This project examines insurance ratemaking with an endogenous deductible. While it has been assumed that the deductible is exogenous to the claim frequency and severity in the classical nonlife insurance ratemaking practices, its endogeneity has been argued in terms of policyholder’s behaviour and asymmetric information to insurers. Moreoever, it is also possible that the existence of deductible can make the insurance claims underreported, which can leads to censored or truncated data.In this context, we consider a joint modeling framework that examines possible dependence between the insurance claims and the endogenous deductible via pair copulas, upon the presence of incomplete insurance claim observations due to underreporting. The proposed framework provides a solid procedure to reflect endogeneity of the deductible depending on the presence of underreporting and correctly conduct risk classifications with incomplete claim observations as validated via a comprehensive simulation study. Lastly, with an empirical analysis using the LGPIF data, it is observed that the endogeneity of the deductible is not ignorable and the usage of the proposed framework allows for a better risk classification model in terms of enhanced goodness-of-fit and prediction performance.