The Construction Of China's Financial Conditions Index Based On Mixed Frequency Model

The recent financial crisis has sparked an interest in the accurate measurement of financial conditions.An important lesson of recent events is that financial conditions,not only necessarily driven by monetary policy actions but also finance itself.The need for policy-makers to closely monitor financial conditions is clear. In response to this need,a recent literature has developed several empirical econometric methods for constructing financial conditions indices. Financial conditions indices are used for several purpose.For instance, measuring the current financial conditions, the effect of monetary policy, and forecast the future economic development. So how to effectively construct the index of Chinese financial conditions is very important.Our article references to the index of Chinese financial conditions which constructed by Goldman Sachs.On the basis of the six indicators we increase the ratio of loan to depositthe six indicators are interest rates,exchange rates,stock prices,the scale of social finance,house prices and money supply.these indicators not only include high frequency time series, such as interest rates and stock prices, but also contain low frequency time series, such as the scale of social financing.the vector autoregressive model is chosed as benchmark model. Finally, we expand indicators from 7 to 21 indicators, then use factor augmented vector autoregression model and the generalized dynamic factor model to construct the index of financial conditions, and compared with the mix state space model which is constructed in this paper.As for the results of the model, the index of financial conditions has upward trend,this mean that the liquidity of financial market is more abundant, Otherwise,this situation means the liquidity of financial market is more insufficient.First, the index of financial conditions is espressed FCIVAR which make use of impulse response function and vector autoregression model.The fluctuation of FCIVAR is more frequent which compared with the trend of CPI, the trend of the fitting effect is generally,Second, the index of financial conditions is espressed FCIDYM which make use of mix state space model.The trend of FCID YM is not only ahead of the trend of CPI, but also the trend of two indexes are relatively consistent.As for the rise or decline of interest rate,the real-time financial conditions index has the right response. Third,financial conditions index and CPI have causal relationship on the basis of Granger causality test.then determine the leading order as for CPI by the circular equation.Fourth, Robustness test-For the sake of the sample predictive ability between the benchmark model and mix state space model superior predictive ability test is used which proposed by Hansen. The null hypothesis is the benchmark model's ability to predict is not bad in the comparison of the model,we refused the nul hypothesis because of P= 0.32. The mix state space model of the financial conditions index has better predictive ability out of sample. Fifth, according to Bernanke (2005) we expand the index options.FAVAR model is proposed to get new financial conditions index which expressed FCIFAVAR. At the same time, financial conditions index is expressed FCIGDFM which use the generalized dynamic factor model. These two generalized factor model and mix state space model make comparison, the results of SPA test show that mix state space model still has certain advantages. Overall, this paper bases on the mix state space model to build the financial condition index can reflect the financial conditions in the future, and has reference value as for the policy makers.