Between record exports and a 19.3% three-month KOSPI decline · the larger repricing was in the multiple, not earnings
Time window · FX · valuation · flows · four layers of the divergence
Over one year, KOSPI broadly kept pace with exports · most of the three-month drop came in July · what broke then was the multiple
Published Oct 3, 2026 · Information cutoff: Oct 2, 2026, 20:00 KST · Includes Korea's Oct 2 market close · Excludes the U.S. September employment report released at 21:30 KST and anything thereafter (the jobs result appears only in the POST-CUTOFF line below the closing note)
snowshagal.comCompared with a year earlier (September 2025), KOSPI and exports moved by broadly similar magnitudes · the ranking flips depending on which export measure is used1, 2, 4, 7, 9
The question "If exports are at a record, why isn't KOSPI?" only really applies to the three months since end-June · break those three months apart and nearly all of the KOSPI decline came in July alone · the gap versus exports widened in two steps: July's index selloff and September's export surge1, 5, 6, 7
Exports are reported in dollars · corporate earnings and KOSPI are measured in won · over Q3, each dollar bought KRW 196.6 less(−12.7%)4, 7, 12
Translate the same three months into dollars and KOSPI's decline narrows to −7.6% · that is simply the arithmetic of a stronger won reducing the loss measured in dollars4 · it approximates the index-level return seen by an unhedged dollar-based investor (actual performance depends on entry point, hedging and dividends) · but this is a difference in measurement units, not proof that FX caused the equity decline
The observable consensus downgrade since late August is around 2% (the change versus end-June cannot be verified) · even the more conservative individual estimate is only a single-digit percentage below consensus · too small on its own to explain a 19.3% three-month drawdown · the larger piece is the multiple on the next page
SK hynix's ADRs began trading on Nasdaq on Jul 10 · total offering size USD 26.5071bn (USD 149 per share)18 · these dollars came from financing, not export receipts · IBK argues that the dollar-supply impulse, or expectations of it, spread from this deal to exporters more broadly17 · Yuanta argues that financial-account dollar flows have become more influential for the exchange rate than physical export flows16 · even as the index fell 22% in July, USD/KRW had declined to 1,443.97 by month-end (monthly-indicator article basis; methodology may differ from the onshore closing series)5
Split the index into earnings and valuation and the three-month picture changes · KOSPI 12-month forward EPS rose from 1,089.5p on Jun 24 to 1,275.84p on Sep 23 · +17.1% · from Jun 30 to Sep 23, KOSPI fell from 8,476.48 to 7,080.92 (−16.5%) · the June-end P/E is only an approximation, pairing the Jun 30 index with Jun 24 EPS (7.78x), but even on that approximation the falling component is the multiple3, 4, 8, 10
With EPS rising while the index fell 22% in July, the multiple compression also points to July · Daishin's assessment is that in September the index rose roughly in step with forward EPS, leaving P/E in a 5.3–5.6x range3 · Q3 earnings forecasts were being cut even as forward EPS rose partly because the 12-month window was rolling toward higher 2027 earnings (the share attributable to this effect cannot be verified) · the 2027 forecast was also nudged higher (KOSPI operating profit KRW 1,250tn → KRW 1,263tn)3
There is no evidence this time for a simple link in which more export dollars translate into equity buying · dollars earned by exporters and dollars moving with foreign equity trades are different flows · estimated Q3 foreign net selling combines KRW 9.9tn in July and KRW 10.2tn in August (Daishin) with September's KRW 21.5108tn(KRX), for roughly KRW 41.6tn in total (approximation combining sources)3, 4, 22
A substantial share of the KRW 134.9195tn sold net over five months occurred even during the index surge in May–June3, 23 · foreign selling and index direction did not always move together
