Bryan S. Graham

Second CAMSE-CLIMB Mini-Conference

Thu, 23 Jan 2025 00:00:00 +0000

On Friday February 21st, 2025 the Center for the Application of Mathematics and Statistics to Economics (CAMSE) and the Center for the Theoretical Foundations of Learning, Inference, Information, Intelligence, Mathematics and Microeconomics at Berkeley (CLIMB) will host one day mini-conference. The goal is to gather campus researchers at the intersection of economics, machine learning and statistics. Attendence is open to anyone from the Berkeley data science communities (broadly and inclusively defined). Registration is not required. Just come!

The conference will be held in room 250 of Sutardja Dai Hall on the north side of the UC Berkeley campus (close to the North Gate of campus).

A preliminary conference program can be found below.

Second CAMSE-CLIMB Mini-Conference

Organizers
Bryan Graham
With special thanks to:
Naomi Yamasaki

Friday, February 21st, 2025

250 Sutardja Dai Hall

Morning Session: Students & Post-Docs Speakers, 9:40AM to 12:00PM

Time	Speaker	Title
9:40AM to 10:00AM	Sara Neff, UC - Berkeley, Economics	Model complexity and restrictiveness
10:00AM to 10:20AM	Lea Bottmer, Stanford, Economics	Synthetic control in disaggregated data settings
10:20AM to 10:40AM	Kunhe Yang, UC - Berkeley, Computer Science	Leakage-robust Bayesian persuasion
10:40AM to 11:00AM	Break
11:00AM to 11:20AM	Kaitlyn J. Lee, UC - Berkeley, Biostatistics	RieszBoost: gradient boosting for Riesz regression
11:20AM to 11:40AM	Amar Venugopal, Stanford, Economics	Causal inference on outcomes learned from text
11:40AM to 12:00PM	Jason Weitze, Stanford, Economics	A predictive approach to structural identification

Afternoon Session: Faculty Speakers, 2:00PM to 5:45PM

Time	Speaker	Title
		Session 1: Policy Analysis and Evaluation
2:00PM to 2:30PM	Guido Imbens, Stanford, GSB & Economics	Causal panel data models
2:30PM to 3:00PM	Quitze Valenzuela-Stookey, UC - Berkeley, Economics	Mechanism reform: an application to child welfare
3:00PM to 3:30PM	Break
		Session 2: Causal Inference and Networks
3:30PM to 4:00PM	Mengsi Gao, Berkeley, Econommics	Endogenous interference in randomized experiments
4:00PM to 4:30PM	Lihua Lei, Stanford, GSB	Causal clustering: design of cluster experiments under network interference
4:30PM to 4:45PM	Break
		Session 3: Distinguished Guest Speaker
4:45PM to 5:45PM	Elena Manresa, NYU, Economics	Adversarial Method of Moments

CAMSE-CLIMB Mini-Conference

Wed, 27 Mar 2024 00:00:00 +0000

On Thursday April 11th, 2024 the Center for the Application of Mathematics and Statistics to Economics (CAMSE) and the Center for the Theoretical Foundations of Learning, Inference, Information, Intelligence, Mathematics and Microeconomics at Berkeley (CLIMB) will host a one day mini-conference. The goal is to gather campus researchers at the intersection of economics, machine learning and statistics. Attendence is open to anyone from the Berkeley data science communities (broadly and inclusively defined). Registration is not required.

The conference will be held in room 250 of Sutardja Dai Hall on the north side of the UC Berkeley campus (close to the North Gate of campus).

A preliminary conference program can be found below.

CAMSE-CLIMB Mini-Conference

Organizers
Bryan Graham
With special thanks to:
Naomi Yamasaki
Michael Jordan

Thursday, April 11th, 2024

250 Sutardja Dai Hall

Morning Session: Students & Post-Docs Speakers, 9:40AM to 11:40AM

Time	Speaker	Title
9:40AM to 10:00AM	Serena Wang, UC - Berkeley, EECS	Information elicitation in agency games
10:00AM to 10:20AM	Dohyeong Ki, UC - Berkeley, Statistics	Totally convex regression
10:20AM to 10:40AM	Anand Kumar Siththaranjan, UC - Berkeley, EECS	When can communication be informative?
10:40AM to 11:00AM	Keaton Ellis, UC - Berkeley, Simons Institute	The predictivity of theories of choice under uncertainty
11:00AM to 11:20AM	Yixiang Luo, UC - Berkeley, Applied Mathematics	Estimating the FDR of variable selection
11:20AM to 11:40AM	Yassine Sbai-Sassi, UC - Berkeley, Economics	Average treatement effects for exchangeable random arrays

Afternoon Session: Faculty Speakers, 1:30PM to 6:30PM

Time	Speaker	Title
		Session 1: Labor Economics Seminar
1:30PM to 2:30PM	Leonard Goff, University of Calgary, Economics	Treatment effects in bunching designs: the impact of mandatory overtime pay on hours
2:30PM to 2:45PM	Break
		Session 2: CS-Econ-Stat @Cal
2:45PM to 3:30PM	Alejandro Schuler, UC - Berkeley, Biostatistics	Lassoed Tree Boosting
3:30PM to 4:15PM	Nika Haghtalab, UC - Berkeley, EECS	Collaborative machine learning: optimization and incentives
4:15PM to 5:00PM	Federico Echenique, UC - Berkeley, Economics	Stable matching as transportation
5:00PM to 5:30PM	Break
		Session 3: Econometrics Seminar
5:30PM to 6:30PM	Whitney Newey, MIT, Economics	Automatic Debiased Machine Learning via Riesz Regression

Conference in Honor of James L. Powell

Sun, 06 Mar 2022 00:00:00 +0000

On Friday and Saturday March 25th and 26th, 2022 the Center for the Application of Mathematics and Statistics to Economics (CAMSE) will host a conference in Honor of the legendary James L. Powell. Attendence is by invitation only.

The conference will be held in room B100 of the Blum Center. A map with the location of Blum Hall can be found here. Information on transit options from SFO and OAK airports can be found here.

