VWRA vs VT: why the choice isn’t just about TER

1. The fee gap, and what it is worth

Start with the two funds as published:

Fund data as published by Vanguard and fund data providers, August 2026. Expense ratios and constituent counts change; the factsheet is the authority, not this page.

Eight basis points. On $100,000 that is $80 a year, charged whether the fund distributes anything or not, every year, on the whole position. It is a real cost and it is genuinely against the Irish wrapper.

It is also the only number in this comparison that behaves itself. It does not depend on where you live, what is on your W-8BEN, what your country does with accumulating funds, or whether your broker will sell you the thing. Everything else on this page does — which is exactly why the fee gap ends up leading every article about these two tickers and deciding almost none of the real cases.

2. Where the dividend is taxed, and how many times

Both funds hold approximately the same companies, so both suffer withholding inside the fund on their non-US holdings, at rates set by the treaties between each source country and the fund’s domicile. Those two amounts are not identical, but they are close enough that the comparison below leaves them out of both columns. What differs structurally is the US-sourced part of the dividend, and what happens to the money on its way to you.

VT is a US fund. It receives US dividends with nothing withheld — a US fund paying a US company’s dividend into a US account is not a cross-border event. But the distribution it then makes to you is US-source income, and your residence’s rate from the IRS treaty table is applied to the entire distribution, including the part that originated in Tokyo and Frankfurt. That is 15% for many treaty countries, 10% for several, 0% for a few, and 30% where there is no treaty or no valid documentation on file.

VWRA is an Irish fund. It suffers 15% US withholding on the US-sourced dividends it receives, under the US–Ireland treaty. Ireland then withholds nothing from a non-resident holder, and because the share class accumulates there is no distribution to withhold from in the first place.

So the comparison is not “15% versus 30%”. It is:

15% on the US slice only   vs   your treaty rate on everything

That asymmetry is the whole story, and it is why a global fund behaves differently from an S&P 500 fund here. In a pure US index every dividend is US-source, the two rates meet on the same base, and the wrapper argument dies exactly at a 15% treaty rate — an independent study using our model worked that case through and found the sign flips below 15%. In a global fund the Irish side’s 15% only ever touches part of the income, so the two sides cross somewhere lower. Finding where is the useful thing this page does.

3. The crossover treaty rate: about 10.7%, not 15%

Assumptions, all stated, all replaceable with your own: $100,000 invested; gross dividend yield of the index 1.7%, so $1,700 a year; US holdings around 60% of the index by weight but lower-yielding than the rest, so roughly 40% of the gross dividend — about $680 — is US-sourced.

Read the last column downwards. At the statutory 30% the Irish wrapper is worth four times the fee gap. At 15% it is worth $73 — still positive, and now the same order of magnitude as the fee it costs you. At 10% it has already gone negative. At 0% the US fund wins by more than twice the fee gap, and no expense ratio anyone is likely to publish would change that.

The condition behind the table, per dollar invested. Let y be the index’s gross distribution yield, s the US-sourced share of that dividend, L1 the 15% fund-level rate an Irish fund suffers on the US slice, r your US treaty rate and ΔTER the Irish fee minus the US fee:

annual difference = y × (r − s·L1) − ΔTER

Set it to zero and solve for the treaty rate at which the two wrappers are worth the same:

r* = s × L1 + ΔTER / y

With s = 0.40, L1 = 15%, ΔTER = 0.08% and y = 1.70%:

r* = 6.00% + 4.71% = 10.71%

Two things are worth pulling out of that formula, because both are invisible when the argument is made in prose.

The first term is structural. Six percent of the crossover is simply the Irish fund’s own 15% applied to 40% of the income — the unavoidable cost of the Irish wrapper on a global index, independent of fees, independent of yield, independent of everything except how much of the dividend comes from the United States. In a pure US fund this term is the full 15% and the crossover sits there. In a global fund it is a fraction of it.

The second term is the fee gap divided by the yield, and that division is the thing most people get wrong by instinct. Eight basis points against a 1.7% yield is not eight basis points of the decision — it is 4.7 percentage points of treaty rate, nearly as much as the structural term. A fee gap is charged on capital; a withholding difference is charged on income; and on a low-yielding global index the capital base is roughly sixty times the income base. That ratio is what makes a number as small as 0.08% matter at all.

Position value cancels out of the condition, so the sign of every row is independent of the $100,000. The dollar amounts are not. The table also excludes tax where you live and any foreign tax credit for US withholding — see the limitations below, because for many readers that omission is larger than everything measured here.

4. How far that number can move

10.71% is a result under four inputs, two of which drift every year. It is worth knowing how sturdy it is before anyone quotes it back at you:

Across every plausible combination the crossover stays in a band of roughly 8% to 12%. That is the useful, robust statement: for a residence at 15% or above the Irish wrapper wins on these two layers under any reasonable input; for a residence at 5% or 0% the US wrapper wins under any reasonable input; and the 10% countries sit in the band where the inputs decide, not the structure.

Notice also which input moves it furthest. Halving the fee gap moves the crossover by 2.4pp. Moving the US dividend share by ten points moves it by 3pp. The expense ratio is not the dominant sensitivity even in the one place it enters the formula.

5. These are not the same index

Everything above treats the two funds as the same portfolio in two wrappers. They are not, and this is the assumption most likely to be doing quiet damage to a comparison.

