The System
We keep using the word. Here it is, defined and dated: not a party, not a person, but a machine — built under colonial rule to extract revenue and keep order over a subject people, and handed to a free country almost unchanged. Four organs, each with a birth certificate.
Three words get used as if they were one. They are not. Politics is the layer that changes hands — the party in the chair, the face on the poster, the five-year rotation. The People are the governed: the ones the machine acts upon. The System is the permanent apparatus underneath the politics — the police station, the Collectorate, the Public Works Department, the statute book that protects them, and the treasury that pays for it all. Elections change the operators. The machine runs on.
That distinction is the whole argument of this archive. When we say system change, not party change, we mean precisely this: swapping the party on top leaves the machine intact, because the machine was never built by any Indian party. It was built to govern a colony — to collect, to control, to keep the peace of the ruler — and at Independence it was inherited rather than redesigned. What follows is the machine, organ by organ, with the documents that date each one and the record of how little has changed.
The party is the variable. The machine is the constant. Changing the constant is the work.
Begin with the wound the system is remembered by. On 30 October 1928, Lala Lajpat Rai led a peaceful "Simon Go Back" march in Lahore against the all-British Simon Commission. The Superintendent of Police, James A. Scott, ordered a lathi charge on the marchers. Documented Lajpat Rai was struck — his own statement, later read into the British Parliament's record, says only that he "received two blows from a police officer who refused to give his name." Documented The popular memory that the blows fell on his head is not borne out by that primary record; several accounts say the chest, and the exact spot is not fixed by any first-hand source. Unsupported That evening he said the blows would be "the last nails in the coffin of British rule in India." Contested — the sentiment is his; the exact wording varies between sources.
He died on 17 November 1928, aged 63. Whether the beating killed him is genuinely disputed, and this archive shows the dispute rather than settling it: nationalists held that the injuries hastened his death, while the colonial account — the New York Times obituary and the government's answer in Parliament — recorded a heart attack and insisted "no evidence" tied the death to the blows. Contested
"Beware, ye bureaucracy." — the HSRA's own poster, Lahore, 18 December 1928
What is not disputed is what the anger did next. Bhagat Singh, Rajguru, Sukhdev and Chandrashekhar Azad of the Hindustan Socialist Republican Association resolved to kill Scott in revenge. On 17 December 1928 the plan misfired: a lookout mistook Assistant Superintendent John P. Saunders for Scott. Rajguru fired first, Bhagat Singh followed, Azad's covering fire killed a pursuing constable. Documented They had meant Scott and killed Saunders — the pre-printed posters claiming the deed had to be corrected by hand. Bhagat Singh, Rajguru and Sukhdev were hanged on 23 March 1931. The point for the system is narrow and exact: the police here were an instrument of the ruler against the ruled, and the men who answered it with a pistol were answering the machine.
The machine that ordered that charge has a name and a date. The Police Act of 1861 was written after the 1857 revolt, when the Crown took India from the East India Company, to build "a disciplined, loyal police force to support colonial administration" — with no provision for citizens' rights, accountability, or grievance redress. Documented Its model was the armed, barracked, centrally-commanded Royal Irish Constabulary — policing built to hold a subject population — not the London "policing by consent" force. Documented (Historians such as Clive Emsley caution that the "Irish model" can be overstated and colonial policing was hybrid; the coercive lineage is nonetheless the scholarly consensus. Contested)
That Act is still in force. Policing is a State subject; most states still run on the 1861 Act or near-copies of it, and it survived even the 2024 overhaul that replaced the Indian Penal Code, the CrPC and the Evidence Act. Documented The Supreme Court tried to change it: in Prakash Singh v. Union of India (2006) it ordered fixed tenures, merit-based chief selection, a Police Complaints Authority, and the separation of investigation from law-and-order; a Model Police Act was drafted the same year. Two decades on, implementation across the states is "sluggish and uneven." Documented
Your instinct about the law is right, and it is written down. To prosecute a police officer for something done "in the discharge of his official duty," a court first needs the government's sanction — Section 197 of the old CrPC, now Section 218 of the 2023 BNSS. Documented It is a procedural shield, not a total immunity — courts have repeatedly held no sanction is needed where the act has no honest connection to duty, such as custodial violence — but it is a real, standing barrier to holding the office to account. Run the other direction and there is no such gate: to assault or even to use force to deter a public servant is itself a defined, readily-charged offence — IPC 332 and 353, now BNS 121 and 132. Documented One side of the encounter is shielded by a sanction requirement; the other is exposed by a ready charge. That asymmetry is not an accident of drafting — it is the inheritance of a police architecture built to protect the state and manage the people.
