Forecasting The Turning Points Of Business Cycle And It’sIinfluencing Factors’ Analysis Based On Leading Indicators

Business cycle refers to an alternative phenomenon of economic expansion and contraction in economic operation. Generally the business cycle is divided into upturned, boom, decline and recessionfour phases. The research on business cycle can be divided into theoretical and empirical studies. The objective of theoretical study is designed to explore the reasons for the business cycle based on theoretical models of certain assumptions, and show that the trend of economic operation according to the model analysis results; while the main task of empirical study is to describe the business cycle, and gives more accurate measurements of the stage of economic operation according to various economic indicators that can be used, this paper is an empirical study on the business cycle.Since the outbreak of the global economic crisis of2007-2009, the prediction problem of business cycle’s turning points has aroused wide public attention. Determining the business cycle turning points is an important problem for families, businesses and investors, because the decisions on consumption, production and investment currently are dependent on their expectations about future economic prospects. For central bankers and policy makers, the accurate and reliable prediction of economic conditions is conducive to the implementation of appropriate and pre-emptive policies.This paper identifies the benchmark indicator first, namely, the coincident composite index released by the National Bureau of Statistics, and determine the reference date of business cycle using the B-B method, and then summarizes the foreign leading indicators system. To select leading indicators from different areas, we use the time difference correlation analysis and peak to valley corresponding methods. In the end, we select twelve leading indicators, namely:crude steel production growth, fertilizer production growth, automobile production growth, pig iron production growth, cement production growth, end of M2of year-on-year growth, investment growth in fixed assets, the number of new projects this year, the cumulative export year-on-year growth, reverse CPI, stock turnover growth and commercial housing sales growth, the data period used in this paper is1999.1-2013.5.Because of the need of empirical research in this paper, we use principal component method based on the correlation matrix to solve factor loadings for the final selected twelve leading indicators, and calculate the eigenvalues, the contribution rate and the cumulative contribution rate at the same time, in order to obtain clear economic implications for the factor loading matrix, we make the initial factor loading matrix orthogonal rotation using the maximum variance method. We extract four factors in the end, respectively is:the main upstream product yield factor, liquidity and investment factor, restrain demand factor and wealth effect factor.This paper based on the four factors and different forms of probit models, the coincident composite index released by the National Bureau of Statistics as the benchmark, a turning point in the economic situation is analyzed and forecast. This paper focuses on the comparison of the prediction effect of the static probit model and the dynamic probit model, when using a single variable static probit model to estimate, the predicted results are not reliable, so further construction of multi variable model is needed. Although the two different forms of probit model contain the same variables, but each explanatory variable lag period is different, we use the highest Mc Fadden R2to select each explanatory variable optimal lag order in different models. The in sample forecasting results show that:four factors extracted by leading indicators can be predicted to a certain degree of economic fluctuation in China, and the dynamic probit model performs better than the static probit model. We also provide probability threshold values of0.5and0.25respectively and found that:the lower probability threshold value improves the prediction of recession months, while increase the recession warning signal error in the same time. For the out-sample forecasting, we use the mean absolute error (MAE) and root mean square error (RMSE) to measure and compare, the results show that:the prediction model we use established by the four factors is reasonable and effective. The dynamic probit model performs better than the static probit model in the out of sample prediction, we use different threshold values to predict the probability of recession months and find that:the probability threshold value of0.25can be more accurate for the out of sample forecasting than the probability threshold value of0.5.Finally, the factors affecting the economy were analyzed according to FAVAR model, the empirical results show that:(1) we give the output of major upstream product a unit positive impact, it will produce a positive response to the economic situation of our country, this effect can probably lasts for about a year or so;(2) the impact of liquidity and investment factors of the economic situation is also positive, increasing the money supply or increasing the fixed assets investment are both have a promoting effect on the economy, and the promoting effect lasts longer than that of the output of major upstream product, and continues for about two years;(3) restrain demand factors have inhibitory effect on the economy, so the demand factors contribute to the economy and therefore increase in demand can promote economic development;(4) the wealth effect has a positive impact on China’s economy. But on the whole, the greatest impact on the economy is liquidity and investment factors, namely, the changes of money supply and investment.The main innovations of this paper are the following three points:(1) in this paper, extraction factors based on leading indicators as corresponding explanatory variables to do empirical analysis.(2) This article uses different forms of multi variable to predict and the prediction results are compared.(3) This article uses the advanced FAVAR model to analyze the impact of the four factors on the economic conditions.