DECEMBER 10, 2020

Everything you always wanted to know about Risk and ESG (but were afraid to ask)

By Objectway

Reading time: 2 min


A risk aware investment process needs to be supported by a solid approach and integrated with risk data feed that are correctly imported, validated and armonised to support the entire investment process.

But what does it really mean?
Let’s talk it over with Elena Giancaspro from Confluence, a global leader in data-driven investment management solutions.

Why isn’t it enough to have “a” risk system, but it is vital to have “a good” one?

Portfolio managers are instinctively by far more interested in performance than in risk. But to generate steady performance, risks cannot be systematically ignored. The regulator does not allow it anymore, imposing to set up risk systems and procedures very seriously. But even imagining a regulation free world, who would entrust a portfolio manager neglecting or misjudging her risks?

That said, it is far from easy to build a “good” risk system. You need years if not decades and expensive quants. One reason being that finance itself is complex, relentlessly giving birth to new investment products, with intricate payoffs that require deep quantitative knowhow to be modelled correctly.

Neither are troops of ingenious quants sufficient. Add technology, that translates formulas into software and therefore the need to hire coding experts able to programme at the technological cutting edge. Even more so when software has to be scalable to millions of portfolios, demanding immense calculation power and the solution of equally immense tech-problems, to which cloud computing is now providing some relief.

Last but not least, to feed a software and provide it as a service to clients, you need high-quality market data. Such data cost a fortune and can be afforded only if there are significant economies of scale. Being small is no longer desirable these days. Expensive data providers increasingly behave like oligopolists, forcing small players to ally, merge or being bought. This makes market data an almost impossible challenge for nearly all in-house built risk systems, no matter how well they are conceived. An exorbitant cost that is more and more difficult to be borne by small players.

How can we say a risk system is a “good one”?

Given that a foolproof risk system has not been invented yet, and that you realize you have a good (or a bad) one only when things go wrong, there are undoubtedly some criteria to determine if there is quality.

A risk system is valid for example if it predicts a maximum loss with some probability and then, after a real downward movement takes place, the effective loss is no greater than the one anticipated. In other words, the level predicted by the system has not breached. It should be always possible to make this “quality check”, which is in jargon called back-testing. The European regulator requires it for all Ucits funds. In the industry you can find many sophisticated and very elegant risk models. But they usually rest on a bunch of assorted assumptions that do not withstand reality, especially when financially turbulent. These models will most likely fail the mentioned back-testing.

Secondly, a good risk system must be able to capture existing correlations between assets and risk factors; for example, if interest rates keep falling due to the persistence of pandemic and a worsening economic outlook, by how much will credit spreads on my high yield bonds widen? These correlations are also the key ingredient of certain Stress Tests, that make them more reliable.

Finally, a good risk system should be able to cover the myriad of asset types that can be found in a portfolio, even those with strange, newly invented, eccentric payoffs like exotic certificates, or structured and illiquid bonds. Think of subordinated callable perpetuals with step-ups. Or derivatives like options, warrants, or simple fx- forwards to hedge against currency risk. These instruments may cause headaches to those who need to value them and often the easy way out is to proxy. But then, when the tempest rages, approximations may turn into nasty losses.

What should a risk system tell?

Among other things, a risk system should tell how much a portfolio can gain or lose in a certain time frames. Or in extreme cases, when we find ourselves in low probability but very high impact scenarios, i.e. in the so-called distribution tails of our negative returns. Sometimes we just want to apply shocks to certain variables. How much is the P&L affected when risk factors move in certain directions? Say the 5 year swap rate moves by 10 bps and the 5 year relevant credit spread widens by 10 bps: how much do we gain or lose on our Italian government bonds?

Another important question a sound risk system should answer is how much a portfolio is diversified, as diversification mitigates risk and involves both assets and risk factors. I believe my assets are diversified, but I find out they aren’t; conversely, I own assets that taken individually are high-risk, but then put together inside a portfolio benefitting from negative correlations, they reduce the overall risk and boost performance over the longer run; if long duration government bonds are added to equities and high yield bonds, could it be that the interest rate factor diversifies in respect to equity and credit risk factors, reducing the overall risk? The answer could be yes!

ESG Risk: what does it really mean?

Put simply, it means to integrate the traditional risk analysis with an analysis which is much more qualitative. The former as we have seen is rigorously quantitative and requires for example to model and populate math functions that return numbers. The latter is instead based on collecting a vastness of public data and financial indicators that by getting crunched by algorithms release ratings or scores. Such outputs like official ESG ratings may range from EEE (the highest degree of sustainability) down to F (unsustainable investment).

Adopting ESG ratings, we can try to ensure that throughout the investment process the issuers we are considering are behaving well from an environment, social and governance point of view and that they are for example not engaging in controversial activities, may it be a corporation, may it be a government.

Also, by gauging ESG risk, we can decide to exclude certain issuers, being them for example outright polluters or ruthless employers or dodgy accountants when producing their balance sheets. A minimized ESG risk not just lightens individual consciences but should also reduce market risk and possibly lift portfolio performance in the longer run.