An estimated ~$60 billion in cash USD sits outside the banking system, alongside 400–500 tonnes of gold held by households, and $100–200 billion in annual crypto flows - none of it recorded on the State Bank of Vietnam's books. This piece breaks down the four transaction channels, the mechanics of their impact on VND liquidity, how the SBV responds, and six historic currency attacks that help expose today's vulnerabilities in Vietnam's national currency.
Scope: This piece describes Vietnam's unofficial FX ecosystem across four layers - scale, transaction channels, macro impact, and policy response. Figures are compiled from the IMF (2024 Article IV), World Bank, FATF/APG 2022, Chainalysis 2024, Tiger Research 2025, BullionStar/SBMA, ISEAS Fulcrum 2025, and financial press.
Note: Most estimates of the unofficial market (cash USD, gold, informal remittances) carry low precision because, by nature, these transactions go unrecorded. The figures presented are rough, calibrated to published IMF/academic ranges, and may deviate meaningfully from reality. Read this as a topographic map, not an engineering blueprint.
Scale Overview - The Hidden Reservoir
Three numbers side by side make the problem plain: the stock of cash USD held by households is larger than official foreign reserves, gold holdings add nearly half again as much, and the black-market spread is at its widest in over a decade. This is no longer "small-scale black-market activity" - it is a parallel ecosystem.
If IMF/academic estimates hold, cash USD held by households ($60B) + gold ($35–40B) ≈ $95–100B - larger than the SBV's own $82B in reserves. Even if the real figure is only half that, the hidden stockpile is still large enough to move the market if exchange-rate expectations reverse in unison.
Decree 88/2019 bans individuals and businesses from trading foreign currency outside the banking system, but fines of 80–100 million VND (individuals) / 400–500 million VND (organizations) are far too low relative to the 5–10% margin on each transaction. The result: the gold streets of Ha Trung (Hanoi) and Le Thanh Ton (Ho Chi Minh City) still operate openly.
There's another layer: since 2015, interest rates on USD bank deposits have been fixed at 0% to kill the incentive for dollarization. The market's response was rational - people did not sell their USD to banks; instead they switched to holding cash USD at home, turning the anti-dollarization measure into a force that pushes people out of the banking system.
| Channel | Estimated Scale | Type | Source |
|---|---|---|---|
| Official interbank FX transactions | ~$800–850B/year | Flow | SBV/MAS report |
| Official remittances | ~$16B/year | Flow | VietnamPlus 2024 |
| SBV foreign exchange reserves | ~$82B (09/2025) | Stock | CEIC / Trading Economics |
| USD held outside the banking system | ~$60B | Stock | Springer / IMF Staff |
| Gold held by households | ~$35–40B | Stock | BullionStar / SBMA |
| Unofficial remittances | ~$4–10B/year | Flow | Migration Policy Institute |
| Crypto/stablecoin flows through Vietnam | ~$100–200B/year | Flow | Tiger Research 2025 |
SBV foreign exchange reserves ($82.2B, 09/2025) cover only ~2.3 months of imports - below the IMF's recommended 3-month threshold. Peak reserves stood at $109.6B (01/2022), meaning ~$27B has evaporated in ~3.5 years. Estimates of household cash USD holdings (~$60B) and gold (~$35–40B) are hard to verify, but even if the real figures are only half that, total non-VND assets held outside the banking system still exceed the official reserve stockpile. Source: Trading Economics · CEIC Data · The Investor.
Four Channels - Not a Single "Market"
The unofficial FX market isn't a single "market" - it's a cluster of interconnected channels, each serving a different set of needs, and together they form a second FX infrastructure running parallel to the banking system.
The most common channel. People bring cash USD (from remittances, travel, or informal income) to sell at gold shops. Cash VND is paid out on the spot, no paperwork. Rates are usually a few hundred dong per USD higher than banks.
- Customer sells USD → shop pays VND on the spot, off the books
- Shop accumulates USD → resells to buyers in need or to larger dealers
- Ha Trung (Hanoi) and Le Thanh Ton (HCMC) are open, well-known hubs
A trust-based network of brokers on both sides of the border. Money doesn't actually move - only information and a payment obligation do. Fees run 1–2% (versus 3–5% for Western Union), rates are better, and settlement takes just hours.
