Monthly Archives: June 2012

I’ve become bored of index trackers.

Trackers are still good for you

It’s now a well known thing that active managers as a group have a hard time beating the index, which makes sense given that to a large extent, the index is the average of active managers and their ilk. For active managers to beat the index, there would need to be a category of investors that systematically does less well. It could be argued that retail mutual funds are not a majority on the market, but the people who manage non-retail funds professionally basically come from the same pool as those who do retail and have no reason to perform differently. Direct retail investors could be that category, but they’re small (single digit percentage of the market at most I would guess) and they’re collectively random. Retail investors underperform via crap timing and transaction cost maximisation through excess churn, not by having consistently negative insight as stock pickers. These errors they can make with index trackers as well, which do not account for bad timing or churn.

So in a perfect world the return of active managers should be the same as the trackers less the excess management fee. Even if they turn out a bit better than the other categories, it is improbable that they could earn the typical 1% annual excess fee in the (compounded) long term.

There has been arguments that the success of index trackers will diminish their efficiency, which is at the limit is true (if 99.9% of the market was trackers, we’d have a problem) but with pricing at the margin and mainstream cap weighted trackers being mostly passive observers of prices — they do not trade in steady state — I doubt this effect could come to dominate. There will always be a sizeable contingent of market participants who think they can add value, regardless of actual ability, in addition to those who actively try to exploit and thus cancel any emerging “index effect”.

So basically index trackers are still probably the best choice for a non professional long term investor who wants exposure to the stock market.

An oblique approach to investing: the index less crap

Yet, I’m bored of them. Every time I go through a tracker’s holdings list, I start thinking, of about a third to a half of the names i recognise, “this is a shite business I don’t want to be invested in”. I have a suspicion that mediocre businesses do, on average and in the long term, less well than good businesses. I have no formal proof for that, beyond anecdotal evidence that famous buy-and-hold long term investors tend to be successful via investing in non-crap businesses, and that businesses that tend to last decades tend to be those that offer some valuable service or product to their customers.

John Kay has been arguing that complex objectives are best attained via indirect means. The idea of the Obliquity Portfolios is to start with the simple concept that shite companies probably underperform, and that it only takes a modicum of common sense to identify a bad company. Basically let’s take an index and strip it from “obviously” crap companies. This should be much easier than picking the very best positively. In addition, it is a narrow objective, that, critically, doesn’t require a pricing model — we just assume that crap companies are on average always a bit underpriced.

What if there was no mispricing? If investors priced in the mediocrity of a business accurately, and that this mediocrity indeed had adverse economic impact on the business itself, they would just be cheaper relative to their good peers, reflecting poorer prospects and accomplishing an efficient market, where basically every instrument returns the same in risk adjusted terms. Why should crap companies be mispriced? I think there’s a chance that similar biases to those that produce the low volatility anomaly are in effect when it comes to the valuation of crap businesses.

In any case, even if crap companies are not mispriced, it is still hard to underperform with a portfolio of good companies. Given broad enough diversification, any sampling technique based on a meaningless criteria will still get pretty close to the index. Filter out companies with an odd number of letters in their company name out of a large index, and you will end up with a random half on the index that will perform very close to the full index: the sample just needs to be big and unbiased enough to approximate the full set. So for a well diversified “good company” strategy to fail, an opposite mispricing would be required: persistent and sizeable overvaluation of good companies.

Large caps as sectoral investment trusts

Another reason to go stock picking is that it’s not in essence that different from asset picking. Investors in index trackers still usually want to mitigate methodology, asset class and tracking implementation risk, so are still left with making decisions on a portfolio of asset classes and indices. There are hundreds of indices to choose from. Nowadays large cap companies are largely sectoral indices, often widely diversified geographically and in business lines within a sector. IBM is like a closed ended fund specialising in IT, BHP Billiton similarly so for mining, etc. In building a full portfolio why not go direct to these and avoid an intermediate layer, with extra fees and implementation risks, however minimal in the case of trackers?

So the portfolio will have a large cap bias, though likely little different than the natural one in whole market cap weighted trackers, for the benefits of in-sector diversification a large company provides. Small business portfolios can only be done either with an edge in a particular industry, or as large samples to recreate diversification statistically, both of which are not realistic for what this is trying to achieve.

Check lists

The actual criteria used to identify a “good business” is basically common sense, but it can still be decomposed in a handful of sub-criteria, the details of which I will expand on in a future post. This seems a perfect fit for Atul Gawande’s checklist idea. This has already proven valuable in the preliminary work, as repetitive churning through lots of potential constituents makes it very easy to forget the boring details. Some companies I would have intuitively let through for passing my basic test failed a more thorough checking of the subcriteria, even for a fairly simple criteria set that any moderately competent person could assess, from public data, in an hour or two for a given stock.