Blum Hall is very close to the North Gate of campus. The nearest BART stop is Downtown Berkeley. The nearest public parking facility is the Lower Hearst / North Gate parking garage at the intersection of Scenic & Hearst (enter from Scenic to access hourly parking spaces). Blum Hall is also about a 10 to 15 minute walk (slightly uphill) from the conference hotel (the Residence Inn by Marriott Berkeley).

Instructions on how to join the conference by Zoom will be sent as we get closer the date.

Campus and the City of Berkeley COVID guidelines evolve regularly. As of March 7th, indoor masking by vaccinnated individuals is no longer required on campus (instead being strongly recommended). However a variety of other COVID related measures remain intact. More information on current campus COVID policies can be found here.

The plan is to set up, weather permitting, coffee breaks and lunch outdoors. Eating outside has a number of advantages, one of which is that formal checking of vaccination status is not required. However all attendees should nevertheless bring proof of vaccination as it will be required to attend the conference dinner (also UC Berkeley has a vaccine mandate, even for conference attendees, so its a good idea to have some sort of proof on hand).

The conference dinner will be held at Agrodolce Berkeley at 7PM (time to be confirmed). This event will be in a large indoor room. In addition to enjoying one anothers company, there will be an opportunity to share stories about, and expressions of gratitude for, Jim. A map with the location of the restuarant can be found here.

A preliminary conference program can be found below (subject to change). Speakers should plan on presenting for approximately 20 minutes, allowing for five minutes of discussion and questions, and an additional five minutes of transition time between speakers.

CAMSE Conference in Honor of James L. Powell Schedule

Organizers
Bryan Graham
Hidehiko Ichimura
Michael Jansson
Shakeeb Khan
Demian Pouzo

Thursday, March 24th, 2022

Reception hosted by Dan and Bev McFadden, 6:30 to 9:30PM (by invitation only)
Ashby Room, Residence Inn by Marriott Berkeley
2121 Center St, Berkeley, CA 94704

Friday, March 25th, 2022

B100 Blum Hall

Morning Session: 9 AM to 12:15 PM

Time	Speaker	Title
		Session 1: Panel Data
9:00AM to 9:30AM	Alex Poirier (remote)	Identification and Estimation of Average Partial Effects in a Semiparametric Binary Response Panel Model
9:30AM to 10:00AM	Ariel Pakes	Unobserved Heterogeneity, State Dependence, and Health Plan Choices
10:00AM to 10:30AM	Whitney Newey	Panel Estimation of Tax Effects with Endogenous Budget Sets
10:30AM to 10:45AM		Coffee Break
		Session 2: Instrumental Variables
10:45AM to 11:15AM	Marcelo Moreira	Efficiency Loss of Asymptotically Efficient Tests in an Instrumental Variables Regression
11:15AM to 11:45AM	Andrew Chesher	IV Methods for Tobit Models
11:45AM to 12:15PM	Jerry Hausman (remote)	Problems with the Control Variable Approach in Achieving Unbiased Estimates in Nonlinear Models in the Presence of Many Instruments
12:15PM to 1:45PM		Lunch

Afternoon Session: 1:45 PM to 4:30 PM

Time	Speaker	Title
		Session 3: Nonparametric analysis
1:45PM to 2:15PM	Matias Cattaneo	Higher-order Refinements of Small Bandwidth Asymptotics for Density-Weighted Average Derivative Estimators
2:15PM to 2:45PM	Ying Zhu	Blessing and Curse of Smoothness and Phase Transitions in Nonparametric Regressions: A Nonasymptotic Perspective
2:45PM to 3:00PM		Coffee Break
		Session 4: Program evaluation
3:00PM to 3:30PM	Songnian Chen (remote)	Quantile Regression with Group-Level Treatments
3:30PM to 4:00PM	Carolina Caetano	Partial Identification of Treatment Effects Using Bunching
4:00PM to 4:30PM	Guido Imbens	Fixed Effects and the Propensity Score
4:30 PM		Adjorn

Conference Dinner (by invitation only): 7PM @ Agrodolce Berkeley 1730 Shattack Avenue (about a 20-25 minute walk from Blum Hall as well as the hotel, or take Taxi/Uber)

Saturday, March 26th, 2022

B100 Blum Hall

Morning Session: 9AM to 12:15PM

Time	Speaker	Title
		Session 5: Dynamic Models
9:00AM to 9:30AM	Sergio Firpo	Uniform Inference for Value Functions
9:30AM to 10:00AM	Konrad Menzel	Structural Sieves
10:00AM to 10:30AM	Lars Hansen (remote)	Risk, Ambiguity and Misspecification: Implications for Policy
10:30AM to 10:45AM		Coffee Break
		Session 6: High-Dimensional Models
10:45AM to 11:15AM	Andres Aradillas-Lopez	Inference with Partially Identified Control Variables
11:15AM to 11:45AM	Cheng Hsiao	Statistical Inference for Low Dimensional Parameters in the Presence of High Dimensional Covariates
11:45AM to 12:15PM	Joel Horowitz	Bootstrap Based Asymptotic Refinements for High-Dimensional Nonlinear Models
12:15PM to 1:15PM		Lunch

Afternoon Session: 1:15PM to 3:30PM

Time	Speaker	Title
		Session 7: Experiments
1:15PM to 1:45PM	Max Kasy	Rationalizing Pre-Analysis Plans: Statistical Decisions Subject to Implementability
1:45PM to 2:15PM	James Heckman (remote)	Orthogonal Designs for Causal Inference in Social Experiments (tentative)
2:15PM to 2:30PM		Coffee Break
		Session 8: Networks
2:30PM to 3:00PM	Eric Auerbach	Spectral Regression for Networks, Panels, and Outcome Matrices
3:00PM to 3:30PM	Fengshi Niu	Error Components Models for Dyadic Data
3:30PM		Adjorn

