- CFA Exams
- CFA Level I Exam
- Topic 1. Quantitative Methods
- Learning Module 5. Time-Series Analysis
- Subject 2. Autoregressive (AR) Time-Series Models
CFA Practice Question
Which equation(s) represent(s) an AR time-series model?
II. yt = b0 + b1t + εt.
III. lnyt = b0 + b1t + εt.
IV. xt = b0 + b1xt-1 + b2xt-2 + ... + bpxt-p +εt.
I. xt = b0 + b1xt-1 + εt.
II. yt = b0 + b1t + εt.
III. lnyt = b0 + b1t + εt.
IV. xt = b0 + b1xt-1 + b2xt-2 + ... + bpxt-p +εt.
Correct Answer: I, II, III and IV
A key feature of an autoregressive model is that the current period values are related to previous period values. The 4 equations all presents the relationship.
I is called AR(1) model and IV is called AR(p) model. Notice that II, a linear trend model, and III, a log-linear trend model, are special cases of AR models.
User Contributed Comments 5
| User | Comment |
|---|---|
| MasterD | Am I mistaken or does II and III not have any t-1 values, only t. If so, then what's all this about all options relating to PREVIOUS period values? |
| ucsbdan | II and III are special cases: check the text book. |
| ericczhang | I got it wrong too, but I'm guessing since t = (t-1)+1 you can rewrite the regression equation in terms of t-1 and thus it's an AR model since then you can rewrite the regression equation in terms of y-1. |
| sahilb7 | In II and III, y is dependent upon b1t which is also related to (t-1)... i.e. b1[(t-1)+1] |
| sahilb7 | II is a Linear trend model III is a Log-Linear trend model |