Projects that use the market systems development (MSD) approach often argue that their most successful interventions create lasting impact for target communities long after donor funding ends. In practice, however, these claims are rarely tested. Ex-post evaluations are uncommon, leaving a significant gap in understanding what truly remains once projects exit.
To address this gap, the International Labour Organization’s Systems Change Initiative revisited its engagement in Afghanistan’s carpet, livestock, and dairy sectors between 2018 and 2021. The review examined two related ILO projects: the Road to Jobs programme, funded by Sweden and implemented from 2015 to 2020, and its successor, Road to Decent Jobs for All Afghans, funded by the United States and running from 2020 to 2022. While the assessment of the carpet sector showed clear evidence of sustained system-level outcomes, a rapid review of the livestock and dairy interventions found that little remained intact for meaningful evaluation. The findings raise important questions about why earlier gains failed to endure and what alternative decisions might have produced more sustainable results.
In the livestock sector, the intervention focused on strengthening access to basic animal health services in rural areas by providing additional training to self-employed, community-based para-veterinarians. This work was carried out in partnership with the Afghan Veterinary Association and addressed a genuine gap, as trusted para-vet services were scarce across northern Afghanistan. Despite initial successes, the benefits for livestock-keeping households proved short-lived.
Political instability played a decisive role. The regime change in 2021 led to widespread emigration, including nearly all para-vets trained under the programme. Those who remained were cut off from professional networks and struggled to access supplies and supporting services, severely limiting their ability to operate. Beyond this, the intervention lacked a strong systemic design and long-term sustainability plan. Training was delivered as a one-off, subsidised activity, departing from core MSD principles. A more sustainable approach might have involved supporting the Afghan Veterinary Association or other actors to develop cost-recovery or commercial training models, potentially with financial contributions from para-vets or farming communities.
The project also missed opportunities to diversify training pathways, such as entry-level courses for new para-vets, advanced modules for experienced practitioners, or reskilling focused on emerging animal health risks. Without a viable business model or pipeline for continued training, the association was not positioned to sustain these services after project closure. At the same time, incentives for behavioural change were weak, as other donor-funded initiatives continued to provide direct support to para-vets, reducing demand for market-based solutions. The intervention also overlooked complementary functions such as professional accreditation, quality assurance, and mechanisms for knowledge sharing, which could have helped higher-quality para-vets differentiate themselves and promote better practices more widely.
In the dairy sector, the intervention supported two processing companies to improve milk aggregation and transportation systems. New collection regimes were introduced in peri-urban areas, where company representatives took responsibility for aggregating, chilling, and recording milk deliveries at converted shop premises known as collection hubs. From there, milk was transported to processing facilities, reducing spoilage and allowing farmers to receive higher prices by selling into more formal value chains.
Despite these gains, little of the physical, institutional, or social infrastructure remains today. A series of interconnected shocks contributed to the decline. Prolonged drought in target districts severely undermined agricultural viability by reducing water availability, grazing land, and animal fodder. As natural pasture degraded, small-scale dairy farming became increasingly unsustainable as a supplementary livelihood. While it is unclear whether investments in climate adaptation would have fully offset these impacts, the absence of such measures is now seen as a critical oversight, particularly as extreme climate events become more frequent.
Demographic changes further weakened the sector. Many cattle-owning households left dairy farming altogether and relocated to urban areas in search of alternative livelihoods. Combined with a broader decline in agricultural labour, these exits reduced milk supply to the point where collection hubs were no longer viable. The loss of agri-extension and animal health workers through emigration compounded these challenges, leaving farmers without essential support to maintain productivity and accelerating the abandonment of small-scale dairy activities.
Market dynamics also shifted unfavourably. An influx of low-cost dairy products from neighbouring countries, especially Iran, undermined the competitiveness of local processors. Several dairy companies are now operating far below capacity and struggling to survive, further limiting opportunities for smallholder farmers to sell milk through formal channels.
Amid this broader decline, there are limited signs that elements of the intervention have endured. One former partner company continues to use milk collection hubs in nearby districts not originally targeted by the project, where smallholder dairy production remains marginally viable. These isolated examples suggest that while most impacts have faded, some innovations can persist where market and environmental conditions remain favourable.
Overall, the findings underline the importance of systemic design, sustainability planning, and resilience to political, climatic, and market shocks. While non-violent development interventions have a long history in Afghanistan, this review highlights how fragile gains can be without deeper market integration and adaptive capacity, offering valuable lessons for future MSD programming in fragile and conflict-affected contexts.
