Input Tax Credit under GST: A Compliance-Workload Study of a Modelled Mid-Size Manufacturing Company

Input tax credit (ITC) is the mechanism that makes GST a tax on value added, but it is conditional credit: Section 16 of the CGST Act attaches conditions — possession of a tax invoice, receipt of goods, tax paid to the government, return filed — Section 16(2)(aa) requires the supplier's invoice to appear in the buyer's GSTR-2B, Rule 36(4) caps availment at the credit shown in that auto-drafted statement, Rule 37 forces reversal when payment to the supplier crosses 180 days, and Section 17(5) blocks certain credits outright. Each condition turns the buyer's accountant into a monitor of other people's filing discipline.

This M.Com project quantifies what that monitoring takes. A modelled mid-size manufacturing company (a fictitious firm, Vaigai Precision Components Private Limited, turnover ₹25 crore, hence within the e-invoicing threshold) is simulated for twelve months (FY 2024–25): about 720 purchase invoices from 36 suppliers in three compliance archetypes — prompt GSTR-1 filers, delayed filers and composition/unregistered vendors — and 480 sales invoices with payment lags that push some past 180 days. The conditions are coded as a rule engine in Excel, applied invoice by invoice, and three result families are measured: ITC deferred or lost by cause, monthly reconciliation workload in man-hours under a manual process versus a template-assisted process, and the association between supplier archetype and mismatch type (chi-square). It uses only synthetic data and public law, fits the University of Madras Semester 4 Project format, and produces a workbook the examiner can interrogate line by line.

A mid-size manufacturer earns input tax credit of several lakh rupees a month, but the credit it may actually take depends on thirty-odd suppliers' filing behaviour and on its own payment discipline. Under Rule 36(4) read with Section 16(2)(aa), invoices that do not appear in the buyer's GSTR-2B cannot be availed; under Rule 37, payments outstanding beyond 180 days trigger reversal with re-credit only on payment; under Section 17(5), some inputs are simply ineligible. The accounts office must therefore reconcile the purchase register with GSTR-2B every month, chase suppliers, track ageing of supplier payments and maintain reversal-and-reclaim records — all before GSTR-3B is filed. Firms do this work, but its scale, its cost in man-hours and the value of credit it protects are rarely measured, so managements treat compliance staffing as overhead rather than as protection of working capital. The problem addressed by this study is to quantify, for a fully specified modelled company, the ITC deferred, reversed or lost under each statutory condition, and the workload of the reconciliation that protects the rest, and to test whether supplier filing behaviour explains the pattern of mismatches.

Abstract

This study simulates the input-tax-credit compliance cycle of a modelled mid-size manufacturing company over twelve months to measure how the statutory conditions in the CGST Act and Rules translate into deferred credit, reversals and accounting workload. A synthetic invoice population with three supplier-compliance archetypes is generated, a rule engine implementing Section 16 conditions, Rule 36(4), Rule 37 and Section 17(5) is built in Excel, and outcomes are measured invoice by invoice. Reconciliation workload is costed in man-hours for manual and template-assisted processes, and the association between supplier archetype and mismatch class is tested with chi-square.

Introduction

GST replaced a web of central and state levies with a credit chain: each registered buyer can deduct the tax charged on its inputs, so the tax is intended to fall finally on consumption. The credit chain, however, is only as strong as the weakest supplier's filing. Since January 2022 the buyer's availment is confined to the invoices that appear in GSTR-2B, the static statement generated from suppliers' GSTR-1/IFF filings, and Rule 36(4) caps availment accordingly. A supplier who files late, under-reports, or drifts into composition silently transfers cost to the buyer — first in cash flow, then in interest and, ultimately, in unwound credit.

Review of literature (summary)

The compliance-burden literature on GST in India is mostly of two kinds: survey-based studies of taxpayers' perceptions, and doctrinal commentaries on Section 16 case law. Industry bodies and professional institutes periodically estimate compliance hours, but at the level of broad averages. Gap: few studies work at invoice level with the actual rule text, showing precisely how each condition produces deferred or lost credit and what the corresponding reconciliation activities cost in time. A rule-coded simulation of a single, fully specified firm makes the mechanism visible and verifiable.

Existing approach vs proposed approach

• Existing practice in projects: descriptive write-ups of GST procedures, or questionnaire studies of awareness and difficulties, without quantified outcomes.
• Proposed: an invoice-level simulation in which every availed, deferred, reversed and blocked rupee of credit is traceable to a coded rule and a synthetic source document, plus a workload model comparing two ways of running the monthly reconciliation.

Feasibility

• Data: entirely synthetic; the invoice generator and archetypes are built by the student from stated assumptions, so no real company's confidential data is used or needed.
• Law: every rule is public — the CGST Act and Rules on official portals — and each engine row cites its section or rule number.
• Tools: Excel and Tally available in any commerce lab; SPSS or Jamovi for the chi-square test.
• Time: sixteen weeks within Semester 4, including the three internal presentations.

In scope

• One modelled company (fictitious), twelve months of synthetic transactions, regular-registered purchase and supply transactions in one state, with the rule engine limited to the named CGST provisions.
• Monthly GSTR-2B-based reconciliation, Rule 37 ageing and reversal, Section 17(5) blocking, GSTR-3B Table 4 mapping, and a full-year GSTR-9-style summary reconciliation.
• Workload costing of two defined processes from stated assumptions; chi-square analysis of archetype versus mismatch.

Out of scope

• Real company or live portal data; no actual GSTR filing.
• Integrated goods and services tax, imports and input service distributor mechanics.
• Litigation and case-law analysis beyond the plain text of the cited provisions.
• Anti-profiteering, refund processes and e-way-bill logistics.

