tirsdag 29. juni 2010

Den virkelige verden

Dette er et innlegg publisert i Kapital nr. 11 2010 som svar på et innlegg i nr. 9 av Petter Berge i North Capital 


I en gjestekommentar i Kapital nr. 9 i år skriver Petter Berge i North Capital at passiv forvaltning ikke er et reelt alternativ fordi det ikke er mulig å oppnå indeksens avkastning. Berge mener i den forbindelse at eksempler jeg har gjengitt på min blogg på passive fond som slår indeksen og indeksfond uten kostnad er lite representative eller direkte feilaktige. En bør derfor velge aktive forvaltere, slik som North Capital antar jeg. Nå er Berges påstander stort sett feil, noe som på en utmerket måte illustrerer fordelene med passiv indeksforvaltning.

Det jeg er enig med Berge i er at aktive investorer er helt nødvendig for å korrigere priser slik at markedet kan fungere. Det er likevel så mange av dem at omtrent all empirisk forskning viser at de taper i forhold til indeksfond. Dette skyldes én ting – høye kostnader. Før kostnader slår faktisk den gjennomsnittlige aktive forvalter markedet så vidt.

Det er derfor ingen grunn for hverken oljefondet eller andre å bekymre seg for at det blir for få aktive forvaltere. Det er bra å ta etisk ansvar for kinesiske tekstilarbeidere, men finanshusene på Wall Street får greie seg selv som best de kan.

Berge argumenterer så videre for at indeksforvaltning i realiteten er dyrt og vanskelig. Han viser til at Vanguards hovedfond, som jeg har brukt som eksempel, kun gjenspeiler en meget likvid del av det amerikanske markedet. Fondet måles imidlertid mot en indeks som representerer 99,5% av amerikanske børsnoterte selskaper. Berge mener at en ikke effektivt kan indeksforvalte små selskaper, men Vanguard har slike fond og slår indeksen også her. Berge mener Deutsche Banks (DB) kostnadsfrie fond i realiteten koster 2,5% fordi DB tar utbyttet selv, men det har DB bekreftet overfor meg at er feil. Bergs forvirring her skyldes at investor kan velge om utbyttet skal utbetales eller legges til fondet. Berge bagatelliserer dette som to enkeltstående eksempler, men for oljefondet er jo investering i Vanguards fond et reelt alternativ. Oljefondet ville bare utgjøre om lag 25% av Vanguards totale portefølje.

6 kommentarer:

  1. Some of the arguments used by Berge and also it is sad to see by the leadership of Oljefondet are somewhat simplistic and ignore a number of significant points.
    Firstly the infantile suggestion that markets would fail to work and price investments properly if everyone indexed. There will always be a market for instance in company stock due to factors such as shareholders looking to liquify their holdings, share options and their excise, IPO's etc. The idea that everyone would index is illogical.
    The arguments around costs, whether dividends are included or not for me ignore a more critical issue and that is what is an investor trying to achieve. Surely it has to be exposure to the types of securities that will give him of her the best return for a chosen level of risk. If one takes into account such factors as size, value, momentum and liquidity it would seem logical that as disciplined approach to gaining such risk exposure as possible would be the best approach. This would also ensure a degree of portfolio risk precision which may be abdicated if using active management.
    Whilst the majority of the debate so far has focussed on commercially available indices the real world has moved on. It may be that Oljefondet should construct its own indices based around the various risk factors it wishes to gain exposure to and manage the portfolio accordingly - this would be in agreement with the work of Ang Schaefer and Goetzman and would mirror the approach being taken already by some fund manager such as Dimensional Fund Advisors who apply the Fama/French three factor model in the portfolio construction.
    The final nail in the coffin is the complete lack of reliability in the performance of managers with an active mandate. In an article I wrote for FTfm dated 10th May 2010 I highlighted the work of Barras, Scaillet and Wermers in 'False Discoveries in Mutual Fund Performance' (Swiss Finance Institute 2008) which shows and astonishing lack of alpha among mutual fund managers over time. This backs up the work of Fama & French, Carhart, Jensen et al. clearly showing that active management is not worth the risk.
    More intelligent levels of index construction and portfolio management are available and functional even 'i den virkelige verden'

