- What a yuan move actually carries
- Two prices for one currency — and why we track the second
- The channel nobody formed a view about
- Relative prices, and what China's factories buy
- Three China tiles, two sections, and one sign that is a judgement call
- Why each tile has its own scale, and what the score is measuring
- Weighted in the score, absent from the correlation grid
- Three ways to misread the yuan tile
- Where to look at this
- Common questions
What a yuan move actually carries
A yuan move reaches an Indian portfolio through two channels that have almost nothing to do with each other. One is plumbing: much foreign money arrives inside an emerging-market allocation, so the complex gets traded as a block. The other is real: China's relative prices and its industrial demand.
Neither is a forecast, and this article makes none. The yuan is not a signal that says where anything goes next. It is an input whose path into Indian share prices can be described, and the value of describing it is that you can then tell the days when the path is open from the days when the tile is just moving.
The two channels also disagree with each other about which Indian companies are helped, which is the thing a one-line summary always loses. A cheaper yuan is a harder competitor for an Indian producer and a cheaper supplier to an Indian buyer, in the same move, on the same morning.
This article takes one input apart. Reading the macro signals together covers how the yuan sits alongside the dollar, yields, crude and flows, and why no single tile is built to be read on its own.
Two prices for one currency — and why we track the second
The yuan trades at two prices. Getting this distinction straight is most of getting the tile straight, because the two are not interchangeable and only one of them is free to move.
Inside mainland China, the currency trades as CNY. China runs a capital account with controls on who may move money across the border, and the central bank publishes a daily reference rate around which onshore trading is constrained to a band. That is a managed price by design. Outside the mainland — chiefly in Hong Kong — the same currency trades as CNH, in a market open to non-residents and not bound by that band.
| CNY — onshore | CNH — offshore | |
|---|---|---|
| Where it trades | Mainland China | Mostly Hong Kong |
| Who can trade it | Access governed by capital-account rules | Open to foreign participants without those approvals |
| What constrains the price | A published daily reference rate and a band around it | Supply and demand among offshore holders of the currency |
| What it tells you | Where policy is willing to let the rate sit | Where offshore money is willing to price it |
The gap between the two is itself the information. When the offshore price sits meaningfully weaker than the onshore one, the offshore market is pricing pressure that the onshore band is absorbing. That is a mechanical reading of a spread and nothing more. It carries no claim about what follows: a spread describes where two sets of buyers stand today, and a band that is absorbing pressure is doing exactly the job it was built to do. Reading a gap as a countdown to the onshore rate closing it is adding a forecast the spread never contained.
Our Currencies section carries the offshore rate, as USD / CNH, quoted to three decimal places because its ordinary daily moves are small enough that two would round most of them away. Read it the way every dollar cross on that row is read: the number rising means more yuan per dollar, which is to say a weaker yuan.
The channel nobody formed a view about
Here is the mechanism that makes the yuan an Indian variable rather than a foreign curiosity, and it has nothing to do with trade.
A large share of the foreign equity money that reaches India does not arrive because someone decided about India. It arrives inside a broad emerging-market or Asia-ex-Japan mandate — a fund whose job is to hold the asset class in roughly benchmark proportions. When that fund receives money it buys the basket; when it is redeemed it sells the basket. Indian shares are inside the basket.
So a decision about the emerging-market complex executes as a trade in Indian shares, with no view about India anywhere in it. That is not a market inefficiency to be exasperated by. It is what an index mandate is for — the same mechanical proportionality that makes index funds and ETFs cheap to run is what makes their flows indiscriminate on the way in and on the way out.
China is where this bites. It has long carried one of the largest single-country weights in the standard broad emerging-market benchmarks, with India another. A currency that is read as a barometer of stress in the largest constituent is therefore read as a barometer of the whole asset class, and an allocator reducing exposure to the class sells everything in it. The full path from an allocation decision to an Indian tape print is in foreign and domestic flows, and the dollar's version of the same story is in why foreign flows follow the dollar.
