Synthetic vs Physical Replication: Does the Fund Own the Things?

Both approaches can track an index accurately; they carry different risks, and a reader who does not know which one they own does not know what they own. This is a distinctly European subject: synthetic replication is far more common in the UCITS market than in the US, so a European reader is materially more likely to encounter it, often without realising. The index-fund article listed the three replication methods; this article takes the synthetic one apart properly, names the counterparty exposure plainly, and then describes fairly the arrangements that mitigate it — because dismissing synthetic replication as simply dangerous is as unhelpful as ignoring the risk.

Physical replication: what most people assume they have

A physically replicating fund buys the securities. Full replication holds every index constituent at its weight; sampling or optimisation holds a representative subset where holding everything is impractical, as the index-fund article described. The reader owns, through the fund, a claim on real securities held by a depositary. Four properties follow. What you see is what you own — the published holdings are the actual portfolio, which makes the fund straightforwardly legible. Tracking depends on implementation: rebalancing costs, cash drag, and unrecovered withholding tax produce the tracking difference that article dissected. The fund may lend its securities to generate revenue, which introduces a separate consideration covered in the next article — so "physical" does not automatically mean "no counterparty exposure." And some exposures are hard or impossible to hold physically: markets with ownership restrictions, quota systems, or settlement arrangements that make direct holding costly, and commodity indices where holding the underlying is not feasible for a fund. That last point is the reason synthetic replication exists.

Synthetic replication: the swap, the collateral, and the exposure

A synthetically replicating fund obtains the index return through a total return swap — a contract under which a counterparty, typically an investment bank, agrees to pay the fund the index's total return in exchange for a fee, and for the return on whatever the fund actually holds. Two structures dominate. In an unfunded swap, the fund uses investor money to buy a substitute basket of securities — often large liquid shares or government bonds, and often bearing no relationship to the index being tracked — and swaps that basket's return for the index return. In a funded swap, the fund transfers cash to the counterparty and receives the index return, with collateral posted to a segregated account in the fund's favour. Either way, three things are true and need stating without softening. The fund does not own the index constituents. A synthetic emerging-market ETF may hold German government bonds and European blue chips; its published holdings will look nothing like its name, and that is the structure working as designed rather than an error. You have counterparty exposure. If the swap counterparty fails, the fund's claim to the index return is a claim against a failed institution, and what the fund recovers depends on the collateral arrangements. This is a genuine, additional risk that a physically replicating fund does not carry in the same form. And UCITS limits it but does not remove it. The framework caps counterparty exposure per counterparty, requires collateral meeting defined criteria, and mandates risk management — which is why the exposure is bounded rather than open-ended. Now the mitigations, described fairly because they are substantial. Collateralisation is the main one: funded structures post collateral, frequently in excess of the exposure, held for the fund's benefit and valued daily. Frequent resets — swaps are commonly reset at defined intervals or when exposure crosses a threshold, crystallising value and reducing accumulated exposure. Multiple counterparties spread the risk in some funds. And collateral eligibility rules restrict what may be posted, addressing the concern that collateral could itself be poor quality. Against those, three honest counterpoints. Collateral quality and correlation matter: collateral that falls in value precisely when the counterparty fails is worth less than its headline figure suggests. The structure is opaque relative to physical — assessing it requires reading the swap arrangements, the counterparty identity, and the collateral schedule, which is more work than reading a holdings list. And the crisis history is relevant: counterparty exposure to major banks was widely regarded as remote before 2008, and the crisis pillar documents what that assumption was worth. None of that makes synthetic replication unacceptable; it makes it a different bargain requiring different diligence.

When each is used, and what to check

Synthetic replication is chosen for reasons that are usually structural rather than opportunistic. Hard-to-access markets where direct holding faces restrictions or prohibitive cost. Commodity indices, where a UCITS fund cannot hold the physical asset. Tax efficiency on dividends in some structures, where a swap can deliver a gross-dividend index return that a physical holder could not achieve after withholding tax — historically a significant driver, and one that has narrowed as tax rules changed. And tracking precision: synthetic funds often show very tight tracking, because the counterparty bears the implementation friction rather than the fund. That last point connects to something the tracking article warned about: a fund with unusually tight tracking may be delivering it through a structure with its own risk, so tight tracking is not automatically evidence of quality. What to check, in four items. Which method — stated in the fund documents, and often in the fund's own name or its KID. If synthetic, who is the counterparty, how many are there, and what is the current exposure. What is the collateral — its composition, its quality, whether it is over-collateralised, and where it is held. And what does the fund actually hold, which for a synthetic fund is the substitute basket rather than the index, and which should be read precisely because it will look surprising. A reader who checks those four knows what they own. A reader who assumes an ETF holds what its name describes may be wrong, and in the synthetic case will be wrong by a wide margin. This portal expresses no preference between the methods: the choice depends on the exposure sought, the alternatives available, and a reader's own view of counterparty risk — which is a judgment, and theirs to make with an adviser if they want one.