There’s been much heated debate about Surjit Bhalla’s August 29 and September 5 Indian Express op-eds in which he argued: “the only explanation for the constancy of the share of the Christian population between 1991 and 2011 (2.32% and 2.3% respectively) is large scale conversion to Christianity”.

The crux of Bhalla’s argument is that because Sikhs and Christians had comparable income and fertility levels in the early 1990s, their respective population growth rates in between 1991 and 2011 should have been similar. Instead, “the Sikh population grew at an average rate of 1.2% per annum, while the population growth rate of Christians was a relatively higher 1.9% per annum… despite [Christians] having the highest per capita consumption, the highest level of female education and the lowest fertility.” The gap is accounted for, says Bhalla, by religious conversions to Christianity, which he quantifies at 1,70,000 per year between 1991 and 2011.

This seems at first like a smart way to isolate the impact of conversion using demographic data. But as the following critique by Rohini Prabha Pande, expert on gender and population dynamics, shows, Bhalla may be too clever by half:

Guest post by Rohini Prabha Pande

Why Bhalla’s argument is half-baked 

The most likely explanation for Bhalla’s “puzzle” might be the simplest one. Even with a slightly lower fertility rate, the Christian population may have grown larger than that of Sikhs because of a larger population base to which this fertility rate is applied, rather than lakhs of conversions. In particular, if the age distribution among Christians favors the reproductive age groups, then even with lower fertility rates the population could grow. This is related to the concept of population momentum and, at minimum, Bhalla needs to examine the age distribution and age-specific fertility rates of Christian versus other groups to rule this out as an alternative explanation.

In another sleight of hand, Bhalla summarily rules out migration – a key component of population growth – as an explanation of why the Sikh population might have grown more slowly than the Christian population. Here too he needs to explain why he is correct, and others – like Aswini Nanda – are wrong.

There are other flaws in his reasoning. I agree with Tony Joseph’s argument that Bhalla’s results depend on which period you pick. To that I would like to add that Bhalla’s focus on 20-year time periods brushes under the rug some key aspects of the relative growth of these two groups’ population, because the more you collapse data, the more information you lose. A quick calculation shows that if you simply parse 1991-2011 into two 10-year census periods (even without going further back in time to 1971, as Joseph does) you get a somewhat different picture: the (compounded) rate of growth of the Christian population across the two periods is slowing faster than the rate of growth of the Sikh population (see Table 1 below).

In other words, while the Christians may have a higher population growth rate than the Sikhs, both groups’ rates are dropping, and the Christians’ growth rate is declining faster than that of the Sikhs. The difference between the Christian and Sikh populations’ growth rates has narrowed from 0.9 percentage points between 1991-2001 to 0.6 percentage points between 2001-2011.

Bhalla then asks “what caused Christian population growth to accelerate from 1.39% (1971-91) to 1.93% a year (1991-2011)?” and concludes that the answer lies in religious conversions. One could equally and legitimately ask what caused Christian population growth to drop from 2% between 1991-2001 to 1.5% between 2001-2011. Framed this way, the question does not – unfortunately for Bhalla – lead to the conclusion that conversions have risen.

So one has to take Bhalla’s assertions with a large pinch of salt, if one were to pay any attention to them at all.

Finally, in his 4 September rebuttal of Joseph, Bhalla asserts that: “The excess Sikh males do not marry; excess Christian women do marry and produce Christian kids.” Aside from the parochialism implicit in this statement, is he assuming that the excess Sikh males don’t marry because there are no Sikh women? His own occasional coauthor, sociologist Ravinder Kaur, is the foremost researcher in India (look here) on the increasing phenomenon of across-region marriage, showing in particular how excess males in Punjab (presumably including Sikhs) import brides from Andhra Pradesh and Assam.

To sum up, Bhalla’s calculations of conversion are at best inconclusive, at worst, disingenuous.