Horizontal collusion or price-fixing has been a major focus of Federal antitrust enforcement since the passage of the Sherman Act, but the effectiveness of these efforts in deterring price-fixing has never been evaluated. This paper first describes a model which considers the effect of antitrust enforcement on the decisions of firms within an industry to fix prices collusively. Particular attention is given to the magnitude of the penalties and the possibility of detection. This theoretical model shows that price-fixing generally is not a discrete choice, but is instead a continuous choice regarding the optimal degree of collusive price markup. The optimal markup decreases as enforcement efforts or penalties increase. In order to test the model's implications, markups on white bread in 20 major cities for 1965-76 and measures of enforcement were developed. This study revealed that the DOJ does create a deterrent effect by bringing a price-fixing case since markups in neighboring cities fell in the wake of a DOJ action. However, analysis also showed that the deterrence generated by DOJ actions was largely due to subsequent class action suits. Additional tests performed on the Supreme Court decision in Eisen IV and on the Arizona Bakery Products litigation supported the independent deterrent impact of consumer class actions. This impact can be attributed to the large settlements involved in class actions and an unexpected aversion to such suits among those accused of collusion. The appendixes contain over 80 footnotes, a glossary, statistical tables, and 75 references. (Author abstract modified)