# Executive summary

## In a few sentences only, please summarize what this proposal is for.

This project will alleviate information asymmetries by identifying, training, and recommending exceptional students from diverse socioeconomic backgrounds to the top 20 economics PhD programs in the world. These students will learn cutting-edge open source computational tools that allow them to maximize their impact while engaging in open, cross-disciplinary research. Through a combination of direct effects and scaling via replication, the project will mitigate the alarming trend away from socioeconomic diversity in the economics profession, while at the same time bringing welcome new perspectives and outstanding talent to bear on sustainable development, poverty alleviation, and other crucial real-world issues.

## In a few sentences only, what are the topline impact targets that this proposal will aim to achieve?

First, this proposal will advance scientific research by identifying brilliant students with an interest in economics and maximizing their potential. Second, the proposal will impact the direction of economic research by increasing intake of students from diverse socioeconomic backgrounds and by training them in open-source computational tools that encourage open science and collaborative, interdisciplinary research. Third, this project will improve the pipeline for identifying highly talented individuals in all fields of science that rely heavily on computation.

# Proposal overview

## What is the problem / need / gap / market failure that this proposal will address?

Recent research shows that, relative to other academic fields, economics is strikingly non-diverse in terms of the socioeconomic background of its PhD candidates. For example, US-born economics PhDs are now 5 times more likely than a similarly-aged member of the general US population to have a parent with a graduate degree. This lack of diversity is extreme even relative to mathematics, engineering and the natural sciences, suggesting that potentially brilliant students from non-elite family backgrounds face large intangible barriers to entry.

Worryingly, the problem described above is intensifying rather than abating. For example, the share of US-born PhD recipients from a household where at least one parent has a graduate degree has increased from 20 percent to 65 percent since 1970, exceeding the rate of growth not only in the general population but also in all other major PhD fields.

Falling socioeconomic diversity is harmful to both students and the discipline. Even setting aside equity considerations, narrow intake implies poor efficiency (i.e., failing to include potentially brilliant students). At the same time, narrow perspectives and leadership by highly concentrated networks tend to stifle innovation, mis-allocate research effort and suppress cross-disciplinary collaboration.

While some might argue that outstanding students who cannot achieve their potential in economics will find other fields to contribute to, this argument does not account for the fact that some of these students will be particularly passionate about solving economic problems. Our proposal will enable exceptionally talented and enthusiastic students from diverse socioeconomic backgrounds to make major contributions to the field of economics.

## What are the specifics of how this program / initiative will function (please strive to be concise, putting additional material in the appendix as needed)?

The project will consist of four stages:

1. Remote learning and assessment.
2. Selection of summer school participants.
3. Summer school in India.
4. Recommendations to top PhD and predoc programs.

Stage 1 will use videos and interactive browser-based lectures that teach quantitative economics and prepare students for summer school intake assessment. Prior to commencement, QuantEcon (QE) develop and deploy jupyter-course, an open-source platform for teaching and grading that builds on the Jupyter Book ecosystem.

(A small number of in-person workshops will also be provided, in order to raise awareness of the program, maximize participation and promote diversity.)

Stage 2 consist of two parts. Part A will implement automated grading to reduce a large pool of applicants to 100 students. Part B will involve human judgment to select the final 25 students according to a weighted combination of talent, potential and socioeconomic diversity.

In Stage 3, QE will run a fully funded, intensive six week in-person summer school program in computational and quantitative economics for the 25 selected students, with Thomas Sargent acting as the lead instructor. The students will have abundant opportunities to directly interact with workshop leaders in class and small-group discussions.

In Stage 4, the best students will receive introductions to scholarship opportunities, recommendation letters for PhD applications, and recommendations to high-quality predoctoral fellowships.

## What is clever, innovative, or new about this approach?

First, by combining existing open-source tools, QE will develop a platform for remote learning and high-volume automated grading that permits operation at scale, engaging and assessing a student base orders of magnitude larger than a comparable program using manual grading. The platform will be open-sourced to facilitate low-cost replication by like-minded teams from economics and other fields.

Second, QE will focus its operation on a single country (India) and run the summer school domestically. This maximize cost-efficiency (relative to an international operation with student airfares and associated expenses) while facilitating intake of students from non-elite backgrounds, since no passport or visa is required. (At the same time, with 30 million tertiary students and strong English background, India has an ample pool of outstanding talent, including many students from non-elite family backgrounds.)

Third, as an indirect means of targeting top PhD programs, QE will emphasize placing students in predoctoral fellowships, which have become a key pathway into elite PhD programs. Targetting will be achieved by training students in in the skills most sought after by supervisors: computational data analytics, statistics, data management and high-performance computing with applications to economics.

Fourth, the focus on training in cutting-edge computational tools will provide foundations that allow students to thrive at top PhD programs while also pushing them towards cross-disciplinary collaboration (since computational tools are typically more advanced in other fields, such as engineering and the natural sciences). Moreover, these tools will all be open source, plugging students into a network of open-source developers and helping to emphasize the vital role played by open, reproducible, democratic scientific research.

Fifth, an important aspect of Stage 1 is that the application process itself will provide significant benefits for students. Even those who are not selected will receive first-rate education in high-demand skills, which will enhance future study and career opportunities.

