By Matthew J. Bartolini, Head of SPDR Americas Research, State Street Global Advisors.
There are many considerations to weigh when choosing from the more than 2,000 US-listed ETFs, not least of which is cost. However, many analyze cost based on expense ratio alone, when the reality is that this is just one piece of the total cost of ownership (TCO) puzzle – trading costs can also be a significant piece. When conducting ETF due diligence, it’s important for investors to understand what TCO really means.
Low expense ratios don’t always mean the lowest TCO
Low fees don’t naturally lead to low trading costs and, therefore, the lowest TCO. The graph below plots the 30-day average bid-ask spread versus the expense ratio for the 100 largest US equity ETFs (based on AUM), clearly illustrating the lack of correlation between ETFs’ expense ratios and trading costs. As a result of differences in secondary market liquidity profiles, some higher-fee funds have lower bid-ask spreads, and some lower-fee funds have higher bid-ask spreads.
Bid-ask spreads are an important ETF trading metric to consider, as they can be an indicator of ETF liquidity – the lower the spread, the greater the liquidity. This is demonstrated by the strong correlation between bid-ask spreads and trading volume, where ETFs with higher daily trading volumes tend to have lower spreads.
Why ETF trading costs matter: A hypothetical example
Since there’s no correlation between expense ratio and liquidity, evaluating ETF costs on expense ratio alone can result in materially inaccurate estimates of what it might cost to own any given ETF. Therefore, much like calculating the money paid in fund fees on an investment, the calculation of trading costs must also be included in any model used to make pre-investment decisions, as well as during the ongoing review of the TCO for a particular ETF.
To demonstrate this concept, the following hypothetical example compares the overall cost of two ETFs that seek to track the same index but have different expense ratios and rebalancing costs (after a hypothetical 20% annual gain on the original investment).
As shown, trading costs make Fund XYZ – the fund with the higher expense ratio – the lower-cost option in this annual rebalance scenario. And the more frequent the rebalancing, and the greater the size of the rebalance, the greater the impact on the total cost of ownership.
To give you a sense of the calculations behind this example, when a position is established, the trading costs are applied to the total notional dollar figure. For Fund ABC, this would be the 0.05% applied to the $10m position. This is similar to the application of an expense ratio. But any rebalancing trades are applied only to the notional value traded – in this case, the $2m, which is just a portion of the total portfolio value.
ETF trading can be nuanced
It’s important to note that this example omits some trading nuances that should be considered. For example, we list commissions as the same, but in reality, that may not be the case, as some funds are offered commission free. Also, depending on the trading platform, some trading commissions are calculated on a per-share basis, not on flat basis points.
We also assume that the trading environment stays the same, with bid-ask spreads near “average” levels when it comes time to trade. Of course, volatility can strike at any time, and an ETF’s liquidity profile can change along with it – a dynamic we saw in Q4 of 2018.
Sometimes you pay higher fees for lower costs
While this example doesn’t incorporate all the nuances inherent in ETF trading, it nonetheless illustrates a key point: buying an ETF based on its headline expense ratio alone may not lead to the most cost-efficient solution. Depending on your rebalancing size and frequency, trading costs can accumulate significantly and have a larger impact on the TCO than any expense ratio difference between two ETFs. This underscores why liquidity analysis has to be a part of any due diligence process prior to implementation.
(The views expressed here are those of the author and do not necessarily reflect those of ETF Strategy.)