Tuesday, October 10, 2017

Prosecuting Public Officials for their Mistakes

(This post first appeared on the Global Anticorruption Blog and has been cross-posted with permission)

In July 2011, Yingluck Shinawatra became Prime Minister of Thailand after her party (founded by her brother, former Prime Minister Thaksin Shinawatra) won a decisive electoral victory. One of her principal campaign promises was to establish a program to purchase rice from farmers at above-market prices then store the rice to reduce supply. The hope was that doing so would increase world prices—because of Thailand’s position as the leading global rice exporter—ultimately allowing the government to sell at a profit. Shortly after the election, Yingluck’s government implemented this program, and it worked well for a few months—until other global players increased their supply of rice, causing Thailand to lose billions of dollars in the process. This economic debacle was entirely predictable—and indeed was predicted by many experts. And the program itself was beset by allegations of fraud and corruption in its implementation.
But should the failure of the rice-buying program be the basis of a criminal charge of corruption and a prison sentence against Yingluck herself, in the absence of evidence that she was directly involved in any embezzlement, bribery, or other more conventional forms of graft? Section 157 of Thailand’s Penal Code allows for just such a prosecution, as this section makes it a crime for a public official to either dishonestly or “wrongfully discharge or omit to discharge a duty so as to expose any person to injury.” And last month, the Thai Supreme Court found Yingluck (out of power since she was deposed by a military coup in 2014) guilty and sentenced her to five years in prison. She fled the country before the verdict.
Thailand is not alone in adopting anticorruption laws that criminalize not only dishonest conduct (bribery, embezzlement, conflict of interest, etc.), but also negligence or incompetence. When India updated its anticorruption law in 1988, it added a new provision that makes it a criminal offense for a public official to “obtain for any person any valuable thing or pecuniary advantage without any public interest.” This broad offense was interpreted by a state High Court to not require any proof of dishonesty or criminal intent, and the Central Bureau of Investigation (India’s premier anticorruption agency) has routinely employed the provision in grand corruption cases to avoid the problem of having to prove corrupt intent. In perhaps the most high-profile such prosecution, the agency went after an ex-Prime Minister, Dr. Manmohan Singh. Dr. Singh was the Minister of Coal at a time when the Government decided to liberalize allocation of coal-blocks and to sell mining rights to private parties. In 2014, the Comptroller and Auditor General’s office reported the policy had caused losses worth billions of dollars because the rights had been sold for too little, through a process that was too ad hoc to be considered legal. Dr. Singh was subsequently charged under India’s broad law, though his trial has currently been stayed while his challenge to the constitutionality the law is pending before India’s Supreme Court. (There are clearly concerns in other quarters about the breadth of this statute: In 2016 a Select Committee of the Upper House of India’s Parliament submitted a report that suggested India eliminate this offense. Parliament hasn’t yet acted on this recommendation, but there are signs that it has some support.)
Is it appropriate to enact broad anticorruption laws that allow government officials to be convicted for dereliction of duty, acting in a manner contrary to the public interest, and the like? Anticorruption activists and prosecutors may find such statutes appealing: It is easier to secure convictions of elected officials who are suspected of corruption, but where it is too difficult to prove the specific intent necessary for traditional corruption offenses. But in fact these broad laws are likely to do more harm than good, and countries like Thailand and India would be better off without them. There are three main reasons for this:
  • First, criminal prosecutions carry immense expressive value and signal to society that certain acts ought to be condemned. The message sent to society, when a court convicts a senior official for a policy failure absent any showing of dishonesty, is not likely to be that the government will not tolerate corruption, but rather that prosecutors and judges have broad power to jail leaders that they don’t like. This is only worsened by the inherently subjective nature of the crime in question. These prosecutions also blur the lines between genuinely dishonest conduct and policy mistakes, undermining the special condemnation that ought to be reserved for the former. On balance, even if prosecutions for dereliction of duty might enable prosecutors to convict leaders suspected of actual but unprovable graft, in the long term such prosecutions will only worsen the respect for rule of law in societies troubled by corruption in governance.
  • Second, and connected to the first point, such broad offenses raise the very real risk of politically motivated prosecutions. In countries where voters are angry about corruption, politicians often stress increased anticorruption prosecutions to show their commitment to reform. What better way to show such commitment then by attacking the alleged corruption of the previous regime? And this is much easier to do when the prosecutors (influenced by the new government) need not actually prove dishonesty, but need only find some alleged failure to fulfill a duty, or some act not in the “public interest.” Not only are partisan witch-hunts contrary to the rule of law, they also don’t always make for sound cases, and can ultimately dilute both the strength of using the courts as an anticorruption force and the faith of the public in the integrity of the institutions of justice.
  • Third, broad anticorruption laws that effectively criminalize negligence are likely to have an undesirable chilling effect on decision-making by public officials. When elected officials take bold moves they already run a risk of alienating their electorate and losing a re-election. Add to that the specter of eventual criminal liability once they’re out of office, and it officials are likely to become too timid to try anything even slightly bold or controversial. This fear of risk trickles down to any official tasked with implementing government policy.
To be sure, elected officials owe a duty to the citizenry to act in the public interest. If they fail to do so, and the failure can be proven to be due to dishonesty or greed, then more narrowly-drawn laws against bribery, embezzlement, and conflict of interest would apply. For serious errors that are made in good faith – or even in those cases where there’s suspicion of wrongdoing but it can’t be proved – there’s an alternative remedy already in place: elections. The political process is of course imperfect. Many Thais were likely frustrated that Yingluck didn’t seem to pay a serious political cost for the failure of the rice-buying scheme, and may have been relieved both by the coup that removed her from power and by her eventual conviction for dereliction of duty. But in the long term, the democratic process offers a better safeguard than allowing for politicians to be sent to jail for mistakes.

