Key Details
CAG Report No. 1 of 2026 examines Delhi Government revenue administration and selected economic, social and general-sector activities. Although formally covering the year ended 31 March 2023, individual audits cover different periods and include departmental action or responses available in some cases up to 2025.
Audit Area | What CAG Found | Why It Matters |
|---|---|---|
GST administration | E-way bill and return data revealed non-filing, transaction mismatches and other high-value inconsistencies, but scrutiny and follow-up were incomplete; physical verification covered only 0.10% of e-way bills | Available GST data were not consistently converted into enforcement action, creating risks of undetected liabilities, delayed recovery and time-barred tax demands |
Electricity subsidy | Subsidies were broadly distributed rather than socio-economically targeted; benefits reached inactive connections and were higher per connection for consumers above 200 units. Regulatory assets also rose sharply | Weak targeting increases fiscal costs, while accumulating unrecovered power costs create future tariff and fiscal risks |
E-procurement | Processes remained partly manual, while weak validation and audit trails allowed shared PANs, email addresses and IP addresses across competing bids; contract-award and evaluation data were often incomplete | Digitisation alone does not ensure procurement integrity without automated controls, complete records and mechanisms to detect potential conflicts or collusion |
The report also contains individual compliance findings concerning stamp duty, recruitment examinations, public works, hospital contracting, industrial land, provident-fund investment and Delhi Jal Board expenditure.
GST Data Were Not Fully Converted Into Enforcement Action
CAG's audit of Delhi's e-way bill system (2018–19 to 2021–22) and GST oversight found that available data could identify potentially non-compliant transactions, but departmental systems did not consistently convert these red flags into timely scrutiny and enforcement.
Key findings include:
15 taxpayers generated 580 e-way bills covering ₹252.66 crore of supplies and potential tax of ₹31.46 crorewithout filing corresponding GSTR-3B returns.
28 taxpayers whose registrations were cancelled had earlier generated 3,629 e-way bills covering ₹734.75 crore, with a tax implication of ₹99.39 crore. Required final returns had not been filed and pending liabilities had not been assessed.
E-way bills were generated using scrapped, stolen, surrendered or registration-cancelled vehicles, while some invoices were used to generate multiple e-way bills.
Only 63,283 of 6.10 crore e-way bills—around 0.10%—were physically verified.
A separate data analysis identified 487 high-value inconsistencies. The Department responded in 127 cases and initiated remedial action in 91, including show-cause notices covering ₹2,656.65 crore and scrutiny notices covering ₹273.87 crore.
Detailed examination of 70 taxpayers also found 344 compliance deficiencies, with a stated revenue implication of ₹3,071.92 crore and turnover mismatches of ₹3,710.17 crore.
These figures are not established revenue losses or recoverable tax. They represent audit-identified mismatches and potential implications requiring departmental scrutiny, taxpayer response and, where necessary, adjudication.
Electricity Subsidy Was Broadly Distributed Rather Than Targeted
Delhi's electricity subsidy increased from ₹2,405.59 crore in 2019–20 to ₹3,161 crore in 2022–23, accounting for roughly 67–70% of the Government's total subsidy expenditure during this period.
CAG found that support extended to nearly the entire domestic consumer base without assessment of household socio-economic status. The 400-unit eligibility threshold covered around 80% of domestic consumers, but the audit could not establish the basis for selecting it.
The distribution of benefits was also uneven. More than 30 lakh consumers using up to 200 units received an average annual subsidy of about ₹6,000 per connection, while around 16.60 lakh consumers consuming above 200 unitsreceived more than ₹10,000 on average—around 70% more per connection.
Analysis of BRPL data identified another anomaly: approximately ₹42.26 crore in subsidy went to connections recording zero consumption for extended periods between April 2019 and March 2023, including ₹17.81 crore for more than 50,000 connections showing no consumption for over a year.