The reform does not need inventing; it needs doing. The 2006 Prakash Singh directions and the Model Police Act are the blueprint the Supreme Court already ordered: fixed tenures so an officer cannot be transferred for doing the job, a complaints authority with teeth, and investigation walled off from the minister's law-and-order desk. A police force accountable to the citizen instead of the ruler is the single change that turns 1861 into 2026.
ANSWERED THE MACHINE WITH A PISTOL; HANGED 23 MARCH 1931 · WIKIMEDIA COMMONS
The most powerful officer in an Indian district is called, to this day, the Collector — and the name is the confession. Warren Hastings created the post in Bengal in 1772, under the East India Company, and it took its name from the plain fact that its holder was "the head of the revenue organization (tax collection) for the district." Documented He collected the land revenue; he was also made magistrate, directing the courts below him and the police. Revenue, justice and force gathered into one hand — a concentration no modern separation of powers would permit. Documented
The honest complication belongs here: it was not a single unbroken fusion. Cornwallis's 1793 code briefly separated revenue from the judiciary, handing justice to a distinct District Judge; the powers re-gathered around the Collector only across the nineteenth century, into the Collector-cum-District-Magistrate we have now. Contested — the continuity is real, but it was assembled, not preserved.
The post did not retire in 1947. The district remained the unit of administration, and the same officer runs it: District Magistrate in Uttar Pradesh and Madhya Pradesh, District Collector in Tamil Nadu and Kerala, Deputy Commissioner in Karnataka, Punjab and the North-East — three names for Hastings's 1772 office. Documented The citizen who needs a certificate, a ration card, a land record, or disaster relief still goes to the Collectorate; the revenue-collector-and-magistrate is still the face of the state in the district.
The direction is the one the Constitution already pointed at. The 73rd Amendment promised self-government to the panchayat; the unfinished work is to make the Collector the servant of elected local government rather than its administrator-in-chief — to unbundle the revenue, magisterial and police powers that 1793 showed can be held apart. See Swaraj for the local-government reality check.
Before there was a Public Works Department, much public work was the community's own charge. In South India the custom had a name — kudimaramat, the customary maintenance of irrigation tanks and channels by the cultivators who used them. It is documented in the colonial record itself: Thomas Munro, as Collector in 1803, wrote that ordinary repairs were done "by the cultivators themselves… either by their own labour, or by an assessment of grain," and the annual clearing of water-courses "was not regarded as any extra duty." Documented
Then the state took it over. A central Public Works Department was raised under Lord Dalhousie in 1854–55, replacing the presidency Military Boards; in Madras (1856–57) "the duty of initiating tank repairs was… removed from the Revenue Officers and transferred to" the new PWD — and the motive was frankly revenue: "the state is so interested in the land revenue dependent on them that it has gradually undertaken their control." Documented The link between the people who used a tank and the duty to keep it was cut, and salaried distance put in its place.
Here the archive has to correct a romance, because the honest version is stronger than the myth. The tanks did decay, and the severing of local responsibility is a documented part of why. Documented But the picture of a self-governing village that "built it themselves and it worked better" does not survive the record. Kudimaramat was not idyllic volunteering: it was compulsory, caste-tied labour, enforced by village officers who kept a "defaulter book," levied fines, and cut off water to those who would not turn out. Documented The British 1858 Compulsory Labour Act had to legislate that compulsion precisely because the custom was already dying — the Irrigation Commission later said it was "well worth resuscitation… if the custom was not dead." Documented And the long decline of tank irrigation had many causes — borewells, encroachment, tenure change — not centralisation alone. Contested
So the claim, stated honestly: local maintenance was real, and displacing it did harm — but it was neither voluntary nor a lost paradise. The recoverable idea is local responsibility, not compulsory labour. Contested
Tamil Nadu has already tried the recovery: its revived Kudimaramathu scheme (2017) funds farmers to de-silt and repair their own tanks with state money and local hands. Responsibility close to the work, funds transparent and accounted — the same model as this project's Action Arm: fix the thing, publish the receipt. Not a return to the corvée; a return of ownership.
