11.11 The rupee: drivers and the 2026 record low

11.11.1Drivers and the balance of payments#

The balance of payments (BoP) identity is: current account + capital account + change in reserves (with sign) = 0. Pressure on the rupee comes from flows that are not absorbed by the RBI.

1. Oil and the current account. India imports close to 90% of the crude it processes (PPAC's monthly Snapshot of India's Oil and Gas Data, https://ppac.gov.in; the September 2026 edition is image-only and the exact figure was not re-read, so check the current edition). The current account deficit (CAD) widens one-for-one with the oil import bill. If $Q$ is net crude imports in barrels a year and $Y$ is nominal GDP in USD, a $\Delta p$ $/bbl rise widens the CAD by $Q\,\Delta p/Y$ as a share of GDP before any demand response. Problem 11.11 asks you to compute this with stated assumptions. Widely cited rules of thumb put the effect at roughly 0.4–0.5% of GDP per $10/bbl, but that figure is not verified here, so derive your own.
2. Portfolio flows. FPI equity and debt flows are the swing factor at weekly horizons (11.12).
3. Rate differential and carry (11.9).
4. RBI intervention. The RBI sells dollars from reserves to slow depreciation and buys in inflow episodes. It also uses forwards (a forward book) and swaps.
5. Reserves set the RBI's capacity. Total reserves reached a record US$785,706 mn as on 4 Sep 2026, after a one-week rise of US$44,903 mn¹. The jump is in the primary data, so it is not a reporting error; its composition shows where it came from: foreign currency assets rose US$47,498 mn while gold fell US$2,594 mn. A 7.9% one-week rise in FCA is too large to be a currency-valuation effect alone, which suggests a large one-off inflow (for example a swap or a large purchase); the RBI release does not give the cause, so do not assume one. By 30 Oct 2026 reserves were US$717,056 mn (FCA US$591,118 mn, gold US$102,591 mn), US$68.7 bn below the September record but US$25.9 bn above end-March 2026². Part of the fall since September may be the reversal of the one-off inflow rather than intervention. Earlier in the year reserves set a record of about $728.5 bn at end-February and fell to about $690 bn by 1 May (M10 · 10.5.2) and about $674 bn in the week of 3 Jul³.

11.11.2The 2026 record low#

The 2026 episode

Date	USDINR	Note
16 Dec 2025	91.03	Record before 20264
27 Mar 2026	94.83	March high on the same series5
20 May 2026	96.82	2026 record low of the rupee6. Intraday and Indian-market closing rates differ slightly (one aggregator's series peaks at about 97.05 on 19 May [S]); a "99.82 in March" figure that circulated in one secondary source is not supported by the H.10 series, whose March maximum is 94.83, and should be disregarded.
8 Oct 2026	96.78	Second peak, after the RBI hike7
5 Nov 2026	94.47	Market rate [S]; FBIL reference rate 94.39968

The FBIL reference rate (published daily by Financial Benchmarks India and reproduced on the RBI homepage and in its Reference Rate Archive, https://www.rbi.org.in/scripts/ReferenceRateArchive.aspx) is the Indian primary benchmark; the H.10 noon buying rate in New York is used above because it has a free downloadable history. The two differ by a few paise on most days.

The rupee gained about 2% in the month to early November as crude fell [V]. Policy responses in 2026 included a special FX window for oil PSUs, an FCNR(B) deposit scheme, a rise in gold import duty from 6% to about 15% in May (to cut the gold import bill, M10), and the October repo hike.

Mechanism in one chain Brent up → oil import bill and CAD up → rupee down → imported inflation (fuel, fertiliser, edible oil) → RBI hawkish → domestic yields up → FPI equity outflows (worse returns in dollars) → Nifty down and the rupee down further. The feedback from FPI selling to the rupee is what makes EM currency crises nonlinear.