THE POLICY LAB (TPL) - 57
JANUARY 03, 2025

By J B Mohapatra

A: ACCURACY, completeness, and consistency being the essential attributes of datasets generated and stored across many systems of many departments of the government, it is essential that decisions affecting the members of the public rendered on the back of those datasets are rendered after ensuring comprehensive analysis including cross matching the information across the relevant datasets to ensure fairness in regulatory actions at an individual level and algorithmic fairness at system level. At times, datasets generated by an agency or a department is comprehensive enough to facilitate informed decision making without aid or assistance of any other datasets of any other agency, while at other times a specific dataset generated by a department or an agency needs a corroborative affirmation or negation based on the datasets generated by another agency or a department. Take datasets of Election Commission of India (ECI) as an example: a mismatch between Form 17C (filled in by the Presiding Officer at the close of polls) and Form 20 filled in by the Returning Officer on the counting day- both forms generated by ECI- can be explained on the basis of information generated within and by ECI without the necessity for any further dataset of any other agency. But for ECI to conduct a complaint based verification for ascertaining claims of falsity of an election affidavit, owing to inadequacy of its internally generated datasets ECI necessarily falls back on inputs of other stakeholder agencies for reaching at a resolution of the complaint.

B: The 2 central tax departments (GST and Income Tax Department- ITD) follow the same strategy: cross verifying datasets whether internally generated or in tandem with datasets generated by official 3 rd party sources for detecting inconsistencies in taxpayers' submissions. Mechanism in GST set-up mainly relies on internally generated datasets from the taxpayers' periodic filings, whereas in ITD, not just the taxpayers' return filings, but many additional layers of information, for example the TDS returns from payers u/s 200 of Income Tax Act (ITA) or statement of financial transactions culled from submissions of reporting entities u/s 285BA of ITA are integrated for detecting any anomalous submission in the tax returns.

C: While one would be more inclined to believe that decisions rendered on the basis of datasets internally generated by a department or an agency without the necessity for corroboration by another agency or department would be less problematic and more definitive qua decisions rendered when datasets of many agencies and departments are required to be integrated, this assumption is not always true. Take GST as an example.

D: Reportedly for FY 23-24, approximately 1.13 lakh GST show-cause notices were issued to tax payers by field formations under the central tax administration of which majority of notices pertained to ‘data discrepancies ' in tax return filings and merely 0.5% were in relation to classification disputes. Considering that the GST's central tax administration issued 1.13 lac show cause notices in 2023-24 alone, and assuming an equal number of notices issued by the states, for all the years beginning from a more mature stage of GST, say 2020-2021, it would be no surprise if one were to find that close to 10 lacs show cause notices have been issued since the advent of GST.

E: ‘Data discrepancies' giving rise to GST notices are information mismatches between and among the taxpayers' filings to the Department. For example, cross matching GSTR 1(the monthly return of outward supplies) and GSTR 3B (monthly summary return) may throw up mismatch in tax liability with taxes paid. Reasons could vary from manipulation of data to genuine cases of mis-recording of supplies, non-reporting of nil rated or exempted supplies or export sales in GSTR 1 or incorrect reporting of the nature of tax itself. Inconsistency in cross matching of GSTR 3B with GSTR 2A (a system generated and auto-populated utility providing a dynamic and real time view of ITC based on supplier's submissions) may be owing to manipulation of data at the registered person or supplier level, but may also be ascribed to delayed filing of GSTR 1 by the supplier or inaccurate reporting by the supplier or failure to correctly report GST paid under the reverse charge mechanism. Discrepancies flagged at the system level are preliminary alerts requiring substantiation of claims by the taxpayers through the mechanism set up by the department.

F: Under the extant law, what should follow at GST is a pre-show cause ‘ intimation of tax ascertained as payable u/s 73(5)/ 74(5) ' issued to the taxpayer in form DRC-01A mentioning inter alia the amount of tax/interest/penalty payable. If it is determined that the inconsistency in data submission is simply erroneous and not tinged with fraud or wilful misrepresentation or suppression of facts, apart from tax due, additionally applicable interest is shown in DRC-01A. But if it were a case of inconsistency in data submission actuated by a fraudulent intent, tax along with interest and penalty are duly reflected in DRC-01A. Taxpayer is required to respond DRC-01A in one of the two ways: (a) pay the demand of tax, interest/penalty by a given date, else face notice u/s 73(1) or 74(1) or (b) make and substantiate through part B of DRC-01A that the liability is not acceptable. In case the taxpayer's version in part B of DRC-01A is found unacceptable, DRC 01 containing summary of show cause notice invoking either section 73 or 74 is issued signalling initiation of a regular proceeding.

