Dilution in Director Risk Assessments: The Effects of Diagnostic and Non-diagnostic Information

Jean L. Seow

This paper reports the results of an experiment conducted post-SOX (2002) involving 107 board directors in Singapore to investigate how directors’ identification and evaluation of diagnostic and non-diagnostic information in a financial statement fraud risk assessment setting affects their risk assessments. The results show that more (less) diluted risk assessments are made when more (fewer) non-diagnostic factors are identified and where more (less) weight is placed on non-diagnostic factors identified as relevant to the judgment decision. Directors with higher (lower) levels of domain knowledge identify more (fewer) diagnostic factors and place more (less) weight on these factors. Directors with higher levels of domain knowledge also place marginally less weight on non-diagnostic factors, resulting in less dilution in risk assessments made. The ‘gap’ in performance on the risk assessment task between directors with higher versus lower levels of domain knowledge is found to persist even with the use of a decision aid. Collectively, these findings suggest that domain knowledge is a key driver in directors’ ability to identify and evaluate risk factors, and lack of domain knowledge is not easily replaced by decision aids. The findings also show that some commonly-used measures of director expertise and competence may be inappropriate proxies for their risk assessment abilities.

JEL Codes: M41, M49

DOI : http://dx.doi.org/10.21102/jbpr.2014.07.91.02