In this paper, we show that standard approaches for valuing non-market amenities like clean air through defensive expenditures are methodologically fragile. Estimates are highly sensitive to assumptions about household use of defensive technologies, and survey data on usage is unreliable. Moreover, because researchers convert a one-time purchase or willingness-to-pay elicitation into a flow value by dividing by assumed hours of use, lower assumed usage can mechanically imply higher valuations. To address this fragility, we develop a revealed- preference approach that estimates the flow value of clean air using high-frequency data on residential air purifier usage in Dhaka, Bangladesh. Leveraging randomized variation in the marginal cost of operation, we recover the households’ trade-off between money and clean air without relying on purchase decisions or stated usage intentions. We find that the marginal willingness to pay (MWTP) for clean air is lower and much more tightly bounded than previous estimates. This low valuation is not driven by traditional market frictions like liquidity constraints or misperceptions about pollution severity and technology effectiveness, suggesting limits to environmental policies that rely on sustained complementary household action.