CEA Economic Forecasting Techniques 2 — Questions and Answers
Question 1: Which forecasting method uses a system of simultaneous equations to model the interdependencies among multiple economic variables?
- Vector Autoregression (VAR)
- Simultaneous Equation Model (SEM) (Correct answer)
- ARIMA
- Exponential Smoothing
Correct answer: Simultaneous Equation Model (SEM)
Simultaneous Equation Models capture bidirectional relationships among variables using a system of equations estimated jointly.
Question 2: In the context of economic forecasting, what does 'nowcasting' refer to?
- Forecasting 30 years into the future
- Estimating the current state of the economy using real-time data (Correct answer)
- Backcasting historical data gaps
- Predicting next quarter's GDP only
Correct answer: Estimating the current state of the economy using real-time data
Nowcasting uses high-frequency real-time data (e.g., weekly jobless claims) to estimate economic conditions for the current or very recent period.
Question 3: The Kalman filter is primarily used in economic forecasting to:
- Remove seasonal patterns from time series
- Estimate unobserved state variables from noisy observations (Correct answer)
- Test for unit roots in panel data
- Identify cointegration among non-stationary series
Correct answer: Estimate unobserved state variables from noisy observations
The Kalman filter recursively updates estimates of hidden state variables (e.g., potential output) as new data arrives.
Question 4: A forecast that minimizes the sum of squared errors is said to be optimal under which loss function?
- Absolute error loss
- Asymmetric LINEX loss
- Quadratic (squared error) loss (Correct answer)
- Zero-one loss
Correct answer: Quadratic (squared error) loss
Quadratic loss penalizes large errors heavily and its minimization yields the conditional mean as the optimal point forecast.
Question 5: When combining forecasts from multiple models, research consistently shows that:
- The most complex model always dominates
- Simple averages of forecasts often outperform individual models (Correct answer)
- Only Bayesian model averaging is valid
- Combination works only for short-horizon forecasts
Correct answer: Simple averages of forecasts often outperform individual models
Forecast combination exploits model diversity and reduces variance, frequently beating any single model including the best ex-post model.
Question 6: Which technique decomposes a time series into trend, seasonal, and irregular components using a multiplicative or additive framework?
- Hodrick-Prescott filter
- Census X-13ARIMA-SEATS (Correct answer)
- Granger causality test
- Phillips curve estimation
Correct answer: Census X-13ARIMA-SEATS
X-13ARIMA-SEATS, developed by the US Census Bureau, is the standard seasonal adjustment and decomposition tool used by statistical agencies.
Question 7: An economic forecaster uses a rolling window of 60 months to re-estimate a model as each new month of data arrives. This approach is called:
- Recursive estimation
- Rolling window estimation (Correct answer)
- Expanding window estimation
- Cross-sectional pooling
Correct answer: Rolling window estimation
Rolling window estimation keeps the sample size fixed, allowing model parameters to evolve over time and accommodating structural change.
CEA Economic Forecasting Techniques 2
Which forecasting method uses a system of simultaneous equations to model the interdependencies among multiple economic variables?