Last week's selloff did not become panic. Demand around 1,780–1,800 allowed the index to rebuild above 1,800.
AMY MARKET CLOSE · 21 SEPTEMBER 2026Post-FTSE opening reverses.
Post-FTSE opening reverses.
1,800 must be reclaimed.
VN-Index rose more than 11 points early but closed at 1,799.67, down 15.99 points (-0.88%). Post-rebalance demand failed to hold the index; 22 September must confirm active buying and breadth near 1,800–1,815.
AMY VIEWSelective defense · await a 1,800–1,815 recovery
Reduce positions losing their bases and retain relative strength. Add only when breadth, banks and brokers confirm; reassess risk if 1,785 fails on heavier supply.
01
21 September close in 60 seconds
- VN-Index closed at 1,799.67, down 15.99 points (-0.88%) from 18 September. It had risen more than 11 points early on before reversing below 1,800.
- HNX-Index closed at 274.38 (-0.91) and UPCoM-Index at 126.35 (-0.05). Combined turnover across the three venues was about VND18.5tn, more than 30% below Friday's ETF-heavy session; that comparison is distorted by the rebalance.
- Across the market, 263 stocks rose, 402 fell and 882 were unchanged; VN30 had 12 risers, 16 fallers and two unchanged. VIC, VHM, VCB, SSB and HPG weighed; CTG, TCB, ACB, VPB, FPT, BSR and VJC rose.
- Foreign investors sold more than VND670bn net, led by VHM, VIC and FPT; MCH, VPB and CTG attracted buying. Friday's rebalance demand did not persist automatically.
- For 22 September, watch whether VN-Index reclaims 1,800 and 1,810–1,815 with broader participation. A break below 1,785–1,790 on heavier selling calls for de-risking.
Close data: Znews, 21 September 2026. Turnover, breadth and foreign flow refer to the market scope described above.
02
14–18 September: real recovery, contaminated close
VN-Index reached 1,822.77; HoSE recorded 200 advancers versus 114 decliners and nearly VND18tn turnover—the week's highest-quality rebound.
VN30 was pressured in ATC while the underlying market held. The 7.11-point drop was more event-specific than broad distribution.
The market proved demand exists at 1,800; it has not proved sufficient force to clear the peak zone. Next week must shift from event support to active money.
Sources: Weekly data · 18 September rebalance.
03
What changed on 21 September?
01 · 1,800 FAILEDOpening strength reversed into a weak close
The index gained more than 11 points early but ended below 1,800. Friday's support thesis failed on the FTSE effective date; the next session must test whether buyers can reclaim it.
02 · TURNOVER FELLAbout VND18.5tn across three venues
More than 30% below Friday, which contained exceptional ETF orders. The comparison alone cannot prove discretionary capital fled; broadening decliners are the clearer concern.
03 · BANKS SPLITCTG, TCB, ACB and VPB rose; VCB and SSB fell
Selective demand remained, but failed to offset large-cap selling. A durable recovery needs broader bank participation and stability in VIC, VHM and VCB.
04 · FOREIGNERS SOLDMore than VND670bn net selling after rebalance
VHM and VIC saw prominent selling, while MCH, VPB and CTG attracted foreign buying. One day does not define long-term flows, but it disproves the assumption that Friday's buying would automatically repeat.
04
22–25 September · reclaim 1,800 before targeting 1,850
POST-CLOSE THESIS1,800 is now a demand test; Monday's decline warns of risk but alone does not establish a new downtrend.
On Tuesday watch whether 1,800–1,815 is reclaimed, breadth improves, and foreign selling in large caps moderates. These are observation zones rather than automatic entries.
RECOVERY20%
Reclaim 1,815Test 1,830–1,850Confirmation: 1,800 holds on a pullback, banks and brokers recover together and advancing volume broadens.
- Action
- Probe only stock-specific bases after confirmation; do not chase opening gaps. A later independent test is needed at 1,850.
BASE50%
1,785–1,815 rangePost-FTSE supply testConfirmation: turnover stays moderate, the index crosses 1,800 repeatedly without holding and money concentrates in a few banks or oil names.
- Action
- Preserve reserves; judge positions against their own bases and relative strength. An index bounce alone is insufficient to add exposure.
RISK30%
Lose 1,785Watch 1,775, then 1,750–1,760Confirmation: a close below 1,785 on expanding sell turnover and broad declines, especially in large banks and property.
- Action
- Reduce positions that lose their bases; avoid averaging down and leverage. Wait for broad demand before re-entering.
05
After 21 September · select stocks, not baskets
READING CAPITAL FLOWSForeign selling resumed; domestic demand still supported selected banks, BSR and FPT.
Friday's turnover was inflated by ETF orders. Tuesday's credible signal is a stock defending its base without fund orders, accompanied by breadth and active volume.
SELECT BANKS · CTG · TCB · ACB · VPBAttracted buying during a down market
CTG +2.5%, TCB +2.1%, ACB +2.3%, VPB +3.6%. Watch whether they retain at least half their gains and form higher lows on Tuesday; do not extrapolate to VCB, SSB or the entire bank sector.
OIL · BSRBSR +4.5%; check sector breadth
Assess whether other oil names confirm and the oil-price impulse persists. One advancing stock does not yet establish sector leadership; avoid chasing opening gaps.
INDEX DRAGS · VIC · VHM · VCBLarge index impact outweighs gains elsewhere
VIC -2.6%, VHM -4.1%, VCB -1.7%; VHM and VIC also faced foreign net selling. They must stabilize alongside broader participation before 1,800 can be considered repaired.
