Three facts decide a Mauritian firm's unrest exposure

Riots in Mauritius from 12 October 2026 were followed by seven nights of curfew. Early reporting measured what this cost the country: Business Mauritius put activity losses at Rs 270–470 million, a modelled 20–35% fall in direct value added on the most affected days, published without its parameters. That figure is not a damage estimate, and it is not any firm's exposure. The study asked the narrower question a board must answer: what should a firm with Mauritius exposure test to judge its own resilience to civil unrest? The answer starts in the Civil Code. Article 1983-46 relieves the insurer of liability for riot losses “sauf convention contraire” — unless the contract provides otherwise. On every Mauritian insurer page that could be read, explicit riot cover appears only on motor products. After 1999, insured firms under-recovered, and the State was held not liable in both reported cases, after litigation lasting twelve to fifteen years. Three mechanisms then set the size of the loss. Restriction reached furthest: about three violent days but seven curfew nights, with buses curtailed and the Metro suspended, and such losses are usually excluded from cover. Concentration set severity: in South Africa in 2021, the retailer that lost two distribution centres carried about R650 million of net loss after insurance. And claims pay slowly: about 14% of expected indemnities had been paid six months after the 2024 New Caledonia riots. Liquidity is therefore part of protection, not an alternative to it. What shortened losses elsewhere lies within a firm's control. Firms that restarted elsewhere traded within weeks; reducing stock and cash on site is cheaper than protecting it. Having a continuity plan did not predict recovery; capacities did. Early Mauritian signals, with no wave of cancellations and no rupee depreciation, are consistent with a short-lived national effect, conditional on the unrest not recurring. Much remains open. No resilience score reviewed has been validated against real outcomes; Mauritian recovery evidence does not yet exist; no data on SME insurance take-up exist. The study's diagnostic is therefore an evidence-calibrated screening instrument, not a validated predictor, and its thresholds are published FalconBridge judgements. A board could consider asking management, within thirty days, for three things the firm already holds: the endorsement wording that expressly covers riot; the share of inventory in its largest node, and whether it could restock elsewhere within fourteen days; and the days of fixed costs reachable within a week, net of deductibles — at least 30, preferably 90. A failure on any one sets the judgement to Exposed, whatever else is strong. Re-scoring at every renewal would keep the answer current.