Television talent shows are ten-a-penny these days. Discerning viewers know when to tune in to enjoy the most drama: the early audition stages, where they can witness all manner of unlikely entrants, and the final round, when the surviving contestants engage in an all-out battle for the popular vote.
If the European regulatory process was such a talent show, the Packaged Retail and Insurance-based Investment Products (Priips) framework would be approaching the final stage. On June 23, the European Supervisory Authorities (ESAs) published a technical discussion paper detailing four options for determining the risk ratings that will need to be disclosed in key information documents (KIDs) for all products captured under the regulation from late next year.
Now, the industry – and national regulators – have until August 17 to nominate their favoured methodology for inclusion in draft regulatory technical standards (RTSs) due to be published in the autumn. It looks set to be a fraught battle.
"The ESAs are representing the national supervisory authorities, and the national supervisory authorities are split on this issue," says Thomas Wulf, secretary-general of the European Structured Investment Products Association (Eusipa). "The decision on the risk factor should be made after they get the feedback and it becomes clear what market participants support. That's a legitimate way of working," he adds.
First impressions from dealers and regulators suggest that two of the four methodologies are in contention for the top spot in the RTSs (see box: What's on the table) – option two, which has the initial support of the UK Structured Products Association and several dealers that Risk.net spoke to for this article, and option three, which counts parts of the German, Austrian and Swiss industries among its fans.
It’s like saying if I have three good tyres on my car and one that is completely bald, I’m going to use the median tyre quality – the good tyre – rather than the one that is going to go pop
The alphanumeric indicator proposed in option two, which shares similarities with the Ucits synthetic risk-and-reward indicator (SRRI), appears to have support from the ESAs, who write that it would be a reliable measure that comes close to using a full valuation approach.
Some lawyers agree that segmenting market and credit risk is sensible, while at least one distributor argues adopting option two could lead to a level playing field between funds and structured products.
David Stuff, chief executive of UK-based Cube Investing (pictured), says: "Volatility is an important measure because it tends to be the one used across the industry to measure risk. If a fund has a risk of 15% and a structured product has a risk of 15%, you want this to be a direct, side-by-side comparison. You have to do your best to have the same calibration of the two numbers."
Meanwhile, one strategic adviser at a major European dealer familiar with the work of the European Securities and Markets Authority's expert group on Priips says an approach similar to option two has performed well under simulations, generating similar risk classification as those that emerged from more complex value-at-risk-based approaches.
Option two also has a powerful backer in the shape of the UKSPA. Zak de Mariveles, chairman of the trade body, says he is pleased to see the option included, as it shares parallels with its own risk ratings for UK financial advisers. The group has stated its strong objection to a completely unified single risk-reward indicator.
Others warn, however, that a bipartite indicator goes against the spirit of the single summary risk indicator promised in the level one text of the Priips framework, while some question whether such an approach would be easily grasped by the typical retail investor.
"The most important thing from our perspective is whatever factor is introduced must be understood by the client. It should [also] be cost effective to produce," says James Chu, head of product development at UK distributor Reyker.
A question mark also hangs over the claim that volatility is an appropriate measure of downside risk. After all, to many investors, the underlying volatility of a structured product matters only in so far as it threatens their initial capital.
"There are some people who use the entire distribution of returns [in contemporary volatility risk indicators]. One problem with that is that structured products are almost by definition designed to give you a non-normal distribution of returns. So if you use an entire distribution you are going to get a volatility number, but it's not going to represent the riskiness of a product because it won't represent the downside risk," says Stuff.
Stuff is also wary of a credit risk score that relies on rating agency disclosures. He says the way in which multiple credit risks in a single product are scored is illogical: "They've done something really bizarre in that they use the median [rating] if the product is linked to multiple credits, which I can't understand at all. It's like saying that if I have three good tyres on my car and one that is completely bald, I'm going to use the median tyre quality – the good tyre – rather than the one that is going to go pop."
Instead, he favours quantitative methods of calculating credit risk, such as credit default swap spreads and implied default scores generated using a Merton model.
The luxury option
Option three is the most heavy-duty quantitative methodology on the roster, with its supporters largely hailing from Germany and Austria. Its popularity in the former is little surprise as its domestic market has been making use of a VAR-driven approach to risk indicators, calculated to 99% confidence over a 10-day time horizon, since 2005.
Supporters of the approach argue that the 10-day calibration strikes the right balance between complexity and utility. "Some risks are harder to cover in a model, because the longer the time horizon, the more you have to input some drift assumptions in the values of the underlying. From the beginning, we chose an approach using a short-term holding period where this drift component could be neglected," says Björn Döhrer, managing director at European Derivatives Group, the firm tasked by the German Derivatives Association to calibrate the VAR model for domestic disclosure requirements.
