simple passive investment - is this OK?

TheTracker

Ryan_Futuristics wrote: »

The fact managed funds often hold more in cash and alternative assets (bonds, REITs, foreign equities) means they probably shouldn't be compared to pure equities indexes

e.g. If a fund hold 5% in property, it should probably be compared to a 95% equities index and a 5% property index

A favourite chart showing Vanguard's own active funds consistently outperforming their indexes, and cheap actives in general doing the same

The Morningstar blog that published this graph warns against easy interpretation. Quite simply the vanguard passive funds are a result of all funds over the period measured. However, the active funds are only that have survived, as many have sounded the death knell,and these are not shown. The blog entry cautions against any interpretation that cheap actives beat cheap passives, in fact it says they are basically equal if you adjust for survivorship bias, even though the author doesn't attempt to. As to the vanguard actives, I have little knowledge of these. I imagine there are only several dozen of them at most and they are unfairly brought out against the thousands that represent the other bar in the graph. The blog author has hardly done a scientific comparison, but I doubt it is statistically significant.

It's a fairly noddy graph tbh - the vanguard passives are an average of all fund performance. All the graph seems to be saying is cheap actives beat expensive actives.

Ryan_Futuristics

masonic wrote: »

Well the herd, in terms of the chart above, will be lumped into the expensive active camp. I don't think there's any danger of your current audience falling into that trap.

What's interesting from that chart is that the results from 'other' cheap actives does not appear materially better than passives, even after the expensive active funds have been removed. I'd be pretty disappointed if the active funds I hold were ranked in the same decile as passive funds, like the chart seems to suggest, but it is probably because we are looking at the average cheap active fund. In fact, you are normally the first to take issue with people comparing the average managed fund to trackers, with the average active fund pretty much constituting the market the passive funds are tracking.

The chart does raise the question: what exactly are these Vanguard Active funds? I'd be inclined to think these are something along the lines of the smart-beta products that are starting to emerge here. I'd suggest whether or not these really should be called active or passive is a moot point. The reason they are outperforming the other active funds in aggregate is likely down to lower charges and lower portfolio turnover. But nobody is going to go out and buy the available funds universe in aggregate, so I don't see the point of the comparison.

I don't think the herd are necessarily in expensive active ... Most fund supermarkets these days penalise expensive funds - so on a site like HL, the top 10 funds are usually in the cheap camp

'Average' is an unhelpful concept until you start breaking it down ... As you say, a lot of those actives are going to be closet trackers ... Remove them (look at 'active share') and you also improve group performance



The active/passive debate's been raging decades - whenever anyone's broken the data down, it's not stood up well for the 'efficient market' theorists


Re: Vanguard Actives

These are actually very actively managed funds with top fund Wall Steet fund managers ... These are things like Vanguard Windsor Fund (VWNDX), Vanguard Selected Value, Vanguard Balanced Fund Admiral Shares

Vanguard's CEO has his pension and savings in their active funds

Ryan_Futuristics

TheTracker wrote: »

The Morningstar blog that published this graph warns against easy interpretation. Quite simply the vanguard passive funds are a result of all funds over the period measured. However, the active funds are only that have survived, as many have sounded the death knell,and these are not shown. The blog entry cautions against any interpretation that cheap actives beat cheap passives, in fact it says they are basically equal if you adjust for survivorship bias, even though the author doesn't attempt to. As to the vanguard actives, I have little knowledge of these. I imagine there are only several dozen of them at most and they are unfairly brought out against the thousands that represent the other bar in the graph. The blog author has hardly done a scientific comparison, but I doubt it is statistically significant.

It's a fairly noddy graph tbh - the vanguard passives are an average of all fund performance. All the graph seems to be saying is cheap actives beat expensive actives.