There is a second, opposite version of the block effect, and it is the one most explanations leave out. Within a fixed emerging-market budget, India and China are not allies — they are competitors for the same money. A China that suddenly looks cheap or newly stimulated can pull weight toward itself out of a budget that does not grow, and the funding for that comes from the rest of the complex. Same plumbing, opposite sign.
Which of the two is operating on a given day is not something the tile can tell you, and the honest position is that both are real. What is structural is this: in both versions, Indian prices move for reasons that are entirely about the container India is held in. A domestic investor watching Indian earnings will find nothing in them that explains the day.
Relative prices, and what China's factories buy
The second channel is a real-economy one, and it splits into two mechanisms that are worth keeping apart because they operate at different speeds.
Relative price. A cheaper yuan lowers the foreign-currency price of everything China sells, without any Chinese producer changing a price list. For an Indian company competing with Chinese imports — in chemicals, in electronics assembly, in a long list of engineering goods — the competitor just got cheaper in rupee terms. For an Indian company buying Chinese intermediates, components or capital equipment, the input just got cheaper by the same arithmetic. One currency move, two groups of Indian companies, opposite signs.
That is the trade-off the headline version always drops. There is no reading of a weaker yuan that is bad for Indian industry in general, because Indian industry sits on both sides of the transaction. It is a rotation between them, and which side any particular holding is on is a bottom-up question about where its revenue and its inputs come from — which no macro tile can answer.
The consumer-price version of the same mechanism runs through imported goods prices: cheaper imports are a disinflationary force in the receiving economy, one input among many into the domestic price level described in inflation.
Industrial demand. The second mechanism is about volumes rather than prices. China is the largest single consumer of most industrial metals, so what its construction and manufacturing sectors are doing sets the marginal price of inputs Indian manufacturers buy and of the commodities Indian producers sell.
Here the honest statement about the yuan is narrower than it is usually made. A weaker yuan and softer Chinese industrial demand get talked about as one story — but the currency is not causing the demand. Both are responses to the same domestic conditions, which is a claim about a common cause, not about a transmission channel. Treating the currency as the demand signal is borrowing a conclusion from a variable that is only standing next to it. If the industrial-demand channel is the one you care about, the metals themselves are the closer read. The page carries copper as a global growth proxy rather than a China one, which makes it the nearest thing on the page to this channel and still not a clean measure of it. The commodity path into India, worked through end to end, is in crude and the Indian economy.
Three China tiles, two sections, and one sign that is a judgement call
China arrives on the Macro page in three places, and they are not next to each other — the currency sits in Currencies, the equity proxies in Global Equities. It is worth laying them out together, because what each one contributes to the composite score is quite different.
| Tile | Section | Weight and sign | What green means on it |
|---|---|---|---|
| USD / CNH | Currencies | 0.04, sign −1 | Green when the number falls — a stronger yuan |
| Hang Seng | Global Equities | 0.03, sign +1 | Green when the index rises |
| Shanghai Composite | Global Equities | No weight — displayed, not scored | Falls through to raw direction: green when it rises |
The composite is a weighted average scaled to a fixed range:
score = 100 × Σ(wₓ · cₓ) ÷ Σ(wₓ),
where each contribution c is clamped to the range −1 to +1. Twenty tiles carry a
weight, summing to 1.33, and foreign portfolio flow is folded in at scoring time at 0.12,
giving 21 contributions and a divisor of 1.45 on a day when every feed reports. A feed
that fails drops its weight from the numerator and the denominator both, so 1.45 is the
all-present maximum rather than a constant.
That gives the arithmetic a reader can redo. The most the yuan tile can move the composite is 100 × 0.04 ÷ 1.45, which is about 2.8 points on a ±100 scale. The Hang Seng adds at most 2.1. Together, everything explicitly about China can shift the headline number by under five points, against 13.8 for the dollar and 8.3 for foreign flows. The yuan is deliberately a small vote in a large committee, and if you are looking for it to swing a regime label on its own, it cannot.