Notwithstanding this concession to simplicity, concerns remain about how difficult this methodology would be to implement market-wide. Running concurrent series of Monte Carlo simulations for different products may be straightforward enough for Europe's largest retail flow houses, but it is a much bigger challenge for smaller issuers, not to mention distributors that may want to cross-test the risk rating of products that go out under their banner.
The European dealer's strategic adviser says the German approach would be "costly to implement" in terms of IT infrastructure and staff time: "If we have to do this for all retail products that we issue then the pricing would quickly become difficult to produce in a short timeframe. So we would have to invest in more computer power."
Others argue the approach is not suited to markets where longer-dated products dominate. "The German market is a market of short-term traders. [The VAR method] is very good for measuring the volatility of a turbo, a mini-future, a call warrant or something that is a short-term trading instrument. It's completely inappropriate for a product with a five- to six-year time horizon," says Cube's Stuff.
Is the hassle worth it? That depends who you ask. Döhrer says the European Derivatives Group approach means firms get a holistic view of all the risks of a product relayed in a single calculation. "With such an approach [you] get the risk fully reflected in the risk factor. When you only do it on a qualitative basis or assumption-based approach, then you would neglect some kind of risk," he says.
One dealer, however, says the VAR approach, despite being well engineered, ends up producing similar results to the SRRI methodology, which brings into question the value of the extra cost inherent in using VAR. "Eusipa did an exercise where they compared a whole set of products with the two methodologies. They reached very similar results. It's a bit like taking a trip in a Rolls Royce or a Dacia; they're both good vehicles, but the luxury of a Rolls Royce is much more expensive."
Quality over quantity
The stragglers in the race for supervisory approval are seen as options one and four. The latter's critics see it as an unwieldy compromise that will do little to foster comparability in the market.
The ESAs themselves comment in the paper that the flexibility offered in the second level of the indicator means it would be difficult for the risks of different products to relate to each other. "Furthermore, it is difficult to compare products if they are first segregated on the basis of a general product characteristic," they go on to state.
The European dealer source adds: "This fourth option has been put in following a discussion in the expert group to probably mimic what is done in other regulation, like Solvency II for insurers, where you have a standard model and an internal model. The standard model can apply to everyone, and the internal model can be more complex and designed by the manufacturer. It's included in the discussion paper as an [additional] option, but compared to option two or three, I would say it's less [viable]."
Option one counts among its supporters those who argue in favour of a qualitative approach to risk ratings, such as Belgian regulator the Financial Services and Markets Authority. Eusipa's Wulf says the watchdog has been seeking supporters at the European Securities and Markets Authority level for just such a methodology, and is lobbying other regulators on its assumed merits.
Its zeal may stem from the fact it has been working on a qualitative indicator for domestic regulatory purposes, and hopes to align this home-baked proposal with the European Priips-KID.
However, the watchdog has a struggle on its hands to keep this initiative on track in Belgium. The domestic risk label was outlined in a Royal Decree of April 25, 2014 – before Priips was ratified at the European level – along with new rules on the advertising of structured products. Though due to come into force on June 12 this year, the country's minister of economic affairs suspended implementation of the risk label methodology after it became clear there was significant overlap with Priips. The future of this section of the Decree is now in doubt.
Spain is another country flying the flag for the qualitative approach. The markets watchdog – Comisión Nacional del Mercado de Valores – was working on a domestic qualitative risk indicator along similar lines to Belgium's until earlier this year, when responsibility for the initiative was handed over to the Ministry of Economy and Competitiveness. The proposal is due to be finalised before entering force by the end of 2015, according to a spokesperson at one Spanish bank – though there is a possibility that it, too, could be postponed.
Hopes among national competent authorities that such methodologies could worm their way into the Priips framework were dashed in the technical paper, however, which explicitly rules out a purely qualitative approach – citing its reliance on certain judgement calls and potential inaccuracies as clear negatives – in favour of the hybrid qualitative-quantitative option floated by the ESAs.
Yet, say market participants, the fact that the other three options have a distinctly quantitative flavour, and the lack of wider support for a qualitative methodology, means even the hybrid approach may struggle to find its way into the RTSs.
One reason cited is that a reliance on judgement means that risk ratings could be tampered with, or drawn up in such a way as to understate the risk. The strategic adviser at a European dealer says: "The key takeaway reflected in the technical paper is that quantitative risk measures are preferred over qualitative risk measures, as it was made very clear that a qualitative measure is easier to manipulate."
The importance of the final risk rating methodology goes beyond determining how the risk of a product is perceived by retail investors, however. It will also determine how other provisions of the incoming Priips framework will affect issuers – most significantly, Article 10 of the level one text, which lays out the obligation for product manufacturers to review and update the information contained in the KID.