Attending in Person Bryan Graham, Maximilian Kasy, Ariel Pakes, Carolina Caetano, Demian Pouzo, Matias Cattaneo, Konrad Menzel, Jack Porter, Chris Shannon, Jinyong Hahn, John Rust, Eric Auerbach, Andres Santos, Vira Semenova, Sergio Firpo, Ying Zhu, Whitney Newey, Laura Chioda, Andrew Chesher, Isaiah Andrews, Eric Mbakop, Thomas Rothenberg, Irene Botosaru, Carlos Flores, Hidehiko Ichimura, Antonio Galvao, Chris Muris, Oscar Jorda, Gregorio Caetano, Joel Horowitz, David Card, Sofia Villas-Boas, Cheng Hsiao, Shakeeb Khan, Graham Elliott, James Powell, Fengshi Niu, Andres Aradillas-Lopez, Tiemen Woutersen, Guido Imbens, George Judge, Bo Honore, Fred Finan, Daniel McFadden, Colin Cameron, Hidehiko Ichimura, Michael Jansson, Mikkel Soelvsten, Rosa Matzkin, James Stock, David Ahn, Liyang Sun, Patrick Kline, Dmitry Arkhangelsky, Peng Ding

Attending in Virtually Esfandiar Maasoumi, Yoosoon Chang, Joon Park, Donald Andrews, Xiaohong Chen, Myoung Jae Lee, Gary Solon, Songnian Chen, Ivana Komunjer, Alexandre Poirier, Roger Moon, Elie Tamer, Jerry Hausman, Lars Hansen, Serena Ng, Cristine Pinto, Elena Manresa, Charles Manski, Geert Ridder, James Heckman, Richard Crump, Richard J. Smith, Tomas Rau, Nese Yildiz, Petra Todd

Fourth Annual Berkeley-Stanford Econometrics Jamboree

Tue, 12 Oct 2021 00:00:00 +0000

On Friday November 19th, 2021 the Berkeley econometrics group will host the fourth annual “Berkeley-Stanford Econometrics Jamboree”. The event will consist of a morning session of student presentations (9AM to 11:30AM) and an afternoon session of faculty presentations (2PM to 6PM). The morning session will take place in the legendary venue 639 Evans Hall. To accommodate a somewat larger group, the afternoon session will take place in 290 Hearst Memorial Mining Building. This magnificient building is adjacent to Evans Hall.

Attendence is open to anyone from the Berkeley and Stanford data science communities (broadly and inclusively defined). Due to campus protocols regarding COVID-19, masks will be worn by all participants and no food or beverages will be provided. Berkeley affiliates are expected to meet all requirements for being on campus. Our Stanford guests should operate as if their own campus guidelines apply. The City of Berkeley requires proof of COVID-19 vaccination for indoor dining; so attendees (self-catered) lunch and dinner options will be greater if you bring such proof.

I do not expect to have to limit attendance, but to ensure a safe and comfortable event there is some chance I will. For this reason I ask that you let me know whether you will attend (and whether for the AM session, PM session or both). The event is otherwise intentionally informal. Please feel free to share these event details with others who may be interested.

A preliminary conference program can be found below. Please contact Bryan Graham with any corrections or possible scheduling conflicts. We’ve got a great group of speakers scheduled.

For our attendees coming from Stanford, Evans Hall is very close to the North Gate of our campus. The nearest BART stop is Downtown Berkeley. The nearest public parking facility is the Lower Hearst / North Gate parking garage at the intersection of Scenic & Hearst (enter from Scenic to access hourly parking spaces).

Jamboree Schedule

Organizers
Bryan Graham

Friday, November 19th, 2021 639 Evans Hall

Morning Session: 9AM to 11:30AM (Student Presentations)

Time	Speaker	Title
		Session 1:
9:00 to 9:20AM	Kevin Dano (Berkeley, Economics)	Identification and Estimation of Social Effects in Endogenous Networks
9:20 to 9:40AM	Jake Soloff (Berkeley, Statistics)	Trimming the Fat: An Empirical Bayes Approach to Local False Discovery Rate Control
9:40 to 10:00AM	Nikos Ignatiadias (Stanford, Statistics)	Noise-Induced Randomization in Regression Discontinuity Designs
10:00AM to 10:30AM		Break
		Session 2:
10:30AM to 10:50AM	Tijana Zrnic (Berkeley, EECS)	Post-Selection Inference via Algorithmic Stability
10:50 to 11:10AM	Sanath Kumar Krishnamurthy (Stanford, Operations Research)	Quantifying and Optimizing the Bias-Variance trade-off in Contextual Bandits
11:10 to 11:30AM	Evan Munro (Stanford, Economics)	Treatment Effects in Market Equilibrium
11:30AM		Break for lunch (on your own)

Afternoon Session: 2PM to 6PM (Faculty Presentations) 290 Hearst Memorial Mining Building

Time	Speaker	Title
		Session 1:
2:00 to 2:45 PM	Susan Athey (Stanford, GSB)	Federated Causal Inference
2:45 to 3:30 PM	Jingshen Wang (Berkeley, Biostatistics)	Breaking the Winner’s Curse in Mendelian Randomization
3:30 to 3:45 PM		Break
		Session 2:
3:45 to 4:30 PM	Liyang Sun (CEMFI, Economics)	Inference with Many Weak Instruments
4:30 to 5:15 PM	Guido Imbens (Stanford, GSB/Economics)	Multiple Randomization Designs
5:15 PM		Adjorn

Third Annual Berkeley-Stanford Econometrics Jamboree

Sun, 06 Oct 2019 00:00:00 +0000

On Friday November 8th, 2019 the Berkeley econometrics group, in partnership with the Statistics Department, will host the third annual “Berkeley-Stanford Econometrics Jamboree”. The event will consist of a morning session of student presentations (9AM to 11:30AM) and an afternoon session of faculty presentations (2PM to 6PM). Both sessions will take place in the legendary Neyman room (1011 Evans Hall).

Attendence is open to anyone from the Berkeley and Stanford data science communities (broadly and inclusively defined) and there is no need to register. The event is intentionally informal. Please feel free to share these event details with others who may be interested.

A conference program can be found below. Please contact Bryan Graham with any corrections or possible scheduling conflicts. We’ve got a great group of speakers scheduled.

If you have any logistical questions please write Fengshi Niu.

