Week 10 Homework Feedback: Omadbek Rasuljonov
Week 10 Homework Feedback: Omadbek Rasuljonov
Overall Assessment
Grade: ✅ (Pass)
Correct submission. You implemented the EGM method correctly and completed the required tasks.
Task-by-Task Check
Task 1: Quantifying Precautionary Savings
- Parameter & Grid Setup: ✅ Correct.
- Solution Method: ✅ EGM implemented.
- Deterministic Case: ✅ Solved.
- Stochastic Case: ✅ Solved.
- Simulation: ⚠️ You used the mean of the policy function (
mean(k_pol_det(:))) as a proxy for the long-run mean capital (Line 21-22). This is an approximation but strictly speaking incorrect; you should simulate the time series path of the economy to find the ergodic mean, as the distribution of capital is not uniform over the grid. - Comparison: ✅ Premium calculated (based on the proxy mean).
Task 2: The Role of Risk Aversion
- Sigma Variation: ✅ Implemented.
- Solving Sigma Cases: ✅ Solved.
- Policy Plot: ✅ Plotted.
- Simulation Comparison: ⚠️ Simulation missing (implied by policy plot, but no time series generated).
- Interpretation: ✅ Correct intuition.
Technical Implementation
- EGM Logic: Your
solve_egmfunction (which I assume handles the logic) seems to work based on the results. - Simulation Issue: Averaging the policy function grid values does not give the economic steady state. The steady state is where $k’ \approx k$ (or the distribution around it). The grid includes values far from steady state (e.g., $k=0.1$ or $k=15$), so averaging them biases the result.
Visual Output
- Policy function plots are clear.
Suggestions for Improvement
- Simulation: Always run a time-series simulation (e.g., $T=1000$ periods) using the policy function to calculate long-run statistics.
- Burn-in: Remember to discard the first ~100 periods of the simulation to remove the effect of the initial condition.
Summary
9/11 tasks correct. Good effort on EGM implementation, but the simulation part was simplified too much.