An ARMA model is a time-series model that explains a variable using both its own past values and past error terms.
$$$$
The model
A standard ARMA((p,q)) process can be written as:
$$
y_t = c + \sum_{i=1}^{p}\phi_i y_{t-i} + \varepsilon_t + \sum_{j=1}^{q}\theta_j \varepsilon_{t-j}
$$
The autoregressive part captures persistence in the series itself, while the moving-average part captures how past shocks continue to affect the series.
Why economists use it
ARMA models are useful when economic data shows both persistence and short-run shock dynamics. They are a standard building block in forecasting and in the study of macroeconomic and financial time series.
What the model helps reveal
Different combinations of (p) and (q) produce different autocorrelation patterns. That is why ARMA models are closely tied to autocorrelation diagnostics and model identification.
Knowledge Check
### An ARMA model combines:
- [x] lagged values of the series and lagged shocks
- [ ] only cross-sectional variation
- [ ] only current-period information
- [ ] tax rates and unemployment
> **Explanation:** The AR part uses lagged values and the MA part uses lagged error terms.
### Why are ARMA models useful in forecasting?
- [x] They capture persistence and short-run shock effects
- [ ] They require no data
- [ ] They eliminate volatility
- [ ] They apply only to panel data
> **Explanation:** Many economic time series have both ongoing momentum and residual shock structure.
### The moving-average part of an ARMA model captures:
- [x] the lingering effect of past shocks
- [ ] only deterministic trends
- [ ] the average tax rate
- [ ] equilibrium unemployment
> **Explanation:** The MA terms model how previous disturbances still affect current outcomes.