The Hon Andrew Leigh MP
Assistant Minister for Productivity, Competition, Charities and Treasury
Randomising Our Way to Better Policy
Asian Evaluation Week 2026,
Asian Development Bank, Manila
Online Address
Tuesday, 1 September 2026
Thank you for the opportunity to speak at this year’s Asian Evaluation Week.
I thank the Asian Development Bank for bringing together policymakers, practitioners and researchers from across our region.
This year's theme, ‘Put Your Money Where Your Evidence Is’, is a powerful reminder that we have a responsibility to direct funding and resources towards programs that make a real difference.
At its core, evaluation is about answering a straightforward question: did a policy or program improve people's lives?
That question is surprisingly difficult to answer. We can observe what happened after an intervention was introduced, but we cannot directly observe what would have happened in its absence. Rigorous evaluation helps bridge that gap, because it gives us a credible counterfactual.
Evaluation moves us beyond assumptions. It helps us distinguish between correlation and causation.
Just as importantly, evaluation encourages humility.
Many policies that appear promising fail to deliver meaningful results when tested rigorously. Equally, some simple and inexpensive interventions produce benefits that exceed expectations.
Rigorous evaluation helps policymakers become more curious.
Governments (and multilateral institutions for that matter) should be learning institutions. As stewards of the public good, we should be constantly improving our understanding of what works and why.
That principle has guided Australia's efforts to strengthen evaluation across government, including through the Australian Centre for Evaluation.
It also underpins the Asian Development Bank's longstanding commitment to improving development effectiveness through rigorous evidence and impact evaluation.
I particularly want to acknowledge the work of Emmanuel (Manny) Jimenez and his inspiring leadership as the Director General of the ADB’s Independent Evaluation Department.
From Generating Rigorous Evidence to Using It
When I spoke at the Asian Development Bank and the Asian Development Bank Institute in Manila in July 2024, I argued that rigorous evaluation is essential for understanding whether development programs improve lives (Leigh 2024a, 2024b).
Today, I want to focus on the next step.
Generating rigorous evidence is important.
Using rigorous evidence is essential.
Too often, evaluations conclude with a report. The big challenge is ensuring that findings influence decisions about which programs should be scrapped, and which should be scaled up.
To achieve that, evidence must be integrated into the way institutions allocate resources.
It should inform budget decisions and investment priorities.
In short, we must put our money where our evidence is.
Three Lessons from Recent Evaluations
Let me offer three examples.
First, evidence can help governments improve the fairness of the tax system. In the Democratic Republic of the Congo, researchers worked with local authorities in a large Congolese city to test different approaches to property taxation (Tourek et al 2026). Neighbourhoods were randomly assigned to either a progressive property tax schedule or a proportional one.
The progressive approach increased revenue by around 56 per cent relative to the proportional schedule, and after it was scaled up alongside targeted enforcement, shifted the effective tax burden towards wealthier property owners.
Most importantly, the findings were used. The progressive tax schedule was subsequently expanded across the city, demonstrating how rigorous evidence can directly inform public policy.
Second, small interventions can sometimes make a big difference.
A study in Zambia considered the budgeting challenge for farmers who derive their income from one annual harvest and then spend it over the course of the year (Augenblick et al 2026).
Researchers found that many households underestimated upcoming costs, not because they lacked information, but because they found it difficult to recall their past spending at the appropriate moment to inform their future spending plans.
The intervention was pretty simple: a randomly selected group of households were guided through a structured exercise to think about future expenses.
That small prompt increased savings by 15 per cent, and allowed households to enter the annual ‘hungry season’ with approximately one extra month of savings.
Households also invested more in farming, resulting in a 9 per cent increase in crop revenue the following year. The researchers replicated the belief-updating effects in a quite different population – low-income Americans – suggesting the intervention has broader applicability. Effective policy does not always require complex solutions.
Third, evaluation helps us understand not only what works, but how long the effect lasts.
In Kenya, researchers looked at the issue of road safety, which is a major public health challenge in many countries where traffic laws exist on paper, but enforcement can be inconsistent (Schönholzer et al 2026).
In this trial, Kenyan minibus drivers were randomly assigned to receive financial incentives for safer driving, measured using vehicle monitoring technology.
The incentives reduced dangerous driving behaviours. For example, drivers who received the incentive reduced speeding by 29 per cent. And there was little evidence that drivers simply shifted their risky behaviour to unmonitored places or times.
But the study also found that the effects diminished once the incentives were removed.
No policy lasts forever, and a rigorous randomised trial can help get at impact duration as well as impact magnitude.
The Role of the Asian Development Bank
The Asian Development Bank has been a leader in building evaluation capability across the region, supporting impact evaluations, developing practical guidance and fostering collaboration between governments and researchers.
This work helps ensure that development decisions are informed by credible evidence.
It also enables institutions to learn and adapt. Learning institutions make better use of scarce resources and ultimately, they deliver better outcomes for communities.
The Asian Development Bank's commitment to rigorous evidence is also reflected in randomised trials published through ADB and ADBI in recent years.
In the Philippines, researchers tested computer-assisted learning in remote schools and found meaningful improvements in mathematics achievement (Glewwe et al 2025).
In the People's Republic of China, a randomised evaluation of a village banking program showed that access to microcredit increased incomes and reduced poverty (Cai et al 2025).
In Bangladesh, phone-based learning interventions demonstrated that parents can play a powerful role in supporting children's education, even when schools are disrupted (Hassan et al 2024).
And another Bangladesh experiment showed that community participation can improve access to safe drinking water, but only when decision-making processes are designed well and local communities have meaningful incentives to engage (Habib et al 2023).
These studies remind us that rigorous evaluation is not confined to one sector. It is a practical tool for improving decisions across the full spectrum of development policy.
