Academic journal article Journal of Business and Entrepreneurship

MODELING BUSINESS FAILURE AMONG SMEs: AN ARTIFICIAL NEURAL NETWORKS AND LOGISTIC REGRESSION ANALYSIS

Academic journal article Journal of Business and Entrepreneurship

MODELING BUSINESS FAILURE AMONG SMEs: AN ARTIFICIAL NEURAL NETWORKS AND LOGISTIC REGRESSION ANALYSIS

Article excerpt

(ProQuest: ... denotes formulae omitted.)

INTRODUCTION

Since the 1980s, there has been a growing literature on business failure as a measure of firm performance (Campbell et al., 2012; Mellahi and Wilkinson, 2004; Thornhill and Amit, 2003; Whetten, 1980). The literature however, remains quite disperse with very few definitive conclusions as to the factors that have led to failure among firms, especially the small and medium-sized enterprises (Mellahi and Wilkinson,2004). This seems to be due mainly to: the less than rigorous and sophisticated methods used to model failure, the divergence in theoretical lenses used to analyze the problem, among other things (Mellahi and Wilkinson, 2004). This study therefore, will depart from the traditional simplified methods such as, linear regression models and qualitative case studies and use a more powerful tool, the Artificial Neural Networks model, which has never been used in the literature before, to predict the factors that impact on business failure. Neural network models are more flexible and less restrictive in their assumptions about the data and as such, will lead to more robust results. In addition to the Neural Networks model, this study will also apply the logistic regression model to the problem in order to validate the findings even more rigorously. This methodological approach to the problem will help to add new insights to the work and also, further advance the ability of theorists in the field to build a general theory on business failure.

The work presented in this paper will try to answer the following questions: a) Which factors are most important in predicting business failure among SMEs? b) Can the neural networks models give a better predictive accuracy than the traditional logit model in predicting business failure?

The first question draws on the resource-based view of the firm as the theoretical lens through which to analyze the research problem. Viewing business failure through the resource-based lens, it argues that the fewer resources that a firm possess is the greater the likelihood that it will fail. In essence, the argument is that there is a negative relationship between failure rates and the availability of resources (Anderson and Tushman, 2001; Barney, 1991; Williams, 2009). The empirical evidence however, is not always consistent. It is this inconsistency that has motivated the research reported in this study. Given the strength of the methods to be applied to the problem in this research, the work is expected to add greater clarity to the findings. Further, the results should help to enhance the external validity of the works in previous studies given that the variables used will be taken from previous studies on the subject.

To shed greater light on the research problem, the remainder of the paper is organized as follows: the next section will give a brief overview of the theoretical lens that is used to focus the study. Following this review of the theory, the next section will present the research variables and hypothesis development. The subsequent sections will present the research method, the findings, a discussion of the findings and ends with some concluding remarks.

THE THEORETICAL LENS

Since Penrose (1959) conceptualization of the firm as a bundle of resources, a steady stream of work has evolved under the rubric, the resource-based view of the firm (Amit and Schoemaker, 1993; Barney, 1991; Cooper et al, 1994, Wernefelt, 1984). The basic idea is that the firm is viewed as a heterogeneous bundle of resources (Thornhill and Amit, 2003). Indeed, Amit and Schoemaker have defined resources as the stocks of available factors that are owned and controlled by the firm. They further went on to identify the difference between resources and capabilities, which in most cases are conflated. They noted that capabilities are information based, tangible or intangible processes that are firm-specific and are developed over time through complex interactions among the firm's resources (1993, ibid pp. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.