Why It Breaks
In 1856–57 the colonial state took the duty of repairing irrigation tanks away from the cultivators who used them and handed it to a salaried department. That severance is not a historical curiosity. This page is what it looks like now — the road opened three weeks after it was laid, the pothole that is really a drainage failure, the treatment plant connected to nothing, and the one budget line that explains most of it. Every finding here is from the government's own audits and parliamentary committees. And every section ends with something that has actually worked, in India.
In 2023–24 the Ministry of Road Transport and Highways was allocated ₹2,600 crore to maintain the national highways — one per cent of the Ministry's total budget. Over the nine years to 2023 that network had grown by 54,858 km. Parliament's own Standing Committee called the maintenance allocation "miniscule," and noted that several committees including NITI Aayog have said maintaining existing roads should be prioritised over building new ones. Documented
You do not need corruption to explain Indian infrastructure. Building is an announcement. Keeping is a line item. Every other finding on this page is a variation on that sentence — and none of it is one party's doing. The audits below cover the Union road ministry, Delhi's PWD, Mumbai's BMC, Bengaluru's BBMP, Uttarakhand's water agencies and twenty-nine states, under many different governments. This is a structure, not a scandal.
Here is the structural cause of the road being opened after it is laid, taken from a real rule. New Delhi's road-cut circular lists the agencies a digging party must separately notify: water supply, sewerage, the electric division, horticulture, CPWD, MTNL, Delhi Transco, BSES, Indraprastha Gas, Delhi Police and Traffic. Documented
Read what that document does and does not do. It creates a duty on the digger to phone everybody else. It does not create a single owner of the corridor. Coordination is delegated to correspondence — and correspondence is exactly what fails when eleven bodies have separate budgets, separate calendars and separate approving authorities.
India is not short of road-cutting rules. The NDMC regime (2021) has real teeth: restoration charges paid upfront by surface type (₹4,430/sqm for bituminous concrete); a monsoon moratorium — no permission between 1 July and 30 September except emergencies; an express presumption against cutting main roads at all, with trenchless technology to be considered first; mandatory ground-penetrating radar survey before permission; a bank guarantee of 50% held for a year; escalating debarment of 3, 6 and 12 months; provision for an FIR; double-rate charges for digging without permission; and a clause compelling utilities to migrate into a common service corridor at their own cost once one exists. Documented
Delhi PWD went further in 2024 — all restoration of PWD roads to be done by PWD only, and implementing agencies must file an advance action plan of roads to be cut, twice a year, in January and September. That is the closest thing to a shared corridor calendar we found in India. Documented
So why does it still happen? Three answers, and the third is the real one:
First, the agencies that dig most are exempted from the bond — the bank guarantee is waived for government departments. Second, the rules are municipal and the diggers are often not. And third — there is no published compliance data anywhere. Not permits issued, not penalties levied, not guarantees forfeited, not a single agency recorded as debarred. The rules are public. The enforcement record is not. Contested — because it is unmeasurable, not because it is disputed.
There is no national dataset on road-cutting incidence in India. No count of how many kilometres are cut per year, by which utility, or how many were cut within months of being laid. Cities hold permit records; nobody aggregates or publishes them. Documented as an absence.
Which means every claim about roads being dug up right after laying — including yours — is currently anecdote. It is correct anecdote. That is precisely the problem: a failure that everybody experiences and nobody measures cannot be managed, budgeted against, or pinned on anyone.
And there is a sharper version. NDMC requires a ground-penetrating radar survey before every dig, at the digger's cost — and does not appear to retain the result as a shared public map. No Indian city publishes an underground utility asset register. Every agency pays, again, to rediscover the same pipes. Documented The national single-window that does exist — GatiShakti Sanchar — covers telecom only: one of the five utilities under the road.
Potholes and "collapses under rain" are not two complaints. They are one. The Indian Roads Congress code on road drainage requires that the pavement structure including subgrade "must be protected from any ingress of water" — and then adds the sentence that matters: the road must be built so that it will drain "in the event of a failure of the integrity of the surfacing layers." Documented
Read that clause carefully. The code assumes the bitumen will eventually crack. It requires the structure underneath to survive that. So a road that disintegrates when water gets in has not suffered bad luck or cheap tar — it has failed a design requirement that was written down decades ago. Camber, the cross-slope, is named in code as structural, not cosmetic.
