Macroeconomic Modelling and Forecasting Using EViews Training Course

Participants can import, clean and document macroeconomic data for econometric modelling.

• Importing spreadsheet, CSV and database extracts into structured EViews workfiles
• Frequency conversion using sum, average, end-of-period and interpolation methods
• Missing observations, outliers and revisions identified before model estimation
• Seasonal adjustment sources and treatment documented for official macroeconomic series
• Transforming levels into logs, growth rates, differences and index numbers
• Data dictionaries record definitions, sources, release dates and transformations

Participants can assess data patterns and form economically plausible model hypotheses.

• Time plots reveal trends, cycles, volatility clusters and potential breaks
• Correlation matrices distinguish contemporaneous association from causal economic relationships
• Scatterplots and lag plots examine plausible transmission delays
• National accounts identities checked against forecast variables and data definitions
• Economic priors stated before selecting regressors and lag structures
• Descriptive statistics identify skewness, dispersion and extreme observations

Participants can test time-series properties and select transformations suitable for modelling.

• Unit roots tested using augmented Dickey-Fuller and Phillips-Perron procedures
• KPSS tests used to examine stationarity under an alternative null hypothesis
• Deterministic trends, intercepts and seasonal dummies specified appropriately
• Seasonal unit roots considered for monthly and quarterly macroeconomic series
• Chow tests identify known policy changes and sample breaks
• Bai-Perron procedures assess multiple structural breaks within long samples

Participants can estimate and interpret theory-led single-equation macroeconomic regressions.

• Ordinary least squares equations specified from economic behavioural relationships
• Coefficient signs, elasticities and semi-elasticities interpreted in economic terms
• Lagged dependent variables model persistence in inflation and output
• Distributed lag specifications capture delayed policy and demand effects
• Dummy variables represent reforms, crises and temporary administrative shocks
• Adjusted R-squared and information criteria support specification comparison

Participants can diagnose common regression failures and revise model specifications responsibly.

• Residual plots assess bias, changing variance and unmodelled temporal patterns
• Breusch-Godfrey tests detect serial correlation in regression residuals
• White and Breusch-Pagan tests assess heteroskedasticity risks
• Variance inflation factors identify harmful multicollinearity among predictors
• Jarque-Bera tests examine residual normality for interval interpretation
• Ramsey RESET tests flag functional-form and omitted-variable concerns

Participants can build and validate ARIMA models for a single macroeconomic series.

• Autocorrelation and partial autocorrelation functions identify candidate ARIMA orders
• Differencing achieves stationarity while preserving interpretable forecast variables
• Seasonal ARIMA terms model recurring monthly and quarterly patterns
• Maximum likelihood estimation produces ARIMA parameter estimates in EViews
• AIC, BIC and residual autocorrelation guide model selection
• Dynamic forecasts separate estimated history from future projected observations

Participants can assess whether a forecasting model performs adequately against alternatives.

• In-sample fit distinguished from genuine out-of-sample forecast performance
• Holdout samples simulate forecast conditions using historical data vintages
• RMSE, MAE and MAPE compare competing forecast specifications
• Theil inequality coefficient decomposes bias, variance and covariance errors
• Forecast encompassing tests assess incremental value from competing models
• Benchmark comparisons include random walk, naïve and consensus forecasts

Participants can estimate VAR models and interpret dynamic interactions among macroeconomic variables.

• VAR systems model jointly determined macroeconomic variables without restrictive ordering
• Lag length selected using AIC, Schwarz and Hannan-Quinn criteria
• Stability conditions checked through inverse roots of characteristic polynomials
• Granger causality tests examine predictive content between series
• Impulse response functions trace shocks through the estimated system
• Forecast error variance decomposition attributes uncertainty across shocks

Participants can apply defensible identification assumptions when analysing macroeconomic shocks.

• Cholesky ordering imposes recursive contemporaneous relationships among VAR innovations
• Economic theory guides ordering of output, prices and policy rates
• Structural VAR restrictions distinguish monetary, demand and supply shocks
• Sign restrictions encode economically defensible shock-response patterns
• Policy shock responses interpreted with confidence intervals and transmission lags
• Identification assumptions documented for review by economic and risk stakeholders

Participants can model long-run equilibrium relationships and short-run adjustment using VECMs.

• Cointegration distinguishes common long-run movements from spurious regression relationships
• Johansen trace and maximum-eigenvalue tests determine cointegration rank
• Restricted deterministic terms selected for trend and intercept treatment
• Vector error-correction models combine equilibrium correction with short-run dynamics
• Adjustment coefficients quantify speed of return toward long-run equilibrium
• Long-run restrictions tested against economic theory and institutional knowledge

Participants can estimate ARDL models for mixed-order macroeconomic variables and derive long-run effects.

• ARDL specifications accommodate regressors integrated at level zero or one
• Bounds testing assesses long-run relationships within an autoregressive distributed-lag framework
• Lag orders selected using information criteria and economic timing
• Error-correction terms measure disequilibrium adjustment after macroeconomic shocks
• Long-run multipliers calculated from estimated ARDL coefficient restrictions
• Dynamic multipliers illustrate adjustment paths following explanatory-variable changes

Participants can develop an inflation forecasting model using relevant monetary and real-economy indicators.

• Inflation measures compared using headline, core, trimmed and administered-price concepts
• Output gaps estimated using trends, filters and capacity indicators
• Exchange-rate pass-through modelled through imported inflation channels
• Policy rates, money growth and credit conditions entered with appropriate lags
• Phillips curve specifications tested against inflation persistence and expectations
• Fan charts communicate inflation uncertainty around central forecast paths

Participants can model external-sector variables and incorporate global assumptions into forecasts.

• Exchange-rate transformations distinguish bilateral, nominal and real effective measures
• Terms-of-trade shocks linked to export prices, imports and fiscal revenues
• Commodity-price assumptions translated into sectoral and balance-of-payments effects
• Interest-rate differentials considered within exchange-rate forecasting equations
• External demand indicators connect trading-partner growth to domestic exports
• Foreign-currency series aligned for timing, units and quotation conventions

Participants can produce scenario-based forecasts and explain their assumptions, ranges and limitations.

• Scenario narratives translate economic shocks into quantified model assumptions
• Baseline assumptions documented for global growth, commodities and policy settings
• Alternative paths constructed for exchange rates, rates and fiscal measures
• Stochastic simulation generates forecast distributions from estimated equation uncertainty
• Sensitivity tables identify assumptions with material effects on forecast targets
• Executive forecast notes separate evidence, judgement, uncertainty and recommended actions

Participants can build, review and present a complete EViews macroeconomic forecasting model with documented scenarios and forecast recommendations.

• Forecast question, target variables and decision horizon defined clearly
• Data workfile assembled with documented sources, transformations and release dates
• Model specification selected using economic theory and diagnostic evidence
• Baseline forecast generated with intervals, assumptions and key model outputs
• Downside scenario quantified through explicit shock paths and transmission channels
• Peer review tests reproducibility, plausibility, diagnostics and communication quality