Abstract:
In addition to its primary role of providing financial protection for other industries the insurance industry also serves as a medium for fund mobilization. In spite of the harsh economic environment in Nigeria, the insuranceindustry has been crucial to the consummation of business plans and wealth creation. However, the continued downturn experienced by many countries, in the last decade, seems to have impacted negatively on the financialhealth of the industry, thereby rendering many insurance companies inherently distressed. Although there is a regulator to monitor the insurance companies in order to prevent insolvency and protect the right of consumersthis oversight function has been made difficult because the regulators appeared to lack the necessary tools that would adequately equip them to perform their oversight functions. One such critical tool is a decision makingmodel that provides early warning signal of distressed firms. This paper constructs an insolvency prediction model based on artificial neural network approach which could be used to evaluate the financial capability ofinsurance companies
Machine summary:
This paper constructs an insolvency prediction model based on artificial neural network approach which could be used to evaluate the financial capability ofinsurance companies .
Keywords: Insurance, Financial protection, Artificial neural network, Insolvency, Ratio analysis, Training, Prediction model, Early warning signal INTRODUCTION The role of insurance in providing financial protection in the economy is well established.
In the property/liability insurance business a number of empirical studies have used statistical models based on insurers’ financial data to predict insolvencies (Trieschmann and Pinches, 1973, 1977; Harmmelink, 1974; Eck, 1982; Hershbarger and Miller, 1986; Harrigton and Nelson, 1986; BarNiv and Smith, 1987; Ambrose and Seward, 1988; BarnNiv and Raveh, 1989; BarNiv, 1989; BarNiv and McDonald, 1992).
Other studies of financial insolvency in property – liability insurance, for example, include those of BarNiv and McDonald (1992), Ambrose and Seward (1988), and Cummins, Grace and Philips (1999).
Our aim in this study, therefore, is to attempt to construct a model that can mitigate this difficulty and be used to predict insurance company insolvency in an artificial neural network framework.
)) Figure 1: A simple artificial neural network architecture Data Collection and Key Ratio Selection The data for the study were obtained from the financial statements of registered insurance companies, the Statistical Reports of the National Insurance Commission (NAICOM), the Nigerian Stock Exchange Fact-Book and the Nigeria Insurance Association (NIA).
The procedural considerations for the neural networks techniques for the simulation of insolvency in insurance industry were, therefore, conducted under the following subheadings: 9 tuning the network 9 data preprocessing 9 training of the network The steps are considered in their order of d − µ application in the ANN algorithm.