In our first money-market piece we covered Xeon, which «becomes» the overnight rate via a swap. C3M gets there by the opposite road: it buys real government bills, at very short maturities, rolling them every week. Same category, opposite philosophy — with measurable differences in costs, taxes and risks. We measure them all, over nine years of data.
- What it is: eurozone government bills under 6 months, actually owned, rolled weekly. Since 2009, €0.94bn, SRRI 1.
- True cost: 0.14% TER; the KID estimates 0.25% all-in, but measured tracking says ~15bp a year.
- Taxes (Italy): 12.64% effective — and above all stable, unlike the synthetic.
- Risks: the category’s — the negative-rate erosion — nearly ten years below the peak (September 2014 → September 2022, −4.86%), with recovery in July 2024; no swap, no securities lending.
- Vs Xeon: slightly behind gross and catching up; nearly even net of taxes. Same category, opposite philosophies.
Every chart and table on this page is an original Rebalix computation on official data: the issuer, FTSE Russell and the issuers’ tax documents — not a rehash of marketing factsheets.
One email when a new article is out, with Rebalix product news at the end. Nothing else: no ads, no data shared with third parties, one click to unsubscribe.
1. What a money-market ETF is (60-second recap)
A money-market ETF is a tool for «parking» cash: it follows the euro’s very short-term interest rates. When ECB rates rise it earns more; when they fall — or go below zero, as they did from 2015 to mid-2022 — it earns little or slowly loses. It is not a deposit account: it is an exchange-traded fund whose price moves every day — very little, but it moves. The full recap, with a yield calculator, is in our Xeon article.
2. What they are for, according to research
Academic finance has a name for very short government paper: near-money. Stefan Nagel (Quarterly Journal of Economics, 2016) measured the premium the market pays for assets liquid enough to be almost cash: holders accept a slightly lower yield in exchange for the ability to exit at any time. Krishnamurthy and Vissing-Jørgensen (Journal of Political Economy, 2012) showed investors consistently pay up for the safety and liquidity of government debt. Translated from the academic: these instruments exist to store money that must stay available — not to grow it. European law says the same: EU Regulation 2017/1131 defines money-market funds as tools for «short-term liquidity management». Using them for long-term goals is a category error — true of C3M, Xeon and any money-market fund.
- Designed for: storing cash over short horizons; parking money between deployments; keeping interest-rate risk minimal (~0.2 duration).
- Not designed for: long-term growth; beating equities; protecting from inflation over long horizons — other instrument categories exist for that.
3. The identity card
| Item | Value |
|---|---|
| Name | Amundi Euro Government Bond 0-6 M UCITS ETF Acc |
| ISIN | FR0010754200 |
| Ticker | depends on the exchange — see venues below |
| TER (management) | 0.14% — but see the costs section |
| Assets (AUM) | €0.94bn |
| Replication | Physical (actually owns the bills) |
| Securities lending | None (per prospectus) |
| Distribution | Accumulating |
| Index | FTSE Eurozone Government Bill 0-6 Month Capped |
| Launch | 22 June 2009 (French FCP) |
| Risk (SRRI) | 1 of 7 |
According to the official lists published by the brokers (downloaded and cross-checked by us), these share classes of the fund appear tradable with zero fees:
FR0010754200C3M- Directazero within the savings plan (buys)
- Finecozero on buys (single orders and savings plan) — promo list renewed monthly: check it is still listed
- Moneyfarmzero within the savings plan (buys), agreement until 15 May 2027
- Moneyfarmzero on single buys from 1,000 €, agreement until 15 May 2027
- XTBzero (up to €100,000 monthly volume, unreachable for a retail plan); fractions available in plans
Official lists verified on 15 October 2026. We record published terms — this is neither an invitation to open accounts nor a recommendation: lists change, the broker’s documents prevail. Trade Republic, Scalable Capital and Trading 212 publish no per-ISIN lists and don’t appear here.
