ISSN (Print) - 0012-9976 | ISSN (Online) - 2349-8846
-A A +A

Issues of Compliance in GST

Reflecting on the CAG’s Report

Sacchidananda Mukherjee ( is with the National Institute of Public Finance and Policy, New Delhi.

Debates on compliance audits of the goods and services tax will have to move beyond mere “invoice matching” to more holistic approaches for data reconciliation that can enable effective tax enforcements, and compliance risk assessment and mitigation.


The Comptroller and Auditor General of India (CAG) has come out with the first assessment report on goods and services tax (GST) for 2018–19 (CAG 2019). The report was tabled before Parliament on 30 July 2019. This is the first assessment of the GST system that has been carried out by any independent institution. There are three broad areas where the assessments have been carried out: GST revenue and return filing trends (Chapter II), information technology (IT) audit of GST network (Chapter III), and compliance audit of GST (Chapter IV). Though this article is mostly restricted to issues related to GST compliance and revenue, yet, before discussing on the main findings of the report in these aspects, it would be important for the readers to know the GST return (GSTR) system and expected interlinkages among returns.

There are two main steps under GST for taxpayers. First, is the origin of tax liability, which is mostly related to outward supplies; but also related to inward supplies in case the supply originates from unregistered entities or composition taxpayers where tax is charged on reverse charge basis. Second, is the tax payment after settlement of available input tax credits (ITC adjustment).

Under the GSTR-1, taxpayers are required to upload invoice-wise details of their outward supplies on a monthly basis, and from there the GSTR-2A is auto-populated for registered recipients of those supplies. In the GSTR-2A, taxpayers get options to accept or reject the available ITC. If any amendment is required in GSTR-2A, the recipients need to approach their suppliers to file the amendment in GSTR-1. After finalising GSTR-2A and adding any additional inward supplies that is not reflected in their GSTR-2A, recipients file the GSTR-2. The details in GSTR-2 are expected to be reflected in the dashboard of suppliers as GSTR-1A. However, the CAG’s report has pointed out that it is not operational as of now.

It was expected that monthly summary return of details of outward supplies (arising of tax liability) and inward supplies (availability of ITC) will be auto-populated from GSTR-1A to GSTR-2 and it will be reflected in GSTR-3. However, in the absence of cross-verification of GSTR-1A and GSTR-1, taxpayers are provisionally required to file GSTR-3B to make tax payment after adjustment of available ITC. Therefore, a part (between GSTR-1 and GSTR-2A) of invoice matching is operational under the present GST system, whereas the other part (GSTR-2 and GSTR-1A) is not yet functional.

GSTR Filing and Revenue

According to the CAG report, GSTR filings revealed declining trend from April to December 2018. This is in contrary to the expectation that GST compliance will improve with the stabilisation of the system.

There is a gap in filing percentage between GSTR-1 (monthly returns on outward supplies) and GSTR-3B (summary self-assessed return).1 Short-filing of GSTR-1 vis-à-vis GSTR-3B has resulted in a full-fledged verification of the ITC claims made through the filing of GSTR-3B. This has also hindered the tax departments to verify the details in GSTR-3B from the invoice-level details given in the GSTR-1 and arrive at the turnover.

The CAG had faced constraints in accessing GST data and therefore the audit was limited. However, in the limited audit, the CAG has identified “serious systemic deficiencies.” First, the Central Board of Indirect Taxes and Customs (CBIC) has identified 50,000 cases where verifications for transitional credit2 claims with respect to central GST (CGST) are pending. However, the CAG has found that the cases have not been efficiently leveraged to identify and reject inadmissible credits.

Second, the systemic deficiencies also point to a lack of coordination between policymaking and implementation. On revenues, the report finds that the growth of indirect taxes slowed down to 5.80% in 2017–18 over 2016–17, while this growth rate was 21.33% during 2016–17. Moreover, during 2017–18, the union government had adopted the finance commission’s tax devolution formula instead of place of supply (POS) rule-based settlement of the integrated GST (IGST) across the states. The CAG report finds that post implementation, there is a 10% decline in the union government’s revenue from taxes subsumed under GST (excluding central excise on petroleum and tobacco) in 2017–18 vis-à-vis 2016–17.

Data Reconciliation

In the presence of gap in compliance (or any difference in filing obligations) between filing of GSTR-1 and GSTR-3B, it is likely that aggregate tax information (such as, turnover, tax liability, available ITC) obtained from GSTR-1 may not match with that derived from GSTR-3B returns. Though the CAG has raised this issue in the report—lacking verification of invoice-level details obtained from GSTR-1 with summary returns filed through GSTR-3B—it fails to identify the expected problems associated with the lack of reconciliation of aggregate information across GST returns and associated revenue thereof. Even for a GSTR (say, GSTR-3B), lack of reconciliation of aggregate information across tables cannot be overruled.

