Objective: The study examined two hypotheses: that being laid off increases the likelihood of violence and that being employed in an industry in which employment is unexpectedly low decreases the likelihood of violence. Methods: Logistic regression analyses were used to examine data for mote than 4,000 persons who participated in initial and follow-up interviews as part of the Epidemiologic Catchment Area survey. Data for persons who were working and not violent at the time of the initial interview but who were unemployed at reinterview were examined, as were data for persons who remained employed in industries with low employment levels. Results: The risk of violent behavior of those who were laid off was nearly six times higher than that of their employed counterparts. Controlling for concurrent psychiatric disorder did not reveal a lower risk. The risk of violent behavior was reduced among those who remained employed in industries where layoffs were occurring. Conclusions: Economic contraction seems to affect violence in the community in two countervailing processes—one process increases violence and one reduces it. However, the net effect may well be that violence decreases in communities experiencing economic contraction.