01 August 2019
Investing in illiquid credit brings about a whole set of challenges in terms of risk oversight and valuation of the assets. In part two of this roundtable from Insurance Asset Risk and NN Investment Partners participants discuss why relying on rating agencies' assessment is not an option and how to reflect the illiquidity premium in the asset valuation.
Conor Sweeney, investment risk and operations manager, Canopius
David Epstein, group ALM director, Aviva
David Walker, senior staff writer and head of projects, Insurance Asset Risk
Erik Vynckier, board member, Foresters Friendly Society
Hayley Rees, head of direct investments, Legal & General Retirement
Ian Coulman, chief investment officer, Pool Reinsurance
Jeev Muthulingam, head of insurance investment solutions, NN Investment Partners
Peter Vasey, head of investment risk, Pension Insurance Corporation
Prasun Mathur, head of shareholder investment, UK & Ireland Life, Aviva
Suresh Hegde, lead PM export finance & government loans, NN Investment Partners
Chaired by Vincent Huck, editor, Insurance Asset Risk
Vincent Huck: What should firms do to manage the risk oversight well? Do you do it in-house, do you rely on third parties such as rating agencies?
Erik Vynckier: I am certainly not relying on rating agencies. I put one particular loan to two smaller rating agencies: one rated it BBB, the other rated it B-.
Hayley Rees: That is a bit of a stretch.
Erik Vynckier: It is a stretch, but you should always check the background of the credit analysts working the rating to begin with. If the rating agency is cheap, that might be a reason not to use them, because it will be a shallow process. One of the unfortunate things is that rating agencies that land the ESMA [European Securities and Markets Authority] certification say, 'With the regulatory certification in our pocket, we are good to work for insurers.'
You have to have a second look at which rating agency you use, because ESMA's list is, in my view, full of unreliable rating agencies.
Hayley Rees: I would agree with that. We have an in-house ratings team, and their backgrounds are from Fitch, S&P, Moody's. If we take an external rating, we will only take it from those three houses; we do not use the other agencies. However, in terms of risk, you have to start by ensuring you actually understand the risk of the assets you are taking on to your book. If you do not, then do not do it.
Conor Sweeney: If you've hired credit analysts to do due diligence on the investment, why from an internal perspective do you need a categorical rating?
Hayley Rees: Because it feeds into our capital model – that is the simple answer.
Erik Vynckier: A fairly large life insurer in the UK will have an internal model and alongside that they have an internal rating process, which they can submit to the regulator as part of the internal model for approval. Non-life insurers probably do not have assets of a scale to give such attention to the asset side in their internal model; the smaller players have the standard formula and that forces them down the road of needing ESMA-certified rating agencies.
Ian Coulman: We would outsource it and totally turn to the expertise of the managers we have outsourced to.
David Epstein: It depends partly on where your origination team sits; if it's within your business, then you'll need the expertise in-house. For us, if the origination is with the asset manager, then we would have a lot of expertise there. Then we also need expertise in the business to ensure what we are receiving through the asset manager is appropriate.
One area of focus at the moment is asking, how does the internal rating process compare to external credit rating agencies? Similarly, if you were on standard formula, what would this look like? So, understanding how different these approaches would be from your internal process.
David Walker: What is the response if what you are being offered from an external agency is different from the internal rating?
Peter Vasey: Where we have comparisons between internal and external ratings we may find a divergence of maybe a notch; you certainly do not get the spread of ratings that you see through some comparisons across agencies. We will go through a process within risk of validating all of the internal ratings that are generated by our credit teams, and one aspect of that will be looking at the comparables in the external environment.
Erik Vynckier: We call the rating for a private placement by the same name, but it is a completely different animal than a rating we would use for a public bond. For illiquid assets you are the only user of that rating at the end of the day, so you have to work on a daily basis with that rating to understand how it was established, and take all of the commentary and all of the judgment around that as part of the rating, instead of going to Bloomberg and finding out what the rating for a public bond is.
Vincent Huck: How do you reflect illiquidity premium in the valuation of the asset?
Conor Sweeney: We try to understand the economic drivers of the asset performance, for example the contractual cash flows and associated maturities. Even if you do not have access to the full list of underlying assets or contracts, at least have certain so called "model points". Then, you can do your own modelling of their behaviour alongside that of your wider portfolio. We have a stochastic model, which also incorporates liquidity demand from volatile claims liability outflows, and a standard function which translates risk to the premium we expect.
Erik Vynckier: For me, a point of reference for the pricing of illiquid debt is the leveraged loan markets, and these will serve mid-sized companies. The leveraged loan market is fairly liquid. The liquidity premium for private debt really starts from there. What many people tend to do is see what interest rate they can push through in the primary markets, see what sort of conditions they can enforce on the borrower. Then, that is their starting point for what the liquidity premium in excess of leveraged loans was at the point of sale. It seems such an approach might underestimate volatility.
