Withholding Tax Leakage
The three-layer chain
Follow one dividend from a US company to a European investor:
Layer 1: source-country withholding. The US withholds on dividends paid to foreign holders: 30% statutory, reduced by treaty. Every country does a version of this: Switzerland 35% statutory, Germany ~26%, France 25–30%, Japan 15%, with treaty reductions varying by counterparty. This is the big leak, and it happens before the fund's NAV ever sees the money.
Layer 2: the fund's domicile. The fund itself is the treaty claimant. An Irish UCITS fund claims the US treaty rate of 15%; a Luxembourg fund typically doesn't and bears 30% (Chapter 50). Some domiciles achieve 0% on certain markets. The fund may also file reclaims where treaties allow: an operational competence that varies by provider and shows up as basis points.
Layer 3: you. Whatever survives arrives in the fund and is either distributed or accumulated, then taxed by your own country under Chapter 52/53's rules; and here is the cruelty of the chain: you generally cannot credit the tax paid at Layer 1 inside a fund. Foreign tax credits are available to direct holders and, in some regimes, to fund investors on distributed income, but the leakage inside a fund is usually gone for good. This is why the structural routes below matter more than any personal tax filing.
What the leak actually costs
Typical magnitudes at mid-2026 index yields (US ≈1.1%, Europe ≈2.9%, World ≈1.5% — S&P DJI and MSCI factsheets, May–Jul 2026; EM approximate):
| Exposure | Gross yield | Effective leak | Annual cost |
|---|---|---|---|
| US equities (Irish fund, 15%) | ~1.1% | 15% | ~0.17% |
| US equities (Lux fund, 30%) | ~1.1% | 30% | ~0.33% |
| European equities | ~2.9% | ~10–15% net of reclaims | ~0.30–0.45% |
| Emerging markets | ~2.5% | ~10–20% | ~0.25–0.50% |
| World blended (Irish) | ~1.5% | blended | ~0.20–0.30% |
Set that beside a 0.15% TER and the point lands: for global equity funds, withholding leakage is usually the single largest cost of ownership, and it is nowhere on the marketing material.
Why good funds "beat" their index
Index providers publish net-return variants that assume a standardized withholding rate: typically the rate a Luxembourg-style investor would suffer, i.e., harsher than what a well-run Irish fund actually pays. A fund that leaks less than the assumption therefore shows positive tracking difference against its own benchmark, before securities lending adds more (Chapter 36).
Two consequences for your fund selection:
- Positive TD is normal for Irish physical funds on US indices and does not indicate anything mysterious.
- Comparisons must use the same benchmark variant. A fund benchmarked to net-return and one to gross-return are not comparable; Chapter 35's first rule exists for exactly this.
The four structural routes
1. Domicile. Ireland for US-heavy exposure (Chapter 50): the 15%-vs-30% lever, worth ~0.2%/yr on US holdings.
2. Synthetic replication. A swap counterparty delivering the gross-return index effectively bypasses the dividend leak on US indices (Chapter 11's mechanism, Chapter 38's checklist), which is why synthetic S&P 500 UCITS funds persistently out-track physical peers by roughly the withholding amount. Price: collateralized, capped counterparty exposure.
3. Fund-level reclaim competence. Where treaties permit refunds (several European markets), the provider must actually file: a boring operational function that separates good trackers from mediocre ones, and one you audit through TD rather than through disclosure.
4. Your own return, where applicable. US investors can often claim a foreign tax credit for foreign withholding passed through by US-domiciled funds (Chapter 52). This is the one layer an individual can sometimes reclaim; and it argues, for Americans, in favor of holding international exposure in taxable accounts, where the credit is usable, rather than in IRAs, where it is wasted. A genuine, actionable asymmetry.
A worked example
Two Dutch investors, €100,000 each in S&P 500 exposure, held 20 years, markets returning 7% with a 1.1% dividend yield.
- Anouk buys a Luxembourg-domiciled physical fund (TER 0.05%): withholding drag ≈0.33%, total drag ~0.38%.
- Bram buys an Irish synthetic fund (TER 0.07%): withholding drag ≈ 0.03%, total drag ~0.10%.
The 0.28%/yr difference compounds to roughly €21,000 over 20 years on the initial €100,000: from a decision neither of them could see on a fee comparison site, and both could have seen in a five-year tracking-difference table. Anouk optimized the number she was shown; Bram optimized the number that mattered.
Key takeaways
- Cross-border dividends are taxed at three layers: source country, fund domicile, and you; and Layer 1's leak happens inside the fund, invisibly and (for most European investors) irrecoverably.
- Typical cost for a global equity fund: ≈0.2–0.3% per year at current yields, frequently exceeding the TER: the largest hidden line in ETF ownership.
- Net-return benchmarks assume harsh withholding, so efficient funds legitimately post positive tracking difference: always compare against the same benchmark variant.
- Four structural routes reduce the toll: Irish domicile (≈0.17%/yr saved on US assets at mid-2026 yields), synthetic replication (near-gross on US indices), provider reclaim competence, and, for US investors only, the foreign tax credit.
- US asset-location corollary: hold international funds in taxable accounts where the foreign tax credit is usable, not in IRAs where it is lost.
FAQ
Can I reclaim the withholding my fund paid?
As a European fund investor, generally no: it happened at fund level. US investors can often claim a credit for foreign taxes passed through by US-domiciled funds. This asymmetry is why Europeans solve it by structure and Americans partly by filing.
Do accumulating funds avoid withholding?
No. Accumulation affects when your country taxes you (Layer 3), not the source-country leak (Layer 1).
Does this apply to bond ETFs?
Much less: interest is exempt from withholding under many treaties (notably US Treasury interest to foreign funds), so bond funds' domicile story is far weaker.
Why doesn't the factsheet disclose the leak?
Because it is embedded in performance rather than charged as a fee. It surfaces as the gap between fund and gross-index returns: which is exactly why Chapter 35 crowned tracking difference.
Is the synthetic route risky enough to refuse over 0.2%?
A legitimate personal call. Run Chapter 38's five-item checklist; investors who read it and still prefer owning physical shares are making an informed trade, which is all this book asks.
Sources & further reading
- US withholding rules for foreign persons (statutory 30%; treaty rates) and the US–Ireland treaty dividend article
- MSCI/FTSE net-return index methodology (the assumed withholding rates)
- Provider disclosures and academic studies of ETF withholding drag by domicile (dated)
- IRS foreign tax credit guidance (Form 1116 / pass-through reporting) for the US route
- This site's Chapters 11, 35, 38, 50, 52–53, 92–93