Module code: EC7059
In this module you will study time series methods and analysis in time domain.
Time series data is the sequence of observations constructed by sampling data at fixed time intervals, thereby tracking the underlying variable’s behaviour over time. The vast majority of economic data are collected and recorded in time series format including inflation rate, real gross domestic product (RGDP) and exchange rate.
What primarily distinguishes time series data from cross section data is that the observations are all connected and interdependent. This can make some conventionally used statistical modelling techniques unusable because they only work when observations are independently and identically distributed ('i.i.d.'). Any model of time series data must be able to capture, or uncover, this interdependence.
Interdependence is a particularly distinguishing feature of economic time series data. In any economic system a wide range of variables influence one another through different channels. Usable modelling techniques need to capture this interdependence, and gauge one variable’s influence over others, over time, accurately.
- Autoregressive integrated moving average (ARIMA) models
- Maximum likelihood estimation technique
- Vector auto-regression (VAR)
- Impulse response functions
- Multivariate least squares
- Vector error correction models (VECM)
- The Granger representation theorem
- 24 hours of lectures
- 9 hours of seminars
- 3 hours of computer classes
- 111 hours of guided independent study
- Coursework (30%)
- Exam, 2 hours (70%)