On Tuesday, I wrote an article about how index trackers go about their task of tracking. Today, I am going to try to use this background information to look at how this might affect your choice of index tracker. If some of the terms in this piece don't make any sense, then I suggest you go back and read Tuesday's article. In the end, the business of choosing trackers comes down to personal attitudes towards the balance between risk and reward. However, as ever with personal finance, the basic points aren't too tricky but it gets difficult as soon as we try and apply it all to our own situations.
Which Index?
A tracker's performance is defined by its overall expected returns and the variation of these returns from year to year. Underlying both of these factors is the choice of index. In other words, most of the ups and downs of a FTSE 100 tracker will be caused by the ups and downs of the FTSE 100 itself.
The purpose of tracking is to get the average return of the market that you want to invest in. By doing this, you forego the best possible results but you also make sure that you avoid the worst possible results. However, choosing the market or index that you want to track is obviously of vital importance. There a trackers all over the world and to go into all of them would not be possible. I am therefore going to base all my comments on just two indices, the FTSE 100 and the FTSE All-Share. These are clearly of most relevance to UK Fools. The comments should generally be relevant to choices between other indices.
The FTSE 100 is an index of the UK's biggest 100 companies. The FTSE All-Share is designed to follow the whole of the stockmarket. To do this it contains about 800 companies. These are mostly, but not necessarily, the 800 biggest companies. However, even the FTSE All-Share avoids around a thousand "tiddlers". Still, as soon as the tiddlers start getting big, they tend to get included, so the FTSE All-Share should never be far away from the average performance of the whole market (to a lesser extent, the same can be said about the FTSE 100). The FTSE 100 companies make up about 80% by value of the FTSE All-Share so the differences between them are reasonably slight. However, they should not be ignored.
Expected Returns
As far as the expected returns from the FTSE 100 and the FTSE All-Share themselves (that is, not necessarily the trackers) are concerned, history has things very slightly in favour of the FTSE All-Share. The argument goes that smaller companies are more risky and, over the long-term, the market prices them to compensate you for taking this on. However, recent history (perhaps the last 10 or 15 years or so) has things on the side of the FTSE 100. The argument for this is that, in the modern globalising economy, big is best (because of things like "economies of scale"). If any threat is posed by the smaller companies then, the story goes, they just get taken over or crushed by the big boys.
All in all, in pure expected return terms, I would say there is very little to choose between the two indices (although, this is not necessarily so for the trackers following them). Other factors are almost certain to make much more of a difference.
Variation in Returns
In terms of variation in return against the average performance of the UK market, the FTSE All-Share has an advantage from being more broadly based than the FTSE 100. In terms of variation in return, then, I'd expect the FTSE All-Share to be very slightly better. However, there will be very little in it.
Deviating from the Index
On top of the expectation and variation in the returns of the underlying index, you also need to consider the amount by which a tracker misses its index. Tracking an index is not as straightforward as is sounds. Well, tracking an index cheaply isn't very easy anyway (some of the reasons for this are given in the article How do Index Trackers Work?). So, we need to look at the reasons why certain types of tracker might deviate from the index they're tracking.
One reason for a deviation that I'll deal with straight away is the question of dividends. Most indices, including the FTSE 100 and the FTSE All-Share, do not included re-invested dividends. As a result, any tracker which includes re-invested income (often called "accumulation units") will look better than its index (or it should do anyway) because they're getting an extra couple of per cent per year. This difference is purely artificial and, as long as you are comparing like with like, it can be ignored.
When companies describe deviations in the performance of their trackers, they seem to use the words "tracking error" to describe any departure from the performance of the underlying index. However, to get to the bottom of things, we Fools need to look at things rather more closely.
Charges
The number one reason for a tracker not performing in line with its index is its charges. It's the number one reason because it is easily quantifiable, operates in one direction (that is, downwards) and is relatively visible. If the index manages an annual rise of 10% and your tracker charges 1% per annum, then your return will only be 9%. It's as simple as that.
Well, not quite. You have to be very careful that you're picking up all the charges. First of all there is the annual management fee which is frequently described (incorrectly) as the total charges. In addition, there will generally be some "admin and trustees fees" or something like that.
Also, don't forget the effect of any initial charge and/or bid-offer spread. As one-off charges then, so long as they are fairly small, they will tend to have less effect over the long term than a higher management charge. There are a couple of posts, here and here, which go into this further.
