The Korea Discount, Measured — 55% of Listed Companies Trade Below Book

"Korean stocks are cheap" is one of the most repeated claims in emerging market investing. It is also true — but the interesting question is not whether the discount exists. It is what the discount is made of, because that determines whether it can close. Here is the whole market, all 2,653 listed companies, measured.

1. The headline numbers

More than half of Korea's listed companies are priced below the accounting value of their own net assets. In most markets that would be a screaming anomaly. Here it is the median condition.

The fourth number explains most of the first three. A median return on equity of 2.7% is below what a Korean savings account pays. If a company's book value earns 2.7% a year, refusing to pay 1x book for it is not a mispricing — it is arithmetic. Capital that compounds slowly is worth less than capital that compounds quickly, and the market prices that correctly.

2. The profitability problem is broad, not concentrated

It would be comforting if low ROE were confined to a few dying industries. It is not. 37.8% of listed companies have negative ROE — they lost money over the trailing twelve months. Another large block earns a low single-digit return.

Only about one company in ten sustains the kind of return that would justify a premium to book. The discount is not applied uniformly to a market of good businesses; it is a weighted average that includes a very long tail of businesses that do not earn their cost of capital.

3. The cash problem — a discount that is real but stuck

Here is where it gets more interesting for a value investor. 1,175 of the 2,479 companies with computable net debt — 47.4% — hold more cash than debt. Korean balance sheets carry an enormous amount of idle liquidity.

On an enterprise-value basis, a meaningful slice of this market is cheaper than the P/E suggests, because the P/E does not net out the cash. A company with a ₩100bn market cap, ₩40bn of net cash and ₩10bn of operating profit trades at 10x earnings but 6x EV/EBIT. The operating business is being valued at roughly half what the headline multiple implies.

The catch: cash that never leaves the balance sheet does not accrue to minority shareholders in any practical sense. This is why the Korean valuation debate is dominated by governance and payout policy rather than by growth forecasts. The assets are there; the question is whether their owners will ever share them.

4. Sector dispersion is enormous

Median P/E by sector (sectors with at least eight companies):

A ten-fold spread. Note that the expensive end is partly a denominator artefact: a company earning almost nothing trades at a huge multiple no matter how you feel about it. Any screen that ranks the whole market on a single multiple is mostly sorting by sector.

5. So can the discount close?

Break it into three parts and the answer differs for each:

• The earnings-power part — companies that genuinely earn 2–3% on equity. This is not a discount to be closed; it is a fair price for weak capital returns. It closes only if the businesses improve.
• The payout part — companies that earn adequately but hoard. This can close, and it is what Korea's corporate value-up programme is aimed at. Watch dividend policy, buybacks, and above all share cancellation, which is the only buyback that permanently raises per-share value.
• The structure part — holding companies and cross-shareholdings, where the value of the underlying assets does not flow to the listed vehicle. See the holding-company discount.

Treating all three as one number is how "Korea is cheap" becomes a thesis that never resolves. They are different problems with different timelines.

Key figures

• 2,653 listed companies; median P/B 0.88x; 55.3% below book.
• Median ROE 2.7%; 37.8% negative; only 10.2% above 15%.
• 47.4% of companies hold net cash — the discount is partly an EV story.
• Sector median P/E ranges from 6.2x to 61.2x. Compare within sectors only.

All figures computed from Korea FSC price data and OpenDART filings as of 2026-08-12. Flow items use trailing four quarters; balance items use the latest quarter-end.