- An overseas Vietnamese hands $10K to broker A in the US
- A contacts broker B in Vietnam (a trusted counterpart)
- B pays out ~250M VND in cash to the recipient
- A and B settle periodically via transfers, goods, or netting
People holding USD sell to those who need it through referrals, Zalo, or Facebook. Individual trades are small ($500–$50K), but millions of transactions a year add up to a meaningful flow. Common in areas with large overseas-Vietnamese populations (HCMC, the Mekong Delta).
- Parents need USD for a child studying abroad → buy from an acquaintance with remittance income
- Freelancers paid via PayPal/Wise → sell to buyers in need
- Small businesses importing from China → buy USD/CNY from black-market dealers
- Zalo/Facebook groups connect buyers and sellers
The newest channel but the fastest-growing. USDT functions as "digital USD" - transferring value across borders instantly, with no bank involved. Since Vietnam has no licensed crypto exchange with a fiat on/off-ramp, P2P is the core infrastructure.
- Crypto remittances: buy USDT abroad → transfer to a wallet → sell for VND via Binance/Bybit/OKX P2P
- International payments: freelancers and dropshippers receive USDT → sell it on P2P
- Capital flight: buy USDT with VND → send to an overseas wallet → sell for USD - completely bypassing capital controls
- Telegram OTC: handles large trades ($50K–$1M+) off-exchange
Why Do People Choose the Unofficial Channels?
The unofficial channels don't exist because of "poor financial literacy" - they exist because they offer real advantages over the official banking system, on four dimensions at once.
The free market reflects real supply and demand, typically buying USD a few hundred dong higher than banks. USDT on P2P carries a +3.35% premium over the official rate (2024). Sellers of USD benefit immediately.
Premium source: Tiger Research 2025
Banks require paperwork proving purpose (contracts, invoices, plane tickets). The free market and P2P ask nothing - transactions rest on a mutual agreement. This removes a lot of friction for ordinary users.
Banks: visit a counter, get approved, business hours only, wait 1–3 days. Gold shops: instant, cash on the spot. Crypto P2P: 24/7, completed in 15–30 minutes, automated transfers.
Some people want to keep transactions private for reasons of tax, asset declaration, or exceeding transfer limits. Official channels leave a complete trail through the banking system and get reported to the SBV.
Remittances - Official vs. Unofficial
Vietnam ranks in the global top 10 for remittance inflows. The real question is: how much flows through banks where the SBV can buy it into reserves, and how much "disappears" into cash stockpiles?
The IMF/World Bank estimate that unofficial remittances in developing countries add roughly ~35% on top of the official figure. If that holds for Vietnam, the official $16B implies a true total of roughly $21–22B/year, with ~$4–10B taking the "long way round" outside the banking system. Every billion dollars of remittances that bypasses commercial banks equals ~25,000 billion VND of liquidity that should have existed but didn't. Source: Migration Policy Institute - Vietnam Remittances.
Crypto - The Parallel Financial System
Stablecoins are becoming the largest unofficial FX channel in Vietnam. Tiger Research calls this a "parallel financial system" - freelancers, dropshippers, and e-commerce businesses settle directly in USDT, recreating unofficial FX in digital form.
Vietnam has no licensed crypto exchange with a fiat on/off-ramp (as of early 2026). Every VND ↔ crypto conversion has to go through P2P trading - buyers and sellers trade directly, settling via VN bank transfer. Main infrastructure: Binance P2P, Bybit P2P, OKX P2P, and OTC Telegram groups.
Legally: Resolution 05/2025 pilots a regulated crypto market starting late 2025, with ~5 exchanges licensed initially. However, official fiat on/off-ramp infrastructure still isn't fully operational - most of the flow still runs through unregulated P2P.
Macro significance: Every USDT-VND transaction on P2P is effectively an unofficial FX transaction. Money flowing out of the country via crypto bypasses capital controls. Money flowing in doesn't create official foreign currency for the SBV. Both directions weaken the reach of monetary policy. Regulatory source: PwC - Vietnam Crypto Framework Dec 2025.
Impact on VND Liquidity
Every US dollar outside the banking system is a dollar the SBV cannot convert into VND. Below are four specific scenarios - each one a channel through which unofficial foreign currency quietly drains liquidity out of the economy.
