Since their introduction in 1990 as a cost-effective means of index replication, exchange-traded funds (ETFs) have grown exponentially in number, variety and asset value. At the end of 2007, before the main market impact of the global financial crisis, there were 1,170 distinct ETFs with a total market value of $851bn. Nine years on, at the end of 2016, the comparable numbers were 6,625 funds valued at $3.546trn (according to sector researcher ETFGI) – an increase in assets of 317% over the period. ETFs’ rising popularity stems from several benefits they offer investors: they are cheap (the total expense ratio on State Street’s $139bn SPDR fund is 9 bps), they provide exposure to numerous asset classes, industries, geographies, factors and indeed combinations of these, and (in theory, because they are listed on exchanges) they offer a liquid means of building, hedging or shorting a position.
However, ETFs are not without their risks.
Liquidity issues have emerged in the past, in periods of market stress, and remain contentious. ETFs are structured to provide liquidity at two levels, namely the trading of the ETF on the secondary market (investors trade shares in the ETF like a normal listed share) and primary market liquidity, when ETFs are liquidated or created from their underlying components. Here, it is instructive to distinguish between “plain vanilla” and exotic ETFs.
In the former grouping, the instrument is closely matched by its underlying components, both in terms of composition and liquidity. Provided the ETF is not so large that its dealings (for example, in response to changes in index constituents) have a market impact, these ETFs have proved to be largely robust in the past. Capacity management (the market impact point) is the main issue to watch. Despite some temporary divergences from their underlying indices, these ETFs have for the most part been true to their stated objective of providing exposure to their underlying holdings.
In contrast, exotic ETFs are characterised by liquidity mismatches, leverage or both – and it is on these products that concerns tend to focus. In the event of a sell-off in a high yield ETF, for example, where liquidity in the underlying bonds has all but disappeared, gaps may emerge between the price of the ETF and that of the index it is trying to replicate. In theory, the action of authorised participants (APs) in the marketplace should prevent this. APs are incentivised, but are not obligated, to make a market in ETF shares and exploit arbitrage opportunities when the price of an ETF diverges from its underlying. However, given that this involves a parallel trade in the underlying securities, APs may withdraw from the market under conditions where the liquidity of the underlying holdings dries up, or there is significant market volatility in the price of the ETF’s components. Under these circumstances, the price of the ETF may diverge significantly from the stated index price due to supply of and demand for the ETF in the secondary market.
Synthetic ETFs, where the ETFs are not backed by physical securities but by derivatives with investment banks as counterparties, also present some issues. In many cases the collateral posted by the counterparties to the derivative arrangements bears no relation to the assets of the underlying index being tracked. At times of stress this mismatch exposes the ETF to credit risk from its counterparties. Now, aversion to holding the collateral basket or dealing with the counterparty bank may cause APs to stop providing primary market liquidity – again giving rise to potential price discrepancies between the ETF and its components.
There are also market structural reasons why the performance of the ETF may not replicate its target index. For example, in the case of the VIX, the ETF will replicate its exposures using forward contracts on the index. Owing to the usual state of contango (upward slope) on the VIX futures curve, long-term holders of the ETF will gradually have their capital eroded (relative to the performance of the index) by paying away the roll yield of the futures.
Leveraged ETFs present other difficulties. Due to the requirement to rebalance leverage daily, investors using a leveraged ETF to match their exposure to an index may find that after three days of market volatility they have not had the gains or losses they expected based on the performance of the index.
Then there are issues relating to ETF operational structures. Given the predictability of ETFs trading in the market when indices are rebalanced or future contracts are rolled over, there is some concern that they are easy targets for speculators, particularly in times of financial stress.
Ultimately, the outcomes from ETFs come down to how they are deployed. Here we invoke our strategies for coping in a complex investment environment. Investors need to be clear on their investment objective (self-understanding), have a clear understanding of the strategy they are deploying to achieve this (adaptability), and be mindful of the other market participants trading in ETFs and how they might be looking to exploit structural features of the products (meta-understanding).
Jeremy Spira