In the months that they were unemployed, those people who experienced both unemployment and employment from 1993 to 1995 had an individual income of approximately less than half the income they had in the months they were employed. This estimate is drawn from the first three waves of the European Panel (1994-1996). It uses a fairly broad concept of individual income, which includes earned income and replacement income (excluding severance pay) as well as supplementary benefits (excluding housing benefit and taxes). The relative deviation between earned income and unemployment income is more moderate when employment precedes unemployment than vice versa. A person who has had a period of employment followed by a period of unemployment over a three-year period has an unemployment income one-third lower than their earned income. However, a person in the inverse situation doubles his unemployment income when he takes a job. The deviation is greater for new arrivals on the labour market. Income deviations also vary a great deal with age. Unemployment income represents only one-quarter of the earned income of under-25s, as opposed to over two-thirds for the over-50s. Other things being equal, the relative deviation between unemployment income and earned income is greater when qualifications and wages are high. The average level of income deviations can be highly sensitive to the way in which they are defined and estimated (monthly or annual deviations, incorporation of severance pay, etc.). Some 10% of the people who had been both employed and unemployed over the three years studied received severance pay. Half of these people used their severance pay to offset the loss of income due to being unemployed for five or more months.