Accused out of proportion

Between the 2017 and 2023 censuses, Pakistan recorded more Christians and Hindus, but fewer Ahmadis. The fall in the recorded Ahmadi population sits beside another large difference in the data. Ahmadis made up 0.07 per cent of the population recorded by religion in the 2023 census, but 26 per cent of the people CSJ recorded as accused of blasphemy between 2020 and 2025. The census decline predates the sharp rise in Ahmadi accused in 2025.

The 2025 total is the highest in the six-year series used here. The Pakistan Penal Code’s chapter on offences relating to religion runs from sections 295 to 298-C. Section 295-C provides for the death penalty, while sections 298-B and 298-C apply specifically to Ahmadis.

We set CSJ’s data beside the 2017 and 2023 censuses, district by district, to assess whether accusations are recorded more often where religious minorities live, and whether recorded minority populations increased more slowly where accusations were recorded. The first has a qualified answer. The available dates and district totals cannot establish whether accusations were followed by a slower increase in a community’s recorded population.

The two datasets

The census figures come from Table 9 of the Pakistan Bureau of Statistics population censuses, which gives each district’s population by religion. We reproduce the census categories as published, including its separate Ahmadi category, and combine Hindus and Scheduled Castes. Boundary changes are reconciled into 124 comparable district units.

The accusation figures were compiled by CSJ, which counts people accused and records the religion of each, whether or not police registered a case. CSJ’s record includes formal and informal accusations and does not claim to be exhaustive, and one case can involve multiple accused people. A district missing from a CSJ list means that CSJ recorded no accusation there, which does not establish that none was made. Of the 1,035 people CSJ recorded between 2023 and 2025, 974 could be matched to one of the 124 districts.

How minority populations changed between the censuses

Any comparison with accusations starts from where each community lives and how its numbers moved. Figure 1 shows the change in each community’s recorded population between 2017 and 2023, against the increase in the Muslim population over the same six years.

A population’s increase and its proportion of the district can move in opposite directions. Lahore’s recorded minority population rose from 589,000 in 2017 to 616,000 in 2023, while its proportion of the district fell from 5.30 to 4.74 per cent because the rest of Lahore’s population rose faster. Across the 124 districts, the minority count rose in 122, while the minority proportion fell in 13: Lahore, Sheikhupura, Kasur, Gujranwala, Sialkot, Rawalpindi, Narowal and Sahiwal in Punjab; Quetta, Sibi and Sherani in Balochistan; Sujawal in Sindh; and Islamabad. The count fell only in Quetta and Sherani.

The Christian population recorded by the census rose by 24.9 per cent, from 2.64mn to 3.30mn. The combined Hindu and Scheduled Caste population rose by 17.4 per cent, while the Muslim population rose by 15.6 per cent. The Christian increase merits a closer look. In 76 districts the count more than doubled in six years. Those districts contained only 55,000 Christians in 2017 but account for about a quarter of the national increase. In Khyber Pakhtunkhwa, the count rose from 50,000 to 135,000. The census tables cannot separate births and migration from changes in enumeration or classification. The 2023 census was Pakistan’s first digital enumeration. If it counted people the 2017 census had missed, part of the recorded increase is the recovery of an earlier undercount rather than people added since 2017. The published tables cannot confirm or rule this out. In Punjab, where three in four Christians live, the Christian count rose by 19.2 per cent while the province’s total population rose by 15.8 per cent.

The recorded Ahmadi population fell by 15.2 per cent, from 192,000 to 163,000, with declines in 94 of the 124 districts. The two census counts cannot separate the effects of births and deaths, migration into or out of the area covered, and changes in identification or enumeration. Movement between districts within that area cannot by itself explain the overall decline.

Who is accused

The census describes communities, and the CSJ record describes individuals. Figure 2 shows the people CSJ recorded as accused in each year from 2020 to 2025, by the religion recorded for them.

The figures count people rather than cases. In 2024, for example, CSJ recorded 32 Hindus accused, of whom 31 were named in a single case in Islamkot, in Tharparkar district.

Muslims are the largest group in every year. They accounted for 213 of the 362 people recorded as accused in 2025. CSJ notes that its Muslim category includes members of smaller sects that the table does not separate. Ahmadis are the second-largest group. Their number rose from 49 in 2024 to 141 in 2025, taking their proportion of the annual total from 14 to 39 per cent. Across the six years, 394 of the 1,498 people recorded were Ahmadi, 61 were Christian and 44 were Hindu.

Ahmadis account for 0.07 per cent of the population recorded by religion in the 2023 census and 26 per cent of the people recorded as accused of blasphemy between 2020 and 2025.

Several 2025 cases involved many Ahmadis at once. Police arrested 22 in Daska in March after they gathered for prayers. Later that month, an FIR in Karachi’s Surjani Town named 19 Ahmadis and more than a dozen unidentified people. In April, Kasur police booked 42 Ahmadis, two named and 40 unidentified. CSJ’s figures for Ahmadi accused draw primarily on community sources, and CSJ takes a conservative approach where counts differ. These reported case counts have not been reconciled with CSJ’s annual total.

Christians account for 1.4 per cent of the 2023 census population and 4 per cent of those accused over the six years. The two sources cover different periods and areas: CSJ’s national series includes Azad Jammu and Kashmir and Gilgit-Baltistan, while the census comparison covers the four provinces and Islamabad.

