Every social programme produces two kinds of results. The first kind is easy to count: houses built, families relocated, budgets disbursed. The second kind only appears when someone goes and asks: did life actually improve? After two decades of evaluating programmes in Mauritius, we keep finding the same gap between the two.
Our evaluation of the La Valette social reintegration experiment illustrates it well. On paper, the project was a success — two hundred concrete houses replaced sheet-metal shelters, and when we surveyed the 154 resident households, two-thirds said they were satisfied. Having a roof of one’s own, in the residents’ words, compensated for almost everything else.
Almost everything. Because the same survey revealed what the delivery figures could not. Transport was cited as a problem by more than six in ten households — the village was simply too far from everything. Nearly half the families’ children had been forced to change schools. A third of respondents had completely lost touch with their former friends. A third of household heads were unemployed or in precarious work, in part because the new village sat far from the employment poles they had left behind.
None of this means the programme failed. It means relocation is not just a housing operation — it is a rupture, professional, educational and social, and programmes that treat it as a housing operation will keep being surprised by their own results.
The evaluation also surfaced something subtler: the role of rules. Residents had signed a “social contract” governing life in the village. The clauses about respecting neighbours and the environment were widely internalised and consensual. But the prohibitions — on running a small business from home, on hosting relatives — were experienced by many as incomprehensible, even punitive, precisely because informal commerce and hosting kin were how these families had survived for decades. A rule that reads as good governance in a project document can read as a restriction of citizenship on the ground.
Three lessons travel well beyond this one project. First, evaluate against lived outcomes, not deliverables: employment, schooling continuity, social networks and perceived dignity tell you whether a programme worked; completion rates tell you only that it was executed. Second, timing matters: an evaluation conducted a few years after delivery captures adaptations and ruptures that a handover-date assessment cannot see. Third, listen for the rules: how beneficiaries receive a programme’s conditions is one of the most reliable predictors of its long-term social sustainability.
Public institutions, NGOs and funders in Mauritius are commissioning more evaluations than ever, and that is good news. The better news would be seeing their findings shape the next programme’s design — because the families of La Valette told us, very precisely, what to do differently. Someone just has to keep asking.