Saturday, September 30, 2017

Guest Post: In Defence of Sonu v State of Haryana

(I am happy to host a Guest Post by Mr. Lakshya Gupta, who is a 2017 graduate of the National Law University, Delhi, and is currently practising criminal law in Delhi

As has already been discussed on this blog, Anvar P.V. v. P.K. Basheer [(2014) 10 SCC 473, 'Anvar'] held that the only way to make electronic records admissible by way of secondary evidence is by adducing a certificate under Section 65-B of the Indian Evidence Act 1872. This blog has argued that:

a)   As per Anvar, the certification under Section 65-B is the only mode of proof for secondary evidence of electronic records;
b)  Sonu @ Amar v. State of Haryana [AIR 2017 SC 3441, 'Sonu'] made a distinction between ‘inherent admissibility’ of evidence and its ‘mode of proof’ and held that the requirement of certification under Section 65-B pertained to the latter and not the former;
c)  Ruling in Anvar must be interpreted to mean that absent a certificate under Section 65-B, secondary evidence of electronic record is rendered inherently inadmissible;
d)   Since Sonu which is a two-judge bench, ruled contrary to point c), it is at fault in not applying the law laid down by Anvar, which was decided by a higher bench of three judges.   

I am in agreement with points a) and b). I, however, contest point c) and consequently d). Further, I argue that Sonu applies Anvar retrospectively and decides a different issue than the one determined by Anvar.

The ruling in Sonu

Sonu relies on a two-judge bench decision of the Supreme Court in R.V.E. Venkatachala Gounder Vs. Arulmigu Viswesaraswami and V.P. Temple and Anr. [(2003) 8 SCC 752, 'Venkatachala'] wherein a distinction was been made between admissibility of a document in itself (inherent admissibility) and the manner or mode through which it is sought to be made admissible. The Court in Venkatachala held that objections with respect to the former could be raised for the first time even at the appellate stage while objections with respect to the latter could not be raised once evidence had been tendered. According to Venkatachala (SCC version, paragraph 20), the:

“… crucial test is whether an objection, if taken at the appropriate point of time, would have enabled the party tendering the evidence to cure the defect and resort to such mode of proof as would be regular ...”

Viewed in light of the reasoning behind Venkatachala, the ruling in Sonu is that since the objection to admissibility (absence of certificate under Section 65-B Evidence Act) of the electronic record (paper printouts comprising Call Detail Records – secondary evidence of their contents) dealt with the mode of proof, such an objection could not be entertained at the appellate stage, if the same was not raised at the time when the electronic record was submitted in evidence at the stage of trial. This is because had an objection been raised at the trial stage, that would have presented an opportunity to the prosecution to cure the defect as to the admissibility at that stage itself. Since no occasion would be available to cure the defect at the appellate stage, it would be unfair to the prosecution if the evidence on record at the time of trial and not objected to then, was omitted from being considered at the appellate stage.

The ruling in Anvar cannot be interpreted to mean that absence of certificate under Section 65-B renders secondary evidence of electronic record inherently inadmissible

From a close reading of Venkatachala and Sonu, it appears that the question of inherent admissibility concerns the nature of the evidence (electronic record) and is separate from the question in what way this evidence may be made admissible in court (as either primary or secondary evidence). So the question that whether printouts comprising Call Detail Records (electronic record in Sonu) or a CD with recorded files (electronic record in Anvar), is something that can be admitted in evidence, decides the inherent admissibility of such CDR or CD. Anvar does not deal with this question, but addresses whether these electronic records can be read in evidence without a certificate under Section 65-B of the Evidence Act. Anvar, as correctly pointed out by this blog, unequivocally answers this question in the negative.

The holding that this is the only manner in which an electronic record by way of secondary evidence may be read in evidence cannot be interpreted as a comment on the nature or inherent admissibility of the electronic record. While a certificate under Section 65-B certainly has a bearing on the authenticity of the electronic record, it does not in any manner, alter or affect the contents of such CD or CDR. Whether or not a certificate under Section 65-B is supplied, the nature of the CD or the CDR remains unchanged. The question of how you establish its authenticity is different – and while State (NCT of Delhi) v. Navjot Sandhu @ Afsan Guru [(2005) 11 SCC 600] provided the option of establishing authenticity with or without a certificate, Anvar held that authenticity can be established only through certification under Section 65-B. This question of establishing the authenticity relates to the mode of proof, the only issue discussed in Anvar. Therefore, contrary to what was argued by this blog, I submit the judgment in Anvar does not deliberate on the issue of inherent admissibility or the nature of an electronic record.  

I must also counter a possible response. An argument may be made that since Section 65-B is a deeming provision, an electronic record can be deemed to be a document only if conditions under Section 65-B are satisfied. If the electronic record fails to meet these conditions, it does not qualify as a document and hence becomes inherently inadmissible in evidence. However, as was held in Anvar itself, the deeming of an electronic record as a document depends only on conditions under Section 65-B(2) and not on the certificate under Section 65-B(4). It must be noted here that conditions under Section 65-B(2) relate to the circumstances of the ‘computer’ and the manner of production of the ‘electronic record’ by such ‘computer’. If these circumstances and manner of production exist, then only the electronic record can be deemed to be a document as per Anvar. Now, the question of inherent admissibility of the electronic record would depend on the existence of these circumstances, and not on the manner in which they can be proved before the Court. Even if the only manner in which they can be proved to exist is through a certificate, as was probably held by Anvar, their existence itself determines the inherent admissibility of the record and not the manner in which their existence is proved. It must also be noted that Anvar nowhere expressly observes that an electronic record cannot be deemed to be a document absent a certificate under Section 65-B.  

Sonu applies Anvar retrospectively and decides a different issue that the one determined by Anvar

Sonu recognises that since the law laid down by Anvar applies retrospectively, requirement of a certificate under Section 65-B was necessary to make secondary evidence of electronic records admissible – and that must be deemed to have been the position of law from the introduction of Section 65-B in the Evidence Act. Since this was the position of law, the objection as to admissibility (failure to submit certificate under Section 65-B) should have been raised at the stage of tendering of evidence. Since the objection had not been raised during trial at the stage of evidence, it could not be entertained at the appellate stage.

So basically, the Court is telling the accused/appellant that – we agree that the position of law is what Anvar held, and so you ought to have argued it at the time when evidence was being lead during trial, and you cannot argue it now, at the appellate stage if you didn’t raise the argument during trial. Hence, the Court in Sonu in fact realises it was bound by in Anvar and reaffirms it.  