Where accusations are recorded

Whether accusations are recorded where minorities live is a question about districts. Figure 3 plots each district’s Christian, Hindu or Ahmadi proportion of the population in 2023 against the number of people of all religions recorded as accused there, per million residents a year, between 2023 and 2025.

At first glance, districts with larger minority population proportions tend to have higher recorded accusation rates. Much of that association reflects differences between provinces. Punjab accounts for 628 of the 974 accused who could be matched to a district, and Sindh for another 270. Both recorded about 1.6 accused people per million residents a year, compared with 0.5 in Khyber Pakhtunkhwa and less than 0.1 in Balochistan. Punjab and Sindh are also home to 91 per cent of Pakistan’s Christians, 99 per cent of its Hindus and Scheduled Castes, and 98 per cent of its Ahmadis.

The national rank comparisons weaken substantially after accounting for province, population and urban share. Within Punjab, districts where Christians form a larger proportion of the population still record slightly more accused people per resident, though the association is modest. Some national count models also retain positive associations for Christian and Hindu proportions, so the strength of the relationship depends on the method. The district rate includes accused people of every religion, so it cannot tell us whether Christians face a higher chance of accusation.

A few cases can move a district sharply. Kurram has the highest recorded rate, 9.3 accused people per million residents a year over 2023 to 2025, from 22 people recorded in the three years. Tharparkar’s rate of 6.6 rests largely on the Islamkot case in which CSJ recorded 31 Hindu accused. Table 1 lists the eight districts with the highest recorded rates, with Lahore and Karachi for comparison, and sets each beside its minority proportion in 2023. The highest rates occur both where minorities are under 1 per cent of the population, as in Kurram, Chakwal and Mandi Bahauddin, and where they are more than a third of it, as in Tharparkar, Tando Allahyar and Umerkot.

What district totals cannot show

The second question is whether accusations are followed by a community shrinking. The available district data cannot establish whether accusations preceded population decline.

Of the available complete annual district lists, only 2021 falls wholly within the 2017 to 2023 census interval; it names 21 of the 124 comparable districts. Four years of population change had already occurred before this snapshot. In Punjab, districts with more accused people per resident in 2021 tended to see their Christian proportion rise less or fall. Their recorded Christian populations, however, did not consistently increase more slowly.

The Christian proportion of the district population fell in six Punjab districts: Lahore, Sheikhupura, Kasur, Rawalpindi, Gujranwala and Sialkot. Four appear in CSJ’s 2021 list. All six still recorded more Christians in 2023 than in 2017. Their recorded Christian counts increased more slowly than the district totals. If the 2023 count recovered people the 2017 census had missed, the comparison is weaker still: a district whose 2017 count was closer to complete would show a smaller increase for that reason alone. District totals cannot show whether families left, moved elsewhere within the district or were recorded differently. A move between two neighbourhoods of Lahore leaves the district total unchanged.

One Lahore case shows how much can happen below that level. In February 2018, Patras Masih, a Christian resident of Shahdara, was accused of posting blasphemous material online. Protesters converged on the locality and threatened to burn it down, and most of its roughly 800 Christian residents left. More than a week later, Al Jazeera found the neighbourhood still largely deserted. By October, Amnesty International reported that Masih’s parents were still displaced because of continuing threats. Public reporting does not show how many residents returned, or whether those who did not stayed within Lahore.

A November 2021 report points to movement within Lahore. After a blasphemy case was registered in Khodi Khushal Singh village, Dawn reported that some Christian residents unconnected to the allegation had moved to relatives elsewhere in the city out of fear. Such a move would leave the district population unchanged.

What remains invisible

Recorded accusations are concentrated in Punjab and Sindh, which record about 1.6 accused people per million residents a year, against 0.5 in Khyber Pakhtunkhwa and less than 0.1 in Balochistan. The association between minority population proportions and recorded accusation rates weakens substantially after adjusting for province, population and urban share, and the results vary by community and method. Ahmadis are 0.07 per cent of the census population and were 26 per cent of the people recorded as accused between 2020 and 2025, and 39 per cent in 2025. That is the clearest pattern in the data.

What happens after an accusation is harder to trace. The census counts the population once every several years, six in this case, so it cannot follow a family that leaves one neighbourhood, stays with relatives elsewhere in the city and later returns. The Shahdara case documents displacement and an uneven return after an accusation, and the 2021 report from Khodi Khushal Singh documents a move within Lahore. These movements cannot be reconstructed from district totals.

The census recorded 29,000 fewer Ahmadis in 2023 than in 2017. The sources examined here cannot establish how much of that decline reflects births, deaths and migration and how much reflects people not being counted, or not being recorded as Ahmadi. That question stays open.

Ghufran Khalid is an economics graduate from NUST and an applied researcher working across economic, financial and policy analysis. He works in equity research and writes independently on Pakistan’s economy, with a particular interest in industrialisation, structural change and data-driven economic research.

Acknowledgements

The accusation data used in this piece were compiled by the Centre for Social Justice (CSJ) and are used with its permission. We are grateful to Peter Jacob, Executive Director of CSJ, for his input into this work.