BROKERS · AWAIT CONFIRMATIONTurnover sensitivity now cuts both ways
Combined turnover fell sharply versus Friday's rebalance and VCK lost 3%. Raise priority for SSI, HCM and VCI only if the group holds bases together and regular matching turnover improves.
LONG-TERM MODEL · FPT · HPG · MWGPreserve the intended investment horizon
FPT +1.8% despite foreign selling, HPG -2.1%, while MWG needs its own signal. A single day's price does not replace 6–24 month business performance or the invalidation conditions below.
Price and foreign-flow data: Znews, 21 September close. Tuesday's conditions are AMY INVEST analysis.
06
Medium–long-term model · 6–24 months
Separate model: 55% equities and 45% cash/short-duration bonds. This portfolio follows earnings and cash-flow conversion—not generic “sector leaders.” Build in tranches; never use this portfolio to warehouse failed short-term trades.
VCB · 15%Credit-cycle upside with an existing provision buffer
What may be underpriced: VCB completed over 63% of its full-year profit plan in H1 2026; CASA was about 32%, while Q1 NPL was only 0.99% and NPL coverage 178%. Faster credit can therefore translate into earnings without taking the same asset-quality risk as high-beta banks.
Re-rating milestones: NIM stops falling for two consecutive quarters; the 1bn-plus share/c.VND10.6tn capital increase completes; credit expands while group-2 loans remain controlled.
Thesis breaks if: group-2 loans and credit cost rise faster than net interest income, CASA falls materially, or delayed capital raising constrains CAR. Do not buy valuation expansion without improving ROE.
FPT · 12%Profit outgrowing revenue is the investable signal
What may be underpriced: H1 2026 revenue rose 12.6% while PBT rose 18.1%, suggesting a better contract mix and productivity. Eight Q1 wins above USD10m each provide evidence of migration from small outsourcing jobs toward larger transformation programs—more useful than a generic “AI” narrative.
Re-rating milestones: double-digit new bookings, further overseas-IT margin expansion, resilient Japan/US demand and conversion of large deals into 2026–2027 revenue.
Thesis breaks if: backlog rises without revenue conversion, PBT trails revenue for two quarters, or staffing/ESOP costs cause EPS to lag profit materially. Accumulate after a base; never pay any P/E merely for the AI label.
GMD · 10%A 2026–2027 capacity gap, not a simple export bet
What may be underpriced: Nam Dinh Vu phase 3 enters a normalized operating year, while Gemalink phase 2 broke ground on 17 April 2026 and should lift total port capacity above 3m TEU annually. Earnings should come from filling new capacity and operating leverage—not just higher tariffs.
Re-rating milestones: sequential volume growth at Nam Dinh Vu 3; Gemalink 2 remains on track for 2027; core port profit outpaces revenue and capex does not impair free cash flow beyond plan.
Thesis breaks if: GMD loses share despite trade growth, capacity fills slowly, Gemalink 2 slips or debt/capex dilutes shareholder returns.
HPG · 10%The Dung Quat 2 absorption test
What may be underpriced: H1 2026 NPAT reached VND15.48tn, or 70% of plan. The near-15m-tonne 2026 steel target—roughly 40% growth—turns the thesis from “steel-price recovery” into whether HPG can sell Dung Quat 2 output while protecting margin as depreciation and interest rise.
Re-rating milestones: HRC/high-grade steel volume rises without an inventory build; gross margin holds at high utilization; import substitution and rail/high-grade steel orders become visible.
Thesis breaks if: capacity outruns demand, inventories and working capital swell, the HRC–ore/coking-coal spread compresses, or imports pressure price before trade remedies work.
MWG · 8%Buy restructuring-to-cash conversion—not the generic “retail leader” label
What may be underpriced: DMX now has an independent listed valuation; H1 2026 revenue reached VND65.28tn (+27%), NPAT VND4.876tn (+73%), net margin 7.5% and SSSG 32% with a stable 3,013-store network. EraBlue erased all accumulated losses: H1 revenue was IDR1.888tn (+92%) and NPAT IDR30bn (+154%); it exceeded 283 stores by July. An Khang delivered four consecutive profitable months and H1 revenue grew 22%.
Re-rating milestones: DMX sustains SSSG and post-listing margins; MWG's DMX stake receives clearer recognition; EraBlue reaches 500 stores in Q1 2027 while new stores still break even within six months; Bach Hoa Xanh and An Khang generate recurring profit and cash rather than subsidy-led growth.
Thesis breaks if: DMX SSSG and margins contract together, EraBlue expansion lengthens store breakeven, or Bach Hoa Xanh/An Khang resume cash burn to buy growth. In that case the subsidiary listing is a one-off valuation event—not a 6–24 month earnings engine.
Company data: DMX H1 2026 · EraBlue · An Khang.
CASH/SHORT BONDS · 45%Capital for mispricing—not idle weight
Deploy 15% when an operating milestone is achieved but price has not reacted; 15% after a successful market support retest; and 15% after quarterly results confirm earnings. Never use reserves to average down after an explicit thesis-break condition appears.
07
Execution discipline · 22 September
Early strength is insufficient after Monday's 11-point opening rise reversed. Watch breadth and buying after 14:00.
Look for defended support, declining sell pressure and a participating sector. Avoid adding broad exposure or leverage while the index struggles below 1,815.
Act on predetermined protection points when a stock loses its latest low on rising sell turnover. Do not convert a failed short-term trade into a long-term holding.
Protect cash and watch 1,775, then 1,750–1,760. Reassess defense if the index recovers 1,815 with banks and brokers.
Updated after the 21 September 2026 close for the 22–25 September strategy. Figures may be rounded. Scenario probabilities are not certainties; this is not personalized advice, a trade recommendation or an SSI house view.