In Anvar, there is no doubt that the plea of non-admissibility of electronic record has been accepted by the Supreme Court at the appellate stage. It is crucial here to note that the reasoning of Sonu does not preclude the Court from entertaining objection as to admissibility of the electronic record at the appellate stage, but it bars the defence from raising that objection at the appellate stage when this was not taken at the time of tendering of evidence. Now, this would be a legitimate course of action available to the Bench in Sonu if it can be established that Anvar did not consider the issue determined by Sonu, which is - whether a plea regarding non-admissibility of electronic records (due to absence of certification under 65-B) could be taken at the appellate stage if the same had not been raised when evidence was being tendered during the trial. A look at the High Court decision in the Anvar case (Election Petition No. 3 of 2011 in High Court of Kerala) shows that the plea regarding non-admissibility of CD’s was raised by the petitioner even in the High Court.[1] 

Sonu also notes that Venkatchala was a civil case, and also places reliance on the three-judge bench decision in PC Purshothama Reddiar v. S. Perumal [(1972) 1 SCC 9, 'Reddiar'] which pertained to admissibility of police reports in a criminal trial. The defence in Reddiar had objected to the admissibility of police reports (marked in evidence without any objection during trial stage) on the ground that the police officials who had covered those meetings had not been examined. The Court held it was not open to the accused to raise an objection about the admissibility of the police reports when no such objection was taken at the time when evidence was being lead during trial. While Reddiar did not make a distinction between inherent admissibility and mode of proof, it is clear that the Court was unwilling to entertain objections pertaining to admissibility when they had not been raised during trial. Hence, it was legitimate for Sonu to decide an issue which was not considered in Anvar and the determination of which is in consonance with a bench co-ordinate to Anvar.

As has been correctly identified on this blog, Sonu is concerned that retrospective application of Anvar is ‘not in the interests of administration of justice’ for a large number of criminal cases that have already become ‘final’. However, I disagree that Sonu was a misstep, and submit it stands on firm legal footing.    

[1] The argument in the High Court was that since the CD’s were secondary evidence of the content of the recordings contained therein and since primary evidence of this content (recordings created and stored on mobile phones, digital camera or the computers to which they were transferred) was not submitted in court, the secondary evidence (CD’s) was not admissible since it could not be relied upon as an authentic source.

Wednesday, September 27, 2017

Mahmood Farooqui's Appeal and a Problem of Labels

On 25 September 2017, a single judge of the Delhi High Court allowed an appeal filed by Mahmood Farooqui challenging his conviction under Section 376 for rape, where he was sentenced to undergo seven years rigorous imprisonment and pay a fine of Rs. 50,000/-. It comes in the wake of the Punjab and Haryana High Court suspending the sentence of three students convicted of gang-rape in another case that attracted significant media attention. The decision has garnered mostly negative criticism from what I gather (see here, and here), and I am certain more scrutiny of the opinion will come during the week. At the outset, while the fast pace of the proceedings must be applauded (the appeal has been decided around two years after filing of the complaint itself), it only reminds us of the other side of that coin which is, unfortunately, the only one that 'have-nots' unfortunate enough to be stuck in the Indian criminal justice shall ever see. How many regular hearings from past years suffered so that Mr. Farooqui's case could be heard, we shall never know.   

To have my two cents worth as a lawyer, I think the decision is questionable and there is a good case in appeal. I certainly do hope that an appeal will be filed soon by either the the State / Victim before the Supreme Court. The High Court decision does not, with certainty, tell us why the conviction is overturned beyond telling us that the prosecution case was not proved beyond reasonable doubt. What I mean is, that the Court does not fully commit to either saying that it (i) the entire allegation of sexual acts was not proven, or (ii) only the non-consensual nature of the sexual acts was not proven. Given the amount of time spent by the decision in explaining the idea of consent, one may think it is the latter. The Court has said in paragraph 102: "But, it remains in doubt as to whether such an incident, as has been narrated by the prosecutrix [victim], took place and if at all it had taken place, it was without the consent / will of the prosecutrix and if it was without the consent of the prosecutrix, whether the appellant could discern / understand the same.

Because of this (and because I know that a lot will be said about this aspect by more competent commentators) I refrain from discussing in detail the lengthy discussion on consent that the High Court engages in. Suffice to say that everyone thought we had moved past a time when courts would tell us that "instances of woman behaviour are not unknown that a feeble 'no' may mean a 'yes'" [see paragraph 78]. Such inferences are precisely what the 2013 amendments to the Indian Penal Code sought to exclude when it added an explanation to state that 'consent' for the Penal Code required an unequivocal agreement. The court, instead, offers us another idea and suggests the consent definition may be flipped to requiring proof that there was not an unequivocal disagreement, in situations depending on various factors such as whether the parties are persons 'of letters' and 'not conservative'. 

All this should convince readers that the decision is very muddled and will make for good arguments in the Supreme Court. That appeal will not raise any discussion on what are, I think, deeper problems that this case highlights: a problem of labels and criminal conduct. Rape, Murder, Robbery, Extortion - all these are labels that have carried on in language to describe certain kinds of acts. Legislatures have the authority to change their meaning but don't do so easily, because they acknowledge the connotations of these labels. For instance, if a corporation dumps toxins polluting rivers which leads to death, calling it murder may not cut it, so you make a different label for that kind of act. The Indian legislature decided to change the meaning rape and expanded the kinds of acts amounting to rape in 2013 to include non-consensual oral sex. It could have done so differently, i.e., by adopting different labels for different kinds of acts (treat penetrative and non-penetrative acts differently, for instance). It could have also shed the label of rape altogether, as has been done in other countries. But it decided to stick to the old label, and in doing so it hoped that the condemnation the law had reserved for particular kinds of acts by labelling them as rape for a hundred years could be extended to other acts. It also hoped that judges who had been trained to not think of certain acts as rape, would change their minds simply because the law said that they had to now. I think Mr. Farooqui's appeal shows us that this experiment is not working, at least not yet with judges where (on an average) they seem to have distinctly different social mores than the parties. It might not have been the best idea to adopt an aggregator term and not adopt a more granular approach for sexual offences. The judgment conveys to me that faced with the binary choice between holding a person guilty of rape or not, the judge could not do it because of how serious the accusation is. Was that illegal? Perhaps. But do judges simply apply the law? Of course not, and pretending otherwise will not help. Judges have biases, and smarter laws should account for them. If we assume as a base position that most judges in India are male and hold gender biases (implicit or explicit), then why create an architecture that only gives them two options and stiffens how the bias operates? A better architecture would factor in that bias, and probably avoid decisions like this one.