Low maintenance

As this portfolio will be invested with real money, I have to take care about transaction costs, and also with being long term maintainable without requiring constant nursing. I don’t have any plans to spend most of my time on this. As it is biased towards “good” and large cap businesses with long term prospects, a classic passive “never sell” portfolio should do. A sale should happen only because a business has changed enough to fail the inclusion criteria, or has outperformed so much as to be overweight in the portfolio in which case it needs to be cut back to size. An annual review of each existing holding should be more than enough.

Open source

It’s quite common practice for the failed investor to realise that, rather than beating the market, it is actually easier to profit from the naivety of the next generation of hopeful investors and become part of the finance industry; by becoming a broker of some description, or getting commissions by channelling the naive to some dubious service or other, or selling subscriptions to a tip sheet packed with one’s random investment ideas. There are platforms to publish one’s portfolio and get followers to pay for the benefit of tracking it. Doing this is a clear sign of failure. Those who can’t invest talk about investment, and charge for it.

In so far as I keep running personal portfolios in the Obliquity style, I will publish the holdings and weights for free, here or on any convenient open platform. The finance industry could do with more open source spirit. It wouldn’t take many vaguely enlightened amateurs — I wish others more enlightened than me did it — to make fee-based active management a vintage niche product.

Disclosure: the Obliquity Portfolio is in design phase and thus 100% in cash.


In “The End of Finance, As We Know It“, the Psy-Fi blogger argues that a common language for financial products will disrupt the industry.

Tim Berners-Lee was a lucky guy

First, on a point of detail, I think the HTML-made-the-web meme is well overdone. It’s very probable that the web and the Internet were doomed to happen once the enabling technologies — hypertext, client computers powerful enough for graphical user interfaces, networking fast and cheap enough for networked hypertext for the median computer user — were available. Indeed, it happened almost as soon as these technologies had developed to be close enough to be practical. Then sooner or later an implementation of the networked hypertext paradigm that hit a sweet point between usability and technical practicality was going to appear, and then catch on and through network effects and become dominant. Some of the qualities of the HTML solution, like openness, are probably only very marginally relevant. It’s something that techies appreciate, but most of the world don’t care about such things. Video happened on the web via Flash, which is the antithesis of HTML: a closed proprietary technology, which in the days that mattered was tightly controlled by a rent-seeking commercial operator.

The focus on HTML as a semantic language also seems overdone, at it’s core it’s just an hypertext protocol, that plays a similar role as over technologies in the stack, such as operating system APIs that support browser software, or the encodings used at the fiber level on operator level data-links. Had it been named a hypertext “protocol” instead of a “language”, which would probably be more accurate description of what it is in essence, there would probably be very little talk about the web being enabled by a common language in a semantically relevant way. Contemporary HTML, known as HTML5, has evolved more towards a platform, which includes Javascript, a programming language, but it’s a technical artefact that doesn’t really change the argument that it’s not really a “language” in the common sense of the word that made the web possible.

Product standardisation without disruption

The main point behind the post, HTML irrelevance’s notwithstanding, remains though valid: having machine-tractable product descriptions enable more efficient markets in the underlying products in the presence of the Internet. The argument is then that the development of standardised descriptions of financial products will revolutionise finance, and displace the incumbent operators.

This might be true for some narrow niches, e.g. treasurers of midsize companies looking for OTC derivatives to hedge their real world business would benefit if they could compare rival investement bank’s proposals from computer-processable term sheets sent to a comparison service or software. There are already advanced initiatives in that field. It may be true of other niches as well.

But what about retail: in the most advanced markets the core useful products of retail finance: current and savings account, electronic payment methods, mortgages, index funds, all are pretty much standardised, and can often be approximately but close enough summarised in more or less one number (APR, interest rates, TER, etc). Some of this has been driven by regulation, and to a large extend it is one domain where it has been successful; even too much perhaps, in the sense that there’s little space for non-standard products due to the difficulty of making them fit in the regulated standardised framework. This is all common “language” that’s needed.

More exotic products may be less standardised but are are of dubious utility to most people, and for those people who misuse them the solution is not to standardise them but to stop using them. In a similar vain, making a peer review platform for churches or religions, however thoroughly, won’t make God exist, and there’s no interesting benefits to be gained in making efficient markets in fundamentally pointless things.

So we really do have readily comparable financial products, and indeed comparison platforms for all of them are widely available. You can use the Internet to find more or less painlessly the best mortgage or savings account. There’s some rough edges, some tangential contract conditions may escape the normalisation, but overall it’s good enough for purpose, and this has been the case for a number of years. So we have the precondition for disruption, and yet nearly everyone is still getting loans, mortgages and current accounts, from the same old banks. Why is this so?