Jamboree Schedule

Organizers
Bryan Graham
Fengshi Niu

Friday, November 8th, 2019 1011 Evans Hall (Jerzy Neyman Room)

Morning Session: 9AM to 11:30AM (Student Presentations)

Time	Speaker	Title
		Session 1:
9:00 to 9:20AM	Evan Rose (Berkeley, Economics)	Who gets a second chance? Effectiveness and equity in supervision of criminal offenders
9:20 to 9:40AM	Eli Ben-Michael (Berkeley, Statistics)	Synthetic control and weighted event study models with staggered adoption
9:40 to 10:00AM	Chelsea Zhang (Berkeley, Statistics)	Active matrix factorization for surveys
10:00AM to 10:30AM		Break
		Session 2:
10:30AM to 10:50AM	Lydia Liu (Berkeley, Computer Science)	Competing bandits in matching markets
10:50 to 11:10AM	Michael Pollman (Stanford, Economics)	Causal inference for spatial treatments
11:10 to 11:30AM	Fengshu Niu (Berkeley, Economics	Kernel density estimation for undirected dyadic data
11:30AM		Break for lunch (on your own)

Afternoon Session: 2PM to 6PM (Faculty Presentations)

Time	Speaker	Title
		Session 1:
2:00 to 2:45 PM	Bin Yu (Berkeley, Statistics)	Three principles of data science: predictability, computability, and stability
2:45 to 3:30 PM	Guillaume Basse (Stanford, Statistics)	Minimax crossover designs for estimating learning and instantaneous effects
3:30 to 3:45 PM		Break
		Session 2:
3:45 to 4:30 PM	Konrad Menzel (New York University, Economics)	Bootstrap with cluster-dependence in two or more dimensions
4:30 to 5:15 PM	Lihua Lei (Stanford, Statistics)	Doubly robust two-way fixed effects regression for panel data
5:15 to 6:00 PM	Will Fithian (Berkeley, Statistics)	Statistical methods for replicability assessment
6:00 PM		Adjorn

Second Annual Berkeley-Stanford Econometrics Jamboree

Tue, 30 Oct 2018 00:00:00 +0000

On Friday November 30th, 2018 the Berkeley econometrics group, in partnership with the Statistics Department (through their NSF Research and Training Grant “Advancing Machine Learning - Causality and Interpretability”) will host the second annual “Berkeley-Stanford Econometrics Jamboree”. The event will run from 2PM to 6PM in room Jerzy Neyman Room (1011 Evans Hall). Attendence is open to anyone from the Berkeley and Stanford data science communities (broadly and inclusively defined) and there is no need to register. The event is intentionally informal. Please feel free to share these event details with others who may be interested.

A conference program can be found below. Please contact Bryan Graham with any corrections or possible scheduling conflicts. We’ve got a great group of speakers scheduled.

This year we will also organize a morning session consisting of short (20 minute) student presentations. Student presentations will be held from 9AM to 11:30AM in the Neyman Room. Confirmed student presenters are Michael Pollmann, Fengshi Niu, Andrin Pelican, Jake Soloff, Sara Stoudt and Jason Wu (exact order to be determined).

If you have any logistical questions please write Ingrid Haegele.

Jamboree Schedule

Organizers
Bryan Graham
Ingrid Haegele

Friday, November 30th, 2018 1011 Evans Hall (Jerzy Neyman Room)

Time	Speaker	Title
		Session 1:
2:00 to 2:45 PM	Art Owen (Stanford University, Statistics)	Optimizing the tie-breaker regression discontinuity design
2:45 to 3:30 PM	Avi Feller and Jesse Rothstein (Berkeley, Goldman/Statistics)	Augmented synthetic control method
3:30 to 3:45 PM		Break
		Session 2:
3:45 to 4:30 PM	Xiaoxia Shi (Wisconsin, Economics)	A simple uniformly valid test for inequalities
4:30 to 5:15 PM	Noureddine El Karoui (Criteo AI Lab, Paris/Palo Alto & Berkeley, Statistics)	Auction theory from the bidder standpoint
5:15 to 6:00 PM	Jann Spiess (Microsoft Research & Stanford, GSB)	Unbiased shrinkage estimation
6:00 PM		Adjorn

Berkeley-Stanford Econometrics Jamboree

Sat, 04 Nov 2017 00:00:00 +0000

On Friday November 17th, 2017 the Berkeley econometrics group will host a “Berkeley-Stanford Econometrics Jamboree”. The event will run from 2PM to 6PM in room 597 of Evans Hall. Attendence is open to anyone from the Berkeley and Stanford data science communities (broadly and inclusively defined) and there is no need to register. The event is intentionally informal. Please feel free to share these event details with others who may be interested.

A conference program can be found below. Please contact Bryan Graham with any corrections or possible scheduling conflicts. We’ve got a great group of speakers scheduled.

If you have any logistical questions please write Ingrid Haegele.

Jamboree Schedule

Organizers
Bryan Graham
Ingrid Haegele

Friday, November 17th, 2017 597 Evans Hall

Time	Speaker	Title
		Session 1: Panel Data & Related Models
2:00 to 2:45 PM	Chris Muris (Simon Fraser, Economics)	Binarization for panel data models with fixed effects (joint with Irene Botosaru)
2:45 to 3:30 PM	Guido Imbens (Stanford, GSB/Economics)	The role of the propensity score in fixed effects models (joint with Dmitry Arkhangelsky)
3:30 to 3:45 PM		Break
		Session 2: High Dimensional Models
3:45 to 4:30 PM	Lihua Lei (Berkeley, Statistics)	Asymptotics for high dimensional regression M-estimates: fixed design results (joint with Peter Bickel and Noureddine El Karoui)
4:30 to 5:15 PM	Markus Pelger (Stanford, Management Science & Engineering)	Estimating latent asset pricing factors from large-dimensional data (joint with Martin Lettau)
5:15 to 6:00 PM	Stefan Wager (Stanford, GSB)	Efficient policy learning (joint with Susan Athey
6:00 PM		Adjorn

Berkeley/CeMMAP Conference on Networks

Tue, 11 Oct 2016 00:00:00 +0000

We are very excited to welcome conference attendees to Berkeley on November 4th and 5th. A preliminary conference program can be found below. Please contact Bryan Graham with any corrections or possible scheduling conflicts. If we have forgetten someone please let us know and we will do our best to make adjustments.

