The single most useful thing to know about Taiwan futures positioning is this: foreign investors’ net open interest in TAIEX futures is almost always negative, and that tells you very little by itself — the signal lives in the change in that position and in whether it agrees or disagrees with what the same investors are doing in the cash market. As of August 12, 2026, foreigners were net short 86,633 TAIEX futures contracts while simultaneously net buying NT$11.02 billion of Taiwanese stocks — a combination that looks contradictory until you understand that most of the short futures book is a hedge on one of the world’s largest foreign-owned equity portfolios.
This guide became topical in mid-August 2026 for a specific reason: foreign institutions were back on the buy side of the cash market — NT$11.02 billion of net buying in the August 12 session alone — and the TAIEX closed at 45,518.07 on August 12, 2026, with the monthly TAIFEX final settlement landing the following Wednesday, August 19. Expiry weeks are exactly when these positioning numbers get quoted most and misread most. What follows is a timeless manual for reading them correctly.
What TAIFEX publishes, and where
The Taiwan Futures Exchange (TAIFEX) discloses, after every trading session, the aggregate futures and options positions of the san da faren (三大法人, the "three major institutional investors"): waizi (外資, foreign investors), touxin (投信, domestic investment trusts, i.e. local mutual funds), and ziyingshang (自營商, dealers — securities firms’ proprietary and hedging desks). The two most-watched numbers are:
- Foreign net open interest in TX — TX is the flagship TAIEX futures contract (notional value of roughly NT$200 per index point, a widely known contract specification). Net open interest (net OI) is simply outstanding long contracts minus outstanding short contracts held by foreigners at the close.
- The options put/call ratio — for TAIEX options (TXO), the exchange publishes both a volume-based ratio (puts traded ÷ calls traded) and an open-interest-based ratio (put OI ÷ call OI), each expressed as a percentage.
Both series are free on the TAIFEX website, updated each afternoon after the session ends. The companion dataset — institutional net buying and selling in the cash equity market — comes from the Taiwan Stock Exchange (TWSE), and reading the two side by side is the whole craft.
Foreign net open interest: level versus change
Why the level is structurally negative
Foreign institutions own an enormous share of Taiwan’s cash market — as of August 2026, foreign ownership stood at 69.17% of TSMC, 55.36% of MediaTek, and around 41% of both Hon Hai and UMC. A portfolio manager holding tens of billions of dollars of Taiwanese equities who wants to trim index exposure without selling stock sells TX futures. Multiply that across hundreds of institutions and you get a persistent structural short base in the futures market that coexists comfortably with long-term bullish equity holdings.
This is the first and most common reading error: treating a large negative foreign net OI as a directional bearish bet. It is better understood as an insurance book. The questions that actually carry information are:
- Is the net short growing or shrinking? The daily net trading volume tells you the direction of travel. A shrinking short (net buying of futures) means hedges are being lifted — a risk-on signal even while the headline level remains deeply negative.
- Is the level extreme relative to its own recent range? Because the structural base drifts with the size of foreign equity holdings, compare today’s net OI to its trailing few months, not to zero.
- Does the futures direction agree with the cash direction? This is covered in the divergence section below.
Worked example: reading the August 12, 2026 print
Here is an actual day’s data and how a Taipei desk would read it, step by step:
| Series (all as of Aug 12, 2026) | Value | Step-by-step read |
|---|---|---|
| Foreign net OI in TX futures | -86,633 contracts | Step 1: large negative level — expected, structural hedge base. No signal yet. |
| Foreign net volume in TX that day | +2,245 contracts | Step 2: they were net buyers of futures on the day. The daily figure alone cannot show whether that was new longs or closed shorts, but the marginal direction is risk-on. |
| Foreign cash-market net flow (TWSE) | NT$+11.02B net buy | Step 3: cash direction also positive. Futures and cash agree. |
| Put/call OI ratio (TXO) | 106.06% | Step 4: slightly more put OI than call OI outstanding — roughly neutral; mildly constructive only under the local convention discussed in the next section. |
Conclusion of the exercise: despite an -86,633 headline that a casual reader might call "massively bearish," the marginal indicators that day pointed the same modestly positive direction — futures being net bought and cash being bought — while the options ratio sat near neutral. The level shouted; the changes whispered; the changes were right to listen to.
The put/call ratio: Taiwan’s counterintuitive convention
In many markets a high put/call ratio is read as fear. In the Taiwan options market, a common local practitioner convention reads the OI-based ratio the opposite way, on the theory that much of the outstanding put open interest is written (sold) by institutional and professional accounts collecting premium — sellers of puts expressing confidence that the index will hold above their strikes. Under that reading, a put/call OI ratio drifting up reflects growing willingness to underwrite downside, while a ratio collapsing well below 100% flags stress. An important caveat before using it: every outstanding option contract has both a buyer and a writer, so the aggregate ratio cannot by itself reveal who initiated the positions or why — rising put OI is equally consistent with bullish premium selling and bearish protection buying, and the headline data are not broken out by investor type. Treat the Taiwan convention as an unverified practitioner heuristic, not something the published numbers can prove.
With that caveat in place, the working orientation goes roughly as follows: readings sustained comfortably above 100% are conventionally described as a supported market, readings sinking well below 100% as deteriorating risk appetite, and the zone around 100% — such as the 106.06% OI-based print of August 12, 2026 — as broadly neutral. These are rules of thumb, not calibrated thresholds. The volume-based ratio (108.38% that same day) is noisier: it captures a single session’s trading, spikes on hedging days, and is best used only to confirm or question a move in the OI-based series.
