Rupee puzzle: Why the currency’s exchange rate is not justified by India's economic fundamentals

Yet the rupee has depreciated by over 13% against the dollar in the past two years and by more than 15% since January 2023. If exchange rates are verdicts on fundamentals, India appears to have received an unjust one.

Consider what the scoreboard actually shows.

India’s real GDP growth has remained the fastest among major economies, running well above 6% annually even in a difficult global environment. Inflation has stayed moderate and within the Reserve Bank of India’s tolerance band. The current account deficit has been contained at low levels relative to GDP. External debt as a share of GDP remains among the lowest in the emerging market universe, insulating India from the rollover vulnerabilities that typically trigger currency crises.

The Union government has been on a credible fiscal consolidation path. And three sovereign rating upgrades in this period reflect precisely the kind of institutional confidence that should, in any textbook world, be supporting the currency.

Something else has been going on, and understanding it requires looking beyond India’s borders.

The Korean won offers a clarifying parallel. South Korea runs a structural current account surplus—it earns more foreign exchange from the rest of the world than it spends by any conventional measure—which should provide a durable floor under the won. Yet the won depreciated by more than 17% against the dollar over the same two-year window as the rupee.

The explanation is not to be found in Korean trade competitiveness or monetary mismanagement. It lies in the extraordinary pull that the artificial intelligence (AI) investment supercycle has exerted on global capital flows, pulling money towards dollar assets and away from economies—however fundamentally sound—that are not at the centre of the AI capital expenditure narrative.

The AI investment cycle of 2023-26 has been, by any sober reckoning, one of the most dramatic episodes of narrative-driven capital misallocation in recent financial history.

When a footwear company rebrands itself with an AI adjacency and sees its market capitalization rise by more than 600% in a single day, the market is no longer performing price discovery—it is running a momentum machine decoupled from any recognizable valuation anchor.

The dollars flowing into this vortex came, in part, from emerging market allocations, compressing currencies from Seoul and Jakarta to Mumbai regardless of whether those economies were well or poorly managed. So exchange rates at this juncture reflect capital flows, not fundamentals.

India was not a beneficiary of the AI narrative because India is not pursuing AI through hyper-scale data centre construction and chip-intensive infrastructure investment. This may yet prove to have been the right instinct rather than a missed opportunity.

India has instead approached AI from the demand and application side, investing in frugal innovation and sovereign digital infrastructure. Nobody predicted that India would build world-class digital public infrastructure—Aadhaar for identity, UPI for payments, U-Win for vaccinations across the life cycle, each conceived for a specific purpose and expanded far beyond it.

The quiet intelligence of that pattern—doing more with less, building for scale through ingenuity rather than capital—may well describe India’s AI trajectory too. The world’s assessment of what constitutes a winning AI strategy could look quite different a year from now.

Against this backdrop, some investors have been raising a second-order concern: that AI structurally undermines India’s labour-cost arbitrage in IT-enabled services, while China continues to squeeze India’s manufacturing space. Both arguments carry a kernel of truth. Neither carries much beyond that.

On IT services, the relevant unit of analysis is no longer the back-office operation but the Global Capability Centre (GCC). India today hosts more than 1,800 GCCs, employing hundreds of thousands of professionals who perform sophisticated work in engineering, analytics, product development and research.

Admittedly, roughly half of what GCCs do is commoditized and therefore genuinely exposed to automation. But the other half represents genuine intellectual value-addition — the kind of complex, context-dependent problem-solving that AI is more likely to augment than replace.

A GCC staffed with engineers designing semiconductor architectures or modelling risk for a global bank is not the same thing as a call centre handling insurance claims. Conflating the two understates how far India’s services export model has already evolved and overstates AI’s capacity to substitute it.

On manufacturing, the critique has more texture but requires more precision. India’s merchandise trade deficit is high, and that is a legitimate structural concern. But a significant portion of it is oil-related—India is a large hydrocarbon importer in a world where energy prices remain elevated.

Strip out oil and its non-oil merchandise trade deficit presents a considerably less alarming picture, with India carving out export niches in pharmaceuticals, chemicals, electronics components, defence equipment and engineering goods that were not part of the story even a decade ago.

The deeper point, however, is philosophical rather than sectoral. Did anyone in the early 1990s seriously give India a chance to build a globally competitive IT-enabled services industry that would generate a structural export surplus and employ millions of highly skilled workers? The candid answer is ‘no.’

The linear projections of that era, working from India’s infrastructure gaps, bureaucratic constraints and capital scarcity, would have produced a confident verdict of impossibility. They were wrong, comprehensively and consequentially. Economies and societies evolve in ways that defy extrapolation.

To make linear structural projections about what India can or cannot achieve in manufacturing, or in AI, or in any other domain, is to repeat the intellectual error of every previous generation of India sceptics. Complexity does not counsel optimism or pessimism; it advises humility about forecasts.

Two further data points deserve to anchor this argument. First, data on real effective exchange rates (REERs) from the Bank for International Settlements shows that through much of 2022–24, India’s REER was above China’s—meaning that, on a trade-weighted, inflation-adjusted basis, Indian exports had become less competitive than Chinese ones.

The rupee’s depreciation has corrected this misalignment, restoring a rough parity in relative competitiveness that had been quietly eroding. This is a silver lining that the nominal depreciation story rarely acknowledges.

Second, India’s gross foreign direct investment (FDI) inflows for the financial year 2025-26 reached $88 billion through February alone and the full-year figure is likely to settle between $90 and $95 billion—a decisive break above the $70-80 billion range that had prevailed for four years.

Unlike portfolio flows, which are driven by short-term sentiment and liquidity cycles, FDI reflects long-term, patient capital making irreversible commitments to India’s productive capacity. That number does not describe an economy that has lost the confidence of investors who think in decades rather than quarters.

The rupee’s current level reflects a global flow environment distorted by AI-driven capital concentration in the US, compounded by structural bear narratives about India that do not survive serious scrutiny. The fundamentals are sounder than the exchange rate implies. When the AI cycle turns—as cycles always do—that gap will close.

The author is chief economic advisor to the Government of India.