You are here


The program's format is 100% virtual.

Day 1 - Monday, August 9, 2021

10am - noon: Introduction

Speaker: Toni Whited, University of Michigan


- What is structural estimation?
- What does structural estimation buy you? When is it needed?
- Motivating a structural estimation paper

1pm - 3pm:  Solving Models (Part 1) 

Speaker:  Stephen Terry, Boston University


- Discrete-time dynamic programming
- Value function iteration
- Simulation

3pm - 3:10pm:   Afternoon Break

3:10pm - 6:30pm:  Project Part 1

Participants will work on the first part of the project. The goal for day one is to numerically solve the model and describe the optimal policies. Faculty and assistants will be on hand to help.

8:30pm - late:  More work on the project + refreshments

Day 2 - Tuesday, August 10, 2021

9:30am - 11:30am:  Solving Models (Part 2)

        Speaker:  Stephen Terry, Boston University


- Advanced topics on dynamic programming
- Efficient computing

11:30am - noon: Project check-in

Speaker:  Stephen Terry, Boston University

We will use these check-in sessions to present a solution to the previous day’s work, discuss challenges people had, and show some tricks of the trade. We will also use these sessions to present a set of solution Matlab code. By the end of the week, participants will have not only their own code but also this set of solution code that can be adapted for future projects.

1pm - 3pm:  Simulated Method of Moments Estimation (Part 1)

Speaker: Toni Whited


- SMM algorithm
- Assessing model fit
- Heterogeneity
- Searching the parameter space
- Identification challenges

3pm - 3:10pm:  Afternoon Break

3:10pm - 6:30pm: Project Part 2

The goal for day two is to develop an identification strategy, choose moments, and write code for computing the moments in both the actual and simulated data.

8:30pm - late:  More work on project 


Day 3 - Wednesday, August 11, 2021

9:30am - 10:00am:  Project check-in

Speaker: Luke Taylor

10am - noon:  Simulated Method of Moments Estimation (Part 2)

Speaker: Toni Whited

Topics / Goals

- Practical advice
- Introduction to weight matrices
- Influence functions
- Getting the standard errors right

1pm - 3 pm:  Simulated Method of Moments Estimation (Part 3)

Speaker: Toni Whited

Topics / Goals

- Discussion of applied papers

3pm - 3:10pm:  Afternoon Break

3:10pm - 6:30pm:  Project Part 3

The goal for day three will be to estimate the covariance matrix of empirical moments using influence functions, compute the optimal weight matrix, and start coding up the SMM estimator.

8:30pm - Late:  work on project

Day 4 - Thursday, August 12, 2021

9:30am - 11:30am: Literature Review

Speaker:  Luke Taylor

We will start with a broad review of the literature, and then we'll do an in-depth discussion of a few papers. The goals will be to understand what has already been done, why structural estimation is needed in these papers, and how to best position a structural estimation paper.

11:30am - 12:30pm: Project wrap up

Speaker:  Luke Taylor

1:30pm - 4:30pm:  Project part 4

Goals: Finish coding up the SMM estimator, using simulated annealing to search the parameter space. Also code of the test of over-identifying restrictions and counterfactual analyses.