EC313 Midterm II

Information

• Time: 60 Minutes.
  • All questions equally weighted.
• Instructions:
  • Carry out tasks using R in RStudio.
  • Save tasks in either .R or .Rmd file, email to j.j.reade@reading.ac.uk.
    • Add comments to your code to explain your steps.
  • Email an .R file with all your commands (from History window), along with any downloaded data files, to me also.

Tasks

1. Direct download task:
   1. Download data from quantmod the following series: Consumer Price Index for All Urban Consumers: All Items, which has the code CPIAUCSL. Create a time series object and plot the resulting object.
   2. Use the functions log() and diff() to create inflation, making use of the approximation that \(\%\Delta x = \Delta \log (x)\).
   3. Create at least two models (e.g. ARIMA, exponential smoothing, time series decomposition) for inflation. Which models appear best and why?
   4. Select an appropriate training period and select the best forecast model for each variable.
   5. Create forecasts for inflation for the next six months and submit at http://goo.gl/forms/ooUgNAzo2O (or see below).
2. Indirect data collection task:
   1. The website http://www.fxstreet.com/economic-calendar/ contains information on upcoming data releases. On Friday, UK “consumer inflation expectations” are released. Find this entry in the calendar, and download the available data (via “History Table”, and “Export”).
   2. Use the isat function from the gets package to carry out both outlier and break detection on each variable, and implement these variables into an autoregressive model.
   3. A theory of expectation formation suggests that inflation expectations are influenced by current actual inflation. Collect data on UK inflation, and construct a linear regression model including actual inflation.
   4. Generate a forecast for inflation expectations based on your models created thus far, and submit it at http://goo.gl/forms/ooUgNAzo2O (or see below).