India, Dec. 23 -- India's rural employment programme, Mahatma Gandhi National Rural Employment Guarantee Scheme (or MGNREGS), has long attracted two criticisms. First, that it is wasteful, and plagued by leakages and corruption; second, that it is distortionary - raising wages without boosting productivity, and hurting employment. Both critiques point to the same policy question: Should the answer be to fix implementation, or to redesign the programme itself? Over the past decade, credible research has delivered a clear lesson. When implementation improves - by making work more readily available, reducing leakages, and ensuring timely wage payments - the gains are substantial. In a large-scale randomised evaluation of improving MGNREGS implementation, covering around 19 million people in Andhra Pradesh and Telangana, we found sharp income increases and meaningful poverty reduction. Crucially, these gains mostly came not from MGNREGS wages themselves, but from higher private labour market earnings. These findings overturned our own prior beliefs. Like many economists, we were sceptical that MGNREGS could raise wages without reducing employment. The data, however, showed the opposite. Better implementation increased both market wages and private employment by strengthening workers' outside options in rural labour markets with employer market power. Higher incomes then boosted local demand, leading to a significant expansion of non-farm enterprises and employment, as confirmed by Economic Census data. Thus, better MGNREGS implementation improved both equity and efficiency, which is rare for a welfare programme. The gains were also broad-based. Ordering households from those with the smallest to largest landholdings, the bottom 93% were better off, and only the top 7% were worse off due to higher wage costs. But many large landowners also own non-farm businesses that benefit from higher local demand. So, when we assess the distributional impacts by households' overall socioeconomic status, and not landholding alone, we find that nearly everyone benefits from a better implemented MGNREGS. The deeper point is that a well-functioning MGNREGS is not just a pro-poor policy; it is also a broad-based rural growth policy. Work-based guarantees also have two other strengths. First, the work requirement helps target the most vulnerable effectively without relying on fine-grained administrative identification. Consistent with this, data shows that landless labourers, women, and Scheduled Castes and Tribes are disproportionately represented among MGNREGS workers. Second, it is a powerful automatic stabiliser. Demand for employment rises during droughts, downturns, and major shocks (as seen in surging MGNREGS demand during Covid-19), and falls in good times. Thus, the key to obtaining the benefits of MGNREGS is effective implementation, which makes it a credible outside option for workers. Yet, implementation quality has varied widely across states over the last two decades. Some - such as Tamil Nadu, AP/Telangana, Kerala, and Rajasthan - have generated high person-days of work and paid wages reliably. Others, including some of India's poorest states, have persistently under-provided work despite high need. The constraint was not a flawed design, but uneven implementation capacity, and political and administrative will. Against this backdrop, Parliament has replaced the Mahatma Gandhi National Rural Employment Guarantee Act with a new law: the Viksit Bharat - Guarantee for Rozgar and Ajeevika Mission (Gramin). Some elements of the reform deserve praise. The statutory ceiling is raised from 100 to 125 days. Greater emphasis is placed on timely wage payments. And there is a welcome effort to align public works with durable assets related to water security, rural infrastructure, livelihoods, and climate resilience. If the reform stopped here, the debate would largely be about implementation - which is where it should be. The concern is that the new law introduces changes that risk weakening the programme's core function as a credible employment guarantee. It replaces a rights-based guarantee that can be demanded by citizens with a centrally designed scheme, and moves from full central financing of wages to a 60:40 Centre-state cost-sharing model (90:10 for some states), with state-wise normative allocations. This risks constraining the programme precisely where it is needed most. Poorer states with weaker fiscal capacity and administrative systems may now use it less, not because need has fallen, but because their costs have risen. Finally, mandating a 60-day pause adds little efficiency - because MGNREGS work already drops naturally in peak seasons through worker self-selection. Yet, it introduces uncertainty about work availability that can weaken the programme's broader labour-market and demand effects. One concern that the reform may be seeking to address is that some richer states draw far more MGNREGS funds than their share of the rural poor. However, the solution should be to improve implementation quality in poorer states, not penalise those that have implemented it well. If the aim is to reduce the Centre's fiscal burden, a better way to operationalise a 60:40 split would be for the Centre to fully finance a basic guarantee of 75 days, while allowing states to top up for an extra 50 days if they choose. This would reduce central outlays to richer states without weakening incentives for poorer states to provide at least 75 days of employment. India does need to modernise its rural safety net. But evidence suggests that the central challenge with MGNREGS was never its basic design; it was inconsistent implementation, especially in states where political will and administrative capacity were weakest. The priority should therefore be to strengthen implementation precisely in those states - through timely wage payments, effective grievance redressal, and administrative support that makes access predictable and credible. For a country aspiring to become both viksit (developed) and atmanirbhar (self-reliant), boosting rural wages and sustaining demand growth is essential for raising productivity, expanding non-farm employment, and deepening domestic markets. A credible, well-implemented employment guarantee programme is one of the few interventions with high-quality evidence that it can deliver these outcomes at scale - improving equity while also supporting economic growth. Reforms should build on this strength, and not weaken the essential features that make it work....