    SvarSlett
  2. I do apologize to Mr. Stott for being not only 'simplistic' but also 'infantile'. My dear wife and children sometimes charge me along similar lines, but they do of course have the advantage of having actually met me. I rather doubt that Messrs Sirnes and Stott really promote anything very efficiently with a starting point that people who do not agree with them must be idiots. High horses can give nasty falls. But I'm sure that their records of excellence are such that this is not a real risk.
    As for the case discussed, I get a little bit confused - but that might just be me. My first confusion is how a market gets efficient without a fairly extensive number of participants analyzing it. 'Shareholders looking to liquify their holdings' do not create efficiency, they merely seek liquidity. If anything, investors looking to sell because of a need for liquidity rather than an equity view, pretty obviously reduce market efficiency. And sorry, IPOs do not work for me either. It might have something to do with my infantility, of course, but I fail to see how it could be possible to complete IPOs of any scale unless the company issuing new equity was well researched. And who would do that if everyone, or even just a considerable number of participants, indexed? There would be no paying interest in independent corporate research without large numbers of active investors, and in the absence of that the only research offered would be from the intermediary performing the IPO - better known as marketing. Unless of course the altruistic friends of Mr Sirnes in Deutsche Bank took up free research in addition to free asset management (and on the Deutsche Bank point, please see my reply to Mr Sirnes in the current issue of Kapital, which shows that there must be some very serious confusion at his end as to the difference between ETFs and index funds. Which in the current case is actually rather funny, given that he's picked a very expensive set of ETFs to promote).
    My second confusion relates to indices that are not 'commercially available', and which somehow should be superior to the commercial ones. I take that to mean indices that are weighted according to some other factors than market cap. It is very difficult for me to understand that their commercial availability or not makes a blind bit of difference, but there must be a set of principles guiding that construction. That set of principles must be in turn be something other than 'the market always knows best'. I quite simply fail to see how that differs from active management. Yes, the 'intelligent index construction' is usually based on mathematical models of some sort rather than some chap with suspenders saying 'I believe that..', ie it is a quant type active management - but it is still active management. 'Enhanced index management' is in my view a complete nonsense - it is a euphemism for ‘hedging the non-bet’, a really silly strategy.
    But as always, the proof of the pudding is in the consumption. Mr Stott refers to the three factor model as a real world example of stuff that works (really, Mr Stott, that is so 90s, surely best practice has passed both 5 and 6 factors by now?). He also points to Dimensional Fund Advisors as a real world example of practitioners of the superior, non-commercial and not-quite-indexation asset management. DFA not only apply the Fama / French model, but pretty much have the founding fathers of the model on board. Surely they then produce outstanding results? Well, actually, no. Their US Core Equity Fund has marginally higher average returns than the peer group, but a markedly higher volatility than the same peers, and it tends to fall out of bed more often than the peers - and thus scores weakly on Sharpe measures and gets a mediocre 3 star by Morningstar. It actually gets beaten silly by some straight index funds - and never mind the best active ones.
    So how was it that this was proven to work in the real world? Or is that a too simplistic and infantile question?

    SvarSlett
  3. I will leave it to Richard Stott to answer for himself, but I would just like to mention that his use of the words “simplistic” and “infantile” appears to me to be aimed at the arguments used and not the person. As for my self I think personal characterization of ones adversaries is counterproductive, which is why I never do it.

    So to your question; as one can read from my response above the market does need some active investors. The point is that there is no sign that we are anywhere near having too few of them. On the contrary, almost all empirical evidence points to the fact that the market is overpopulated with active investors and there is no good reason why that should change in foreseeable future. Your extreme example of no informed investors is just not very realistic.

    Whether we call it enhanced index management or active management is completely irrelevant for me, that is just semantics. It is what the oilfund actually do that is important, and the fund is currently doing futile expensive betting. I call it active vs. passive just to label it something. If the fund chose to change its management strategy to intelligent indexing I would support it fully and they could call it what they wanted for my part. They are however not doing that at present.

    Your statements on the DB index fund is incorrect though, and I will post something on that here shortly.

    SvarSlett
  4. I very much look forward to finding out how and why my statements on the DB Index Trackers are incorrect. I see that you still refer to them as index funds, while they most certainly are derivatives based ETFs. What's more, they belong to a family of index linked structured products that for years have been available as inter alia ETFs, straight swaps and index linked bonds.

    I've made reference to the Eurostox 50 version of the ETF. The prospectus is available on
    http://www.etf.db.com/SE/SE/pdf/SE/prospectus/prospectusdbxtrackers1_2010_03.PDF .

    Page 84 shows the composition of the index followed, and I refer to paragraph 3 from the bottom, which reads:

    'The indices are calculated as either price or total return indices. The difference is based on the different treatment
    of dividend payments on the Index Securities. Price indices only include cash dividends larger than 10% of the
    equity price and special dividends from non-operating income. Other dividend payments are not included. Total
    return indices include all dividend payments. All dividend payments are included in the both price indices and total
    return indices as net dividends, being the declared dividend less the respective country withholding tax. The Index
    is a price index.'

    It is difficult to see much uncertainty there?

    SvarSlett
  5. And just to complete the documentation; DB themselves do not make any attempts to describe their ETFs than anything other than what they are - derivatives.
    Their ETF brochure for the Swedish market is to be found at http://www.etf.db.com/SE/SE/binaer_view.asp?BinaerNr=1175.
    Please see page 5. The second paragraph explains the ETFs can be produced through replication or through derivatives. And then the fun part, I quote:
    Båda metoderna används av de
    ledande europeiska leverantörerna av
    ETFer. db x-trackers ETFer följer sitt
    index genom en syntetisk replikering
    och faller således inom den andra
    typen av ETFer. db x-trackers ETFer
    placerar i minst 90 % överlåtbara
    värdepapper (obligationer eller aktier)
    och vidare i OTC-derivat som inte
    överstiger 10 % av fondens NAV.

    This is precisely the same structure as the index linked bonds you do not like, it is just called a fund. It is perfectly absurd that you suggest that Oljefondet should get their market exposure through prepackaged structures based on OTC derivatives - that is precisely what you and other have called unsuitable for private investors (and you've used stronger words than that - do you want the links?). And now it is suddenly suitable for our largest investor?
    I think you would do your own credibility a lot of good by just admitting that you got this one badly wrong. And quite honestly, nobody could possibly believe that DB really offers free asset management? Yes, they have ETFs on total return indices, but they have management charges. Yes, they have ETFs without management charges, but they are not on total return indices.

    SvarSlett
  6. I do not contest that the ETF is heavily using derivatives. However that the fund uses the wrong index and that this costs the investors 2,5% a year is incorrect.

    SvarSlett