Vanguard actually have 30 active funds - some of which have been around since the 70s - I believe only one's been closed (but they've also closed at least one passive .. That Morningstar article may have better figures)

And almost all have consistently beaten their indexes ... They don't tout it though ... They advertise their passives to regular investors, and let financial advisors (and their own board), etc use the actives ... Odd situation

http://uk.reuters.com/article/2014/12/16/us-funds-active-vanguard-analysis-idUSKBN0JU25L20141216

TheTracker

Ryan_Futuristics wrote: »

Vanguard actually have 30 active funds - some of which have been around since the 70s - I believe only one's been closed (but they've also closed at least one passive .. That Morningstar article may have better figures)

And almost all have consistently beaten their indexes ... They don't tout it though ... They advertise their passives to regular investors, and let financial advisors (and their own board), etc use the actives ... Odd situation

http://uk.reuters.com/article/2014/12/16/us-funds-active-vanguard-analysis-idUSKBN0JU25L20141216

For those interested here is the vanguard white paper on their active initiative. In figure 2 they show they've achieved a 0.24% pa premium over benchmark indices. Me, I'll stick with passives.

redux

I understand the often made ,argument that many fund managers fail to beat index trackers, due to higher management fees and dealing charges on churned holdings, but surely not all investment decisions are right, some managers are better than others, just as the same is true of companies, so the investor's ideal is to pick the best ones

Ryan_Futuristics

TheTracker wrote: »

For those interested here is the vanguard white paper on their active initiative. In figure 2 they show they've achieved a 0.24% pa premium over benchmark indices. Me, I'll stick with passives.

And lower volatility, in the most overanalysed and difficult to beat market in the world

That 0.24% even taken in isolation is still the difference between an 'expensive' fund platform and a cheap one

By all accounts, they're better deals

Ryan_Futuristics

redux wrote: »

I understand the often made ,argument that many fund managers fail to beat index trackers, due to higher management fees and dealing charges on churned holdings, but surely not all investment decisions are right, some managers are better than others, just as the same is true of companies, so the investor's ideal is to pick the best ones

Well I do like ways to remove human error from the equation - I appreciate that people who make good decisions can also make bad ones

e.g. I can't guarantee Neil Woodford will continue to beat the index ... But the fact his funds are low fee (as low as trackers were not long ago), and have a high active share, puts them in a category which has generally been unlikely to underperform

But I always think it's worth remembering that statistically your choice of fund tends to exert less than a 1% influence on your returns ... Woodford Equity Income's about 10% ahead of the index since launch, but it's only 10% of my portfolio ... so ... 1%

masonic

Ryan_Futuristics wrote: »

These are actually very actively managed funds with top fund Wall Steet fund managers ... These are things like Vanguard Windsor Fund (VWNDX), Vanguard Selected Value, Vanguard Balanced Fund Admiral Shares

I haven't looked at all of them, but the Vanguard Windsor Fund is highly correlated with the Russell 1000 Value Index (R squared = 94% over 3 years), and the returns are very similar over 1, 3, 5 and 10 years, as is the performance chart over those 10 years. The same seems to be true of Vanguard Selected Value and its benchmark, the Russell Mid-Cap Value Index.

So these look very much to me like trackers, or low cost closet trackers. It's just a shame we don't have similar UK-listed products (covering US equities) available to us, not even expensive ones.

bowlhead99

Ryan_Futuristics wrote: »

What you have there is a graph of how different levels of 'activeness' appeared to drive returns, although of course correlation doesn't imply causation. With 'active share' measuring a deviation from the benchmark holdings. Presumably we're saying that the bar on the right must contains the most passive funds and closet trackers which between them are doing 300bps worse than the most active which are on the left.

But don't we see other research pieces, graphs and league tables showing passive funds coming mid table between the actives that get it right and the actives that get it wrong? So if we accept that the trackers are in the 0 to -100bps range, then in the far right columns we must have a mix of all those trackers with mid table performance, and all the high fee active closet trackers with terrible performance of -250 to -300, which is driving the overall performance of that graph column, so that on average that group has -150bps below benchmark (and -300bps below the most active funds on the left).

While in the far left column we have the most active funds that deviated furthest from the index. This must include the active funds which got it right, together with the ones that also deviated from the index and got it wrong. We're saying that the net result of mashing those high deviation funds together is +1.5%. But we know that on the upside a high active fund that gets it right can more than double your money; it could deliver 5x while a bad one cannot lose more than all of your money and then 4x more. So putting both good and bad active funds next to each other may well yield a positive average result providing a natural boost in the goal of reaching or beating the sector average.

You then have survivorship bias because all the good high active funds are recorded but the bad high active funds may have died during the 14 year period and will not be included. So it's quite easy to see reasons why the most highly active quintile of funds end up with a positive position in the league table.

What if any conclusions can be drawn here? Well, basically, higher returns are available if you select a high active fund that gets it right. N.S., Sherlock. The practical problem will be that the high active funds that get it right are rather less visible in advance than with hindsight.

