Strota/ Research

FII Flows Don't Predict the Nifty — They Chase It. And Since 2020, DIIs Broke the Link

By Strota Editorial · Published 2026-06-16 · How Strota reports

FIIDIINiftysmart moneymarket structure
The reflex that 'FIIs buy, market up; FIIs sell, market down' fails three separate tests. The causality runs the other way, the correlation is weak, and India's SIP-fed domestic buyers have decoupled the large-cap index from foreign flows. Here's what the data supports — and what it doesn't.

Open any market broadcast on a heavy selling day and you will hear the same sentence: foreign investors are pulling money out, so the market is falling. The reverse is offered just as confidently on green days. It is the most repeated cause-and-effect story in Indian markets — "FIIs buy, the Nifty rises; FIIs sell, it falls" — and it is treated as something close to a law of nature.

It is not. Tested properly, the belief fails in three separate ways. The causality mostly runs the other direction. The correlation that does exist is weak and coincident, not predictive. And since around 2020 a structural change in who owns India's market has broken the link for large-caps almost entirely. None of that means FII data is useless — it means it tells you something quite different from what the headline implies.

Test one: which way does the arrow point?

The cleanest way to ask "do flows move the market, or does the market move flows?" is a Granger-causality test — does knowing yesterday's flows help predict today's returns, beyond what past returns already tell you, and vice versa? Indian researchers have run this repeatedly, and the answer is consistent.

The foundational study is Rajesh Chakrabarti and Sandeep Batra's work for ICRIER (Working Paper 109), using daily Sensex data from January 2000 to December 2002. It finds one-way causality running from returns to flows, not flows to returns: the hypothesis that flows do not cause returns could not be rejected, while the reverse — that returns drive flows — was rejected strongly. The coefficient on the prior day's return was large and highly significant, the statistical signature of investors who buy after good news. FIIs, in other words, are positive-feedback or momentum traders. They are reacting to the market, not setting it in motion.

A later daily study covering April 2014 to November 2019 reaches the same place from the other end: past Nifty returns Granger-cause FII flows at several lags, while any reverse effect from flows to returns shows up only weakly, at a single day's lag. And tellingly, the original ICRIER monthly regressions show no significant feedback at all — the lead-lag is a short-horizon, day-to-day phenomenon that washes out over weeks. So even where flows and returns appear linked, it is largely the market pulling the flows along, and only over the very short run.

Test two: the correlation is real, but weak — and it's a coincidence indicator

None of this means FII flows and the Nifty are unrelated. They are positively correlated in direction: days of net foreign buying skew green, days of net selling skew red, and that positive sign holds across daily and monthly windows. That much is well established and not in dispute.

What the evidence does not support is the strength the headline implies. The relationship is modest — explanatory power in the mid-teens percent at best — and, crucially, it is contemporaneous. Flows and returns tend to move together on the same day, which is exactly what you would expect if both are responding to the same news, and if foreign desks are buying into strength as they go. A variable that moves with the market on the same day is a coincidence indicator. It is not a leading signal you can trade ahead of. Treating "FIIs were net buyers today" as a forecast for tomorrow is reading a thermometer as if it were a weather forecast.

This is also why the academic literature pushes back on the popular framing directly. A 2013 study by Bhanu Murthy and Amit Singh set out to test "the popular hypothesis that FIIs dominate the Indian stock market" and concluded the opposite: FIIs are "opportunistic agents who do not cause any fundamental change," and domestic institutions exert an independent influence of their own. That was true even on pre-2020 data. What happened next made it far more true.

Test three: since 2020, domestic money broke the link

The single biggest change in Indian market structure over the last five years is the rise of the systematic investment plan. Monthly SIP contributions have risen roughly seven-fold from under ₹4,000 crore a month in FY17 to north of ₹28,000 crore, touching about ₹32,000 crore in early 2026 across nearly 9.7 crore SIP accounts (AMFI). That is a large, price-insensitive bid that arrives every month regardless of what foreign investors are doing.

