About

Joshua Blumenstock is the Chancellor’s Associate Professor at the UC Berkeley School of Information and the Goldman School of Public Policy. His work focuses on applying analytic rigor to serve the public good, particularly in the context of policy research and impact. As a faculty member, he contributes to the academic community through his research and teaching, emphasizing the use of data-driven approaches to address social and economic issues.

Research topics

• Computer Science
• Economics
• Computer Security
• Medicine
• Political Science
• Economic growth
• Artificial Intelligence
• Business
• Finance
• Internet privacy
• Psychology
• Engineering
• Law
• Demographic economics
• Telecommunications
• Criminology
• Data science
• Econometrics
• Geography
• Environmental health

Selected publications

• Predicting Well-Being with Mobile Phone Data: Evidence from Four Countries

  Open MIND · 2026-02-02

  preprint

  We provide systematic evidence on the potential for estimating household well-being from mobile phone data. Using data from four countries - Afghanistan, Cote d'Ivoire, Malawi, and Togo - we conduct parallel, standardized machine learning experiments to assess which measures of welfare can be most accurately predicted, which types of phone data are most useful, and how much training data is required. We find that long-term poverty measures such as wealth indices (Pearson's rho = 0.20-0.59) and multidimensional poverty (rho = 0.29-0.57) can be predicted more accurately than consumption (rho = 0.04 - 0.54); transient vulnerability measures like food security and mental health are very difficult to predict. Models using calls and text message behavior are more predictive than those using metadata on mobile internet usage, mobile money transactions, and airtime top-ups. Predictive accuracy improves rapidly through the first 1,000-2,000 training observations, with continued gains beyond 4,500 observations. Model performance depends strongly on sample heterogeneity: nationally-representative samples yield 20-70 percent higher accuracy than urban-only or rural-only samples.

  DOI

• Insecurity and Firm Displacement: Evidence from Afghan Corporate Phone Records

  American Economic Journal Economic Policy · 2026-01-29

  article

  We provide empirical evidence on how insecurity affects firm behavior by linking data on deadly terrorist attacks in Afghanistan to geolocated data on corporate mobile phone activity. We first develop an approach to estimate the geographic footprint of firms based on employee locations. Using these measures, our main analysis shows that violent shocks reduce local firm presence by both increasing firm exit and decreasing entry. Firms respond most strongly to violence in their “headquarters” districts. We also find suggestive evidence of persistence; stronger impacts in more secure districts; and spillovers, whereby attacks in provincial capitals reduce firm presence in surrounding rural districts. (JEL D22, K42, L11, L96, O18, R32)

  PublisherDOI

• Predicting Well-Being with Mobile Phone Data: Evidence from Four Countries

  ArXiv.org · 2026-02-02

  articleOpen access

  We provide systematic evidence on the potential for estimating household well-being from mobile phone data. Using data from four countries - Afghanistan, Cote d'Ivoire, Malawi, and Togo - we conduct parallel, standardized machine learning experiments to assess which measures of welfare can be most accurately predicted, which types of phone data are most useful, and how much training data is required. We find that long-term poverty measures such as wealth indices (Pearson's rho = 0.20-0.59) and multidimensional poverty (rho = 0.29-0.57) can be predicted more accurately than consumption (rho = 0.04 - 0.54); transient vulnerability measures like food security and mental health are very difficult to predict. Models using calls and text message behavior are more predictive than those using metadata on mobile internet usage, mobile money transactions, and airtime top-ups. Predictive accuracy improves rapidly through the first 1,000-2,000 training observations, with continued gains beyond 4,500 observations. Model performance depends strongly on sample heterogeneity: nationally-representative samples yield 20-70 percent higher accuracy than urban-only or rural-only samples.

  PublisherOA PDF

• Measuring religion from behavior: violence, climate shocks and religious adherence in Afghanistan

  London School of Economics and Political Science Research Online (London School of Economics and Political Science) · 2026-03-03

  article

  Religion plays a fundamental role in society but is often difficult to measure. We develop a novel method for measuring religious adherence that is based on decreases in digital activity during periods set aside for prayer. We apply this approach to a dataset of roughly 23 billion phone calls to study the determinants of religious practice in Afghanistan. We find that religious adherence declines after violent attacks by Islamist insurgents but increases in response to droughts in agricultural regions. This approach creates new avenues for studying religious behavior in contexts where conventional data are unavailable or unreliable.