CAG recommended data-based targeting of beneficiaries, consideration of sanctioned load as an eligibility criterion and examination of Direct Benefit Transfer (DBT). The Power Department argued that DBT was not feasible because subsidy payments were also being adjusted against certain outstanding dues of distribution companies.
Unrecovered Power Costs Create a Longer-Term Fiscal Risk
The audit places subsidy policy within a wider problem of electricity-sector financial sustainability. Tariffs had not been increased after 2014–15, while the gap between recognised distribution costs and tariff recovery contributed to accumulating regulatory assets.
CAG reported that Delhi distribution companies' regulatory assets increased from ₹9,063 crore in 2019–20 to ₹27,200.37 crore as of 31 March 2021. Their eventual liquidation could therefore require substantial tariff increases or additional fiscal support.
The audit consequently raises two connected questions: who should receive electricity subsidies, and how will costs not recovered through present tariffs ultimately be financed?
Digital Tendering Retained Manual Processes and Weak Controls
Delhi's e-procurement platform was intended to reduce human intervention and create a transparent digital trail. CAG found that manual processes remained and important validation controls were missing, creating vulnerabilities in procurement oversight.
Among the audit red flags:
1,185 tenders were floated manually despite the e-procurement framework.
45 tenders worth ₹530.25 crore included competing bidders registered under a common PAN.
In 38 tenders, all bidders shared the same registered email address, raising suspected cover-bidding concerns.
The same IP address was used for 14,527 valid bids across 7,181 tenders worth ₹3,217.44 crore—a conflict-of-interest and collusion risk, though not proof of collusion.
Only 5,363 of around 1.55 lakh tenders (3.47%) contained Award of Contract details, while financial-evaluation information was updated for only around 34%.
The platform also allowed financial bids to be decrypted before technical evaluation was completed in three tenders. Missing logs prevented CAG from determining whether back-end manipulation had occurred.
CAG recommended stronger automated validation, integration of procurement rules into the platform, alerts for shared bidder credentials and vigilance examination of suspected collusion. The Government's response to the principal findings was awaited when the report was finalised.
The Audit Also Identifies Avoidable and Unfruitful Expenditure
Separate audit findings include:
₹29.45 crore in unfruitful expenditure on land for a Gems and Jewellery Park and Fashion Design Hub affected by changing plans over 18 years;
₹5.63 crore in avoidable expenditure from underestimated dewatering quantities;
a potential ₹5.50 crore interest loss from delayed investment of Delhi Transport Corporation employees' provident-fund surplus;
₹1.93 crore in avoidable payment linked to deficiencies in a hospital manpower contract;
₹1.56 crore in avoidable electricity expenditure by the Delhi Jal Board; and
₹1.05 crore in wasteful examination expenditure relating to withdrawn engineering vacancies.
What Is a Regulatory Asset?
Regulatory Asset represents distribution costs recognised by the regulator for recovery through future tariffs rather than immediately from consumers. A growing balance can eventually require higher tariffs, government support or a prolonged repayment mechanism.
Policy Relevance
The report shows that digital systems strengthen governance only when their data are actively used, their controls reflect the applicable rules and audit findings lead to time-bound action.
Data-led tax administration: E-way bills and GST returns generate usable compliance signals, but their value depends on automated matching, risk-based scrutiny and action before statutory deadlines expire.
Better subsidy targeting: Consumption and sanctioned-load data can help distinguish vulnerable consumers from inactive connections and relatively high-benefit households without necessarily removing support from those who need it.
Transparent fiscal choices: Electricity subsidies should be assessed alongside regulatory assets, foregone taxes and surcharges, and the possible effect on future tariffs.
Controls before digitisation: Procurement portals should prevent shared identities, invalid credentials and improper bid sequencing rather than merely recording them after the event.
Track audit resolution: Show-cause notices and departmental acceptance are intermediate steps. Public accountability requires disclosure of adjudication, recovery, system correction and closure of audit observations.
Follow the Full Report Here: Report of the Comptroller and Auditor General of India for the year ended 31 March 2023, Government of NCT of Delhi, Report No. 1 of 2026