The recent development of indexed catastrophe (CAT) securities has been a concern in insurance literature. We refer to the two existing prominent explanations as the systematic risk approach and the moral hazard approach. In the systematic risk approach, the systematic risk portion is hedged by index-triggered securities, and the remaining unsystematic risk is hedged through indemnity-triggered vehicles including traditional insurance. In the moral hazard approach, indexing protects firms from the loss without incurring moral hazard problems. We argue that indexing seems to be supplementary, rather than domainant, in those approaches. We suggest two alternative rationales for indexing CAT securities. First, We argue that if firms are concerned with downside risk rather than variation, then indexing is optimal. Second, we argue that the observability of loss is another key factor for indexing, even when firms are concerned with variability. We identify the important sources of high observation costs as (i) the inherent difficulty in identifying CAT losses; (ii) the non-separability of cash flows between CAT event and other operations of the firm, and (iii) the impossibility of talking over the firm by the bondholders under a CAT event. In both cases, we show that indexing is a dominant tool, in contrast to existing approaches.