Adjusted vs Unadjusted Prices: What Splits and Dividends Do to a Chart

Usually neither site is wrong: they are applying different adjustment conventions to the same raw record. Raw (unadjusted) prices are what actually printed on each historical day; adjusted prices rescale that history so it remains comparable after the company changed its own share arithmetic — splitting its stock, or paying out cash as dividends. This article teaches the two adjustments, the arithmetic of each, and the reading skills that follow: why unadjusted long-run charts contain cliffs that never happened to shareholders, why performance calculated without dividend adjustment understates what an investor actually earned, and why "the price in 2010" is a question with more than one honest answer. The corporate actions themselves — why companies split, how dividends work — are the equities and corporate-actions pillars' territory; this article is about what they do to the data.

Split adjustment: same pie, different slices

A stock split changes the share count without changing the business: in a 2-for-1 split every shareholder wakes up with twice the shares at half the price, and the total value is untouched. The raw price series, though, now contains a 50% overnight "drop" that no investor experienced — and a company that has split several times over decades has a raw history full of such phantom cliffs, with early-years prices that look absurdly high. Split adjustment repairs this by dividing all pre-split prices by the split ratio (and multiplying volumes by it), so the series shows what the price path would have looked like had today's share count always existed. The arithmetic is exact and uncontroversial — a share that traded at a raw $300 before a 2-for-1 split shows as $150 split-adjusted — and essentially every chart you normally see is at least split-adjusted, because unadjusted long histories are unreadable. Two literacy notes follow. First, split adjustment is why historical per-share figures quoted from old sources ("the stock was $600 in 1995") often don't match today's chart: both are right — one is raw, one is restated in today's share units. Second, reverse splits adjust identically in the other direction, and the same logic covers other share-arithmetic events (bonus issues, and — with somewhat messier conventions — rights issues and spin-offs, where data vendors' treatments genuinely differ and are worth checking in any serious historical work).

Dividend adjustment: price return vs total return

Dividends are trickier, because unlike a split, a dividend really does move the price: on the ex-dividend date, the share trades without the right to the declared payment, and — mechanically, all else equal — the price opens lower by roughly the dividend amount, because that cash has left the share and is on its way to the holder. The raw price series records this small step down; over decades of quarterly payments, those steps compound into a large gap between what the price did and what a shareholder earned. Hence two distinct series. The price-return series (split-adjusted only) answers "what did the quoted price do?" The total-return series (adjusted for splits and dividends, conventionally by assuming each dividend was reinvested in the stock on its ex-date) answers "what did an investment in this stock earn?" The gap is not academic: for steady dividend payers over long horizons, reinvested dividends historically account for a large share of total return, and comparing a dividend-paying stock's price chart against a non-payer's is quietly unfair to the payer — the payer has been handing part of its return out in cash the price chart doesn't show (the dividend-policy article covers the decision behind the payment, and the buyback article the alternative route). The same distinction runs through index data (major index families publish price-return and total-return versions, and headline index quotes are conventionally price-return — the index-methodology article picks this up) and through every performance claim a reader will ever evaluate: which series is the first literacy question, because a "10-year return" can differ by tens of percentage points between the two, both honestly computed.

Reading charts across sources: the practical decoder

The conventions explain most cross-source chart disagreements, and a short decoder covers the rest. Which adjustments does this source apply? Common configurations: raw (rare, mostly professional contexts); split-adjusted only (typical for "price" charts); split-and-dividend-adjusted (typical for "adjusted close" columns and performance tools). The same instrument's "close" on one source and "adjusted close" on another can legitimately differ on every historical date. When was the series last re-based? Adjusted history is recomputed after each new event — yesterday's adjusted 2010 price changes slightly after today's dividend — so downloaded historical data goes stale in a way raw data doesn't: a subtle but real trap in spreadsheets and backtests. Which convention fits which question? Stated as mechanics: raw prices answer "what printed that day" (the right series for checking a historical quote or an old news story); split-adjusted answers "how has the quoted price travelled in today's units" ; total-return answers "what did holding it earn" (the right basis for any performance comparison — and the one this portal's risk-and-return foundations implicitly assume). And one forward pointer: adjustment fixes the arithmetic of the companies in the data — it cannot fix which companies are in the data at all, which is the separate, larger honesty problem (survivorship bias) this pillar closes on.