Financing Health under Climate Shocks: Risk Pricing, Data Poverty and Financial Protection in Pakistan

Abstract

Climate shocks in low-income settings are usually accounted for as damage to physical assets. This paper argues that in health systems financed at the point of use, the larger and more persistent welfare loss falls on household budgets through out-of-pocket medical spending, and that this loss is systematically invisible to the post-disaster accounting on which reconstruction finance is based. Pakistan’s 2022 floods are the motivating case: the Post-Disaster Needs Assessment valued damage to health facilities at USD 109 million, roughly seven tenths of one per cent of assessed physical damage, while destroying an eighth of the country’s health infrastructure and coinciding with a fivefold rise in malaria incidence in a system where households already financed most care directly. We develop a framework in which two frictions govern whether such losses are pooled. The first is a pricing friction: where the price of climate risk is administratively suppressed or never established, cover is not written. The second is a data friction: where screening is delegated to statistical models, applicants with sparse records are rationed out, a mechanism documented for venture capital by Bonelli (2026) and for clinical risk prediction by Obermeyer et al. (2019). Our central proposition is that in a thin-data health economy these frictions are sequential rather than parallel. A price never charged generates no claims record, and the absent record is precisely what a later algorithm would require, so the same districts are excluded twice. We derive three hypotheses and specify an identification strategy that is executable with existing public microdata: a difference-in-differences design on district flood exposure using the Household Integrated Economic Survey rounds of 2018–19 and 2024–25, a triple-difference exploiting Punjab’s 2025 termination of Sehat Sahulat cover in public hospitals, and a prospective audit of algorithmic underwriting. We set out a four-part design response that couples actuarially priced parametric cover to Pakistan’s existing social health insurance and instant-payment infrastructure, and assess its fiscal incidence. The paper is a research design and policy analysis; it reports no estimates of its own, and labels its propositions accordingly.

Keywords: health financing; catastrophic health expenditure; financial protection; climate shocks; social health insurance; parametric insurance; algorithmic underwriting

JEL classification: I13, I18, H51, H84, Q54, O16

https://doi.org/10.5281/zenodo.22642436