Analysis of MNC36 Index Movements Using the Autoregressive Integrated Moving Average (ARIMA) Method During and After the COVID-19 Pandemic

Abstract

This study analyzes the movement of the MNC36 index during the COVID-19 pandemic (2020–2022) and after the pandemic (2023–2025) using the Autoregressive Integrated Moving Average (ARIMA) method. Daily closing-price observations were obtained from Investing.com for the period 2 January 2020 to 30 December 2025, comprising 1,449 observations, and were processed using EViews 12. The analysis followed the Box–Jenkins procedure, including descriptive analysis, Augmented Dickey–Fuller (ADF) stationarity testing, model identification using ACF and PACF, parameter estimation, information criteria comparison, and residual diagnostics. The results show that MNC36 was more volatile during the pandemic, with a mean of 314 and standard deviation of 31, compared with a mean of 343 and standard deviation of 24 after the pandemic. The series in both periods was non-stationary at level and became stationary after first differencing. Based on AIC and SIC, ARIMA (3,1,3) was selected for the pandemic period, while ARIMA (2,1,13) was selected for the post-pandemic period. The latter produced lower AIC and SIC values and a Durbin–Watson statistic closer to 2, indicating relatively better fit and milder residual autocorrelation. However, residuals from both selected models did not fully satisfy the white-noise criterion. The findings indicate that ARIMA remains useful for modeling MNC36, but its performance depends on market conditions and the model should be interpreted alongside residual diagnostics.