How did that change the way you invested?
Based on the assumption that we could now vary exposures, that led us towards an approach which does two things: it seeks to attract returns from market trends, but does so in a way that is not constrained by market cap indices, but instead examines how you maximise returns. This in turn led us in the direction of factor-based portfolio construction. I wouldn’t say we are entirely there yet, but it’s where we believe the future is. It also led us to seek returns that are both uncorrelated but also significant because they are driven by market inefficiencies. This has opened the door to more idiosyncratic investments, alternative strategies that are high-performance which try to extract risk premia which are either behavioural or non-economic.
An organisation like CERN must have a culture of innovation and bold ideas and clearly risk management is a vital part of what it does, but what is it difficult convincing the board to adopt this model?
Yes and no. On the one hand it’s always a hard sell for a board to accept significant change, but at the same time the board demanded change, so we went in with that mandate, but without a clear solution demanded by the board. Having such a brilliant board as we do at CERN is difficult in a way because you really do get challenged very thoroughly on risk models, control loops and such things. You are talking about people who are used to juggling numbers and complexity and risk. That part was a real challenge, because you have such a demanding board, but at the same time we had an understanding that change was needed. It was pushing for change, but needed to be reassured that the change being made really made sense. The other issue which was wonderful was how we mapped the CERN risk control model into the investment world. CERN is basically operating a nuclear particle accelerator, so it is regulated like a nuclear facility: extremely stringently. It uses as much electricity as a city of one million people, which means any short circuit could potentially cause enormous damage. At the same time it operates at the frontier of science, which by definition means doing things which have not been done before.
So they recognised the similarities there?
So yes, the question is: how do you operate a nuclear facility that has to innovate while being regulated with the most stringent standards possible? This is the equation they cracked at CERN and for us the major breakthrough came when I was discussing this challenge with the board. I said: “Look, we need a risk budget and we need staff to be able to innovate and be flexible and try things which have never been done before, but with controls, so you on the board can feel comfortable.” And the board said: “Well, this is exactly the challenge we have.” So finally we ended up bringing from CERN the person who had designed and was responsible for the safety infrastructure for the organisation. When we explained the problem to her and where we wanted to go, she was able to help us map the CERN model and transfer it over to the investment process. So this was a big breakthrough and the investment governance model is very much inspired by the CERN safety management process.
Your approach is very much about being ‘liability aware’, which is somewhat different from the ‘liability driven’ approach adopted by so many UK schemes. Are they being restricted by this focus?
I think yes, it is restricting them, but at the same time they are not being given much of a choice, because the regulator has in many cases adopted that LDI framework and imposes punitive consequences if funds depart from certain criteria based on that framework. We’ve seen that in the UK and in Holland. So trustees have to work with that and it’s difficult to escape. The CERN model is also fully compatible with an LDI approach, but the risk, instead of being defined in terms of absolute asset losses, gets defined in terms of funded status risk. The risk budget is a funded status budget as opposed to a pure asset loss budget. Instead of defining the risk as the maximum loss of assets you could incur and still survive, you define it as the maximum funded status deterioration you could incur and survive. It complicates the calculations, but notionally it’s the same approach.
Can the model be easily-replicated by other investors wishing to go down this route?
Absolutely. We are a research institution and sharing research is part of our soul. We wrote a research paper on the model (http://www.geog.ox.ac.uk/events/130926/economou-etal-2013.pdf ) and got some very good feedback on that from pension funds around the world. There’s an extraordinary amount of talent among CIOs in this industry and the challenge is how do you free up that talent to be as effective as possible? I think what people saw in that paper was a framework for how you can do just that. In a single sentence what the CERN model does from a management standpoint is to move you from managing a single dimension, which is return versus the benchmark, to managing several dimensions: return versus the benchmark, risk versus the risk budget and how good you are at converting risk-taking into return.
Comments