Drawdown and Recovery Time, and How to Read Them

Why this is the statistic worth learning

Ramesh can read a standard deviation and it tells him nothing he can picture. A drawdown he can picture immediately: this is how far it fell from the best it had been.

That is the case for learning this one properly. Of all the risk numbers on a fact sheet, drawdown is the only one stated in the terms a person actually experiences — a fall from a high they remember, lasting a length of time they had to sit through. The rest are summaries of a price series; this is a description of an episode.

Which makes it worth knowing exactly how it is computed, because there is a choice hidden in it.

How a drawdown is measured

Take the running high — the highest value reached at any point up to now. A drawdown is the fall from that high to a later low, expressed as a percentage of the high.

Maximum drawdown is the worst such fall in a period. It is one observation: the single worst episode inside whatever window was chosen.

Two consequences follow directly. A longer window will generally produce a worse maximum drawdown, so the figure is not comparable between holdings measured over different periods. And because it is the worst thing that happened, it is not the worst thing that can happen — a window without a crash in it reports a comfortable number for something that would have been savaged by one.

The two clocks, and why the difference matters

Recovery time is where the flattering convention lives.

Measured from the low point, recovery time is how long it took to get back to the old high after the bottom. This is the figure usually printed, and it is shorter.

Measured from the high point, it is the fall and the recovery added together — the whole time between the last good day and the day of being level again.

The second is the one that describes a person's experience, and the reason is simple: nobody can identify the low point while it is happening. Somebody invested at the high did not step in at the bottom; they were already there, and they waited both halves. The from-the-low figure describes the experience of an investor who bought precisely at the bottom, which is an investor who does not exist.

On the Indian record the two conventions differ substantially, and the falls are laid out both ways in why stock markets crash, which also carries the distribution of how long the wait actually ran.

The asymmetry that makes drawdowns worse than they look

A fall and the rise needed to undo it are not the same size.

Lose a quarter and you need a third to get back. Lose half and you need to double. Lose two-thirds and you need to triple. The percentage required to recover grows faster than the percentage lost, and it grows without limit as the loss approaches everything.

This is why the depth of a drawdown matters disproportionately rather than proportionately, and it is the same arithmetic that makes an average return overstate what money did — the mechanism worked through in why an average return can mislead you.

It has a practical consequence: avoiding the deepest falls is worth more than it appears from the percentages, which is exactly the argument used to sell protection strategies. Whether any of them reliably deliver it is a separate question, and the honest state of the evidence is in does past downside protection predict future protection.

Reading a drawdown figure somebody shows you

Four questions, in order.

Over what window? Without it the number means nothing and cannot be compared with any other.

From the low or from the high? If the recovery time is not defined, assume the flattering one.

Of what? A drawdown of a fund, of an index, or of a portfolio including its safer holdings are three different things. Portfolio-level drawdown is the one that bears on you.

Daily or monthly data? A drawdown computed from month-end values misses everything that happened inside a month, so it will report a milder fall than actually occurred. This is not a trick; it is what the data allowed. It is worth knowing when comparing two figures computed differently.

What it still does not tell you

It does not say how likely a fall of that size is, because one observation carries no probability.

It does not say anything about your risk, which depends on when you need the money and whether you could be forced to sell during the fall — the distinction in risk against volatility.

And it does not describe the ordinary condition of holding the thing. Maximum drawdown is the worst episode; most of the time is spent somewhere in between, below a previous high but not at the bottom, and that is where an investor actually lives.

What to take away

Drawdown is the most readable risk number available, provided you know the window it came from and which clock the recovery was measured on.

For planning, use the from-the-high figure. It is longer, less flattering and it is the one that describes what you would have had to sit through — which is the only version that helps you decide whether you could.