Adverse selection and moral hazard in insurance: Can dynamic data help to distinguish?