  Publisher

• Welfare Effects of Digital Credit: A Randomized Evaluation in Nigeria

  Economic Development and Cultural Change · 2025-02-05

  article

  Digital loans, which provide short-term, high-interest credit via mobile phones, have exploded in popularity across low- and middle-income countries. This paper reports the results of a randomized evaluation of a digital loan product in Nigeria. Being randomly approved for a loan (among those who otherwise would have been denied) substantially increases subjective well-being after 3 months, but being randomly approved for a larger loan does not have any effect. Neither intervention significantly affects other measures of welfare, and we can rule out large effects—either positive or negative—on income and expenditures, resilience, and women’s economic empowerment.

  PublisherDOI

• Scalable Targeting of Social Protection: When Do Algorithms Out-Perform Surveys and Community Knowledge?

  SSRN Electronic Journal · 2025-01-01

  articleOpen access
  PublisherDOI

• Scalable Targeting of Social Protection: When Do Algorithms Out-Perform Surveys and Community Knowledge?

  National Bureau of Economic Research · 2025-06-01 · 1 citations

  reportOpen access

  Innovations in big data and algorithms are enabling new approaches to target interventions at scale.We compare the accuracy of three different systems for identifying the poor to receive benefit transfers-proxy means-testing, nominations from community members, and an algorithmic approach using machine learning to predict poverty using mobile phone usage behavior -and study how their cost-effectiveness varies with the scale and scope of the program.We collect mobile phone records from all major telecom operators in Bangladesh and conduct communitybased wealth rankings and detailed consumption surveys of 5,000 households, to select the 22,000 poorest households for $300 transfers from 106,000 listed households.While proxy-means testing is most accurate, algorithmic targeting becomes more cost-effective for national-scale programs where large numbers of households have to be screened.We explore the external validity of these insights using survey data and mobile phone records data from Togo, and cross-country information on benefit transfer programs from the World Bank.

  PublisherDOI

• Probing the limits of mobile phone metadata for poverty prediction and impact evaluation

  Journal of Development Economics · 2025-02-04 · 3 citations

  articleOpen access

  A series of recent papers demonstrate that mobile phone metadata can, together with machine learning, estimate the wealth of individual subscribers and accurately target cash transfer programs. In the context of an emergency cash transfer program in Haiti, we combine surveys and mobile phone call detail records (CDR) to test whether such methods can be used to estimate the program’s impact on household expenditures . We find that CDR-based predictions of total and food expenditures are much less accurate than predictions of wealth—particularly when estimated on a relatively homogeneous sample of rural communities eligible for the program. While impact estimates based on conventional survey data are positive and statistically significant, estimates based on CDR predictions are not statistically significant. In a postmortem discussion, we assess reasons for this failure and discuss the implications for using big data in poverty measurement and impact evaluation. • Mobile phone data and machine learning are better predictors of wealth than expenditure. • Prediction accuracy goes down when sampling from primarily poor households. • A tradeoff exists between data useful for program evaluation and for welfare prediction.

  PublisherDOI

• Expanding Perspectives on Data Privacy: Insights from Rural Togo

  Proceedings of the ACM on Human-Computer Interaction · 2025-05-02

  articleOpen accessSenior author

  Passively collected "big" data sources are increasingly used to inform critical development policy decisions in low- and middle-income countries. While prior work highlights how such approaches may reveal sensitive information, enable surveillance, and centralize power, less is known about the corresponding privacy concerns, hopes, and fears of the people directly impacted by these policies --- people sometimes referred to as experiential experts. To understand the perspectives of experiential experts, we conducted semi-structured interviews with people living in rural villages in Togo shortly after an entirely digital cash transfer program was launched that used machine learning and mobile phone metadata to determine program eligibility. This paper documents participants' privacy concerns surrounding the introduction of big data approaches in development policy. We find that the privacy concerns of our experiential experts differ from those raised by privacy and development domain experts. To facilitate a more robust and constructive account of privacy, we discuss implications for policies and designs that take seriously the privacy concerns raised by both experiential experts and domain experts.

  PublisherDOI

• What Would it Cost to End Extreme Poverty?

  SSRN Electronic Journal · 2025-01-01

  preprintOpen access
  PublisherDOI

Recent grants

• CAREER: Welfare-Centric Machine Learning

  NSF · $451k · 2020–2025

Frequent coauthors

• Michael Callen

  21 shared

• Daniel Björkegren

  19 shared

• Emily Aiken

  Berkeley College

  19 shared

• Tarek Ghani

  18 shared

• Suzanne Bellue

  14 shared

• Dean Karlan

  10 shared

• Nitin Kohli

  8 shared

• Nathan Eagle

  Tallaght University Hospital

  8 shared

Awards & honors

• NSF CAREER Award
• Intel Faculty Early Career Honor
• Gates Millennium Grand Challenge award
• Google Faculty Research Award
• UC Berkeley Chancellor's Award for Public Service