We evaluate an intervention aimed at reducing household air pollution in Burkina Faso by encouraging households to switch from wood or charcoal to liquefied petroleum gas (LPG) as their main cooking fuel. We randomize 805 urban households to a subsidy treatment, a credit treatment and a control group. Treated households receive an offer to purchase a gas stove from a local retailer either at a discounted price or at the market price with the option to pay in three installments. We estimate the effects of these capital costs subsidies and consumption loans on the adoption and intensity of use of LPG over a six-month period following treatment. Subsidies and access to credit increase the take-up of gas stoves by 54 and 28 percentage points respectively, suggesting both credit constraints and high price elasticity of demand. We find evidence of either selection or sunk-cost effects: higher usage rates in the credit group nearly offset the difference in take-up with the subsidy group. 6 months after the start of the interventions, 30% of the LPG stoves are in use on a given day but LPG use does not result in a decrease in the consumption of wood, the main baseline fuel. We also find no effect on cooks’ exposure to fine particulate matter. We show that total useful energy consumption increases among treated households, suggesting that LPG acts as a complement rather than as a substitute to wood. Heterogeneity analyses show that the interventions lead to small reductions in both wood consumption and exposure among households which purchased their wood at baseline but have no effect among exclusive wood collectors. We conclude that financial incentives for LPG appliances should target populations without access to fuel collection areas.