#1 Remittances that never enter the banking system Hawala + crypto bypass banks · flow loss Liquidity loss
When remittances travel through Hawala or crypto instead of Western Union/banks: the USD reaches the recipient, but never passes through a commercial bank → the SBV can't buy that USD into reserves → no new VND gets issued into the system.
If $4–10 billion in unofficial remittances/year bypass the banks, the system loses ~100–250 trillion VND in potential liquidity every year (at ~25,000 VND/USD). This isn't money that "disappears" from the economy - it's money the SBV never creates in the first place because the USD never enters the vault.
#2 Record trade surplus, but the USD doesn't stay in Vietnam FDI profit repatriation + Errors & Omissions Fiscal evaporation
USD from exports must be deposited into an account at a VN bank (Decree 70/2014). Businesses cannot withdraw USD cash to sell on the black market. However, the mandatory surrender ratio - the share of USD firms must sell to banks - has fallen from 80% (1998) to 0% today - businesses can hold USD in their accounts without ever converting to VND. The core problem isn't the black market - it's export structure. Source: Decree 70/2014 · Vietnam Briefing - Profit Repatriation 2025.
The core problem: The FDI sector (Samsung, Intel, LG, Nike, Adidas…) accounts for ~73% of total exports and generates 100% of the trade surplus. Domestic businesses actually run a trade deficit (~$17–22 billion/year). When FDI firms earn USD from exports, they legally repatriate profits to their home country - fully permitted under VN law once tax obligations are settled.
| Trade surplus (goods) | +$28B |
| Services (deficit) | −$5 to −$8B |
| Primary income (FDI profit out) | −$22.5B |
| Secondary income (remittances) | +$14B |
| Current account surplus | ~+$25.8B |
Without the −$22.5B in primary income, the surplus would be ~$48B. FDI profit repatriation "cuts" roughly 47% of the surplus. Source: World Bank - Current Account VN.
End of 2023: reserves ~$100B → end of 2024: ~$81B → a $20B drop. The SBV sold ~$9.4B spot in 2024 to defend the exchange rate. Errors & Omissions: 2023 = 4% of GDP (~$17B), 2024 = 6.9% of GDP (~$30B). The IMF classifies E&O as "unrecorded imports and short-term capital outflows" - including FDI transfer pricing, illegal gold imports, crypto flows, and other unofficial channels. Source: IMF - 2024 Article IV.
11/2024: the overnight interbank rate spiked to 6.2–6.5%, forcing emergency SBV injections of $3.3B. 12/2025: interbank rates at 4.6–7.3%, with more than 20 banks raising deposit rates. Loan-to-deposit ratio (LDR) at ~146%. Source: VietnamNet · VietnamNews.
#3 Speculative USD hoarding Self-fulfilling prophecy · expectations spiral Self-fulfilling spiral
When expectations of VND depreciation spread → households and businesses rush to buy and hoard USD (through both banks and the black market) → USD gets "locked" in safes instead of circulating → the market's supply of foreign currency dries up → the exchange rate rises further (exactly as originally expected). The SBV has to sell reserves to meet demand → pulling VND back out of circulation → a double hit to liquidity.
Expectation of VND depreciation → hoard USD → USD grows scarce → VND actually depreciates → even more people hoard USD. With the hidden reservoir already this large (~$95–100B), a small shock is enough to trigger the spiral - which is why "confidence" is the single most important macro asset a developing-country central bank has.
#4 Crypto bypassing capital controls Stablecoin flows · data gap + policy gap Double loss of control
Money flowing out of the country via stablecoins (buy USDT with VND → transfer wallet → sell for USD abroad) bypasses official capital controls entirely. Conversely, money flowing in through crypto also creates no official foreign-currency inflow. Both directions cost the SBV data visibility and the ability to intervene on the actual flow of money.
The SBV loses both data (no visibility into how much money is flowing in/out) and tools (capital controls can't block crypto flows). At $100–200B in crypto flow per year, the scale is too large to dismiss.
SBV Response - Spot, Forward & Administrative Tools
The scale of SBV intervention in 2024–2025 shows just how large the FX pressure has become. Two main market tools - spot selling and forward selling - serve two different purposes, and choosing between them means choosing between "burning real bullets" and "psychological signaling".
