The Psy-Fi Blog: Mental Models, Arrayed on Charlie Munger’s Latticework

Old Ideas, New Context

One of the origins of this e-rag was the idea that investors might be able to learn something from the vast array of information in the world not commonly presented to them in a useful form. Much of this information isn’t about investing, although it can often be usefully related to it.

This idea, of course, isn't new. It’s nothing more than a slightly updated take on an idea which will never get old: Charlie Munger’s concept of arraying knowledge on a latticework of mental models. However, it’s one thing to produce individual ideas, as we do here, it’s quite another to put them all together and make them usable.

The Man With a Hammer

Munger’s idea is that learning stuff is relatively easy, it's finding a way of using it that's hard. To be able to do this effectively you need some mental model that allows you to structure and place the information, a way of arraying the information in a usable form. But the idea goes beyond this, because to avoid what he calls “man with a hammer syndrome” you need multiple models and these too have to fit within an overall structure – a latticework – that allows you to make sense of them.

Probably the best decription of this was given by Munger himself in this speech on the application of models to investment, a piece of essential reading for intelligent investors if there ever was one:

"Well, the first rule is that you've got to have multiple models—because if you just have one or two that you're using, the nature of human psychology is such that you'll torture reality so that it fits your models, or at least you'll think it does."

Underlying this is the idea that learning is a lifetime’s activity, one you need to constantly engage in, to develop these models and modify them in the light of experience. In many domains people’s knowledge ossifies as they get older: it’s often said that science advances one death at a time, because the old guard has to die before new ideas can come to the fore. As individual investors that's not a luxury we can afford.

Physics Models

So, to start, models of physics are important for investors, because a lot of the way markets behave can be related back to physical systems, which ultimately underpin everything. In particular the thermodynamic concept of equilibrium is central to a lot of economic models that predict that markets are efficient. Of course, we more or less know that that’s a busted flush as regards the way markets actually work – but a lot of what you’ll hear about markets from traditional economists is based around this theory, and learning to recognise this will help you distinguish the faint signal from the overwhelming noise.

Physics is relevant in other ways, as well, in particular in the development of the sets of systems involved in positive feedback. Positive feedback is exactly what happens in markets when they adapt to human behaviour changing – becoming more risk adverse, say – which then feeds back into the psychology of the people. Markets are adaptive, people are reflexive and they spin around in a dance that never ends.

Evolutionary Models

In fact, in investing we’re not dealing with a system of universal constants (which physicists desperately hope govern the cosmos) but with a system characterised by constant change. Life and markets are characterised by adaptation and modification: they are, in the parlance, complex adaptive systems. And this is a major model that people need to get their heads around: the model of evolution, in which changes happens, often randomly, some which accidentally happen to work given the current state of the world.

So people and businesses are constantly experimenting with what works. And sometimes, often by accident, they find it. Appreciating that there’s a lot of randomness in the system we call the stockmarket is an important piece of the model, because otherwise we’ll spend too much time overanalysing. Sweating the big stuff is more important.

Psychological Models, Statistical Models

Biological models are important in other ways, because it’s possible to map some biological behaviour onto psychology and, as we’re increasingly finding out, many of the odder happenings in markets and by market participants, are based in the slightly off-centre way we estimate our way around the world. These approximations are then labelled irrational by economists when really they're nothing more than our limited brainpower making the best sense of the world that it can, and generally doing a pretty good job of it.

Most of the time this doesn’t matter, but when we’re exposed to the pressures of the markets it can cause all sorts of anomalous behaviours, which we call behavioral biases. In particular we’re reactive because we change our behaviour in response to the market's latest manic-depressive state. Munger is particularly good on this in his speech on the Psychology of Human Misjudgement.

These problems seem to be partially based in our limited capabilities to process statistical information. The underlying issue is that we evolved to handle observed frequencies – judging risk on the basis of how often we observed people getting stomped on by woolly mammoths, rather than by looking at the numbers about how often this actually happened (which, of course, we didn't have access to prior to the modern obsession with counting everything) and figuring out what the real probabilities of death at the feet of a pachyderm with a terrible haircut really were. So an understanding of the basic statistics of mean reversion and the types of mathematics that Cardano, Pascal and Bernoulli developed, are also models we need to have a handle on.

Mathematical Models, Historical Models

Mathematics, of course, is at the heart of a lot of business in the form of accounting. It’s not always obvious that accountancy is another form of mental model but, of course, that’s exactly what it is. To invest without a reasonable understanding of accounting is a dangerous thing, and to invest without an understanding of the horrors that accounting can hide is a suicidal one: for example, appreciating the difference between cashflow and accruals is a pretty important step for investors looking at earnings ratios.

The history of all of these developments is also a critical part of the raft of mental models that investors need to develop. History is frequently used to predict the future, to prophesise doom or boom, but is usually a poor basis for any kind of detailed futurology. In particular investors in the USA and the UK are unusual in having an unbroken history of stockmarket returns, and they’re typically biased as a result towards a belief in higher returns than are likely, on average.

Economic Models

All of these ideas are, in some form or another, addressed by the dismal science of economics. In fact economics has borrowed heavily and frequently from these other domains, but has tended to do it by stealth, without always making it clear what it’s doing. The problem with this is that when the knowledge in other domains advances economics tends to stagnate. Still, ideas like the Law of Supply and Demand and the Law of One Price need to be part of an investor’s toolkit: all things being equal the invisible hand of the market will depress excess profits, so we’d like to ensure that all things don’t stay equal: moated businesses are best, if you can get 'em at the right price.

Also, appreciating the fact that markets are not the solution to everything is important: a dogmatic belief that markets can solve everything is as useless a position as the belief that government is always bad. They nearly always can and it nearly always is, but nearly is not the same as always. So an understanding of market failures like Akerloff’s Lemons and Hardin’s Tragedy of the Commons need to be added to the dataset; and, once again, Munger has a lot of wordly wisdom to dispense in this speech ruminating on the madness of academic economics.

The Latticework

There are, of course, other models that can be usefully developed, but one of the things we’re trying to do here is to show how these ideas relate together. The physics and the biology, the psychology and the statistics, the history, the economics and the accountancy are not totally separate domains of knowledge, they’re all part of an interlinked mesh: the latticework.

Without this mesh of models we're just men with hammers, hitting things in the hope they might be nails. The alternative, to develop what Charlie Munger calls "wordly wisdom" is perhaps a daunting task, at least it doesn't have to be a boring one.

You’ll find the outlines of many of these models, and the way they link together, by following the links in the Psy-Fi Blog's Latticework.