(this post was updated on 27 September, 2017. The reference to Section 114-A Evidence Act earlier was erroneous, and has been corrected.)

Saturday, September 16, 2017

The Indian Evidence Act - Ringing the Bell for Reform

As an undergraduate student, my professor stressed on the beauty of the Indian Evidence Act 1872 [IEA] before leading us down a traumatic three months. I did not know it then, but he was echoing a very commonly held sentiment of reverence for the statute (for instance, the Indian Law Commission in its 69th Report thought the IEA was entitled to a 'place of pride' in the statute book). During my brief experience in the trial courts, I saw the IEA in operation and have also had opportunities to talk about some issues under its specific provisions on this Blog. But the more I think about, I can't help but arrive at the conclusion that the IEA is a horribly antiquated piece of legislation that has served us well beyond its time. It must go. I am hardly the first or the last person to come to this conclusion - but what surprised me is most of my predecessors have not been from India. Instead, I was exposed to rich argumentation from Singapore and Tanzania, countries where the IEA was transplanted by colonialism. Here, I first take up structural issues and then address functionality concerns, both of which strike at the heart of the statute's existence. My comment heavily borrows from these debates and while the articles are not all freely available, I have indicated sources at the end for reference.*

No Jury, No Problem? 
The 69th Report of the Law Commission (referred to above) was a momentous effort. The Report is massive and provides introductory material on the law of evidence, then considers all lengthy 167 provisions in minute detail, and records notes of dissent. But the context of its birth is as impressive as its comprehensiveness. Democracy had just been revived in India after being thrown into a comatose during a two year long Emergency. The Janata Government had won elections, and riding a wave of popularity undertook a project to review various laws. The 69th Report was part of this project and we get a sense of the urgency from the Introduction which the Chairman ends by saying: "At present the Commission is engaged on the study of Transfer of Property Act; and this task again is arduous and exacting. But let me assure you that the Commission has undertaken this task with the full confidence that it will be able to forward to the Union Government its report on the subject before its tenure expires on the 31st of August this year [the introduction is dated May 9, 1977].** 
In this massive, 915 pdf page document, I was dumbstruck to find not a single discussion about the abolition of the jury system in India and its effects on the IEA. The only mention of jurors and assessors is on printed page 869 which recommends deletion of Section 166 IEA, which referred to the role of jurors and assessors where they were part of a trial. All the Commission said was "the system of trial by jury has been abolished ... this section should therefore be deleted." Is this the limit to which a jury system affected the IEA? Absolutely not. 

In a jury system, judges are not triers of fact but decide issues of the law being applied to the dispute. Laypersons decided the factual matrix and whether innocence or guilt flow from this consideration. Admittedly, the IEA was not only written with that idea in mind and was designed to be as effective for judges trying the facts as for cases where juries were triers. But, logically, Stephen was required to cater to the lowest minima rather than design rules for judges in designing the legislation. While Stephen thought judges in India were not top notch, like most Englishmen he thought worse of juries, which explains the strictness of his rules if we think of them as being designed for jurors of poor ability. Moreover, certain rules were clearly designed to limit their prejudicial effect on the minds of jurors that have no basis in a case being tried solely by judges. This includes the 'similar fact' rule, the desirability of which has been questioned in several articles written in the Singaporean context.

Trials had the judges screen the material before it came before a jury to ensure it complied with the rules in the IEA. This system makes objections central: since you can't 'unring the bell', you have to ensure the dubious material never reaches the jury in the first place. Replacing the jury with the judge does not reduce the force of that argument at all for, after all, judges are only human. In retaining the same structure of evidence rules we continue in foolishly hoping for judges to 'unring the bell' by deciding on relevance objections themselves after having seen the material. In fact, this has only been worsened after the Indian Supreme Court held in Bipin Shantilal Panchal v State of  Gujarat [(2002) 10 SCC 529] that objections have to be recorded and then decided at the end of trial during final arguments. Effectively, we now let judges see the evidence, touch it, keep it with them while the case develops, and hope that in the end they can exclude that evidence from their consideration of the case.

Conflating Relevance and Admissibility - Stephen's Splendid Mistake 
That is how Pollock described the idea most central to the IEA. He was joined by Thayer who noted that inverting the English position and fitting in all of evidence law within narrow rules of relevance is what deprived Stephen's work of 'permanent value'. Let me give some context. Traditionally, the basic rule of evidence is that everything relevant is admissible. The Federal Rules of Evidence in the United States explain relevant evidence through Rule 401 as anything making a fact more or less probative, which fact is material to the case at hand. This loose formulation leaves it for the trier to decide whether things are relevant. This enquiry is governed by a framework which excludes relevant evidence for policy concerns, for instance, excluding confessions elicited through use of prohibited force.

In the IEA, Stephen reversed this logic. He prescribed strict rules for determining what is, and is not, relevant. He had his reasons as I briefly mentioned above - judges and lawyers being poorly trained in India, juries needing more careful handing. But beyond this reversal, Stephen did not change much of the English law of evidence - what were, traditionally, inquiries of admissibility, had been converted to be understood as questions of relevance and bundled together through Sections 5-57. Out of these, I would term some as hard rules and others being soft. For example, the exclusion of confessions to police officers by accused persons under Section 25 was a hard rule because it did not allow for the evidence to be considered at all. Whereas provisions such as Section 7 (explaining how certain facts are relevant if they establish cause and / or effect) are soft rules because they allow for facts to be looked at, and then applied.     

