ETC3450 Applied Time Series Econometrics
The Complete Exam Bible · S2 2026ETC3450 Overview
- Monash University
- Semester Two offering
- Applied econometric modelling
- Raw mathematical evidence
- 5 concept chapters
Applied Time Series Econometrics treats ordering as information. A series is not an unordered sample with dates attached: dependence across lags changes the objects that must be modelled, the meaning of uncertainty and the forecasts available at each horizon.
- Index before modelling Fix the observation interval, frequency and information set first.
- Derive the moments Use recursion to distinguish conditional from unconditional behaviour.
- Diagnose persistence Read ACF and PACF patterns with stationarity conditions.
- Reconcile the forecast Check horizon, uncertainty and calibration alongside point accuracy.
How ETC3450 is assessed
| Component | Weight | Format |
|---|---|---|
| Presentation | 10% | Current Moodle assessment section; linked to the group-assignment material |
| Written | 30% | Current Moodle assessment section; linked to the individual assignment |
| Examination | 60% | End-of-semester examination |
The live Moodle assessment sections label the components Presentation (10%), Written (30%) and Examination (60%). The Unit information summary instead calls the 10% item Assignment 1 (individual) and the 30% item Assignment 2 (group), reversing the apparent individual/group mapping. Keep both readings visible and confirm the current task mapping on Moodle. The hurdle section directs students to the Handbook; confirm its current rule there.
Applied Time Series Econometrics assessment structure
Use the published weights as a planning map; the current learning site controls instructions, submission settings and any stated pass condition.
What ETC3450 covers
The sequence moves from time indexing and stochastic-process notation through stationarity, random walks and autocorrelation to autoregressive forecasting and lag-polynomial roots.
Time-Ordered Data and Stochastic Processes
Define the index and sampling rule before treating observations as a series02Moments, Dependence and Stationarity
Keep expected value, variance and covariance attached to their time indexes03White Noise, Random Walks and Differencing
Combine zero mean, finite constant variance and zero nonzero-lag covariance04Autoregression, Moments and Forecasts
Separate intercept, persistence and innovation before deriving moments05Higher-Order Dynamics and Lag Polynomials
Read each coefficient inside a joint dynamic systemThe current detailed material begins with stochastic processes, moments, conditional expectations and stationarity. It then develops white noise and random walks, demonstrates why accumulated shocks produce growing variance, uses differencing to transform persistent levels, and introduces autocovariance and autocorrelation as dependence summaries.
Autoregressive models connect those foundations to conditional distributions, long-run moments, forecast error, calibration, ACF and PACF identification. Higher-order dynamics are organised with lag polynomials and root conditions. Every formula in this guide is tied to intact lecture or solution evidence, and calculations are independently checked.
The live assessment labels conflict with the prose summary about which assignment is individual or group, so the guide preserves the published weights while directing students to Moodle for the current mapping. For Semester 2, 2026, Applied Time Series Econometrics at Monash University publishes this assessment map: Presentation (10%); Written (30%); Examination (60%).
The live Moodle assessment sections label the components Presentation (10%), Written (30%) and Examination (60%). The Unit information summary instead calls the 10% item Assignment 1 (individual) and the 30% item Assignment 2 (group), reversing the apparent individual/group mapping. Keep both readings visible and confirm the current task mapping on Moodle.
The hurdle section directs students to the Handbook; confirm its current rule there. The sequence moves from time indexing and stochastic-process notation through stationarity, random walks and autocorrelation to autoregressive forecasting and lag-polynomial roots. A time series records the same kind of quantity at ordered points.
Frequency may be hourly, monthly, quarterly or irregular, but the ordering permits dependence, trend, seasonality and response delays. Shuffling preserves the marginal values while destroying the lag relationships that make dynamic modelling possible. Check for duplicate times, gaps, irregular spacing and changes in measurement before computing correlations.