In many instances though, field formations resort to DRC-01 at the first instance without going through the process of pre-show cause notice as laid out in the law.

In either case whether DRC 01A or DRC 01 is issued, the process to get at a resolution of the issue/s starts and ends with long and protracted interface between the field officer and the taxpayer- a work method mostly decried on grounds of transparency and efficiency.

G: Similar to GST, the ITD processes its own internally generated datasets anchored on third party information sources, such information obtained in terms of 285BA of ITA and shared with the taxpayer by way of Annual Information Statement (AIS). However unlike in GST, processing of those information is through an enabling legislative provision and an approved Scheme. E Verification Scheme, 2021 notified under powers vested u/s 135A of the ITA coalesces the information seeking powers already vested u/s 133, 133B, 133C, 134 and 135 of ITA and facilitates the utilisation of information stored with DG Systems or made available to the DG Systems by the DG (Intelligence and Criminal Investigation), CIT in charge of CPC- ITR, CIT in charge of CPC-TDS or any other authority, body or person. DG System makes available taxpayer information to CIT-E-Verification, information which was uploaded to the registered account or mobile device of the taxpayer, and in respect of which the taxpayer either did not accept the information or where the taxpayer did not respond within 90 days. CIT E-Verification conducts the initial verification through automated query to the information source for cross checking the fidelity of the particular information. If mismatch persists, then depending on the parameters of a risk management strategy set out by the Board and closure of low risk cases, the remainder of the mismatch cases are examined at the level of DG I&CI, who completes a preliminary verification in accordance with a Board approved process. Thereafter CIT E-Verification cross matches the preliminary verification report of DG I&CI with the return of income and prepares a final verification report. This final verification report is made available to the assessing officer if matter pertains to a pending scrutiny assessment under process. In case there is no pending scrutiny of the case, actions on the final verification report in accordance with provisions of the ITA including issuance of reassessment notices are taken.

H: Differences in the approaches to tackling data mismatch by the 2 central tax departments are hard to miss. One, while the ITD has opted for a more elaborate process to handle data mismatch through a dedicated set-up (most significantly, after affording to the taxpayer an opportunity to cross verify, give feedback, deny or confirm the information displayed in the AIS) through multi -level iteration involving the taxpayer and the data source, GST has chosen a shorter route, that is transmitting the data mismatch cases to the GST field formations for exploring the mode and manner of possible and required remedial action. Two, in majority of data mismatch cases in GST, field formations invariably fall back on the issuance of pre-show cause DRC-01A intimations to elicit taxpayer response or directly issue DRC-01. Three, unlike in ITD's approved SOP which neither necessitates nor enjoins upon its officers to communicate to the taxpayer the likely demand of tax etc at the show cause stage, DRC-01A intimations or DRC-01 notices carry the quantum of additional demand of tax, interest and penalty. Four, ITD does not at the show cause stage in any manner decipher the taxpayer's mental state in submitting an erroneous statement, unlike in GST which segregates data mismatch cases to (a) erroneous filing sans mala fide u/s 73 and (b) filing with an intent to defraud and thereby invoking section 74 at the pre-show cause stage based on ‘available information' with the proper officer. To many, segregation of data mismatch cases in the manner as provided in DRC-01A between bona-fide cases of omission and mala fide attempts to defraud at the very threshold stage of a proceeding is far too premature, going by standards of common prudence and experience.

I: It is essential for a tax system relying extensively on technology, as GST does, to operate with appropriate data mining and precise data analytics to locate discrepancies in taxpayers' filings. However data-mismatch cases are required to be resolved through process driven and iterative SOPs under an enabling legislation, like the procedure adopted by the ITD through its E-Verification Scheme. Pre-show cause demand notices by GST at the ‘red flag' stage does in that context appears too much too soon.

While the GSTN architecture does provide for a procedure alerting the taxpayers about data mismatch, the gap thereafter is lack of a system driven process for their resolution thereby triggering issuance of pre- show cause demand notices on a mass scale. A process similar to what ITD has come up with for handling data mismatch cases through its E-Verification Scheme may provide GST a non-adversarial pathway for dispute resolution and improved compliance.