“Science tells us what we can know, but what we can know is little, and if we forget how much we cannot know we become insensitive to many things of very great importance.” 
Bertrand Russell

When IPE was founded in 1997, the mainstream investment approach was based on the Modern Portfolio Theory (MPT). It was founded on the belief in rational, risk-averse ‘agents’, and a general equilibrium in markets that is sometimes temporarily dis-equilibrated by external shocks. This resulted in a strong belief in statistical parameters such as volatility and correlation as measures of risk.  

The merit of this approach to its users was that you could apply statistical and mathematical protocols to construct portfolios, which was convenient. In the 1990s, value at risk (VAR), closely related to MPT, was introduced and became popular as measure for portfolio risk. It was also adopted by regulators. Everyone was working with the same ideology, the same belief in how markets worked and how risk could be quantified. Everyone used the same models. 

It seemed to work well – at least on the surface. Despite serious crises in the 1980s and 1990s – both in western and emerging markets – MPT became more popular. However, it became abundantly clear to this author and other practitioners that although the approach was working 90% of the time, the other 10% involved periods of turbulence when the theory was so flawed and the consequences so severe, that its predictions of maximum harm could be off by a factor of 30 or more.

In the 1990s, it often happened during crises, such as in emerging markets, that parameters such as volatility and correlation could drop to very low levels and provide a feeling of low risk when true risk was actually rising and instability increasing. But the low-risk perception, the result of the low volatility, invited more risk, increasing the actual real risk that remained unnoticed in the MPT risk quantifiers. The year 2007, with the lowest volatility in decades, just before the markets crashed, is one of many examples. 

Despite its unrealistic assumptions and repeated failures, MPT remained the dominant approach into this century. One of the key reasons is a behavioural bias called ambiguity aversion. People want to use mathematical models as explanations for how the world works, even if the model does not work perfectly (and even more uncomfortably, if it fails totally). As Nietzsche said: “Any explanation is better than none.” Seeking causality even when it is absent is one of the drivers of men.  

Paradigm at risk

In the last two decades, many academics and practitioners have become frustrated by the failures of MPT, efficient market hypotheses and all related equilibrium-theories. Inspiration for new approaches has come among others from advances in behavioural economics (Thaler, Shiller, Slovic, Tversky, Kahneman), ideas around fundamental uncertainty (going back to Knight, Keynes and Hayek) supported by new ideas around chaos (Mandelbrot) and complexity theory (Shelling, Holland, Farmer).

“People want to use mathematical models as explanations for how the world works, even if the model does not work perfectly”

Behavioural finance made it clear people can behave irrationally, and even collectively irrationally, especially when operating in the abstract area of investing and dealing with financial assets. This can create reflexive feedback loops, leading to euphoria and unstable markets. 

Complexity theory – which has its foundations in the Santa Fe institute in the mid-1980s – models the way people interact with each other and react to market developments to see how this micro behaviour leads to emergent (macro) consequences. It investigates bottom-up emergence of collective behaviour, so called endogenous processes, quite different from the neo-classical worldview of exogenous shocks. Often these endogenous consequences are in the form of instability and dangerous tipping points. Hyman Minsky’s Financial Instability

Hypothesis describes such a process, but is just one of society’s many possible processes. 

Complexity theory explains that we cannot predict the timing of any tipping point. The modelling of endogenous, often chaotic processes reveals the bottom-up behaviour of people and all reflexive processes between these people can never lead to reliable predictions. It can, at best, lead to some understanding of how the system works in general, not specifically parameterised, as Friedrich Hayek explained in his Nobel speech in 1974. A very inconvenient truth at first glance. In a world filled with billions of not-so-rational people – especially when it comes to making decisions in complex financial markets – the end result is that we need to steer between order, chaos and randomness. We need to accept that the world switches between these states. What does this imply for investing?

Being confronted with fundamental uncertainty does not mean we can not deal with it. Yes, we have to accept we can not ‘control’ uncertainty completely, as we used to think for the past half century under strongly refuted assumptions. But we are better able to steer our portfolio if we accept the fundamental notions of uncertainty and complexity. According to Stephen Hawking, complexity is the science of the 21st century, and it is time investors start to adopt its insights.

