Korean won

The won's rise came before the rate hold, not after it

Nearly all of the won's gain since its July low came before the Bank of Korea held rates on 22 October, and in January–August the current account brought in 2.8 times as many dollars as foreigners took out of Korean stocks.

datasteg desk · AI-assisted

12 min read · 23 Oct 2026

PDF Cite

Key numbers

97%^{Source 1}
of the won's rise since its 2 Jul low had come by 16 Oct, before the meeting week
−60bp^{Source 2,Source 3}
change in the gap between Korea's 3-year and the US 2-year government bond rates, 2 Jul–21 Oct, all of the net change after late August (negative = against Korea)
2.8×^{Source 4,Source 5}
the dollars the current account brought in, against those foreigners took out of Korean stocks, January–August 2026

Key findings

The won did not speed up after the hold: on decision day it moved 0.08% against the dollar, less than on 20 of the 22 trading days before, and 97% of its rise since the July low had come by 16 Oct⁠^{Source 1}.
On balance the gap between Korean and US market interest rates moved against Korea only from late August, 60bp in all from 2 Jul to 21 Oct⁠^{Source 2,Source 3}. Korea's two hikes had narrowed the policy-rate gap and the Fed's rise widened it again, so the hold left Korea's base rate 100bp below the top of the Fed's range⁠^{Source 6,Source 7}.
Studies of other currencies find that decisions move exchange rates through their surprise, and on Fed and ECB decision days mostly through news about the future path of rates; in 1996 a model whose power came almost wholly from net buying and selling explained more than half of daily moves in the mark-dollar rate⁠^{Source 8,Source 9,Source 10}. Korea's 3-year government bond yield had fallen 30bp from its late-September high before the meeting and moved 1.8bp on decision day, which suggests the hold was largely expected⁠^{Source 2,Source 6}.
This was the won, not a weak dollar: from 2 Jul to 16 Oct a dollar came to cost 13.3% fewer won, while the broad dollar index rose 0.6% and the Taiwan dollar, a fellow chip exporter's currency, weakened⁠^{Source 11,Source 12,Source 13}.
In January–August 2026 the current account brought in USD 279bn, 2.8 times the USD 101bn foreigners took out of Korean stocks, and in the balance of payments foreigners were small net buyers of Korean shares in July–August, though daily exchange data, which we cannot republish, show net selling in every month from July; how much export income was turned into won is not measured⁠^{Source 4,Source 5}.

The question

The question we answered

Why did the won keep strengthening through the Bank of Korea's October 2026 rate hold, though US rates rose further above Korea's and foreigners sold Korean stocks heavily this year, and what part did export dollars and other flows play?

What we looked at

How the won moved around the Bank of Korea's October rate hold and this year's two hikes, whether interest rates or the dollars flowing into and out of Korea explain the move, and what research finds about currencies on central-bank decision days.

A note on the question. At the daily close the won did not rise faster after the decision than before it—nearly all of its gain since July came earlier—and over that rise the gap with US market interest rates moved against the won, not for it, on balance only from late August.

Almost all of the won's rise came before the meeting week

Almost all of the won's gain since its July low came before the meeting week. We measure the won as won per dollar, so a lower number means a stronger won. At the Bank of Korea's afternoon close a dollar cost 1,555.8 won on 2 Jul 2026, the won's weakest of 2026, and 1,324.6 won on 23 Oct, a fall of 14.9%⁠^{Source 1}. On decision day the won closed at its strongest since 2 Oct 2024.

Meanwhile the Bank raised its base rate from 2.50% to 3.00%, on 16 Jul 2026 and 27 Aug 2026⁠^{Source 6}. By Friday 16 Oct, the last trading day before the meeting week, 97% of the won's rise had already happened⁠^{Source 1}. That Friday alone a dollar got 13.2 won cheaper. The chart below shows the year: the won weakened into early July, then rose, most of it by mid-October, with a setback in mid-September.

Figure 1

A dollar cost 1,555.8 won at the won's 2 Jul low and 1,324.6 won on 23 Oct; 97% of that fall had come by 16 Oct, before the meeting week

Won per US dollar in 2026Won per dollar, Bank of Korea 15:30 close, daily, 2 Jan–23 Oct 2026

How to read this chart. Lower means a stronger won: fewer won buy one dollar. Vertical lines mark the Bank of Korea's two rate rises, to 2.75% on 16 Jul and to 3.00% on 27 Aug, and its hold at 3.00% on 22 Oct; the shaded strip is the meeting week, 19–23 Oct.

By the 16 Oct close of 1,331.4 won, the last before the meeting week, 97% of the 231.2-won fall from 2 Jul to 23 Oct had happened.

The Bank of Korea's 15:30 close. Since July 2026, with round-the-clock onshore trading, ECOS also publishes closes on some Korean holidays; every published close is shown. The rate dates are the days the base-rate series shows each change.

Left out: the extended-session close, which tells the same story (in the text), and New York noon rates (in the peer-currency chart below).

Source: Bank of Korea, ECOS · Data to 23 Oct 2026^{Source 1,Source 6}

About this chart

Definition: Won per US dollar, 15:30 close in the Seoul market, as published by the Bank of Korea; Bank of Korea base rate.
Calculation: As published. Share of the fall by 16 Oct = (close 2 Jul − close 16 Oct) ÷ (close 2 Jul − close 23 Oct) × 100.
Provider: Bank of Korea, ECOS
Series: 731Y003/D/0000003, 722Y001/D/0101000
Period: 2 Jan 2026 – 23 Oct 2026
Last observation: 23 Oct 2026
Retrieved: 23 Oct 2026
Licence: ECOS Open API terms: reuse allowed with attribution
Sources: Won per US dollar, 15:30 close, Bank of Korea base rate

Download CSV

The hold moved the won less than an ordinary day

In the meeting week the won kept the pace it had already set, and on decision day it moved less than on almost any day of the month before. Over the four days to Thursday 22 Oct it gained 7.2 won, about 1.8 won a day⁠^{Source 1}. Over the 22 trading days before, it had gained 1.7 won a day on average⁠^{Source 1}. On Friday it slipped back 0.4 won.

