A Study Of Methods Of Inflation Forecasting Based On Survey Data

After transforming quarterly qualitative survey data publicized from the Consumers'Saving Survey System conducted by the People's Bank of China into quantitative expected inflation data, and get quarterly inflation rate out of monthly CPI, this thesis studies the methods of how to use inflation expectation to forecast real inflation rate. It shows that simplified survey data model cannot fit the real data well, and the forecasting error is much bigger than traditional time series ARMA models. To promote the result, this thesis combines inflation expectation and real inflation rate together to do the forecasting which turns out to be a ADL model. The result of ADL turns out to be much better, and it gets smaller error. But ADL is not the best model in forecasting, many other variables and factors need to be considered in order to get a perfect model.