Amnesty Lied On Caste Violence And It Lies Again

Amnesty Lied On Caste Violence And It Lies AgainDalits stage a protest against a Supreme Court order which allegedly dilutes the SC/ST (Prevention of Atrocities) Act in Lucknow, India. (Subhankar Chakraborty/Hindustan Times via GettyImages)
  • Amnesty International’s theories of caste violence are based on manipulations of data, spurious allegations, straw man arguments and pompous expressions of personal incredulity - reflecting a complete disregard for data and facts.

In my previous article, I had called into question the claims about excessive violence against Scheduled Castes (SCs) and Scheduled Tribes (STs) that were made by groups such as Amnesty International. Amnesty International India’s Leah Verghese responded to my article trying to defend Amnesty and calling my criticisms a “criminal mistake”. But rather than bringing forth any credible justifications, Verghese repeated some of Amnesty’s earlier assertions and brought in more errors, manipulations of data and baseless allegations. Here is why her defence of Amnesty exposes the group’s bias and shoddy work even further.

Amnesty Indeed Fudged Conviction Rates

In my previous article, we had discussed how Amnesty International had made an erroneous assertion that the conviction rate for the crimes under the SC/ST Prevention of Atrocities (POA) Act is only 15 per cent. The alleged lower conviction rates under the POA Act have often been presented as due to the failure of the police and the judiciary ostensibly on account of some “deeply entrenched prejudice” against SCs and STs and as a reflection of a general apathy towards the gruesome violence against the communities.

Verghese responded to my article saying that I was wrong in claiming that Amnesty had fudged data and that the figure of 15.4% as the conviction rate for crimes under the SC/ST Prevention of Atrocities Act “is cited by the NCRB 2016 report itself on pages XXV and 571.”

So now let us fact-check this claim. It is indeed true that the NCRB’s CII 2016 report mentions the conviction rate of 15.4% on page 571. But what crimes is this crime-heading referring to? The explanation has been provided in the ‘Executive Summary’ (page 5 of the NCRB report), which Amnesty seems to have either missed or deliberately glossed over. And here is what it says:

“Includes cases registered under SC/ST (PoA) Act,1988 only(SCs+STs). Figure does not include cases registered under IPC along with SC/ST (PoA) Act as those cases have already been counted under IPC Crimes. Details are available in the Chapter 7 - 'Crimes against SCs/STs'.” (emphasis mine)

So these crimes, for which the conviction rate is mentioned as 15.4 per cent refer to only 4,595 out of the 18,109 (or about 25 per cent of) cases disposed off by the courts under the POA Act in 2016. As the above clarification excerpted from the report suggests, the details for all the (18,109) cases disposed off by the courts under the POA Act for 2016 are available in Chapter 7 of the report. And the correct conviction rates (mentioned on page 303 and 347 of NCRB CII 2016 Report) are 25.8 per cent and 20.8 per cent for crimes against SCs and STs respectively (or 25 per cent for SC+ST) and not 15 per cent (or 15.4 per cent) as claimed by Amnesty.

So, the question is this: did Amnesty and Verghese miss reading this clarification by NCRB on what the number meant? Or are they deliberately trying to gloss over, while they keep their fingers crossed, hoping nobody finds out? Nobody could, perhaps, have faulted had it been a mere calculation error or an oversight which can be corrected at least when the error is pointed out. But that does not seem to be the case here as Amnesty is defending the error, perhaps because its whole narrative is dependent upon it?

Verghese writes, “Sashittal is welcome to disbelieve this statistic [of 15% or 15.4%], or offer an alternative analysis”. But sadly for Amnesty, these are not a matter of subjective opinions or beliefs and there are no ‘alternative facts’. The report unambiguously mentions what this number is and it is clear that Amnesty mined a number and misrepresented it to build an entire narrative around it.

So what are the crimes for which the conviction rate is mentioned as 15.4 per cent in the NCRB report? This may need a bit of an explanation. The POA Act is extensively used along with the provisions within the Indian Penal Code (IPC). For example, for a murder reported under the POA Act, the charges would invariably be registered under the sections of IPC along with the clauses of the POA Act. This is true for the most violent crimes under the POA Act like murder, rape, etc., and even less violent crimes like hurt, intimidation and trespass. Consider the accompanying Venn diagram that details the categories of cases within the POA Act.

Figure: POA cases disposed off by courts in 2016. (Only for illustration not to scale)*IPC provisions include the ones for violent crimes and crimes like hurt, intimidation, trespass etc.
Figure: POA cases disposed off by courts in 2016. (Only for illustration not to scale)*IPC provisions include the ones for violent crimes and crimes like hurt, intimidation, trespass etc.

