Active funds vs Trackers (again...)

koru

Loughton_Monkey wrote: »

I'm really not sure about this. Well resourced? Probably. Clever? Not so sure.

Well, I am not saying that all fund managers are clever. But a lot of them are. If almost all managers were stupid and were making easy decisions, then a clever fund manager would have no difficulty at all in outperforming them by a considerable margin. But in fact, it is extremely difficult to outperform the market consistently.

koru

jamesd wrote: »

The efficient markets theory is obviously false for the general case, since there are people and companies that have long records of outperforming trackers, beyond what can be attributed to chance, and have become famous for doing so.

Well, it is certainly true that those who manage to outperform the market for more than a year or two do become famous, but this is because they are so rare. However, is this really beyond what can be attributed to chance? How many years of outperformance would count as being beyond chance? In practice, a manager who outperforms the market for four years on the trot is likely to become very famous indeed, but four years could easily be attributable to pure luck.

In any single year, roughly 50% of fund managers will beat the average and 50% will underperform. So, the chances of a random manager outperforming in a single year is one in two. Pure chance would suggest that there is a one in four chance of outperforming for two consecutive years. If you carry on the maths, there is a one in 16 chance that a random manager will outperform for four years. Put another way, for every 16 managers you would expect that as a result of pure luck, one is likely to outperform for four years on the trot. Since there are hundreds of fund managers, you'd expect lots to outperform for four years on the trot, just as a result of pure chance.

jamesd

Luck or lucky speculation can and surely does happen. Pick a high risk strategy and you'll either fail and not be noticed or succeed and be noticed, a perverse reward system since it's contrary to the advantage of consumers, even though it's arguably in the interests of fund houses.

Longer term it's harder because the approaches needed to do better than average change over time, so if you take just one approach that works well during a boom market you're likely to under-perform during a bear or level market. Or a medium size company approach will fail to outperform when large companies do well.

The unusually good cases shouldn't be surprising, since investing is a human activity and in all fields there are some humans who are better at it than others. Even in fields where there are many excellent performers. You still might do better to avoid the most competitive areas and pick ones where there are more of the less capable performers who can be exploited by the better ones.

It's known, and I assume you believe it, that there are active managers who consistently do less well than average. If half outperform you can eliminate the bad ones and on average your expectation is now greater than average.

You can do the same with trackers. If you want to consistently outperform the average tracker, just pick one with good index correlation and lower fees. That'll leave you consistently beating the ones with fees that are 0.9% higher.

It's tougher with active managers because you're eliminating worse human and company skills but outperformance might not persevere either, if it happens to be the result of a style choice that ceases to work.

In the US markets it's even harder for active management to succeed, because there are higher taxes on shares that are held for less than a year, so you can get the situation where the average active manager outperforms before tax but not after tax, in fully taxed investing. Hence the popular US approach of using holding turnover to try to eliminate those who are less likely to do well. There are still turnover costs here but at least they aren't so large.

vectistim

The relevant index for a particular sector is effectively the average of all the trades. Tracker funds aim to match the constituents of that particular index and so, due to charges should provide performance just short of the index's overall return. The active funds are then distributed either side of the average, so if you pick the right active fund you can exceed the market average, if you pick any passive fund you don't get that chance. Its then up to you whether the risk of a larger underperformance is worth it for the chance to exceed the market average return.

koru

jamesd wrote: »

It's known, and I assume you believe it, that there are active managers who consistently do less well than average. If half outperform you can eliminate the bad ones and on average your expectation is now greater than average.

Just as pure chance can explain sustained outperformance, so it can explain sustained underperformance over a given period. If it is down to chance, then it tells you nothing about likely future performance. You can certainly identify funds that have underperformed, but I don't think that provides a rational basis to assume they will continue to underperform.

The evidence that I am aware of shows that poor performance in the past is not a reliable indication that the manager will perform poorly in the future. The October 2002 report by CRA for the Investment Management Association doesn't seem to me to indicate that it is easy to identify managers who will continue to perform poorly. For instance, looking at the managers in the UK All Companies sector (page 26), more than half of the managers who were in the bottom quarter of performance for the previous three-year period were in the top two quartiles of performance in the following three year period.

The results vary a little depending on what time period you look over, and what sector you are looking at, but my overall conclusion from looking at these figures is that past performance (whether good or bad) tells you virtually nothing about likely future performance. If your aim is to choose a fund that will perform better than average in the future, eliminating past underperformers does not increase your chances. You might as well close your eyes and stab a pin in a list of all the fund managers in the sector.