Filing GSTR-3B is mandatory for all taxpayers, unless special provision has allowed them not to do so. For a considerable section of taxpayers (other than the composition taxpayers) quarterly filing of GSTR-1 has been allowed; whereas they have to pay tax (through GSTR-3B) on a monthly basis. It is likely that aggregate turnover and tax liability reported in GSTR-1 may not necessarily match with the same in GSTR-3B. In the absence of invoice-level details available for all taxpayers (from GSTR-1), it is impossible for tax departments to verify the ITC claims in GSTR-3B and derive the turnover of individual taxpayers. Therefore, information on turnover, associated tax liabilities and utilisation of ITC obtained from GSTR-3B are all not necessarily obtained from invoice-level details.

In the presence of differences in tax information (GSTR-1) and tax payment (GSTR-3B), effective tax enforcement will be a challenge in the GST regime. So, in the absence of clean data on tax information, it will be over-ambitious to expect from the tax departments to administer GST efficiently. The experience of previous tax regimes shows that tax departments mostly rely on voluntary compliance to generate revenue. In other words, a large section of taxpayers complies voluntarily. The lack of information for the entire population of taxpayers and all tax-related information may force the tax departments to adopt selective tax enforcement. This may go against the “equity principle of taxation,” provided those selected by the tax departments are not involved in tax evasion. It will be important for the CAG to explore the problems associated with data reconciliations across GST returns and also provide clear guidance on what will be the preferred information to use for policy analysis as long as the data reconciliation problem persists.

IGST Settlement

IGST settlement between the union and the state governments is needed mostly on account of four types of transactions: (i) business-to-business (B2B) interstate transactions; (ii) business-to-consumer (B2C) interstate transactions where the value of the supply (invoice value) is more than ₹2.5 lakh; (iii) B2C interstate transactions where the value of the supply is less than or equal to ₹2.5 lakh; and (iv) B2B imports.

The first three transactions are related to domestic trade, whereas the last transaction is related to international trade. For the first two transactions, the required information for IGST settlement is captured through the GSTRs—the GST Identification Number (GSTIN) of exporting and importing taxpayers for the first type of transactions, and the POS information for the second type of transactions. None of the present format of GSTRs can capture the information of the destination state for the third type of transactions. It is expected that the revised GSTR formats will capture this information.

In the absence of POS information related to the third type of transactions, ad-hoc settlement of IGST will continue. The information related to IGST payment on account of imports is captured in the Indian Customs Electronic Data Interchange System (ICES) through the GSTIN of importers. Since GSTIN is state-specific, identification of destination state may not be a problem for imports. It is expected that the IGST paid by inbound passengers for importing goods for personal use above the prescribed baggage rule will be accrued to the state where they disembark. The integration of ICES data with the GST network (GSTN) is warranted.

The following are the formulae for IGST settlement for union (CGST account) and state (SGST account) governments:

(i) IGST settlement for CGST account = IGST credit utilised to pay CGST liability–CGST credit utilised to pay IGST liability;

(ii) IGST settlement for SGST account = IGST credit utilised to pay SGST liability–SGST credit utilised to pay IGST liability;

(iii) IGST collection = IGST paid in cash for domestic trade + IGST collected from imports (including passengers coming from abroad) + (CGST credit utilised to pay IGST liability + SGST credit utilised to pay IGST liability–IGST credit utilised to pay IGST, CGST and SGST liabilities).

For efficient settlement of IGST across the union and state governments, capturing information from GST returns and aggregating the same require a robust IT system. The CAG has carried out an in-depth IT audit and found several faults in the present system. So far, any over (under)-settlement related to government accounts (either CGST or SGST) may not be a major issue, as the same could be rectified in the next round of settlement. Identification of faults in the present system of IGST settlement for taxpayers, as the CAG report has pointed out, is very important to take corrective measures. However, fraudulent claims of IGST credit by a taxpayer, as the CAG pointed out, show that the system is vulnerable to ITC frauds and it must be rectified as soon as possible. IT system is the backbone of any modern tax administration and therefore strengthening it will not only help in effective tax enforcement but also facilitate identification of tax frauds.

