Kautilya On Administration Of Justice During The Fourth Century B.C.

Click to increase image sizeClick to decrease image size

Acknowledgments

I am deeply indebted to the two referees for enhancing clarity and content of the paper. I am solely responsible for any remaining errors.

Notes

¹A. K. Sen (1987, p. 5) believes that Kautilya's Arthashastra is the first book on economics. He states:

The “engineering” approach also connects with those studies of economics which developed from the technique-oriented analyses of statecraft. Indeed, in what was almost certainly the first book ever written with anything like the title “Economics,” namely, Kautilya's Arthashastra (translated from Sanskrit, this would stand for something like “instructions on material prosperity”), the logistic approach to statecraft, including economic policy, is prominent.

²Joseph J. Spengler (1971) makes a special note of legal rules regarding partnership. He (p. 79) writes: “Rules for the distribution of remuneration when work was done jointly not only were laid out by Kautilya but also found expression in commercial arithmetic. When workmen, guild members, or others engage in cooperative undertakings, they shall divide the wages as agreed upon or in equal proportions” (3.14.18).

³Since Gary S. Becker's (1968) seminal work, hundreds of articles have appeared dealing with many aspects of law enforcement. These works analyze various aspects of law enforcement and deterrence. These may be classified as: (i) rent-seeking behavior or corruption by the enforcers and its impact on economic growth and crime deterrence, (ii) judicial fairness and the minimization of legal errors in the disposition of criminal cases, (iii) the form of punishment that whether it should be monetary or non-monetary, and (iv) the time inconsistency or the credibility problems, that is, the society may not find it optimal to carry out the punishment once the crime has been committed, and the related issue of judicial discretion.

⁴Only recently has this issue drawn attention from economists. Pranab Bardhan (1997) reviews the issues related to corruption and economic growth.

⁵Kautilya's contribution is discussed in Sihag (2005, 2007a).

⁶Early Roman law derived from custom and statutes, but the emperor asserted his authority as the ultimate source of law. His edicts, judgments, administrative instructions, and responses to petitions were collected with the comments of legal scholars. As one 3rd-century jurist said, “What pleases the emperor has the force of law.” As the law and scholarly commentaries on it expanded, the need grew to codify and to regularize conflicting opinions. It was not until much later in the 6th century AD that the emperor Justinian I, who ruled over the Byzantine Empire in the east, began to publish a comprehensive code of laws, collectively known as the Corpus Juris Civils, but more familiarly as the Justinian Code.” http://www.crystalinks.com/romelaw.html.

⁷Additional analysis on this issue is provided in Sihag (2004).

⁸A. Mitchell Polinsky and Steven Shavell (2000), pp. 72–73 survey the field on law enforcement. In the last section of their article, under the sub-heading “future research” they recommend:

The behavior and compensation of enforcement agents have not been examined in this article, but this topic is important and should be studied for two reasons. First, the incentive of enforcement agents to discover violations is affected by the structure of their payments. Secondly, enforcement agents may be corrupted: they may accept bribes, or demand payments, in exchange for not reporting violations. Corruption tends to reduce deterrence, and therefore its presence obviously will affect the theory of optimal law enforcement.

In the light of Kautilya's contribution their suggestion amounts to: “going back to the future.” Similarly, David D. Friedman (1999, p. 5261) describes the various elements of an efficient system of criminal punishment, which includes “penal slavery for criminals who can produce more than it costs to guard and feed them.” He summarizes his findings as: “Hence imprisonment is always dominated by execution and both are dominated by fines and other alternatives. Modern legal systems do not fit that pattern. One possible explanation is that the ability of enforcers to profit by convictions can produce costly rent seeking.” Friedman believes that the real reason for the existence of inefficient system is to curb the possibility of rent seeking on the part of the enforcers.

⁹Charles Drekmeier (1962, p. 167) observes: “The administrative organization and regulations of Kautilya are generally taken to be a description of the Mauryan system. However, Kautilya never purports to give an account of a specific polity. It is a theoretical work, and any attempt to deduce more than the broad outlines of the Mauryan administrative system from it must bear this in mind.” It is a well-established fact that the Arthashastra is a theoretical treatise.

 Pushpendra Kumar (1989, p. xxv) also notes: “Thus he stands out as the foremost theorist of ancient India and the first to prepare a scientific treatise on state-craft with economics as the basic factor.”

¹⁰Hal R. Varian (1993, p. 162) notes: “When Markowitz defended his dissertation at the University of Chicago, Milton Friedman gave him a hard time, arguing that portfolio theory was not a part of economics, and therefore that Markowitz should not receive a Ph.D. in economics. Markowitz (1991) says, ‘this point I am now willing to concede: at the time I defended my dissertation, portfolio theory was not part of Economics. But now it is’.”

¹¹Joseph E. Stiglitz (1987, p. 966) credits Stephen Ross (1973) for coining the term principal-agent.

¹²For example, Thomas J. Miceli (1990) remarks that, “For instance, an important question of fairness relates to the incidence of errors by the criminal process.”

¹³Becker (1968) discussed only the prevention of crimes but did not suggest anything if a crime was committed. Miceli (1991) proposes a comprehensive model of fairness and deterrence, which presumably combines Becker's crime prevention model and Miceli's (1990) fairness model. However, Kautilya implicitly provided a more comprehensive approach with many additional insights. The following table 1b captures Kautilya's conceptual framework.