The result is that domestic institutional investors have become the market's shock absorber. By the standard ownership data (Motilal Oswal, for the quarter to March 2026), DII ownership of the Nifty 500 reached a record ~20.9%, edging above FIIs at ~17.1% — a crossover that would have been unthinkable a decade ago. Through recent bouts of heavy foreign selling, domestic buying has absorbed the large majority of it (analysts put the figure near 90%; treat that as an estimate, not a precise constant).

The decoupling shows up plainly in price. Through 2025, foreign investors were persistent net sellers, yet the Nifty still rose around 6.5% year-to-date by mid-October. The sharper illustration came in early 2026: after a steep, FII-selling-driven drawdown in March, the Nifty rallied about 7.5% in April 2026 — its biggest monthly gain in over two years — even as foreign investors kept selling tens of thousands of crore. The headline "FIIs are still selling" and the tape "the index is ripping higher" were true at the same time. The old reflex would have had you short into one of the best months in years.

Two honest caveats keep this from becoming a new myth in the opposite direction. First, the absorption is a large-cap, index-level phenomenon; broader mid- and small-caps, where domestic flows are thinner relative to froth, stayed weak through the same stretch. Second, this is not the same as saying "DII flows now drive long-term returns" — that stronger claim is not supported. Domestic money has neutralised foreign selling as a directional driver of the index; it has not turned flow-watching into a buy signal. The SIP figures are also gross, before redemptions and cash buffers, so the true deployable cushion is smaller than the headline run-rate.

So what about the derivatives data?

If cash flows are a coincidence indicator, surely the positioning data — the FII long-short ratio in index futures, the options book — does better? It is richer, but not in the way most retail traders hope.

The most robust empirical finding on FII derivatives is that their positioning leads implied volatility, not price direction. A 2012-2021 study in the Journal of Risk and Financial Management found one-way causality running from FII derivatives activity to India VIX: FII buying pressure helped predict changes in implied volatility, while implied volatility did not predict FII behaviour. In plain terms, when foreign desks are active in the derivatives book, the useful, evidence-backed inference is about the volatility regime — how choppy the next stretch is likely to be — not about which way the index goes.

The popular use of positioning as a contrarian price signal does not survive testing. In August 2025 the FII index-futures long-short ratio collapsed to roughly 0.09 — about ten short positions for every long, the most bearish since March 2023. The retail script says such extremes precede a short-covering rally. Instead, the shorts stayed intact even as the Nifty rose nearly 700 points off its lows. Extreme positioning is a real-time readout of sentiment and exposure; it is not a validated timing signal, and trading it as one is a good way to be early and wrong.

Cash versus derivatives: which is the better signal?

Neither is a price-direction predictor. But they answer different questions, and that is how to use them:

The highest-value read is not any single number but the divergence between layers — when the cash headline and the derivatives book disagree, the book is usually telling the more honest story. That is a structural insight, not a market-timing one.

The regulatory backdrop — and what we still don't know

Two policy shifts frame all of this. SEBI's evolving FPI framework continues to tighten disclosure and position rules for foreign investors, and the 2024-25 derivatives overhaul — larger contract sizes, fewer weekly index expiries, tighter intraday position-limit monitoring — has reshaped who can hold what in the F&O market. What we cannot yet say, honestly, is how that overhaul has changed the informational content of FII positioning: no rigorous post-overhaul study has measured it. Likewise, the strongest causality evidence predates 2020, so whether the "returns lead flows" dynamic still holds unchanged under today's DII-dominated structure is an open question rather than a settled one. Anyone who tells you precisely how the new regime maps to next week's Nifty is guessing.

What to actually do with FII/DII data

The data is genuinely valuable — it is simply valuable for understanding positioning, hedging and volatility, not for the one thing everyone tries to use it for. The traders who lose money on FII data are not the ones who ignore it. They are the ones who read a coincidence indicator as a crystal ball.

Educational and informational content, not investment advice. Figures on ownership, SIP run-rate and positioning are point-in-time and drift; causality results denote statistical predictive precedence, not proof of mechanism. Sources are linked above.

Sources

Strota Research is written by the Strota editorial desk using exchange and public data, drafted with AI assistance and edited by a human before publication. See our editorial standards, sourcing and AI-use disclosure. Found an error? Tell us — we correct transparently.