Other Corporations were September's largest net buyers · Daishin attributes most of that to Samsung Electronics and SK hynix buybacks and argues those purchases helped firm the index floor · cumulative purchases: Samsung Electronics KRW 10.31tn (74.7% of plan), SK hynix KRW 24.39tn (58.8%, all to be cancelled) · net buying exceeded KRW 1tn a day on 20 of 22 trading days since Aug 20 in the buyback context3
Daishin-estimated buyback completion · Samsung Electronics around Oct 8; SK hynix around Oct 163 · the same supporting bid discussed in NO.04, "The Supporting Bid Runs Out First" · Daishin expects foreigners and financial-investment firms to fill the gap if rates and FX stabilize3 · whether they actually do can be checked in daily flows after mid-October
iM Securities argues that KOSPI underperformed exports relative to Taiwan · Korea's market-cap-to-trailing-12-month-exports ratio is lower than Taiwan's, which iM attributes to a more diversified export mix with FX adding part of the gap · Q3 USD/KRW fell 12%, while USD/TWD was roughly flat24
Market cap ÷ trailing-12-month exports is not a valuation multiple for estimating fair value; it is a supplementary gauge of the relative gap between the two markets
A weaker won would support earnings translated back into KRW · but IBK says foreign equity flows have been the variable most closely linked with USD/KRW since last year17 · FX direction alone cannot determine the index direction
The comparison base drops to USD 59.52bn in Oct 2025 and USD 60.77bn in Nov 2025, while the Chuseok shift also creates a working-day reverse base effect2 · even if high growth rates persist, the new information is not the year-on-year base comparison but changes in earnings estimates
September commodity DRAM export unit value −3.5% m/m, NAND −22.0% m/m (Korea Customs Service, weight-based)21 · exports can be at a record while the market focuses first on a rollover in unit-value momentum · Hanwha's main text, however, characterizes commodity DRAM as "unit prices steady, volumes up" and puts high-value memory (MCP) unit prices at +2.7% m/m21
In iM's chart, Korea's market-cap-to-exports ratio sits well below Taiwan's · iM also points to Korea's more diversified export mix24 · an underlying difference remains that Q3 events alone cannot explain
Does forward P/E move back above 6x while forward EPS holds or rises? · Daishin maps 6x P/E to KOSPI 7,6553 · if preliminary results miss consensus but forward EPS estimates hold, the signal is in the multiple
The buyback completion dates marked as estimates are brokerage calculations based on the daily purchase pace, not official filing periods · actual completion should show up in daily Other Corporations net buying before the DART completion reports
Next checks: preliminary results on Oct 7–8, then daily flows after the buyback bid disappears.
POST-CUTOFF · Oct 2 21:30 KST: U.S. September payrolls +29k · unemployment 4.2% · July and August revised down by a combined 60k · not incorporated into the main analysis31
This document examines the divergence between exports and the index over a defined period · it is not a forecast of any asset's direction or a trading recommendation · Data basis: exports use the ministry's Oct 1 preliminary release · KOSPI and USD/KRW use reported closes (USD/KRW 15:30 onshore close, through Oct 2) · forward EPS/P/E use Daishin Securities aggregates (Jun 24 and Sep 23, Sep 28 report; Aug 31 P/E is this note's calculation using end-August EPS of 1,240p) · September investor-group net buying and five-month foreign net selling use reports citing KRX data through Sep 30; July–August foreign flows use Daishin · Oct 2 foreign net selling is shown as ~KRW 140bn because closing tallies differ by outlet · earnings consensus uses FnGuide (reported Sep 30) and Daishin · U.S. 10-year: Jun 30/Jul 31 from a monthly-indicator article, September change from Daishin, Sep 29/30 from U.S. Treasury daily par yields · brokerage interpretations: DS Sep 29; Yuanta, Hanwha and Eugene Oct 1; iM and IBK Oct 2 · ADR size from SEC prospectus dated Jul 9 · rebasing, Index in USD terms, June-end forward P/E and decomposition, and Q3 foreign-flow total are this note's calculations · all may change with subsequent data and market moves
Information cutoff · Oct 2, 2026, 20:00 KST · includes Korea's Oct 2 market close · excludes the U.S. September employment report released at 21:30 KST and anything thereafter