跳至主內容 Skip to main content

Tools Classroom, Lesson Five: How to Read Past Returns Without Fooling Yourself — Cumulative vs Annualised, the Calendar-Year Traps, and the “Since Launch” Mirage

2026-11-20
Marcus Tang
Marcus Tang

Welcome back to the mpf.hk Tools Classroom. Lessons one to four built your tray and taught you to read a fund page cell by cell and to catch a column that had slipped. This lesson stays on the return rows themselves — the numbers most readers look at first and misread most often. Phase One of full portability opens on 21 December, 31 days from today, and each eligible member gets one transfer per calendar year. Our specimen is the same fund as lesson three, the Manulife MPF Hong Kong Equity Fund under the Manulife Global Select (MPF) Scheme, because you already know its page. Every figure below was checked on the live page on 20 November; the page’s data is as of 30 June 2026.

Two rows, one table — and they are not the same unit

Scroll to Trailing Returns and you will see two rows under the same column headings: Cumulative and Annualised. For this fund they read:

3 Months1 Year3 Years5 Years10 YearsSince LaunchYTD
Cumulative+0.65%+7.95%+31.98%−21.95%+42.86%+141.01%−1.94%
Annualised——+9.69%−4.84%+3.63%+3.50%—

The cumulative row is the total move over the whole period. The annualised row is the same move restated as a compound rate per year — the steady yearly rate that would produce that total. Five years at −4.84% a year compounds to −21.95% in total; three years at +9.69% a year compounds to +31.98%. The two rows agree with each other; they are two views of one fact.

Three reading rules follow. First, never compare a cumulative number with an annualised one. “+42.86% over ten years” and “+3.63% a year” are the same result — ranking a fund’s ten-year cumulative against another fund’s annualised figure is comparing a decade with a year. Second, the annualised row is the honest comparator across periods: one year at +7.95%, three years at +9.69% a year, five years at −4.84% a year, ten years at +3.63% a year. Read down the annualised row and the story is plain — a strong recent run sitting on a five-year hole that has not yet been filled. Third, notice the blanks: the annualised row shows a dash for 3 months, 1 year and YTD, because annualising a period of a year or less would be inventing a rate that never happened. A blank is the page refusing to guess. Respect it.

The one-year leaderboard is a photograph, not a film

The homepage’s first ranking table sorts by one-year return, and right now it is spectacular: the Haitong Korea Fund at +84.85%, the BEA (Industry Scheme) Asian Equity Fund at +82.03%, the Sun Life MPF Asian Equity Fund at +58.15%. All three are real figures, all three are risk class 6 or 7, and all three will tempt you to chase them. Before you do, read the five-year row we verified in lesson four: BEA’s five-year annualised return is +10.85% a year, Sun Life’s is +4.63%. A fund can double in a year and still average under 5% a year over five, because the one-year number says nothing about the four years before it. The homepage also runs a second table sorted by five-year return, and a third sorted by fee — the page itself offers the longer view, but the one-year table sits on top. If you only ever read the top table, you are reading the noisiest number on the site. Use the one-year column to spot what moved; use the five- and ten-year annualised columns to judge whether it lasted.

The calendar-year table: five years, and the worst year is not in it

Under Calendar Year Returns the page shows five full years:

20252024202320222021
Return+32.02%+14.12%−14.97%−22.39%−15.79%

This table is the best bad-year detector on the page, and it needs two cautions. First, compound the bad years before you forgive them. 2021, 2022 and 2023 together are not “about −53%” — they compound: HK$100 became HK$84.21, then HK$65.43, then HK$55.56, a fall of −44.4% across three years. The two good years that followed, +14.12% and +32.02%, lift that HK$55.56 to only HK$83.72 — still −16.28% over the five calendar years 2021–25. A +32% year feels like a recovery; against a −44% hole it is a partial one. Second, the table only covers five years, so the worst year on record may not be in it. The Fund Commentator on this very page calls 2018 (−18%) the worst calendar year on record — yet the table’s own worst year, 2022 at −22.39%, is worse than that, and 2018 itself sits outside the five-year window entirely. Lesson three already caught that mismatch; the lesson here is structural: a five-year calendar table cannot show you a bad year from eight years ago. If your horizon is twenty years, five calendar years is a sample, not a history.

And do not expect the calendar table to add up to the trailing five-year figure. The trailing five years stand at −21.95% cumulative; the five calendar years 2021–25 compound to −16.28%. The gap is not an error — the trailing window runs to 30 June 2026 and includes the first half of 2026, which the calendar table does not. Different windows, different answers; always check which window a number belongs to before you compare it with another.