Andrew Sulston, partner at law firm Allen & Overy (pictured), explains: "What's chosen for the summary risk indicator will be really important for determining whether the updating requirements are going to be viable or not. Depending on how sensitive the SRRI is to specific changes in the market, that would actually determine how onerous it becomes and how often you have to update."
Somewhat counterintuitively, this facet of the regulation strengthens the case of those supporters of the qualitative approach. "The question of how often you have to update the KID is not clarified yet, but the argument is being made that purely qualitative models would find it easier to deal with as they wouldn't have to update as frequently. That's a procedural argument, though, not a technical justification, given that risk parameters of products can obviously change during their lifetime," says Wulf.
For an insight into how the volatility-based approach recommended in option two might perform in relation to this provision, it is instructive to look at how the SRRI for Ucits has behaved to date, says Mike Gould, retail markets specialist at the Investment Management Association in London.
"The relevant regulation requires a revision of the SRRI if the relevant volatility of the Ucits has fallen outside the bucket corresponding to its previous risk category on each weekly or monthly data reference point over the preceding four months. Although we don't keep any figures on this, in the UK this is a fairly rare occurrence. The buckets are quite broad and portfolio managers usually manage the risk profile of their funds so that volatility does not vary significantly enough to warrant a change in the SRRI," says Gould.
Although the applicability to Priips may be limited because the methodology will be different, this could be interpreted as a further fillip for option two.
However, none of the options on offer have erased doubts that a one-size-fits-all methodology for all Priips can be found. An all-inclusive approach that could embrace buy-to-hold insurance-based products such as unit-linked policies as well as short-term punts like turbo warrants is a big challenge, and Sulston says the ESAs have to be wary of not "pushing square pegs into round holes".
Wulf says that while "some [approaches] are more suited to maintaining a broader product landscape than others," Eusipa does not have a final recommendation as yet.
Others are already in a position to make their preferences known. "The best compromise – and at some point we are looking for a compromise – is a volatility-based risk measure," says the source at the European dealer.
The road towards that compromise is fast running out, though. With the ESAs having promised draft RTSs for the autumn, that gives market participants and regulators alike only a few short months to thrash out an accord.
Box 1: What's on the table?
The ESAs have recommended four possible methodologies for determining the Priips-KID risk rating. The industry has until August 17 to submit comments.
• Option one: Products are classified into risk categories based purely on their qualitative characteristics - for example, whether they are capital protected. For some types of product, such as investment funds, a quantitative risk measure will be used to supplement the rating.
• Option two: Products are classified via an alphanumeric indicator. Market risk is rated on a scale of one to seven based on a quantitative measure of a product's underlying volatility, and credit risk is rated from A to G using a methodology based on credit rating agency disclosures.
The technical paper says this volatility score could be based "to a very big extent" on the methodology currently used for Ucits funds, modified to reflect the characteristics of structured products. For example, the ESAs propose calculating the volatility of a product's bond component and risk component separately. The latter would be computed using a "delta approximation" approach, which makes use of both historical volatility and implied volatility to generate the score.
• Option three: Market and credit risk are scored together using a value-at-risk methodology. Two variations are proposed. The first recommends a VAR calibration set at a 99% confidence level over a short holding period, using a Monte Carlo simulation to generate the distribution of returns. Products will be ranked on a risk indicator from one to five depending on the number churned out by the VAR calculation. The second variation also uses VAR, but with variable time horizons to account for the different maturities of different products.
• Option four: A two-level approach using a mixture of quantitative and qualitative tools. The first level classifies products "in a very broad or simplified dimension", states the technical discussion paper, while the second level would provide a more in-depth analysis using volatility or probability-of-loss measures, as in options two and three.
Box 2: Who's backing what?
The number of recommendations laid out in the technical discussion paper may look like diplomacy - but in fact it reflects the divided opinions of European regulators. The option that wins the most political support will be in a strong position to get policy-makers' seal of approval in the final draft regulatory technical standards. Although whichever methodology wins out, it will probably cause grumblings from those lobbies that missed out.
Regulators in Belgium and Spain, and to a lesser degree those in Portugal and Italy, are said by people familiar with the negotiations to favour a predominantly qualitative approach, as reflected in option one.
The UK and France, meanwhile, are said to favour option two, which is similar to the UK Structured Products Association's own alphanumeric risk indicator and the synthetic risk-and-reward indictor (SRRI) used to classify Ucits funds.
Option three – the VAR-based methodology – is popular in Germany and Austria. Option four is regarded as a compromise option that has yet to stir strong feelings among regulators.