The report is available here:
http://www.investmentfunds.org.uk/assets/files/press/2002/20021014-01.pdf

koru

vectistim wrote: »

The relevant index for a particular sector is effectively the average of all the trades. Tracker funds aim to match the constituents of that particular index and so, due to charges should provide performance just short of the index's overall return. The active funds are then distributed either side of the average, so if you pick the right active fund you can exceed the market average, if you pick any passive fund you don't get that chance. Its then up to you whether the risk of a larger underperformance is worth it for the chance to exceed the market average return.

Yes, assuming the tracker tracks the index perfectly, it will underperform the index by an amount equal to its charges. Active funds will be distributed either side of the average, but they will tend to be below the average because they have much higher charges, so their performance net of charges will, on average, be poorer than a low charge tracker. Of course, if you can spot the active funds that will outperform, this would be better, but I've never seen any evidence that there is a reliable way to do this. (And I've read a lot of academic studies in the hope of finding such evidence.)

dunstonh

if you can spot the active funds that will outperform, this would be better, but I've never seen any evidence that there is a reliable way to do this.

If you strip out the passive managed funds (and probably bank and insurance company funds with it), then you are left with a much smaller list of managed funds. That will improve your odds.

Dont forget that the managed fund will often not have the same objective as the tracker. So, you are never going to get like for like comparisons.

The fact is that no-one can tell what is going to be best. The trackers will typically be mid table but consistent (unless focused areas where they could jump from top to bottom if the focused area within that sector is doing well/bad). For that knowledge that you are getting consistency, you pay less. Then you have the managed funds which offer the potential to outperform but with the possibility that they will under perform. For that you pay more.

Which is going to be best in future? No-one can tell. Its very much a personal decision.

Personally, I like using both depending on the sector and objective. I dont see why you should rule either out.

jamesd

koru wrote: »

my overall conclusion from looking at these figures is that past performance (whether good or bad) tells you virtually nothing about likely future performance.

Thanks for providing the source. Your conclusion is somewhat different from the executive summary of the report on page 1 which states:

"* Performance broadly persisted in UK equity based unit trusts between 1981 and 2001. ...
* Based on this information it is possible for retail consumers (and their advisers) to use this performance information to aid their decision-making. ...
* The importance of persistence depends on both the time horizon and the sector in which the fund is invested. ...
* Choosing a top quartile fund, as opposed to a bottom quartile fund will, on average, add to an investor's potential return. ...
* The results are not sensitive to charges. We find that the persistence of unit trusts is not counteracted by charges. ...

Therefore we conclude that, based on our dataset, consumers (and their advisers) can use past performance information as a beneficial part of their investment decision-making process."

I'm still not greatly keen on the report, for it has the systematic error that is common in such reports, ignoring manager (the human team) performance and using fund products instead, so it misses performance changes that move with individuals.

The period you chose is an interesting one where past performance may well fail, because there was a radical change in the market within the period: the start and then end of the tech boom, along with the market change from boom to bust. Such changes can be expected to favour changes in style and may invalidate past absolute outperformance, though perhaps not relative outperformance compared to other funds in the same sector (all tech may drop but some managers may drop more or less than others, but most important was an asset allocation that was low in tech).

koru

jamesd wrote: »

Thanks for providing the source. Your conclusion is somewhat different from the executive summary of the report on page 1...

I'm still not greatly keen on the report, for it has the systematic error that is common in such reports, ignoring manager (the human team) performance and using fund products instead, so it misses performance changes that move with individuals.

The period you chose is an interesting one where past performance may well fail, because there was a radical change in the market within the period: the start and then end of the tech boom, along with the market change from boom to bust. Such changes can be expected to favour changes in style and may invalidate past absolute outperformance, though perhaps not relative outperformance compared to other funds in the same sector (all tech may drop but some managers may drop more or less than others, but most important was an asset allocation that was low in tech).

Well, CRA were hired to support a certain conclusion, so that's what they concluded. I guess we can all look at the data and form our own conclusions. My own view is that if you take particular sectors over particular periods, you can convince yourself that performance tends to persist, but this seems equally balanced by other periods/sectors which suggest that performance tends to reverse. However, in almost all cases, the odds do not seem to vary much from what you would expect if performance is purely random.

I agree with your reservations about the report. Unfortunately, it seems to be about the best that is available. However, given that it was intended to prove that performance does exist, I think it is quite significant that they were unable to find better evidence in support of that proposition than what is set out in the report.

koru

dunstonh wrote: »

If you strip out the passive managed funds (and probably bank and insurance company funds with it), then you are left with a much smaller list of managed funds. That will improve your odds.

Why do you say that? Is it because you have some sort of analysis which shows that these categories of funds perform poorly more often than the specialist truly active funds or just that you think it stands to reason? You may be right, but if there is some sort of analysis that backs this up, I would genuinely love to see it.