YTD: the shortest window on the page, and the most quoted

The trailing table’s last column is YTD — year to date. For this fund it reads −1.94%, sitting next to a one-year return of +7.95%. Both are true at once: the fund rose over the twelve months to 30 June 2026, and fell in the first six months of 2026. YTD is a partial year — here, January to June — and it moves with every month that passes. Never rank funds on YTD, never annualise it, and never treat a negative YTD as a verdict on a fund whose one-, three- and ten-year numbers tell a different story. It is a progress report on a year that is not finished.

“Since Launch”: the biggest number on the page, and the least useful

The trailing table’s most eye-catching cell is Since Launch: +141.01%. It looks like a triumph — your money more than doubled. Read the annualised cell underneath it: +3.50% a year. The fund launched on 1 December 2000, so that +141% took nearly 26 years to build, at a compound rate lower than the ten-year annualised figure of +3.63%. Since-launch figures span different eras for different funds — a fund launched in 2000 and a fund launched in 2012 are not running the same race — so the cumulative since-launch column should never be ranked, and the annualised since-launch column only tells you the average speed over a lifetime that may include markets no longer relevant. Use it to sense the fund’s age and its long-run average. Nothing more.

The DCA table: the same fund, a different investor

Below the calendar table sits Dollar Cost Averaging Return — the same fund seen through monthly contributions instead of a single lump sum. The contrast is the sharpest on the page:

1 Year3 Years5 Years10 Years
Lump sum, cumulative+7.95%+31.98%−21.95%+42.86%
DCA, cumulative−2.37%+22.75%+19.00%+11.93%
DCA, annualised—+13.76%+6.89%+2.22%

Over one year the lump sum gained +7.95% while the monthly contributor lost −2.37% — a 10.32-point gap, because money paid in month by month bought in at prices that then fell. Over five years the sign flips entirely: the lump sum lost −4.84% a year while the monthly contributor gained +6.89% a year, because regular buying through the 2021–23 falls bought more units when they were cheap. The 2022 calendar year shows the mechanism in miniature: lump sum −22.39%, DCA −4.44%. Neither table is the “true” return — each is true for a different way of investing. If you contribute monthly, as most MPF members do, the DCA table is your table; the lump-sum rows describe someone else’s experience. And note the N/A cells: DCA shows nothing for 3 months or YTD, because the page will not pretend a few monthly payments make a meaningful return series.

FER History: a one-line summary, not a decade of fee cuts

Finally, scroll to Fees & Charges. The fund expense ratio stands at 1.77859%, and beneath it a single line reads “FER History ▼ 0.01pp since 2021 (−0.6%)”. That line is the whole of the page’s fee history: one direction, one window, one number. It is not a chart and it does not span a decade — the window is since 2021, five years, and the move is 0.01 percentage points, which is essentially flat. A fund can hold its fee steady for five years and still be expensive; the history line tells you direction, the level tells you cost. This fund’s level is 1.77859% against the Hang Seng Index ESG Fund’s 0.92494% from lesson one — a gap of 0.85 percentage points, about HK$2,134 a year on a HK$250,000 balance, and on the desk’s illustrative basis (HK$5,000 a month at 7% a year before fees, compounded monthly) roughly HK$37,065 over 10 years, HK$222,317 over 20 and HK$759,409 over 30. When you read a fee history, ask two questions: how far has it moved, and from what level? A flat line at a high level is not a fee cut.

Your next step

This week, take the fund you already hold and read its Trailing Returns table the way this lesson reads it: annualised down the rows, calendar years compounded across the bad ones, YTD ignored as a ranking, since-launch admired and set aside, and the DCA table if you contribute monthly. Then do the same for one challenger in the same category and compare only matching windows — five-year annualised against five-year annualised, 2022 against 2022. Lesson six will look at what the two funds actually own: the top-10 holdings, and how to spot two “different” funds buying the same basket twice. And the standing reminder: 21 December is an opening date, not a deadline. A return you have read properly beats a transfer you have not. See you in lesson six.

Figures verified on the live mpf.hk homepage and the Manulife MPF Hong Kong Equity Fund detail page on 20 November 2026; that fund page’s data is as of 30 June 2026, the browse page’s data as of 30 September 2026. The fee illustration assumes HK$5,000 a month and a 7% annual return before fees, compounded monthly. The worked example teaches the reading method and is not a recommendation for or against any fund.

    Related articles

    funds to compare