We’ve got a great group of speakers, from the statistics, econometrics and machine learning communities, scheduled.

The conference will be held in Blum Hall Plaza at UC Berkeley. The venue is very close to the North Gate of campus. The nearest BART stop is Downtown Berkeley. The nearest public parking facility is the Lower Hearst / North Gate parking garage at the intersection of Scenic & Hearst (enter from Scenic to access hourly parking spaces).

If you have any logistical questions please write Eric Auerbach. I will post additional information to this page over the coming weeks.

IMPORTANT: We have allocated 40 minutes for each talk. Please anticipate questions from the audience when planning your talk. You are very welcome to use less than forty minutes and decidely not welcome to use more!

Provisional Schedule

Organizers
Aureo de Paula
Bryan Graham
Jasjeet Sekhon

Friday, November 4th, 2016

Time	Speaker	Title
11:40 to 12:30PM		Registration & Lunch
		Session 1: Block models & Heterogeneity
12:30 to 1:10PM	Peter Bickel (Berkeley)	Block Models with Covariates: Likelihood Methods of Fitting (joint with Purna Sarkar, U.T. Austin, Soumendu Mukherjee, Berkeley, and Sharmodeep Bhattacharyya, Oregon State)
1:10 to 1:50PM	Stephane Bonhomme (Chicago)	Discretizing Unobserved Heterogeneity (joint with Thibaut Lamadon, Chicago, and Elena Manresa, NYU)
1:50 to 2:10PM		Coffee Break
		Session 2: Dynamic models
2:10 to 2:50PM	Angelo Mele (Hopkins)	Variational Approximations for Large Strategic ERGMs
2:50 to 3:30PM	Peter Hoff (Duke)	Tensor Regression for Dynamic Network Data
3:30 to 3:50PM		Coffee Break
		Session 3: Experiments & Spillovers
3:50 to 4:30PM	Edoardo M. Airoldi (Harvard)	Two Model-Assisted Strategies for Designing Experiments on Networks
4:30 to 5:10PM	Dean Eckles (MIT)	Estimating Peer Effects in Social Networks with Peer Encouragement Designs (joint with René Kizilcec, Stanford, and Eytan Bakshy, Facebook
5:10 to 5:50PM	Chris Rose (Toulouse)	Social Effects when the Network is Unobserved and Sparse (joint with Eric Gautier, Toulouse)
6:45PM		Speakers’ dinner¹ at Soi 4 5421 College Avenue, Oakland, CA 94618

Saturday, November 5th, 2016

Time	Speaker	Title
8:30 to 9:00AM		Breakfast
		Session 4: Strategic interaction
9:00 to 9:40AM	Michael Leung (USC)	Normal Approximation in Strategic Network Formation
9:40 to 10:20AM	Konrad Menzel (NYU)	Many Player Asymptotics for Large Network Formation Problems
10:20 to 11:00AM	Geert Ridder (USC)	Estimation of Large Network Formation Games (joint with Shuyang Sheng, UCLA)
11:00 to 11:20AM		Coffee Break
		Session 5: Link formation with Heterogeneity
11:20 to 12:00PM	Andreas Dzemski (Gothenburg)	An Empirical Model of Dyadic Link Formation in a Network with Unobserved Heterogeneity
12:00 to 12:40PM	Ivan Fernandez-Val (BU)	Panel Quantile Effects Via Distribution Regression (joint with Victor Chernozhukov, MIT, and Martin Weidner, UCL)
12:40 to 2:00PM		Lunch
		Session 6: High dimensional models
2:00 to 2:40PM	Roger Moon (USC)	Estimation of High-Dimensional Graphical Models with Shape Restriction (joint with Khai Xiang Chiong, USC Dornsife INET)
2:40 to 3:20PM	Martin Weidner (UCL)	Fixed-Effect Regressions on Network Data (joint with Koen Jochmans, Sciences Po)

Unfortunately, due to very tight physical space constraints at the venue, the dinner is by invitation only. Soi 4 is a pleasant 2.5 mile walk from campus through the Elwood and Rockridge neighborhoods. Directions can be found here. For those not wishing to walk, it is just a short Taxi/Uber ride away. ↩

netrics: a Python module for the econometric analysis of networks

Thu, 15 Sep 2016 00:00:00 +0000

Over the past few years I have devoted a large amount of my research time to developing econometric methods for the analysis of network data. This an interesting area; the statistics, math, economics and empirical work are all very interesting (and not coincidently challenging).

The first full paper I’ve completed is this area is called “An econometric model of link formation with degree heterogeneity.” The latest revision of the paper can be found here. The paper proposes and analyzes two estimators for the parameter indexing a simple (but nevertheless interesting) model of network formation.

To describe the model let $ \mathbf{D} $ be an $ N \times N$ adjacency matrix. The $(i,j)^{th}$ element of $ \mathbf{D} $ equals one if agents $i$ and $j$ are linked and zero otherwise. The paper considers the undirected case so that $D_{ij}=D_{ji}$ . Assume there are $i=1,...,N$ agents in the network to be analyzed. Let $\mathbf{X}$ be an $N \times J$ matrix of agent-level covariates observed by the econometrician.

Let $n=\tbinom{N}{2}$ denote the number of dyads, or distinct pairs of agents, in the network. Using the rows of $\mathbf{X}$ we can construct the $n \times K$ matrix $\mathbf{W}$ of dyad-specific covariates. Because the network is undirected these covariates are constructed to be invariant to permutations of their indices (i.e., $W_{ij}=W_{ji}$ ). Examples of such covariates include the physical distance between two agents, the absolute difference in their income levels, whether they belong to the same religion, are blood relatives and so on.

I also assume that each agent is characterized by an unobserved individual-specific parameter $A_i$ . Let $ \mathbf{A} $ be the $N \times 1$ vector of these parameters. For reasons that will become apparent shortly I call these degree heterogeneity parameters.