Keeping aside the slurs on judges in Indian courts, one can imagine a trial working with a jury in largely the traditional manner, albeit with tighter restraints on what is fed to the jury. The judge still exercises oversight, and lawyers raise objections to prevent bad material from being considered. But take the jury away, and it simply does not work anymore. What we have then, is judges ignoring all the soft rules on relevance with all hard rules being conflated with admissibility. The ignorance of soft rules is compounded by the extremely convoluted and overlapping nature of those provisions, which in effect are treated as statutory verbiage for the traditional rule that whatever the judge thinks is relevant, is admissible. I am not making heretical claims here - a bare perusal of appellate decisions confirms how judges have a free reign on bringing in evidence as long as it is not barred. Relevance, as Stephen originally created in the IEA, has thus ceased to function.

The Law Commission reviewed the IEA again in the 185th Report (access it from the link) which was even longer than the 69th Report. Yet, there was no consideration of these issues and all the Commission did was express its agreement that Section 166 ought to have been deleted (which it was). The IEA is a bad statute. Its core is rotten, its context altered, and several provisions - the ones on judicial notice and presumptions especially - either just don't make sense or can be significantly reduced in size. And this post does not utter a word about how the Supreme Court has tortured some of the text over time. 

Why, then, must we continue to carry the burdens of this clunky, nearly 150 year old document? The easy answer is the cynical one - lawyers don't want things to change because they will go out of business soon enough if the law became simpler. But that can hardly suffice. The other answer can be found in the same 69th Report where the Law Commission gave the IEA 'place of pride'. It went on and noted that "respect for its excellence should not amount to blind adoration bordering on deference." The deference, as I have tried to show, is entirely unwarranted.



The Singapore Evidence Act first came in as the Straits Evidence Ordinance in 1893 and, barring a few modifications, is a nearly complete copy of the Indian Evidence Act 1872. I have relied upon the following pieces from the Singapore context: (i) Robert Margolis, Evidence of Similar Facts, the Evidence Act, And the Judge as Trier-of-Fact, 9 Sing. L. Rev. 103 (1988); (ii) Robert Margolis, The Concept of Relevance: In the Evidence Act and the Modern View, 11 Sing. L. Rev. 24 (1990).

Tanzania has the Tanzania Evidence Act 1967 which, again, is a nearly complete copy of the Indian Evidence Act 1872. In 2011, the Tanzanian Government initiated a project to reform the law and the team included Professor Ronald Allen. Their efforts were published in a three part series run in the Boston University International Law Journal which I would strongly urge everyone to read: Part One, Part Two, and Part Three is the Draft Law prepared.

The Janata Party failed to get any of the criminal law reform bills passed through parliament. 

Wednesday, September 13, 2017

Criminal Procedure and Clogged Dockets - Reverse Engineering Solutions?

The delay and docket logjams in the Indian criminal justice system have been so abysmal for so long that, at some level, they've lost their shock value. Year on year, local and international reports will be dutifully published highlighting the plight of the system. The government has done its bit too, with the Indian Law Commission having reviewed the issue of delays on multiple occasions (the reports can be accessed on its website). Beyond losing their shock value, while reading most of these reports I also saw that the pervasive nature of the problem has also rendered discussions on solutions stale, with the same drum being beaten incessantly (for exceptions, see this piece by Vrinda Bhandari).

The solutions (give or take a few) range from (i) increasing courts, (ii) increasing fast-track courts, (iii) increasing alternate dispute mechanisms like Lok Adalats, (iv) make judges work more, (v) introduce time-limits for cases, (v) reinvigorate the dormant plea-bargaining system, (vi) enhance quality of investigations. If the criminal justice system were to be viewed as a funnel (to borrow an old and often-repeated imagery) with the persons who don't commit crimes at the broad top of the funnel and those who go to court at the narrow tip, then we see that these solutions are all focused at the tip. Everyone is talking about what happens when we reach the end of the funnel - why not look at what happens at the start itself?

If you think that there are no steps from the top of the funnel till its tip, you're quite mistaken. There are various steps, and each involves an exercise in discretion. Take speeding on the road as an example. First, not everyone who breaks the law gets apprehended - so the first filter is how many people get caught. Out of those who get caught, the second filter is how many of those are arrested and proceeded against at all (a talking to is all that you get). Next, out of those proceeded against, how many are actually made accused persons with a charge-sheet against them. After which, there is the possibility of discharge / dropping proceedings / settlement. Only after this do we get to cases that stay in courts.* So, what I am suggesting is that the current debates are (at best) only looking at when the courts get involved. But there is so much that happens before that needs to be looked at as well.   

At this stage you might have two conjoint objections arguing basically that everyone who gets caught in the web of law enforcement should be prosecuted and taken to court. Should they, really? The view is based on a gross assumption that the criminal justice system is capable of handling so many cases, which it is not. And in our already bloated criminal justice system, the capability to add fresh cases each year is especially reduced. Do the police know this? Of course they do. Are they encouraged to exercise greater restraint in taking cases forward? I don't think so. Should they be the ones making the decision to take the case forward? That is the big question that I think is not being asked often enough in our system. Most reports on police reform that I saw did not discuss this either and were more focused on improving administrative structures and the quality of investigations, rather than discuss the exercise of discretion (see, Model Police Bill 2015, PRS Report on Police Reforms 2017, Model Police Act 2006, Human Rights Initiative Report

If I am a police officer who invested months in an investigation, and then another few months in preparing the file of the case (all thanks to our cumbersome processes), will I ever think that the case should not be tried and justice not be done? If my track record rests on how many charge-sheets I file, am I not incentivised to add cases to the courts? The case is different from the crime and since the police are experts at dealing with the latter, they often make bad decisions about the former. This why the Criminal Procedure Code 1973 [Cr.P.C] stipulations in Section 169 requiring police officers to not take cases forward when there is deficiency of evidence are rarely applied in practice. Cases may not get taken ahead for many reasons when a crime is certainly committed - lack of evidence is obvious, but equally important are cost-benefit considerations since each trial imposes sever costs on the exchequer (as a judge constantly reminded me in the District Courts at Saket in New Delhi, you lawyers make citizens pay for each adjournment you take).  

