Out of sample forecasting

During the last decade, several studies have argued that sticky information model proposed by Mankiw and Reis (2002), in which firms update their information occasionally rather than instantaneously, explains some stylized facts about the inflation dynamics. Sticky information pricing model successfully captures the sluggish movement of aggregate prices in response to monetary policy shocks. Despite the importance of sticky information, no empirical studies have been done yet to estimate sticky information Philips Curve (SIPC) and its key parameter - the degree of information rigidity - in Iran. This paper is the first attempt to estimate the degree of information stickiness in Iran using the two stage empirical approach proposed by Khan and Zhu (2006). Having the correct structural parameter allows a better understanding of the dynamics of inflation. Results show that the average duration of information stickiness ranges from 3.2 to 4 quarters in Iran. In addition, the existence of threshold effects in SIPC is also tested in this paper. Based on the estimation of TAR model over 2002Q2- 2015Q1, firms update information faster when inflation is higher. This evidence suggests that firms are more aware of macroeconomic conditions when inflation is higher; that is, missing information during high inflation periods is costly. JEL Classifications: E31, E37, C53, D84

The aims of the present study are developing a financial stability index (FSI) using banking indices to measure financial stability in Iran, and examining the relationship between financial stability and macroeconomic variables for policymaking. To these ends, we have employed principal-component analysis, out of sample forecasting, Autoregressive Integrated Moving Average (ARIMA) method, and Vector Error Correction Model (VECM). The monthly data period is spanning 2007:3 through 2017:2. We find evidence of one cointegrating vector. According to the cointegration test, there is a long-run relationship running from inflation, Gross Domestic Product (GDP) growth rate, and unemployment to FSI. Also, the results of the Engle-Granger test indicate bidirectional causality between FSI and unemployment. Forecast evaluation shows that VECM-based FSI prediction is more accurate than the ARIMA model.