# Joshua Catalano

*launch*Personal Website

*file_download*Download CV

##### Research Area

##### Education

The University of British Columbia, Vancouver School of Economics, Sept. 2017 - Present

Program: PhD in Economics, expected graduation in June 2024

Fields: Econometric Theory

The University of Colorado at Boulder, College of Arts and Sciences, Aug. 2010 - Aug. 2014

Degree: BA in Economics & Mathematics, with distinction

Honors: Summa Cum Laude in Economics (Honors Thesis)

## About

My research lies at the intersection of econometric theory and machine learning. Currently, I am investigating how a novel algorithm I developed can be used in tandem with classical regression tree algorithms to consistently select and estimate geometrically flexible partition models. In my job market paper, I apply this partition estimator to the problem of parallel trend violations in difference-in-differences estimation. I expect to graduate in June, 2024.

## Research

**Job Market Paper**

We propose and demonstrate an alternative difference-in-difference framework to be used when observed pre-trends cast doubt on the standard parallel trends assumption. Modeling trends in the outcome variable as a partition model on a support of observable variables, we impose that the parallel trends assumption holds within at least one of the subsets (or groups) constituting that partition. That is, units whose covariates fall into such a group have untreated potential outcomes that evolve in parallel. Group-level Average Treament Effects on the Treated (ATT) are identified for such subsets. If the group-level ATT is well-identified for each subset in the partition, then so is the usual ATT. Otherwise, we recover identification of what we call the partial ATT -- an ATT conditional on being in a well-identified subset. To estimate the partition model and ultimately these ATTs, we propose a two-step estimator. First, the partition model is consistently estimated on pre-treatment data using standard regression tree algorithms in tandem with a novel joining algorithm. Subsequently, the usual two-way fixed effects regression is estimated within each group separately -- that is, only using units whose covariates fall into that subset. We suggest two ways to select which groups ostensibly follow the parallel trends assumption based on the pre-trends of their treated and untreated units. Selected group-level ATT estimates can be aggregated to estimate either the usual ATT (if all groups are selected) or the partial ATT.

**Working Papers**

## Awards

**The University of British Columbia**

Faculty of Arts Graduate Award, 2017 & 2020-2021

British Columbia Graduate Scholarship, 2019

President’s Academic Excellence Initiative PhD Award, 2020-2023

**The University of Colorado at Boulder**

Dean’s List, 2011-2014

Sieglinde Talbott Haller Endowed Economics Scholarship, 2013

## Teaching

__As Lecturer__

#### Working with Big Data

This is a 40-hour, month-long data science course I taught and designed in July-August of 2023 for the **Vancouver Summer Program**. I modeled the structure of the course after **QuantEcon’s** course, **Introduction to Economic Modeling and Data Science**. Due to the timeframe, lack of pre-requisites, and lab-based course structure, I had to create all new materials, occasionally borrowing code and examples from QuantEcon.

**Relevant Links:** **Course Github** and **Student Evaluations**.

__As Teaching Assistant__

All of my teaching assistantships were for courses at the **University of British Columbia**.

#### Undergraduate Courses

ECON 101 — Principles of Microeconomics

ECON 221 — Introduction to Strategic Thinking

ECON 311 — Principles of Macroeconomics (upper-year students/non-majors)

ECON 323 — Quantitative Economic Modelling with Data Science Applications** [Evals: ****1** **2**, **3**]

ECON 345 — Money and Banking

ECON 365 — Topics in Industrial Organization

ECON 370 — Benefit-Cost Analysis and the Economics of Project Evaluation

#### Graduate Courses

PPGA 500a & 500b — Economics for Policy [**Evaluation**]

ECON 573 — Environmental Economics

ECON 622 — Computational Economics with Data Science Applications [**Evaluation**]