Real bullets - immediate impact
The SBV sells USD immediately to commercial banks on the interbank market. Banks pay VND, receive USD within 1–2 business days.
- Reserves: drop immediately. Every USD sold spot is one USD withdrawn from the vault.
- Liquidity: pulls VND out right away → an immediate drop in VND liquidity.
- Upside: USD supply rises instantly, the exchange rate stabilizes immediately.
- Downside: genuinely "burns" reserves. Prolonged pressure means the vault runs dry.
Buying time + psychological signaling
The SBV signs a contract committing to sell USD at a fixed rate on a future date (1–3 months out).
- Reserves: don't drop immediately. But real USD must be delivered at maturity.
- Liquidity: doesn't pull VND out right away (only at maturity).
- Upside: (a) "buys time" for pressure to cool on its own, (b) signals to the market → reduces speculative USD hoarding, (c) preserves reserves on the books.
- Downside: (a) it's "future debt", (b) if pressure doesn't ease, the SBV ends up delivering on forwards and continuing to sell spot → a double squeeze, (c) if the market senses the SBV is just bluffing, the psychological effect stops working.
When reserves are already thin and the SBV wants to keep its "last bullet" in reserve, but still needs to intervene. Forward selling creates a psychological (signaling) effect without spending reserves right away - a "buy time" tactic, betting that by the time the forward matures, market conditions will have improved (the Fed cutting rates, stronger exports). If conditions don't improve, the SBV has to face both the maturing forward and fresh pressure at once. Source: The Investor - SBV sells $1.5B · VietnamPlus - SBV curbs dollarisation.
Decree 88/2019: bans individuals and businesses from buying/selling foreign currency outside the banking system. Fines of 80–100 million VND (individuals), 400–500 million (organizations).
In practice: the fine is low relative to the profit margin on these trades → limited deterrent effect. The major gold streets still operate openly.
Since 2015, the SBV has fixed the interest rate on USD bank deposits at 0%. Purpose: kill the incentive to hold USD in banks, encourage selling USD for VND (which earns interest).
Unintended reaction: people switched to holding physical USD cash outside the banking system instead of selling - growing the hidden reservoir.
Resolution 05/2025: Vietnam begins piloting a regulated crypto market starting late 2025. If successful, it could bring part of the crypto flow into official channels - giving the SBV data visibility and the ability to intervene.
Six Vulnerabilities That Make a Currency an Easy Target
A currency attack (speculative attack) happens when hedge funds or international capital flows realize a central bank lacks the resources to defend the exchange rate. History has left behind a classic checklist of vulnerabilities - worth comparing against VND's current state.
Reserves under 3 months of imports = not enough "ammunition". Once the market knows a central bank is about to run dry, it sells the local currency en masse → creating pressure the bank can't absorb.
Creates a "clear target" for speculators. If the market believes the rate is mispriced, shorting the local currency is close to win-win - either the central bank devalues (big profit) or holds the line (small loss from borrowing costs).
A persistent trade deficit means money keeps flowing out (demand for foreign currency exceeds supply). The central bank has to burn reserves to compensate - a classic sign of crisis.
Businesses/government borrow heavily in short-term USD → when the dollar strengthens, debt-servicing costs spike → sell local currency to buy USD to repay → the exchange rate rises further (a spiral).
Holding a currency that's losing real value (interest rate < inflation) naturally drives money toward foreign currency or other assets. The central bank has to hike rates very aggressively to hold it back - killing the economy in the process.
When most foreign currency circulates outside the banking system, the central bank loses the ability to control real supply and demand. A wide black-market spread signals that control is already failing.
Six Historical Currency Attacks - For Comparison
Vulnerabilities on paper are one thing. What do they look like when the market actually attacks a currency? The six cases below show the common playbooks - and what VND could learn from each one.
'92 Soros vs. the British Pound - "Black Wednesday" Sep 16, 1992 · Quantum Fund made ~$1B in a single day Fixed rate breaks
Background: The UK had joined the European Exchange Rate Mechanism (ERM) - committing to keep GBP within a narrow band against the DEM. But the UK economy was weak (high inflation, rising unemployment) and needed low rates, while Germany was raising rates after reunification. The UK had to hold rates high to defend the exchange rate → suffocating its own economy.
