VAR Modelling of Insurance Premiums: Case of Life Insurance in India Kumar Jitendra*, Afifa Umme**, Sharma Nikita*** *Department of Statistics, Central University of Rajasthan, Kishangarh, Rajasthan, India **Department of Statistics, Central University of Rajasthan, Kishangarh, Rajasthan, India ***Department of Statistics, Central University of Rajasthan, Kishangarh, Rajasthan, India Online published on 10 May, 2016. Abstract Insurance is a tool of future risk management by transferring the financial liabilities to the insurance companies if any uncertain loss occurs under the covered risks. For the coverage of equivalent financial loss upto maximum sum assured for which insured is liable to pay the premium. Time series is chronologically recorded data and multivariate time series models are used to model multiple variables, where the cross sectional dependency of variables are important. As Insurance is long term planning/investment, therefore to know the future premium is equally important for the company as well as administration. Present study is carried out to model all four types of life insurance premiums together using vector Auto Regressive Time Series model for the seven insurance companies, which are doing business in India to know the future premium. We have taken total premiums of different companies from IRDA monthly news updates for the period from April 2006 to June 2013. Using VAR approach best models are selected in respect to optimum value of HQ and then tested the stationarity of the Model. Best suitable model for all selected companies is VAR(1) except Max New York for which best model is VAR(3). After the selection of best model future premium for the next 25 months is also forecasted. Top Keywords Life Insurance, VAR Modelling, Eigen Values, Stationarity. Top |