FORECASTING COOPERATIVE MEMBERSHIP GROWTH: A COMPARATIVE STUDY OF THREE TREND METHODS

Abstract

Forecasting membership growth is important for savings and loan cooperatives to establish membership targets and plan operational requirements systematically. This study aims to compare the forecasting performance of three trend methods: the least squares method, the semi-average method, and the time-series least squares method, in estimating the number of members of Mekar Sai Savings and Loan Cooperative (KSP Mekar Sai) in Bandar Lampung, Indonesia. This study employs a quantitative approach using time-series analysis based on annual membership data for the 2011–2024 period obtained from the cooperative’s official records. Each method is applied to the historical data to construct a trend equation and generate an estimate of membership in 2025. Forecasting performance is evaluated using the mean squared error (MSE) and mean absolute percentage error (MAPE), where lower values indicate smaller forecasting errors. The results show that the least squares method estimates 24,123 members, the semi-average method estimates 24,344 members, and the time-series least squares method estimates 23,956 members for 2025. Based on the reported evaluation results, the time-series least squares method produces the lowest MAPE of 2.393%, whereas the least squares method yields the lowest MSE of 124,263. These findings indicate that the assessment of forecasting performance depends on the accuracy measure employed. In practical terms, the forecasts can provide a reference for KSP Mekar Sai’s membership planning and operational resource allocation. However, the findings should be interpreted in light of the study’s limitations, particularly its reliance on historical membership data from a single cooperative without incorporating external factors that may influence membership growth.