The following considerations are a recipe of how to deal with complexity as an investor –including some recommendations how not to deal with it – that could gradually become the new paradigm of investing.

The complexity world view

In a world of investing according to the principles and insights of complexity, thinking might change radically.

Instead of looking at the output of markets – returns, volatility, correlation – investors must concentrate on the underlying processes. The processes and developments investors have to focus on include: private debt accumulations; divergence in financial versus real asset prices; demographics; technological changes; flows of money; tone in the media and so on. These are the drivers of the dynamics of financial market valuations. The approach Hyman Minsky chose is one example: looking at increases in private debt combined with financial asset valuations running out of line with the real economy, and overconfidence as indicated by the tone of the market, may be indicators of instability. Even when a risk measure such as volatility tells us everything is fine.

Investors will search for true diversity, which is not the same as classic, mechanical correlation-driven diversification. Diversification is based on the output in financial markets, such as stocks and bonds in various indices. But these financial securities in the index may all be part of a market that is visited by many index-seeking investors and therefore extremely connected. 

True diversity is the quest for investments that will suffer, or prosper, from different risk drivers in different markets – different regions, liquidity and asset classes, including often neglected long-term investments in firms and infrastructure projects. Diversity does not require a huge amount of stocks in a possibly interlinked market, it requires a limited amount of truly different strategies in less connected markets. Understanding connectivity is key to the complexity game.

Diversity under this new world view will also seek diversity in tools. Investors can no longer rely on one approach or one big conceptual framework such as the equilibrium models. The new slogan is don’t put all your models in one basket – agent-based models, network theory, scenario thinking and so on.

Instead of looking at (unknown) numerical probabilities, investors will increasingly think in terms of consequences to avoid surprises and damage. New thinking in consequences might entail that scenario-thinking becomes an effective investment tool. Scenario-thinking is a process based on fundamental changes in underlying geopolitical, demographic, technological, ecological drivers, translated into a narrative around the change that also explains the financial market implications. It creates imaginative mental models of the future.

During the process of scenario-building, investors will conduct two main functions:

• They are able to anticipate the consequences of these possibilities and take action now that makes them more robust. The aim is to make sure we survive all of them, not just the one that we think is most likely, because we cannot afford to ignore those others knowing the consequences could be lethal. The scenarios can be used to test the diversity of the portfolio and its robustness. A example of such a robust approach is Ray Dalio’s all-weather strategy, preparing the investor for many possible futures – inflation, deflation, rising or slowing economic growth;

• By imagining the path that leads to these scenario outcomes, scenario-thinking makes investors more alert if fundamentals change. It makes us ‘memorise the future’ and adapt based on signals in the scenarios. People who applied this way of thinking in the early part of this century were able to imagine serious declines in interest rates, recognise them as a possible structural trend instead of dismissing them as irrelevant deviations from an assumed mean-reversion pattern, and could react to these risks accordingly when interest rates did trend downward. People who had the Minsky debt-deleverage in mind as a possible scenario were not able to see the 2007-08 crash coming, but they were able to understand the signals in 2007. When credits went down, this looked like a Minsky tipping point and action was taken to reduce equity exposure in a timely manner. In both cases, the difference is that scenario thinkers, by being imaginative, were not better predictors but were able to escape narrow thinking and recognise new trends or instabilities and take actions earlier.

The Bertrand Russell quote at the beginning of this article says  “.…if we forget how much we cannot know, we become insensitive to many things of very great importance”. Scenario-thinking increases our sensitivity to important developments that are often ignored when we use probability models about the output of financial markets. 

In the next few decades investment will hopefully develop more around the concepts of complexity and diversity and rely little on MPT.  

By applying a more humble approach that uses different tools that try to understand the underlying endogenous trends, assess the interconnectivity of institutions, markets and investments, and trying to create mental models of different futures, pension funds can become more robust (acting now based on consequences they cannot afford) and adaptive (recognising signals earlier thanks to mental models of the future). This avoids the funding ratios of pension funds riding the rollercoaster of financial markets, and enables them to create more stability. 

The uncomfortable news is that this is not a simple recipe for success hidden in a few prescribed formulae. The good news is that it will avoid the recipe for disaster built in the current one-dimensional paradigm of investing. 

Theo Kocken is professor of risk management, VU Amsterdam, and co-CEO of Cardano Group