“A robust education system provides learning opportunities for all children. Education is a pathway to social mobility, especially for those from disadvantaged backgrounds. While the education sector has historically been slow to evolve, disruptions like AI and the climate crisis present opportunities for transformation.”

Dayoung Lee is a Partner at Dalberg Advisors, based in Goa, India, where she co-leads the firm’s global Education to Employment Practice. Her work spans designing an outcomes fund to accelerate foundational learning through systemic change, advising a leading Asian philanthropy on how they can equip the next generation to be AI-ready, and supporting the World Bank and the Indian state of Maharashtra with an ambitious reform of its skilling ecosystem. She has been instrumental in designing gender and disability-inclusive solutions, including a playbook for mainstreaming inclusive education and a strategy to address gender norms for a nonprofit advancing career guidance and exploration. 

Before joining Dalberg, Dayoung was a consultant with The Parthenon Group (now EY-Parthenon), focusing on the education sector across the United States and Asia. She holds a joint MPP/MBA from Harvard Kennedy School and Harvard Business School and graduated with Honors in Economics and International Relations from Stanford University. 

In this interview, Dayoung talks about upcoming trends that can reshape the education sector, the opportunities and challenges that come with technology, and effective impact measurement metrics. 

1. You’ve led one of the world’s largest education development impact bonds. Are you seeing a lot more outcomes-based financing in the education to employment space? Why or why not?  

Yes, we are seeing more outcomes-based financing mechanisms that reward individuals or institutions after the credible verification of an achieved result to support and invest in the education to employment space. This sector is a leader in using innovative and outcomes-based finance solutions such as impact bonds, impact-linked loans/finance, performance-based contracts, risk-mitigation facilities like guarantees, and income share agreements. Three factors are driving this growth. First and perhaps most importantly, funders are increasingly interested in ensuring outcomes and are looking for evidence of impact. Many of these alternative models allow for greater accountability towards the investment and greater learning to enhance impact. Second, there is demand for more flexible funding, and for private companies, funding that allows servicing communities with lower ability to pay. Models like impact bonds that are de-linked from inputs and activities allow for organizations to innovate further and refine their models to maximize impact. Impact linked loans allow private companies to access cheaper capital to serve lower income groups who otherwise may not be viable economically. Finally, mechanisms like impact bonds and guarantees can allow right sizing the risk across parties with different risk profiles for more innovative solutions. However, it is still a small share of overall education financing.

The first barrier is that too often, funders must give out a certain amount every year. For example, CSRs in India are mandated to spend certain portion of their profits every year. If a local government saves too much money because certain outcomes were not achieved, their budgets from the central government may reduce in the following year, which disincentivizes them from paying based on outcomes. Philanthropies, multi and bilateral donors also have commitments, and budgets they are responsible for deploying each year. While there are ways to get around it, it can be easier to stick with grants or input-based payments with reliable expenditures. Second, it can be difficult and costly to measure outcomes and structure the funding. It can be challenging to identify the right metrics and measure them at a reasonable cost, assess the value of the outcomes and price them as well as the risk of not meeting outcomes, get the right stakeholders around the table, and then get them to align on the terms of the agreement. Such costs are often seen as transaction costs, even though the alternative may be spending 20% less and getting no outcomes at all, and there are often many positive externalities (e.g., organization wide learning around what works that can be replicated outside of the outcomes-based contract). Finally, the pipeline of providers that are focused on outcomes and can adapt to changing needs to manage them well may be slim, depending on the market and outcome area. More funders asking for outcomes and helping build implementing organizations’ capacity to manage them will solve this challenge over time. 

2. Given the complexity of measuring educational outcomes, how do you establish clear and robust impact metrics when you design outcomes-based financing mechanisms in this space? 

Some educational outcomes, such as foundational literacy and numeracy (FLN), have well established and widely accepted metrics for success. There are validated tools like Early Grade Reading Assessment (EGRA) and Early Grade Math Assessment (EGMA) that are simple, low-cost, and orally administered. Our tools in the LiftEd Development Impact Bond are modeled after these but adapted for our specific needs.  

In our LiftEd DIB, we wanted to do more—we wanted to incentivize not only one-time student level learning gains but lasting systemic change that will sustain impact for future generations. These shifts are much harder to measure as there is far less consensus on what systemic levers lead to student learning, and what level of change to expect in a specific time frame. Hence our approach was to use the well-established student level FLN outcome measures alongside developing and refining systemic shift indicators (SSIs). These include indicators like quality review meetings between teachers and principals on classroom practices. While on the one hand, these review meetings are just a step towards students in those classrooms improving FLN, on the other hand, they are important signals that key stakeholders’ behaviors have shifted in a way that will last. For example, if high quality discussions between principals and teachers on their instructional practices have shown to improve FLN, the education system now mandates and expects these meetings, and teachers and principals are consistent holding them, that signals the system has shifted fundamentally in a way that can continue improving FLN of students that come after the DIB investment.  

We have co-designed these indicators in extensive consultation with our implementing partners and had a learning year to adjust them based on their experience of using them. This approach provides a more holistic view of progress on strengthening the system at multiple levels. It has been our key innovation and contribution to the education ecosystem. 

More innovation is required to measure other important outcomes that education systems should be held accountable to, such as socioemotional wellbeing and 21st century skills like creativity and critical thinking. While there has been a lot of progress in measuring these skills, we need greater consensus on what success looks like, and more cost-effective, scalable ways to measure them. I am excited about a new outcomes-based instrument we are designing at Dalberg that will push this agenda—tying these types of outcomes to funding raises the stakes, so it will accelerate the robustness of measuring these important outcomes. This will in turn enable organizations to better understand whether and how well they are achieving them and improve their programs.  

