The Behavior Gap Ch. 2: Two Ways to Measure, Only One of Which Is About You

Investment Background

There are two legitimate ways to measure the return on an investment, both correct, answering different questions. Almost every investor has seen only one of them, and assumes it answers the other's question.

This chapter is about what the divergence between them reveals — not about how to compute either one. The formulas are a spreadsheet function and belong in a manual. The interpretation is the part with money attached, and it is the part nobody explains.

The distinction is simple enough to state in one line each:

• Time-weighted return asks: how did this strategy perform?
• Dollar-weighted return asks: how did my money do?

When those two numbers are far apart, something has been learned — and what has been learned is not that the fund was good or bad.

The Wall Street Translation

Why the Industry Reports One and Not the Other

Time-weighted return deliberately strips out the effect of deposits and withdrawals. That is not a flaw. It is the entire design goal, and for its intended purpose it is exactly right.

A fund manager controls the strategy. They do not control when you deposit. Grading a manager on a number that moves when unitholders happen to add money would be grading them on something outside their control — so the industry standard removes that effect and reports the pure performance of the strategy.

This is correct, defensible, and the right number for the job it was built for. It is also, for you, a number about someone else's performance.

	Time-weighted	Dollar-weighted
Question answered	how did the strategy do?	how did my money do?
Sensitive to my deposits?	no, by design	yes — that is the point
Right tool for	judging a manager or fund	judging my own outcome
Who publishes it	every fund, everywhere	nobody

The asymmetry in the final row is not a conspiracy. A fund cannot compute your dollar-weighted return, because it does not have a single one to report — there is a different one for every investor, determined by each person's own deposit history. The number is unpublishable in principle, not withheld in practice.

But the consequence stands regardless of intent: the number describing the strategy is everywhere, and the number describing you exists nowhere until you produce it.

What a Large Divergence Actually Tells You

This is the interpretive core of the chapter, and the part worth reading carefully.

When the two numbers differ substantially, the temptation is to conclude the fund disappointed. The arithmetic says something different and more useful: the divergence is generated by the relationship between your contribution timing and the return path.

Three cases, and only three:

Dollar-weighted materially below time-weighted. More of your capital was present during weaker stretches and less during stronger ones. The strategy delivered; you were underweight when it did. This is the common case, the population-level case, and the one the rest of this book is about.

Dollar-weighted materially above time-weighted. More of your capital was present during stronger stretches. Worth being honest about: this is usually luck rather than skill — most often it means you happened to be contributing steadily into a decline, and were rewarded for continuing rather than for predicting. Chapter 3 argues that this is what disciplined automatic investing looks like from the inside.

The two roughly equal. Your timing was approximately neutral. For most investors this is the realistic target, and it is achieved by contributing on a schedule rather than by forecasting well.

Notice what is absent from all three cases: any information about whether the fund is good. Time-weighted return answers that question. Dollar-weighted return cannot, and reading fund quality out of it is the specific error this chapter exists to prevent.

The Misattribution That Follows

A disappointing decade demands an explanation, and the wrong one is close at hand.

An investor compares a remembered fund return against their actual balance, finds a shortfall, and concludes the fund failed. The response is to switch funds — and the switch is very often executed at precisely the moment described in chapter 4, converting a timing shortfall into a realised loss and starting the same cycle in a new vehicle.

The two numbers are diagnostic, and pointing them at the wrong subject produces a specific and costly mistake.

Observation	Wrong conclusion	What the arithmetic actually supports
My balance lags the published return	the fund underperformed	my capital was underweight during the strong stretches
A friend earned more in the same fund	they picked a better share class	their contribution timing differed from mine
My return improved after switching	the new fund is better	the market path changed; the vehicle may be irrelevant

The right-hand column is not more flattering than the left. It is simply the one the numbers support — and it points at a fixable input rather than an unfixable grievance.

Division of Labor With the Rest of the Library

This chapter is the one with mechanics-drift risk in this book, and the boundary is worth stating explicitly.

What this chapter does not do: teach you to compute an internal rate of return, explain the mathematics of cash-flow discounting, or walk through spreadsheet functions. That is instrument mechanics, and the standing editorial rule of this library excludes it. Any competent reference or the XIRR function covers it in a paragraph.

What this chapter does: establish that two numbers exist, that only one is published, that only the unpublished one describes you, and that their divergence is a diagnostic about your own timing rather than a verdict on the fund.

Question	Book that owns it
How do I judge whether a fund is any good?	winning-the-losers-game-ellis, a-random-walk-down-wall-street
What is inside an index product?	index-fund-machine
How should I think about a manager's track record?	what-works-on-wall-street
Which number describes me rather than the fund?	This book ch02

Executable Trading Rules

1. Compute your dollar-weighted return once, then compare it to the fund's published figure. The gap between them is the subject of this book, denominated in your own money. Until you have both numbers you are reasoning about a quantity you have never measured.
2. When the two diverge, interrogate your contribution timing before you interrogate the fund. The arithmetic points at timing first. Switching funds on this evidence treats a symptom that will follow you to the new vehicle.
3. Do not use dollar-weighted return to grade a manager. It is the wrong tool for that job and will mislead you in both directions. Use time-weighted for the fund and dollar-weighted for yourself.
4. Treat a dollar-weighted return above the fund's as luck, not skill. Repeating a result that came from fortunate timing requires forecasting ability the rest of this library argues you do not have.

Relevance to a Retirement Portfolio

A retirement projection is built on an assumed return, and the assumption is invariably a time-weighted number — pulled from a fund's history, an index's history, or a planning tool's default.

The plan then applies that number to a stream of contributions and withdrawals, which is a dollar-weighted situation. The projection quietly assumes those two are the same. For a disciplined contributor they are close enough. For anyone whose contributions have responded to market conditions, they are not — and the plan has been compounding an optimistic assumption for years.

The correction is not to lower your return assumption out of pessimism. It is to measure the gap once, learn whether your own timing has been costing you, and — if it has — apply the fixes in chapter 5, which are structural and require less attention rather than more.

For a reader holding a low-cost diversified core, this measurement is usually reassuring. Automatic contributions into a broad fund produce a dollar-weighted return close to the time-weighted one, which is the entire argument for automation and the reason this book does not end up recommending anything more elaborate.