The relevant standards, by number: IRC:SP:42 (road drainage, 2014 revision) · IRC:SP:50 (urban drainage) · IRC:37 (flexible pavement design) · IRC:86 (urban roads and streets) · IRC:103 (pedestrian facilities, 2022 revision). India does not lack standards.
The CAG's performance audit of storm-water management in Bengaluru, conducted with technical support from ISRO, is the most complete account of how this fails. Documented
Drain length in the Koramangala valley fell from 113.2 km on 1900s cadastral maps to 62.8 km by 2016–17; in the Vrushabhavathi valley from 226.3 km to 111.7 km. Two drains merging before Bellandur lake shrank from 338 m to 136 m between 2008 and 2016, and construction followed. About 780 million litres a day — 54% of the city's untreated sewage — was discharged into storm-water drains and lakes, with sewer lines laid inside storm-water drains in breach of two statutes.
And then the part that explains the rest: the city has no complete map or database of its drainage network. A Master Plan of Drains costing ₹3.6 crore classified primary and secondary drains and did not map tertiary and roadside drains at all — which are precisely the ones beside the road that flood it. BBMP told auditors that the records of the tender, process and payments to that consultant were lost, and detailed volumes existed for only two of eight zones. Because the master plan and the project reports were commissioned simultaneously, the detailed reports were written before the base data they were meant to follow existed. Works were later abandoned when contractors found there was not enough land beside the drains, and ₹83.6 crore in central grants was lost. Documented
Missing data is not a neutral absence. It is the condition under which the spending continues unchallenged.
44% of Delhi's roads have no footpath at all. Of those that exist, only 26% meet the IRC minimum width of 1.8 metres, and effective pedestrian width across the city has fallen below 0.8 m — less than half the legal minimum. Documented The Supreme Court has held that use of a demarcated footpath is protected under Articles 21 and 19(1)(d), and that pedestrian movement takes legal priority over motorised vehicles. Documented
Two widely-shared figures are not on this page because we could not trace them to any primary document: "84% of Indian footpaths fail basic engineering standards," and "72% of Ahmedabad's roads have no footpath." Both circulate confidently. Neither survives checking. Unsupported
Lowest-bidder selection. Central procurement runs under the General Financial Rules 2017, where the default award is to the lowest evaluated bid — L1. Quality-cum-Cost Based Selection now exists for works, in the Finance Ministry's own 2025 procurement manual — but it is permissive: it "may be used… wherever applicable." L1 remains the residual rule. Documented ⚠️ What share of Indian road works is actually awarded on quality rather than price is not published. That is a real measurement gap.
Defect liability exists — and goes unenforced. Rural roads under PMGSY carry a five-year defect liability, chained to a further five years of zonal maintenance. Highway contracts were announced in 2024 as moving from five years to ten. On paper this is a decade of contractual responsibility. Then the audit: Documented
In 12 states, maintenance during the defect liability period was not carried out in 1,590 road works.— CAG, Performance Audit of PMGSY, Report No. 23 of 2016, covering 29 states
The same audit found roads completed without the required bridges or cross-drainage structures, making them unusable as all-weather roads — note which component gets value-engineered away — and quality-control failures with no field laboratories, no equipment and no trained staff in twelve states. Parliament's Standing Committee on Rural Development found 21% of completed PMGSY work was not properly maintained, and that only 15 states had a rural roads maintenance policy at all. Documented
India does not lack a defect liability rule. It has one, and does not enforce it.
Under lowest-bid selection with a construction-only contract, a contractor's rational move is to bid low and build to whatever survives inspection. Under a performance-based maintenance contract — paid to hold a road at a defined level of service for years — the cheapest available action becomes building it properly the first time. The incentive is not exhorted into existence; it is inverted by the contract.
India already has the template. The Indian Academy of Highway Engineers — the road ministry's own training arm — runs a formal course on performance-based maintenance contracts covering level of service, key performance indicators, and awards and penalties. Documented ⚠️ What India has not published is whether it works here — we found no Indian evaluation comparing performance-contracted roads against conventionally maintained ones. Unsupported
| Stage | Capacity | Share of generation |
|---|---|---|
| Urban sewage generated | 72,368 MLD | 100% |
| Installed treatment capacity | 31,841 MLD | 44.0% |
| Operational capacity | 26,869 MLD | 37.1% |
| Actually treated | 20,235 MLD | 28.0% |
Public debate knows about the first gap — sewage with no plant. There are two more. 4,972 MLD of built capacity is not running (376 of 1,469 plants under construction or non-operational). And a further 6,634 MLD of working capacity sits idle because the sewage never arrives — the pipes don't reach the plant. Documented
India treats 28% of its urban sewage. It has paid for 44%.