With this notation established, the paper assumes the conditional likelihood of the event $ \mathbf{D=d} $ is:

$$ \Pr\left(\left.\mathbf{D}=\mathbf{d}\right|\mathbf{X},\mathbf{A}\right)=\prod_{i<j}\left[\frac{\exp\left(W_{ij}'\beta_{0}+A_{i}+A_{j}\right)}{1+\exp\left(W_{ij}'\beta_{0}+A_{i}+A_{j}\right)}\right]^{d_{ij}}\left[\frac{1}{1+\exp\left(W_{ij}'\beta_{0}+A_{i}+A_{j}\right)}\right]^{1-d_{ij}} $$

In words: links form independently conditional on $ \mathbf{X} $ and $ \mathbf{A} $ . The latter, however, is unobserved by the econometrician. Unconditional on $ \mathbf{A} $ links will generally be dependent. The $A_i$ parameters allow individuals to vary in the generic surplus they generate when forming links. Practically this allows for rich patterns of degree heterogeneity. A common feature of real work networks are degree distributions characterized by many agents with only a handful of links combined with a few so called “hub” agents with many links. For a popular account of this phenomena you can read this article by Albert-Lászlo Barabási and Eric Bonabau in the Scientific American.

The model of network formation outlined above rules out strategic behavior, whereby the return to two agents forming a link may vary with the presence or absence of links elsewhere in the network (a phenomena admittedly central to some economic settings), but it is a reasonably rich starting point for empirical analysis. It provides a nice set-up for studying homophily, the tendency of individuals with similar attributes to form links, in a context that also allows for degree heterogeneity. This type of heterogeneity appears to be important in practice and, furthermore, can confound inferences about the presence or absence of homophily (as discussed in the paper via a few examples). For a theoretical take on homophily and degree heterogeneity, with some connections to my paper, and also, incidentally, in a non-strategic setting, see this paper by Yann Bramoulle, Sergio Currarini, Matthew Jackson, Paolo Pin and Brian Rogers in the Journal of Economic Theory.

As a quick aside, I am also very interested in models which do allow for strategic link formation. See this working paper. There’s also been very interesting work in this area by Konrad Menzel, Aureo de Paula, Seth Richards-Shubik, Elie Tamer and Shuyang Sheng among others. Much of this work is very recent. A few examples are here, here and here.

My paper proposes two estimators for the common parameter $\beta_{0}$ in the model above. The first I call the tetrad logit estimator. A “tetrad” is a group of four agents. It can be “wired” in $2^6=64$ different ways. The tetrad logit estimator looks at the probability of different tetrad configurations conditional on the configuration belonging to a subset of the 64 possible configurations. These subsets are carefully chosen such that the $A_i$ degree heterogeneity parameters “drop out” of the resulting criterion function. The argument is based on sufficiency results for exponential families. These arguments have proved very useful in panel data analyses (most famously in a classic 1980 paper by Gary Chamberlain in the Review of Economic Studies). It probably goes without saying that I drew my inspiration from these panel data applications (Gary was my dissertation advisor and has deeply influenced how I think about econometric problems). In independent work, a similar sufficiency argument was introduced by Karyne Charbonneau in the related context of gravity models of trade.

It turns out that the distribution theory for the Tetrad Logit estimator was difficult to work out. It involved many trips to the library and discussions with my colleagues at Berkeley and elsewhere. The tetrad logit estimator is a 4th order U-Process minimizer with a particular degeneracy and dependence structure. This results in some fun twists relative to the normal analysis of such estimators (as in, for example, Honore and Powell (1994, Journal of Econometrics)).

Subsequent to my work Koen Jochmans at Sciences Po has worked out distribution theory for the related set-up initially considered by Charbonneau. Concretely this means a parallel set of results for directed networks is also now available to empirical researchers.

The second estimator I introduced drew its inspiration from so called large-N, large-T nonlinear panel data models. A key reference for my work here is the paper by Hahn and Newey (1994, Econometrica). My second estimator jointly estimates the common parameter along with the $N$ individual-specific degree heterogeneity parameters. It turns out that if the network is sufficiently dense, such that in the limit each agent has many links, such a procedure can work (despite the growing dimension of the incidental parameters). However the limit distribution for the estimate of $\beta_{0}$ will have a bias term. This can be fixed by bias correction. My work here also draws on results from the probability literature on random graphs; especially a paper by Chatterjee, Dianconis and Sly (2011, Annals of Applied Probability). Their work, in turn, builds on an older literature on the Bradley-Terry model in statistics.

I think both estimators have a role to play in empirical work. Right now I am probably most partial to the tetrad logit approach. It delivers a consistent estimate of $\beta_{0}$ under weaker assumptions than the joint estimator. In particular it can work well in sparse networks (including very sparse networks). Such networks are very common in practice. Also theory, introspection, and Monte Carlo experiments suggest that the joint fixed effects estimator may work very poorly in sparse networks. At the same time researchers will sometimes like to have estimates of the heterogeneity parameters as well as the common parameter (e.g., to compute marginal effects). Andreas Dzemski shows how such estimates can be used for specification testing in a way likely to be very attractive to network researchers. A copy of his paper can be found here. So I think both approaches have a role to play in empirical work and there is more methodological work to be done to develop each method more fully.

The Monte Carlo results reported in the current (publically available) draft were done using a Matlab script I wrote last year. This worked fairly effectively. However, on my desktop computer, the tetrad logit estimator would konk out at just over 100 agents due to memory problems. Exact computation of the tetrad logit estimate is difficult because there are $\tbinom{N}{4}$ tetrads in a network and this number gets big very quickly. There are a variety of ways one could approximately compute the tetrad logit estimator “at scale”, but I wanted to be able to provide code that could work reliably on networks with at least a few hundred agents on a good desktop machine (and perhaps somewhat larger networks when computing on a cluster). To this end I recoded the procedure in Python over the summer and this exercise resulted in a much more successful, and user friendly, implementation.

I also recoded the joint fixed effects estimator in Python. To make all this code a bit more accesible to empirical researchers I put everything together into an add-on Python package.

The Python package is called netrics for “NETwork economeRICS” (it could be worse). Currently the package only includes a implementation of the tetrad logit and joint logit fixed effects procedures as well as a few auxiliary helper functions. I hope to add more functionality over time (e.g., an implementation of my dynamic network formation work mentioned above).

