4.Empirical Study

4.1 Methodology

We firstly test the stability of each series.Tab.10 reports the results from ADF Fisher unit root tests.All the series have not unit root and are stationary.Then we establish three equations for estimation.Eq.（18）based on Bruno Jetin（2012）study explains China's imported inputs of production for export（CIM）by domestic demand（Y）, foreign demand（Ywp）, real effective exchange rate（E）and price deflator, including export price deflator（Px）, import price deflator（Pm）and inflation price（P）.Eq.（19）explains China's total export by macroeconomic approach by relative labor cost compared with China's main trade partners（RW）, relative labor productivity（RProd）, foreign demand, price deflator and foreign content.Because the relative labor cost and productivity are already converted into U.S.dollar, we do not use the variable of exchange rate in this equation for avoiding autocorrelation.Eq.（20）does panel data estimation to re-estimate the relationship between China's export and foreign content.The HAUSMAN test was used to select a preferable model between fixed effects（FE）and random effects（RE）. Since HAUSMAN test statistic under RE model is not significant at 5% and the adjusted R-squared is only 0.03, the FE model was selected.EXPit is China's export of sector i in time t, CIMit is China's intermediate inputs'import of sector i in time t, λ is unobserved common effect, γi is sector-effect, μt is unobserved time-effect, and εi t is the error term.

Figure 10 share in China's import by source country（area）and tech level

4.2 Results

Tab.11 reports the estimated results of Eq.（18）.Column（1）and（2）have too high DW coefficient thus column（3）,（4）,（5）are preferable.The coefficients are similar in these columns.Since intermediate inputs'import is just for export, it does not depend on domestic demand.On contrary, foreign demand and export price deflator have positive effect on foreign content.Besides,10% appreciation of yuan conducts 8.9% diminish of foreign value added embodied in export.Because an appreciation will reduce china's export then the needs of imported inputs for export will absolutely go down.Different from ordinary import, the import price has a positive effect on intermediate imports.This part reveals the correlation between variable of CIM and others, which permits us to ameliorate the following estimation.For avoiding autocorrelation among explicative variables, it is better to not use variable of Ywp, E and price when CIM is taken into account.

In Tab.12 concerning eq.（19）, it is showed that the higher labor productivity and foreign demand, the more export China wins.Meanwhile, an increase of labor compensation and of export price will reduce export.China profited cost advantages and price competitiveness for export progression during recent years.Yet the CIM has no significant impact on export, which is opposite to our proposition.Explanation is eq.（19）analyzes the overall economy's export.The CIM has different effects for each sector.When all the sectors are taken together, its effects become bias.This explanation is proved in follows.

Tab.13 illustrates the results of eq.（20）, which utilizes panel data estimation of each sector.Column（1）shows the result of overall economy's estimation.Column（2）concentrates on manufacturing sectors and Column（3）,（4）add variables of real effective exchange rate or relative labor cost and productivity as well.All the coefficients of.ln（CIM）have expected signs and are statistically significant.Generally, intermediate inputs'import has positive impacts on Chinese export progression.In column（1）, a 10% imported inputs augment leads an increase of 9.5% in exports.When considering technological level, the effect is less for low qualified sectors（9.4% and 8.3%）, while it is bigger for high qualified one（11.8% and 24.2%）.Besides, effect became more significant during the decade of 1997-2007.The coefficient rose from 9.9% to 55.4%.Column（2）shows the general effect for manufacture is higher than that for overall economy.10% CIM augment conducts an increase of 13.1% in export of manufacture.This indicates the robustness of previous inference that Chinese manufacturing export relies more on imported intermediate inputs compared with overall export.The higher tech level, the more dependent on intermediate inputs'import and the dependence increased over times.

Table 1 1 Δl n（CIM）

Note:∗∗∗Significant at 1% level; ∗∗Significant at 5% level; ∗ Significant at 10% level.

Table 12 Δl n（EXP）

Note:∗∗∗Significant at 1% level; ∗∗Significant at 5% level; ∗ Significant at 10% level.

Table 13 Δl n（EXP）

Note:∗∗∗Significant at 1% level; ∗∗Significant at 5% level; ∗ Significant at 10% level.