Those weights, signs and cut-offs are FNOTrader's design judgement, not measurements. Nobody has established that the yuan is worth 0.04 of anything. The number encodes a considered view — that most of what a yuan move does to India arrives through the dollar and through flows, both of which already carry their own weights, so scoring the yuan heavily would count the same force twice. A different reasonable view would set it higher and would not be wrong.
Now the part that is genuinely awkward, and worth stating rather than smoothing over. The Hang Seng carries sign +1, so a rising Hong Kong index scores as supportive for Indian equities — on the reading that Chinese risk appetite proxies for emerging-market risk appetite generally. But the tile's own “why it matters” text, which the page shows you on request, names the competing channel in the same breath: a strong Hang Seng can pull a fixed emerging-market allocation toward China at India's expense. Both are the block mechanism from the section above, and they run in opposite directions.
A single sign cannot carry both. The model picks the one it judges to dominate and applies it consistently; the tile text preserves the one it did not pick. That is a modelling choice sitting in plain sight, and a reader who knows it is there will read a green Hang Seng differently from a reader who does not.
One convention governs every coloured tile on the page, and it is the most counter-intuitive thing about it: green and red show the effect on Indian equities, not the direction the number moved. A falling USD/CNH shows green, because a stronger yuan is scored as supportive. A falling dollar shows green. A rising USD/JPY also shows green, because a weak yen is the condition under which the yen-funded carry trade stays intact. There is no direction rule to learn — each weighted tile carries a sign we chose, and the sign is the entire content of the colour.
The wrinkle is on the tiles that carry no weight. With no sign to apply, they fall through to raw direction, which is a different rule on the same page. For the Shanghai Composite that lands where a positive sign would have put it anyway, so nothing looks odd. On the unweighted WTI crude tile it does not: cheaper oil is supportive for India, and a falling WTI tile still colours red. Known, and worth knowing before you read a row of colours as a verdict.
Why each tile has its own scale, and what the score is measuring
A 0.4% day in the yuan and a 0.4% day in the VIX are not comparable events. The composite handles that with a per-tile scale: the daily move at which that input's contribution saturates at its maximum.
USD/CNH saturates at a 0.4% daily move — the tightest setting on the page, shared with the rupee. Brent's is 3%, the VIX's is 8%. A managed currency and a fear gauge are being put on one ruler, and the ruler is what makes that possible. The two volatility indices get wide scales for the same reason in reverse: a 3% day in the VIX is noise.
The clamp is the point, and it has a consequence most composite scores hide. Because every input is squashed to the range −1 to +1 before it is weighted, a catastrophic move in one channel scores identically to a merely decisive one. The score measures agreement across channels, not severity within any one of them. A day when eighteen tiles lean the same way modestly outscores a day when the yuan gaps and everything else is flat — by construction, not by accident.
Which is exactly why the stress label cannot be left to the average. The page overrides it on any one of three conditions: the score at or below −35, the VIX up 20% or more in a day, or USD/JPY down 1.2% or more. None of the three is a yuan condition. An override exists to catch the severity the averaging throws away, and the two instruments carrying it were picked on our reading of which fast moves reach everything else quickest — the reasoning behind the yen one is in the yen carry trade. Those thresholds are chosen numbers too, not thresholds anyone measured.
There is one more place the small weight shows up, and it surprises people. Below the score the page writes out its largest few drivers in plain English, ranked by the size of each term — weight multiplied by contribution — keeping the top four and dropping anything under 0.005. Run the arithmetic on the yuan: even fully saturated, its term is 0.04, which a mere 0.12% move in the dollar index matches. A yuan tile at its maximum can be crowded out of the written explanation by a dollar move you would not otherwise have noticed, and a USD/CNH move under roughly 0.05% never earns a line at all. The tile is still in the score. It is simply not in the sentence.
