Home / Sectors
Taiwan stock sectors
1971 Taiwan-listed tickers grouped into 34 official sector classifications. Each sector page lists every ticker in it with the P/E, P/B and dividend yield from that day's official-source snapshot.
Valuation levels across sectors
Snapshot date 2026-08-03. Each median covers only the tickers in that sector with a positive value on that date; the sample size is in the right-hand column, and any sector with fewer than five valid observations shows a dash. Rows are ordered by ticker count, not by valuation — this is not a ranking.
How this table is computed
Classifications are taken from official filings, not assigned by us
The sector of each ticker comes from the industry field on its official monthly revenue disclosure. We do not reclassify or merge adjacent sectors — the single exception being two source labels that share one English classification — so the grouping here matches what you see on the filings themselves.
Per-ticker snapshot values first, then aggregate
For each sector we take the P/E, P/B and dividend yield that were available for every constituent on the snapshot date, then take the median of those. Every input is visible row by row on that sector's own page, so you can recompute the figure rather than trust it.
Meaningless observations are dropped, and the drop is disclosed
P/E values at or below zero — meaning a trailing-four-quarter loss — and undisclosed values are excluded. The resulting sample size sits in the right-hand column next to the full ticker count, so you can judge representativeness instead of being told it is fine.
Too few observations means no number
Any sector with fewer than five valid observations shows a dash. The median of three stocks is a coincidence, not a sector level, and once published it gets quoted as fact. A missing cell beats a wrong one.
Four terms to read this page
Frequently asked questions
Who defines these sector classifications?
They come from the industry field that companies file on their official monthly revenue disclosures — not from anything we assign. Around 250 listed tickers carry no industry label in the source, and we neither force them into a sector nor invent an unclassified page. Two source labels map to the same English classification and are merged into one page rather than split into two near-identical ones.
Why medians instead of averages?
P/E distributions inside a sector have very heavy tails. One company that just crossed back into profit can carry a four-digit P/E, and an average would then be that single name rather than the sector. The median says where half the sector sits, which is the question worth asking at this level.
How are negative or missing P/E values handled?
They are excluded from the median and reflected honestly in the sample-size column. A P/E at or below zero means the company lost money over the trailing four quarters — that is not cheapness, it is the ratio losing its meaning. Blanks mean nothing was disclosed that day, and nothing here is ever filled in or carried forward.
The sample is much smaller than the ticker count — is the median still useful?
The two columns sit side by side precisely so you can judge that yourself. The gap is mostly loss-making companies plus names that had not disclosed. Research-heavy sectors such as biotech show the widest gap, so read their median as describing only the profitable half. Below five valid observations we publish no number at all.
Can I use this table to pick a sector?
No — we publish no signals and no recommendations. This is a cross-section on one snapshot date with no judgement about whether a level is cheap: a low multiple may correctly price a decline, and a high one may be fair for growth. To judge cheapness you want a ticker against its own history, which is what the valuation river page does.