Should, then, we have greater discretion being invested in prosecutors to take the decisions of what cases reach the court? This is common practice across the globe. Owing to certain issues with political interference in prosecution office across states with several posts in the Directorate of Public Prosecutions lying vacant without court intervention, I don't know how useful will it prove to have prosecutors make those calls. But I don't merely proceed on an abuse basis to make a claim - there are legal objections towards such a re-orientation. This comes from the primarily inquisitorial model that the Cr.P.C. contains in which judges have supervisory powers over the investigations being conducted by the police. The Supreme Court has already construed this to include a pro-active duty to monitor investigations in Sakiri Vasu and so I argue that this duty requires courts to throw bad cases away as well. This opportunity will not only present itself at the start when a police officer has to report to a magistrate upon beginning an investigation under Section 157, but will come periodically if a person is in custody due to the limits placed upon pre-trial detention under Section 167. Moreover, if the case is sent to the Magistrate after completion of investigation then the Magistrate can only recommend cases be sent for trial after considering various factors such as evidentiary satisfaction, costs, docket control etc. This would, naturally, mandate that the trial not proceed before the magistrates.     

This is a germ of an idea which is riddled with problems that I hope get exposed in the comments. But I do think that in reversing the focus and looking at the top of the funnel rather than the end, we might be able to come up with better solutions for our ailing criminal justice system. 

* The funnel analogy and this discussion are not my own ideas and were exposed to me as part of a class that I am currently enrolled in for my LLM at Harvard Law School. 

Monday, August 28, 2017

Supreme Court Clarifies Default Bail and Section 167 Cr.P.C. - Or Not?

If criminal procedure is an exercise of the law regulating how power (the State) can take away an individual's liberty to compel her to participate in the criminal process (as Foucault suggests), then Section 167 perhaps is the most important provision in the Indian Criminal Procedure Code of 1973 [Cr.P.C.]. In a country rampant for high rates of pre-trial detention, this provision limits how long such detention can (ordinarily) last during an investigation after arrest. The primary limit of 24 hours is honoured only in breach, and from there we move to steep cliffs: limits are placed at sixty days and ninety days depending on the kind of allegations initially levelled. If the police do not complete an investigation within this time by filing a Final Report / Chargesheet, then the accused is entitled to bail. This is popularly called 'default bail'. 

The Questions
Here, the devil is in the detail, and I've fleshed it out below to highlight the interpretation issue that arises:
  • Detention can last up to ninety days for offences punishable with death, imprisonment for life, or imprisonment for a term not less than ten years
  • Detention can last up to sixty days for any other offence.
Now, how do we construe the phrase "imprisonment for a term not less than ten years"? Does it include offences punishable with a term from zero to a maximum of ten years as well, or does it only cover offences with at least a ten year prison term? I have briefly discussed this issue on the Blog, and very recently it was taken up by a three judges bench of the Supreme Court in Rakesh Kumar Paul v. State of Assam [SLP (Crl.) 2009 of 2017, hereafter Rakesh Kumar Paul]. A two judge majority held that only offences punishable with at least a term of ten years imprisonment could lead to detention for up to ninety days. The majority also held that it is not really necessary to file a specific application under Section 167 Cr.P.C. for being released on default bail, oral arguments could cover it too without specific written averments. In fact, the majority suggested that a court was required to assess the circumstances and apprise an accused if she was entitled to default bail and grant bail if an accused could furnish sureties. In this post, we focus on the first of these issues in detail.

Procedural History 
Why was a bench of three judges hearing the issue? Because two different division benches had come to seemingly different conclusions in Rajeev Chaudhary v. State [AIR 2001 SC 2369] and Bhupinder Singh v. Jarnail Singh [(2006) 6 SCC 277]. In Rajeev Chaudhary, it appears that the Court held that offences must have at least a ten year term (I say 'appears' as the procedural history is dodgy and the majority and minority in Rakesh Kumar Paul fence over what the decision actually held; the majority found it prescribed a mandatory minimum while Pant, J. suggested the holding was ambiguous). As against this, in Bhupinder Singh the bench made obiter remarks implying that a mandatory minimum was the incorrect interpretation (Obiter, since the offence there was punishable under Section 304-B IPC with punishments ranging from a mandatory minimum of seven years to life imprisonment, thus placing it squarely within the ninety days column).

In Rakesh Kumar Paul, the allegations were under Section 13(1)(c) of the Prevention of Corruption Act 1988 which carries punishments ranging from a mandatory minimum of four years to a maximum of ten years. The Petitioner had filed a bail application after completing over sixty days in custody, which was dismissed. No Final Report / Chargesheet had been filed yet, but one was filed before expiry of the ninety day period. The High Court also dismissed the plea for bail. No specific averments were initially made in the bail application on 'default bail' but the plea was purportedly raised during oral argument, and addressed by the courts which held that an offence punishable up to ten years imprisonment was part of the ninety days category. Since the police filed a Final Report / Chargesheet within ninety days, there was no merit in the default bail claims.*

Majority v. Minority in Rakesh Kumar Paul
Lokur, J. has written the main majority opinion with which Gupta, J. concurs, while Pant, J. penned a dissent. Their conclusions have been mentioned above, but how did they arrive at these? The majority bases the holding on two prongs: (i) tracing the statutory history of Section 167, and (ii) locating this within a larger argument that personal liberty is prized under the Constitution and the Cr.P.C. The development of Section 167 Cr.P.C. unarguably shows how the provision came in 1973 to remedy a problem caused by the tight limits of a fifteen day detention period under the 1898 Code which were never honoured. The sixty - ninety day distinction came in through subsequent amendments in 1978, and Lokur, J. considered the text of the Amendment Bill (ultimately not fully incorporated) which read "imprisonment for not less than ten years or more" to conclude that the legislative intent was to only have offences with a clear sentence of at least ten years. He then grounds his argument in the idea that the Constitution and Cr.P.C. stress on preserving individual liberty, and a broader reading of Section 167 Cr.P.C. would defeat this larger objective. Both judges suggest that their interpretation honours the idea that only serious offences get the possibility of lengthy detention. 

