If, in the criminal justice context, "mercy" is defined as forgoing punishment that is deserved, then much of what passes for mercy is not. Giving only minor punishment to a first-time youthful offender, for example, might be seen as an exercise of mercy but in fact may be simply the application of standard blameworthiness principles, under which the offender's lack of maturity may dramatically reduce his blameworthiness for even a serious offense. Desert is a nuanced and rich concept that takes account of a wide variety of factors. The more a writer misperceives desert as wooden and objective, the more likely the writer is to mistake judgments of blamelessness for exercises of mercy. Should a criminal justice system exercise mercy in its real sense (of giving an offender less punishment than he deserves, using a fully nuanced and rich account of desert)? One can imagine enormous benefits to the exercise of mercy by individuals in their personal dealings with others. A tendency toward mercy seems an admirable personal trait. However, it does not follow that mercy would be a desirable practice for a criminal justice system. Our strong interest in equality of treatment of like offenders and offenses suggests that mercy, if used, would need to be regularized in its application; punishment ought not depend upon the tendency toward mercy, or lack thereof, of the particular decisionmaker in the case at hand. But to institutionalize mercy is to create an expectation and right to it that may be inconsistent with its fundamental character of giving a relief or mitigation from punishment to which an offender is not entitled. Further, one can imagine serious effects detrimental to the effective operation of the criminal justice system were mercy to be institutionalized. Classic arguments against it would cite its effect in undermining deterrence and the incapacitation of dangerous offenders. While some of us might find these arguments unpersuasive, even the desert advocate would have reason to be concerned. A "mercy program" would seem to similarly undermine both deotological and empirical desert, failing to do justice both as moral philosophers and as the community's shared intuitions of justice would assess it. On the other hand, what if it were determined – as recent research suggests – that community intuitions tend to support some exercise of what might be seen as mercy? If one sought to distribute criminal liability and punishment in a way to maximized the criminal law's moral credibility, might such evidence of principles of mercy shared by the community suggest that such principles ought to be instantiated in law?
Robinson, Paul H., "Mercy, Crime Control & Moral Credibility" (2011). Faculty Scholarship at Penn Law. 327.