Return to Ken Judd’s Home Page | Return to Ken Judd’s Penn State Numerical Methods Course Home Page
March 16 – March 27, 2026
Dates, Times and Locations
I will be visiting Penn State March 16 – March 27, 2026. My lecture schedule is
Tuesday, March 17, 6pm-7:30pm in 107 Chambers
Thursday, March 19, 6pm-7:30pm in 107 Chambers
Tuesday, March 24, 6pm-7:30pm in 107 Chambers
Thursday, March 26, 6pm-7:30pm in 105 Chambers
Office
I am in room 508.
Overview of Lectures
Tuesday, March 17: Optimization and Approximation
Discussion of unconstrained optimization methods used for maximum likelihood and moment estimators:
A systematic approach of how to proceed when you need to find an optimum and don’t have a good initial guess.
Thursday, March 19: Structural estimation, NFXP, MPEC and Surrogate methods:
Discussion of efficient methods for NFXP.
Introduce mathematical programming with equilibrium constraints (MPEC) and compare the two approaches.
Define the surrogate method for approximation and optimization
Tuesday, March 24: BLP models
The advantages of MPEC (Dube, Fox and Su).
The importance of quadrature rules (Rangel and Judd).
The Judd-Rangel generalization of BLP
Thursday, March 26: The Evolution of Artificial Intelligence and Machine Learning
Lecture Details
Below are the topics I will cover. The links below take you to the papers related to the topics. I certainly cannot cover all of them in detail. I have included links presentations by Che-Lin Su and others which do present the details.
These topic descriptions are incomplete. I will add slides and code soon. I am also flexible. I can adjust my lectures to suit student interest.
1: Tuesday, March 17: Practical Ideas for Unconstrained Optimization
This is a basic lecture describing unconstrained optimization methods such as maximum likelihood methods and moment estimators. Many people write down their problem and then send it to their optimizer, only to find that things don’t work. I will focus on giving you a practical approach to your estimation problems:
Step one: write the code for computing the objective. Thoroughly test it before sending it to a solver
Step two: Use a sequence of optimization methods, starting with simple search methods, then derivative-free methods, then gradient methods and then use Newton-style state-of-the-art methods
Approximation problems are examples of optimization problems and can be used, via surrogate methods, to solve optimization problems.
Computer power is now based on parallelism. Use methods that exploit parallelism.
2: Thursday, March 19: Structural Estimation: NFXP vs MPEC vs Surrogate
We first compare nested methods to basic constrained optimization.
Nested Methods and a Superior Alternative
Evolution of Su-Judd MPEC approach
The SJ papers are versions of the Su-Judd paper on MPEC, a small piece of which was published in Econometrica as a note. Here are the four versions of Su-Judd, leading to the 2012 Econometrica paper.
2008 SJ Original submission to ECTA
2010 SJ First revision for ECTA
Che-Lin Su Presentations
Surrogate Method
Markus Trunschke, Gregor Reich and I have applied the surrogate method to the Zurcher bus model.
Rust comment:
In the 2014 and 2015 versions, Rust et al. compares NFXP and MPEC. They argue that NFXP is more stable and faster than MPEC. However, Che-Lin Su showed that the comparisons were not valid.
Rust Iskhakov Schjerning June 2014
3: Tuesday, March 24: BLP
The BLP model is widely used in empirical IO. However, the standard methods are not reliable and make unnecessary simplifications.
Stopping Rules
Nevo’s A Research Assistant’s Guide to Random Coefficients Discrete Choice Models of Demand describes the BLP model and a commonly used computational method. Che-Lin Su and I looked at Nevo’s code and found some obvious weaknesses. In particular, the stopping rules were too loose. I had Che-Lin use MPEC to solve the Nevo problem. He found a better value for the objective and was happy because he was trained as a computational mathematician. I told him that was not enough, and that he should check to see if Nevo’s estimate was statistically different than his result. It was. I decided my job was done. Che-Lin went on to write a paper with Dube and Fox.
AMPL code for MPEC applied to BLP.
Dubé, Fox, & Su (2012) compares Nevo to an MPEC approach
Problems with Monte Carlo Integration
Skrainka-Judd (2012) and Rangel-Judd (2025) shows that Monte Carlo quadrature methods cause serious for estimating BLP models, but also shows that numerical quadrature formulas are far more accurate and reliable. Essentially, Monte Carlos methods have a tendence to underestimate standard errors. Using Monte Carlo integration helps with publishing papers because (1) you are more likely to get identification, and (2) you can try different seeds for your RNG and use the one that gives you results the editor will like. Editors will likely detect “regression fishing” but I doubt that they are as good at detecting “seed fishing”.
Portfolio problem PSU 2019 nb.pdf
The Judd-Rangel generalization of BLP
The standard description of BLP has always confused me. Carlos Rangel and I have worked on a cleaner presentation that makes clear the computational details. It also generalizes the BLP model
4: Thursday, March 26: Evolution of AI and Machine Learning
There is a lot of discussion of AI and machine learning, and their application to economics. Standard descriptions, much of which is best described as propaganda, make this all sound magical. In fact, neural networks are nothing more than nonlinear least squares fitting, a fact that our late colleague Ron Gallant knew and wrote about over 30 years ago. I will describe the evolution of AI and machine learning, and the connections with numerical methods and econometrics.
Office Hours:
I will also be available for office visits. I ask that you first send me something written that describes the computational aspects of what you are working on or want to work on. Do not assume that I have read the papers related to your research. You should be able to explain the mathematical and computational structure of your problems in a way that I can understand without knowing published economic papers (which are often poorly written).
Basically, the part of your research I can help with is the computational methods you use. Feel free to send me your writeup NOW so that I will be acquainted with your work when we meet.
Looking forward to seeing you soon.