Dike surveillance for portfolios
Let’s start with the basics: why don’t more common deterministic and statistical models suffice when drawing up an adequate risk policy? Both methods are based on excessively simplistic analyses. Therefore they fail to generate a clear overview of the potential for highly negative results. The problem with deterministic models is that they often sketch worst-case scenarios. They show how much a portfolio will decline if certain investments plummet in value, but they say nothing about the risk of this occurring. Statistical models do give an indication of this risk, but focus mainly on ‘typical changes’ via, for example, standard deviation and the use of normal distribution. This works well in 90 to 95 percent of cases, but this is not good enough.
The biggest error in statistical risk models is that the risk of unexpected, extreme events is grossly underestimated. It is precisely in times such as these, that this is perilous. Equity prices were higher than ever this year and the last genuinely large downturn on the markets was a long time ago. At Kempen, we believe effective risk management which also examines the risk of extreme events, or tail risks, is crucial. The acute stress on the markets in early March corroborates this picture.
The greatest storms
It’s perfectly normal to take extreme risks into account outside the world of investment. A good example is the dikes for which the Netherlands is so famous. When constructing these, engineers don’t work on the basis of normal water levels, but on very extreme levels. Our dikes need to be able to withstand these extreme levels, even if they occur very rarely. Only a dike that is built according to that principle, offers sound long-term protection.
Nothing lasts forever
In the same way that engineers take extreme conditions into account, we believe that risk management for investment portfolios also needs to consider extreme scenarios. These extremes can cause existing patterns to look completely different.
For instance, we’re used to equity prices rising when interest rates fall, as central banks have been prepared to act when necessary over the past few years. The question is whether this mechanism will continue to exist in the future. Yet traditional risk models assume that it will. We’ve seen this mechanism function consistently over the past ten or fifteen years, but that doesn’t mean it’s a given. Therefore it makes more sense to incorporate into risk models the fact that the situation can change drastically, and that the elements can collectively turn against you.
Extreme interest rate scenarios
Let’s look at interest rates as an example of an extreme scenario. Given their current levels, almost everyone assumes that central banks will not cut rates further. Yet this is a possibility. After all, ten years ago interest rates already stood at historically low levels of around 4 percent. If interest rates are cut further, this will have a huge impact on pension funds. If yields of -1 percent don’t put investors off buying government bonds, why would yields of -2 percent do so?
Traditional models don’t analyse these types of extremes, and they ought to. And what if, as is often predicted, interest rates suddenly rise sharply? This would hit private investors hardest, especially if equities are also affected. Pension funds will be hit less hard in this case, as their liabilities will fall more sharply in value than their portfolios.
Focus on extremes
How can we analyse extreme investment scenarios so that we can anticipate them? The answer lies in the tail risk model, which Kempen developed last year in conjunction with Casper de Vries, Professor of Monetary Economics at Erasmus University in Rotterdam.
This model doesn’t work on the basis of all outcomes and the usual normal distribution, but instead focuses purely on the worst outcomes. To illustrate this, let’s return to the example of dikes. The fact that a dike is able to withstand 99 percent of water levels, is irrelevant. It only takes a single storm to breach a dike and wipe out all the other days by causing lasting damage. That’s why you need to focus on the biggest outcomes you’ve observed. And extrapolate the trend you can identify within these, to even less likely risks.
For use in shockwaves
The tail risk model enables us to generate predictions about extreme downturns that occur in the wake of sudden major events, such as policy shocks. Take a radical Green Deal that could have a huge impact on the European Union’s economy. Another highly topical example of a tail risk is the coronavirus outbreak that erupted at the start of this year. At the time of writing, it was already having a significant impact on the investment markets and no end was yet in sight. It’s precisely this type of shockwave that the model focuses on.
The tail risk model is not suited to events that evolve more slowly, such as climate change. Everyone understands the problem there, but it’s enormous and plays out across the very long term. This makes it difficult to incorporate into investment policy. That doesn’t alter the fact that extreme climate risk ought to be part of investment policy, but forecasts of long-term trends are probably more significant here. The upward potential of such trends is important as well. The best opportunities in terms of climate can currently be found among European equities, as the European Union is a global leader via its climate policy.
Important to the investment mix
Both private and institutional investors benefit from the tail risk model, although they are affected in different ways by sharp downturns on the markets. Private investors are hit by both falling equity and bond markets in the event of acute market stress. Institutional investors suffer most during a flight from equities to bonds, as the lower yields make their liabilities extremely expensive.
Both dynamics can be analysed using the tail risk model. We believe the model can play an especially significant role in the strategic phase of portfolio construction. It helps all kinds of investors to clarify where their tolerance limits for extreme outcomes lie.
The use of tail risk models is new to the financial industry. The model has not been introduced before because asset managers find it too complex, or simply think that existing methods suffice. We now take a different view. Kempen is the first asset manager to use the tail risk model. In that sense, its application can certainly be described as innovative. We are incorporating the tail risk model into our investment advice and have recently started to apply it in practice.
Reducing stress levels
We use the tail risk model at the earliest possible stage in our investment decision-making, so that clients can see early on what could happen to their portfolios, if extreme events occur. The model can therefore affect the portfolio composition. It may also help to reduce investors’ stress levels. Over a longer horizon, from about one year, the model shows that the impact of tail risks is often not that severe, as markets often bounce back after a downturn. That’s another aspect that traditional models have difficulty mastering.
Our goal is to help investors keep a cool head, even during stressful periods. And this should enable them to keep it above water as well.
This article has been published in Kempen Insight- May 2020
Kempen Capital Management N.V. (KCM) is licensed as a manager of various UCITS and AIFs and authorised to provide investment services and as such is subject to supervision by the Netherlands Authority for the Financial Markets. This document is for information purposes only and provides insufficient information for an investment decision. This document does not contain investment advice, no investment recommendation, no research, or an invitation to buy or sell any financial instruments, and should not be interpreted as such. The opinions expressed in this document are our opinions and views as of such date only. These may be subject to change at any given time, without prior notice.