Before moving on to the minority opinion, a word about the concurrence by Gupta, J. According to him there is no ambiguity in the text of Section 167 Cr.P.C. which bars resorting to external aids to interpret the provision (for some reason the opinion keeps referring to the 1898 Cr.P.C. in this context and how 117 years have passed since the enactment without appreciating that Section 167 never existed in the erstwhile Code). However, he expressed concurrence with the reasoning of Lokur, J. which did not think so. This makes it difficult to understand the place of the concurrence itself.

The minority opinion does not think that the text is clear and again revisits the legislative history to draw legislative intent. But now Pant, J. turns to the speeches of the particular member of Parliament who moved the Amendment Bill in both houses of Parliament. He emphasises a count given by S.D. Patil that specifies to which offences would the ninety days clause apply - which lists out thirty-six offences as punishable under the category of "imprisonment for not less than ten years". Scanning the Indian Penal Code 1860 [IPC], Pant. J. demonstrates how this could only be if we include offences that also have a maximum of ten years imprisonment. Therefore, he disagrees with Lokur, J. that the legislative intent pointed to the opposite conclusion and finds that the plea for default bail ought to be dismissed in the present case.

Evaluating the Decision
The majority decision in Rakesh Kumar Paul is, unfortunately, rather problematic when one goes beyond the rhetoric of individual liberty that it offers. For this comes at the expense of engaging with the legal issue at hand. A broader problem is that while Gupta, J. agrees with the conclusions of Lokur, J., his reasoning on this issue differs materially but this remains unacknowledged and thus makes his concurrence carry a different weight. I highlighted one conflict above. The other comes when the judges compare the views in Rajeev Chaudhary and Bhupinder Singh. Lokur, J. suggests there is a conflict and specifically agrees with the view in Rajeev Chaudhary while Gupta, J. (correctly) finds there was no conflict, as the remarks in Bhupinder Singh were at best obiter

The deeper problem with the majority opinion is the lack of intellectual rigour on display.  Any law student will tell you that Pant, J. did not uphold the canon of statutory interpretation which says that the views of a large legislature cannot be attributed to the voice of a single participant in the house. Further, the minority opinion does not offer an explanations on why the words were dropped either. But the mere presence of those speeches in the minority opinion severely lessens the credit of the majority opinion. Not only does it convincingly show there is an ambiguity, but also shows us that Lokur J. did not fully engage with the legislative history before speculating on legislative intent. That doubt is damaging and one then can't stop wondering what else was missed out by the majority.

While I firmly believe that the possibility of pre-trial detention must be narrowly construed, the sad truth is that currently Section 167 Cr.P.C. does not allow this. The majority does not engage with the most important point made by the minority - that no offences in the IPC ever prescribed a minimum of ten years imprisonment. Which begs the question that why would the legislature use a language to describe the kind of offences that our penal code has never seen. Following the majority logic leaves out some serious offences - Attempt to Murder punishable under Section 307 IPC for instance - and therefore betrays the logic of the sixty - ninety day distinction being tied to the gravity of the offence which both majority opinions adopt.

Given that there was no prevailing conflict between the views of the Supreme Court in Rajeev Chaudhary and Bhupinder Singh, we can say that the Court has simply upheld the view in Rajeev Chaudhary by a 2-1 split in the present case. Sadly for us, the majority opinions lack conviction on the issue that was under discussion. Both Lokur, J. and Pant, J. hinted at the language of Section 167 Cr.P.C. being problematic, and it is apparent that the problems will persist without clear legislative intervention. Given how this opinion has simply upheld status quo - for Rajeev Chaudhary had already curbed the powers of investigative agencies on this front, it is unlikely that this will happen. Teleologically, perhaps that is for the best after all.

On another note, the majority opinion does warrant applause for liberalising the regime of default bail by placing a greater onus on the courts. It has long been acknowledged how accused persons continue to languish in custody despite being entitled to default bail, only because of poor legal assistance. While I have often seen trial courts apprising accused person's of their rights, a strict honouring of the Court's observations in Rakesh Kumar Paul can certainly further the cause. 

*The Petitioner had also made an argument trying to construe the maximum period of punishment as seven years using the Lokpal and Lokayuktas Act 2013, which was roundly rejected by all courts hearing the case. 

Wednesday, July 26, 2017

Guest Post - Same Day Sentencing in Death Penalty Jurisprudence

(I am happy to host this Guest Post by Ms Nivedita Mukhija, an advocate practising in New Delhi)

235. Judgment of acquittal or conviction.(1) After hearing arguments and points of law (if any), the Judge shall give a judgment in the case.(2) If the accused is convicted, the Judge shall, unless he proceeds in accordance with the provisions of Section 360, hear the accused on the question of sentence, and then pass sentence on him according to law.

This provision is an explicit recognition of the idea that the relevant considerations for sentencing are different from that for conviction/acquittal, and thus, sentencing should be treated as a separate phase in a trial. It has assumed considerable significance in death penalty jurisprudence, where lower courts often downplay the importance of hearing, in detail, all aggravating and mitigating factors awarding the death penalty, erroneously assuming that these factors have already been sketched out during the trial. Bachan Singh v. State of Punjab [(1980) 2 SCC 684] called Section 235(2) as something enabling a “bifurcated trial”, wherein the accused submits evidence possibly without bearing on the crime, but which may still be linked to the “special reasons” to be provided while awarding a death penalty as per Section 354(3) of the CrPC

Conflicting Lines of Judicial Opinion
The question, however, is whether a separate phase implies that a different date be set for hearing on sentencing, or, whether sentencing can be conducted right after the determination of guilt has been concluded. There seem to be two lines of cases emanating from the Supreme Court. One states that Section 235(2) cannot be done justice to unless a different date is set, to afford ample time for the accused to prepare a case on sentencing. The second holds that Section 235(2) merely implies that the accused must be effectively heard, which does not require setting a different date for hearing. I argue here that sentencing should always be deferred to a later date in death penalty cases to grant an effective hearing to the accused. I also contend that if this has not been done, the appellate court should remand the case, as a fact-intensive exercise like sentencing should be done by a court of first instance, which also rightly extends the right to appeal against that decision to an accused. While doing so, the appellate court must ensure that the case is disposed off in a timely manner by the lower court to maintain judicial efficacy and reduce the suffering of the accused through further delays.

