Investors who don’t want to pick stocks have a simple and attractive investment option in the form of index trackers. An index tracker is a fund designed to investors exposure to a basket of stocks that comprises an index. This means investors get the benefits of diversification at (usually) relatively low cost. For tracking the popular Straits Times Index (STI), the lowest cost options are two ETFs traded on the SGX – the SPDR Straits Times Index Fund (SPDR STI) and the Nikko AM Singapore STI Fund. The average total expense ratios (TER) of these two ETFs in the most recent 5 years (to Dec 2017) are 0.30% and 0.37% respectively. All things equal, low TERs mean higher net returns.
A less well known STI tracker is the Singapore Index Fund (SIF). The SIF is not an ETF but a unit trust whose investment objective is to replicate the returns of the benchmark FTSE Straits Times Index. While this sounds a lot like what the SPDR STI and Nikko AM Singapore STI funds are doing, investors should bear in mind that unit trusts charge higher TERs than comparable ETFs. Another factor that may pull down a unit trust’s return is the ‘cash drag’ resulting from the need to hold a portion of the trust’s assets in cash to meet fund redemptions. On the flip side, being a unit trust gives the manager the flexibility to settle for less than perfect index replication either to reduce replication cost or as a deliberate strategy to profit by deviating from the index. In view of these complexities, I will let the data speak for itself by comparing the returns of the more established ETF, the SPDR STI against the SIF.
The sample period I use for this comparison is from May 2002 to Dec 2017. I start the data in May 2002 since the inception date of the SIF is April of that year. I exclude the Nikko AM STI as its inception date is later (February 2009). My data is downloaded from Datastream. The data is monthly and comprise total returns with dividends reinvested. All returns are net of costs.
The table below shows the mean and standard deviation of returns (annualized) for the two index trackers relative to the STI. I also show the ratio of mean return to volatility, which is roughly a fund’s ‘Sharpe ratio’ (SR) even no risk-free rate is actually subtracted from the mean. Finally, to see whether performance characteristics are robust over time, I report results for two sub-periods: 2002-2007 and 2008-2017 in addition to the full sample period.
The main results can be summarized quite easily:
- The SIF is consistently more volatile than the STI.
- Over the full sample period, this higher volatility did not translate to higher mean returns compared to the index. The SIF’s mean return (8.7%) is only marginally higher than that of the SPDR STI (8.4%) but in terms of the Sharp ratio, it is worse than either the index or the SPDR STI.
- The SIF delivered a slightly higher mean return than the STI in the first sub-period. But it took too much risk to achieve this. Consequently, it continued to under-perform both the STI and the SPDR STI in terms of the Sharpe ratio.
- The second sub-period saw the worst performance for the SIF. Despite having almost the same volatility as the STI (19.5% versus 19.2%), the SIF’s mean return was only 4% compared to 5.1% for the STI and 4.6% for the SPDR STI.
- Relative to the index, the STI under-performed by an average of 1.1% per annum in the second sub-period. Its under-performance was twice that of the SPDR STI.
- As all three ‘indices’ had very similar risk profile in the second sub-period, the stark performance gap of the SIF in this period is consistent with its higher expense ratio compared to the SPDR STI. This is what investors of unit trust index trackers should expect if the fund manager made no attempts to deviate from the index.
Conclusion: Clearly, not all index trackers are created equal!