New rules require tracking errors to be disclosed in index-tracking Ucits funds. Judith Evans looks at the difficulties involved and discovers which of the methods used suit different types of investor, the short-term trader, or the long-term holder.
As investors pile into exchange-traded funds, they are increasingly scrutinising performance. They are also recognising that the process of tracking an index is not so simple as it seemed.
In a crowded European market, indicators of tracking performance can be key to investors’ choices, says Marlene Hassine, ETF strategist at Lyxor. “ETFs have to deliver the best performance, because that is what investors want. That’s the focus of the market today,” she says.
An ETF’s tracking efficiency depends on several factors. Since many indices are weighted by market capitalisation, their contents can fluctuate constantly with share prices, an effect that is difficult for ETFs to mirror; meanwhile, costs and fees generally cause a tracker to slightly underperform its benchmark.
The focus on performance is forcing a re-assessment of ways to measure trackers’ efficiency, with some market players arguing new metrics and standards are needed.
ERROR AND DIFFERENCE
ETFs’ performance in tracking their indexes has traditionally been measured using two metrics: tracking difference and tracking error.
Tracking difference measures the annualised difference between a fund’s return and that of its benchmark, while tracking error measures the volatility of this gap – important for investors who may be operating on shorter timeframes or using the ETF as part of a complex strategy. In technical terms, tracking error is the annualised standard deviation of the tracking difference.
But tracking error and tracking difference can be calculated in different ways using various time horizons, making it difficult to accurately compare one fund with another using the figures issued by providers.
From 2013, index-tracking Ucits funds will be required to disclose tracking difference and tracking error, and to estimate them for the year ahead – but the regulation doesn’t specify how they should be calculated.
Ben Johnson, director of passive fund research at Morningstar, says standardising these measures was “essential in order for investors to be able to make like-for-like comparisons amongst ETPs [exchange-traded products] and other index-tracking funds”.
Some analysts argue that in any case, the two figures are not enough. Morningstar and fund provider Lyxor have both put forward new metrics this year, arguing that markets need a fresh approach.
LIQUIDITY
“One of the specificities of this traded, listed product is that you have to take into account the cost of liquidity,” says Hassine.
“There can be a very big difference between the liquidity of different products. You can have an ETF with a [bid-offer] spread of 10 percentage points and another with a spread of 20, 25 or 30, depending on the liquidity that is offered by the provider.”
Robert Rushe, head of ETF servicing for Ireland at State Street Global Servicing, adds: “The majority of ETF trading [in Europe] is over-the-counter, so it’s difficult to see how liquid it is.”
Inexperienced ETF investors, or those lacking contacts among the best market makers, could hit performance by up to 100 basis points in one trade, said Rushe, whose firm is the biggest asset servicer for ETFs in Europe.
The second instalment of the Markets in Financial Instruments Directive (Mifid II), expected in 2015, is set to change this situation by requiring that all ETF trades are reported, he says.
Lyxor’s metric, christened the ETF Efficiency Indicator, uses a value-at-risk formula to incorporate tracking difference, tracking error and liquidity spread; the resulting figure represents the probability that the investor will face a loss or gain compared with the index’s performance.
SMOOTHING OUT
Analysts also point out that tracking error is a static measure, relying on one set of returns to judge the fund’s performance.
Morningstar’s new metric, called Estimated Holding Cost, seeks to improve on this by calculating 20 return ratios between the fund and its benchmark over set periods, then taking an average for a “smoothed, annualised estimation of the ETF’s divergence from its index over a year”.
This figure takes into account all the costs associated with holding an ETF – total expense ratio, securities lending revenues, rebalancing costs and so on – but in contrast with the Lyxor formula, it excludes costs associated with buying and selling units in the fund.
The company says its metric is aimed at “buy-to-hold investors”, who should look for the estimated holding cost to be as low as possible.
Johnson says: “Keeping the estimated holding cost from our market impact measure [which incorporates bid-offer spreads] allows us to create tools that enable investors to run bespoke total cost of ownership calculations that reflect their unique circumstances.”
TRACKING ERROR USE
Analysing tracking error can still be useful to investors; for example, a higher figure can indicate a lack of liquidity in the underlying securities.
In general, Morningstar found, tracking error was “limited” among the ETFs under study. Funds tracking the FTSE 100, Dax, MSCI Japan and S&P 500 indices were among the closest to their benchmarks, with tracking errors of about 4 to 7 basis points over periods that the firm assessed (June 12, 2010 to September 30, 2012).
Those tracking the MSCI Emerging Markets Index were the most volatile, with annualised tracking error of 0.81% on average.
Synthetic ETFs, which use derivatives to track an index rather than holding the relevant securities, generally have lower tracking errors than physical ETFs, Morningstar’s research found – but they do come with extra counterparty risk.
Tracking error is of most concern to short-term investors and those using the ETF for hedging purposes, while long-term investors should be more concerned with tracking difference, Morningstar said.
Tracking difference can be affected by total expense ratio, securities lending income and tax optimisation, as well as the costs of rebalancing physical ETFs and swap fees for synthetic vehicles.
Tracking differences of FTSE 100 trackers averaged -0.55% for physical replicators and -0.39% for synthetic trackers from July 2010 to September 28 2012, Morningstar says.
Rushe is sceptical that a new measure of ETF performance will become standard.
“It’s difficult to find something that works for everybody and it’s difficult to move away from what the market understands these things to be,” he says.
But in terms of how ETFs work, he adds, “people are educating themselves in leaps and bounds”.
©2013 funds europe