On top of all this, there are the dealing charges involved in buying and selling shares to keep the tracker in line with its index. Quite often, the dealing charges which relate to new money coming into the fund (or old money leaving) will come through as a "bid-offer spread" (for more on this see How do Index Trackers Work?). Where there is no bid offer spread, the costs will be reflected in the annual fees. Overall it probably makes little difference how it is done, except that you need to be aware of it and include it in any comparison. Dealing to keep on top of the changes in the companies included in the index or dealing required by the statistical sampling technique being used should be very small indeed and are probably the same for all trackers (although maybe slightly worse for statistical samplers).
Tracking error
There are basically two types of tracking error. Either the error is as likely to be towards the upside as the downside or it is weighted in one particular direction. For want of better expressions, I will call these unsystematic error and systematic error. Not surprisingly, a downside systematic error is much worse than an unsystematic error (of the same amount), since it has definitely gone against you.
Systematic Tracking Error
Systematic error has a tendency to be in one particular direction. In this way, like charges (which, I suppose, are themselves a very pure form of systematic error), they affect the expected return of a tracker rather than just the variation in those returns.
All the time, a unit trust tracker has people putting money into it and taking money out. It therefore has to buy and sell shares to maintain its weightings in the index. However, in amongst all this buying and selling, the fund may, for instance, tend to have a very small permanent cash balance. In a rising market, this will tend to cause underperformance. In a falling market this will tend to cause outperformance. Since markets tend to go up over the longer term, then this would be a systematic error towards the downside. This type of systematic error would be a problem for all types of trackers (unit trust ones anyway).
Unsystematic Tracking Error
I can think of two main sources of unsystematic tracking error. The first is when new companies come into and leave the index. The tracker generally has some flexibility to decide when to sell the old companies and buy the new ones. Sometimes it will get it right, sometimes it will get it wrong. On the whole it won't make much difference. Having said that, it could be argued that shares which are about to join the index are heavily in demand and that trackers always have to pay too much for them. This would turn part of this error into a systematic downside error. However, since we are talking about the shares joining the index we are generally talking about very small weightings and the effect will very small. This will mainly be a problem for the FTSE 100 trackers, since the relative weighting of the shares coming into and leaving the index is much greater for the FTSE 100 than for the FTSE All-Share. It is also more of a problem for full replication trackers than statistical sampling trackers, since the latter don't necessarily have to buy the new entries.
The second type of unsystematic error is almost entirely a problem for the statistical samplers. This makes it more of a problem for the more widely spread indices like the FTSE All-Share (which are more likely to use this technique). It simply comes down to the fact that they do not actually own (or attempt to own) all the shares in the relevant index. So, if the shares they don't own happen to do very well, then the tracker will underperform. Of course, it works both ways and if the shares which the tracker doesn't own do relatively badly, then the tracker will outperform. How big a variation you get here will be down to the index tracked, how good the manager is at tracking and how much (of your) money the manager throws at the problem. If you're prepared to pay the extra, then you could fully replicate the FTSE All-Share, but would it really be worth it? In the end it will always be a bit of a balance.
The Moral of the Tale
The trouble with all this tracking error stuff is that you need a very long performance history to have any real confidence about the level of overall tracking error. You'd need longer still to have any confidence about how much of this was systematic error and how much was unsystematic. So you end up making guesses about a lot of unquantifiables. Still, it's quite fun really and good to get tracker fans doing some hard core guesswork. The direct equity Fools thought this was exclusively their preserve, so it's nice to prove them wrong!
The moral, then, is to analyse a tracker just as you would anything else. Look at the risk and the expected return. Of course, the beauty of investing in trackers is that, so long as you follow a few fairly obvious fundamentals, it shouldn't make a whole heap of difference.
First of all, you want to avoid high (apparent and hidden) charges. Secondly, you want something which isn't likely to vary too far from the index (within reason). Finally, if there is to be any variation, then you'd obviously prefer it to be even-handed. The second and third points are likely to come down to thinking about the index that you're tracking, thinking about how it is being tracked, and thinking about the reputation of the company doing the tracking. As trackers get a longer and longer history (they're a comparatively new breed), we'll be able to make better and better judgements about who's any good at doing it. For the time being we'll have to feel our way a bit. Still, we can be sure that our expected return is going to be higher than for a managed fund.
Questions and comments on index trackers should go to the Index Trackers discussion board. More specific matters relating to this article should go to the Personal Finance discussion board.
Relevant Links
• How do Index Trackers Work?
• Mega Mergers and Index Tracking -- Part Two
• Mega Mergers and Index Tracking
• A tale of Two Fund Managers
• What Do You Trust?
• Twisting By the Pool
• The Trouble With Trackers
• Great Expectations
• Financial Deviance
• Index Trackers Discussion Board
• Personal Finance Discussion Board
Skip Navigation
Fool USA