Two rows are worth the whole article. Physical replication: C3M owns the bills, not a contract promising their return — the exact opposite of Xeon (the trade-offs of each route are explained in Physical or synthetic replication). And securities lending: none — we verified it in the prospectus («transactions involving the temporary acquisition and/or disposal of securities: none», April 2026 edition): no extra lending income, but also zero counterparty risk from that door (in fairness: in UCITS swaps that risk is heavily mitigated anyway — collateral, exposure limits, oversight — as our replication pillar explains: the difference is one of nature, not an alarm). Live data source (NAV, assets, authorisations): Amundi’s official API, which we query daily — this page always shows the latest published figure, with its date.
A distinction that trips up even seasoned investors: authorised does not mean listed. The fund is authorised for sale in 7 countries (the UCITS «passport»: regulatory permission to offer it to that country’s investors) but listed on just 3 exchanges — Xetra, Paris and Milan. A Finnish or Dutch investor simply buys it on one of those venues through their broker: no home listing needed. Switzerland is the special case: authorised, yes, but for institutional investors only — Swiss retail cannot buy it.
Method — source: official Amundi NAV and share count · processing: net flows reconstructed by Rebalix (change in shares × same-day NAV) · updates: automatic, monthly.
4. Inside the basket: the «T-bills» of 7 countries
The index buys bills — discount Treasury paper — issued by 7 eurozone states: France, Germany, Spain, Italy, Belgium, the Netherlands and Portugal, each with at least two short-term investment-grade ratings. Maturities under 6 months. The «Capped» in the name is a ceiling: no issuer above 34.5% — without it, France and Italy (the largest bill issuers) would dominate. Amundi publishes the complete basket — 55 instruments at the most recent position date (15/10/2026), each with ISIN and weight, downloadable from the fund page. We check and archive it every day: each new position date lands in the archive — and within a few weeks the archive itself will tell us how often the issuer actually updates. Here, for readability, the country mix and the top 10 holdings:
Method — source: full composition published by Amundi (ISIN and weight of every holding) · processing: per-country aggregation and top-10 ranking by Rebalix · updates: automatic daily check and archiving.
The technical detail that explains half the article: the index rebalances weekly (bills mature constantly and must be replaced). A physical fund tracking it must therefore trade every week. Keep that in mind: it returns in the costs section. Source: the index’s official ground rules (FTSE Russell, v3.3, May 2026 — public document).
Three points of due precision. First: bills are zero-coupon paper — no coupons, bought below par, paying at maturity; today the basket is 100% zero-coupon (measured: none of the 54 holdings carries a coupon), although the prospectus admits, in principle, coupon-bearing government paper with under 6 months to run. Second: duration — the prospectus caps interest-rate sensitivity between 0 and 0.5; in practice the June 2026 factsheet measures a modified duration of 0.22: a sudden one-point rate rise would cost the basket about 0.2% — a fingernail. Third: no «expected dividend» — the fund accumulates and distributes nothing; the number that does exist is the basket’s yield to maturity, 2.38% per the June 2026 factsheet, which tracks ECB rates and is why this fund earns what the money market earns, no more, no less.
Method — source: Rebalix archive of the official compositions · processing: each new composition automatically compared with the last reported state (2–5 point thresholds) · updates: daily check; the register only writes when a threshold trips.
5. Performance — and the true cost of tracking
Method — source: official Amundi NAV and the index as republished by the issuer; XEON from the official DWS NAV · processing: cumulative series rebased by Rebalix, with automatic checks against the declared calendar-year returns · updates: automatic, daily.
| Year | Fund | Index | TD |
|---|---|---|---|
| 2018 | −0.72% | −0.51% | -21 bp |
| 2019 | −0.58% | −0.38% | -20 bp |
| 2020 | −0.60% | −0.47% | -14 bp |
| 2021 | −0.75% | −0.64% | -11 bp |
| 2022 | −0.46% | −0.29% | -17 bp |
| 2023 | +2.89% | +2.99% | -11 bp |
| 2024 | +3.53% | +3.65% | -12 bp |
| 2025 | +2.09% | +2.23% | -14 bp |
| 2026 · partial | +1.61% | +1.69% | -8 bp |
| Whole period (annualised) | +0.71% | +0.85% | -15 bp |
Method — source: official fund and index returns · processing: fund–index gap per calendar year, annualised, measured by Rebalix · updates: automatic, daily.