In percent, which keeps days at different levels comparable, the decision-day move was 0.08%, smaller than on 20 of those 22 trading days⁠^{Source 1}. It was also smaller than on the two days the Bank raised rates, 0.29% and 0.28%. The chart below compares them with the average day.

Figure 2

On the day of the hold the won moved 0.08%, less than on either hike day and less than on 20 of the 22 trading days in the month before

How far the won moved against the dollar on decision days, and on an average dayChange in won per dollar from the previous 15:30 close, %, whichever the direction, 2026

How to read this chart. Each bar is how far the won moved against the dollar in one day, whichever the direction; on all three decision days it strengthened. The average day is the mean size of the 22 daily moves from the 16 Sep close to the 16 Oct close, gains and losses alike.

In won, a dollar got 1.0 won cheaper on decision day, against 4.3 won on 16 Jul and 3.9 won on 27 Aug; percent keeps days at different levels comparable. The month's biggest day was 22 Sep, when a dollar got 22.8 won cheaper.

On 23 Oct, the day after the hold, the won slipped 0.4 won. In the table, a negative change means fewer won per dollar, a stronger won; the close is won per dollar.

Source: Bank of Korea, ECOS; datasteg calculation · Data to 23 Oct 2026^{Source 1,Source 6}

About this chart

Definition: Won per US dollar, Bank of Korea 15:30 close; daily % change from the previous published close.
Calculation: Daily move = (close ÷ previous close − 1) × 100, shown as a size (absolute value). Average day = mean size of the 22 daily moves from the 16 Sep close to the 16 Oct close.
Provider: Bank of Korea, ECOS; datasteg calculation
Series: 731Y003/D/0000003, 722Y001/D/0101000
Period: 16 Jul 2026 – 23 Oct 2026
Last observation: 23 Oct 2026
Retrieved: 23 Oct 2026
Licence: ECOS Open API terms: reuse allowed with attribution
Sources: Won per US dollar, 15:30 close, Bank of Korea base rate

Download CSV

In June Governor Shin Hyun Song said that round-the-clock trading in the onshore currency market would open in July⁠^{Source 14}. The won's extended-session close tells the same story: on decision day it ended 0.3 won weaker than the day before, at 1,326.5 won⁠^{Source 15}.

The market-rate gap moved against the won late in its rise

Over the won's rise as a whole the gap that markets trade moved 60bp against Korea, all of the net change after 26 Aug, by when 74% of the rise had already happened⁠^{Source 1,Source 2,Source 3}. Two gaps matter here, both measured in basis points. The policy-rate gap is Korea's base rate minus the top of the Federal Reserve's target range. The market-rate gap, Korea's 3-year government bond yield minus the US 2-year Treasury yield, moves with where investors think policy rates are heading.

The policy gap moved both ways: Korea's two hikes narrowed it from 125bp to 75bp below the Fed's⁠^{Source 6,Source 7}, and the won rose before, between and after them. The Fed's rise, in effect from 17 Sep 2026, widened it to 100bp, where the hold left it⁠^{Source 6,Source 7}. On 17 Sep 2026 the won fell 0.99%, its largest one-day fall since the July low⁠^{Source 1}. It came within a setback that had begun on 9 Sep, before the Fed moved: a dollar cost 1,336.1 won on 9 Sep and 1,383.3 won on 18 Sep⁠^{Source 1}.

The market gap moved against Korea, from −39bp at the won's July low to −99bp on 21 Oct, its widest since 30 Sep 2025⁠^{Source 2,Source 3}. Until late August it swung, between −26.4bp on 15 Jul and −52.2bp on 31 Jul, and on 26 Aug it stood at −37.4bp, close to where it started. US 2-year yields rose 67bp, while Korea's 3-year yield rose only 8bp⁠^{Source 2,Source 3}. The chart below shows both gaps through the year.

Figure 3

The gap between Korean and US market rates moved 60bp against Korea from 2 Jul to 21 Oct, on balance all of it after late August; the hold left the policy-rate gap at -100bp

Korea–US interest-rate gaps in 2026Basis points, Korean rate minus US rate, daily, 2 Jan–22 Oct 2026

How to read this chart. Below zero, Korean rates are lower than US rates; a falling line means the gap is moving against Korea. The policy-rate gap stepped up when Korea raised its rate on 16 Jul and 27 Aug and down when the Fed raised on 17 Sep. The shading is the won's rise, from its 2 Jul low.

Market-rate gap: Korea's 3-year government bond yield minus the US 2-year Treasury yield, on days with both a Seoul and a New York close, to 21 Oct (latest US data). From 2 Jul to 21 Oct US 2-year yields rose 67bp and Korea's 3-year yield 8bp.

Policy-rate gap: the Bank of Korea's base rate minus the top of the Fed's target range, to 22 Oct. It narrowed on balance from -125bp to -100bp over the won's rise; on balance only the market-rate gap moved against Korea.

Left out: the 10-year gap, which tells the same story (-31bp on 2 Jul, -97bp on 21 Oct); the won itself is in a different unit and is marked here by the shading, never a second axis.

