Erik van Dijk's  Instablog

Erik van Dijk
Send Message
Hi! I am Erik L. van Dijk, principal at LMG Emerge. LMG Emerge is an internationally-operating institutional investment consultant with offices in the Netherlands and at Mauritius. Our clients are pension plans and other institutional investors, family offices and HNW individuals. In close... More
My company:
LMG Emerge
My blog:
LMG Emerge - New Economies
My book:
Asset Pricing and Risk in Emerging and Frontier Markets
  • What chess has to do with asset management; a reflection on passive versus active fund strategies 5 comments
    Mar 12, 2011 9:21 AM

    This note is not about a new topic. It is not even something we never wrote about before. But sometimes you have a bright moment, which can then be translated in a more precise representation of your worries when it comes to discussions about 'active' versus 'passive' fund management.


    Finding top chess players really easier than finding top investors? We doubt it.

    Market Efficiency
    Markets are not 100 percent efficient. This implies that there are inefficiencies that could be translated into above-average return/risk ratios. Obviously, since we are talking about markets here: what the smart investor is winning through an above-average result is lost by others. Similar to chess, where there is always 1 point to be distributed per game via either a 1-0, 0.5-0.5 or 0-1 result. The only difference with the financial world is that the market game sometimes awards 1 point per game, sometimes 2, sometimes 3. There are good and bad periods. But the average player always gets approximately have the points.

    For financial markets this translates into the average result being therefore equal to the market return/risk ratio.

    Not every market is as inefficient as others. The larger the information flow about a market and its constituent financial securities, the more liquid its trading and the better the information quality, the less inefficient it is. And also: the more volatile a market, the larger the potential for catastrophe for those that don't understand it. Bad news for them, but great news for their opponents: the grandmasters.

    Therefore: the probability of beating the average is larger in the Emerging Markets than it is in - let's say - US Large Cap Value strategies.

    But in neither case it is impossible to outperform: compare it with different openings in chess. It can be proven that one opening leads far more often to draws than the other, irrespective of the level of play.

    Taking all this into account,  the likelihood of the opposite statement being true - asset managers cannot beat their benchmark structurally - is more outrageous and extreme! Compare it with chess again. There are grandmasters who beat the average club player. There are also club players who are lousier than others. And there are also rating systems to keep track of all of this.

    Grandmaster players and grandmaster selectors

    So basically, when hearing the well-known statements that asset managers cannot outperform and that you should opt for cheaper ETFs (=exchange traded funds) with a passive strategy and not for more expensive (gross) active strategies that try to beat the index, this is only true to the extent that we cannot FIND the grandmaster asset managers. But our not finding the grandmasters doesn't imply that they are not there. Just like the average club player in chess wouldn't be able to really analyze performance qualities in a specific opening when comparing grandmaster X with grandmaster Y.

    In other words our not finding the winners means that we didn't know where to look for them and how to find them. It is first and foremost a statement about our search qualities and not about grandmasters being there or not.

    After the crisis the call for passive management has grown. Normal! After a period of disappointment one tends to stop believing in excellent result. Just like the little chess prodigy comes back home crying, after losing against that shrewd, older and more experienced top player. This notwithstanding the fact that overall he is far more talented than the old guy! The market is not an easy opponent, and there can be periods in which we have to sit still and wait for our revenge. Again: no big difference with chess.

    Now, isn't it true that those who want us to go passive, are probably not active management specialists themselves. Otherwise they wouldn't come up with that suggestion, knowing that fees can be higher in active management. So they want us to follow their passive - index - strategy assuming that those active managers who are good this year will underperform next year and also end up with their same gross average return, albeit with higher costs.

    Makes sense? Not really: the more people will go for just the average (i.e. buy the index), the easier it will be to beat the average for those that do have above-average knowledge! And/or those that continue to do research and create new strategies that incorporate the fact that growing groups of investors don't even look for outperformance anymore.


