Your IndustryMar 19 2014

Replication strategies for exchange-traded funds

twitter-iconfacebook-iconlinkedin-iconmail-iconprint-icon
Search supported by

There are two main types of exchange-traded fund: physical and synthetic.

Ben Thompson, director of business development, listed products and ETF UK of Lyxor, says the difference between the two types of ETF is not as great as many might say.

Both start with a portfolio of physical assets, which is owned by the fund and held in a segregated account. Where the two types differ is in the role of these physical assets: one uses the physical assets for performance, the other for security.

However, Mr Thompson says even this is an overly simple explanation and there are many important techniques and strategies used by both types of ETF.

Full replication

Full replication physically-backed funds purchase the stocks of the benchmark index in order to replicate the performance of the index, minus fees. As such, Mr Thompson says the holdings of the ETF must be the same, or very similar in proportion, to that of the benchmark index they track.

Purely physical funds work well for large, liquid markets where the component stocks or bonds can be easily purchased and traded. The ETF provider makes no attempt to optimise the fund’s performance or improve its tracking efficiency.

However, Mr Thompson says some indices are not so simple and can not be replicated by buying all the assets of the benchmark index.

Sampling

In response, Mr Thompson says ETF providers have developed a number of alternatives to pure physical replication.

This might be the use of sampling techniques; whereby stocks which can’t be bought physically are substituted with larger, more liquid stocks. In general, a fund which utilises sampling would replicate the index in its sector weightings, but without holding every security.

In the case of very large indices such as the MSCI World index, which has more than 1,600 constituent stocks, Mr Thompson says buying all the stocks is inefficient and as such stocks with a small weighting may simply be ignored in the ETF basket.

In either case, Mr Thompson says as the ETF does not hold the same stocks as the index, sampling may lead to discrepancies in performance between the ETF and the index it is tracking.

Securities lending

Mr Thompson says a technique used by fund managers to improve the performance of a physical fund is to lend out the fund’s assets to other financial institutions. In return, the financial institution provides collateral against the stock and pays a fee for the service.

Mr Thompson says this practice is used throughout the asset management industry and it is called securities lending.

By reinvesting the revenues from securities lending, Mr Thompson says physical fund managers can improve the degree to which the ETF tracks its benchmark index.

“The risk is that the institution borrowing the securities could default, in which case the securities could be lost, and the fund value would suffer. However, with adequate collateral and tight management this risk can be significantly reduced.”

Performance swaps

Mr Thompson says the other way that a fund manager can look to optimise the performance of their ETF is through the use of a performance swap contract.

These ‘synthetic’ ETFs also invest in a basket of physical assets such as equities or bonds, which are owned directly by the fund and appear on the fund’s balance sheet.

However, Mr Thompson says this time the role of the physical assets is not to provide the performance of the fund, but purely to act as security. The performance of the fund is instead based on dervatives known as ‘performance swaps’.

This swap is a contractual agreement which is negotiated between two parties: the ETF and the swap counterparty. The swap counterparty commits to pay the precise daily performance of the benchmark index, including any dividends to the ETF.

In return, Mr Thompson says the ETF pays the swap counterparty a fee for the swap arrangement, and the performance of the physical assets it holds, including any dividends.

Mr Thompson says: “Using performance swaps or securities lending initiatives involves additional risk for investors.

“This risk may still be acceptable if it delivers sufficient additional returns; either through a performance improvement or a reduction in tracking error.

“If this is not the case, the best replication method would be a physical structure without any swap or securities lending.”

Assessing risk

While there is a general tendency to view physical replication as structurally ‘less risky’ than synthetic, Ben Seager-Scott, senior research analyst at Bestinvest, says this is not always the case and the arguments for and against different replication methods can be fairly nuanced.

For synthetic exposure, Mr Seager-Scott says it is important to note that in all of these cases exposure to counter-parties are limited where Ucits rules apply, and most groups will ensure counter-party risk is mitigated through collateralisation policies.

Since physical replication involves buying and selling index components, Hortense Bioy, director of passive funds research at Morningstar, says this strategy is inherently more labour intensive and potentially more costly than synthetic replication, depending on the market being tracked.

Ms Bioy says: “ETFs using physical replication may also exhibit larger tracking error than ETFs using synthetic replication.”

In order to mitigate the risks of the existing swap-based ETF structures, a spokesman for iShares says various ETF providers started launching new swap-based ETFs which would tackle some or all of the flaws of the swap-based funds.

The spokesman says: “Multi-counterparty swap models with an over collateralised counterparty exposure model mean that the aggregate market value of collateral taken will exceed the overall counterparty exposure.

“The quality of the collateral and over collateralisation helps investors to protect from counterparty default risk.”

The first swap-based ETFs were launched in 2001, according to iShares, and involved a single counterparty underwriting the swap. The swap exposure was limited to 10 per cent of net asset value in line with the Ucits rules.

According to iShares this model lacked transparency in the fund’s holdings (the reference or substitute basket), which are often purchased from the fund promoter’s investment bank affiliate, and it carried undiversified counterparty risk exposure.

To address counterparty risk, a spokesman for iShares says the next generation of swap-based ETFs were introduced at the end of 2008. The structure used multiple swap counterparties, which allowed diversification of counterparty risk.

A spokesman for iShares says: “Having multiple swap counterparties can decrease the costs of the swap as counterparties are not affiliated with the ETF provider and compete with each other on monthly swap spreads.”