Source: Bank of Korea, ECOS; Federal Reserve via FRED; datasteg calculation · Data to 22 Oct 2026^{Source 2,Source 3,Source 6,Source 7}

About this chart

Definition: Market-rate gap: Korea 3-year government bond yield minus US 2-year Treasury yield (constant maturity). Policy-rate gap: Bank of Korea base rate minus the upper limit of the Fed funds target range.
Calculation: (Korean rate − US rate) × 100, in basis points, for the same calendar date (Seoul and New York closes of that date).
Provider: Bank of Korea, ECOS; Federal Reserve Board via FRED; datasteg calculation
Series: 817Y002/D/010200000, DGS2, 722Y001/D/0101000, DFEDTARU
Period: 2 Jan 2026 – 22 Oct 2026
Last observation: 22 Oct 2026
Retrieved: 23 Oct 2026
Licence: ECOS Open API terms: reuse allowed with attribution; FRED: public domain, citation requested
Sources: Korea 3-year government bond yield, US 2-year Treasury yield, Bank of Korea base rate, US federal funds target range, upper limit

Download CSV

Day to day, the won ignored the market gap. Over 70 days the correlation between their daily moves was 0.05, close to none, and the won moved the way the gap suggested on 47% of days⁠^{Source 1,Source 2,Source 3}. These are descriptions, not a test of cause.

Research suggests why a hold need not move a currency. Under uncovered interest parity a currency reflects the whole expected path of rate gaps⁠^{Source 16}, but in US, euro-area and UK data expected rate gaps in the near future matter much more than the theory implies⁠^{Source 16}. By that finding, the market gap's move against Korea would have been expected to weigh on the won⁠^{Source 2,Source 3,Source 16}. On Fed and ECB decision days currencies react to the surprise, mostly about the expected path of rates⁠^{Source 9}. In a study of advanced and emerging economies, surprises in the current rate averaged close to zero⁠^{Source 8}.

Korea's 3-year yield suggests the hold was expected. It fell 30bp from its 28 Sep 2026 high to 21 Oct, the day before the meeting⁠^{Source 2}. Its margin over the base rate shrank from 112bp to 82bp over the same weeks, and on decision day it moved 1.8bp⁠^{Source 2,Source 6}.

Rates still matter: in the same study a surprise rise in the policy rate strengthened the currency⁠^{Source 8}, though less so in advanced economies since 2021⁠^{Source 8}. In a study of emerging economies from 2000 to 2019, the tendency of currencies to weaken after rate rises disappeared or shrank once forward-looking expectations were included⁠^{Source 17}.

This was the won, not a weak dollar

The won rose far more than the broad dollar moved, and more than its Asian neighbours, so the explanation points to Korea. In emerging economies capital inflows rise when the dollar weakens, and the dollar remains an important global factor⁠^{Source 18}. From 2 Jul to 16 Oct, at New York noon rates, a dollar came to cost 13.3% fewer won while the broad dollar index rose 0.6%⁠^{Source 11,Source 12}. A dollar bought 1.8% fewer yen but 1.7% more Taiwan dollars, the currency of a fellow chip exporter⁠^{Source 13,Source 19}.

The chart below shows the three currencies from 1 Jun, each set to one hundred on 2 Jul.

Figure 4

From 2 Jul to 16 Oct the won gained far more against the dollar than the yen did, while the Taiwan dollar, a fellow chip exporter's currency, weakened

Value of the won, the yen and the Taiwan dollar against the US dollarIndex, 2 Jul 2026 = 100, daily, New York noon rates, 1 Jun–16 Oct 2026

How to read this chart. Each line is how many dollars a unit of the currency buys, set to 100 on 2 Jul, the day of the won's weakest Seoul close of 2026. Above 100, the currency buys more dollars than on 2 Jul; higher means stronger.

Over the same span a dollar came to cost 13.3% fewer won, while the broad dollar index, the dollar against a wide basket of currencies, rose 0.6%: the won's rise was not dollar weakness.

New York noon rates (Federal Reserve H.10), to 16 Oct; the meeting week arrives with the 26 Oct release. The broad dollar index measures the dollar's strength, the opposite direction to these lines, so it is in the notes and the text rather than on the chart.

Left out: Seoul-market rates for the yen and the Taiwan dollar, which cannot be republished; the yuan, which we have only as the Bank of Korea's won-yuan close, is in the text.

Source: Federal Reserve via FRED; datasteg calculation · Data to 16 Oct 2026^{Source 11,Source 12,Source 13,Source 19}

About this chart

Definition: Noon buying rates in New York for won, yen and Taiwan dollars per US dollar, turned into each currency's value in dollars; nominal broad US dollar index (notes only).
Calculation: Index = 100 × (currency per dollar on 2 Jul 2026) ÷ (currency per dollar on the day); higher = stronger currency.
Provider: Federal Reserve Board (H.10), via FRED; datasteg calculation
Series: DEXKOUS, DEXJPUS, DEXTAUS, DTWEXBGS
Period: 1 Jun 2026 – 16 Oct 2026
Last observation: 16 Oct 2026
Retrieved: 23 Oct 2026
Licence: FRED: public domain, citation requested
Sources: Korean won per US dollar (New York noon rate), Japanese yen per US dollar (New York noon rate), Taiwan dollars per US dollar (New York noon rate), Broad US dollar index (Federal Reserve)

Download CSV

The month before the meeting week looks the same. From 16 Sep to 16 Oct the broad dollar rose 2.1%, and the yen and the Taiwan dollar each lost about 1.9% against it, while a dollar came to cost 2.5% fewer won⁠^{Source 11,Source 12,Source 13,Source 19}. US data for the meeting week itself are not yet out. Against the yuan, the one comparison we can publish, the won gained 0.43% from 16 Oct to 22 Oct⁠^{Source 20}.

Calmer markets do not explain the timing either. The VIX, a gauge of expected swings in US stocks, closed at 16.46 on 15 Oct, the US close before the won's best day that fortnight. That was above every close from 2 Oct to 13 Oct and only a little below 14 Oct's⁠^{Source 21}.

