"the Facts" Thoughts In The Application Of The Macro Policy Effect Assessment

Formulating and implementing policies is an important means for governments to realizetheir function. Recently, Chinese economy is in a stage of rapid development, and social reform isgradually deepening, policies and rules change frequently. In order to reduce the cost of policy,promote the scientific nature of government decision-making, and enhance the government’sability of macroeconomic control, evaluation the effect of implemented policies is tremendouslyneeded. Evaluating policy effect through "counterfactual" constructing is a hot topic in the studyof policy evaluation, but this idea has not made much progress in macroeconomic field, whichrestricts the theory and practice development of macroeconomic policy evaluation.As the characteristics of Micro and Macro data are different, the widely used methods for"counterfactual" construction in microeconomics, such as Difference-in-differences and matching,are not possible to be applied to macroeconomic policy evaluation. While the main methodrecently used in macroeconomic policy evaluation, such as vector autoregressive model, dynamicstochastic general equilibrium model also have some shortcomings. For example, they requires toset up several structure models, collecting large amounts of data, solved by solving complex, theworkload is huge and there exists difficulty in causal inference. Therefore, we need to take the"counterfactual" thought as the guide, and to explore new methods suitable for macroeconomicpolicy evaluation. Under this background, this paper introduced a new method proposed by Hsiao(2010), which used the correlation between cross-sections of panel dada to build "counterfactual"and estimated the policy effect, and conducted a series of theoretical and empirical analysis aroundthe method.Based on the introduction of Hsiao’s theoretical model, this paper focused on the analysis ofthe key assumption in the model--conditional independence assumption, and used Monte Carlosimulation techniques to see how sensitive the empirical result was to the assumption. Secondly,this paper took Western Development Program and China’s "Four Trillion" Stimulus Package asexamples, elaborated the application steps of this method and the problems that should be noticed.The case of Western Development Program met the key assumptions, while the latter violated thisassumption more or less, but we made it satisfy the critical assumption through a variety oftechniques. In the analysis of the former case, we focused on the inspection of key assumptions. As is regard to the latter case, we designed comparison model, and used the real data to figure outthe influence of the key assumption on policy effect calculation. Through the theoretical analysisand empirical analysis, the paper draws the following conclusion:The theoretical analysis results show that, the result of the Hsiao’s method is very sensitive tothe critical assumptions, even if slight violation would have great impact on the empirical results.Therefore, we should consider the applicability of the method before using the model. At the sametime, because the actual data generation mechanism is unknown, the key hypothesis cannot betested by statistical methods, only be judged by fully detailed proof. In practice, if the keyassumption is not satisfied, one should adopt DID and other methods to evaluate the policy effect.Compared with the structural models, this method does not have to establish structure equations,and avoids solving the problem of endogeneity that’s common in the field of policy evaluation. Inaddition, it’s easy to reflect the impact of external shocks.After all, this model is simple, requiressmall amount of data, and easy to solve. As there are so many advantages, Hsiao’s method willhave important influence in the practice of macroeconomic policy evaluation.The results of the empirical analysis show that, Western Development Program is essential innarrowing the gap between western regions and the other areas. Western Development Programhas raised the real GDP growth rate of Chongqing by2.24percentage during the year2000and2007, and this effect is steady around years. This result is similar to that of Liu Shenglong (2009),indicating that Hsiao’s method is effective. But contrary to the research conclusions of otherscholars, empirical analysis of the effect of China’s "Four Trillion" Stimulus Package shows thatthe investment raised the real GDP growth rate by0.93percentage in2009,-0.575percentage in2010, and0.021percentage in2011. In the whole study interval, the overall effect is0.138percentage points, almost zero. The results indicate that for China, the Stimulus Package hasplayed a considerable role in stopping its recession in the global financial crisis, but it will notaffect the economy significantly in the long run. This Indicates that the government should takethe crowding-out effect into consideration, and focus on side effects of inflation and other factorscaused by this policy while implementing expansionary fiscal policies.