The impact of health literacy on healthcare spending: theory and evidence from China.

We develop a theoretical model where health literacy endogenously shapes healthcare spending decisions. The model predicts an inverted U-shaped relationship between literacy and spending, driven by two channels: the 'attention' effect, where increased health literacy enhances health awareness, and the 'productivity' effect, where health literacy improves health production efficiency. We empirically test these predictions based on an adult subsample of 2020 China Family Panel Studies, where individual-level health literacy is measured using a self-design questionnaire (the Chinese Health Literacy Survey) following internationally comparable and established instruments. To address the potential endogeneity, we apply Natural Language Processing on 2.16 million news articles to construct an instrumental variable for health literacy coverage.​ This instrument exploits province–month variation in news coverage of concepts related to understanding and navigating health information. Our quantitative analysis reveals that a 10% increase in health literacy correlates with a 5% reduction in healthcare expenditure. Through mechanism analysis, we provide evidence for the attention and the productivity effects, respectively. For robustness, we conduct alternative IV construction and use Double Machine Learning to control for observable and high-dimensional confounding. Our findings imply that improving health literacy can be an effective policy measure to curb the rising healthcare spending and promote efficient healthcare utilization around the globe.