Three cautions specific to the ratio
- Expiry distortion. Open interest mechanically collapses into monthly final settlement (the third Wednesday) and rebuilds in the new series. Ratio swings in the two or three sessions around settlement are usually roll mechanics, not sentiment.
- Strike composition is invisible. A 106% ratio built on far out-of-the-money puts means something different from one built near the money. The headline number hides this; serious readers check the OI distribution by strike, also published by TAIFEX.
- It aggregates everyone. Unlike the futures data, the headline ratio is not broken out cleanly by investor type, so retail lottery-ticket buying and institutional premium selling are mixed together — which is exactly why the "put writers" interpretation above remains a heuristic rather than a fact.
When cash and futures diverge
The highest-value signal in this entire dataset is disagreement between what foreigners do in stocks and what they do in futures. A simple two-by-two:
| Cash flow | Futures (change in net OI) | Common interpretation |
|---|---|---|
| Net buying | Covering shorts / adding longs | Conviction risk-on. Both channels agree. |
| Net buying | Adding shorts | Hedged buying — accumulating stock but insuring it. Often seen when buying is stock-specific (e.g. concentrated in one heavyweight) rather than a market call. |
| Net selling | Covering shorts | Frequently mechanical: reducing both the portfolio and its hedge, or index-arbitrage unwinds. Weak directional signal. |
| Net selling | Adding shorts | Conviction risk-off. The combination practitioners conventionally treat as the clearest warning configuration. |
Two practical notes. First, judge the futures leg by the change, never the level, for the reasons above. Second, persistence beats magnitude: three consecutive sessions of the same quadrant is far more informative than one large day, because single-day futures flow is contaminated by arbitrage, market-making, and roll activity.
Building a positioning scoreboard
Positioning is best read as a panel, not a single number. A workable daily scoreboard — shown here with the August 12, 2026 values — combines five public series:
| Input | Source | Aug 12, 2026 reading |
|---|---|---|
| Foreign cash-market net flow | TWSE | NT$+11.02B (net buy) |
| Foreign TX net OI (level + daily change) | TAIFEX | -86,633 contracts; day’s net volume +2,245 |
| Put/call OI ratio | TAIFEX | 106.06% |
| Market-wide margin loan balance (rongzi, 融資 — retail buying on broker credit) | TWSE | NT$548.1B, +NT$3.93B on the day |
| SBL short-balance breadth (securities borrowing and lending — the institutional short channel, distinct from retail margin shorts) | TWSE | Balance up in 465 issues, down in 394 |
The margin balance adds the retail-leverage dimension the institutional data misses; SBL breadth shows whether institutional shorting is broadening or narrowing beneath the index. On this particular panel, the institutional flow inputs — cash buying and net buying of futures — leaned constructive, but the other legs did not simply agree: the rising margin balance means more retail leverage in the system (a growing vulnerability, not a bullish vote), and SBL balances rose in more issues than they fell, indicating institutional short exposure was broadening beneath the index. A genuinely mixed panel — which is precisely why positioning should be read as a panel rather than a single headline.
Limits and failure modes
- Hedges, arbitrage, and market-making are all in the same bucket. Foreign net OI aggregates directional bets with index arbitrage against ETFs and baskets, delta hedges from options desks, and pure liquidity provision. There is no clean way to separate them from public data.
- Settlement week is noise week. Into the third-Wednesday final settlement, positions roll to the next month; OI and the put/call ratio both gyrate mechanically. Wait for the new series to build before drawing conclusions.
- TX is not the whole book. Foreigners also trade the mini contract (MTX), single-stock futures, and offshore Taiwan index products; the TX number is the biggest but not the complete picture.
- Positioning is a condition, not a timer. Extreme readings describe vulnerability — a crowded hedge that could unwind violently — but say nothing about when. Use them to size and frame risk, not to time entries.
FAQ
Is a negative foreign net open interest in TAIEX futures bearish?
Not by itself. Foreigners hold a structurally large short futures base as a hedge on very large cash equity holdings (as of August 2026, roughly 69% foreign ownership in TSMC alone). The bearish configuration is a negative position that is growing rapidly while the same investors also sell cash equities — the level alone is close to meaningless.
What is the difference between the OI-based and volume-based put/call ratios?
The OI-based ratio compares outstanding put and call contracts and moves slowly — it is the positioning gauge. The volume-based ratio compares a single session’s trading and is noisy — useful mainly to confirm whether a shift in the OI ratio came with active repositioning. When the two diverge sharply, trust the OI series.
Where do I find these numbers each day?
TAIFEX publishes institutional futures and options positions, the put/call ratios, and OI-by-strike tables on its website after each session, in both Chinese and English. The matching cash-market institutional flows, margin balances, and SBL data are published by TWSE after its close. Both are free and require no registration.
Does the put/call ratio work the same way as in US markets?
By local convention, often the opposite — but treat that convention as an unverified practitioner heuristic. Taiwanese commentary often reads a rising OI-based ratio as confidence, on the theory that put OI is heavily supplied by premium-selling institutions, whereas in the US a rising ratio is usually read as fear. Because every outstanding contract has both a buyer and a writer, the published aggregate ratio cannot itself confirm which side initiated the positions. Always check which convention a commentator is using before acting on a quoted number.
Sources
- TAIFEX — institutional traders’ futures and options positions, put/call ratios: www.taifex.com.tw
- TWSE — institutional investor flows, margin balances, SBL data: www.twse.com.tw/en and openapi.twse.com.tw
- TPEx — OTC-market institutional and margin data: www.tpex.org.tw
- MOPS — company filings: emops.twse.com.tw
Disclaimer: This is an information and analysis publication, not investment advice. See our Methodology for data sources, standards, and our corrections policy.