The issues of persistence in the observed labour market status of men are investigated using the British Household Panel Survey for the period 1991-97. The paper extends previous work in many directions. In particular, problems of endogenous initial conditions, and unobserved heterogeneity, are addressed within the context of different definitions of unemployment. In addition, allowance is also made to accommodate the ‘stayer’ phenomenon in the state of employment. All these were found to be very important in the estimation of the effect of scarring.