effect intensity on different industries and choose the optimal implementation path.In addition,the market could be divided into periods of high volatility and low volatility.Compared with the period of low volatility,the negative impact of uncertainty caused by economic policy releasing in the period of high volatility will be offset by the beneficial effect of the policy,which shows that the government's “market rescue” policy has significant positive significance in practice.In terms of stock returns,Markov switching model is used to divide the market into two major blocks.Economic policy uncertainty has a negative effect on the returns of each block.Therefore,the government should minimize policy intervention to market Stabilization. |
PDF Full Text Request