Screeners divide by total equity. For holding companies that includes money that isn't the shareholders'. Here's the same 183 companies measured both ways.
I kept running into the same problem screening Korean equities: a holding company shows up at 0.2x or 0.5x price-to-book, and when you open the filing a large part of that equity turns out to belong to the minority shareholders of its subsidiaries, not to the company whose shares you'd be buying.
Consolidated statements bring 100% of a controlled subsidiary onto the parent's balance sheet and park the outside ownership in non-controlling interests. Market cap never included that part. So the standard P/B has a numerator and a denominator describing different sets of shareholders.
I rebuilt the calculation for 183 Korean listed companies straight from DART (Korea's mandatory disclosure system), measuring each one twice — once against total equity, once against equity attributable to owners of the parent.
What came out:
\- 70 companies read below 1.0x on total equity. On the owners' share, 66 do.
\- Four cross the line entirely: SK 0.50x → 1.68x, HD Hyundai 0.61x → 1.84x, Misto Holdings 0.87x → 1.17x, Hyosung 0.99x → 1.16x.
\- It's concentrated, not general. Median gap across the 26 holding companies is +28.90%; across the other 157 it's +1.56%. Median across all 183 is +2.41% while the mean is +21.76% — a few very wide cases drag the average.
\- Widest is Doosan at 1.69x vs 13.25x.
Some notes on how it's built, since that's usually the first question:
\- Universe is the top 200 KOSPI common shares by market cap plus ten smaller names, financials removed (a bank's "total borrowings" can't be separated from operating liabilities, so half the schema would be meaningless). 183 after exclusions.
\- Every figure carries the DART receipt number of its filing, and the table links each one to the original document. If a filing doesn't carry a number, the cell is empty — I don't fill gaps with estimates.
\- Two exceptions, both arithmetic identities inside the same statement rather than estimates: where a company prints no controlling-interest line because it has no non-controlling interests, I complete it as total equity less NCI, or net profit in full. Nine companies, all flagged so you can exclude them.
\- The interim-report trap that cost me a day: in a Korean half-year filing the headline income-statement column holds the three-month figure, not the cumulative. Read the wrong field and you halve the result.
None of this says anything about whether a share is cheap. A wide gap means large non-controlling interests, full stop. I'm posting it because the input is more useful than my opinion about it.
Free, no signup to view, and the full CSV with both receipt-number columns is on the same page. Methodology is written out in full — universe, exclusions, the identity rules, how total borrowings are summed, preliminary vs filed figures.
[https://accidentalorder.com/en/data/adjusted-valuation/](https://accidentalorder.com/en/data/adjusted-valuation/)
Happy to answer questions about the pipeline, and if you spot a row that looks wrong, the receipt link is right there — tell me and I'll fix it.
