PERFORMANCE OF MUTUAL FUNDS ON SUSTAINABLE MICRO-HEALTH INSURANCE SCHEMES IN GHANA: The Role of Fast Adaptive Neural Network Classifier

Abstract

Micro-health insurance models have emerged (in different forms) as a more reliable source of seeking financial protection for a significant proportion of Ghanaians against the downside of medical cost. Most micro-health insurance contributions end up with mutual funds hence the performance of the mutual fund determines to a large extent the sustainability of the scheme. This makes the mutual fund market an indispensable factor in stimulating or stifling healthcare access in Ghana with health equity implications. Recently, machine learning performance evaluation approaches such as Artificial Neural Networks (ANN) have proven valuable in evaluating the performance of mutual funds due to superior forecasting and calculating abilities relative to native algorithms. We applied a novel Fast Adaptive Neural Network Classifier (FANNC) to publicly available historical financial performance data from the Ghana Stock Exchange. We benchmarked our results against the outcome of a backpropagation neural network model (BPN) and measured speed of processing performance information for micro-health insurance managers looking for high earning but less risky investment destination for their vulnerable funds. The FANNC tool proved superior in terms of prediction error and processing time to existing robust models such as the Backpropagation neural network. Employing effective tools such as FANNC tools can quicken the pace of investment performance assessment, rank a hierarchy of desirable investment options for micro-health insurance schemes to explore. This can guarantee fast processing time and accurate investment decision of micro-health insurance to enable it play the essential role of enhancing health risk management of the members of the general public especially the poor and vulnerable that are often the victims of catastrophic healthcare expenditure.