In law as in all professions, perverse incentives threatening the proper fulfillment of professional duties are a persistent problem. Commonly referred to as "conflicts of interest", they can arise from something as routine as an insurance company employing a lawyer to defend a policy holder or as extraordinary as Marcia Clark's 4.2 million-dollar deal for a book about the Simpson case. Sources of such perverse incentives seem limitless. Relationships with family, friends, or other clients, simple greed, even the laudable ideal of reforming the law to better serve society, all may tempt the lawyer to disserve a client.

The legal profession takes conflict of interest seriously. There are ethical rules, published cases and bar opinions by the hundreds, and an expanding academic literature all devoted to conflict of interest. Lawyers who violate conflict rules may incur serious sanctions-money damages, loss or reduction of a fee, discipline by the bar, or disqualification by a judge from representing a client.

Despite all this, the subject of conflict of interest remains needlessly confused in part at least because of the frequent failure to distinguish clearly between harm to the fulfillment of professional duties and risk of such harm. To understand this failure and my proposed remedies, we must grasp the difference between harm rules and risk rules.

To see how a harm rule differs from a risk rule, consider the following problem far from lawyering. Player brawls-fights involving more than one player from each team-have marred the playoffs of the National Basketball Association (NBA) in recent years and received extensive media coverage. How might the NBA stop player brawling?

One strategy would be to adopt a rule that any player who brawls is suspended or fined. This is a harm rule-it requires a player actually to engage in harmful conduct for sanctions to be imposed. Like the crimes of homicide, rape, or robbery, such a rule directly proscribes what it seeks to prevent; it requires occurrence of that harm to trigger its operation. In that sense, a harm rule is reactive.

The NBA might also seek to prevent brawling through a more indirect strategy, penalizing conduct that increases the risk of a brawl. The NBA might decide, for example, that players leaving their team's bench when a fight begins on the basketball court increase the risk that the fight will escalate into a brawl. The NBA would then adopt a rule that any player who leaves the bench during a fight is suspended or fined, regardless of whether he actually brawls. Like the crimes of conspiracy, attempt, and solicitation, this "leaving the bench" rule is proactive. It anticipates harmful conduct which may or may not in fact occur.

Both the harm rule and the risk rule just described have the same ultimate objective: preventing brawling. But they pursue it in different ways. The harm rule directly proscribes the target harmful conduct. The risk rule seeks to prevent the target conduct by proscribing conduct that has a good chance of leading to the target conduct. A harm rule is about sin, a risk rule about temptation.

An Identity Crisis
The ultimate objective of conflict of interest rules is clear: preventing failures to carry out fundamental professional duties. But how do rules about conflict of interest pursue this objective? Is the prohibition against conflict of interest essentially a harm rule or a risk rule? Unfortunately, it tries to be both.

In law and other fields, the term "conflict of interest" is commonly used loosely to encompass both harm rules and risk rules dealing with perverse incentives, often without distinguishing clearly between them. This unresolved "identity crisis" about the strategic thrust of rules against conflict of interest can result in considerable confusion.

Take, for example, the recent exchange in the news between former Secretary of Agriculture Michael Espy's defenders and Special Prosecutor Donald Smaltz, who recently indicted Espy on "conflict of interest" charges. Espy's defenders emphasize that any favors he received did not influence performance of his duties as Secretary of Agriculture, that the perverse incentives those favors created actually did no harm. Implicit in that defense is a view that conflict of interest regulations are harm rules. Smaltz, in contrast, emphasizes that Espy violated those regulations even if he did nothing in return for those favors. For Smaltz, the prohibition of conflict of interest is a risk rule. Though both Espy's defenders and Smaltz are talking about conflict of interest, one side focuses on harm, the other on risk; they therefore simply talk past each other.

To get a better sense in a legal context of the contrast between the harm interpretation of conflict of interest and the risk interpretation, consider a hypothetical case in which a defense lawyer in a high-profile criminal case enters into a deal to write a book about the case. Assume the prosecution offers the defendant a guilty plea, acceptance of which is in the defendant's best interest. Since a trial will generate publicity increasing the value of the book deal, however, the lawyer advises the client to decline the prosecution's offer. The lawyer also leaks confidential information about the case to generate publicity likely to increase future book revenues.

There are three problems here. The lawyer: (1) engaged in risky conduct-the book deal; (2) gave advice serving his own interests rather than the client's; and (3) leaked confidential information.

In legal ethics it is relatively easy to distinguish problems (2) and (3). They involve substantive harms we place in separate conceptual compartments. We lawyers refer to giving guilty plea advice tainted by self-interest as a violation of the professional duty to exercise independent judgment and the information leak as a violation of the professional duty to preserve client confidences. We talk and think about problems (2) and (3) as distinct conflict of interest and confidentiality problems.

But in legal ethics distinguishing problem (1) from (2) is difficult because our standard concepts and vocabulary do not allow it. We refer to both the high-risk situation created by the book deal's perverse incentives and the resulting tainted advice as conflict of interest and view both as falling under the duty to exercise independent professional judgment. We collapse problems (1) and (2) risk and harm-into the phrase "conflict of interest" and the conceptual compartment of independent professional judgment.

There are times when rules of legal ethics clearly distinguish between risk and harm, as in the rules about lawyers suing former clients. These are explicitly based on avoiding risk of misuse of confidential information. But more typical is the blending of risk and harm. For example, the conflict of interest provisions of the 1969 Model Code of Professional Responsibility begin with Canon 5's general harm rule, "a lawyer should exercise independent professional judgment on behalf of a client", which says nothing about risk. But the provisions which follow, both advisory Ethical Considerations and mandatory Disciplinary Rules, though supposedly derived from Canon 5, flip flop between harm and risk prohibitions. Similarly, Rule 1.8 of the 1983 Model Rules of Professional Conduct, "Conflict of Interest: Prohibited Transactions", contains a mixture of harm and risk rules.

Possible Remedies
How could we better distinguish risk from harm in conflict of interest?

We could begin with the semantically ambiguous phrase "conflict of interest", jettisoning it on the ground that it is freighted beyond salvage with ambiguity. Alternatively, and perhaps more practically, we could restrict the term "conflict of interest" to risk rules governing perverse incentives, referring to harm rules by the underlying duty each is based on, such as independent professional judgment, confidentiality, and competence.

We could also separate risk rules more clearly from harm rules. For example, the Model Penal Code reserves a section for treatment of the anticipatory offenses of conspiracy, attempt, and solicitation, clearly distinguishing them from harm offenses. Similarly, risk rules could be given a distinct place in an ethics code, rather than being placed, as in the Model Code of Professional Responsibility, under the duty of independent professional judgment. Putting both risk and harm under the duty of independent judgment, as in the Model Code, makes it difficult to distinguish between them. Putting them both there also erroneously suggests that conflict of interest is concerned only with protecting independent professional judgment, not with protecting other duties such as confidentiality and competence.

Even after we succeed in distinguishing risk from harm, questions remain. How much risk is acceptable? What reasons for taking that risk are legitimate? Who should decide whether to take the risk? These are hard questions, of course, but distinguishing risk from harm is a necessary first step to confronting them squarely and to making conflict of interest more understandable.