We study the suitability of average peak ground acceleration (𝑃𝑃𝑃𝑃𝑃𝑃) as a ground-motion proxy for parametric CAT bond design. We tie the selection of 𝑃𝑃𝑃𝑃𝑃𝑃 (as a triggering parameter for CAT bonds) to computational convenience (fast retrieval from the recorded ground motion) and loss correlation (optimum monetary return on the investor side). Our case studies advocate that 𝑃𝑃𝑃𝑃𝑃𝑃, as a candidate ground-motion proxy, can be used confidently for parametric CAT bonds, in particular, applications associated with dense coverage of seismic networks. It is still a compelling ground-motion proxy even if the seismic network coverage is sparse provided that the accelerometric stations are deployed in the vicinity of assets that financially represent the most significant portion of the insured building stock. We establish an event-based risk model of the Istanbul city (via Monte Carlo simulations) to depict the rationale behind our proposition and compare its performance with other competing (more sophisticated) proxies in terms of accelerometric network density, spatial distribution as well as the different risk levels used in risk management.
