Synthetic ETFs and Swap-Based Replication

Most ETFs track an index the obvious way: they buy the actual stocks or bonds in the index, in the right proportions. A synthetic ETF does something less obvious — it holds a basket of collateral that may have nothing to do with the index, and separately pays a bank to hand over the index's return via a contract called a total return swap. The fund never touches most of the securities it claims to track.

How the structure works

Step	What happens
1. Fund raises cash	Investors buy ETF shares, cash comes into the fund
2. Fund buys collateral basket	The cash buys a basket of securities — often unrelated to the tracked index, chosen for safety and liquidity
3. Fund enters a swap	A bank agrees to pay the fund the index's total return in exchange for the return on the collateral basket
4. Index goes up	The bank pays the fund the difference; fund NAV rises to match the index
5. Index goes down	The fund pays the bank the difference; fund NAV falls to match the index

The collateral basket sits there mostly as insurance — if the swap counterparty defaults, the fund still owns something.

Worked example

A synthetic ETF is meant to track a foreign stock index that is hard and expensive to buy directly (illiquid, foreign withholding taxes, market-access restrictions). Investors put in $100 million.

1. The fund uses the $100 million to buy a basket of liquid, investment-grade bonds as collateral, worth $100 million.
2. The fund enters a swap with a bank: the bank will pay the fund the return of the foreign index, and the fund will pay the bank the return of the bond collateral.
3. Over the year, the foreign index rises 12 percent while the bond collateral returns 3 percent.
4. Settlement: the bank owes the fund the 12 percent index gain, netted against the 3 percent bond return the fund owes back — a net payment from the bank to the fund of roughly 9 percent of notional, about $9 million.
5. The fund's NAV rises by that $9 million, tracking the 12 percent index gain almost exactly, even though the fund never directly owned a single foreign stock.

Why use this structure at all

Synthetic replication solves real problems physical replication cannot: it sidesteps foreign withholding taxes on dividends that a direct holding would owe, avoids trying to trade illiquid or restricted local markets, and can track an index more precisely because the swap is contractually defined to deliver the exact return rather than approximated by imperfect physical trading.

The trade-off

Physical ETFs face tracking error — small day-to-day gaps between the ETF and its index caused by trading costs, cash drag, or sampling instead of fully replicating. Synthetic ETFs largely eliminate that tracking error, but introduce something physical funds do not have: counterparty risk. If the swap bank fails, the fund is left holding whatever the collateral basket is worth, which may not match the index the fund promised to track. Regulation typically caps a swap's exposure at a fraction of the fund's assets (commonly 10 percent in European UCITS rules) and requires the collateral basket be diversified and marked to market frequently, but the risk is structurally different, not eliminated.