More dollars came in than foreign stock selling took out

In the short run a currency follows the net of all buying and selling, and in January–August 2026 the dollars the current account brought in far exceeded those foreign stock sales took out, though how many of them reached the currency market is not measured⁠^{Source 4,Source 5}. Like the price at a market stall, a currency moves with who is buying and who is selling today, whatever their reasons; studies call this order flow⁠^{Source 10}. In mark and yen trading in 1996, a model with order flow and the rate gap explained more than half of daily moves in the mark-dollar rate, almost wholly through order flow⁠^{Source 10}. In a study of major currencies and the Mexican peso from 2006 to 2024, changes in dealer banks' currency positions accounted for most of the monthly moves in exchange rates, and much of the adjustment came through the exchange rate rather than interest rates⁠^{Source 22}.

Foreign selling was large: in the balance of payments, foreigners sold a net USD 101bn of Korean shares in January–August 2026, including USD 32bn in June, the most in a month since at least 2015⁠^{Source 5}. What the data measure is foreigners' net selling of shares, which is how we read the question's "redemptions". In June Governor Shin said such outflows kept the won weak despite a large surplus⁠^{Source 14}. In a study of Thailand in 2005 and 2006, net purchases of Thai shares went with a stronger baht, and foreign investors did not appear to hedge their currency risk⁠^{Source 23}.

But the current account brought in USD 279bn over the same months, 2.8 times as much⁠^{Source 4,Source 5}. From March it ran at between USD 28bn and USD 50bn a month⁠^{Source 4}. Foreigners also bought USD 15bn of Korean bonds, while Korean residents' purchases of foreign securities took USD 74bn out; then-Governor Rhee Chang Yong had said such purchases weighed on the won in late 2025⁠^{Source 24,Source 25,Source 26}. Netting these four flows leaves USD 119bn coming in, as the chart below shows; direct investment, derivatives, other investment and reserves are left out, so this is not the whole balance of payments⁠^{Source 4,Source 5,Source 24,Source 25}.

Figure 5

From January to August 2026 the current account brought in USD 279bn, nearly three times the USD 101bn foreigners took out of Korean stocks

Dollars into and out of Korea through four flows, January–August 2026USD billion, sum of monthly balance-of-payments flows, size of each flow

How to read this chart. Each bar is a total for January–August 2026, whichever the direction; blue bars brought dollars into Korea, grey bars took them out. Stock outflow: foreigners' net sales of Korean shares. Buying abroad: Korean residents' net purchases of foreign shares and bonds. Bond inflow: foreigners' net purchases of Korean bonds.

These four flows net to USD 119bn into Korea. Direct investment, derivatives, other investment and reserves are left out, so this is not the whole balance of payments.

Left out: the monthly timing. Foreign stock selling was heaviest in March–June (USD 93bn) and turned to net buying of USD 6bn in July–August; this is in the text. Daily exchange data on foreign stock trades for October cannot be republished. Balance-of-payments data end in August.

Source: Bank of Korea, ECOS; datasteg calculation · Data to Aug 2026^{Source 4,Source 5,Source 24,Source 25}

About this chart

Definition: Balance of payments, monthly, USD: current account; equity liabilities (foreign investors' net purchases of Korean shares); portfolio investment assets (Korean residents' net purchases of foreign securities); debt-securities liabilities (foreign investors' net purchases of Korean bonds).
Calculation: Sum of monthly values, January–August 2026, USD million ÷ 1,000, shown as sizes. Net of the four = current account + foreign purchases of Korean shares and bonds − residents' purchases of foreign securities.
Provider: Bank of Korea, ECOS; datasteg calculation
Series: 301Y013/M/000000, 301Y013/M/BOPF22100000, 301Y013/M/BOPF21000000, 301Y013/M/BOPF22200000
Period: Jan 2026 – Aug 2026
Last observation: Aug 2026
Retrieved: 23 Oct 2026
Licence: ECOS Open API terms: reuse allowed with attribution
Sources: Korea current account balance, Foreign investors' portfolio equity investment in Korea (BoP equity liabilities), Korean residents' portfolio investment abroad (BoP portfolio assets), Foreign investors' portfolio debt-securities investment in Korea (BoP debt liabilities)

Download CSV

For foreign stock flows the balance-of-payments timing fits, up to August. Foreign selling was heaviest in March–June, USD 93bn, when the won was weakest, and turned to small net buying of USD 6bn in July–August as the won rose⁠^{Source 5}. The net of the four flows was positive before the turn too, USD 51bn in March–June while the won weakened, but faster in July–August, USD 61bn in two months⁠^{Source 4,Source 5,Source 24,Source 25}.

Daily exchange data, which we cannot republish, tell a different story from July. In them foreigners went on selling Korean shares on balance in every month, more heavily in September and October while the won kept rising. Their net buying came only in short spells, such as in mid-July and mid-August, and from 16 Oct, the won's best day in October, to decision day, with one selling day in between. The balance-of-payments data end in August, so the two records cannot yet be compared for the autumn.

Export dollars arrive, but not on a timetable

The export boom created a large supply of dollars, but when exporters sell them for won is not in the data we could use. Export growth turned up at the end of 2025, reaching 14.7% a year in December, and surged from January 2026, when it reached 37.1%⁠^{Source 27}. By September 2026 export values were 89.2% higher than a year before, with higher prices explaining 58% of the rise⁠^{Source 27,Source 28,Source 29}. Then-Governor Rhee said the chip boom had helped steady the currency market⁠^{Source 30}.

