All other things being equal, the key test of the effectiveness of an ETF is its “tracking difference” which is the term for any divergence between the performance of the ETF and the performance of its underlying index. ETF providers publish the tracking difference of their funds regularly so they can be closely monitored.
There are several potential sources of tracking difference which include:
- Transaction costs – each trade within an ETF involves a set of costs, including the spread between the bid and ask prices.
- Annual fees – ETFs charge an annual fee that includes the cost of portfolio management and custody of the securities in the funds.
- Rebalancing costs – index providers, such as FTSE or S&P, regularly rebalance their indices to reflect securities entering or departing from their indices. In rebalancing, index providers do not take into account the costs and timing considerations of buying and selling securities.
- Stock lending revenue – some ETFs generate additional revenue by lending out some of their investments. Stock lending is a well-established and big business. It began as a way of facilitating the prompt settlement of trades but many stock borrowers are now hedge funds who borrow stock in order to sell it short. Stock lending is capable of generating worthwhile revenues though it should be noted that it introduces an element of counterparty risk.
- Optimisation leakage – where it is not possible to exactly replicate the index portfolio and ETF managers employ optimisation there can be divergences between the index return and the returns from the securities actually held. This problem does not arise when synthetic replication is used since this tracking methodology is capable of eliminating all forms of tracking difference other than annual fees.
It can be seen that the key factors in tracking difference revolve around costs since an index is a theoretical construction that does not bear any costs.
These costs, however modest, mean that ETFs are bound to underperform their respective indices to an extent. However, this will not matter if the original asset allocation judgement is successful because studies show that this is what really makes the money. It is better to slightly underperform a strong index performance due to moderate costs than risk seeing that successful asset allocation decision undermined by poor stock selection and much higher costs which is the risk with active stock picking.
