Test it, prove it, scale it, Nobel winning economist bats for innovation

India stands at a historic opportunity to transform its development outcomes, Nobel laureate and economist Professor Michael Kremer told a packed auditorium in Chandigarh on Friday. The transformation, he said, was contingent to policymakers replacing intuition with rigorous evidence, embracing artificial intelligence (AI) and treating public programmes as experiments that must be tested, refined and proven before being scaled.

Delivering a public lecture on “Economics and Policy Innovation in India” at Mahatma Gandhi State Institute of Public Administration (MGSIPA), Kremer — awarded the 2019 Nobel Prize in Economics for his experimental approach to fighting global poverty — drew on a wide body of research spanning agriculture, road safety, drinking water and education to make a single, compelling case: when governments combine data, technology and institutional accountability, the results are measurable, significant and replicable.

The event was jointly organised by the Chandigarh Citizens Foundation (CCF), the Centre for Research in Rural and Industrial Development (CRRID) and MGSIPA.

Former Punjab Chief Secretary Vini Mahajan, who is the vice-president of CCF, moderated the session that drew active participation from policymakers, academics, administrators and students.

Among the most striking findings Prof Kremer shared was the economic impact of AI-driven weather forecasting, which was 80,000 times faster than conventional systems. He said AI-based forecasts accurately predicted an unusual pause in the monsoon’s northward advance last year — a detail that traditional models missed. Farmers who received timely advisories adjusted their nursery timings, transplant decisions and seed choices. The result: an average economic gain of approximately Rs 2,000 per hectare per year.

With over 1.5 lakh road deaths annually, India has one of the world’s most lethal highway networks. Kremer highlighted how the HAMS platform — now deployed across 56 Regional Transport Offices — uses AI-based visual analytics to standardise and objectify the driver licensing process, reducing the scope for bias and corruption. Evidence from a partnership with Ola showed that drivers certified through these AI-supported systems reported 20 to 30 per cent fewer instances of unsafe driving behaviour, with the sharpest improvements among younger drivers.

He also presented findings from 19 randomised studies showing that water treatment interventions can reduce child mortality by 21 per cent. Applied to India’s rural population, it has the potential to prevent over one lakh under-five deaths every year. Field trials in Odisha’s Rayagada district using in-line chlorination systems showed sharp reductions in bacterial contamination. However, Kremer added that technology needs to be accompanied with maintenance and community involvement.

In education, Kremer turned to Pratham’s initiative of sending daily math activities to parents via SMS and WhatsApp. Tablet-based teacher support programmes, meanwhile, delivered gains equivalent to one additional year of schooling in some studies. The message, he stressed, was not that technology replaces teachers — it was that technology, when designed thoughtfully and evaluated honestly, can make teachers more effective.

He closed by urging governments to institutionalise what he called “the innovation cycle” — identify a problem, design candidate solutions, test them rigorously, refine based on evidence and scale only what demonstrably works.

The auditorium was filled to capacity, a rare sight for a policy lecture, reflecting the depth of interest among Chandigarh’s academic and administrative community in evidence-based approaches to governance.