Pre-Doctoral Fellowship (Prof Breza and Viviano)

PhD @Harvard University posted 2 months ago

Job Description

Position Description

The fellow will work on projects for both Breza and Viviano spanning development economics and econometrics. We provide three examples below of projects the fellow may work on.

1. “Meta-analysis and evidence aggregation” (Breza and Viviano)

Description:

The goal of this project is to develop tools for meta-analysis to be able to pool evidence across experiments conducted in different contexts (e.g., across different countries or with different individuals). The project’s goal is to develop new statistical techniques and implement these across a wide range of datasets.

Specifically, we aim to study the identification of generalizability: the ability to take a dataset and, at a given resolution, partition the data into a set of observations for which effects are generalizable and observations that exhibit non-predictable heterogeneity. For instance, we can ask whether in a given location, patterns exhibited by a certain sub-population in a given country generalize to other countries, or whether patterns in small scale experiments generalize to certain large-scale experiments. Our goal then is to form predictions of scientific mechanisms, by allowing for ignorance: perhaps for some individuals and some environments, we are unable to claim that effects generalize and instead prefer recommending further scientific investigation.

We will then apply our theoretical framework to the analysis of existing experiments to draw conclusions that are generalizable across a wide variety of interventions in development economics (e.g., cash transfer interventions, training programs, etc.)

Role of the predoctoral fellow:
through statistical software and the (meta-)analysis of existing interventions using our proposed tools. The fellow will also be involved in parts of the literature review and data collection.

2. “Using remote sensed variables for program evaluation” (Viviano)
Description:
While traditional program evaluations directly measure outcomes, certain economic outcomes such as environmental quality or living standards may be infeasible to directly measure. Empirical researchers often estimate treatment effects using remotely sensed variables (RSVs), such satellite images or mobile phone activity, instead of direct measurements of the economic outcome. Common practice predicts the economic outcome from the RSV, using an auxiliary sample of labeled RSVs, and then uses such predictions as the outcome in the experiment. As we show, this approach is biased when the RSV is a post-outcome variable. Motivated by this result, we provide a novel method that identifies and estimates the treatment effect, using an assumption inspired by common practice: the conditional distribution of the RSV remains stable across both samples. We then propose a statistical technique to efficiently represent (compress) the RSV such as satellite images into simple objects that can be used for program evaluation. We study the properties of the method in a large-scale public program in India using satellite images.
 
Role of the predoctoral fellow:
The fellow will be implementing our statistical technique by developing a statistical software (e.g., R package). In addition, he will be involved in the data-analysis through data collection (e.g., collecting satellite images) and analysis. We expect the fellow to become familiar throughout this project with statistical techniques for causal inference and high dimensional data.

3. “Impacts of weather insurance on adaptation, social networks and migration” (Breza)
Despite its appeal as a policy to cope with adverse climate shocks, weather insurance at scale remains elusive in most LMICs. We study the Indian government’s crop insurance program (PMFBY), where enrollment rates remain low seven years into the policy’s rollout. The Odisha state government recently increased the insurance subsidy, bringing farmer premiums down to Rs.1. In summer 2024, we ran an experiment across more than 230 villages in flood-prone areas identified via satellite data. In treated villages where we distributed flyers and facilitated community meetings to discuss enrollment procedures, administrative data show enrollment increased by 40% and insured area increased by 30% relative to control. These gains were concentrated among small-holder farmers, and treatment drastically increased enrollment rates among SC/ST farmers, a group often lacking access to formal financial instruments. In spring 2025, we will survey these households to measure impacts on agricultural investments and harvest as well as on migration, risk sharing and social network relationships.

Role of the predoctoral fellow:
The pre-doctoral fellow will play a central role in helping with the analysis of this experimental data. Moreover, we plan to use publicly available village-level data at a national level to understand insurance take-up dynamics and broader impacts on rural development and migration. Changes to state-level subsidies and weather conditions yield natural experiment-type variation for a differences-in-differences analysis. The fellow will also work on assembling and analyzing this data.
Basic Qualifications
  • Undergraduate degree in Statistics/Economics/Applied Math/Computer Science or related fields
  • Knowledge of regression analysis, statistical inference, familiarity with machine learning/prediction tools
  • Experience with statistical software (e.g., R/Python/Matlab)
  • Writing sample
  • At least 1 reference who can speak to your research skills
Additional Qualifications
  • Topical exposure/interest in development economics
  • Experience with web-scraping
  • Topical exposure/interest in statistical learning
Special Instructions
Applications will be accepted and reviewed on a rolling basis.
Please include the following:
1. CV
2. Official or unofficial transcript
3. Cover letter
4. Writing Sample
Contact Information
Shree M
Contact Email econacademicappointments@fas.harvard.edu
Equal Opportunity Employer
We are an equal opportunity employer and all qualified applicants will receive consideration for employment without regard to race, color, religion, sex, national origin, disability status, protected veteran status, gender identity, sexual orientation, pregnancy and pregnancy-related conditions or any other characteristic protected by law.
Minimum Number of References Required 1
Maximum Number of References Allowed 3
Keywords

Supplemental Questions

Required fields are indicated with an asterisk (*).

Applicant Documents

Required Documents

  1. Curriculum Vitae
  2. Cover Letter
  3. Transcript
  4. Writing Sample 1

Optional Documents

familiarity with machine learning/prediction nowledge of regression analysis statistical inference

Related Jobs

 
Required 'Candidate' login to applying this job. Click here to logout And try again
 
 
 
 

Apply for this Job

  • Drop a resume file or click to upload.
    To upload file size is (Max 5Mb) and allowed file types are (.doc, .docx, .pdf)
    or
    Upload Resume
  • Upload Cover Letter

    To upload file size is (Max 1Mb) and allowed file types are (.doc, .docx, .pdf)

  • By clicking checkbox, you agree to our Terms and Conditions and Privacy Policy

 

Answers

 

Job Alerts

 

Account Activation

Before you can login, you must activate your account with the code sent to your email address. If you did not receive this email, please check your junk/spam folder. Click here to resend the activation email. If you entered an incorrect email address, you will need to re-register with the correct email address.