ABSTRACT
Misspecification of an underlying relationship may contribute to an observed lack of statistical normality in distributions of particular ratios. One factor which can induce non-normality in ratio distributions in the incidence of outliers. Many researchers choose to exclude such observations, but even where this is done distributions may be non-normal. Manipulation can refer to a number of accounting indicators - e.g. gearing or liquidity ratios, as well as profitability measures. Moreover, it is really necessary to try to distinguish the motives. In a signalling context 'profit improvement' is essentially a short term strategy which may be employed when a business is in financial trouble. In terms of accounting manipulation, a laboratory experiment by B. S. Koch interestingly found that profit smoothing was a favoured strategy amongst managers working for companies whose shares were widely held - in essence where there is a clear need to signal future intentions.
