Impulse response of export price to production cost based on data mining

Influence of production cost on export price is always the hot topic of foreign trade research. Previous researches just use original series to establish models but not fully discover information content of time series signals. Taking China's citrus trade for example, the paper uses wavelet analysis to decompose production cost and export price into four series, which are trend, cycle, volatility and noise. After filtering noise, the paper forecasts the influence of trend, cycle and volatility of production cost on that of export price by using impulse response and variance decomposition which are based on VAR. Results show that in the long run, change trend and cycle of production cost's influence on that of export price is very significant, but the volatility of production cost's influence on that of citrus export price is very small.