A lag means a number of index steps; its calendar duration changes when frequency changes. The mean locates the process, variance measures dispersion at a time, and autocovariance connects observations separated by a lag. Time indexes cannot be dropped casually: a nonstationary process can have a mean or variance that changes with the date, making one global summary misleading. Write the pair of times behind every covariance.
Before interpreting a lag statistic, check that the relevant moments exist and that stationarity assumptions justify pooling across dates. Under the standard white-noise conditions, past values do not alter the conditional mean, which remains zero. The process is stationary and provides the innovation building block for richer models.
Uncorrelated does not always mean independent unless stronger distributional assumptions are supplied. Check the residual plot and ACF as a pattern. Systematic clusters, decaying correlations or changing spread challenge white noise more strongly than a lone fluctuation. The AR(1) model combines an intercept, a lagged observation and a new innovation.
Conditional on current information, the next mean is the intercept plus the autoregressive coefficient times the latest value. The innovation variance is conditional forecast uncertainty for one step, not the unconditional variance of the process. Derive the long-run mean by equating expected values only after asserting stationarity.
For variance, keep innovation independence explicit and reject a denominator that is nonpositive. An AR(p) model relates the current value to several past values and an innovation. Coefficients describe a conditional recursion holding the other lags in the equation. Dynamic response emerges from their combined roots, so signs and magnitudes should not be interpreted one at a time without the system.
Write the full lag vector, derive the implied characteristic polynomial and inspect residual dependence after estimation. Compare candidate orders out of sample rather than selecting the best in-sample fit automatically.
Before interpreting the result from Lag-polynomial roots decide stability
- 1Write the process, information set and maintained parameter restrictions.
- 2Derive the required moment, dynamic response or forecast before substitution.
- 2Use residual or limiting behaviour to test the specification.
- 1Interpret the result and state which assumption would invalidate it.
Key terms
- Time Index
- Time Index — A time series records the same kind of quantity at ordered points. Frequency may be hourly, monthly, quarterly or irregular, but the ordering permits dependence, trend, seasonality and response delays. Shuffling preserves the marginal values while destroying the lag relationships that make dynamic modelling possible.
- Frequency
- Frequency — A monthly average cannot identify hourly peaks, and an annual series cannot reveal within-year seasonality. Aggregation can smooth volatility or change apparent dependence. State the observational clock, timestamp convention and missing-time rule before fitting a stochastic process.
- Lag
- Lag — Check for duplicate times, gaps, irregular spacing and changes in measurement before computing correlations. A lag means a number of index steps; its calendar duration changes when frequency changes.
- Realisation
- Realisation — The observed path is a finite realisation. A stochastic process is a collection of random variables indexed by time and describes possible paths through its joint distribution. Data summaries describe what occurred; model statements encode assumptions about what could occur and how observations depend.
- Stochastic Process
- Stochastic Process — A smooth plot does not prove a deterministic trend, and a model equation does not guarantee fit. Econometric reasoning moves from pattern to candidate process, derives implications, then returns to data for diagnostics and comparison.
- Implication
- Implication — Write model implications before examining the matching statistic. This prevents an attractive plot from deciding the model. Record which discrepancies would reject or revise the candidate.
- Information Set
- Information Set — A forecast of the next observation is conditional on an information set available at the forecast origin. Including the target or later values creates look-ahead bias. The notation should make timing explicit so the conditional mean and forecast error are defined against the same information.
- Forecast Origin
- Forecast Origin — White noise has the same conditional and unconditional mean under its defining assumptions, so past values do not improve the point forecast. Autoregression changes that result because the latest observation enters the conditional expectation.
- Look-ahead Bias
- Look-ahead Bias — Label origin, horizon and target beside every forecast. Then verify each predictor's timestamp. A clean formula cannot rescue an invalid information set.
- Time Plot
- Time Plot — A time plot can reveal trend, changing variance, breaks, seasonality, outliers and persistence. These features suggest transformations and model families. They do not uniquely identify a process, because different mechanisms can produce visually similar finite paths.
ETC3450 FAQ
What is gained by treating a time series as a stochastic process?
