Skip to the content

‘Survivorship bias’ is not always the argument some seem to believe it is

27 September 2016

Kevin Murphy takes a closer look at survivorship bias and asks how concerned investors should be about it.

By Kevin Murphy ,

Schroders

In the year we celebrate a decade at the helm of our flagship fund, it is perhaps understandable that we have been hearing a lot of the phrase ‘survivorship bias’ – the idea the past returns of existing mutual funds are flattered by the fact poorly-performing funds will often be closed or merged away by their management groups. Understandable, yes – but is it fair?

Certainly the view any fund sector or indeed group of assets needs to be ‘adjusted’ equally for survivorship bias is not universally applicable.

Take the UK market as an example and imagine a large company falls on hard times, dwindles slowly and eventually goes bust – only a small amount of its decline is registered by the FTSE 100 before it drops out of that index. Clearly the index is flattered by jettisoning the declining company and an adjustment would be appropriate.                                                   

There is also survivorship bias as the company continues its decline through and out of the FTSE 250 – though not so much as that index has a lot more constituents, meaning more of the decline is retained before the company exits the index.

Generally speaking, the more constituents in an index, the less affected it is by survivorship bias. As such, there is still some survivorship bias in the much larger FTSE All Share, as companies usually delist before they go to zero, but less than with the FTSE 100.

There is little doubt, though, that survivorship bias is more of an issue for fund managers – a point clearly illustrated by way of an example used by David Swensen, who has been chief investment officer of Yale University’s hugely successful endowment fund for more than three decades, in his 2005 book Unconventional Success.

It concerns Long Term Capital Management (LTCM), a hedge fund that enjoyed three bumper years before it – and its investors – were pretty much wiped out by the 1997 Asian financial crisis and the 1998 Russian debt crisis. LTCM and other hedge funds used to submit their performance numbers to a company called Tremont, which would use this data to calculate benchmark indices for the sector.

From March 1994 to October 1997, the performance of the wider hedge fund fraternity was greatly boosted by LTCM, which made an impressive 32.4 per cent net of fees for its clients. From October 1997 to October 1998, however, when LTCM to all intents and purposes ceased to exist, those positive returns became a 91.8 per cent loss – yet this was never reported to Tremont and so never appeared in its data.

Performance of fund vs sector and index over 10yrs

 

Source: FE Analytics

In this way, then, survivorship bias represents a difference between reality and perception – and you can see something similar going on in the context of the 10-year anniversary we mentioned earlier. When we took over managing the fund in July 2006, its one-year performance numbers were ranked against a peer group of 300-odd competitors and, 10 years on, this remains the case.

 

Initial competition

The fund’s 10-year numbers are, however, ranked out of 194 competitors, which at first glance seems strange.

It had 300-odd competitors in 2006 and a similar number in 2016 – they are just, to a large extent, a different 300-odd funds. Over the last decade, around one-third of our initial competition have been closed down, merged away or whatever by the investment houses that ran them, to be replaced with brand new funds offering some fresh allure of success.

Now, one thing you can say with a fair amount of certainty is that investment houses tend not to close down or merge away funds that perform well. The funds that have ceased to exist and whose records have consequently been expunged from the record of our own fund’s peer group are therefore likely to have been the poor performers.

What this means is that, when you are checking to see whether a fund is, say, a first-quartile or a fourth-quartile performer over one year, you do not really need to think about survivorship bias. In contrast, when it comes to a fund’s quartile ranking over five or 10 years, there can be huge survivorship bias to consider – and two different ways of thinking about this.

Before we go any further we should stress, as you will realise should you look up our own quartile ranking over 10 years, that there is no special pleading going on here. That out of the way, the next time you see a fund that is ‘only’ a second-quartile performer, it is worth thinking about where it might have stood had not a long tail of poorly-performing funds been erased from history.

Although that of course leads on to the second way of thinking about survivorship bias – and one much favoured, unsurprisingly, by champions of passive investing – which is that the average performance of actively-managed funds over time would be materially worse if all the ex-funds were included in the calculations.

And you know what? We would not take issue with that point because, well, you cannot argue with the facts.

What we would point out, however, is that that take on survivorship bias is no real argument for passives but rather for picking a fund that can boast a durable, rigorous and repeatable investment process that history has shown can outperform over the longer term – such as, for example, a valuation-based approach.

An alternative – albeit one that is the antithesis of a value strategy – is to pick a fund that just swings hard for the fences every time. That is all very well as long as they are connecting but, if they miss, they strike out – and such funds tend not to be the ones that survive long enough to have to worry how their 10-year performance numbers look.

 

Kevin Murphy is a fund manager in Schroders’ UK value team and writes at The Value Perspective. The views expressed here are his own and should not be viewed as investment advice. 

Editor's Picks

Loading...

Videos from BNY Mellon Investment Management

Loading...

Data provided by FE fundinfo. Care has been taken to ensure that the information is correct, but FE fundinfo neither warrants, represents nor guarantees the contents of information, nor does it accept any responsibility for errors, inaccuracies, omissions or any inconsistencies herein. Past performance does not predict future performance, it should not be the main or sole reason for making an investment decision. The value of investments and any income from them can fall as well as rise.