Abstract | As the China’s economy shift to new normal and the external environments more complicated, the pressure of the economic downturn is increasing. Under the objective requirements of maintaining stable economic development, the importance of stabilizing expectations has become increasingly prominent. Since the changes in asset prices are generally considered to reflect the expectations of economic development, this paper attempts to clarify whether asset prices help to understand the role of expectations in the China’s economy by exploring the relationship between asset prices, anticipated shocks and macroeconomic fluctuations.
This paper builds a small DSGE model that considers the long-run anticipated shock process and asset prices data for Bayesian estimation. First, the result of the unconditional variance decomposition has changed greatly in the case of changing only the set of observed variables. When asset prices data are not included, the model overestimates the effect of anticipated shocks on output growth and consumption growth. Including asset prices does not bring on the deviation between theoretical moments of the model and standard deviation of actual data, so the changes in variance decomposition is not the result of introducing high volatility data. The estimation results of the simulation series show that only the asset price data can be used to recover the contribution of the exogenous shock. Secondly, this paper also analyzes the short-run anticipated shock process to test the rationality of the long-run anticipated shock process, and detailed the reason of long-run anticipated shock fit actual data better. Finally, we clarified the importance of anticipated shocks to each macroeconomic variable by forecast error variance decomposition and historical variance decomposition. The results indicate that anticipated shock is the most important factor for fluctuations of total market valuation. The theoretical basis of this paper is the idea of news driving business cycle, originated from Pigou (1926) and Clark (1934), Cochrane (1994) and Beaudry and Portier (2004) re-emphasize the important role of expectations for the economic cycle. Beaudry and Portier (2006), Matsumoto et al. (2011), Malkhozov and Tamoni (2015), and Avdjiev (2016) provide the theory foundation for the introduction of asset prices in our model, Matsumoto et al. (2011), especially, proved that in a standard closed DSGE model, the introduction of asset prices and anticipated shocks does not inhibit the ability of the model to fit actual data. In addition, the paper draws on the long-run risks strand of the asset pricing literature. This type of literature proved that sustained and expectable exogenous shocks can help to explain key asset pricing phenomena, such as the equity premium puzzle and the risk-free rate puzzle. Chen Guojin et al. (2014) used the Chinese stock market data to test the macro long-run risk model. They concluded that the long-run macro risk and macroeconomic uncertainty can also help explain the key asset pricing phenomenon in China's stock market.
This article uses data from the first quarter of 1993 to the fourth quarter of 2017. The data are from the Wind database and the Center for Macro Finance Research (CMF) (http://cmf.cafr.cn/data/listpage). Quarterly data including GDP by expenditure, total retail sales of social consumer goods, total investment in fixed assets, relative price of investment products, total stock market valuation, seven-day repurchase rate of inter-bank bond pledge and consumer price index were selected.
The results we obtain indicate that: (1) Anticipated shock is the main explanation for asset price fluctuations, and asset prices contain market expectations for future economic fundamentals. (2)The information contained in the asset price helps to correctly understand China's macroeconomic fluctuations and the role of expectation, and to catch the main channels of the anticipated shock function. Therefore, under the framework of the news driving business cycle, the asset price data should be fully considered when using Bayesian method to study the economic cycle. (3) Through the comparison of the models, when the asset price data is included, we find that the setting of the long-run anticipated shock process fit better for the China’s economy. (4) Anticipated shock has a certain influence on the economy, while it contributes less to economic fluctuations than previous literature.
This paper makes following contributions. First, it fills the gap in domestic research on the relationship between asset prices and expectations, and proves that China's asset prices can reflect agent’s expectation. Second, it expands the model structure of domestic DSGE research. Domestic research on the economic cycle rarely takes into account the long-run shock process. This paper provides a model framework for future research by comparing the long-run and short-run shock processes and studying the relationships between the long-run shock process and asset price. Thirdly, it helps to expand the research perspective. The comparison of variance decomposition proves that asset prices help us understand the role of expectations in the China’s economy. Finally, from the perspective of policy significance, this paper argues that in the current context, policy authorities should pay attention to capital markets for expectation management, stabilize the relative prices of investment products, and promote the development of the real economy by ensuring the prosperity and stability of the capital market and conducing the formation of positive public expectations.
|