New mini-series on ML and Econometrics!
Covers ML basics, double/debiased ML, post-double-selection LASSO, and other applications. Main focus is on preserving appropriate econometric inference when utilizing ML.
9-videos playlist here:
https://t.co/5qT4OH3sxW
New mini-series on sensitivity analysis and partial identification! Covers limits of IV and how to bound causal effects in OLS regression models. Very useful; give it a look if you're interested.
https://t.co/BFH5BZBYCp
Brand-new mini-series: this is an 8-parter on learning models in economics.
Topics include:
- What is learning
- What is Bayesian updating
- Generalizations beyond Bayesian updating
- Multidimensional learning
- Recent empirical papers about learning
https://t.co/Il9PN9TfUH
New 8-part YouTube series on factor models & measurement error is now posted! Covers PCA, factor analysis, identification, maximum likelihood estimation, and Heckman's work on cognitive/non-cognitive skills + frontier papers. Check it out!
https://t.co/hVIdoMslNZ
Just dropped a new 4-part video series covering model assessment and counterfactual simulation. Covers model fit, validation, and counterfactual simulation.
Playlist here: https://t.co/lQgnE0NbW4
New mini-series on Conditional Choice Probability (CCP) estimation for dynamic discrete choice models! Covers Hotz-Miller inversion, renewal actions, finite dependence, unobserved heterogeneity, and more. Watch here: https://t.co/HyJE8wUHWV
Finished another video series!
This one: dynamic discrete choice models and optimal stopping.
Bellman equations, backward recursion, value function iteration, and why John Rust's 1987 bus paper is a masterpiece.
Watch here: https://t.co/stDyEG1uoe
Another mini-series is live! This one is on mixed logit & preference heterogeneity.
It covers:
• random coefficients
• identification
• simulation methods (quadrature vs MC)
• finite mixtures
• EM/MM algorithms
• practical implementation
Enjoy!
https://t.co/oGPGY8gYkh
Another mini-series is live! This one is on mixed logit & preference heterogeneity.
It covers:
• random coefficients
• identification
• simulation methods (quadrature vs MC)
• finite mixtures
• EM/MM algorithms
• practical implementation
Enjoy!
https://t.co/oGPGY8gYkh
New 7-part series on nested logit models! It covers the full pipeline from GEV error structures through estimation and real-world applications. Ideal if you've always wondered about nested logit but felt too intimidated to jump in!
https://t.co/Aneue3ieUq
I just finished an 8-part YouTube series on discrete choice models! It covers the full pipeline from random utility → logit derivation → MLE → welfare analysis.
Each video builds on the last with both math and intuition.
Enjoy!
https://t.co/hMUCFrwfKj
I just posted a new 8-video YT playlist covering the process of producing an empirical structural model. In each video, I go through a (contrived) example of each step in hopes of elucidating the process.
https://t.co/cVC0est14W
I had a lot of fun with my latest video tracing the structural vs. reduced form "war" in applied microeconomics! (credit: @causalinf's history series).
Check out how the war got started, who the key figures were, and why I believe it's (mostly) over:
https://t.co/YDaHHGt75M
@joorosa12185462@causalinf I've collected my resources from many different sources. Mostly lecture notes from leading practitioners. These videos are serving as "my textbook"
The reduced form vs. structural debate is a non-starter.
A school voucher study in India shows why: neither method alone could answer the key policy question.
When researchers combined both approaches, they found *new* mechanisms!
Video: https://t.co/PpAsIORLBP
My latest video explores a paradox of empirical economics: "credibility" ultimately rests on untestable assumptions.
Parallel pre-trends ≠ proof of parallel post-trends
Model fit ≠ proof of correct specification
New video exploring why 👇
https://t.co/2HVPt6Vcjz
🚀 What do Galileo, Newton & Einstein teach us about economics?
Galileo experimented, Newton modeled, and Einstein refined.
Economics evolves the same way.
🎥 https://t.co/s7yyVgXdYZ
#Economics#Science#EmpiricalMethods
When should you use a structural model vs. a simple RCT or diff-in-diff? It depends on whether you're evaluating existing policies or designing new ones. New video breaking down the tradeoffs:
https://t.co/LdQ0J3uWWZ