CEA Policy Evaluation & Economic Forecasting 3 — Questions and Answers
Question 1: The Congressional Budget Office (CBO) uses 'dynamic scoring' to:
- Adjust inflation projections using market expectations
- Incorporate macroeconomic feedback effects of tax changes into revenue estimates (Correct answer)
- Score regulatory costs separately from budget costs
- Discount future spending using the social discount rate
Correct answer: Incorporate macroeconomic feedback effects of tax changes into revenue estimates
Dynamic scoring accounts for how tax policy changes alter economic behavior (GDP, labor supply), affecting revenue beyond static calculation.
Question 2: In program evaluation, 'deadweight loss' refers to:
- Administrative costs that reduce program efficiency
- Benefits that would have occurred even without the program
- The net welfare loss from market distortions caused by the policy (Correct answer)
- Transfers that reduce recipient work effort
Correct answer: The net welfare loss from market distortions caused by the policy
Deadweight loss is the welfare cost of market inefficiency created by taxes or subsidies—value destroyed that neither party captures.
Question 3: A Vector Autoregression (VAR) model is particularly useful for economic forecasting because it:
- Requires only one endogenous variable
- Captures interdependencies among multiple economic variables simultaneously (Correct answer)
- Eliminates the need to specify causal direction
- Relies exclusively on theoretical priors
Correct answer: Captures interdependencies among multiple economic variables simultaneously
VAR models jointly forecast multiple variables by allowing each to depend on lagged values of all variables in the system, capturing dynamic feedback.
Question 4: Which method best handles the selection bias problem when evaluating a voluntary job-training program?
- Ordinary least squares regression on participants
- Randomized control trial or instrumental variable estimation (Correct answer)
- Before-after comparison for participants only
- Cross-sectional comparison at program end
Correct answer: Randomized control trial or instrumental variable estimation
Selection bias arises because participants self-select; RCTs or IV estimation isolate causal effects by removing the correlation between treatment and unobserved characteristics.
Question 5: A high fiscal multiplier for government spending implies that:
- Government spending crowds out private investment completely
- An increase in spending generates a proportionally larger increase in GDP (Correct answer)
- Spending has no effect on employment
- Tax cuts are more effective than spending
Correct answer: An increase in spending generates a proportionally larger increase in GDP
A multiplier above 1 means the GDP increase exceeds the initial spending injection, typically because of induced rounds of consumption spending.
Question 6: In Bayesian forecasting, the 'prior distribution' represents:
- The distribution of forecast errors in past models
- Beliefs about parameters before observing new data (Correct answer)
- The posterior estimate after model updating
- The likelihood function of the observed data
Correct answer: Beliefs about parameters before observing new data
The prior encodes what is known or believed about parameters before incorporating new evidence, which is then updated via Bayes' theorem.
Question 7: The 'policy lag' that refers to the time between recognizing a problem and implementing a policy response is called the:
- Outside lag
- Impact lag
- Inside lag (Correct answer)
- Transmission lag
Correct answer: Inside lag
The inside lag covers recognition, decision, and implementation delays within government before a policy is enacted.
CEA Policy Evaluation & Economic Forecasting 3
The Congressional Budget Office (CBO) uses 'dynamic scoring' to: