Analysing crime data by implying that it is bad to have a high number of recorded crimes is not only wrong, it could lead harmful police practices.
Representative image. Credit: Reuters/Danish Siddiqui
At the end of August, ‘Crime in India 2015‘ was released by National Crime Records Bureau (NCRB). The print and electronic media has since been analysing the data presented in the report ad nauseum. Much of such analysis is facile, declaring which state has come first and which has come second in a particular crime, or where and which crime has increased or declined compared to the previous year. Implied in that (or sometimes explicit) is praise for states with low crime recording and criticism for those with high crime recording. Some writings analyse data of crime rates (crimes per 100,000 population) rather than crime recordings.
Not only is this wrong way of analysing this data, such analysis could encourage wrong or pernicious police practices.
To understand crime statistics, we must distinguish between three different concepts – incidence of crime, reporting of crime and recording of crime. It is well known that there is data loss between incidence and reporting as well as between reporting and recording of crimes.
Not all victims report crimes to the police. Sometimes, it is because the crime is trivial (say theft of a wallet) and the effort in pursuing the matter would be disproportionate. Sometimes, the victim fears reprisal. A large number of offences (for instance rape, stalking, harassment or molestation) against women in our country are not reported because the victim feels that reporting such an offence may stigmatise her or bring unwarranted public attention to her. Many hesitate to report crimes fearing a loss of reputation or due to lack of confidence in the capacity of law enforcement agencies.
Similarly, there is difference between crimes reported and crimes recorded. A large (probably largest) number of complaints/grievances against the police in our country is about non-registration of an FIR.
Big increase
Let us take the example of the Delhi police. Total IPC offences registered by the Delhi police was more or less stable from 2001 to 2012 at around 53,000 (lowest 44,404 in 2003, highest 56,065 in 2005). This is very surprising in view of the increase in population and other relevant factors. Sometime in 2013, leadership of the Delhi police seems to have decided to make registration easier. Total IPC crimes recorded raced from 54,287 in 2012 to 80,184 in 2013 to 155,654 in 2014 and 191,377 in 2015 (an increase of over 250% in three years).
The trend of crime being stable from 2001 to 2012 and then the sudden big jump from 2012 to 2015 in Delhi is seen in most major crime heads, including offences against women. But crimes of murder or attempt to murder have been stable (or growing at a moderate rate), indicating that the mismatch between reporting and recording in murder and attempt to murder is probably minimal. Robbery has seen most substantial jump. Recorded robbery was around 550 in 2001-2012 (lowest 441 in 2003, highest 624 in 2001). This saw a dramatic jump from 608 in 2012 to 1,245 in 2013, to 6,464 in 2014 and to 7,407 in 2015 (increase of over 1,350% in three years).
So how should we discuss crime? Should the Delhi police be lauded for faithfully recording robbery or ridiculed and reprimanded for the increase in robbery and declared the ‘robbery capital of India’? The example of Delhi may be more dramatic than others, but it is not the only one of its kind.
Big decline
On the other hand, there are statistics showing a steep decline in crime. S.I. Wilkinson, an American academic, relying on some research done by N.S. Saxena, UP police chief in the 1970s, found that “In Uttar Pradesh for instance there were 335.5 Indian Penal Code (IPC) crimes per 100,000 (persons) in the decade 1905-14 but only 86.9 crimes in 1955-60 and only 73.2 in 2004”. (The crime rate for IPC offences in UP in 2015 is 112.1 per 100,000 persons, up from an unbelievably low figure of 2004 but still less than half of national crime rate of 234.2.) No wonder Wilkinson concluded, “Since Independence, there has been professional incentive for police officers to minimize crime figures in order to make their performance and that of their political masters look better than it is.”
So would it be proper to conclude that UP was an oasis of peace and tranquility in 2004 and has deteriorated since then? An honest answer would be that we don’t know.
Difference between two government agencies
Another interesting example is of cyber crime. A total of 11,592 cases were registered by the police in 2015, compared to 9,622 in 2014, according to the NCRB. Some in the media are concerned about the increase in cyber crime. Now CERT-In (Indian Computer Emergency Response Team), the government agency “responsible for responding to computer security incidents”, in its annual report mentions handling 26,244 cases of websites defacements alone in 2015, compared to 25,037 in 2014. CERT-In handled many more incidents, some of which may not be categorised as cognisable crime. A large number in the annual report are categorised as “others”. It is not possible to do a meaningful comparison in one article. But the difference in data between the NCRB and CERT-In by such a margin substantiates the argument that crime recorded is not the same as incidence of crime.
The National Police Commission, writing in 1981, said, “Since the crime situation is discussed every year … primarily on the basis of crime statistics … State Governments and senior police leaders frequently connive at underreporting of cases”. In fact, some scholars have argued that police crime data in India is nothing more than indication of recording practices of the police agencies.
The problem of loss of data between the incidence of a crime and reporting and reporting to recording is not unique to India. It has been discovered in nearly all jurisdictions where research to this effect has been conducted. For example, in a survey conducted by Statistics Canada in 2009, it was found that only 31% of crime is reported to police. A survey in England and Wales found that only one in four crime gets reported.
However, there are many who would argue with evidence that the magnitude of data loss between reporting and recording is much higher in India. We should thus be cautious in ranking the states in crime on the basis of recorded data (FIRs) alone.
Are we then helpless in ascertaining whether crime is increasing or on the decline? Is there no way to do a comparative analysis of crime between states? The answer, fortunately, is no.
Crime victimisation surveys
To estimate the data loss and get more information, most advanced democracies have found an identical solution – crime victimisation surveys. A sample of the population is selected by using sophisticated sampling techniques. A reputed agency conducts the survey by asking citizens about their experience as crime victims. The survey is done at regular intervals. The crucial value of the survey is its ability to find out about crimes which do not get reported to or recorded by the police.
Such a survey is conducted every year in the US and the UK, every five years in Canada and every five or six years in Australia. The survey in all these countries is conducted by government agencies or under the supervision of government agencies. The specific crimes covered vary from country to country, as does the name of the survey.
Government agencies, researchers and the public thus have two kinds of data, one recorded by the police and the other collected by the victimisation survey. With this, a holistic picture on crime in the region emerges.
For crime victimisation surveys to have meaning, these must be conducted at regular intervals and cover the entire country. This is possible in India only with government support. Moreover, such surveys in other countries have been conducted or sponsored by government agencies only. Governments in India thus have their task cut out for them.
Abhay is an IPS officer. The views expressed are personal.