Pedestrian counts sampled hourly reveal a daily rhythm that vanishes in a daily total. Both datasets describe traffic, yet they support different forecasts and interventions because their indexes retain different temporal information. Check for duplicate times, gaps, irregular spacing and changes in measurement before computing correlations.
A lag means a number of index steps; its calendar duration changes when frequency changes. The algebra, parameter restriction and economic interpretation should agree; a contradiction among them is a specification warning.
When does covariance stationarity make lag relationships reusable?
One random-walk realisation may drift upward for most of the sample even though each innovation has zero mean. Reading the path as evidence of a positive deterministic trend confuses realised accumulation with the process expectation. Write model implications before examining the matching statistic. This prevents an attractive plot from deciding the model. Record which discrepancies would reject or revise the candidate.
State whether uncertainty concerns parameters, innovations or model choice because each source changes the forecast differently.
Why does a random walk resist a fixed long-run mean?
A backtest accidentally constructs a lag after sorting data in reverse order. The apparent predictor contains future values and produces unrealistically low error. Restoring chronological order removes the artificial performance. Label origin, horizon and target beside every forecast. Then verify each predictor's timestamp. A clean formula cannot rescue an invalid information set.
A residual plot is part of the argument when it tests assumptions that the reported coefficient or forecast requires.
How should an autoregressive parameter change a shock interpretation?
A price index rises steadily in levels while its first differences fluctuate around a stable mean. The two plots support different modelling objects: a persistent level and a potentially stationary change. Pair visual diagnosis with a stated implication and a statistic able to challenge it. Mark unusual periods instead of deleting them silently, since an intervention or measurement change may explain the break.
The algebra, parameter restriction and economic interpretation should agree; a contradiction among them is a specification warning.
Which residual pattern would challenge a fitted dynamic specification?
Two series share the same marginal mean and variance, yet one alternates signs and the other forms long runs. Their autocovariances expose dynamics that unordered summaries cannot see. Write the pair of times behind every covariance. Before interpreting a lag statistic, check that the relevant moments exist and that stationarity assumptions justify pooling across dates.
State whether uncertainty concerns parameters, innovations or model choice because each source changes the forecast differently.
Why must forecast uncertainty be separated from the point projection?
An AR process with a coefficient close to one may appear to wander for a finite sample although it is stationary. A random walk may temporarily oscillate near zero although its variance grows with time. Test the stated model conditions before judging the plot. If moments depend on initial conditions or time, say which stationarity requirement fails rather than using 'nonstationary' as an unexplained label.
A residual plot is part of the argument when it tests assumptions that the reported coefficient or forecast requires.
How should the conflicting assignment labels be handled?
For a stationary AR(1), the next expected value depends on the current observation, while the long-run mean is constant. Forecasting every date with the long-run mean ignores useful lag information even though the mean itself is correct. Write conditioning bars and information sets until the distinction becomes automatic. Compare a forecast error with conditional variance, not with the wider unconditional variance.
The algebra, parameter restriction and economic interpretation should agree; a contradiction among them is a specification warning.
How to study for the exam
Work time-series problems in an invariant order. First write the model, index and information set. Second state the assumptions on the innovation and decide whether the requested object is conditional or unconditional. Third derive symbolically before substituting numbers: recursive substitution often reveals the mean, variance accumulation and forecast horizon more clearly than a memorised formula.
Fourth inspect the stationarity condition. If it fails, do not quote a long-run moment that does not exist. Fifth compare the theoretical pattern with a plot, ACF or PACF, remembering that sample correlations fluctuate. For forecasts, keep the point prediction, forecast-error variance, interval and horizon on separate lines. Reinsert special parameter values to recover white noise or a random walk as a check.
Maintain a small error log for wrong lag, wrong information set, confusing innovation variance with process variance, and treating calibration as accuracy. Those mistakes have different repairs. Confirm the current individual/group assessment mapping on Moodle because the two available descriptions conflict.