Export dollars need not reach the currency market on any given day. Exporters can hold on to dollars instead of converting them⁠^{Source 31}, and Bank of Korea economists write that Korea's persistent surpluses have been directed more towards overseas investment⁠^{Source 18}. In Brazil from 2002 to 2011, daily flows from non-financial customers, mostly trade, were much more strongly linked to moves in the real than flows from financial customers⁠^{Source 32}. One published hint in the data we used, residents' foreign-currency deposits at Korean banks, rose about USD 29bn from end-June to end-August, so residents were adding to their foreign-currency deposits, not running them down, as the won rose⁠^{Source 1,Source 33}.

Taiwan offers a contrast: this year a dollar came to cost 3.2% more Taiwan dollars but 7.6% fewer won⁠^{Source 11,Source 13}. We found no study of why the two diverged in 2026. One difference the BIS noted in 2021 is hedging: Korean life insurers hedged close to all of their foreign-currency bond portfolios⁠^{Source 34}, while the share hedged by Taiwanese life insurers had fallen to just over half in 2019⁠^{Source 34}. In April 2025 the won was reportedly used as a stand-in hedge for the Taiwan dollar⁠^{Source 18}.

Whether the authorities bought or sold dollars is not in our sources. Foreign-exchange reserves rose USD 13bn from end-June to end-September, but changes in the value of reserve assets blur that⁠^{Source 35}.

Our read

The question's two intuitions, that rate gaps set the won and that foreign stock selling pushes it down, describe real forces, but in this episode neither was in charge. Korea's 3-year yield moved little on decision day, which suggests the hold was largely expected⁠^{Source 2,Source 6}. The gap with US market rates moved against the won on balance from late August, and the won rose far more than the broad dollar moved, which points to something specific to Korea, such as the flows into and out of it⁠^{Source 2,Source 3,Source 12}. The four flows we measured brought in more dollars than they took out, with a current account surplus 2.8 times the size of foreign stock selling in January–August, though they are not the whole balance of payments⁠^{Source 4,Source 5}. But they did so on balance in March–June too, when the won weakened⁠^{Source 4,Source 5,Source 24,Source 25}. When exporters convert dollars, how institutions hedge, whether the authorities intervened and what moved in the meeting week are not in the data we could use or publish. US data due on 26 Oct and September's balance of payments will test this read.

Terms in this article

base rate: The policy interest rate a central bank sets to steer borrowing costs across the economy. In Korea it is the Bank of Korea's Base Rate.Here: the Bank of Korea's base rate, raised twice in 2026 and then held in October.
bond yield: The yearly return an investor gets by buying a government bond at today's market price and holding it until it is repaid. When bond prices fall, yields rise.Here: the yields on Korea's 3-year government bonds and US 2-year Treasuries, compared day by day.
broad dollar index: The US dollar's value against the currencies of many trading partners at once. A rise means the dollar has strengthened against the world as a whole.Here: the Federal Reserve's nominal broad index, used to test whether the won's rise was just dollar weakness.
current account: A country's earnings from selling goods and services abroad and from investments abroad, minus what it pays for the same. A surplus means more foreign currency comes in from these transactions than goes out.
basis point: A hundredth of a percentage point, the usual unit for small gaps and changes in interest rates.
policy-rate gap: The difference between the interest rates two central banks have set. It says where policy stands today, not where markets expect it to go.Here: the Bank of Korea's base rate minus the top of the US Federal Reserve's target range.
uncovered interest parity: The textbook idea that investors expect a higher-interest currency to lose value by just enough to offset its extra interest. It implies that a currency reflects the whole path of interest-rate gaps investors expect.
expected policy path: Where investors expect a central bank to set its policy rate over the coming months and years. It shows up in market yields of one to three years.
order flow: The difference between orders to buy and orders to sell a currency or other asset over a period, whatever the reason for each trade. More buying than selling tends to push the price up.
balance of payments: A country's record of all money crossing its borders: trade and income in the current account, and investment flows such as purchases of shares and bonds in the financial account.Here: the Bank of Korea's monthly figures in dollars, to August 2026.
currency hedging: Contracts that protect an investor holding foreign assets against losses if the foreign currency falls. Hedging does not mean selling the assets.Here: how much of their foreign bonds Korean and Taiwanese life insurers protect against currency moves.

