چکیده:
Inflation forecast is one of the tools in targeting inflation by the central bank. The most important problem of previous models to forecast the inflation is that they could not provide a correct prediction over time. However, the central bank policymakers shall seek to create economic stability by ignoring the short-term and temporary changes in price and regarding steady inflation. On this basis, in the present paper, it has been aimed to provide nonlinear dynamic models to simulate the inflation in the economy of Iran using quarterly data referring to 1988- 2012 as well as TVP-DMA and TVP-DMS models. These models can provide changes in input variables as well as changes in the coefficients of the model over time. Based on the results, the possibility of growth of currency in circulation, economic growth, also the growth of deposits either visual or non-visual variables, is more remarkable in modeling of inflation in economy of Iran. In addition, the predictive power of dynamic models presented in this study is more than other models
خلاصه ماشینی:
On this basis, in the present paper, it has been aimed to provide nonlinear dynamic models to simulate the inflation in the economy of Iran using quarterly data referring to 1988-2012 as well as TVP-DMA and TVP-DMS models.
Overall, considering the Phillips curve in the past half-century review suggests the important point that relationships between variables have changed over time; according to Stock and Watson (2008) one of the problems that previous models had in prediction was that they could not correctly predict in all periods of time, and sometimes it was observed that some models could predict the estimation of recession well, and some others could predict the estimation of the boom better.
in recent years, major studies conducted in the field of inflation forecast have often been in the form of time varying parameters (TVP) models, Monte Carlo Markov Chain (MCMC) (Nakajima, 2011).
(2007) in combination with TVP model, and applying the method of Stock and Watson (1999 and 2008), the power of approved variables has been investigated through the theoretical foundations of the Phillips curve and the main variables in the domestic empirical studies that had significant impact on inflation, and non-linear impact on inflation in Iran.
In terms of structural breaks and cycle changes in time series (which is the main feature of time series in Iran’s economy), the conventional methods are not enough to calculate the parameters, in this condition Kalman filter provides the possibility of modeling of the above facts with variable coefficients over time, (Stock and Watson, 2008).