Performance Chasing

Performance chasing is moving money toward whatever has done well lately and away from whatever has not, on the implicit assumption that the recent record predicts the next period. It is a behavior rather than a strategy, and the mental habit that produces it is recency bias, which has its own page and which describes the general tendency to over-weight recent experience when forming an expectation. This page is about what the behavior is and what it does to a portfolio. It is also called return chasing. The behavior should be distinguished from momentum investing, which uses past returns deliberately, systematically and on defined horizons; the resemblance is superficial and the difference is discussed below.

The decision rule feels reasonable because in almost every other domain it is reasonable. A restaurant that has been excellent for three years is a good bet for Friday. A surgeon with a strong recent record is a sensible choice. Investment returns are the unusual case where a good run can itself be the reason to expect less: the price has already risen, so the same claim on the same future earnings now costs more. Nothing in ordinary life trains anyone for a domain in which quality and price move together and only one of them is visible on the chart.

The mechanism by which chasing costs money is worth stating precisely, because it is frequently described as though the investor simply picked bad funds. The problem is not usually selection. It is presence. A fund's reported return assumes one dollar was invested at the start of the period and held to the end of it, with distributions reinvested, which is the correct way to measure the manager's result because it removes the effect of money arriving and leaving. An investor's own result depends on how much was invested at each moment. If money arrives after a strong stretch and leaves after a weak one, the two numbers separate, and both are computed correctly. The distinction is the one between time-weighted return and money-weighted return, and performance chasing is the behavior that drives a wedge between them.

A measured gap of this kind between fund returns and investors' dollar-weighted returns exists and is reported annually by fund research firms. The size of it is not settled: credible published estimates of how much of the gap is attributable to poor timing differ by more than tenfold, and the firm that publishes the best-known measurement cautions against treating it as a proxy for the average investor's return, noting that ordinary and sensible practices such as contributing from every paycheck can open a gap on their own. This page therefore prints no figure for what chasing costs. The pattern is not in dispute; the magnitude is.

What survives the disagreement about size is the cross-sectional pattern, and that is the actionable half. The measured gaps are smallest for broadly diversified all-in-one allocation funds, which are difficult to trade in and out of piecemeal, and largest for narrow, volatile categories, which invite it. More striking, the same underlying strategy has shown a materially different gap depending on whether investors held it as a traditional fund or as an exchange-traded fund. That points at how a holding is used rather than at what it contains, which is a more useful conclusion than any headline number: the fund choice and the trading behavior are separate decisions, and the second one is doing much of the damage.

Several ordinary features of the investing environment push in the same direction, without anyone intending it. Fund flows follow returns, so the crowded funds are the ones that just did well. New funds and new exchange-traded funds are launched into categories that have recently performed, because that is when they can be sold. Ratings and rankings are computed from past returns by construction. And a workplace plan menu or a brokerage "top performers" screen is a list sorted by exactly the variable the investor should be least moved by.

The honest counter-argument deserves stating rather than dismissing. Momentum is a documented pattern in academic finance, so "recent returns carry information" is not a claim that can be waved away. But the documented effect operates on specific horizons, applies systematically across many securities with disciplined rebalancing, and is implemented as a rule decided in advance. A household moving its retirement contributions into last year's best fund shares none of those properties. It is a single, undiversified, discretionary decision made at the moment enthusiasm is highest, which is close to the opposite of the systematic version.

The behavior follows a repeating four-step loop, and each step is individually reasonable.

1. A category or fund has a strong stretch, and becomes visible: in rankings, in coverage, in conversation, in a plan menu sorted by return.
2. Money moves in, at prices that already reflect the strong stretch.
3. Results come back toward the ordinary, which feels like the fund breaking rather than like the run ending.
4. Money moves out, into whatever is now leading, and the loop restarts.

A hypothetical example, showing why the fund's number and the investor's number can both be right. A fund returns 40 percent in year one and falls 25 percent in year two. Over the two years together, one dollar held throughout becomes 1.40 times 0.75, which is 1.05, so the fund reports a two-year gain of 5 percent and can say so accurately.

Ana invests $50,000 at the start of year two, having seen the 40 percent. Her money is present only for the fall, so $50,000 times 0.75 leaves her with $37,500, a loss of 25 percent. Nothing has been misreported. The fund's two-year figure describes a dollar that was there for both years; Ana's describes a dollar that was there for one.

The countermeasures are all mechanical, because a rule decided in advance is not available to be re-argued at the moment it is tested:

• A written target allocation, so the question at any review is whether the portfolio matches the target rather than which holding did best.
• Contributions on a fixed schedule into fixed proportions, which removes recent performance from the buying decision entirely.
• Rebalancing on a date or a drift threshold, which mechanically sells what has risen and buys what has fallen. That it feels wrong is the point: it is the exact inverse of the chasing loop.
• Looking less often, since each review is an opportunity to re-forecast from the newest data.

Pros

• The impulse behind it is a normally reliable inference, that recent performance signals quality, so recognizing it does not require assuming anyone is careless.
• The behavior is unusually easy to detect in one's own record, because the dates of purchases and sales can be compared against what preceded them.
• The countermeasures are cheap, mechanical and do not require any forecast.
• Momentum research means the underlying idea is not simply false, which makes the distinction between a systematic rule and an impulse worth understanding rather than dismissing.

Cons

• Buys after prices have risen and sells after they have fallen, which is the order that does the most damage.
• Generates trading costs and, in a taxable account, short-term gains taxed at ordinary rates.
• Concentrates a portfolio into whatever recently performed, which is often a single narrow category, so it usually reduces diversification at the same time.
• Cannot be measured against the fund's published return, since the fund's number describes a dollar held throughout, so an investor may never see what the behavior cost them.
• The specific dollar cost commonly quoted for it is disputed, so the behavior is frequently argued against with a figure weaker than the argument itself.