References

Research

8
Emeksiz, E. O., Kamber, G., Otten, J., Pasricha, G. K. (2026). Still Packing a Punch: Monetary Policy Transmission in a New Cross-Country High-Frequency Dataset. IMF Working Papers 2026/018. International Monetary Fund
Read the source ↗PDF ↗Cited in: ↑ Key findings, ↑ The market-rate gap moved against the won late in its rise
9
Gürkaynak, R. S., Kara, A. H., Kısacıkoğlu, B., Lee, S. S. (2020). Monetary Policy Surprises and Exchange Rate Behavior. NBER Working Papers w27819. National Bureau of Economic Research
Read the source ↗PDF ↗Cited in: ↑ Key findings, ↑ The market-rate gap moved against the won late in its rise
10
Evans, M. D. D., Lyons, R. K. (1999). Order Flow and Exchange Rate Dynamics. NBER Working Papers w7317. National Bureau of Economic Research
Read the source ↗PDF ↗Cited in: ↑ Key findings, ↑ More dollars came in than foreign stock selling took out
16
Galí, J. (2020). Uncovered Interest Parity, Forward Guidance, and the Exchange Rate. NBER Working Papers w26797. National Bureau of Economic Research
Read the source ↗PDF ↗Cited in: ↑ The market-rate gap moved against the won late in its rise
17
Ha, J., Kim, D., Kose, M. A., Prasad, E. S. (2024). Resolving Puzzles of Monetary Policy Transmission in Emerging Markets. NBER Working Papers w33133. National Bureau of Economic Research
Read the source ↗PDF ↗Cited in: ↑ The market-rate gap moved against the won late in its rise
18
Cavallino, P., Ehlers, T., Gamboa, F., Guerra, R., Martinez, B., Shim, I. (overview); Kim, J., Kim, M. (Bank of Korea chapter); Central Bank of Chile (2026). Capital flows, exchange rates and financial conditions in EMEs in an evolving international monetary system. BIS Papers 171. BIS
Read the source ↗PDF ↗Cited in: ↑ This was the won, not a weak dollar, ↑ Export dollars arrive, but not on a timetable
22
Dao, M., Gourinchas, P.-O., Itskhoki, O. (2025). Breaking Parity: Equilibrium Exchange Rates and Currency Premia. IMF Working Papers 2025/153. International Monetary Fund
Read the source ↗PDF ↗Cited in: ↑ More dollars came in than foreign stock selling took out
23
Gyntelberg, J., Loretan, M., Subhanij, T. (2012). Exchange Rate Fluctuations and International Portfolio Rebalancing in Thailand. IMF Working Papers 2012/214. International Monetary Fund
Read the source ↗PDF ↗Cited in: ↑ More dollars came in than foreign stock selling took out
31
Juselius, M., Xia, D., Wooldridge, P. (2025). Financial channel implications of a weaker dollar for emerging market economies. BIS Bulletin 114. BIS
Read the source ↗Cited in: ↑ Export dollars arrive, but not on a timetable
32
Kohlscheen, E. (2013). Order flow and the real: indirect evidence of the effectiveness of sterilized interventions. BIS Working Papers 426. BIS
Read the source ↗PDF ↗Cited in: ↑ Export dollars arrive, but not on a timetable
34
McGuire, P., Shim, I., Shin, H. S., Sushko, V. (2021). Outward portfolio investment and dollar funding in emerging Asia. BIS Quarterly Review December 2021. BIS
Read the source ↗PDF ↗Cited in: ↑ Export dollars arrive, but not on a timetable

Speeches and statements

14
Shin, H. S. (2026). Hyun Song Shin: Speech - 76th Anniversary of the Bank of Korea. cb_speech review_r260616d. Bank of Korea
Read the source ↗PDF ↗Cited in: ↑ The hold moved the won less than an ordinary day, ↑ More dollars came in than foreign stock selling took out
26
Rhee, C. Y. (2026). Chang Yong Rhee: New Year Speech. cb_speech review_r260105a. Bank of Korea
Read the source ↗PDF ↗Cited in: ↑ More dollars came in than foreign stock selling took out
30
Rhee, C. Y. (2026). Chang Yong Rhee: Farewell address. cb_speech review_r260505m. Bank of Korea
Read the source ↗PDF ↗Cited in: ↑ Export dollars arrive, but not on a timetable

