The Importance of Infrastructure for Scalable Policy Solutions

In “Will It Scale?” (Issues, Fall 2024), Omar Al-Ubaydli and John A. List rightly underscore the complex nature of scaling successful programs. In light of this thoughtful work, I would like to expand on the critical role that investing in program infrastructure plays in supporting scalable solutions.

Infrastructure is essential for scaling interventions effectively because it helps programs adapt to variability in policy and economic contexts while maintaining program fidelity. Robust infrastructure—including technological platforms, distribution networks, and trained personnel—is necessary to ensure that interventions can be deployed consistently and flexibly at scale. For example, digital tools that streamline implementation and facilitate real-time feedback can significantly enhance a program’s ability to adjust to diverse policy environments without compromising its core evidence base. This adaptability is crucial for maintaining program impact across different regions.

Consider the case of large-scale vaccination drives. These efforts rely heavily on infrastructure such as cold chain logistics for keeping vaccines viable, data management systems for tracking distribution, and trained health workers to administer vaccines and address local concerns. Similarly, educational interventions that scale successfully often utilize online learning platforms that provide standardized content while allowing for localized teaching adjustments.

Equally important are the economic factors underpinning scalability. Solutions must be designed with sustainability in mind, balancing initial investments with long-term cost efficiency. Public-private partnerships can provide vital support, bridging resource gaps and fostering shared accountability. Policies that incentivize investment in scalable solutions, such as tax credits or grant mechanisms, can encourage broader adoption and financial commitment.

Investing in infrastructure also involves ensuring that training programs for personnel are in place to support the scale-up process. Well-trained staff contribute to the consistency and quality of program delivery, which helps uphold program fidelity even as interventions are adapted to new contexts. This investment in human resources is as critical as technological or logistical infrastructure for achieving sustainable scaling.

Collaboration across sectors is often necessary to build this comprehensive infrastructure. Partnerships between governments, private entities, and nonprofit organizations can establish the foundations required to support large-scale implementation. Such collaborations can foster innovation, resulting in tools and systems tailored to meet the varied needs of different communities.

To achieve sustainable scaling, it is imperative that policymakers and practitioners prioritize investment in program infrastructure during the design phase. Infrastructure provides the adaptability needed to navigate policy and economic variability while ensuring that interventions maintain their integrity and effectiveness. By embedding adaptive strategies, robust financial planning, and infrastructure support from the outset, we can create a pathway that extends the reach of effective programs and sustains their impact.

Max Crowley

We join Omar Al-Ubaydli and John A. List in their call for more rigorous scaling research. The complexities they outline—external validity considerations (will the intervention work in a different context?), spillover effects to nonbeneficiaries, and implementation issues—are key to the success of scaling interventions.

We have had experience with these issues. In 2008, one of us (Mobarak) ran a pilot study in Bangladesh to encourage rural-urban migration and prevent seasonal hunger during the agricultural lean season. Researchers distributed subsidies of $11.50 (US dollars) to study participants—enough to pay for the round-trip bus fare plus a few days of food. The effects were startling: In treatment groups, migration percentages increased to 58% versus 36% in the control group, and families of migrants consumed more than 600 extra calories per person per day versus the control group. Subsequent studies showed promising results. Then in 2017, a large nongovernmental organization scaled the program to 150,000 households. The results were disappointing and the program was discontinued.

As social scientists, we viewed the experiment’s failure as an opportunity to delve deeper for answers—and thus was born the Yale Research Initiative on Innovation & Scale (Y-RISE), with a mission to advance the science of scaling development interventions. Here, we offer a few more lessons from our research in scaling policy interventions in low- and middle-income countries.

First, public-sector programs interact more directly with the political landscape than the private-sector interventions that al-Ubaydli and List describe. The path to scale often runs through governments, which are both the “doers” delivering the services and the “payers” at scale. Governments are made up of political leaders with their own constituents and incentives and may therefore reallocate effort or financial resources in a program, potentially enhancing or undermining it. In a large-scale study focused on investments in improved sanitation, local leaders spent more time in communities that had received latrine subsidies. Constituents expressed greater satisfaction with leader performance, ostensibly attributing the program performance to leaders. Understanding these political considerations is critical to scaling policy interventions.

Second, in policy interventions, broad social changes, in addition to economic changes, are part of the overall cost-benefit calculus. In the case of migration, there are numerous social effects, including impacts on risk-sharing, gender norms, political beliefs, and costs such as family separation and risk of divorce.

Third, scaling programs requires a deep understanding of implementation considerations. One such is of partners’ incentive structures. The migration program failed at scale because the NGO was unable to target the right beneficiaries, instead offering subsidies to those who would have migrated anyway. Another implementation consideration is that policy programs are often not single-mechanism interventions tested in research but rather a “bundle” of interventions implemented together, so we need to better understand how the entire policy works together, not just individual components.

These complexities require us to expand our methodological toolbox. In the migration example, the harm of family separation may not be directly observed; thus, we may need to borrow from macroeconomics to mathematically model disutility (capturing harmful or adverse effects) and infer its magnitude to calibrate experimental data.

We urge researchers and policymakers alike to think of scaling as a process rather than an outcome of a successful pilot. Scaling should be thought of in stages, with researchers developing new experimental designs at each step to understand broader effects. The challenges of this approach are that the “sunk cost fallacy” makes it psychologically difficult to stop programs once started, unless agreed to in advance.

Al-Ubaydli and List have made it clear why we need scaling research. At Y-RISE, we see this call as urgent: the 700 million people globally who live in extreme poverty demand no less.

Neela A. Saldanha

Ahmed Mushfiq Mobarak