Weighted in the score, absent from the correlation grid
The same page carries a cross-asset correlation section, and the yuan is not in it. That is not an oversight worth fixing quietly — it changes what you can check.
The grid computes Pearson correlation on daily returns between Nifty and a fixed list of ten drivers, over a window you pick: 30, 60 or 90 days. The ten are the dollar index, the US 10-year yield, Brent, USD/JPY, gold, the US VIX, copper, the S&P 500, the Nasdaq and USD/INR. USD/CNH is not among them. Neither is the Hang Seng, and neither is India VIX.
So the yuan channel is scored on the page and cannot be verified on the page. Two things are available instead. Clicking the tile opens its own historical chart, which shows what the series did without relating it to anything. And copper is on the grid — the closest available proxy for the Chinese industrial-demand mechanism described above, which makes it the row to read when that is the channel you are asking about.
Whatever the grid does show, it is worth being precise about what a coefficient is. Pearson on daily returns measures one thing: whether two series' day-to-day moves leaned the same way, in a straight-line sense, across the days in the window. It is silent on which moved first, on whether either caused the other, and on whether a third thing moved both.
A correlation is an observation about a window; the two channels above are mechanisms. The first is a fact about a stretch of history and expires with it. The second is a claim about how the wiring runs, and survives the window ending. When a coefficient flips sign between the 30-day and the 90-day view, the first thing to suspect is the sample rather than the world. Thirty observations can be set by three of them, and the two windows do not even contain the same days. That is a reading, not a proof — wiring does change, and a mechanism that has genuinely stopped operating would look the same on the grid. What settles it is the mechanism, not the coefficient.
The usable discipline is short, and it runs one way only: name the channel first, then check whether the data is consistent with it. A coefficient with no proposed mechanism behind it is a coincidence measured to two decimal places. Where this matters most is on a broad de-risking day, when everything correlates with everything — the shape of those days is in what a risk-off day looks like.
Three ways to misread the yuan tile
Each of these is a specific error rather than general caution, and each is easy to walk into.
1. Reading a USD/CNH move as China news. It is a dollar-quoted cross, so a dollar move shows up in it with no Chinese content whatsoever. The check takes ten seconds: look along the Currencies row. If the dollar index, the yen, the rupee and the yuan all moved the same way and in proportion — with the euro moving the opposite way, because EUR/USD is the one cross on that row quoted with the dollar on the bottom — the dollar did the moving and the yuan tile is reporting it second-hand. That asymmetry is why the euro tile carries the opposite sign to the other four. The China-specific content is whatever is left over after the dollar is accounted for — which is why the dollar index is read before the crosses, not after them.
2. Reading the yuan as a statement about the rupee. They are different currencies with different central banks, different capital-account regimes and different balance-of-payments positions. They can and do lean the same way, because the dollar sits on one side of both quotes and because a broad emerging-market de-risking touches both. They are not the same question, and the one that decides what a foreign holder actually earns on Indian shares is the rupee — worked through in the rupee and your portfolio.
3. Counting one event three times. On a day of broad emerging-market selling, USD/CNH, the Hang Seng and the foreign-flow figure are not three independent pieces of evidence — one allocation decision moves all three. Watching them agree feels like confirmation and is arithmetic. Two tiles pointing the same way tell you more when the mechanisms behind them are separate; here they are the same mechanism observed in three places. The overnight-lead version of this trap, where Asian indices and the Indian open are read as independent, is in how global indices lead the Indian open.
Where to look at this
Everything above is a way of reading, and the reading has to be yours. The mechanical part — holding the series together, on one scale, over comparable windows — is what the Macro page in FNOTrader's Options Analytics app does.
It shows USD/CNH in the Currencies section and the Hong Kong and mainland China indices in Global Equities, each tile coloured by its scored effect on Indian equities rather than by the direction of the number, each carrying a written explanation of its transmission channel on request. Above them sit the weighted composite, its regime label and the few largest drivers written out in words. Below them, a Pearson correlation grid on daily returns against a fixed set of drivers, with a 30, 60 or 90-day window so a relationship can be checked against more than one sample before it is believed.