The package is registered on PyPi. The source code is available at this GitHub repository. The netrics package has the following dependencies: numpy, scipy, pandas, numba, numexpr and itertools. These are standard libraries and are included in most scientific Python distributions. For example they are included in the highly recommended Anaconda distribution of Python. If you are using the Anaconda distribution of Python, then you can follow the (straightforward but tedious) instructions here to learn how install the netrics package from PyPi and make it available in Anaconda using the “conda” package manager. For users who anticipate only infrequent use, permanent installation of the netrics package may not be worth the trouble. One possibility is to just clone (ie., copy) the GitHub repository, which contains the latest version of netrics. Then append the path pointing to the location of the netrics package (on your local machine) to your sys directory. This is what is done in the snippet of code below.

For example if you download the repository into a directory called “netrics” on your local machine and navigate there, you should observe the following basic structure (perhaps with more .py files in the netrics/ folder)

README.txt
LICENSE
MANIFEST.in
setup.py
netrics/__init__.py
netrics/logit.py
netrics/tetrad_logit.py

Joachim de Weerdt has kindly made his Nyakatoke network dataset freely available on his website here. To use the netrics package to fit a simple model of network formation to the Nyakatoke risk-sharing network collected by Joachim de Weerdt you can then run the following code snippet:

# Import numpy in order to correctly read test data
import numpy as np

# Import urllib in order to download test data from Github repo
import urllib

# Append location of netrics module base directory to system path
# NOTE: only required if permanent install not made (see comments above)
import sys
sys.path.append('/Users/bgraham/Dropbox/Sites/software/netrics/')

# Load netrics module
import netrics as netrics

# Download Nyakatoke test dataset from GitHub
# Edit to download location on your local machine   
download =  '/Users/bgraham/Dropbox/'
url = 'https://github.com/bryangraham/netrics/blob/master/Notebooks/Nyakatoke_Example.npz?raw=true'
urllib.urlretrieve(url, download + "Nyakatoke_Example.npz")

# Open dataset
NyakatokeTestDataset = np.load(download + "Nyakatoke_Example.npz")

# Extract adjacency matrix
D = NyakatokeTestDataset['D']

# Initialize list of dyad-specific covariates as elements
# W = [W0, W1, W2,...WK-1]
W = []

# Initialize list with covariate labels
cov_names = []

# Construct list of regressor matrices and corresponding variable names
for matrix in NyakatokeTestDataset.files:
    if matrix != 'D':
        W.append(NyakatokeTestDataset[matrix])
        cov_names.append(matrix)   

# Fit tetrad logit to Nyakatoke
[beta_TL, vcov_beta_TL, tetrad_frac_TL, success] = \
	netrics.tetrad_logit(D, W, dtcon=None, silent=False, W_names=cov_names)

The netrics.tetrad_logit() function, depending on your machine, may take a few minutes to churn out an answer. The default output is pretty basic. It reports network size, the number of tetrads and the number of tetrads with identifying content (typically only a small fraction of all tetrads). It also reports coefficient and (asymptotically valid) standard error estimates.

For more illustrations of the netrics.tetrad_logit() command, and also of my implementation of the joint fixed effects estimator, see this iPython Notebook on GitHub.

While I would appreciate bug reports, suggestions for improvements and so on, I am unable to provide any meaningful user-support for the package. I hope to add additional functionality for the analysis of networks to the package over time. I hope you find it useful. If you do use it for your own research please let me know, I would be very curious to see how it gets deployed in practice.

ipt Module for Program Evaluation

Sun, 15 May 2016 00:00:00 +0000

Earlier this spring my co-authors and I finally published the paper “Efficient estimation of data combination models by the method of auxiliary-to-study tilting (AST)” in the Journal of Business and Economic Statistics . A copy of this paper can be found on my research page here.

Publishing this material was almost a decade long process. On the whole I’ve been reasonably-to-very lucky when it comes to publishing, but this paper is an exception that proves the rule. The material dates to 2006 (the “Eureka moment” came shortly after the birth of my son), with the first widely circulated version appearing in Section 4 of the initial version of our 2008 NBER Working Paper “Inverse probability tilting and missing data problems.”

The bulk of the material in the 2008 NBER paper was published in the Review of Economic Studies in 2012 (after many very challenging revisions). A copy of that paper is available here. The “Section 4” material, which just appeared in print this spring, went through a long “revise and resubmit”, “reject and resubmit” and then straight “reject” editorial process at another journal before we turned to the Journal of Business and Economic Statistics.

The upshot of these twists and turns is that the end research product went through numerous improvements prior to publication. For both papers we were pushed hard by the editor and referees to improve our research. Unfortunately, while not especially esoteric by the standards of modern econometrics, neither the Review of Economic Studies nor the Journal of Business and Economic Statistics papers are especially easy to read for a non-specialist.

The two papers are respectively organized around a rather general semiparametric missing data and data combination problem. Many interesting econometric problems fall within these two classes. Perhaps the leading two estimands covered by our work are the Average Treatment Effect (ATE) and the Average Treatment Effect on the Treated (ATT). Unfortunately readers interested in the implications of our work for program evaluation applications have to struggle through papers with much more generality than they probably need or want.

A key idea in both papers is to construct a propensity score estimate that re-weights the treatment and control subsamples to exactly match/balance moments of the pre-treatment covariates across the two samples. Researchers often present a table of average mean differences in pre-treatment covariates both before and after undertaking some sort of adjustment procedure (e.g., matching). Researchers using our estimators could present a table showing zero mean differences in the pre-treatment covariates across treatment and control units after re-weighting. When presenting this work to applied audiences I have generally found researchers receptive and attracted to this feature of our method.

The whole idea of invoking “selection on observables” is that, if one adjusts for differences in the observed covariates across treatment and control units, all selection bias can be removed. The key idea in our paper is to directly reweight the treatment and control subsamples to impose distributional balance in observed covariates.

The approach we developed in the two papers was inspired by older ideas on contingency table calibration which go back at least to the 1940s, as well as, more subtlety, by the structure of the semiparametric efficiency bound for the missing data and data combination problems. See, for example, this paper. We also used ideas from the literature on the efficient estimation of expectations, particularly as it appears in research on generalized empirical likelihood (e.g., Imbens (1997, Review of Economic Studies, Newey and Smith (2004, Econometrica)).