Set One: Mandatory Separation of Trial and Sentencing Phases

With respect to the first line of cases, the most prominent example is perhaps Allauddin Mian and Ors. v. State of Bihar [(1989) 3 SCC 5], where three judges of the Supreme Court recognised the importance of giving the accused an effective opportunity of placing all relevant mitigating factors, antecedents and background information, and other extenuating factors on record. The bench held in context of Section 235(2) that as “… a general rule the trial courts should after recording the conviction adjourn the matter to a future date ...”. It was believed that the accused must be allowed to absorb and overcome the shock of conviction before being heard on sentence. Since this was not done in that case, the Supreme Court went on to convert the death sentence into life imprisonment, finding insufficient material on sentencing to warrant awarding the death penalty. This decision was followed in Malkiat Singh v. State of Punjab [(1991) 4 SCC 341], where another three-judge bench noted that usually a failure to give sufficient time to adduce evidence on sentencing would lead to remanding the case. However, in that case, the Court held that given the long period of incarceration (it had now been six years since the accused was first convicted), it was more expedient to convert the death penalty into life imprisonment.

Set Two: Same Day Sentencing

The second line of cases is concerned more with hearing the accused sufficiently on sentence, and holds this can be done on the same day as well. On the point of remanding the case, these cases suggest that sentencing hearings need not involve remands. These views take their root in yet-another three judge bench decision in Dagdu v. State of Maharashtra [(1977) 3 SCC 68], wherein the Court held that a failure to be heard on sentence would not automatically entail a remand, and the appellate court could hear the accused on sentence at the later stage as well. The Court called remand the exception, in the interests of judicial efficacy.

A slew of recent decisions has come in support of same-day sentencing, to which I now return. In 2016, yet another three judge bench held in B. A. Umesh v. State of Karnataka [2016 (9) SCALE 600] that there was no mandate under Section 235(2) to fix a separate date for hearing on sentence, and the Court was required to determine whether an effective hearing was granted after looking the facts and circumstances of a given case instead. Merely because no separate date was fixed was no ground to allow the review petition, the Court held in the death penalty matter. The Court also noted that the parties had not objected to not being given a separate date of hearing in the lower courts, and since then several courts including the Supreme Court had heard the parties on the question of sentence in sufficient detail. This was reiterated by three judges in Vasanta Sampath Dupare v. State of Maharashtra [2017 (5) SCALE 724]. 

Neither B. A. Umesh nor Vasanta Sampath Dupare explain the inconsistencies arising by comparing their logic with the Allauddin Mian line of cases. Instead, they simply rely on Dagdu. Recently, in Mukesh and Ors. v. State of NCT of Delhi and Ors. [2017 (3) SCALE 356], the Court over-simplified the conundrum by stating that these decisions reflect two modes that can be afforded to the accused to present a case: either to remand it before the High Court, or be heard on sentence before the Supreme Court itself [the Court believed, in that case, that “regard being had to the nature of the case”, the second mode was preferable].

Making the Case for 'Bifurcation of Trial' 
It is submitted that this second line of decisions disregards the ‘bifurcation-of-trial’ approach adopted by Bachan Singh. These decisions do not explain why the exercise of sentencing, which is governed by different considerations, and may involve new evidence and witnesses requiring different preparation altogether, should be undertaken on the same day. While many of these decisions rely on rhetoric of ensuring an accused has been granted ‘effective hearing’, the Courts must also recognise that adequate time to prepare on a different limb of a trial is a pre-requisite to effective hearing.

The recent decisions I highlighted are also problematic on another level: they imply that the failure of the lower court to hear effectively on sentencing is remedied when an appellate court hears the same. However, our judicial system clarifies that the lower Courts are the courts of evidence with such exercises being exceptional during appellate hearings. Thus, original evidence must as per rule first go through the lower Courts, which are better equipped to deal with the same. Further, under Article 131 and the Supreme Court Rules, 2013, original jurisdiction is only limited to civil suits between the union and the states, or inter se states. For a factually-intensive exercise such as sentencing, often requiring cross-examination, the Supreme Court is not the ideal place for a first (and often final) determination. (On this, see, http://lawandotherthings.com/2017/05/death-penalty-sentencing-the-supreme-court-as-the-first-and-final-arbiter-of-facts/#_ftnref7).

Another important reason to question the second approach is that an automatic bypass in such cases would deny an accused a right to appeal against the ruling in the sentencing stage before the Supreme Court (see http://indianexpress.com/article/opinion/columns/death-penalty-in-india-supreme-court-cases-law-mitigating-evidence-4551607/). This leads to there being no chance of re-appreciation of evidence on sentencing in case of an error in judicial reasoning. In a death penalty case, where the stakes involved in the sentencing process are highest, we must ensure that there is the protective layer of multiple levels of Courts having the option of hearing on sentence.

Conclusions and Some Ways Out of the Bind
Since B. A. Umesh states that same day sentencing was not an issue there because it was not raised before lower courts, defence lawyers should ensure that the need for a separate date for sentencing is adequately highlighted before the lower court itself. Defence lawyers can perhaps also use the caveat given in Malkiat Singh, and argue that prima facie errors in sentencing coupled with lengthy delay for someone on death row should result in converting those cases to life imprisonment, instead of conducting a mini-trial for sentencing yet again. The Supreme Court's current treatment of 'delay' as a factor requires more nuance but. Consider Yakub Abdul Razak Memon v. State of Maharashtra [(2013) 13 SCC 1], where the Court concluded that sentencing hearings must involve remand unless it is likely to cause delay. But, I ask, given the pervading backlog in case-disposal, when would remand not involve some delay? Moreover, the lack of nuance uses delay to take away a right of appeal at the cost of judicial efficacy. Bland arguments suggest delay only lengthens the suffering of those on death row. Of course delay is bad, but it does not entail cutting corners to arrive at the end. This trade-off is unconscionable. Thus, I conclude with suggesting it is time that the Supreme Court clarifies the inconsistency between the two lines of cases by stating the true scope and import of Section 235(2), and while doing so pays heed to the adverse effects of not remanding a case in a case where the judicial power to take a life is at stake.