A pricing war in the mainstream exchange traded funds (ETFs), like the FTSE 100, the S&P 500 and the major bond indices trackers, has one major winner: the end investor.
FTSE 100 products in particular have seen significant falls in the total fees charged to investors, with the latest entrant, the Lyxor Core UK ETF, charging a tiny 0.04% a year, or £400 for every million pounds of assets. This makes it 0.03% cheaper than the closest competition. Lyxor has managed to do this in part by benchmarking to the Morningstar UK index, which more than likely charges slightly less in index licensing fees than the FTSE. For investors this is good to see; there is abundant fee competition both from both the index providers and the ETF managers.
Low fees are here to stay, but what many investors do not realise is that buying the ETF with the lowest fee will not mean that you will automatically outperform an ETF with a slightly higher total expense ratio.
Common ways to compare ETFs
The most commonly analysed metric when buying an ETF is the NAV tracking error. The lower the tracking error, the closer the ETF tracks the underlying market. For example, Bloomberg reports that HSBC FTSE 100 (HUKX) has a tracking error of 0.04%, which is very small indeed.
Securities lending revenue plays a part in your total return, and cannot be offset against the management fee in the total expense ratio (TER) calculation. Some ETF firms do not lend out stock, and others keep different amounts for themselves to cover costs, but it is a source of return nonetheless. The iShares Core MSCI Emerging Markets ETF (EMIM) has a TER of 0.25%, but securities lending revenue of 0.14% covers a good proportion of it.
Tracking difference reflects how closely the ETF’s net asset value has performed, relative to the benchmark. This can be positive or negative, depending on fees, securities lending and manager skill in re-balancing the portfolio to account for benchmark changes. Tracking difference is very important when looking at returns over time, particularly for index funds.
ETFs have two other variables which determine the price you pay. Bid-ask spread and share price premium, and it is the impact of the latter which can be many multiples higher than the headline management fee in the lowest cost ETFs.
Share price premium and discount volatility
This is an issue that is rarely addressed by ETF providers, in essence because it is out of their control. But in a world of very small pricing differences it has by far the largest potential impact on investor returns relative to the benchmark.
Because ETFs trade on exchange, they are subject to price movements from supply and demand. If there are more buyers than sellers, the price could rise above net asset value and vice versa. An ETF that trades at the same time as the market it tracks - such as the FTSE 100 - should have small deviations in price, but there are deviations nonetheless.
In the chart below we can see the daily closing discounts and premiums for iShares Core FTSE 100 (ISF) and Vanguard FTSE 100 (VUKE) over approximately a four year period. Both ETFs distribute their income, but VUKE charges 0.09% a year, whereas ISF charges 0.07%. The ETFs should always trade at a small premium to reflect the creation costs in buying the underlying stock in the market, and in a perfect world there would be no volatility in the premium.
The reality is different. Their relative premiums (calculated by subtracting one from the other) can fluctuate by as much as plus or minus 0.5%. For the investor who buys an ETF on a premium and then sells it on a discount, a low fee is really rather academic. The damage has been done to their relative performance.