The takeaway: the KID declares ~0.25% a year in total costs (0.14 management + 0.11 estimated transactions), yet the measured fund–index gap is 15 basis points a year over 9.3 years — essentially the TER alone. Nobody is wrong: that is the difference between a prudent regulatory estimate and a measurement. It is also why, when comparing funds, we look at measured tracking rather than sums of declared costs.
6. Drawdown: what «losing» means for a money-market fund
Method — source: official Amundi NAV since launch (June 2009; 2009–2017 series provided by Amundi ETF at Rebalix’s request), month-end; Eurostat/ONS inflation and ECB FX rates for the real version · processing: nominal and real drawdowns per country computed by Rebalix · updates: automatic, daily.
Read the number for what it is. The negative-rate episode — the window in which the fund’s index returned below zero over three months, December 2014 → November 2022 — is worth −4.86%, touched in September 2022: built not in a crash but in a slow descent lasting about 8 years: the peak dates to September 2014, then below-zero rates eroded a few tenths of a point per year, by construction, until the recovery in July 2024 with the hikes. 118 months below the previous peak. It is also the maximum drawdown of the full history. The negative-rate erosion is not a C3M flaw (Xeon lived the exact same story — its chart is in the twin article): it is what happens to any cash-parking vehicle when the reference rate sits below zero. Inflation is a separate matter — it hits deposits, money funds and idle cash alike — and it lives in the «net of inflation» toggle above: there the descent gets far deeper, because the cost of living adds to the rates.
7. Taxes: where physical changes the game
Fair question: if the basket is all eurozone government paper, why is the white-list share 98.9% and not 100%? Because the share does not photograph today’s securities basket (which really is ~100% bills: 54 holdings plus one cash line, visible in the full composition), but the whole fund on the two semi-annual reporting dates — 30 June and 31 December — averaged (Revenue Agency Circular 11/E/2012). On those dates the fund always carries some operational cash: liquidity for redemptions, receivables in settlement. It is fund property, but it is not a white-list government security — that is the missing percentage point. The same reason makes the share wobble slightly semester to semester.
Method — source: the issuers’ official white lists (our database) and official NAVs · processing: the sale-semester tax rate applied to the same gross return, computed by Rebalix · updates: gross returns daily (latest point = latest common date), tax rates with each new white list.
This chart is why we keep a semester-by-semester archive of white-list documents: Xeon’s rate has swung between 12.87% and 15.46% across 11 semesters — the same gain sold in different semesters paid different taxes, discovered only at the moment of sale. C3M, by contrast, is fiscally predictable: physically holding government bills pins the share at 98-99%. A virtue no factsheet mentions.
8. C3M vs Xeon: head to head
| Year | C3M | XEON | C3M − XEON |
|---|---|---|---|
| 2018 | −0.72% | −0.48% | −24 bp |
| 2019 | −0.58% | −0.50% | −8 bp |
| 2020 | −0.60% | −0.57% | −3 bp |
| 2021 | −0.75% | −0.59% | −16 bp |
| 2022 | −0.46% | −0.03% | −43 bp |
| 2023 | +2.89% | +3.27% | −38 bp |
| 2024 | +3.53% | +3.79% | −26 bp |
| 2025 | +2.09% | +2.23% | −14 bp |
| 2026 (to 15 Oct) | +1.61% | +1.65% | −4 bp |
Gross returns say: Xeon ahead, but with a narrowing gap — from +43 basis points in 2022 to +14 in 2025. Two structural reasons. First: Xeon’s index embeds a +8.5bp annual premium over €STR. Second: the two curves react to rate cycles at different speeds — the overnight adjusts instantly, 0-6 month bills price ECB moves in advance. In practice: when the ECB cuts, the bills already in C3M’s basket (~0.2 duration) book a small capital gain before rolling, while XEON steps to the new rate the same day; when it hikes, C3M is the one with a small adjustment lag. In 2022, amid rapid hikes, overnight was the place to be; as the cycle normalises the gap shrinks. On measured costs: Xeon loses ~10bp a year to its index, C3M ~15 — a few basis points apart, not the 15 the KIDs would suggest. Net of taxes the distance narrows further, as the «what you keep» chart showed, where the sale-semester rate does the work. For scale:2.8 points of lower tax rate, on a ~2.4% gross yield (June 2026 factsheet), are worth ~7 basis points a year — the same order as the measured cost gap: which is why the net race is a dead heat. No verdict, then — two instruments of the same category with different profiles: minimal cost and reactivity (Xeon) versus simplicity, real ownership and tax predictability (C3M).