In the year 2016, about 75 per cent of all cases disposed off by the courts under the POA Act were charged along with the provisions of the IPC. The intersection of the two circles in the Venn diagram represents these 75 per cent of the cases that include the most violent crimes and has a conviction rate of 28.3 per cent. The remaining about 25 per cent part of the blue circle on the right side of the diagram represents cases that are solely charged under the provisions of the POA Act and hence it excludes the most violent crimes. This 25 per cent part of POA cases has a conviction rate of 15.4 per cent and includes the less severe crimes. The conviction rates for all the cases under the POA Act put together, represented by the entire blue circle is 25 per cent.

From a data analysis standpoint, the IPC data on violent crimes is significant because there is comparable data available for the individual crime-headings like murder, rape etc against SCs, STs and the overall population. Therefore it is easy to compare and see how the conviction rates match for the same crimes in the populations considered. As I have shown in my earlier article, the conviction rates for violent crimes against victims belonging to these three categories are about 29 per cent for SCs, 22 per cent for STs and 24 per cent for the overall population i.e. they are not disproportionately low for SCs and STs as implied by Amnesty.

Verghese, as part of what seems to be the strategy of ‘argument in the alternative’ or second line of polemical defence, tries implying that some gruesome crimes against SCs and STs like “dispossessing them of their lands”, “making them eat inedible substances” etc have low conviction rates. She calls them “specific targeted crimes” as against the other violent crimes charged under the IPC to imply that these are the crimes that belong to the 15.4 per cent conviction rate category within the POA crimes. But this argument falls through as soon as one looks at the grounds on which it tries to stand. Firstly there is no specific data in the NCRB reports on these individual clauses of crimes to conclude that their conviction rates are low. Secondly, there is no evidence to suggest that these crimes are charged only under POA and not IPC also. In fact, there is some evidence to the contrary: there are instances from court cases that show that many of the crimes quoted by Verghese are registered along with the sections under the IPC for trespass, grievous hurt etc.

Shoddy Allegations Against The Supreme Court

I had questioned Amnesty’s allegations of dubiousness and cherry-picking against the Supreme Court of India and had asked as to what data was the Apex Court deliberately overlooking. Ironically, Verghese has come back having picked some more erroneous numbers and trying to draw a false equivalence. She claims that “The Supreme Court in its judgment had relied on NCRB data from 2015 which showed that around 9.6% of cases involving crimes against SCs and STs were found to be ‘false’ by the police” and laments that crimes of kidnapping and forgery too have a high incidence of false cases and that there is no similar outrage against the misuse of these crimes. Let us consider both these claims.

Firstly, the figure 9.6% as the percentage of false cases is completely false. It is either weaved out of thin air or arrived at by some erroneous calculations by Amnesty. There is no mention of it in the Supreme Court’s judgment either. Where the cases in 2015 data are discussed, the judgment notes from submissions that “in almost 15-16% cases, the competent police authorities had filed closure reports.” The total percentages for all closure reports are much higher but these numbers are accurate for closure reports filed for cases found to be false after investigations. Below is the data from the 2015 report of NCRB that gives the values of 15.32 per cent and 16.82 per cent for percentages of false cases for total crimes against SCs and STs.

Cases Found To Be False After Police Investigation

(From NCRB’s Crime In India Report 2015, Chapter 7, ‘Crime Against Persons Belonging to SCs / STs’, sections 7.3 and 7.9, Col.4 has been calculated based on the formula mentioned).
(From NCRB’s Crime In India Report 2015, Chapter 7, ‘Crime Against Persons Belonging to SCs / STs’, sections 7.3 and 7.9, Col.4 has been calculated based on the formula mentioned).

So, Amnesty seems to have not only misreported this number but has also made a false claim that the Supreme Court relied on it in its judgment. So much for Amnesty’s allegations of dubiousness against the Supreme Court.

Apart from the erroneous numbers, Verghese’s lament of selective outrage is also based on a false equivalence. The Supreme Court was looking into concerns about the misuse mainly with respect to the provisions within the POA Act that result in arbitrary arrests and denial of anticipatory bail even for very minor offenses that are based merely on accusations. The Supreme Court observed that these provisions infringe upon the right to personal liberty guaranteed under Article 21 of the Constitution. The laws on kidnapping and forgery do not have such draconian provisions.

For example, the section 438 of Code of Criminal Procedure (Cr.P.C), otherwise allows courts to grant anticipatory bail even to a person accused of most heinous violent crimes like murder, rape, dacoity etc. But this facility was barred even for the most minor crimes under the POA Act even if the court was to find that the accusations were prima facie false or the arrests were unwarranted. The Supreme Court ruled against such a restriction on the grant of anticipatory bail if the accused is able to convince the court that the case is false or malafide. Section 41 of Cr.P.C requires 'credible' information’ or ‘reasonable suspicion’ about the complicity of the accused before making an arrest and the Court directed that its principle and procedures be adhered to for the crimes under POA Act also. The Court was not against arrests where justification is available like to facilitate investigation, prevent further crime or violence, to prevent the accused from absconding or in grave offenses even to inspire the confidence of the victim. But it was against the practice of arbitrary arrests that are not accompanied by a reasonable process and hence place unrestricted power in the hands of the police.

In a field survey conducted by the Dr B R Ambedkar Research Institute in 2008, and funded by the Karnataka Government, to evaluate the compensation given to the victims under the POA Act, a significant portion of the accused of the cases reported, came from economically and socially very backward communities classified under the ‘Other Backward Class’ (OBC). Many of the offenses were minor and reported to have been from trivial altercations. There are also instances where cases against people from Scheduled Castes themselves have been registered based on allegations and counter allegations of atrocities. It is not as if the accused are always wealthy or powerful, or the crimes are always violent in nature, as the simplistic propaganda on caste violence makes it out to be.

Amnesty International’s opposition to safeguards against arbitrary arrests, flies in the face of Article 9 of ‘International Charters Universal Declaration of Human Rights’ that states that no one shall be subjected to arbitrary arrests. It also runs contrary to its own stand as stated on its website that it is against detention “with no legitimate reason or without legal process”. Is that lip service then? Amnesty is seeking the dilution of the very safeguards in India that it claims to advocate the world over.

Baseless Allegations And Attacks

What Amnesty’s allegations lack in substance or statistical judgment they try to make up through spurious allegations, building straw man arguments and by pompous expressions of personal incredulity like “disingenuous” ,“absurd” as if they are all substitute to facts.

For example, I had said in my previous article that the claims of under-reporting of crimes cannot explain away the lower rates of violent crimes against SCs and STs in the crime data. Verghese twists this by claiming that I had meant the under-reporting of crimes against SCs and STs itself is unfounded and goes at length to refute this substituted position. In fact, my article clearly said that crimes are “known to be under-reported in all populations.”

Also, my previous article had shown from the NCRB data for the period between 1995 and 2013 that the rates for violent crimes against SCs and STs were far lesser than those for the overall population. Verghese tries to discredit the data by making a baseless and false charge that the data fallaciously compares rates of crimes “even though the NCRB had calculated the crime rate” “using the population of India, rather than the SC/ST population.” - basically implying that the crime rates presented in my previous article are based on wrong denominators.

This is a completely spurious allegation. Here is the background. The NCRB yearly reports till 2011 mention the crime rates against SCs and STs using the population of India as the denominator. This was changed in 2012 when NCRB started calculating the rates based on SC and ST populations as the denominators respectively. I had come across this change when I had first studied the crime data in NCRB’s yearly reports. I had also pointed out the errors by a portal named Indiaspend which had failed to detect this change in the denominators and claimed that crimes against SCs and STs had very sharply increased. Had I been blind to this, the data in my article would have had the problems that Verghese alleges. But sadly for Verghese and Amnesty, I did not make that silly mistake, least because I was the one who pointed out that discrepancy. If Amnesty had checked the data carefully and verified before making this false allegation, it would have come to see that the rates presented in the previous article were re-calculated from the raw data with the respective estimated individual populations as the denominators for all the years. In fact, if Amnesty had closely looked at the NCRB statistics for any of the years (before or after 2011), it would have realized that the observation about lower rates of violent crimes against SCs and STs is inescapable from the data and that it cannot wish it away through spurious allegations, word games or manipulations of data.

Again nice try Amnesty! Would it have been better had you graciously accepted the errors pointed out and made amends rather than entangling yourself into more fallacious arguments and manipulations of data in their support?

As I had pointed out in my earlier article, the bane that afflicts the discourse on caste today is the complete disdain for data and facts. It is as if data can be culled, suppressed, and trampled upon at will. And this seems only to be reaffirmed by Amnesty’s hasty retort to even the mathematical and logical errors that were pointed out in its discourse.


Latest Articles

    Artboard 4Created with Sketch.