¹⁴A judicial process is initiated to find the guilt or innocence of a person arrested for an alleged crime. For example, Miceli (1991) defines the probabilities of legal errors as follows: He sets δ = P (G/A)=[P (A ∩ G)/P (A)] = probability that an arrested person is guilty; P_g = [P(C ∩ A ∩ G)/ P (A ∩ G)] = probability of convicting a guilty person (i.e., (1 − P_g) is the probability of not convicting a guilty person); probability that an arrested person is guilty and is convicted = δP_g = P (C ∩ A ∩ G)/ P (A). Type II legal error probability = δ (1 − P_g). Probability of convicting an innocent person = P_i = P (C ∩ A ∩ G_c)/ P (A ∩ G_c), and Type I legal error probability = (1 − δ) P_i = P (C ∩ A ∩ G_c) / P (A)=probability of arresting and convicting an innocent person.

Miceli's definitions based on the numbers: Type I error probability = 5/110 and type II error probability = 20/110. If the objective is to assess the performance of the judiciary only, Miceli's definitions are sufficient since his analysis is confined only to those who have been arrested. However, his definitions are not relevant if the objective is to deter crimes. For example, if the enforcement authorities arrest just one criminal person (out of the 1000) and convict him, that is, δ = 1 and P_g = 1. According to Miceli's definition, the probability of conviction = δ P_g = 1. But that cannot be correct since the probability of conviction of a guilty person would be = 1/1000 (=δ P_g A/G = A/G), which is very small to deter any crime. It means that Miceli's model did not achieve its goal of combining prevention of crimes and judicial fairness.

Polinsky and Shavell (2000) do not define the various probabilities explicitly. It seems that they define the legal errors in the following way. Let the probability of detection, P be defined as P = A/G = 110/ 1000, the Type I error probability (they call it Type II error), ϵ₂ = (1 − δ) A/G = 10/1000; and Type II error probability, ϵ₁ = δ (1 − P_g) A/G = 20/1000. That means in the presence of legal errors, the effective probability of detection = P (1 − ϵ₁ − ϵ₂)=δ P_g A/G = 80/1000. This is precisely, the probability of arresting and convicting a guilty person and is relevant for deterring crimes.

They present an alternative insightful interpretation of these errors. They consider the negative impact of Type I error (contrary to tradition, they call it Type II error) on crime deterrence, and they note, “The second type of error, mistaken liability, also lowers deterrence because it reduces the difference between the expected fine from violating the law and not violating it. In other words, the greater is ϵ₂, the smaller the increase in the expected fine if one violates the law, making a violation less costly to the individual.”

¹⁵See Thomas H. Wonnacott and Ronald J. Wonnacott (1977, pp. 259–60).

¹⁶On the other hand, Miceli (1990) assumes that an increase in efforts by the prosecutor to collect more evidence shifts the distributions to the right implying an increase in the probability of Type I error. He notes that prosecutors generally try to shift the distributions to the right. That is clearly against the collective sense of justice.

¹⁷Drekmeier (1962, p 254) remarks, “Kautilya: holds that danda must be applied with justice if authority is to have the respect of the people—which amounts to saying that justice is what transforms power into “authority.” Danda means punishment.

Adam Smith holds a similar view. He states, “Justice is the main pillar that upholds the whole edifice, if it is removed, the great, the immense fabric of human society must in a moment crumble into atoms.”

¹⁸Recently, Jennifer F. Reinganum (2000, p. 63) discusses the establishment of the United States Sentencing Commission to develop the sentencing guidelines for achieving certain social goals. These are very similar, as mentioned above, to those specified by Kautilya. She states

 The motivation for such guidelines included at least the following arguments. First, the then-current system of indeterminate sentencing with parole made it difficult for either the offender or the state to form a reasonable estimate of the actual sentence; definitive sentencing guidelines were believed to provide honesty in sentencing. Secondly, the sentencing guidelines were intended to reduce observed disparity in sentencing across apparently similar cases. Finally, the sentencing guidelines would build in proportionality in sentencing by conditioning the prescribed sentence on offense and offender characteristics.

¹⁹Becker (1968) reaches the conclusion that monetary fines are merely transfers and do not use real resources and, therefore, are preferable. However, Becker's suggestion has been found to be impractical and the society has to incur some cost in the collection of fines. Based on an empirical study, Robert W. Gillespie (1988-89) finds “The relatively low enforcement success achieved for large fines, particularly drug fines larger than $1000.” Gillespie casts doubt on “the use of fines as a criminal sanction in terms of lower social costs of punishment.”

²⁰Polinsky and Shavell (2000, p. 68) remark: “The implications of injurers' imperfect knowledge are straightforward. First, to predict how individuals behave, what is relevant, of course, is not the actual probability and magnitude of a sanction, but the perceived levels or distributions of these variables.” David M. Levy (1999) points out that approbation and disapprobation figure very prominently in Adam Smith's Moral Sentiments and these could have a significant effect on the behavior of potential thieves. (Incidentally, Adam Smith's Katallactic model as presented by Levy might provide a more convincing explanation of the kink in the loss-aversion function than in Amos Tversky and Daniel Kahneman (1991)). On the other hand, in recent years, the U.S. public has been demanding (from their respective state governments) the right to know if any sex offender lives in their neighborhood. This may serve as a warning to the parents so that they keep a close watch on their children. Recently, some states have passed legislation requiring the registration of sex offenders. Doron Teichman (2004, abstract) argues “That such policies have limited preventative value, yet they might be justified as an efficient way to sanction sex offenders.”