Time series analysis of net premiums received from non-life insurance in Uganda (2005-2017): application of auto regressive integrated moving average
Abstract
The purpose of this research was to look into the time series analysis of yearly net premiums received from non-life insurance in Uganda over the last 13 years (2005-2017) and to build an autoregressive model of a suitable order for the process. Secondary data of net premiums received from non-life insurance between 2005 and 2017, collected by Uganda data portal was analyzed. In this study, the net premiums received from non-life insurance were plotted on a line plot and before an appropriate ARIMA model was fitted. Augmented dickey fuller test was used to check for stationary of the series. The study also statistically tested and validated using the train test split and the residuals appeared to be normally distributed with mean zero and constant variance. The ARIMA (2, 2, 1) model was chosen as the best candidate model for making predictions for up to 6 years for net premiums received from non-life insurance in Uganda using 13 years of time series data. The model was then used to forecast yearly net premiums received from non-life insurance for the following period where the series continues to show an upward trend, indicating an increase in net premiums received from non-life insurance. The study recommends the Insurance Regulatory Authority (IRA) of Uganda to support its members by providing the general public with knowledge about non-life insurance's fundamentals and advantages. This will contribute to increasing non-life insurance consumption.