Content area

Full text

Headnote

Enforcing compliance with the National Minimum Wage (NMW) in the British economy requires the identification of those establishments that are likely to pay a substantial proportion of their employees less than the NMW. In this paper, a fuzzy decision support system is constructed to aid in the elucidation of such an establishment and their proportion of employees paid less than the NMW. Moreover, through an inductive fuzzy decision tree approach, a set of fuzzy rules enables the prediction of this proportion value. These fuzzy rules allow a more human linguistic approach to the problem, which can be interpreted by nontechnical individuals whose role it is to target certain establishments for further inspection. Through a semi-automated procedure for the construction of the fuzzy set theory related membership functions, the whole process mitigates the need for the influence of subjective expert opinion. With rule description and a comparison to multivariate discriminant analysis models, these fuzzy rules are shown to be an appropriate tool for the elicitation of establishments not in compliance with the NMW.

Keywords: decision support system, fuzzy decision trees, inductive rules, low pay, membership functions

(ProQuest: ... denotes formulae omitted.)

1. INTRODUCTION

Britain introduced its National Minimum Wage (NMW) in April 1999. The aim is to prohibit the payment of low hourly wage-rates in the British economy. A major problem faced by those implementing the minimum wage is monitoring its operation. This involves the use of targeted inspections by the Inland Revenue (IR), the body charged with the responsibility for monitoring the NMW (Low Pay Commission, 2000). These are based on (ibid., p. 11);

".. a model aimed at identifying those sectors and geographical areas where non-compliance is likely to be most prevalent, as well as those types of workers most at risk of non-payment."

One fifth of IR inspections result from complaints by workers or by third parties. The rest are based on pro-action by the IR, either on the basis of information emanating from the records for the Working Families' Tax Credit scheme or other sources (DTI/Inland Revenue, 2001), see Table 1. This indicates that a substantial proportion of inspections are based on the IR identifying potential non-compliers from administrative records collected for other purposes. The system has proved reasonably effective...

Show less