3. Based on your experience advising on India’s EdTech policy and designing inclusive technology solutions, what elements do you believe are essential for creating an effective and inclusive EdTech ecosystem at the national level?

There needs to be a shift in focus towards what drives quality. Relative to its hype, studies show that investments in EdTech are too often failing to deliver any outcomes. Currently, much of the emphasis and budget allocation in school systems is on hardware—devices and infrastructure. However, it is the quality of the software and people that truly drive educational outcomes, and this often comes at a cost. Prioritizing investment in inclusive, high-quality software such as Personalized Adaptive Learning tools and a strong layer of teacher training and on-ground staff support is key to ensuring that EdTech is both effective and accessible to all. 

It’s also important to recognize that neutral is not enough when it comes to EdTech. Unless we design technology with an inclusion lens from the outset, the digital divide in gender, children with special education needs, children in lower income and remote rural areas and so on will only exacerbate learning inequalities. Our study on the next half billion (NHB) users—300 million people who have come online since 2017—highlighted persistent significant gaps, especially along gender lines and for persons with disabilities. For instance, in a recent impact study for an investor in an Indian EdTech company, we found that girls in India from tier 2 and tier 3 cities who face challenges traveling long distances to tutoring centers, are among the biggest beneficiaries of the company’s services. However, one major lesson from our NHB study is that simply providing access to technology isn’t enough—systems must be designed to account for the social norms that restrict access for women and girls. For example, concerns about online safety, such as fears that daughters might be exposed to inappropriate content or contacted by men with ill intentions, are prevalent among parents and community members.

Moreover, there are genuine safety and security concerns that must be solved. For example, we learned that during COVID-19, many young female teachers had their phone numbers widely shared on WhatsApp groups of parent-teacher communities—however, this also left them vulnerable to harassment. So, we might suggest an innovation like phone number masking which would allow teachers to communicate more securely with students’ families. This innovation might only empower women teachers, but it also sets a precedent for other sectors where digital access must be balanced with security. Another challenge arises with the advent of algorithms on shared devices. Since content is often prioritized based on usage, women and girls, as minority users of shared devices, may find the content most relevant to them deprioritized in favor of content more relevant to male users. It’s important to develop features that allow for the pinning of important content for women and girls or creating separate account sign-ins that recognize different users. In our study supported by the Ford Foundation and IWWAGE (Initiative for What Works to Advance Women and Girls in the Economy), we captured such design principles for making digital tools work for women. 

For people with disabilities, many current software and apps fail to provide inclusive assistive features, such as multimodal sensory options (auditory or visual captions) that cater to diverse needs. Any digital public goods should have accessibility features that allow persons with disabilities to use these tools. Simple yet powerful design elements, like using photos instead of abstract icons, can make technology more accessible. 

4. Looking ahead, what trends or developments are you most excited or concerned about, and how do you plan to incorporate them into your work?

I am most concerned about climate change, cautiously optimistic about AI, and excited about the momentum in diversity, equity, and inclusion (DEI) efforts.  

Climate change is no longer a problem for the next generation—it is already deeply affecting our children’s learning. It is shortening school days, damaging school infrastructure, increasing stress levels and in extreme cases, leading to climate refugees. Hence, educational interventions need to account for such climate-induced learning discontinuities and losses. On the other hand, education and skilling ecosystems have an important role in ensuring that our young people have the necessary green mindsets, behaviors, and skills to both mitigate and adapt to climate change.  

AI is a major technological innovation that is rapidly transforming how we learn, live and work. AI is becoming more powerful, outperforming humans on not only basic cognitive tasks but also highly analytical and even creative tasks that were previously considered uniquely human. We need to equip our young people to be more AI-ready so they can figure out what humans can do best vs AI and work with AI in way that elevates the quality and speed of what needs to be done. They also need strong ethical principles to discern what AI should (vs. can) do and understand issues like bias and privacy to protect themselves and use AI responsibly and equitably. AI is also transforming how students learn – by personalizing and adapting instruction to each student’s learning level, reducing the administrative burdens on teachers and so on. Yet, AI is exacerbating existing inequalities – those without access to these tools and training are likely to lose from this advancement. For example, AI systems are primarily trained on few high-resource languages (e.g., English, French, Mandarin), excluding speakers of underrepresented languages. Only 1 in 20 students in low-income countries have reliable internet at home vs. 90% in developed countries, making it much harder for low-income students to have access to GenAI tools. Even with equal access, women and girls are less likely to use these tools – women use GenAI tools up to 20% less than their male peers in same jobs. At Dalberg, we’re exploring how AI can be responsibly integrated into public education systems with a specific focus on underserved communities, ensuring that AI benefits all students and educators. 

While both climate and AI will worsen inequalities if we are not far more intentional, I’m hopeful that there is growing focus on diversity, equity, and inclusion (DEI) amongst funders and in the education system. This momentum aligns with Dalberg’s longstanding commitment towards equitable growth. While the gender equality movement started to gain traction in the past 50 years, the momentum around DEI more widely, particularly in light of the US racial justice movement, is now making waves globally. There’s now greater policy recognition of the rights of marginalized children, including children with disabilities in low- and middle-income countries (LMICs). We aim to build on this momentum, with the private and public sector enabling states to build capacity with an inclusion lens. We are actively working on initiatives like the Inclusive Education Fund with experts such as Sol’s ARC, which aims to mainstream DEI into national education agendas, moving beyond treating it as a niche or last-minute consideration. 

By centering the DEI lens into everything we do, I hope we can better protect our most vulnerable young people from the emerging risks and help them take advantage of new opportunities like green and AI-powered jobs. 

Connect with Dayoung Lee to learn more about Dalberg’s work in education and employment: