What do police statistics hide (and what do they reveal) about crime in India?

Sometime in 2013, the Mumbai police decided it had to do something about Meow Meow. Mephedrone, known on the street as Meow Meow, was cheap and widely available, and the police were having a hard time keeping up. Possession of the drug had not been made illegal under the stringent Narcotic Drugs and Psychotropic Substances (NDPS) Act, despite the Mumbai police making multiple representations to various Central agencies for the substance to be banned.

So, in 2013 then city police commissioner Rakesh Maria directed his force to start arresting pedlars of mephedrone under the next-best option: Section 328 of the IPC, “causing hurt using a poisonous substance”. Quite apart from the Mumbai police’s inability to correctly collect, label and present evidence – as borne out by the observations of district court judges in every single case – the basic problem was clearly the application of the wrong law. “You need to show intention to cause hurt. How hard is that to understand and how many times did I need to tell the police that? No effect,” one judge who had seen dozens of such cases said.

The Mumbai police knew this. “See, the hands of the police were tied. So they had to use whatever tool was available,” Shivdeep Lande, deputy inspector general of police from the anti-narcotics cell of the Mumbai police, who arrived in Mumbai with the nickname “Bihar’s Chulbul Pandey” for his upright conduct and penchant for tight-fitted short-sleeved shirts in his previous posting, told me in an interview.

An assistant commissioner of police was more explicit: “We wanted to throw them in and create some fear. It didn’t really matter that they were getting acquitted; anyway courts always find fault with our evidence.”

For a conviction under IPC 328, a court needs to see proof of intention to cause hurt, and the law is typically invoked in cases of food poisoning or the knowing adulteration of food. Because of the consensual nature of most drug transactions, the intention to cause hurt is by definition absent. Between June 2015 and November 2016, Mumbai’s sessions courts pronounced judgments in at least hundred such cases. Not a single one ended in a conviction. All the charge sheets from that time followed essentially the same script.

The complainant’s brother, son or friend had become increasingly ill. On enquiry, the complainant found that the victim had been forced to consume a drug by the accused. The police swooped in to arrest the accused, who was found with a white powder on him. This powder was later “discovered” to be mephedrone.

Except, none of this ever stood in court. In no more than three cases were the police even able to prove possession of the drug. Time and time again, the basic evidence to show that the drug was found on the accused was not produced, and in dozens of cases, the judge accepted that the panchas (eyewitness to the recovery) were “habitual panchas” set up by the police. And then there was the matter of proving intention to cause hurt, which the police made virtually no attempt to do, in the cases I investigated. The result: an acquittal rate of 100 per cent.

In all, nearly 150 young people spent more than a year in jail on charges that the police knew early on were bad in law, but persisted with nonetheless. Yet, no one – not senior police officials, prosecutors or annoyed judges – stopped what was going on. Ultimately, the inclusion of mephedrone in the NDPS in February 2015 brought this to an end.

Of the seven acquitted men who spoke to me, two admitted to petty peddling and drug use, but disputed all the stated facts of the police case. The five others said they had nothing to do with drugs. IPC 328 is a non-bailable offence; the accused in the cases I saw spent between one year and twenty months in custody. An analysis of their names shows that 119 of the 148 acquitted were Muslim.

For anyone looking at crime statistics of the time, as I was for what began as unrelated research on sexual assault FIRs in Mumbai, it would have seemed as if the city was awash with incidents of poisoning, mostly committed by Muslims, and the courts were failing victims. What would have remained hidden between the lines is that forming any assessment of India from police reports or from media reports of them is an exercise in futility – or worse, in deception.

When more crime is a good thing

Not every journalist in the country can investigate every crime statistic for six months before she publishes it, and a reading public cannot be expected to be doing their research before forming views on the statistics they’re reading. But understanding the universe that produces these statistics is essential, and the only way to derive greater use out of them. Indian crime statistics begin from a point of significant under-reporting, and not just for sexual crime.

The country’s officially recorded crime rates are lower than the global average, substantially lower than developed countries, and even low by middle-income country standards. In 2019, the highest crime rates in India were in the two most developed states – Delhi and Kerala.

The rates of recorded crime in most other Indian cities and states, quite frankly, defy belief. Uttar Pradesh with over 200 million people recorded just over 10,000 cases of “grievous hurt” in 2014, while London with under 9 million people recorded over 70,000 cases of “assault with injury offences”, according to its police statistics for 2014-15.

Some crimes suffer greater under-reporting than others. In [a] Rajasthan survey, self-reported crime rates were substantially higher than police-recorded rates for property crimes in particular: survey theft rates were over 800 per cent higher than registered rates, and robbery rates were over 1,100 per cent times higher than in police records.

The differential decreased for violent crimes: rape or molestation surveyed rates were just 43 per cent higher than the rates reported to the police, and assault was only 11 per cent higher in the survey. Motorcycle thefts, where a police complaint becomes essential for insurance claims, were most likely to be recorded by the police.

It isn’t just some crimes that are more likely to be reported than others; some places are also more likely to report crimes than others, with reasons that have relatively little to do with the crime itself. This is likely less of a problem for crimes like burglary or theft, where the incentives to report the crime are high, and the incentives to not record them low. But for other types of crimes, there are indications both from interviews with police officials and from some data, that the states with higher reported crime might actually be the ones doing a better job of ensuring full reporting, rather than being the ones that are the most unsafe, particularly for some types of crimes.

Comparing the rates of sexual crime reported in household surveys with police records for the same year shows that Delhi had consistently higher reporting and lower incidence of actual violence than other states, while Bihar had low rates of reporting and high actual incidence.

Among larger states, both actual violence and the extent of under-reporting were higher in northern states with poor gender indicators than in southern states with better gender indicators. Not that you would know this from the media reporting. “Uttar Pradesh safer for women than many big states: NCRB report,” read the headline in The Times of India, India’s most widely read English newspaper, the day the 2019 NCRB statistics were released.

The police establishment knows this, and also knows of the perverse incentives it creates for the police of a state to suppress numbers. “The unreliability of crime statistics in India is well known...Whenever a genuine effort was made to register all crime...the figures showed such fantastic jumps as were impossible with any normal increase in one year,” the National Police Commission said, as far back as 1979.

Police statistics as a piece of the puzzle

India needs more and better qualitative research into crime than currently exists; unusually high numbers, sudden declines, high conviction or acquittal rates, all need more nuanced investigation before being declared in the media as crime waves or the failure of the police or judiciary. On their own, India’s crime statistics might be problematic, but combined with a deeper understanding of the country itself, they offer an insight into the churning within.

They need triangulation with other data, including indicators of levels of development, female empowerment and state capacity, as well as independent household surveys. For citizens to get a true picture of crime and safety in their district or city, we will have to take that next step – providing them a framework with which to understand crime statistics. Perhaps then they can use this data to create real democratic pressure for the right to live their lives without fear.

Excerpted with permission from Whole Numbers and Half Truths: What Data Can and Cannot Tell Us About Modern India, Rukmini S, Context.