Bihar generates 2,276 MLD, has 10 MLD installed and 0 MLD operational. Assam, Arunachal, Manipur, Meghalaya and Nagaland: zero. Kerala generates 4,256 and treats 47. Documented
This is a forty-year, many-government story and must be read as one: Ganga Action Plan Phase I from 1985, Phase II from 1993, the National River Conservation Plan from 1996, the National Ganga River Basin Authority in 2009, Namami Gange from 2014. Six programmes, five decades, many parties — and a constant failure mode.
The Uttarakhand audit is the one that inspected plants physically rather than reading returns. Of 44 treatment plants, 12 were discharging untreated sewage straight into the Ganga. In one inspection round only 3 of 44 met standards and 36 failed. And the finding that names the mechanism: Documented
The CAG described 21 plants across seven towns as "symbolic" — built, and not linked to any functioning sewage system. In Joshimath, ₹42.73 crore of infrastructure did not connect a single household. Sewer connectivity in Chamoli-Gopeshwar was 354 of 5,510 households — 6.4%. Rishikesh's 5 MLD plant was receiving about 17 MLD; Devprayag's ran at 3–4% of capacity for want of inflow. The maintenance agency refused to take over 18 of the 44 plants, citing construction and safety defects. Of ₹20.59 crore in liquidated damages due from contractors, ₹0.89 crore was recovered.
Money buys concrete. Concrete without a sewer network reaching households, without an operations contract, without power and without enforcement is a monument, not a treatment plant.
⚠️ A correction on the money, because it is distorted in both directions: ₹42,019 crore is the sanctioned project cost for Namami Gange, spread over years and much still under construction. ₹23,424.86 crore is what was actually budgeted and released to 2024–25. Commentators routinely quote the larger number as if spent. And real gains exist alongside the failures: third-party inspections cut industrial BOD load from 26 to 13.73 tonnes per day between 2017 and 2023, and the Ganga dolphin population rose to 6,327 in the 2021–23 survey. Documented
The 22-km Delhi stretch is about 2% of the river's length and carries over 80% of its total pollution load. ₹6,500 crore was spent between 2017 and 2021. Dissolved oxygen beyond Palla is almost zero — the river is, in the assessing body's phrase, "officially without life." Documented
Four documented reasons, and none of them is treatment capacity: of 22 drains, only 9 are tapped; 92% of the load comes from five drains and 84% from just two; treated water is discharged back into the same drains carrying raw sewage, so the cost of cleaning is negated; and over 30% of Delhi's urban population lives in unauthorised colonies whose toilets are not connected to sewers. Documented
Ethanol blending reached 19.24% in 2024–25 and 20% in 2025–26; universal E20 became a legal mandate on 1 April 2026. The programme runs from a 2001 pilot through the 2003 Auto Fuel Policy and the 2018 Biofuels Policy — many governments. Documented
| Vehicle cohort | Efficiency loss | Source |
|---|---|---|
| 4-wheelers designed E10, calibrated E20 | 1–2% | NITI Aayog 2021; MoPNG 2025 |
| Fleet-wide, controlled roller tests | 2–6% | ARAI, 2026 |
| 2-wheelers designed E0, calibrated E10 | 3–4% | NITI Aayog 2021 |
| 4-wheelers designed E0, calibrated E10 | 6–7% | NITI Aayog 2021 |
Both public extremes are wrong. The viral "30% mileage loss" is Unsupported — it misreads ethanol's lower calorific value as a mileage figure. But "no mileage loss at all" is also unsupported: the government's own 2021 NITI Aayog roadmap and its 2025 statements both quantify a loss. The honest range is 1–2% for E20-tuned cars up to 6–7% for older cars designed before E10. Documented
The ministry concedes a real, bounded materials effect too: in "certain older vehicles, some rubber parts and gaskets may require replacement earlier… once in the lifetime of the vehicle." Documented
And here is the part worth knowing: NITI Aayog's own 2021 roadmap recommended compensating consumers — "tax incentives on E10 and E20 fuel may be considered" to offset the efficiency drop. It was not implemented. Documented
Reported at roughly ₹167–170/litre in mid-2026, around 60% above regular petrol — but ⚠️ no official price is published by the oil company or the petroleum planning cell, so we print it as reported and not as a rate. Contested It is available at about 142 outlets nationwide, so availability binds harder than price. And ⚠️ the claim that premium grades are ethanol-free is not confirmed — one company discloses ethanol in its own premium grades. Contested
The real grievance is not the blend. It is the absence of disclosure. We could identify no requirement to display blend percentage at the dispenser, and no mandated E0/E10 option — a Supreme Court petition seeking exactly that was dismissed in September 2025. The stickers seen at some pumps are voluntary. Documented
Mandate blend disclosure at the dispenser — the same instrument that mandated E20 can mandate a label; the US mandates E15 labelling down to the typography, and Brazil publishes pump-level disclosure. Put compatibility on the fuel flap at registration, plus a registration-number lookup: the three regulatory cohorts already exist, they are simply not visible to the owner. Publish machine-readable OEM compatibility lists — and say who pays for that once-in-a-lifetime gasket change. And use the existing proviso allowing regional lower-blend sale to put one E10 outlet in each district, rather than a ₹170 premium grade at 142 pumps.
