Futures v ETFs – the debate continues

by , 5th February 2019

Societe Generale has thrown its two-pennyworth into the debate regarding which is a better trading strategy, futures or ETFs.

In truth the report is pretty arcane; it certainly wouldn’t qualify as your favourite book at bedtime. But for those interested in the costs of both strategies, it is required reading.

SG begins by summarising how the competition between futures and ETFs has intensified as the latter has grown in popularity and interest rates have been maintained at historically low levels.

“Against this backdrop, the question of which instrument to select for gaining exposure to equity beta has emerged and has logically focused on a comparative analysis of ETFs and futures total costs,” the authors state.

Beyond trading costs, both ETFs and futures are subject to holding costs over the time of the transaction. These reflect the cost of manufacturing the instrument incurred either by market makers in the case of futures or investment managers in the case of ETFs.

These charges account for the slippage between the instrument and the index return, or the holding costs. However, the holding costs between futures and ETFs are very different and depend on the nature of the instrument.

“Although futures and ETFs may provide investors with the same economic exposure, the difference in instrument nature means they have different features and characteristics. These structural differences explain to what extent some products may be more suitable than others depending on investment considerations (exposure, transaction size and maturity) and investor nature (ability to manage unfunded derivatives contracts),” the report states.

Given this backdrop, unsurprisingly it means that one conclusion about the debate between futures and ETFs is that, as in beauty and the eye of the beholder, the best instrument will depend on who is investing and for which exposure.

“In the simplest cases, the type of exposure and investor may make one instrument the only choice. For instance, retail investors may only have access to ETFs but not to unfunded derivatives contracts. Similarly, less popular benchmarks may not be accessible through futures but only through ETFs,” the report says.

“The instrument selection becomes more complex in the case where an institutional investor with no trading restrictions is willing to gain exposure to a popular equity benchmark. A rational investor would not eliminate one instrument at first glance but rather apply objective criteria to identify the instrument that is best suited to him or her.”

Cost comparison

An obvious point about the holdings costs of ETFs is that when you are buying an equity beta instrument, you will not achieve the exact underlying return of the index as the pricing of ETFs relies on other parameters than the spot price of the index. Meanwhile, holders of ETFs also have to consider liquidity costs – the costs of buying and selling the instrument – and counterparty risk.

With futures, the holding costs is the difference between the futures price and the underlying asset. However, each ‘roll’ of the future incurs trading costs.

The report goes into detail on the calculations that have to be made around the pricing of and cost of financing (COF) of futures and ETFs and the net asset value (NAV). This author won’t pretend to understand the algebra involved, but jump to the end of the report, there is a comparison between futures and ETFs around the S&P 500 and the Euro Stoxx 50.

S&P 500 example

For a short-term position -3 months – the futures contract looks cheaper than ETFs, owing to the current negative COF. For longer-term exposure (12 months), the costs of the futures logically increase with the rolls and exceed the costs of several ETFs. This is the case with US-domiciled ETFs, which all capture 100% of the S&P 500 gross dividends and with one synthetic structure in Europe that enjoys a super low annual TER (5bp).

“All other ETFs are expected to be less efficient (before taxation at the investor level). In the European space, swap-based structures outperform physical ones on due to favourable swap spreads. The differentiation among swap-based ETFs is driven by the level of management fees and the substitute basket composition.”

But, to complicate matter further, the report shows that these conclusions are volatile and can be impacted by market parameters.

To illustrate this point, the report’s authors ran the calculations as of 11 January 2018, one year later. “The estimated COF was then much higher than now and the cost projections for the S&P 500 futures exceeded the ETF ones in most cases, both in the short and medium term. The hierarchy among ETFs showed little change, by contrast.”

“In the case of a $10m transaction, the relative efficiency of the futures improves thanks to lower transaction cost assumptions,” the report states. “However, the advantage of the futures was not as great one year ago for a 12m investment timeframe, reflecting a higher COF.”

Euro Stoxx 50 example

In the case of the Euro Stoxx 50, a $100,000 Euro Stoxx 50 exposure is “expected to be less costly via futures.” Specifically looking at a scenario based on the index on the 19 January and with the same index one year previously, SG shows the implied COF was negative in both cases.

“The second most efficient solution looks to be the synthetic ETF, which is paradoxical considering the recent move of ETF providers, i.e. nearly all Euro Stoxx 50 swap- based ETFs went physical in recent years. The swap-based ETF exhibits lower costs thanks to super low management fees and favourable swap enhancements.”

A year previous, the futures were expected to outperform over three months, followed by the synthetic ETF.

“That was not the case for a 12-month investment period, with the futures only third to the synthetic ETF and one physical ETF (both domiciled in Europe).

“In all cases, the US- domiciled ETF (physical replication) was expected to underperform, on the back of a relatively high TER and less favourable tax treatment on the underlying dividends.”

In the case of a $10m transaction, the relative efficiency of the futures improves thanks to lower transaction cost assumptions. “However, the advantage of the futures was not as great one year ago for a 12-month investment timeframe, reflecting a higher COF.”

To sign off, SG makes one more point worth considering: “Our total cost estimates are a necessary but insufficient condition to select the best suitable product.” By our estimates, that means the debate will rumble on depending on which salesperson you happen to be listening to. Happy trading!