Every organ above is paid for by the taxpayer, and the reasonable citizen's question is: where does it go? A common answer — "most of it goes on government advertisements and freebies" — is worth stating because it is wrong, and the true figures make the sharper point. The whole of the Union Budget for 2025–26 is about ₹50.65 lakh crore. Documented Here is the government's own "where the rupee goes," in paise:
| Where each rupee of spending goes | paise |
|---|---|
| States’ share of taxes & duties | 22 |
| Interest on past debt | 20 |
| Central Sector Schemes | 16 |
| Defence | 8 |
| Centrally Sponsored Schemes | 8 |
| Finance Commission & other transfers | 8 |
| Other expenditure | 8 |
| Major subsidies | 6 |
| Pensions | 4 |
Government of India, Budget at a Glance 2025–26 (paise-in-the-rupee, gross basis). PRS Legislative Research reports interest as 25% of the Centre's own ₹50.65-lakh-crore expenditure — the same fact on a narrower base; both are correct, the pie above being gross of the 22 paise devolved to states. Documented
Advertising is a rounding error. The Centre's entire publicity bill through its Central Bureau of Communication was about ₹2,586 crore across six years (2020–2026, per an RTI reply) — roughly 0.01–0.02% of a single year's budget, one to two paise in every hundred rupees, too small to appear on the pie at all. That six-year total is smaller than one year's food subsidy (~₹2.03 lakh crore) by about seventy-eight times. Documented Ads are not where the money goes; the biggest slices are the boring ones — money owed to the states, interest on past borrowing, and the running schemes, defence and subsidies.
"Freebies" is the real argument, and this archive keeps it two-sided because it genuinely is. State subsidy spending rose from ₹2.13 lakh crore (2019–20) to ₹4.71 lakh crore (2024–25) — up 121% — with five states across the political spectrum (Tamil Nadu, Chhattisgarh, Karnataka, Madhya Pradesh, Gujarat) accounting for about 65% of it. Documented The RBI has flagged rising subsidies as an area of "incipient stress" that can crowd out productive spending; the Supreme Court has heard PILs on pre-poll freebies since 2022 without resolving them; economists remain split between "welfare investment" and "fiscal risk." Contested And this is systemic, not partisan: free ration for ~80 crore people and PM-KISAN cash come from the Centre; free power, bus travel, cash-to-women and loan waivers come from governments led by every major party. Whatever one calls them, no single side owns them.
The honest fix is not to chase the rounding error. It is transparency: every rupee traceable, every welfare scheme argued on its evidence rather than its sponsor. That is exactly what the Action Arm is built to prove in miniature — money raised and spent in public view, every fix closed with a published receipt. The cure for money spent out of sight is money spent in sight.
A police built to control, dated 1861. A Collector built to extract, dated 1772. Public works pulled away from the people who used them, dated 1854–57. Statutes that shield the office and expose the citizen. All of it funded by a tax rupee spent where few can see it. Five years at a time we change who sits on top of this machine — and call it change. System change, not party change is simply the refusal to keep mistaking the operator for the engine.
Elections rotate the driver. The archive is about the vehicle.
Sources
- Police: UK Parliament Hansard, "Late Lala Lajpat Rai" (26 Nov 1928); New York Times obituary (18 Nov 1928); HSRA poster (18 Dec 1928, marxists.org); Police Act 1861 & MHA bare act; PRS, Police Reforms in India (Prakash Singh 2006); §197 CrPC / §218 BNSS; BNS 121/132.
- Collector: District Magistrate / Collector (Hastings 1772; etymology; nomenclature); Cornwallis Code 1793 (revenue–judiciary separation).
- PWD & kudimaramat: Munro (1803) & Madras 1856–57, in "Land and Water Policy Under British Colonial Rule"; Madras Compulsory Labour Act 1858; Mosse, The Rule of Water (2003); Aubriot & Prabhakar, Water Alternatives (2011); TN Kudimaramathu revival (2017).
- The Money: PRS, Union Budget 2025–26 Analysis; Government of India Budget at a Glance 2025–26; RBI, State Finances: A Study of Budgets 2024–25; FACTLY (DAVP/CBC ad spend, RTI).
- Full tagging in the Claims Ledger. Contested claims are shown as contested; the fabrications this case does not need are left out.
This page describes the machine. Three others take it apart further. The Extraction Machine is the revenue engine underneath it — how the land was taken and taxed, the class the settlements manufactured, the statutes that kept order, and the excise revenue the state quietly came to depend on. The Interval That Never Ended asks who, at independence, argued for keeping this apparatus and who argued for dismantling it — and finds that the most-quoted line in that story belongs to a British prime minister. And the Constitution records what the Assembly argued about while writing a document on top of it all.