In a comprehensive study extending prior research, Prince and Rubin (2002) use the event study methodology, and find negative market reaction to a sample of 15 initial filings of product liability litigation and 29 other litigation events against U.S. automakers between 1973 and 1995. They conclude that the event study methodology is a useful way to measure the costs of litigation. In contrast, after examination of a new sample of 144 initial filing events and 465 other litigation events for six major automobile firms from 1985 to 2000, and after re-examining Prince and Rubin’s data, we find that the market reaction to all but the most extreme and infrequent events is generally not significant. We suggest that the event study methodology may not generally be useful to study the social costs of litigation, but may be useful for unexpected abnormal litigation events where the potential liabilities (including reputation and other losses triggered by litigation) may far exceed the legal liability reserves set up by firms. We find mixed results for the market impact of litigation against a competitor. When a product liability lawsuit is first filed against a U.S. firm, the market values of the Japanese firms significantly decline. When a Japanese firm is sued for product liability, the U.S. firms register a significant increase in market value. However, these spillover results have to be interpreted with caution because of small sample sizes and possible confounding events.