Census effect: Cities will see ‘less crime’ next year

In 2027, most Indian cities will show a decline in crime rate. That is a safe prediction to make. Historical precedent suggests crime rates in large Indian cities fall sharply in a Census year.

This is not because of better policing, but because of a mathematical glitch — one that the National Crime Records Bureau mentions in its fine print, but which is often missed in reports based on its data.





NCRB updates the number of crimes every year. But for cities, it uses the last Census population as the denominator for calculating crime rate. That population number remains unchanged until the next Census. So, while crimes recorded each year may rise or fall, the population used to calculate the rate stays frozen — usually for a decade, and this time for more than 15 years.

This is why comparisons such as “Delhi’s crime rate has risen” or “Delhi’s crime rate is higher than Mumbai’s” can be misleading unless the population base is understood.

Take Delhi. In the latest NCRB report, the population used for Delhi City is still 1.6 crore, based on the 2011 Census. But the projected population of Delhi NCT is 2.2 crore. Since Delhi City accounts for almost the whole of the NCT, the city clearly cannot still have only 1.6 crore people if NCT has 2.2 crore.

This difference changes the crime rate sharply. The number of IPC crimes is the same for Delhi and Delhi City — 2,75,402. But because NCRB uses different population bases, Delhi’s crime rate is shown as 1,259 per lakh, while Delhi City’s rate is 1,688 per lakh — 34% higher.

There is a precedent for what happens when the population denominator is updated. In the 2001 Census, India had 35 cities with a million-plus population.

When city populations were updated in 2011, crime rates fell in 27 of them. Kochi’s rate dropped from 1,898 per lakh to 1,636. Visakhapatnam, Bengaluru, Indore, Ahmedabad and Bhopal also saw drops of over 150 points. The steepest 'falls' came in fast-growing cities, where the old population number had become most outdated.

That is the Census effect. NCRB uses the same city population until the next Census because official city-wise projections are not available.

Juvenile crime rates may be slightly understated in NCRB’s calculations

A constant divisor makes the rate look progressively inflated. When the Census finally updates population, the rate can fall dramatically even if the actual number of crimes has not fallen.

Similar issues appear in other NCRB categories. Crime rates against children, and crimes committed by juveniles, use India’s under-18 population from the 2011 Census — 44.4 crore. But with fertility rates falling, UN estimates suggest the under-18 population is now around 43.3 crore. That means juvenile crime rates and crimes against children may be slightly understated in NCRB’s calculations.

The reverse may be true for senior citizens. NCRB uses the 2011 senior citizen population — 10.4 crore people aged 60 or above. But India is ageing, and UN estimates put the 60-plus population in 2024 at 15.7 crore. If so, NCRB may be substantially overstating the rate of crime against senior citizens.

The lesson: crime numbers and crime rates do not always tell the same story. The number of crimes is an annual count. The rate depends heavily on the population used in the denominator. In Census years, that denominator changes — and with it, the crime rate can appear to improve overnight.