Working with governments to rapidly deliver masks to millions of people | Maha Rehman (2021)

Project Overview and Current Focus

• Maha Raymond led the rapid scale-up of the "NORM" mask-wearing promotion program in Lahore, Pakistan, in June 2021, prior to the Delta variant surge.
• The program adapts a successful Randomized Control Trial (RCT) from Bangladesh, funded by GiveWell and the Effective Altruist Funds, to an urban setting.
• Raymond is currently leading a data-driven social mobilization drive to increase vaccine uptake in Lahore, utilizing 1,000 volunteers trained from 25 local universities.
• The vaccine campaign leverages an existing supply of seven vaccine types with no availability bottlenecks, shifting focus from access barriers to behavioral compliance and complacency.

Bangladesh RCT Findings (The "NORM" Model)

• Core Intervention: The model relies on four components: No cost distribution, Offering information, Reinforcing messages in the field, and Modeling behavior via role models.
• Mask Adoption: The intervention increased mask-wearing from 10% to 40% among 350,000 people across 600 villages.
• Health Impact: The 30% increase in mask wearing reduced symptomatic COVID-19 infections by 10% in the community.
• Demographic Specifics: The infection reduction effect was 35% for individuals aged 60 and older, and surgical masks proved approximately 50% more effective than cloth masks.
• Behavioral Insights: Mask wearing was associated with increased social distancing, contradicting "risk compensation" theories; individuals wearing masks were also more likely to be vaccinated first.
• Failed Interventions: Providing financial incentives to officials and using police for reinforcement did not show statistically significant increases in mask uptake.
• Aesthetic Factors: The color of the mask (e.g., purple vs. green) was a significant driver of adoption, suggesting fashion trends influence compliance.
• Sustainability: The increase in mask-wearing behavior persisted for 10 weeks after the intervention ended.

Urban Adaptation in Lahore

• Logistics Shift: Unlike the rural Bangladesh model, Lahore utilized the postal system to deliver masks to households, using billing data from Lahore Electric Supply Company (LESCO) for verified addresses.
• Funding Model: No masks were purchased; all were donated by corporate and philanthropic donors in Lahore.
• Enforcement Strategy: The campaign utilized "flag marches" (administrative officers patrolling markets) and mobile trucks with announcements rather than household-to-household visits due to high mobility costs.
• Mosque Strategy: To address low compliance inside mosques, promoters distributed masks post-wudu (ablution) and required explicit announcements before prayers began.
• Data Collection: Observers recorded mask-wearing rates every 10 minutes in markets and mosques using a mobile app with built-in quality checks (audio recording, time-stamping) to ensure data integrity.
• Scalability: The success led to a formal directive from Pakistan's National Command and Operation Centre (NCOC) ordering other districts to replicate the model.
• Upcoming Scale: A shipment of 24 million masks is en route to Lahore, the largest in Pakistan's history, aiming to cover every household.

Government Procurement Research

• Study Scope: A field experiment across five government departments in Punjab measured waste in petty procurement (under 100,000 PKR), finding up to 200% price variation for identical goods in the same district.
• Root Causes: Inefficiency stemmed less from active corruption and more from passive waste caused by complex rules, lack of financial literacy, and delays by Accountant General offices holding up payments.
• Intervention Results: Relaxing rules and simplifying procedures (e.g., the "AG Checklist") significantly reduced prices.
• Incentive Limitations: Performance incentives for officers only worked when the incentives of the supervisors (monitors) were aligned; if monitors were obstructive, incentives failed.
• Systemic Solution: The experiment highlighted that automating procurement to reduce face-to-face interaction would eliminate delays and bribery risks, though legal changes are required to implement this.
• Implementation Barrier: Overcoming bureaucratic resistance to automation required a smart coalition and legal advocacy, taking 1.5 years to secure a single permission.

Operational Lessons and Challenges

• Stakeholder Management: Success relied on identifying a single "owner" (Lahore Commissioner) with political capital and aligning incentives across multiple departments.
• Delegation: Key decisions remained with the core team, but implementation (e.g., flag marches, media events) was delegated to the government to avoid bottlenecks.
• Data Quality: Rapid "action over perfection" was necessary; underperforming data firms were quickly replaced, and monitoring systems included random audits.
• Mental Load: The high-stakes nature required rigorous personal time management, including running and mindfulness, to prevent burnout during the crisis.
• Failure Management: The team accepted that not all battles could be won, focusing resources on willing partners rather than forcing resistant officials.

Future Directions

• The "NORM" infrastructure is being repurposed to formalize volunteer networks for vaccination drives, with universities offering academic credit for participation.
• The management information system developed for the procurement study is now being adopted by the government for broader training.
• Raymond plans to shift focus to energy sector loss reduction in Pakistan, applying the same data-driven, field-experiment methodology to new sectors.