Nowadays, non-life insurance is being very popular in Bangladesh. Strategy of a company depends on occurrence of claims. Our main intention of the study is to find out the factors responsible for occurrence of claims. As regression models are used to find the dependence of response variable on different covariates, we used the Cox regression model which is one of the most popular models for analyzing time-to-event data. The number of claims and claiming status (claimed or not claimed) are also found in non-life-insurance claims data. To find out the influence of the factors on the number of claims claimed by policyholders and for finding the difference of the influencing factors of claiming status we used the Poisson regression model and the Logistic regression model respectively. Different covariates were found to be as location of policyholder, cause of loss, type of insured object, use of vehicle, risk covered and type of carrier used for shipping goods, etc. We used the claims data of The Federal Insurance Company Ltd. Bangladesh for the years 2006-2008. A total of 721 individual policyholders’ information was recorded in a SPSS data sheet from the company’s lodge register books.

Using Cox regression model, for motor claims, location of policyholder and use of vehicle which was insured were found to have significant effect on hazard of claim. The hazard was highest for Chittagong and for privately used vehicles. For property insurance claims, location of policyholder and type of insured object were found to be significant. In this case the hazard was highest for Barisal and furniture type objects. For cargo insurance claims, location of policyholder, risk covered and type of carrier used for shipping cargo were found to have significant consequence on hazard of claim. The hazard was highest for the other locations than Dhaka and Chittagong. Also the hazard was highest for medium and high risk covered policies and for steamer. Using Poisson regression model, for property insurance claims, cause of loss and for cargo insurance, risk covered were found to be statistically significant for the number of claims occurred. Using Logistic regression model, for motor insurance, cause of loss and time of claim after starting policy were found to have significant effect on claiming status.