     1) Passive instead of active is the wrong discussion. Go passive when you don't know how to beat the benchmark. I.e. when you can't find the grandmaster. Just like chess grandmaster would opt for a draw when they don't really like their position and don't know what to do.

     2) When markets are inefficient, the larger the inefficiency, the larger your allocation to a potential grandmaster - or a few of them if you want to avoid putting all your eggs in one basket - has to be. And with Emerging Markets becoming more important in the world, the trend is definitely not one toward a growing necessity for a larger share of 'passive' in your overall, global portfolio.

    3) Analyze things on a net basis. If costs are the issue, then why not analyze all strategies after deducting management fees and trading costs from the gross results?

    4) Having no success so far with active vis-a-vis passive strategies, basically doesn't change the conclusion that real grandmasters exist. It just leads to the conclusion that your manager selection adviser was no grandmaster in his field.

    5) The experience in the chess world indicates that it is possible to define grandmasters. The financial world needs a better rating system.

    6) In the end you will end up with a mix of passive and active strategies in a well-diversified good investment portfolio

    7) The manager selection process is an ongoing one. It is not just about selecting the winners, but also about kicking them out when they stop performing. Just like in chess: the world champions and top grandmasters of 10-20 years ago are not necessarily the best kids-on-the-block today. You have to continue doing the necessary research.

    8) Expected rewards? Top managers can outperform by 3-5 percent on a net basis. But OK, if we assume that you can't pick the winners all the time and once in a while will be simply unlucky, then it is still likely that - if-and-only-if you were fair to yourself and critical in your selection process - you can achieve a net 1-3 percent outperformance on selected active managers.

    9) But as we already stated, there will at any point in time be asset classes where you simply don't know how to find the grandmasters. Assuming that you have to go passive in those asset classes, the overall net expected outperformance can be about 1 percent in case you assume a 50-50 active-passive mix.

    10) Not important? If your passive mixed portfolio would return 6 percent per annum, this would transform USD 10,000 into USD 17,908 after 10 years. Earning 7 percent instead would lead to an end result of USD 19,672. A nice difference, especially when this 1 percent becomes relatively more important the lower the expected returns are (e.g. like for instance now when interest rates are very low).

    11) Still not convinced? Then do also not forget that if the 'go passive' group grows, our initial calculation of 3-5 percent (see point 8) will show a tendency to go upward as a result of which you will be left with more than 1 percent outperformance in step 10.

    12) Anything known about manager types when it comes to picking grandmasters? Yes, one thing. Size matters, but it is definitely not true that the bigger asset managers are the best ones.

    In other words: keep checking out those ETFs by Vanguard and others, but never ever stop looking for gemstone managers in your Morningstar or other manager databases. And also: don't overestimate yourself. The grandmaster managers are there, but are you also a grandmaster manager selector!


    Disclosure: I have no positions in any stocks mentioned, and no plans to initiate any positions within the next 72 hours.

    Additional disclosure: The author is an adviser to pension plans in the Netherlands. During the last 10 years he has worked together with Nobel Prize Laureate dr Markowitz on a Modern approach to Modern Portfolio Theory. These reflection are directly embedded in this 'Modern' Modern investment approach.
Back To Erik van Dijk's Instablog HomePage »

Instablogs are blogs which are instantly set up and networked within the Seeking Alpha community. Instablog posts are not selected, edited or screened by Seeking Alpha editors, in contrast to contributors' articles.

Comments (5)
Track new comments
  • Amerlafrance
    , contributor
    Comments (794) | Send Message
    The problem is that chess rankings are developed over finite time frames with multiple results over a fairly short period of time. The investing grandmasters have to be evaluated over a loner period (except day traders), but what period do you pick - 1 year, 5 year a market cycle. Any of these preclude being able to evaluate a statistically significant sample to weed out luck versus the grandmaster. This is an intersting conundrum, I prefer the hubris of thinking I can outperform the index by using common sense and study of what I can understand. I won't beat the grandmasters, but will hopefully beat those who blindly push around the pieces.
    21 Mar 2011, 09:19 PM Reply Like
  • Erik van Dijk
    , contributor
    Comments (20) | Send Message
    Author’s reply » Hi thanks for your feedback.