Research design

Quantitative, rule-based simulation using secondary legal material and synthetic primary data; analytical-comparative in the M.Com tradition, with a statistical test of association.

Step 1 — Rule table (weeks 1–3)

Each condition is written as an IF-THEN rule with its citation: R1 Section 16(2)(a) — valid tax invoice; R2 Section 16(2)(b) — goods received; R3 Section 16(2)(aa) — invoice in GSTR-2B; R4 Rule 36(4) — availment capped at GSTR-2B credit; R5 Section 16(4) — claim by 30 November following the year-end; R6 Rule 37 — reversal if supplier unpaid after 180 days, re-credit on payment; R7 Section 17(5) — blocked credits (motor vehicles, club membership, works-contract immovable property inputs). The table states each rule's data requirement, so the invoice fields fall out of the law itself.

Step 2 — Modelled company and invoice population (weeks 4–6)

The firm: fictitious Vaigai Precision Components Private Limited, Tamil Nadu, aggregate turnover ₹25 crore (above the ₹5 crore e-invoicing threshold), monthly purchases ₹1.6 crore at 18 per cent GST. Thirty-six suppliers are assigned archetypes: A (24 suppliers, GSTR-1 filed within the 11th of the following month), B (8 suppliers, GSTR-1 delayed 1–8 weeks at random), C (4 suppliers, composition or unregistered, with no credit on those invoices). Purchase lines carry date, supplier, HSN block, taxable value, GST rate, payment date (drawn from a lag distribution calibrated so roughly 3 per cent of payments cross 180 days), and a flag for blocked-credit items. Sales lines (about 480, all e-invoiced with IRN reference) carry similar fields. Ten per cent of the population is regenerated and re-checked for internal consistency.

Step 3 — GSTR-2B simulation and rule engine (weeks 7–9)

Each month, supplier A and B invoices that the archetype says were filed enter the simulated GSTR-2B; B invoices filed after the statement's cut-off slip to the next month. The Excel rule engine evaluates R1–R7 per invoice and stamps a status: availed, deferred (next 2B), reversed (Rule 37), blocked, or lapsed (Section 16(4)). Deferred credit is tracked until it enters a later 2B or lapses at the time limit.

Step 4 — Workload model (weeks 10–11)

Six reconciliation activities (download and map 2B; line match; supplier follow-up; Rule 37 ageing tracker; reversal-and-reclaim entries; 3B preparation) carry time estimates per 100 invoices from published benchmarking and the student's own pilot run in the college lab; these assumptions are stated and later sensitivity-tested. Two scenarios are costed monthly: S1 manual matching, S2 template-assisted matching with a keyed supplier master and exception list.

Step 5 — Analysis and testing (weeks 12–13)

Descriptives: value of ITC by status by month; cause-wise deferred credit. The chi-square test (α = 0.05, in SPSS/Jamovi) examines H1₀: supplier archetype is not associated with mismatch class (missing-in-2B versus amount-verified). A sensitivity table re-runs the workload model at ±25 per cent time estimates.

Timeline (16 weeks)

Weeks 1–3 rule table and review · 4–6 population build · 7–9 engine and 2B simulation · 10–11 workload model · presentations 1–3 at weeks 4, 9 and 13 · 12–14 analysis and writing · 15–16 corrections, binding, viva.

The study is a pipeline from law to numbers: every figure in the results sheet is traceable to a rule row and a synthetic invoice.

flowchart TD
  A["CGST Act and Rules on official portals"] --> B["Rule table R1-R7 with citations"]
  B --> C["Data dictionary: invoice fields from the rules"]
  C --> D["Invoice population: 720 purchases, 480 sales, 12 months"]
  D --> E["Supplier archetypes A, B, C"]
  E --> F["Simulated monthly GSTR-2B statements"]
  D --> G["Tally purchase and sales registers"]
  F --> H["Excel rule engine R1-R7 per invoice"]
  G --> H
  H --> I{"Status per invoice"}
  I -->|availed| J["ITC availed by month"]
  I -->|deferred or lapsed| K["Deferred and lost credit by cause"]
  I -->|reversed Rule 37| L["Reversal and re-credit tracker"]
  I -->|"blocked s.17(5)"| M["Blocked-credit register"]
  H --> N["Workload model: S1 manual vs S2 template, man-hours"]
  J --> O["Cross-tab and chi-square: archetype x mismatch"]
  K --> O
  O --> P["Findings and recommendations"]

Excel workbook design

Sheet 1 Assumptions (firm profile, archetype probabilities, payment-lag distribution, time estimates) · Sheet 2 Suppliers (archetype, filing behaviour) · Sheet 3 Purchases (one row per invoice, 12 months) · Sheet 4 Sales · Sheet 5 GSTR-2B-Sim (monthly statements generated from archetypes) · Sheet 6 RuleEngine (R1–R7 columns, one status per invoice, each header citing its provision) · Sheet 7 ITC-Ledger (availed, deferred, reversed, blocked, lapsed by month) · Sheet 8 Workload (activity × volume × minutes, two scenarios) · Sheet 9 Dashboard (monthly charts and the chi-square contingency table export).

Rule-engine extract (illustrative)

• Extend the engine to multi-state supplies, imports and input-service distribution, where credit flows through additional forms.
• Add an interest module under Section 50 on wrongly availed credit, valuing the cost of aggressive availment.
• Build the GSTR-9 and GSTR-9C annual reconciliation layer and measure the year-end catch-up workload.
• Convert the archetype behaviour into a Monte Carlo model to produce confidence bands for deferred credit rather than single figures.
• Survey a small panel of real accountants (anonymously) to validate the activity-time assumptions against practice.