This issue is examined within the context of conventional ideas about impartiality of the administration of law and the politics of bureaucratic survival. The author discusses some key difficulties in ascertaining "what happened" and relative culpability in serious fraud trials, as well as the personal interest-inspired factors that can influence testimony. The study concludes that in the United Kingdom there is relative autonomy in decisionmaking in serious fraud cases, but that in "politically sensitive areas," such as the supply of arms to Iraq, there has been overt interference on purported "public interest" grounds with the information made available to the defense and even to the prosecution itself. The author notes that without access to the grounds for decisions not to prosecute, their "purely legal" justification is difficult to challenge, and this is of most social significance when members of social elites are concerned. The study concludes by addressing some difficulties in producing procedural models that provide genuine accountability for white-collar prosecutions and non-prosecutions in various countries. 23 notes and 46 references