It is expected that the CAG, in the next round of GST audit, will take up an IT system audit to see how effectively the GST information is passed on to the grassroots level tax administrators. For all states—either using GSTN back office or the Application Programming Interface (API) to integrate with the GST system—information collected through GSTRs will be passed on to the central and state tax administrations by the GSTN. After two years of introduction of GST, 16,115 tax officials across 27 states and union territories are using back office so far.3 Therefore, it may not be an exaggeration to say that the back-end process of the IT system for GST administration is not yet deployed for all tax administrators. As a result, some tax administrators are still facing problems to access the GST data. The problem is also associated with the granularity of the information accessible by individual tax administrators for their tax jurisdiction and the portability of the data to regular formats like MS Excel and/or MS Access.

Compliance Audit

The CAG has observed that: “In absence of access to GST data, the conclusions in this chapter on compliance audit were based on limited audits carried out in the field.” Except for the monthly press releases by the Press Information Bureau and occasional answers of questions raised in the Lok Sabha and Rajya Sabha, there is no data on GST revenue collection available in the public domain. The secrecy over GST data, even at the aggregate level, is hampering policy research on GST.

Issues related to transitional credit and the possibility of revenue loss on account of inappropriate credit claims is an area that has been highlighted in the CAG’s report for the first time since the introduction of GST. Perhaps an in-depth investigation may lead to identification of the causes of gap between the CGST and SGST collections. The entire gap cannot be attributed to the sequence of adjustment of IGST credit alone; transitional credit may have played a role there.

The CAG has verified 2,119 transitional credit cases across 47 central GST commissionerates and found 309 instances (15% of the sample cases) of omissions in the verification of transitional credit, which amounted to ₹392.91 core. Even if the CBIC has identified (red-flagged) 50,000 cases where verification is required for transitional credit claims in respect of CGST, the cases were not shared with the CAG.

If we assume that instances of omissions (15%) and average amount involved per case (₹1.27 crore) remain unchanged between sample (2,119 cases) and population (50,000 cases), possible leakage of revenue of ₹9,537 crore on account of transitional credit cannot be an exaggeration. The CAG report has observed that till September 2017, ₹65,000 crore was claimed by the registered persons as CGST transitional credit; whereas the collection of CGST during July and August 2017 was ₹29,296 crore. In the absence of access to required data, CAG has carried out limited audit of transitional credits based on 42 significant observations relating to 39 taxpayers in 13 commissionerate, involving an amount of ₹107.39 crore. The details are in Table 1.

Conservation of information or profiles of taxpayers related to pre-GST regime and mapping of the same with their post-GST profile may help the tax administrations to carry out policy analysis like the long-run sectoral performance assessment of businesses, tax incidence analysis and tax investigation/assessment. It may also help in compliance risk assessment of individual taxpayers and effective tax enforcement.

One issue that has been raised by the CAG and highlighted by media is the lack of invoice matching and the possible scope of ITC frauds thereof. Invoice matching is an effective tool to identify fraudulent claims of ITC. However, it may not be an effective instrument to curb tax evasion on account of under-invoicing. Under-invoicing is a result of collusive tax evasion practices where sellers and purchasers collude and mutually agree to under-report the value of transaction (Mukherjee 2019). Identification of tax evasion on account of under-invoicing requires in-depth data mining and data analytics. Although the system of self-populating of purchases through GSTR-2 may help tax administrations to curb tax evasion through false ITC claims, but not the collusive practices of tax evasion. The mechanism of invoice matching may not be of much help to deter tax evasion on account of pre-empting sales to consumers instead of retailers. This practice erodes the tax base and associated tax revenues.


1 For the period April 2018 to December 2018, average difference in return filing percentage between GSTR-1 and GSTR-3B (by the due date and after due date) is 21% (Table 7,
CAG 2019).

2 Provision for transition of pre-GST era credit of central excise and service tax are contained in Section 140 of the Central Goods and Services Act, 2017. Transitional credit is the Input Tax Credit (ITC) available with taxpayers for adjustment with respect to taxes paid on purchases prior to introduction of GST.

3, viewed on 5 August 2019.


CAG (2019): “Report of the Comptroller and Auditor General of India for the Year Ended March 2018 for Union Government, Department of Revenue, Indirect Taxes—Goods and Services Tax,” Report No 11 of 2019, the Comptroller and Auditor General of India, New Delhi.

Mukherjee, S (2019): “Inter-governmental Fiscal Transfers in the Presence of Revenue Uncertainty: The Case of Goods and Services Tax (GST) in India,” Working Paper No 255, National Institute of Public Finance and Policy (NIPFP), New Delhi, March.

Updated On : 29th Nov, 2019


(-) Hide

EPW looks forward to your comments. Please note that comments are moderated as per our comments policy. They may take some time to appear. A comment, if suitable, may be selected for publication in the Letters pages of EPW.

Back to Top