Mandal Commission explained: how 11 indicators produced the 27% OBC quota

A caste that scored 11 points was a backward class. A caste that scored 10 was not.

The scale ran to 22, so 11 was exactly half of it. ¹

One point was the whole difference between the two sides of that line. Run that test across a country and it produces 3,743 castes identified as socially and educationally backward,¹ and a population estimate putting the Other Backward Classes, excluding Scheduled Castes and Scheduled Tribes, at nearly 52% of India.¹

This is how that number was built, and what was done with it.

Eleven indicators, twenty-two points

The Second Backward Classes Commission evolved eleven indicators for determining social and educational backwardness, grouped under three heads: social, educational and economic.¹

The heads were not weighted equally. Social indicators carried 3 points each, educational 2, economic 1, and the Commission said why in its own arithmetic: ¹

So more than half the scale sits in one head. A caste could be poor and under-schooled and still fall short of 11 if it was not socially marked, and that is a design decision rather than an accident. The Commission was measuring a social category and said so in the weights.

What the survey actually was

Out of the 406 districts in the country, the survey covered 405 districts.

That number sounds like completeness and is not quite what it seems. In each district the survey took two villages and one urban block, and surveyed every household in them.

Then a floor. Where fewer than 20 households of a caste were covered, that caste was left out of consideration, the sample being treated as too small to carry any dependable inference.¹

Read that carefully, because it decides who could be found at all. A caste thinly spread across the country, present everywhere in small numbers and concentrated nowhere, could clear no district's threshold and drop out of the frame entirely. The rule is defensible as statistics and it is not neutral in its effects.

The scoring threshold itself was plain: ¹

52%, and the number it could not ask for

The Commission's derived figure was an approximate derived population of Other Backward Classes, including non-Hindu communities, of 52%,¹ set against Scheduled Castes at 15.05 per cent and Scheduled Tribes at 7.51 per cent.²

52 per cent is not a finding about a minority. On the Commission's own numbers it is most of the country.

And that is where the report ran into a line it had not drawn. ¹ What it recommended instead was a reservation of 27% only, even though the population of the Other Backward Classes was almost twice that figure, in deference to legal limitation.¹

In deference to legal limitation. The Commission is saying, in its own report, that the figure it recommended is not the figure its findings supported. It is the figure the courts left available:

In view of the decisions of the Supreme Court limiting the total reservation to 50 per cent, the Commission recommended 27 per cent reservation in favour of OBCs (in addition to 22.5 per cent already existing

22.5 plus 27 is 49.5. Half a percentage point under the ceiling built in Balaji out of a sentence that said speaking generally and in a broad way.

So the most consequential number in the report is not a measurement at all. It is the largest quantity that would survive judicial review, and the petitioners later said so in court: ¹

The attack on the method

The challenge, when it came, went at the survey rather than at the arithmetic.

One line of attack was the count itself: ¹

Three exercises, three answers: 2,399, 3,743 and 1,057. The judgment records the argument that this was a steep increase over the Kaka Kalelkar Commission, which had put the number at 2,733.¹

The second was the sample. Those challenging the report criticised the manner and method of ascertaining backwardness by questionnaire, issued to what they put at hardly 1% of the population.¹

The third was the evidence the Commission fell back on, which included the caste census of 1931, described in argument as obsolete material.¹

And the framing that tied them together came from counsel: ¹

Those are the criticisms as the judgment records them, made by parties challenging the report. They are arguments placed before a court, not findings of one.

12.55 and 4.69

Against the method sat a different kind of number, and it is the one the government eventually acted on.

The counter affidavit carried a table drawn from the Commission's employee survey: ¹. On that table, ¹ And higher up the structure the share thins further: ¹

Notice what kind of evidence that is, because it changes what the argument rests on. It is not a count of who belongs to which caste. It is a count of employees surveyed in government service. The indicators described who was backward; this table described what that backwardness looked like inside the state's own payroll.

Whatever one thinks of 11 points out of 22, the representation figures are a separate measurement, and they fall as the grade rises.

Six days in August

On 6 August 1990, records the National Commission's own account, ²

The next day it reached Parliament, announced there by the Prime Minister on 7 August 1990.² Then the file itself, on 13 August: ² August 1990, providing reservation for Socially and Educationally Backward Classes of 27%.

A decision on the 6th, an announcement on the 7th, a memorandum on the 13th. Ten years after the report was submitted, and seven years after a Home Minister told Parliament the government desired to implement it, the instrument was one office memorandum running to a little over a page.

The rewrite, and the stay

The memorandum of 13 August 1990 contained no income test. A later government amended it, adding a preference for the poorer sections within the 27 per cent and carving out a further 10 per cent for other economically backward sections, with criteria to be issued separately.¹

So within about a year the same quota carried two different theories of who it was for: a social category identified by 11 indicators, and an economic one to be defined later.

Writ petitions and stay applications followed, and the whole thing was frozen before it operated.¹ What had been decided in six days in August waited for nine judges.

What the counting was for

Strip the argument back and the report did two separable things.

It built a test, and the test is contestable in every direction: the weights put 12 of 22 points in one head, the 20-household floor decided who could be seen, and three serious exercises produced 1,057, 2,399 and 3,743.

And it produced a representation figure, 12.55 per cent of government employment and 4.69 per cent of Class I, for a group it put at 52 per cent of the population.

The recommendation that followed was neither of those numbers. It was 27, chosen in deference to a legal limitation, landing half a percentage point below a ceiling the courts had never written into a statute.

Which is the question the nine judges were left with. Not whether 3,743 was the right count, but whether a quota set by the size of a judicial estimate rather than by the size of a finding could stand at all.

MangoPOV. "Mandal Commission explained: how 11 indicators produced the 27% OBC quota." 14 September 2026. https://mangopov.in/stories/mandal-commission-eleven-indicators/

@misc{mangopov-mandal-commission-eleven-indicators,
  title        = {Mandal Commission explained: how 11 indicators produced the 27% OBC quota},
  author       = {{MangoPOV}},
  year         = {2026},
  howpublished = {\url{https://mangopov.in/stories/mandal-commission-eleven-indicators/}},
}