Wednesday, May 11, 2016

Delays at the police

by Renuka Sane and Neha Sinha.

The Chief Justice of India's recent emotional outburst over judicial delays has reignited the discussion on India's criminal justice system. The focus of much of this discussion has been on judicial delays. Policy reform discussions range from increasing the number of judges, and improving the basic infrastructure in courts, to deeper reforms in the form of separating out the administrative function from the judicial function. These discussions are important.

Our ultimate objective, however, is to create deterrence against crime. The criminal justice system is the interlocking institutional infrastructure of laws, police, courts, public prosecutors and prisons. We must, therefore, look at all delays from the date of the crime to the date of the conviction. Judicial delays are one part of this. Delays at the police are an equally important problem.

After a crime takes place, the very ability to conduct an investigation and identify the criminal degrades with time. Witnesses forget, evidence is lost. Delays in investigation drive up the probability of a failed investigation. The urgency of reducing delays here is, in some sense, even greater than the problem of delays at courts.

What do we wish to measure?

In an ideal world, we should see the delays associated with all the steps from the date of the crime to the date of the conviction. The units for this measurement are in days. Fine-grained data is required, at the level of a sample of cases or the population of cases, in order to make statements such as `On average, when a complaint is filed under IPC S.300 (murder), there is a conviction with $x$% probability, and this takes place with a lag of $y$ days.' Two useful metrics are:

  1. The age of the case: measured from the date of the FIR registration.

  2. The amount of time devoted to each process: measured as the start and end date for each step from crime to conviction. For ongoing cases, many elements of this pipeline would be NA, and one would be censored.

Such measurement will help understand where delays occur, what kinds of cases take more time, how lags in the process vary based on the crime involved, and how police resourcing can change various elements of the lag. Answers to such questions can be the basis for improved management of the police.

The current state of measurement on police pendency

Data on reported crime and response by the police is published by the National Crime Records Bureau (NCRB). Information from individual police stations registered as the First Information Report (FIR) is reported to the District Crime Records Bureaus, flowing up to the State Crime Record Bureau which is finally consolidated by the NCRB. The consolidated data processed by the NCRB is published in their annual publication "Crime in India" providing consolidated numbers for crime registered and status of police as well as judiciary responses at the state and national level. Within this, "Disposal of IPC cases by Police" for the states and union territories provide data on cases reported to the police, pending cases, and cases charge sheeted during the year.

Police pendency in year $t$ in the NCRB is defined as the number of cases pending investigation at the end of year $t$ divided by the sum of cases pending investigation from the previous year and cases reported during the year. This pendency rate was 28.10 per cent at an all-India level in 2014 (Table 4.1). As the previous discussion suggests, this is a pretty bad way to measure pendency at the police.

The North Eastern states of Manipur, Meghalaya, Arunachal Pradesh and Assam reported highest pendency rates amongst all states at 84.75 per cent, 69.91 per cent, 57.91 per cent and 55.29 per cent respectively. On the lowest end of reported pendency were Rajasthan, Madhya Pradesh and Chhattisgarh with pendency rates of 7.13 per cent, 7.27 per cent and 9.88 per cent. Does this mean the latter states have a better functioning police force? When data is so aggregated, it is hard to make such inferences.


What you measure is what you can manage. In India, we do not do a good job of collecting and disseminating facts about the criminal justice system, down the full pipeline from the crime that citizens experience through all the steps of government systems going all the way to convictions. What data is available is highly aggregated, and therefore not very useful in studying police actions. For a contrast, consider the data portal of the police in UK. This provides detailed information on each crime that took place in each county in each month, and the latest outcome on the crime. We need to move towards better data collection and dissemination as we embark on the reform of our criminal justice system.

Renuka Sane is a researcher at the Indian Statistical Institute, Delhi. Neha Sinha is a researcher at IDFC Institute in Bombay. We thank Rithika Kumar for valuable discussions.

No comments:

Post a Comment

Please note: Comments are moderated. Only civilised conversation is permitted on this blog. Criticising me is perfectly okay; uncivilised language is not. I delete any comment which is spam, has personal attacks against anyone, or uses foul language. I delete any comment which does not contribute to the intellectual discussion about the blog article in question.

Please note: LaTeX mathematics works. This means that if you want to say $10 you have to say \$10.