Comparison of Time-Series Forecasting Models in Predicting Stock Prices of Insurance Company

Abstract

The insurance industry has always been closely related to people's lives. In recent years, people have paid more attention to personal and property safety and given more consideration when choosing insurance products. At the same time, more insurance products have been developed to meet people's needs. The impact of such a situation is unclear, but it has certainly become more difficult to anticipate the stock price of insurance companies. This paper aims to predict the stock prices of four leading insurance companies using three different time-series analysis models, namely moving average, exponential smoothing, and autoregressive integrated moving average. In this paper, ten-year monthly adjusted closing prices of four insurance companies are obtained from Yahoo Finance. Then, each method is used to create predictive models. The predicted values from different models will be compared with the actual values to yield the forecasted error. The least forecast error would suggest the superior model in predicting future stock prices. Empirical results show that out of the three time-series forecasting model, the autoregressive integrated moving averages provide better approximations in estimating the stock price for the following month. Additionally, the autoregressive integrated moving average is superior in way that it can forecast stock values for a longer period of time. This paper will offer some insights for researchers in further study of forecasting stock prices of insurance companies.