    We have looked into that and I agree partly. Any top level chess specialist can tell you that grandmaster X or Y is less good or better in opening X or Y or middle game position A, B or C.


    However, most manager selection specialists don't get much further than looking at past performance and one or a few risk indicators. Digging deeper into portfolio compositions in comparisons with the firm's philosophy etc is always done on paper but never really practiced.


    Being - personally - quite a bit involved with top chess and - as a firm - with manager selection, we believe that the RELATIVE level of of analysis in chess is better than in manager selection. So it is to a large extent a problem of the selectors. Maybe just as much as of the asset managers.


    But I agree with your points that time frames do play an important role. We basically solve that by looking at rolling windows of three time frames (short-term - up until 12 months rolling window); middle-long term (1-5 years) and long-term (more than 5 years).


    When a manager's strategy is relatively new (i.e. not 5 or even 3 year track record) we will look to indirect measures (who are in the team? where did they work? how was the track record there?) or simple decide to wait a bit. It works quite well.


    We will co-launch a fund incorporating our methodology later this year.


    Thanks again for your very deep remark.


    Kind regards,
    Erik van Dijk
    LMG Emerge
    10 Apr 2011, 08:44 AM Reply Like
  • Amerlafrance
    , contributor
    Comments (794) | Send Message
    Very true on the ability to play different openings, black vs. white, closed vs. open game etc. With the complexity of understanding the business of energy vs. chip design vs. banking, perhaps there is a need for a team approach that includes the business as well as overall market analysis. A team with the proper composition and defined process should be able to be evaluated over time frames sufficient to make a rating. But the problem is still measuring that kind of performance against a 5 year history in the press which we know some percentage of people will be at the top of by dumb luck. Your point about comparing composition to philosophy and then weighting that comparison to performance is the only way to weed out the lucky and focus on the repeatable perfromance from the philosophy. If the philosophy is well dcoumented, then the ability to fill vacancies and maintain perfromance is also enhanced.


    Very interesting inquiry, I am still interested in the time frames you use to measure performance. If you measure short time frames, don't you have to take into account starting and ending conditions and not just look at performance? For instance, a short term trader may be good when the fed is easing (playing white) but terrible in a fed tightening environment (playing balck). If I was a team captain, I would certainly pick a different grandmaster for a final tounrament game depending upon color. Also, if you use rolling periods, does the time component remove the ability to properly weight starting conditions?


    PS, I am a golden knight in postal chess.
    10 Apr 2011, 01:48 PM Reply Like
  • Amerlafrance
    , contributor
    Comments (794) | Send Message
    As I thought about this further, I am thinking that perhaps the time frame for performance measurment is not the key issue. Any time frame chosen will be arbitrary (there is not a checkmate or draw to annouce the completion). To measure against market cycles just begs the question of which cycle, (US vs Worldwide, inflation vs. deflation, dollar reserve vs. pound reserve etc.) and assumes that we know the start and finish of a cycle in time to make an investment decision. Your view of measuring adherence to a the investment philosophy is the only way to measure a manager; with many back tests of the investment philosophy to create the data points necessary to evaluate the philosophy (rather than the manager). That philosophy can be measured against different real world siutautions to determine which philosophy to pursue and then it is up to the advisor to choose a manager with a high correlation of adherence. Ultimately, the advisor's job in picking the manager is to be alert to divergence from the philospphy (investment policy creep) and to evalaute where in the cycle we are for a given client's goals and risk tolerance to choose a philosophy to guide their investments.