⚠️ One gap worth naming: no published Indian calculation nets the 2–6% efficiency loss against the claimed foreign-exchange saving. That arithmetic is simply missing from the public record — and it is the single largest fixable trust deficit in this file. Also: the viral "10,000 litres of water per litre of ethanol" is Unsupported; the government's own roadmap gives about 2,860 litres, which is damning enough and is their number.
Open defecation fell from 39% of households to 19% between the two national family health surveys. That is a large, real, fast change. Documented
But the finding that matters is what happened inside households that received toilets. Survey research in rural north India found that among households that own a latrine, around 40% had at least one member who still regularly defecated in the open. Contested — and the reason is not laziness. It is that the affordable government pit latrine eventually requires manual pit-emptying, work associated with Dalit castes and untouchability. Households that would have to empty their own pit, or cannot find anyone to, reject the latrine as ritually polluting to keep near the home. Documented
So the structural case is real and incomplete. Infrastructure removed the access barrier, and a large share of the behaviour gap remained — held in place by a caste-linked norm that concrete cannot touch.
⚠️ Honest measurement dispute: the national survey reports only ~2% of latrine-owning households still practising open defecation, against the ~40% above. A randomised experiment showed the estimate depends heavily on whether you ask about the household or about each individual. Both numbers are on this page because neither is settled.
I could find no rigorous evidence that Indians specifically behave more civically abroad than at home. As usually made, the claim is anecdote, and we label it as such. Unsupported
What does exist is sharper. A study of UN diplomats' unpaid parking violations in Manhattan — people with diplomatic immunity, and therefore effectively zero enforcement — found violations correlated strongly with home-country corruption measures. Norms travel with people, which is the opposite of the folk claim. Then New York gained the power to strip diplomatic plates, and violations dropped sharply across all nationalities. Documented
Norms persist across borders, and enforcement changes behaviour regardless of norms. Indians abroad do not become different people. They encounter different enforcement.
So the honest position refuses both comforts: infrastructure and enforcement are necessary; norms are independently real. Neither "Indians are just like that" nor "give them bins and it fixes itself" survives the evidence. And a bin provided but never emptied is worse than no bin — a visibly littered space signals that littering is normal here.
Indore reached 100% door-to-door collection across all 85 wards, with GPS-tracked vehicles on fixed routes and reported full source segregation. The usual telling credits awareness drives. The mechanism is better than that: Documented
Indore built a bio-CNG plant that requires roughly 95% segregation quality to operate at all — and instrumented the chain so that poorly-segregated waste is traced back to its source. Documented
Indore did not achieve segregation by asking. It built something downstream that physically cannot run on bad input, and made failure traceable. The plant is the enforcement mechanism.
Households segregate when they believe segregation will be honoured downstream rather than remixed in the truck — so collection reliability comes first, then enforcement, then campaigns, in that order.