Rigorous Evidence as a Regional Public Good
In the health sector, it would be untenable to imagine that every randomised trial needed to be replicated in every country.
People are different, but there is a lot we can learn from other nations.
Likewise, many of the social policy challenges faced by countries across Asia and the Pacific are shared.
Questions about tax capacity, road safety, poverty reduction, service delivery and economic opportunity arise in many different contexts.
Not every result will transfer perfectly from one country to another. But a rigorous study from overseas can often be more useful than low-quality evidence produced locally.
When countries share findings and replicate successful approaches, evidence becomes a regional public good.
A sign of the expansion in this public good is the growth of randomised trials. In the 1990s there were fewer than 25 randomised trials from developing countries published globally each year. By 2012, there were 274 published (Banerjee and others 2016).
All up, the Abdul Latif Jameel Poverty Action Lab at MIT has led more than 2,600 randomised trials (J PAL 2026). And J-PAL is just one of many organisations running randomised trials. The American Economic Association’s randomised trials registry lists more than 12,000 registered social science trials, with over 1,600 added last year alone.
Evaluation in the Australian Government
The Australian Government, through the Australian Centre for Evaluation, is also doing its bit to generate – and use – rigorous evaluation evidence.
So far ACE has directly supported seven randomised trials, with four more in progress, and more planned. These evaluations have provided evidence on a range of policy problems, from the design of employment services to improving government service delivery and reducing regulatory red tape. And in each case, the evidence fed back into the government’s decision-making process.
ACE has also built a professional network for public servants – the Evaluation Profession – which has grown to more than 8,000 members. A public repository of Australian government evaluations now hosts over 200 publications, and growing.
Conclusion
Fifteen years ago, randomised experiments dramatically improved the way aid workers distribute bed nets to prevent malaria. A critical question was whether it was better to charge a small co-payment because people who paid would be more likely to use the nets.
Researchers established that people who received free bed nets were just as likely to use them as those who made a co payment. But because they were free, many more people took them up (Cohen and Dupas 2010).
That translated into practice and has helped save thousands of lives across Africa and the rest of the developing world.
This is just one example of why we need to put our money where the evidence is.
The goal of evaluation is better lives.
Whether that’s improving public services or making more effective development investments, we can achieve better outcomes for the communities we serve.
If we are serious about putting our money where our evidence is, then evidence cannot be treated as an afterthought.
It should inform the decisions that matter most.
Because the burden of ineffective policy falls upon people.
We have a responsibility to families who depend on public services.
To communities that need opportunities.
And to citizens who rightly expect governments to be constantly learning and improving.
Evaluation will not eliminate uncertainty, but it gives us a rigorous method to reduce uncertainty over time.
Evidence matters - because it helps us build a better future for those who need government most.
Thank you.
Acknowledgements
My thanks to officials from the Australian Centre for Evaluation for valuable assistance in preparing these remarks.
References
Augenblick, N, Jack, K, Kaur, S, Masiye, F, and Swanson, N (2026) ‘Retrieval Failures and Consumption Smoothing: A Field Experiment on Seasonal Poverty’, NBER Working Paper 35430, https://doi.org/10.3386/w35430.
Banerjee, A. V., Duflo, E., & Kremer, M. (2016). The Influence of Randomized Controlled Trials on Development Economics Research and on Development Policy. Prepared for The State of Economics, The State of the World. povertyactionlab.org
Cai, S, Park, A, and Wang, W (2025) ‘Microfinance Can Raise Incomes: Evidence from a Randomized Controlled Trial in the People’s Republic of China’, ADB Economics Working Paper Series No. 812, http://dx.doi.org/10.22617/WPS250409-2.
Cohen J and Dupas P (2010) ‘Free distribution or cost sharing? Evidence from a randomized malaria prevention experiment’, Quarterly Journal of Economics, 125(1):1 45.
Glewwe, P, Raitzer, D, Sharma, U, Chua, K, and Thomas, M (2025) ‘Learning at the Last Mile: Evidence from a Randomized Controlled Trial of Computer-Assisted Instruction in Remote Schools of the Philippines’, ADB Economics Working Paper Series No. 829, http://dx.doi.org/10.22617/WPS250488-2
Habib, M.A, Madajewicz, M, and Tompsett, A (2023) ‘When Does Community Participation in Decision-Making Improve Outcomes? Evidence from a Field Experiment in Bangladesh’, ADBI Working Paper No. 1406. Tokyo: Asian Development Bank Institute, https://doi.org/10.56506/FEQW6636.
Hassan, H, Islam, A, Kayes, I, and Wang, L.C (2024) ‘Building Education Resilience through Parenting Style and Out-of-School Learning: Field Experimental Evidence from Rural Bangladesh’, ADBI Working Paper No. 1454. Tokyo: Asian Development Bank Institute, https://doi.org/10.56506/DHLW9675.
J PAL (2026) About Us, J PAL website, accessed 26 August 2026.
Leigh A (2024a) ‘Making development count: evidence for real change’ Address to Asian Development Bank, Manila, Philippines, 23 July
Leigh A (2024b) 'Evidence based development’ Address to Asian Development Bank Institute Research Conference, Manila, Philippines, 24 July
Schönholzer, D, Lane, G, and Kelley, E (2026) ‘Safer Driving for a Price: Evidence on Behavior and Habits in Kenyan Minibuses’, NBER Working Paper 35394 (2026), https://doi.org/10.3386/w35394.
Tourek, G, Laroche, A, Bergeron, A, Naritomi, J, Weigel, J, and Mavungu Ngoma, M (2026) ‘Does Progressivity Raise Tax Capacity? Experimental Evidence from the D.R. Congo’, NBER Working Paper 35536 (2026), https://doi.org/10.3386/w35536.