In order to provide a framework for reviewing the voluminous literature on unemployment and suicidal behaviour, the author distinguishes between two categories of deliberately self-harmful act: those with fatal outcome (suicide) and those with non-fatal outcome (parasuicide); and differentiates four major types of quantitative research report: individual--cross-sectional; aggregate--cross-sectional; individual--longitudinal; and aggregate--longitudinal. Methodological issues and empirical research findings are discussed separately for each type of study and each category of deliberate self-harm. Cross-sectional individual studies reveal that significantly more parasuicides and suicides are unemployed than would be expected among general population samples. Likewise, parasuicide and suicide rates among the unemployed are always considerably higher than among the employed. Aggregate--cross-sectional studies provide no evidence of a consistent relationship between unemployment and completed suicide, but a significant geographical association between unemployment and parasuicide was found. Results from all but one of the individual longitudinal studies point to significantly more unemployment, job instability and occupational problems among suicides compared to non-suicides. The aggregate longitudinal analyses reveal a significant positive association between unemployment and suicide in the United States of America and some European countries. The negative relationship in Great Britain during the 1960s and early 1970s has been shown to result from a unique decline in suicide rates due to the unavailability of the most common method of suicide. However, despite the firm evidence of an association between unemployment and suicidal behaviour, the nature of this association remains highly problematic. On the basis of the available date, the author suggests that macro-economic conditions, although not directly influencing the suicide rate, may nevertheless constitute an important antecedent variable in the causal chain leading to self-harmful behaviour. Further empirical research based on a longitudinal design is recommended as a matter of urgency so that a more definitive assessment of the aetiological significance of unemployment in parasuicide may be made.