9. Official documents
- KID (PRIIPs) — the key document: costs, scenarios, recommended holding period (1 year)
- Prospectus — the source of the securities-lending verification (section 7)
- Index ground rules (FTSE Russell) — eligible countries, 34.5% cap, weekly rebalancing
- Amundi product page — NAV, breakdowns and monthly report (search ISIN FR0010754200 on amundietf.it)
Frequently asked questions
What is C3M?
C3M is a physically replicated money-market ETF: it invests in eurozone government paper maturing within 6 months (the «T-bills» of 7 countries), accumulates income and costs 0.14% in TER. It is a tool for storing short-term cash — not for long-term growth — with favourable Italian taxation (~12.64%) thanks to its 98.9% white-list share.
Does C3M pay dividends?
No: it accumulates. Gains stay inside the fund and show up in the share price — no coupon ever lands on your account. The yield quietly compounds into the NAV.
What does C3M hold?
Very short-dated government paper (0-6 months) from 7 eurozone countries — Italian, French, German, Spanish «T-bills» and the like — all zero coupon, plus a small operational cash line. The full basket (55 instruments with ISIN and weight) is published by the issuer and we verify and archive it daily.
Which index does C3M track?
The FTSE Eurozone Government Bill 0-6 Month Capped Index: government bills of 7 eurozone countries with at least two investment-grade ratings, a 34.5% issuer cap, weekly rebalancing. The index ground rules are public (FTSE Russell).
What does C3M really cost?
The TER is 0.14%. The KID estimates ~0.25% a year adding the weekly roll's transaction costs. But the measured cost — the fund-index gap on official NAV, year by year — is ~15 basis points a year over 9.3 years: essentially the TER alone.
How much does C3M yield?
Whatever the euro money market yields, no more, no less: the basket's yield to maturity was 2.38% per the June 2026 factsheet and tracks ECB rates. When rates fall it earns less; below zero (2015-2022) it slowly lost.
Is C3M synthetic or leveraged?
Neither: direct physical replication (it actually owns the bills), no leverage, and per the prospectus no securities lending. The swap counterparty risk typical of synthetic money-market ETFs simply does not exist here.
Is C3M safe?
It is among the least volatile instruments around (SRRI 1 of 7): investment-grade government paper at very short maturities, physical replication with no swap and no securities lending. «Safe» does not mean «guaranteed» though: it is not a deposit account, capital is not protected, and in the negative-rate years it slowly lost value. The risks to understand are below-zero rates and inflation, not crashes.
Can C3M lose value?
Yes, in two measurable ways. Slowly, when euro rates sit below zero: from its September 2014 peak it lost ~4.86% over about 8 years, recovered in July 2024 with the hikes. And in real terms, when inflation outpaces the yield. Sudden crashes, no: with a ~0.2 duration, even a sharp one-point rate rise costs about 0.2%.
How is C3M taxed for an Italian investor?
The «white list» share (government bonds) is 98.9%: the effective rate on gains is ~12.64% instead of 26% (2° sem. 2026). What counts is the share of the semester in which you sell — and C3M's is historically stable, between 98% and 99.5%. Always double-check with your broker.
Where does C3M trade and under which ticker?
On Borsa Italiana and Euronext Paris as C3M, on Xetra as C3MGR. It is authorised for sale in 7 countries (Switzerland: institutional only). Some Italian brokers offer it commission-free: the up-to-date list is in the article.
What is the difference between C3M and XEON?
Same category, opposite philosophies: C3M buys real government bills (physical), XEON gets the overnight rate via a swap (synthetic). Similar measured costs (~13-15 vs ~10 bp a year), lower and above all more stable taxation for C3M, slightly different rate sensitivity. The full data-driven comparison is in the article.
Does C3M trade at a discount or premium to NAV?
As a very liquid money-market fund, its exchange price typically sits within a few hundredths of a point of NAV. Remember though: buying and selling you still pay the bid-ask spread of the moment.