Buy and Hold Against Market Timing: What the Evidence Shows

Scope. This article sets out what each position claims and what can be established arithmetically about the hurdle a timer faces. It does not recommend either, supplies no indicator, rule or signal, and tells no reader when to hold or change anything. Deciding how much of a portfolio to hold in what, and when circumstances warrant changing it, belongs to Pillar 31 and is a different question from the one here.

Almost everyone does some of both, and the argument between them is usually conducted as though they were incompatible identities rather than points on a scale.

The useful version of the question is not which is better. It is how good a timer would have to be for timing to be worth doing — because that has an answer, and the answer is arithmetic rather than opinion.

The threshold

The question was formalised in 1975. William Sharpe examined a timer switching annually between equities and cash, and concluded that such a timer would need to be correct about 74% of the time — roughly seven years in ten — merely to match a buy-and-hold outcome.

The mechanism behind that threshold is the part worth understanding, and it is an asymmetry. When a timer correctly moves to cash before a bad year, the gain is the difference between a modest cash return and a negative one — real but bounded. When a timer wrongly moves to cash before a good year, the loss is the difference between a modest cash return and a large positive one. The penalty for being wrong is on average larger than the reward for being right, so a coin-flipping timer does not break even — they lose, and the required accuracy is far above half.

And the timer has to be right twice, not once. Every exit implies a re-entry, and the re-entry decision is at least as hard as the exit — arguably harder, because it is made while the reasons for leaving still feel compelling. A timer with a 74% chance of being right on each of two independent decisions is right on both about 55% of the time. This portal does not treat that as a refutation, since the decisions are not independent and no real process resembles this model. It is offered as the reason the threshold in the literature is a floor rather than a ceiling.

What time out of the market costs

The cost of standing aside is not the drop avoided. It is the return foregone, and on the portal's canonical parameters — 9.0% for equities, 4.0% risk-free — that gap is the 5.0% equity risk premium for every year spent out.

The argument everybody quotes, and what is wrong with it

The most repeated case against timing is that missing a handful of the best days over a long period devastates the outcome. The arithmetic is real — returns are concentrated in very few sessions.

But the argument as usually deployed is weak, and a reader should know why. It compares against a strawman: no timer proposes to miss the best days, and the relevant comparison is a timer who misses some best days and avoids some worst days, not one who misses only the good ones. And the best and worst days cluster together, arriving in the same volatile stretches, so a strategy that avoids the worst will very often miss the best as well. The honest version of the statistic is that both tails are concentrated and difficult to separate — which supports the threshold argument above without needing the rhetorical framing.

What each side is actually claiming

The buy-and-hold position is not that markets are unpredictable. It is that the accuracy required to profit from prediction after costs is higher than most people can sustain, and that the cost of being out accrues continuously while the opportunities to be right arrive rarely.

The timing position is not that anyone can call turns. Its stronger forms hold that expected returns vary over time in ways that are partly observable, so exposure need not be constant — a claim about conditional expectations rather than about forecasting events.

Those are different propositions and the evidence bears on them differently. This portal reports the threshold, the asymmetry and the cost of absence, and does not tell any reader which position to hold or what to do about it.