The only block you desire to be in on the graph is the one on the left because the outperformance looks the best. But the outperforming area is a result net of underperforming sharks and funds that fail, all crammed into that same quintile. Active share definitely tells you if a manager is consistently different, because that's what 'active share' measures, but that still doesn’t mean they’re actually good. So that column with the positive result is fraught with risk because the good high active managers for the next economic cycle are only one part of the general bucket of all high active managers with conviction portfolios, some who will deliver a terrible result.

To lift from someone else's article, "Active share, like tracking error, is a measure of conviction but conviction is not guaranteed to produce results. Big bets can lead to big wins or big losses. To win, managers need conviction and (this is the big caveat) they need to be right." The headline from the graph is that you can get the highest outperformance from high conviction portfolios. Well yes, a high risk that pays off, you'd hope for a high reward.

So, basically you can try to pick the high conviction funds and hope they get it right - selecting a fund from the far left column of highly active managers, but crossing your fingers to hope it is a good one rather than a bad one. Or you can quite happily pick an inactive tracker with a low fee from the far right column, knowing their performance will be much better than the higher fee managers in the same column who are dragging the implied return downwards. You can get middling benchmark returns with no worries from a fund in that column.

So, what have we learned? If you go high active, using non-standard choices to drive investment decisions, and get it right, you can make more money than if you take the average return. If you go low active and keep it low fee, you can make money (receiving a few bps below benchmark as a result of a 0.1% AMC is still making money). But neither of those concepts is new to anyone, they are common sense.

The challenge if you want the absolute top return is to ensure that when you are going active you pick the star fund rather than the dog, for the prevailing market conditions. That is tricky (a fraction of one quintile of the available managers) and that warning is what passive investors have been saying for years: stick to what you know to be a decent long term result rather than gambling for a better one.

All the chart is telling you is that the set of managers producing the top performance include the most active managers. That has to be obvious even without any graphs because we know the tracker funds are geared up to provide middle of the road solid performance without taking gambles. The ones that take gambles and follow (good) convictions correctly, can get better results. They are just hard to find.

Just for fun, the below comment comes from the same blog hosting Ryan's graph:

Active Share and excess return: The most contentious point in the industry’s debate on Active Share is the headline that managers with higher Active Share measures tend to earn higher excess returns. Research subsequent to Cremers and Petajisto shows that some, but not all, high Active Share managers outperform in certain market regimes, and that the relationship is not a simple one. In the eVestment universe, we have found a strong and consistent relationship between Active Share and tracking error; however, high Active Share is not necessarily associated with high excess return.

For both our U.S. large cap core and U.S. small cap core manager groups, for the three years ended March 2014, tracking error and Active Share tended to rise together (the dark blue bars in the chart below, versus the Active Share scale on the left).

The relationship between Active Share and excess return was inconsistent (shown by the light blue bars). In U.S. large cap core, the highest excess return for that period accrued to managers with Active Share scores of 90 to 95 and to those with scores of 55 to 65, but not to those in between. For the eVestment U.S. small cap core group, excess return rose as Active Share increased from 60 through 90, but dropped off at the highest levels. Moreover, some of the group’s highest excess returns were earned by managers with low Active Share.

So, with other datasets it's not even clear how correlated Active Share was with good performance, let alone whether it caused it.

masonic

bowlhead99 wrote: »

While in the far left column we have the most active funds that deviated furthest from the index. This must include the active funds which got it right, together with the ones that also deviated from the index and got it wrong. We're saying that the net result of mashing those high deviation funds together is +1.5%. But we know that on the upside a high active fund that gets it right can more than double your money; it could deliver 5x while a bad one cannot lose more than all of your money and then 4x more. So putting both good and bad active funds next to each other may well yield a positive average result providing a natural boost in the goal of reaching or beating the sector average.

I was thinking along the same lines and was going to mock up an 'artist's impression' of what the error bars might look like, so for those that prefer to see things graphically...



It's not hard to see why some would gladly give up average excess returns of a little under 1.5% in exchange for the confidence of not picking a conviction play that goes badly wrong.

Edit: Incidentally, how does one achieve 100% active share? Does that mean the fund does not invest at all in any of the constituents of its benchmark? Doesn't that just mean it's being compared to the wrong benchmark?