The attack: Soros recognized the Bank of England didn't have enough reserves to defend the rate indefinitely. He borrowed ~$10B GBP and sold it all into the market (shorting). The BoE bought GBP to prop up the price, burning reserves. Soros kept selling. On the afternoon of September 16, 1992, the UK raised rates from 10% to 12%, then 15% - still couldn't save it. At 7pm: the UK announced its withdrawal from the ERM, and GBP fell freely by ~15%.
Fixed exchange rate + a weak economy needing low rates + finite reserves = an impossible trinity laid bare. Soros didn't "create" the crisis - he simply bet on what was already inevitable.
'97 The Asian Financial Crisis Jul 1997 → Jan 1998 · Baht/Won/Rupiah lost 40–80% Domino effect
Background: The "Asian Tigers" grew hot on short-term foreign-currency borrowing, pouring it into real estate and infrastructure. Fixed exchange rates created an illusion of stability → businesses borrowed USD without hedging. Current account deficits persisted, real estate bubbles inflated. FX reserves were thin relative to short-term debt.
How it unfolded: It started in Thailand - the Baht was pegged at 25/USD but was far weaker in reality. Hedge funds shorted the Baht, and the Bank of Thailand burned through nearly all of its $30B in reserves plus $10B in forwards. On July 2, 1997: Thailand floated the Baht, losing 50% of its value within weeks. The domino spread to the Philippines, Malaysia, Indonesia, South Korea. The Won lost ~50%, the Rupiah lost ~80%. South Korea came close to sovereign default, taking a $57B IMF package.
FX reserves near the danger threshold (~2.3 months of imports), a narrow managed float, real estate credit accounting for a large share of the total. Differences: Vietnam runs a stable trade surplus (Thailand ran a deficit), has tighter capital controls (capital can't exit as suddenly), and lower short-term FX debt.
'97 Mahathir vs. Soros - Malaysia's Contrarian Choice Full capital controls · Ringgit pegged at 3.80/USD Capital controls
Background: When the crisis spread to Malaysia, Prime Minister Mahathir Mohamad refused an IMF bailout (unlike Thailand, Korea, and Indonesia). He accused Soros and Western hedge funds of deliberately sabotaging Asian economies.
A different response: Instead of following the IMF playbook (austerity + floating the exchange rate + opening up capital), Mahathir did the complete opposite: (a) imposed full capital controls - banning the transfer of Ringgit abroad, (b) pegged the Ringgit at 3.80/USD, (c) cut rates instead of raising them, (d) pumped money to stimulate domestic demand.
Outcome: Highly controversial at the time, but in hindsight Malaysia recovered faster than countries that followed the IMF. Its economy stabilized sooner than Thailand's or Indonesia's. However, this approach was only viable because Malaysia had a stronger macro foundation to begin with (less FX debt, healthier fiscal position).
Capital controls aren't always bad - used at the right time with sound macro fundamentals behind them, they can stop speculation cold. But if the fundamentals are weak (high bad debt, fiscal deficits), capital controls are just a lid on a pressure cooker - it will eventually blow.
'01 Argentina - The Currency Board Collapses Peso pegged 1:1 to USD from 1991 · $95B default Hard peg breaks
Background: Argentina adopted a Currency Board in 1991 - a hard peg of 1 Peso = 1 USD. It worked at first: inflation fell from 3,000%/year to near zero. But throughout the 90s, public spending rose, government debt ballooned, and productivity failed to keep pace → the Peso became overvalued relative to the real economy.
How it unfolded: In the late 90s, Brazil (Argentina's largest trading partner) devalued the Real → Argentine goods lost competitiveness. Foreign capital began pulling out. The government froze bank accounts ("corralito") - banning withdrawals over $250/week. Riots broke out, the president resigned. January 2002: Argentina abandoned the peg, the Peso lost ~75% of its value. It defaulted on $95B - the largest sovereign default in history at the time.
Hard peg + fiscal deficit + large FX debt + lost competitiveness. When a hard peg is held too long while economic fundamentals deteriorate, collapse is only a matter of time.
'14 Russia - the Ruble & Shock Therapy Dec 16, 2014 "Black Tuesday" · rates 10.5% → 17% overnight Float + shock rate hike
Background: Oil prices collapsed (from $110 to $50/barrel) + Western sanctions after the annexation of Crimea. Russia's economy depends heavily on oil exports → foreign-currency revenue plunged. Russia's central bank had already burned ~$80B in reserves over a few months trying to defend the Ruble, without success.