# Impact

## What is the high-level impact we expect this proposal to have? On what timeline?

One impact of this project will be better methods for identifying highly talented individuals in economics and other fields of science that rely heavily on computation. The proposed activities will maximize impact along this dimension by open-sourcing the platform and encouraging other academics and interested parties to replicate project methodology in other locations and fields. These impacts will be realized over a one to five year timeline.

A second high-level impact will be advancement of scientific research through identification of brilliant students and maximization of their potential (as well as training in open-source computational tools that encourage open science and collaborative, cross-disciplinary research). These impacts will be realized on a ten to thirty year timeline.

A third high-level impact will be improved socioeconomic diversity in the economics profession, partly through the direct effects of this program and primarily through replication by other teams. These impacts will be realized on a ten to thirty year timeline.

## What is the likelihood that this investment will have the intended impact?

The likelihood is high, due to careful planning and having the right team for the proposed activities. Suitability of the team is discussed below.

## Summary of intended targets

**(In other words, how will you know if this project is successful? As much as possible, please include specific, quantitative targets for both project outputs (i.e., activities delivered and tasks completed) and project outcomes (i.e., the end benefits achieved). Please also note the timing of these outputs and outcomes (e.g., by what date would you expect to reach the specified targets)(Note: if there is a more detailed project work plan you’d like to include, please add it to the appendix)**

## Outputs (e.g., “hire X new staffers,” “publish X articles,” etc.)

1. Course creation
1. Content generation, including 25 new videos and 20 new lectures
on quantitative methods and computational economic modeling.
2. Development of jupyter-course, a platform for teaching and automated
grading built on the Jupyter Book ecosystem.
3. Documentation and open source publication of the platform to facilitate
adoption and re-use.
4. One new full-time hire at the level of junior developer.
2. Remote learning
1. 5,000 students enrolling in the Stage 1 online course.
1. 2,500 students completing the Stage 1 online course, including assessment.
3. Summer school
1. 25 students selected for summer school.
4. Academic PhD applications
1. 20 students applying for pre-docs and PhD programs.
2. 20 students receiving academic recommendation.
5. Additional funding
1. At least one new grant or gift to facilitate repetition and replication
of the program.

## Outcomes (e.g., “create X new jobs,” “register X new voters,” etc.)

The first year of the program will produce the following outcomes.

1. Remote learning
1. 50% completion of remote learning and assessment.
2. Summer school
1. 100% completion of summer school.
3. Academic PhD applications
3. 15 students accepted and attending top predoc or PhD programs.
3. 5 students accepted to non-elite PhD programs or positions as
professional economists.

# Team

## Of the various teams working on this problem, why is this the one we should bet on? How does this team compare to / differentiate itself from others doing similar or related work?

The QuantEcon team is ideally suited to deliver this program with maximum long-run impact for the following reasons:

First, QE members have deep academic networks reaching into top PhD programs in economics, as well as extensive knowledge of the current state of PhD programs at top schools.

Second, QE members have very high credibility in terms of writing letters to top PhD programs.

Third, QE team members are at at the frontier of research in computational economics, with publications on computationally intensive problems in leading journals in economics, finance and operations research.

Fourth, QE members have extensive experience teaching computation methods for economists using open source tools, having run lecture series and conducted workshops within many top PhD programs in economics (NYU, MIT, Harvard, Yale, Princeton, Stanford, Columbia, Berkeley, UCLA, etc.)

Fifth, QE members have strong connections to academic institutions in India.

Sixth, QE has a network of potential teaching assistants, including current and former PhD students who have been involved in previous QE workshops.

Seventh, QE has demonstrated experience in building open source tools, software, and infrastructure to assist with delivering education in open source computing. For example, QE is currently working with and plans to hire a talented young Indian developer who has deep knowledge of the Jupyter-Book project.

## Have any entities involved in this proposal (grantee, PI, etc.) received funding from us before? If yes, please provide details (date, amount, conditions).

QE received a $275,000 USD gift via the Schmidt Futures Donor Advisor Fund (from Charles Schwab), awarded on the 7th of March 2020. This funding has been used to develop textbooks on network economics and dynamic programming that will form part of the curriculum for the current proposal.

# Budget

## What is the total budget of the effort described, and how much of that would be provided by us under this proposal? How much of the remaining budget is confirmed from other funders?

The one year total budget is $726,663 USD. Although the details are not yet confirmed, QE is negotiating to organize low-cost provision of facilities, coordination and administration from local partners.

## What would be the schedule of our payments, including milestones and/or contingencies for each payment?

TO BE ADDED.

## Will our funding be used by the recipient to generate income or develop saleable assets?

As part of the budget, Thomas Sargent and John Stachurski will receive three months salary support to enable them to suspend other commitments and focus entirely on teaching and curriculum development prior to and during the summer school.

## Please provide specific detail on the proposed uses of our funds by the recipient.

Please see the attached budget.

## Risk / Conflict

**Generally speaking, what else could keep this project from succeeding? What can we do to mitigate these risks (e.g., are there conditions that should be put on our support?)?**

Pandemic-related flight and visa issues may cause disruption for inbound travel to India.