Inter-class co-movement of retail electricity tariffs and the residential adjustment gradient: evidence from a new Latin American panel

How do the published retail electricity tariffs of different consumer classes move together, and do the schedules of captive residential users adjust more slowly than those of large consumers? This paper assembles a new harmonised panel of tariff schedules for fourteen Latin American countries, covering 91 per cent of regional consumption, annually (2011–2024) and quarterly (2013–2025). Estimated by a mean-group error-correction model, the three schedules move together in the long run, each correcting a quarter to a third of any disequilibrium per quarter. Error correction is bidirectional, so neither class is weakly exogenous: the estimates identify long-run co-movement between published schedules rather than structural pass-through of wholesale, network or policy costs. The long-run co-movement elasticity declines along the industrial–commercial–residential gradient, from 0.88 between the two large-consumer segments to 0.73 and 0.70 for the residential schedule; no relation reaches one-for-one movement, and the residential shortfall is significant when economies are paired (0.21, p = 0.021, against the industrial schedule). Within the residential class the rate of error correction rises with the consumption block, the least-subsidised block adjusting twice as fast as the lifeline block (−0.21; p = 0.024), a gradient consistent with cross-subsidy design though not evidence of it. Splitting the panel by revision frequency shows the calendar governs the pace of convergence, not its long-run extent. The inter-class differential arises not from residential schedules being revised less often but from each revision moving them less far. The results inform cross-subsidy reform, lifeline design and affordability policy.