    Of course, then we have to move from evaluating the performance of the investment manager to the advisor (from the grandmaster to the team captain). I think I was confusing the investment manager who makes the ground level decisions of what to buy/sell and the investment advisor that chooses the make up of the team of managers and who will play at any given time (you were more clear about this than I was, it took me time to see the distincition, one of the problems of fulfilling both roles myself). Of course, in evaluating the advisor (team captain), the won/loss record is ultimately what determines if you keep your job and there absolute returns will matter more than relative performance.
    11 Apr 2011, 12:55 PM Reply Like
  • Erik van Dijk
    , contributor
    Comments (20) | Send Message
    Author’s reply » Good to see that you are also a practitioner. Also played postal chess (at the ICCF) in the past. But business and management roles for the Chess Federation (Netherlands and FIDE) replaced my own activist stand in this respect.


    But good to see that we both know quite well what chess is about. And it is clear from your deep answers that you do.


    Hope that you can accept that we cannot dig too deep here with respect to proprietary firm knowledge, but basically what we also do is run regressions in which we analyze the alpha of managers against sets of fundamental variables from our asset allocation models and databases so as to see to what extent their excess returns are sensitive to specific aspects of asset pricing.


    We also repeat this exercise with a set of country-asset class indices. We added the latter 'loop' to the process when we found out that some regional and global managers basically scored their alpha through structural overweights in one or the other country. Often not at all the result of well-funded investment decisions but more related to coincidental factors.


    Example: some MENA managers do not include Turkey in their MENA definition whereas others do. And differential strategic weights are another issue. By linking alpha to asset-class / country indices as well we can pick up the most remarkable and deviant sensitivities.


    But you are right: in the end you are of course still lift with uncertainty. But never so big as to believe that when faced the decision to invest in Ashmore, Dimension or Acadian (3 GMs in Global Emerging Market Equities) OR a local Dutch big bank trying to present a nice Global Emerging Market Equities product, the former isn't better. Of course, in individual months our Dutch friends could outperform.


    But just like you and me - based on our chess strengths - might be so lucky as to score one or two successes against a top GM in our favorite opening, our problem is that when we play more over a longer period of time, our good luck will fade away.


    Being an insider in investing, there is unfortunately one huge factor blurring the picture. One that I am faced with time and again:




    Very often clients demand manager selection decisions that are not solely based on performance. Performance is of course an important factor, but there is a lot more.


    'Costs' is another one. With the 'cost' being 'certain' and the performance 'uncertain' some clients do tend to have an extreme cost focus. Another one is 'domicile'. Sometimes institutional clients request that you stick to local parties or at best Anglo-Saxon parties that they are used to. Even when a specific local managers from let's say Brazil was better, they might still want to stick with their good friends from Dutch Bank X with whom they play nice games of golf.


    Of course, there is a 'price' for all those non-performance related factors and when the difference is too big between performances of managers, they won't help anymore.


    And you can also say that when tournament organizers have to set up fantastic chess tournaments those factors will play a role as well, but believe me: on average, the non-performance related factors in investments are of larger importance than those in chess.


    I am optimistic about the future. There is a trend, forced upon us by growing problems of pension plans (returns are more difficult to generate in a world of relatively low interest rates), compliance, transparency and unfavorable demographics. That will give better asset managers more chances vis-a-vis their peers and it will also as a result of that trigger the beginning of an increased professionalization of performance measurement in investments.


    Chess has their Elo Rating, but Investments have many rating systems none of which is very good when it comes to forward looking performance. That is why we developed our own approach in cooperation with a large UK-based institutional manager selection database and an innovative manager selector from the same country (Camradata and BFinance respectively).
    11 Apr 2011, 02:28 PM Reply Like
Full index of posts »
Latest Followers
Instablogs are Seeking Alpha's free blogging platform customized for finance, with instant set up and exposure to millions of readers interested in the financial markets. Publish your own instablog in minutes.