Dow Theory

Charles Dow never wrote a book about his theory. He wrote editorials in the Wall Street Journal between 1900 and 1902, died in 1902, and other people assembled his observations into something they called Dow's Theory afterwards.

That matters, because it means there is no single authoritative statement of it. What we have is S. A. Nelson's 1902 book, which first collected the editorials and coined the name, and William Peter Hamilton's 1922 book, written by Dow's successor at the Journal, which turned the observations into a system. Both are out of copyright and hosted on this site.

Dow's central image was tidal. A tide runs for months or years; waves run against it for weeks; ripples come and go by the hour. All three are happening at once, and confusing one for another is how people get hurt.

The practical warning is that a secondary reaction feels exactly like a primary reversal while you are inside it. A two-thirds retracement of a year-long advance is brutal to sit through, and it is not a trend change.

Dow's second idea was that a primary trend is not uniform. It passes through phases defined by who is doing the buying.

1. 1Accumulation. The news is still bad and most people are not interested. Informed buyers take stock from discouraged holders. Prices go almost nowhere, which is precisely why nobody notices.
2. 2Public participation. The trend becomes visible, earnings improve, and the wider market joins. This is the longest phase and where most of the move happens.
3. 3Distribution. The news is uniformly good, everyone is confident, and the people who bought during accumulation are selling into that confidence.

• The averages discount everything. Every known fact, hope and fear is already in the price. You are not competing with the news; you are competing with everyone else's reading of it.
• Volume confirms the trend. Volume should expand in the direction of the primary trend and dry up on reactions. A rally on thinning volume is suspect.
• One average must confirm the other. Dow watched the industrial and rail averages. A new high in one, unmatched by the other, was not a signal.
• A trend is assumed to continue until there is clear evidence of reversal. The burden of proof is on the reversal, not the trend.

The confirmation rule is the one that made Dow Theory distinctive. His reasoning was economic rather than statistical: if factories are genuinely producing more, the railways carrying the goods must eventually show it too. A rise in one without the other described a story that did not add up.

Dow Theory signals late. By the time a primary reversal is confirmed under its rules, a substantial part of the move has already happened. Practitioners have always known this; it is a framework for understanding what kind of market you are in, not for catching turns.

It is also built on an economy that no longer exists. The industrial-and-rail confirmation made sense when goods moved by rail. It transfers awkwardly to a service economy, and not at all to a market with one index.

What survives is the part that was never about railways: that trends exist at multiple scales at once, that volume should agree with price, that the crowd is most confident at the worst moment, and that a single indicator agreeing with itself proves nothing. Every framework on this page inherits something from that.