Data

1
Bank of Korea, ECOS. Won per US dollar, 15:30 close. 731Y003/D/0000003. Retrieved 23 Oct 2026.
Finding the series: On ECOS (ecos.bok.or.kr), open 통계검색 (Statistics search) and type the table name “원화의 대미달러, 원화의 대위안/대엔 환율” or its code 731Y003 in the search box, or follow the statistics list 3. 환율/통관수출입/외환보유액 > 3.1. 환율 > 3.1.1. 일일환율 > 3.1.1.3. 원화의 대미달러, 원화의 대위안/대엔 환율. In the table, select 계정항목: 원/달러(종가 15:30). Set the cycle (주기) to daily, choose the period, and click 조회 (Search).
Data source ↗Cited in: ↑ Key numbers, ↑ Key findings, ↑ Almost all of the won's rise came before the meeting week, ↑ The hold moved the won less than an ordinary day, ↑ The market-rate gap moved against the won late in its rise, ↑ Export dollars arrive, but not on a timetable
2
Bank of Korea, ECOS. Korea 3-year government bond yield. 817Y002/D/010200000. Retrieved 23 Oct 2026.
Finding the series: On ECOS (ecos.bok.or.kr), open 통계검색 (Statistics search) and type the table name “시장금리(일별)” or its code 817Y002 in the search box, or follow the statistics list 1. 통화/금융 > 1.3. 금리 > 1.3.2. 시장금리 > 1.3.2.1. 시장금리(일별). In the table, select 계정항목: 국고채(3년). Set the cycle (주기) to daily, choose the period, and click 조회 (Search).
Data source ↗Cited in: ↑ Key numbers, ↑ Key findings, ↑ The market-rate gap moved against the won late in its rise, ↑ Our read
3
FRED, Federal Reserve Bank of St. Louis. US 2-year Treasury yield. DGS2. Retrieved 23 Oct 2026.
Finding the series: The link opens the FRED series DGS2 directly. Use “Edit Graph” to change the dates or units, and “Download” for the data.
Data source ↗Cited in: ↑ Key numbers, ↑ Key findings, ↑ The market-rate gap moved against the won late in its rise, ↑ Our read
4
Bank of Korea, ECOS. Korea current account balance. 301Y013/M/000000. Retrieved 23 Oct 2026.
Finding the series: On ECOS (ecos.bok.or.kr), open 통계검색 (Statistics search) and type the table name “국제수지” or its code 301Y013 in the search box, or follow the statistics list 2. 국민계정 > 2.5. 국제수지표 > 2.5.1. 국제수지 > 2.5.1.1. 국제수지. In the table, select 국제수지(BPM6): 경상수지. Set the cycle (주기) to monthly, choose the period, and click 조회 (Search).
Data source ↗Cited in: ↑ Key numbers, ↑ Key findings, ↑ More dollars came in than foreign stock selling took out, ↑ Our read
5
Bank of Korea, ECOS. Foreign investors' portfolio equity investment in Korea (BoP equity liabilities). 301Y013/M/BOPF22100000. Retrieved 23 Oct 2026.
Finding the series: On ECOS (ecos.bok.or.kr), open 통계검색 (Statistics search) and type the table name “국제수지” or its code 301Y013 in the search box, or follow the statistics list 2. 국민계정 > 2.5. 국제수지표 > 2.5.1. 국제수지 > 2.5.1.1. 국제수지. In the table, select 국제수지(BPM6): 주식(부채). Set the cycle (주기) to monthly, choose the period, and click 조회 (Search).
Data source ↗Cited in: ↑ Key numbers, ↑ Key findings, ↑ More dollars came in than foreign stock selling took out, ↑ Our read
6
Bank of Korea, ECOS. Bank of Korea base rate. 722Y001/D/0101000. Retrieved 23 Oct 2026.
Finding the series: On ECOS (ecos.bok.or.kr), open 통계검색 (Statistics search) and type the table name “한국은행 기준금리 및 여수신금리” or its code 722Y001 in the search box, or follow the statistics list 1. 통화/금융 > 1.3. 금리 > 1.3.1. 한국은행 기준금리 및 여수신금리. In the table, select 계정항목: 한국은행 기준금리. Set the cycle (주기) to daily, choose the period, and click 조회 (Search).
Data source ↗Cited in: ↑ Key findings, ↑ Almost all of the won's rise came before the meeting week, ↑ The hold moved the won less than an ordinary day, ↑ The market-rate gap moved against the won late in its rise, ↑ Our read
7
FRED, Federal Reserve Bank of St. Louis. US federal funds target range, upper limit. DFEDTARU. Retrieved 23 Oct 2026.
Finding the series: The link opens the FRED series DFEDTARU directly. Use “Edit Graph” to change the dates or units, and “Download” for the data.
Data source ↗Cited in: ↑ Key findings, ↑ The market-rate gap moved against the won late in its rise
11
FRED, Federal Reserve Bank of St. Louis. Korean won per US dollar (New York noon rate). DEXKOUS. Retrieved 23 Oct 2026.
Finding the series: The link opens the FRED series DEXKOUS directly. Use “Edit Graph” to change the dates or units, and “Download” for the data.
Data source ↗Cited in: ↑ Key findings, ↑ This was the won, not a weak dollar, ↑ Export dollars arrive, but not on a timetable
12
FRED, Federal Reserve Bank of St. Louis. Broad US dollar index (Federal Reserve). DTWEXBGS. Retrieved 23 Oct 2026.
Finding the series: The link opens the FRED series DTWEXBGS directly. Use “Edit Graph” to change the dates or units, and “Download” for the data.
Data source ↗Cited in: ↑ Key findings, ↑ This was the won, not a weak dollar, ↑ Our read
13
FRED, Federal Reserve Bank of St. Louis. Taiwan dollars per US dollar (New York noon rate). DEXTAUS. Retrieved 23 Oct 2026.
Finding the series: The link opens the FRED series DEXTAUS directly. Use “Edit Graph” to change the dates or units, and “Download” for the data.
Data source ↗Cited in: ↑ Key findings, ↑ This was the won, not a weak dollar, ↑ Export dollars arrive, but not on a timetable
15
Bank of Korea, ECOS. Won per US dollar, extended-session close. 731Y003/D/0000013. Retrieved 23 Oct 2026.
Finding the series: On ECOS (ecos.bok.or.kr), open 통계검색 (Statistics search) and type the table name “원화의 대미달러, 원화의 대위안/대엔 환율” or its code 731Y003 in the search box, or follow the statistics list 3. 환율/통관수출입/외환보유액 > 3.1. 환율 > 3.1.1. 일일환율 > 3.1.1.3. 원화의 대미달러, 원화의 대위안/대엔 환율. In the table, select 계정항목: 원/달러(종가). Set the cycle (주기) to daily, choose the period, and click 조회 (Search).
Data source ↗Cited in: ↑ The hold moved the won less than an ordinary day
19
FRED, Federal Reserve Bank of St. Louis. Japanese yen per US dollar (New York noon rate). DEXJPUS. Retrieved 23 Oct 2026.
Finding the series: The link opens the FRED series DEXJPUS directly. Use “Edit Graph” to change the dates or units, and “Download” for the data.
Data source ↗Cited in: ↑ This was the won, not a weak dollar
20