The weights and signs are ours, and this article states them so they can be argued with — the page names which inputs weigh heaviest in words rather than printing the numbers on the tiles. The correlations are arithmetic on price series and carry the limits described above whoever computes them.
Common questions
What is the difference between CNY and CNH?
They are the same currency at two prices. CNY is the onshore rate, traded inside mainland China, where the central bank publishes a daily reference rate and trading is constrained to a band around it. CNH is the offshore rate, traded chiefly in Hong Kong and open to foreign participants without mainland capital-account approvals, so it is not bound by that band.
Why does the Macro page track the offshore yuan rather than the onshore one?
Because the offshore price is the one that is free to move. The onshore rate tells you where policy is willing to let the currency sit; the offshore rate tells you where money outside the controls is willing to price it. The gap between the two is itself a reading — when the offshore price sits weaker, the offshore market is pricing pressure the onshore band is absorbing.
Why would a yuan move touch Indian shares at all?
Through two channels. The first is index composition: a lot of foreign money reaches India inside broad emerging-market mandates, so a decision about the asset class executes as a trade in Indian shares with no view about India in it. The second is real — China's relative prices set the competition and the input cost for different Indian companies, and its industrial demand sets the marginal price of metals.
Is a weaker yuan bad for Indian companies?
Not uniformly, and treating it as a single lever on the whole market is where the reasoning usually goes wrong. An Indian producer competing with Chinese imports faces a cheaper competitor; an Indian buyer of Chinese components faces a cheaper input. It is a rotation between the two groups rather than a move in the level, and which side a holding sits on is a bottom-up question about where its revenue and its inputs come from.
How much can the yuan move the composite score?
At most about 2.8 points on a scale that runs from −100 to +100, computed as 100 × 0.04 ÷ 1.45, where 1.45 is the sum of every weight on a day when all of them report. The Hang Seng adds at most 2.1. That weight is FNOTrader's design judgement, not a measured constant — the reasoning is that most of what a yuan move does to India already arrives through the dollar and flow tiles, which carry their own weights.
If the USD/CNH tile is green, does that mean the number went up?
No. Colour shows the scored effect on Indian equities, not the direction of the underlying number. USD/CNH carries a negative sign, so the tile shows green when the number falls — fewer yuan per dollar, which is a stronger yuan. There is no direction rule to learn on this page: each weighted tile carries a sign we chose, and unweighted tiles fall through to raw direction, which is a known inconsistency worth knowing about.
Why isn't the yuan on the correlation grid?
The grid runs against a fixed list of ten drivers — the dollar index, the US 10-year, Brent, USD/JPY, gold, the US VIX, copper, the S&P 500, the Nasdaq and USD/INR — and USD/CNH is not one of them, nor is the Hang Seng or India VIX. So the yuan is scored on the page but cannot be checked against Nifty there. Copper is the closest proxy on the grid for the Chinese industrial-demand channel.
Does the yuan feature in the stress override?
No. The page labels a regime Stress on any of three conditions — the composite at or below −35, the VIX up 20% or more in a day, or USD/JPY down 1.2% or more — and none of them involves the yuan. The override exists because every input is clamped before weighting, so the average measures agreement across channels rather than severity within one. Which two instruments get to bypass it is our judgement about which fast moves reach everything else quickest, not a measured ranking.
I hold an India fund, not an emerging-market fund. Why does this reach me?
Because the price you own is set at the margin by whoever is trading, and a large part of the foreign money trading Indian shares is doing so inside an allocation to the asset class. When that allocation is cut, Indian shares are sold whether or not any Indian holder wanted to sell. The mechanism does not require you to hold anything Chinese — it only requires that someone setting the price does.
Continue reading
More in Macro & Markets · App: Options Analytics · Definitions: glossary · Free tools: calculators · All: every article