The product of our particular mix of all these ingredients is a rather good estimator for the ATE as well as the ATT. One that, in addition to having good first-order asymptotic properties (namely local efficiency and double robustness), also has good higher-order asymptotic properties.

Since our work first appeared, many other authors have developed various direct covariate balancing estimators of the propensity score. The Entropy Balancing approach of Jens Hainmueller is one which I believe developed independently (and concurrently) with our work. The Covariate Balancing Propensity Score method of Kosuke Imai and Marc Ratkovic, which appeared after our (and Hainmueller’s) work, has proved especially popular in the field of Political Science.

There is also some nice work extending the idea of exact balancing to high-dimensional settings. This extension is non-trivial since with many controls there may be no re-weighting which exactly balances, say the means, of the control variables across the treated and non-treated subsamples. In such cases only approximate balancing is possible and there are a variety of delicate decisions involved here. See this working paper by Susan Athey, Guido Imbens and Stefan Wager for some interesting ideas in this area.

There is a lot more additional work in this area. Too much to properly survey here. There are also several important antecedents to our work; not all of them obvious. For example, my favorite is the clever approach to adjusting for non-ignorable attrition using refreshment samples introduced by Hirano, Imbens, Ridder and Rubin (1998, Econometrica). That paper implicitly uses calibration ideas which influenced our own work.

Nevertheless, despite all the recent work on “covariate balancing” propensity scores, our two papers remain distinctive, not just for being “first”, but in terms of the theoretical development of the estimators. We show that our ATE and ATT estimators are locally efficient, doubly robust and have attractive higher order properties relative to other first order equivalent estimators. In the case of our ATT procedure I am aware of no other locally efficient estimator in the literature (although I could be wrong!). Furthermore, while I have not formally analyzed them, the structure of some recent suggestions for estimation of covariate balancing propensity scores suggest, by analogy with known results on GMM estimation of over-identified models, poor finite sample properties (at least relative to our proposals).

While we wrote and made available a fair amount of code in connection with our two papers (mostly in MATLAB) we did not do a tremendous amount to promote it, or make it easy for other researchers to use in their own work. When our second paper finally came out this spring I decided I would try to write some more user friendly code. This is part of a bigger resolution to make my work more accessible to researchers.

While I would like to make a Stata and R implementations of our estimators available, I decided to first prepare a Python 2.7 implementation. I like Python a lot. It is widely-used in the Data Science and Machine Learning communities. While senior researchers in statistics and econometrics may be less familiar with Python, chances are their students are very familiar with it. Python is probably the most common language used to teach introductory computer science. At Berkeley Python is also used to teach our general education Foundations of Data Science course.

The Python package is called ipt for “inverse probability tilting”. Currently it only includes a implementation of our ATT estimator, although I intend to incorporate an ATE estimator into the package in the future. This package is registered on PyPi. The source code is available at this GitHub repository. The ipt package has the following dependencies: numpy, numpy.linalg, scipy, scipy.optimize and scipy.stats. These are standard libraries and are included in most scientific Python distributions. For example they are included in the highly recommended Anaconda distribution of Python. If you are using the Anaconda distribution of Python, then you can follow the (straightforward but tedious) instructions here to learn how install the ipt package from PyPi and make it available in Anaconda using the “conda” package manager. For users who anticipate only infrequent use, permanent installation of the ipt package may not be worth the trouble. One possibility is to just clone (ie., copy) the GitHub repository, which contains the latest version of ipt. Then append the path pointing to the location of the ipt package (on your local machine) to your sys directory. This is what is done in the snippet of code below.

For example if you download the repository into a directory called “ipt” on your local machine and navigate there, you should observe the following basic structure (perhaps with more .py files in the ipt/ folder as additional functionality is added to the module over time)

README.txt
LICENSE
MANIFEST.in
setup.py
ipt/__init__.py
ipt/logit.py
ipt/att.py

To use the package to estimate the ATT using the NSW evaluation dataset used by Dehejia and Wahba (1999, Journal of the American Statistical Association) you can then run the following code snippet:

# Append location of ipt module root directory to systems path
# NOTE: Only required if ipt module not "permanently" installed via "pip", "conda" etc.
import sys
sys.path.append('/Users/bgraham/Dropbox/Sites/software/ipt/')

# Load ipt package
import ipt as ipt

# Read help file for ipt.att()
help(ipt.att)

# Read nsw data directly from Rajeev Dehejia's webpage into a
# Pandas dataframe
import numpy as np
import pandas as pd

nsw=pd.read_stata("http://www.nber.org/~rdehejia/data/nsw_dw.dta")

# Make some adjustments to variable definitions in experimental dataframe
nsw['constant'] = 1                # Add constant to observational dataframe
nsw['age']      = nsw['age']/10    # Rescale age to be in decades
nsw['re74']     = nsw['re74']/1000 # Recale earnings to be in thousands
nsw['re75']     = nsw['re75']/1000 # Recale earnings to be in thousands

# Treatment indicator
D = nsw['treat']

# Balancing moments
t_W = nsw[['constant','black','hispanic','education','age','re74','re75']]

# Propensity score variables
r_W = nsw[['constant']]

# Outcome
Y = nsw['re78']

# Compute AST estimate of ATT
[gamma_as, vcov_gamma_ast, study_test, auxiliary_test, pi_eff_nsw, pi_s_nsw, pi_a_nsw, exitflag] = \
                                                                ipt.att(D, Y, r_W, t_W, study_tilt=True)

The ipt.att() command will spit out lots of useful and interesting diagnostic output (set silent=True to suppress this). For a reasonably well-narrated guided tour of the ipt.att() command see this iPython Notebook on GitHub.

While I would appreciate bug reports, suggestions for improvements and so on, I am unable to provide any meaningful user-support for the package. When I manage to code up the ATE estimator (or any additional features) I’ll post an update on this blog.

This posted was lightly edited on 5/29/16 to reflect a few bug fixes and initial module improvements.