VT tracks FTSE Global All Cap: large, mid and small cap, around 10,000 holdings. VWRA tracks FTSE All-World: large and mid only, around 3,760. The small-cap tier VT reaches and VWRA does not is a mid-single-digit to high-single-digit share of global investable market capitalisation, and it is a segment with its own risk and return profile rather than a rounding residual.

The honest way to say this: the difference between the two indices is an exposure decision, and the difference between the two wrappers is a cost decision, and they have no business being priced in the same sentence. A year in which global small caps beat large caps by two percentage points moves the VT side by more than a decade of the tax and fee terms combined — in either direction, unpredictably, which is precisely why it cannot be netted against them.

The study we cited earlier refuses to run a fee-and-withholding comparison on any pair that does not track the same published index, for exactly this reason, and it names this pair as one it excludes. That is the right discipline. What this page is really measuring, then, is the annual cost of the wrapper — a number worth knowing precisely because it is the part you can compute, sitting next to an index difference you cannot.

A related non-issue: VWRA and VWCE are the same fund and the same ISIN, listed in USD in London and in EUR on Xetra and Borsa Italiana. Choosing the EUR line does not hedge anything; the fund holds the same global equities either way. The listing currency affects the FX you pay at the trade, not the currency exposure you carry.

6. Estate situs: the term with no annual number

VT shares are US-situs assets. For a non-US, non-resident estate, US-situs assets above USD 60,000 can attract US estate tax at rates reaching 40% on the excess, unless an estate tax treaty raises that threshold — and the United States has estate tax treaties with only a limited number of countries. An Irish-domiciled UCITS is not a US-situs asset for this purpose.

This term has no annual dollar figure, which is why it slides out of comparisons built on expense ratios, and it is the largest number on this page by an order of magnitude when it applies. A long-term index investor crosses $60,000 in an ordinary month, with no notification from the broker and no state change anywhere on any dashboard. Whether it ever bites depends on the size of the estate, on how the assets are held and on the treaty position — none of which a portfolio tracker can compute for you. Knowing which side of the line your holdings are on is the part that is knowable, and most people holding VT from outside the US have never checked.

For a great many non-US investors this single asymmetry is the reason they hold the Irish wrapper, and the annual arithmetic in section 3 is a secondary consideration they would accept a worse answer on.

7. Accumulating is not automatically simpler

VWRA reinvests internally: no cash lands, nothing to re-buy, no drag from dividends sitting idle, and — usefully if you track contributions separately from returns — nothing arriving in the account that could be mistaken for a deposit. VT distributes quarterly, visibly, and you reinvest by hand.

That reads like a straight win for accumulation, and in several countries it is the opposite. Some residences tax accumulating funds on a deemed or imputed basis regardless of whether anything was paid out — the German Vorabpauschale and the Dutch box 3 system are the usual examples — and a few treat them less favourably than distributing funds outright. Others tax only what is actually received, which makes accumulation a genuine deferral. The same structural feature is an advantage, a neutrality or a penalty depending on nothing but the address on your tax return.

Which is better depends on the residence you actually have, not on which one is tidier to administer. And that second step — tax where you live, the credits available for what was already withheld, and the caps on those credits — is excluded from every number in section 3. For a resident of a country that taxes dividend income and grants full credit for US withholding, much of VT’s disadvantage is absorbed: the tax you would have paid at home is partly paid in Washington instead. For a resident of a country that taxes nothing there is no credit to claim, which is exactly why the no-treaty, no-income-tax case is where the gap is widest and most permanent.

8. Availability settles it more often than arithmetic

Retail investors in the EEA and the UK generally cannot buy VT at all, because US-domiciled ETFs do not publish the Key Information Document that PRIIPs requires — the broker blocks the order rather than the regulator blocking the fund. Investors in most other jurisdictions — the Gulf, South Africa, much of Asia and Latin America — can typically buy either through an international broker, subject to that broker’s own rules.

If only one of the two is purchasable where you are, every preceding section becomes background rather than decision. The useful question then is not which wrapper to pick but what the wrapper you can actually buy costs you per year — which is the number section 3 produces, and which is worth having whether or not you had a choice.

9. The decision in the order the terms actually matter

Ranked by how much each one moves the answer, for a non-US investor on a $100,000 position:

The expense ratio is last on that list and first in every comparison, and the reason is not stupidity — it is that the TER is the only term that is the same number for every reader. It generalises, so it gets written down. Everything above it requires knowing something about the person asking, so it gets replaced by a rule of thumb that is right for some readers and backwards for others.

The rule of thumb here is “UCITS for non-US investors”, and on these assumptions it is correct down to a treaty rate of about 10.7% and wrong below it. That is not a criticism of the rule. It is the reason the crossover is worth computing with your own yield, your own fee gap and your own rate rather than inherited from a forum post.

Limitations

• The example is an illustration, not your portfolio. Yields, US index weights and the split between US- and foreign-sourced dividends move every year, and all three enter the crossover.
• Treaty rates depend on paperwork. Without valid documentation on file the statutory 30% applies whatever the treaty table says.
• The second step is not modelled. Tax where you live, foreign tax credits and their caps, and the treatment of accumulating funds are residence-specific and can be larger than everything measured here.
• Foreign withholding inside the fund is netted out of both columns. The two funds’ treaty networks differ on some source markets; the difference is small relative to the terms above but it is not exactly zero.
• Switching an existing position is a different question. Realising a capital gain today to save a few basis points a year rarely pays — on an existing holding the embedded gain and your capital-gains rate dominate the entire comparison. The switch study has the payback grids.