Bank of Korea, ECOS. Won per Chinese yuan, Seoul direct market close. 731Y003/D/0000010. Retrieved 23 Oct 2026.
Finding the series: On ECOS (ecos.bok.or.kr), open 통계검색 (Statistics search) and type the table name “원화의 대미달러, 원화의 대위안/대엔 환율” or its code 731Y003 in the search box, or follow the statistics list 3. 환율/통관수출입/외환보유액 > 3.1. 환율 > 3.1.1. 일일환율 > 3.1.1.3. 원화의 대미달러, 원화의 대위안/대엔 환율. In the table, select 계정항목: 원/위안(종가). Set the cycle (주기) to daily, choose the period, and click 조회 (Search).
Data source ↗Cited in: ↑ This was the won, not a weak dollar
21
FRED, Federal Reserve Bank of St. Louis. VIX, US stock-market volatility index. VIXCLS. Retrieved 23 Oct 2026.
Finding the series: The link opens the FRED series VIXCLS directly. Use “Edit Graph” to change the dates or units, and “Download” for the data.
Data source ↗Cited in: ↑ This was the won, not a weak dollar
24
Bank of Korea, ECOS. Foreign investors' portfolio debt-securities investment in Korea (BoP debt liabilities). 301Y013/M/BOPF22200000. Retrieved 23 Oct 2026.
Finding the series: On ECOS (ecos.bok.or.kr), open 통계검색 (Statistics search) and type the table name “국제수지” or its code 301Y013 in the search box, or follow the statistics list 2. 국민계정 > 2.5. 국제수지표 > 2.5.1. 국제수지 > 2.5.1.1. 국제수지. In the table, select 국제수지(BPM6): 부채성증권(부채). Set the cycle (주기) to monthly, choose the period, and click 조회 (Search).
Data source ↗Cited in: ↑ More dollars came in than foreign stock selling took out, ↑ Our read
25
Bank of Korea, ECOS. Korean residents' portfolio investment abroad (BoP portfolio assets). 301Y013/M/BOPF21000000. Retrieved 23 Oct 2026.
Finding the series: On ECOS (ecos.bok.or.kr), open 통계검색 (Statistics search) and type the table name “국제수지” or its code 301Y013 in the search box, or follow the statistics list 2. 국민계정 > 2.5. 국제수지표 > 2.5.1. 국제수지 > 2.5.1.1. 국제수지. In the table, select 국제수지(BPM6): 증권투자(자산). Set the cycle (주기) to monthly, choose the period, and click 조회 (Search).
Data source ↗Cited in: ↑ More dollars came in than foreign stock selling took out, ↑ Our read
27
Bank of Korea, ECOS. Korea export value index (customs, US dollars). 403Y001/M/*AA. Retrieved 23 Oct 2026.
Finding the series: On ECOS (ecos.bok.or.kr), open 통계검색 (Statistics search) and type the table name “수출금액지수” or its code 403Y001 in the search box, or follow the statistics list 3. 환율/통관수출입/외환보유액 > 3.3. 무역지수/교역조건지수(2020=100) > 3.3.1. 수출입 금액/물량지수 > 3.3.1.1. 수출 금액/물량지수 > 3.3.1.1.1. 수출금액지수. In the table, select 계정코드: 총지수. Set the cycle (주기) to monthly, choose the period, and click 조회 (Search).
Data source ↗Cited in: ↑ Export dollars arrive, but not on a timetable
28
Bank of Korea, ECOS. Korea export price index, US dollar basis. 402Y014/M/*AA/D. Retrieved 23 Oct 2026.
Finding the series: On ECOS (ecos.bok.or.kr), open 통계검색 (Statistics search) and type the table name “수출물가지수(기본분류)” or its code 402Y014 in the search box, or follow the statistics list 4. 물가 > 4.3. 수출/수입물가지수(2020=100) > 4.3.1. 수출물가지수 > 4.3.1.1. 수출물가지수(기본분류). In the table, select 계정코드: 총지수; 통화계약구분코드: 달러기준. Set the cycle (주기) to monthly, choose the period, and click 조회 (Search).
Data source ↗Cited in: ↑ Export dollars arrive, but not on a timetable
29
Bank of Korea, ECOS. Korea export volume index. 403Y002/M/*AA. Retrieved 23 Oct 2026.
Finding the series: On ECOS (ecos.bok.or.kr), open 통계검색 (Statistics search) and type the table name “수출물량지수” or its code 403Y002 in the search box, or follow the statistics list 3. 환율/통관수출입/외환보유액 > 3.3. 무역지수/교역조건지수(2020=100) > 3.3.1. 수출입 금액/물량지수 > 3.3.1.1. 수출 금액/물량지수 > 3.3.1.1.2. 수출물량지수. In the table, select 계정코드: 총지수. Set the cycle (주기) to monthly, choose the period, and click 조회 (Search).
Data source ↗Cited in: ↑ Export dollars arrive, but not on a timetable
33
Bank of Korea, ECOS. 1.2.1.1.2. 예금은행 총수신(말잔): 외화예금 (KR). 104Y013/M/BCB2, 104Y013/M/BCB902. Retrieved 23 Oct 2026.
Finding the series: On ECOS (ecos.bok.or.kr), open 통계검색 (Statistics search) and type the table name “예금은행 총수신(말잔)” or its code 104Y013 in the search box, or follow the statistics list 1. 통화/금융 > 1.2. 금융기관 여수신 > 1.2.1. 수신 > 1.2.1.1. 예금은행 수신 > 1.2.1.1.2. 예금은행 총수신(말잔). In the table, select 계정항목: 외화예금. Set the cycle (주기) to monthly, choose the period, and click 조회 (Search).
Data source ↗Cited in: ↑ Export dollars arrive, but not on a timetable
35
Bank of Korea, ECOS. Korea official foreign-exchange reserves. 732Y001/M/99. Retrieved 23 Oct 2026.
Finding the series: On ECOS (ecos.bok.or.kr), open 통계검색 (Statistics search) and type the table name “외환보유액” or its code 732Y001 in the search box, or follow the statistics list 3. 환율/통관수출입/외환보유액 > 3.5. 외환보유액. In the table, select 계정항목: 합계. Set the cycle (주기) to monthly, choose the period, and click 조회 (Search).
Data source ↗Cited in: ↑ Export dollars arrive, but not on a timetable

Cite this article

If you use this article in research, a report or the news, please cite it as follows.

datasteg (2026), “The won's rise came before the rate hold, not after it”, datasteg articles, 23 October 2026, https://articles.datasteg.com/en/a/won-after-rate-hold.

BibTeX

@online{datasteg:won-after-rate-hold,
  author       = {{datasteg}},
  title        = {The won's rise came before the rate hold, not after it},
  organization = {datasteg articles},
  date         = {2026-10-23},
  url          = {https://articles.datasteg.com/en/a/won-after-rate-hold}
}

PDF edition. A print layout of this article, with every figure, the references and this citation. English (11 pages, 0.4 MB) · Korean (10 pages, 0.7 MB)

Key numbers

Key findings

The question

Almost all of the won's rise came before the meeting week

The hold moved the won less than an ordinary day

The market-rate gap moved against the won late in its rise

This was the won, not a weak dollar

More dollars came in than foreign stock selling took out

Export dollars arrive, but not on a timetable

Terms in this article

References

Research

Speeches and statements

Data

Related reading

Cite this article