What doesn’t kill you make you stronger (sometimes)

A possibly salient angle may be to look how the Internet disrupted other industries. There are definitely several industries, from recorded music to travel agents, where incumbents have been marginalised by the internet, being either made irrelevant or replaced by new Internet-era companies taking over their traditional business.

But there are yet other industries where incumbents have successfully taken on the Internet, and improved or re-centred their product successfully. Telecommunication operators are a prime example. When home Internet access became possible, all Internet service providers were startups. The incumbent telcos were into providing voice telephone, and data infrastructure only to businesses, including the ISP startups, and were totally uninterested in doing something else. The ISP scene bloomed, some better operators emerged and became market leaders. Nevertheless, in a short few years the incumbent telco operators got their act together, used their size advantage and the fact that they were anyway operating the underlying infrastructure, and killed, through acquisition or attrition, the startup ISP scene, which are now as niche and retro as vinyl records.

You can observe a similar trend in airlines: there are no “Internet airlines” to speak of, they just all adapted. The last wave of business model innovation in the airline industry was low-cost, which started in the telephone sales era — remember planes with phone numbers on them? — and just adapted successfully to moving onto doing the same on websites.

Any hope?

Whatever the reason that made banking be more like telcos than music labels, so far, what I think is pretty likely is that if the industry does in the end get disrupted beyond recognition, it will not be through (retail) product standardisation. It would probably be unwise to care too much about something that is essentially, unusually for finance, a solved problem.

Here is the the Synergy plane project, via designboom. There is a Kickstarter project.

Rendering of the synergy plane

It sure looks good, and interestingly different, albeit a bit reminiscent of the commercially failed Beechcraft Starship and of what must be the nicest private plane in production today, the forever young, despite being a design more than 30 years old, Piaggio Avanti.

I am not competent to judge whether the funky shape also delivers the claimed aerodynamic advantages, but let’s suppose that form follows function and the promoters know what they’re doing on this.

What I find interesting is the economic aspects:

Designed to show ten times the fuel economy of a small jet at ten percent of the cost,

First I do doubt the claim that a mere funky shape of wing is going to change the physics so much as to change fuel economy by an order of magnitude, notwithstanding the unfair comparison with a jet — it’s a slower propeller plane.

The idea that it could also cost ten times less does not seem credible. The shape might help build a composite unibody, but composite material are not probably not cheaper to manufacture than aluminium bodies, or everybody would be doing it, and surely the price of the structure is a small component of the cost of a full aircraft, which requires the motorisation, cockpit equipment, and costly certification process so that they don’t fall out of the sky on innocent civilians, and a backing infrastructure for maintenance — why they had to stop flying Concorde.

In a less drastic way, this seems a frequent feature of even more pedestrian, pardon the pun, Kickstarter projects which are priced from DIY supplies, assuming quasi-free work by the project initiators and their mates for manufacturing, and without the productisation overhead of running a making a fully validated and regulation-compliant product, backed by a support organisation handing repairs and maintenance and allowing for a sale and marketing structure, which is never free even with internet-based methods.

Another classic naive startup sin is to try to fix too many problems at the same time:

Far greater economy, using better fuels.

OK, we’ve got some unproven wing and aerodynamic technology to sell, and let’s bundle it with some other new unproven technologies, because they sound like magic. Get your thing going with state of the art, but no better, versions of all the supporting technologies; and let, in this case, motor and fuel people innovate at their own pace. Whatever they find can easily be adopted by such a project, that doesn’t need to have multiple points of innovative failure.

As a product it may also suffer from being too low end for the bottom of the jet market (5 people, no moving space, what if we need a pee?) and too unrealistic for the really low end, which seems to cater to the old dream of the 1970s, traffic jams in the sky:

In this second century of flight, we believe that ordinary families should have fast options to travel where they want, when they want, in quiet safety, with better economy than a car. Without the exhausting airport hassle.

Yeah, right. So give all moderately wealthy dwellers of a medium size city that sort of aircraft and you don’t get air traffic control issues and jammed airports? The technology for “family airplanes” has been there for a few decades. It’s just vastly unrealistic, in any but the most remote areas, to have thousands of planes flying simultaneously, let alone the nuisance for people on the ground. Even if the technology was quasi free you would either get rationing via regulation, or else severe congestion.

It’s also something that’s way past the size something a Kickstarter campaign could sustain, though the promoters seem to be aware of that. In this case, it seems more like a charity fundraising. This project will have a chance to come reality, of one shape or other, only if they at some point associate themselves with an incumbent airplane manufacturer with the know-how to do all the work that is beyond the aerodynamics and has realistic expectations on what the market can take.

But it does look good, and $5 or $25 for a bit of dream may be fair value for money. I could even be tempted if they offered a (3D printed?) paper model, deliverable now.