To compare and select exchange-traded funds (ETFs) managed against the same index, investors can use a range of criteria including the cost of the fund or the extent to which its investment portfolio deviates from the list of index components.
Among the most commonly used are
- the total expense ratio (TER or OCR)
- the tracking error (i.e., the volatility in differences of performance between the ETF and the replicated index)
- measures of liquidity such as the market order book (particularly the bid-ask spread) at the stock exchange on which the ETF is listed.
Other criteria may be taken into consideration such as
- where the fund is listed or registered
- its assets under management
- the service providers used by the portfolio management firm.
These, however, do not reflect the fundamental characteristics of the ETF.
Analysing ETFs on the basis of tracking difference
Recently, more and more investors have been analysing ETFs on the basis of tracking difference, or the ‘performance gap’, a measure which is highly complementary to tracking error.
Tracking difference can be explained by several factors:
- Tracking error concerns the volatility of the incremental return and as such, it is a measure of the stability of the tracking difference over time.
- The quality of index portfolio management is intuitively measured by the capacity of an ETF to track as closely as possible and on an ongoing basis the performance of the underlying index such that there are no big diversions in performance, meaning tracking error is as low as possible.
- As index portfolio management has been a mature business for several years now, we believe that the crucial criterion of tracking error is optimised by all the main ETF issuers. As a result, it has essentially become the same from one issuer to another.
- Tracking difference (TD) is calculated by measuring the difference between the returns of an index and those of the ETF replicating it. It measures to what extent the ETF has outperformed or underperformed the index. It may be calculated over a pre-determined period of time. Generally, it is negative. Depending on which index is being replicated, it can be positive too.
Negative tracking differences
Several factors can generate negative tracking differences (TD). The first among these are management fees or other fees to which the ETF is subject. If they are significant, structural underperformance may result. Typically, fees vary depending on the asset class and region being replicated.
Fees for buying and selling securities to maintain target asset allocations are a cost to the fund. They are either one-off in nature (upon each rebalancing of the index in the case of physical replication) or, in the case of a swap and synthetic replication, spread out over several months as they are included in the price of the swap.
Execution costs may arise when investing subscriptions or raising cash to meet redemptions. This would involve the buying or selling of assets where an ETF is managed using physical replication or an adjustment of the size of a swap to maintain 100% exposure in the case of synthetic replication. However, this cost is systematically offset by subscription or redemption fees paid by the market maker authorised to create or redeem shares in the ETF.
Improving the tracking difference
There are ways to enhance the tracking difference. For example, tax incurred on actual dividends compared to the tax applied to an index by an index provider (who calculates the index returns using pre-determined rules) can provide a source of additional returns to the fund.
Other sources of additional income may be securities lending or optimisation of security management. In the case of synthetic replication, judicious management may result in advantageous financing terms from counterparty banks, along with their special access to the market.
Ultimately, this can result in additional returns for the end-investor. The use of open architecture by counterparties for swaps can also generate competitive prices for swaps and may improve a fund’s tracking difference even more.
Reporting on and calculating the tracking difference
In any case, the period of measurement must be sufficiently long, that’s to say at least a year or more, to be considered representative.
Over a short period, tracking difference can be skewed by, for example, the seasonality of dividends. It should be noted that tracking difference is easily calculated: just compile an ETF’s performance data over the desired time period (using the ETF’s NAV and not transaction prices) and those of the index being replicated (based on the index’s benchmark level as stated in the prospectus).
Any bias that could lead to a miscalculation and, hence, an incorrect measure must be eliminated. To compare ETFs, it is essential to calculate tracking differences on the basis of the same index of replication to retain the same basis for calculation.
For example, the tracking difference of an ETF relative to a net total return index (i.e., with reinvested dividends and on an after-tax basis) cannot be compared with the tracking difference of an ETF managed against a price return or gross total return index (i.e., with reinvested dividends and on a pre-tax basis).
As with any calculation methodology, special attention must be paid to the periods under consideration, the price sources used (NAV versus last market price) and the exchange rate if the indices or ETFs are denominated in different currencies.
Tracking difference: essential when comparing ETFfs
When properly calculated, the tracking difference is an easy-to-use and direct measure and essential when comparing ETFs. In terms of total portfolio return, tracking difference is in our view a measure of the quality of the management of a tracker.
In active management, an investor selects a fund by analysing, among other things, its performance track record. In index management, we believe any comparison of performance track records must include an in-depth analysis of the tracking difference.
As managers seek to maximise the Sharpe ratio, optimising tracking difference, tracking error and the information ratio must be their primary goal.
For more on BNP Paribas Easy – ETF & index solutions, click here