An EV manufacturer operating under PLI faces 5 percent GST on its output and 18 percent on batteries, steel, engineering services, and packaging. The Section 54(3) refund mechanism returns part of what was paid on input goods; nothing on input services or capital goods, by statute and confirmed in the Supreme Court's VKC Footsteps ruling. The Fitment Committee that recommended the rate reduction was not asked to model the ITC accumulation it would create. The company finances the policy intent on its own balance sheet. What does this reveal about how rate decisions are actually processed?
The GST Council meets periodically to rationalise rates. A proposal to reduce the rate on a specific output; an electric vehicle, a solar module, a corrugated box, a food delivery service; is evaluated by the Fitment Committee before it reaches the Council. The Fitment Committee's mandate is specific: assess the revenue impact of the proposed rate change, evaluate whether the reduction aligns with the Council's broader rate rationalisation objectives, and recommend accordingly. What the Fitment Committee does not evaluate, because it is not within its terms of reference, is the ITC architecture the rate change will create downstream.
This is where the institutional mismatch sits. The officer preparing the Fitment Committee's recommendation calculates the revenue forgone from reducing the output rate. The officer does not model the ITC accumulation that will result from the widened distance between input and output rates. The Council approves the reduction. The notification is issued. The company discovers the inversion when it files its first return and realises that the GST it paid on batteries, steel, engineering services, packaging, and technology infrastructure at 18% cannot be fully recovered against output supplies taxed at 5%.
The refund mechanism exists under Section 54(3) of the CGST Act. It permits refund of unutilised ITC arising from inverted duty structure. But the mechanism was designed with a structural limitation that reveals how the system thinks about credit: refund is available only on input goods. Input services and capital goods are excluded. The Supreme Court in VKC Footsteps confirmed this in 2021, holding that Parliament had consciously chosen to restrict IDS refunds to input goods, and that extending the definition to include services would "do violence" to the legislative text. The institutional logic is clear: the refund mechanism was never intended to make the business whole. It was intended to address a specific, narrow category of inversion arising from goods-to-goods rate mismatches.
This creates a structural problem for business models the government is actively promoting. A company that qualifies under PLI for manufacturing may find its GST input credits inverted because the output rate was reduced without adjusting the input rate. The scheme incentivises production; the tax architecture penalises the cost structure that production requires.
The same pattern has played out in renewable energy, where solar modules and wind turbines attract 5% while steel, glass, and engineering services attract 18%. In food processing, where GST 2.0 moved finished goods from 12% to 5% while packaging and cold storage remained at 18%. In textiles and footwear, where raw materials are taxed more than the finished garment. In each case, the rate reduction was a deliberate policy choice. In each case, the ITC architecture was not recalibrated to match.
The institutional pattern is consistent. The GST Council processes rate reductions through the Fitment Committee, which evaluates revenue impact. It processes refund policy through a separate track, governed by Section 54 and Rule 89(5). It processes interpretational disputes through the DGGI and the appellate architecture. Each track operates with institutional competence within its own mandate. None is designed to evaluate the combined effect on the business that sits at the intersection of all three.
For a company navigating this architecture, the practical consequence is specific. A rate reduction on your output is not a cost reduction; it is a restructuring of where the cost sits. The cost moves from your customer's invoice to your electronic credit ledger, where it remains until the refund mechanism, operating within its statutory limitations, returns a portion of it. The 90% provisional refund introduced from October 2025 has improved the speed. It has not changed the scope. Input services and capital goods remain excluded. The working capital remains locked.
The question is not whether the GST Council understands the problem. It does. The GoM on rate rationalisation has acknowledged the inversion. The Council has recommended a simplified two-rate structure of 18% and 5% to minimise future inversions. But the structural fix; allowing ITC refund on input services and capital goods, or designing rate reductions with the full ITC architecture modelled in advance; requires an amendment to Section 54(3) of the CGST Act that no Council meeting has yet recommended. The institutional architecture processes the rate reduction and the refund limitation as two separate decisions, taken in two separate meetings, governed by two separate sections of the Act; the company experiences them as one.