The response: December 16, 2014 ("Black Tuesday"): the Ruble lost 20% in a single day. Russia's central bank abruptly raised rates from 10.5% to 17% (in the middle of the night). At the same time, it fully floated the Ruble. This extreme measure was shocking but effective - the Ruble stabilized within weeks, and reserves stopped bleeding.
Sometimes "letting go" (floating + a shock rate hike) works better than "gritting it out" (burning reserves to hold the rate). Russia recovered quickly because it accepted short-term pain. By contrast, defending an exchange rate at all costs (as Thailand did in 1997) often leads to a worse collapse.
'21 Turkey - a Central Bank Loses Its Independence 2018–2021 · Lira lost ~80% · inflation peaked at 85% Politicized monetary policy
Background: President Erdogan believed in the contrarian theory that "high interest rates cause inflation" (the opposite of standard economics). He fired three central bank governors in succession over two years because they wanted to raise rates. Turkey's central bank was forced to cut rates from 19% to 14% while inflation ran above 20%.
Consequences: Deeply negative real rates → people rushed to buy USD and gold → the Lira went into freefall. Inflation peaked at 85% (Oct 2022). The central bank secretly sold ~$128B in reserves through state banks to defend the exchange rate (discovered later). Eventually, Erdogan had to reverse course - appointing a new governor, raising rates to 50%.
A central bank loses independence → monetary policy serves politics instead of economics. Once the market loses faith in the central bank, no amount of reserves is enough. This is the "self-inflicted attack" scenario - no Soros required, just bad policy.
The Macro View - What Is the Hidden Reservoir?
The stock of USD + gold held outside the banking system (~$95–100B) is a "hidden reservoir" that can trigger instability when the public buys or sells it in unison. Every USD traded outside the banking system is one the SBV cannot convert into new VND. The larger the unofficial channel grows, the weaker the SBV's ability to regulate liquidity becomes.
The 5–10% black-market spread seen in late 2025 is a concrete signal that the official system is failing to meet demand. Crypto/stablecoins are rapidly becoming the largest unofficial FX channel - both a challenge (bypassing capital controls) and an opportunity: if brought under orderly regulation, part of that flow could be formalized - giving the SBV data while preserving the channel's appeal relative to traditional gold shops.
- Springer - Dollarization and De-dollarization in Vietnam / IMF Staff Documents - estimates $60B USD outside the banking system (~15% of 2022 GDP)
- BullionStar / SBMA - Vietnam's Gold Market - 400–500 tonnes of gold
- ISEAS Fulcrum - Exchange Rate Stability in Vietnam (2025) - 5–10% black-market spread
- VietnamPlus - Remittances $16B (2024)
- VIR - HCMC $10.5B remittance (2025)
- Migration Policy Institute - Vietnam Remittances Profile
- Tiger Research - Stablecoins as Vietnam's Parallel Financial System - 7.8% of remittances via stablecoin, USDT premium +3.35%
- Tiger Research - Vietnam Crypto Market 2025 - $100–200B annual flows
- Chainalysis - 2024 Global Crypto Adoption Index - Vietnam ranked 5th
- PwC - Vietnam Crypto Framework (Dec 2025) - Resolution 05/2025
- The Investor - SBV sells $1.5B to stabilize the exchange rate
- CafeF - SBV sells $2.8B forward
- VietnamPlus - SBV curbs dollarisation
- USTR - Section 301 Currency Report Vietnam (2021)
- IMF - Vietnam 2024 Article IV Consultation
- World Bank - Current Account VN · Net Primary Income VN
- Trading Economics / CEIC / The Investor - FX Reserves
- Viet Tonkin - FDI Insights 2023–2024 · VietnamPlus - Trade surplus 2024
- Tuoi Tre - Samsung VN profit 2024
- Decree 70/2014 · Vietnam Briefing - Profit Repatriation 2025
- VietnamNet - Liquidity pressure · VietnamNews - Rates exceed 7%
03 Discussion
Leave a note
A considered space for questions, counterpoints, and useful additions. Civil, on-topic, signed.
Reader notes
...Loading notes...