Prasun Mathur: Every asset type will be different. For example, equity release mortgages could be priced based on the going rate of mortgages in the market. Assets like infrastructure debt will often have public credit equivalents that are used in valuing them.
David Walker: So it is deal by deal? It is not possible to say for all lifetime mortgages or all parts of illiquid credit, 'here is the premium I expect across the board?'
Prasun Mathur: No, not all. It depends on the asset class and the nuances of the individual investment that you are making. You need to figure out what kinds of risks that investment is creating, and you price for that.
David Epstein: There is an origination pricing valuation question, and then there is how you value this transaction through its life.
From our perspective, we consider benchmarking, so having a good understanding of primary and secondary markets, allows us to see how some of these variables are changing. You can factor these into your own valuations for specific assets. It is not enough just to see what the market looks like at purchase and say, 'I have this premium, I am going to use that for the rest of time'.
Erik Vynckier: Obviously if the market is opaque it is quite difficult to do. If you are not in the primary market on an ongoing basis essentially you won't know, so that is another franchise issue. If you want to come up with your own data and assumptions, you have to be in primary market to see where the market is trading now. There is a little bit of opportunity on the regular mortgages where you can get prices from platforms and see where the market is trading.
Vincent Huck: In the decision-making process, asset managers tend to say it is a question of being nimble and being quick to react. Is that a hurdle for insurers, whose business model is usually described as being slow?
Hayley Rees: We work with LGIM as our asset managers in the UK and we can be nimble and quick because we have been doing this for a number of years. We understand what we are looking for from the market, and what premium. It is a fallacy to say that all insurers cannot be nimble and quick when it comes to executing and originating these assets. We are always considered in terms of our investment choices, but some assets, because we know them a lot more, we can be agile and quick.
Now, if we go into new asset classes, we are not just going to jump straight in without fully analysing the new asset class. Trade finance was mentioned earlier – nobody has really jumped into that because people are analysing it. However, if an infrastructure asset comes to us tomorrow, that is something we are able to analyse quickly.
Ian Coulman: Yes, I would echo that. If it is a new asset class we have to go through a thorough process to justify an investment and ensure our investment committee are comfortable that it is an asset class that's going to add value.
Prasun Mathur: We are strategic investors and to enter a new asset class we need to see scalability from that opportunity.
Erik Vynckier: The way insurers invest in these assets changed. Fifteen years ago, you had some excess capital, you looked at individual opportunities, and you did it on an ad-hoc basis. Now it is very systematic and invests policyholder funds. The difference? Quantitative easing!
Conor Sweeney: I think as part of a general insurance company, investment performance is perceived to be a very secondary consideration compared with the underwriting side, from a value-add perspective. We have a lot of work to do to understand what our most efficient operating model is in terms of extracting value from the asset side.
The underwriting side is a well-travelled path. Here is a given risk type or line of business, rates might be attractive or there is growth expected, let us go hire this underwriting team, maybe having to bring them across from another insurer, which is a common occurrence.
Whereas on the investment side of things, there is perhaps not the same understanding as to what the operating model should be. For example, certain credit risk exposure that is written on the underwriting side gets a lot more management attention and investment, essentially in capabilities. Whereas we have to find our place in terms of, what is the right level of due diligence or asset allocation granularity? What level of capability should we target in order to be able to access a given asset type? I do not have an answer (yet).
We have an exercise, and I suspect other general insurance companies should, perhaps, go through and see whether investment in an investment function is worthwhile given the relative attractiveness of the so-called liability side versus the assets.
David Walker: Would you expect that to change for general insurers after the last two years, so they may be relying more on investing in some higher yielding investments to turn a profit - given what has happened with natural catastrophes?
Conor Sweeney: My original background is capital modelling. We used to get requests to run certain portfolios through the internal model to see what the impact on the SCR was. Because we have been a very cat-heavy writer, the diversification between underwriting performance and investment market risk means we do not have to hold any more capital when taking more investment risk. Hence, with the increased expected return on equity, surely it is a no-brainer?
However, it completely misses the other side of the portfolio construction coin where adding financial asset risk in itself - for example, by holding more equities on a GI balance sheet - really does not add any value whatsoever to the franchise value of the company, as the economic theory says.
As a shareholder, why would you take the risk on a GI balance sheet when you could hypothetically just go and obtain the same investment leverage by transacting in equity futures on your personal account? While investment returns can potentially be more attractive compared to the underwriting side, it potentially misses the wider question, what risks should be on our balance sheet?
To read part one of this round table click here. Part three will be published on 15th August.