⚠️ And the honest counterweight, because a page that only praised Indore would be doing what we criticise: the Swachh Survekshan rankings rest substantially on self-reported municipal data; Indore's solid-waste record far outruns its liquid-waste record, with much of its sewage in the Kanh and Saraswati; and a 2025 drinking-water contamination incident in the city killed at least fifteen people. Solid-waste excellence did not extend to water safety. Contested
Surat covers 99% of habitable area and 99.5% of its population with sewerage, collects and treats about 1,018 MLD — and reuses 330 MLD of treated wastewater, selling it to textile and diamond industry, agriculture, construction and the metro. That earns the corporation over ₹120 crore a year, with a target of 70% reuse by 2030. Documented
This is the transferable lesson, and it goes directly at the 4,972 MLD of built-but-idle capacity. Surat gave treated water a paying customer. A plant with a revenue stream gets maintained; a plant that is a pure cost centre gets abandoned. Cities without a textile cluster need their own buyer — agriculture, construction, power-plant cooling, municipal non-potable use — but the principle holds.
- Connect households before building plants. Twenty-one "symbolic" plants and Delhi's 30% unsewered population say the network is the binding constraint, not treatment capacity. Sequencing is the cheapest fix available.
- Intercept the top drains first. 84% of Delhi's Yamuna load comes from two drains. Concentrated load means a concentrated solution.
- Give one body the road corridor — an owner, not eleven correspondents — and publish the permit register: cuts granted, by agency, per year.
- Retain the utility survey. If a GPR survey is mandatory before every dig, keep the result as a public map. Stop paying to rediscover the same pipes.
- Ring-fence maintenance as a named budget line, and actually enforce defect liability. 1,590 works went unmaintained inside their liability period; Uttarakhand recovered 4% of the damages it was owed. Pay contractors for measured road condition, not for construction completion.
- Publish real-time monitoring. One audit found state and central boards reporting faecal coliform of 58 and 14,000 for the same parameter. Without trustworthy data nothing else is enforceable.
- Sell the output, and mandate disclosure at the pump. Surat's ₹120 crore a year, and a fuel label that costs the price of a sticker.
None of this requires a new department, and none of it requires better citizens. It requires an obligation that somebody's money or licence actually depends on — which is exactly what was removed in 1856–57, when the people who used the tank stopped being the people responsible for it.
Sources
- Roads & procurement: Parliamentary Standing Committee on Transport, Tourism and Culture, Operation and Maintenance of National Highways (Feb 2024) — the 1% maintenance allocation; CAG, Performance Audit of PMGSY, Report No. 23 of 2016 (29 states); Standing Committee on Rural Development, PMGSY report (2017); General Financial Rules 2017; Ministry of Finance, Manual for Procurement of Works (2025); Indian Academy of Highway Engineers (MoRTH) course materials on performance-based maintenance contracts.
- Road-cutting regimes: NDMC circular No. D/17/AO(W-I)/2021 (19 Jan 2021); Delhi PWD restoration order (2024); BMC Comprehensive Trench Policy (2023) — full text not retrieved; charge schedule unverified.
- Drainage & streets: IRC:SP:42-2014, IRC:SP:50, IRC:37, IRC:86, IRC:103-2022; CAG Report No. 2 of 2021, Management of Storm Water in Bengaluru Urban Area (with ISRO technical support); India Status Report on Road Safety 2024, TRIPC, IIT Delhi.
- Rivers: CPCB, National Inventory of Sewage Treatment Plants (2021, 2020 data), via Rajya Sabha US Q.3741 (03.04.2025); CAG Report No. 2 of 2025, Government of Uttarakhand; CAG performance audit of Namami Gange (2017); CSE, The State of Yamuna (2025); PIB/NMCG budget answers (2025); Surat Municipal Corporation presentation to NMCG (2024).
- Fuel: NITI Aayog, Roadmap for Ethanol Blending in India 2020–25 (2021); MoPNG and PIB statements (2025–26); Gazette S.O. 860(E) (17 Feb 2026); SIAM (Dec 2022); ARAI (2026).
- Civic behaviour: NFHS-4 and NFHS-5; SQUAT survey and r.i.c.e. research on latrine ownership vs use; Fisman & Miguel, "Corruption, Norms, and Legal Enforcement: Evidence from Diplomatic Parking Tickets," Journal of Political Economy (2007); Ostrom, Governing the Commons (1990); Indore case studies (SSRN; Elementa).
- Deliberately absent: the "84% of footpaths fail standards" and "72% of Ahmedabad roads have no footpath" figures (untraceable to any primary document); an official XP100 price (none published); and any claim that common utility ducts have measurably reduced road-cutting in India (no evaluation exists). Every claim is tagged in the Claims Ledger.