CLO analytics approaches discussed

Representatives from Moody's Analytics and ING Investment Management discussed how to leverage the latest analytics and data for CLO portfolio management, valuations and benchmarking during a live webinar, hosted by SCI in May (view the webinar here). Topics included manager style, differences between CLO 1.0 and 2.0 deals, and which data points to look out for. This Q&A article highlights some of the main talking points from the session.

Q: This year has so far been characterised by high levels of new issuance, suggesting that US CLO volumes will exceed most 2012 projections. Does this represent a return of the market to pre-crisis levels?
Marc Boatwright, svp, portfolio manager and team leader in the ING US Investment Management senior loan group: Market participants are certainly keeping track of new issuance levels. Volumes increased strongly before the financial crisis and so there is some concern that the market may be mimicking the past.

But it's important to understand the drivers of the prior CLO boom: it wasn't just the availability of cheap credit; it was driven by a surge in private equity transactions. Attractive financing was available for buyouts, due to fundamental economic growth in the country.

These days, the rise of CLOs is being driven by enormous investor appetite because the structures performed well through the downturn. However, the same large buyouts are not occurring, which means the supply of leveraged loans is more constrained. We're seeing a resurgence of the sector, but it's unlikely to return to its previous heights.

Luis Amador, senior director at Moody's Analytics: We're also seeing a resurgence in European CLOs, mid-market CLOs and project finance CLOs. Investment banks, commercial banks, credit hedge funds, loan funds and insurance and pension funds are all investing in the asset class. These entities are investing throughout the capital stack, whereas historically they may only have looked at senior bonds or at the mezzanine level and below.

Q: As at end-April, the largest portion of 2013 deals has been launched by repeat managers with over 10 deals, accounting for 51% of volume. First-time managers accounted for 26% of volume. How should investors differentiate between new managers or smaller repeat managers and larger managers?
Boatwright: Track record is crucial. New managers typically are led by teams that previously worked at firms with track records in the past.

Large managers can point to not only a certain style, but also a track record that demonstrates adherence to that style. Scale managers also tend to have other funds (not just CLOs) under management, which reduces the franchise risk - investors are buying into stability.

That isn't to say that there aren't compelling reasons to invest in smaller managers or first-time managers. What bothers me is that some people are increasingly investing in the structure of a deal, not the manager. They're investing in CLOs simply as a product.

At the triple-A level, it's possible to earn a bit more spread with new managers. But, as you move down the capital stack, it's important to have a handle on the specific style of the manager.

It may be that the manager has a unique edge; for example, in middle market loans. Whatever it may be, investors need to understand in their own minds what their investment drivers are and what a manager is really marketing.

Q: How significant are manager differences and which other exposures are important to monitor?
Boatwright: Individual manager performance pre-crisis looks similar to today's performance, but this doesn't account for collateral overlap. A low diversity score, for instance, is a good indicator of poor performance.

It's important to focus on how a portfolio would perform in a stress case. The reality is that a low diversity score and high concentration indicates that noteholders could suffer chunky hits in terms of defaults, with the potential to shut cashflows down.

I'd recommend analysing nuances in manager style over time and how it migrates. By looking at the data, it's possible to to identify trends that differentiate a manager across different portfolios.

Q: Which metrics can aid investors in determining whether a manager is debt-friendly or equity-friendly?
Amador: Whether a manager is perceived to be debt-friendly or equity-friendly is an interpretational issue. However, investors can look at factors such as disruptions in equity cashflows or fees as a guide. Another way of gauging manager performance is to analyse the standard deviation of cashflows over a given period.

Q: There are a variety of differences between CLO 1.0 and 2.0 structures. What are the key factors to consider in secondary market deals versus primary and as a debt holder versus equity holder?
Boatwright: The differences between CLO 1.0 and 2.0 structures are well-known: newer deals feature lower leverage, higher spreads, shorter average lives and mostly high quality first-lien collateral. Baskets are limited, so managers don't have as much flexibility to create alternative asset portfolios.

As a debt holder, the aim is to constrain the manager as much as possible regarding reinvestment and flexibility to move into alternative assets. Yet managers need enough flexibility to move into safer assets when necessary. If a manager is too constrained, a deal effectively becomes a static structure.

Amador: The focus on comparative analytics and benchmarking has increased. There is also increased emphasis on isolating 1.0 deals versus 2.0, with reporting and benchmarking being undertaken against their own separate universes.

Q: Covenant-lite percentage levels within CLOs are at an all-time high. Are there interpretational issues in the market when it comes to these instruments?
Boatwright: Much of the focus on the size of covenant-lite baskets in 2.0 deals is due to the concern they caused in 2006-2007. It's important for investors to recognise that leveraged loan managers have had to compete with high yield managers: offering cov-lite loans expands the access that high quality issuers have to public bond markets.

The definition of cov-lite in CLO indentures is nuanced. In order to manage the size of the basket, more scrutiny is emerging over what constitutes a cov-lite loan and ultimately which additional protections could be gained through other loan facilities of the issuer. For instance, we've seen the definition change to reference cross-default language to another facility or a facility that is pari passu to a bank loan.

I don't think trustees are tracking this at the moment, so it's up to individual managers to report and flag such activity. But it can dramatically reduce the number of cov-lite loans under the definition and create more room in the cove-lite basket.

Q: The volume of leveraged loan repricings is also at peak levels post-crisis, with first-quarter repricings totalling US$126.7bn. This is the highest quarterly total since 2007 and tops the US$73bn seen for all of last year. How is this impacting analytics frameworks?
Amador: Loan repricings are driven by cost of funds, so it's important from a valuations perspective to project costs of funds and interest rates. From a pure functionality standpoint, the ability to flag assets, look at the terms of individual loans and come up with assumptions on whether they have a higher potential to reprice is key. An additional wrinkle to be aware of is that CLO 2.0 deals often contain features that allow for tranche repricings.

The cost of funds environment is also driving managers to call their deals. At least 22 transactions were called in the first quarter by equity investors looking to invest in new CLOs.

Boatwright: Focusing on 2.0 deals, repricings have had the most dramatic impact on spread and floors. At the beginning of 2013, newer issue deals carried about a 440bp spread and 120bp floor for a 560 coupon; in April, this had declined to 400bp spread and 110bp floor. This is having a minimal impact on older vintages because they aren't invested in higher spread loans.

However, there is an impact on cash too, which many people don't tend to take into account in their modelling. The amount of cash drag on a portfolio has a significant impact on vintage deals because they comprise of older loans, which are being paid off or extended beyond the legal final.

Structures are also facing pressure on WAL tests. The net effect is that it's increasingly difficult to use cash quickly and efficiently for reinvestment. Settlement times are slowing down, so visibility around the real cash number is worsening.

Q: Leveraged loan defaults have experienced a slight up-tick in the US and a slight decrease in the UK. Is this a cause for concern and should it impact modelling assumptions?
Boatwright: The biggest impact of the up-tick is on 1.0 deals because most of the defaults are in connection with carry-over loans, where issuers have kicked the can down the road. It's a case of drilling down into the single names that are pretty well-known in the market.

Amador: Against this backdrop, Moody's Analytics is focused on comparing the fundamental value of a deal versus its market value. It's a question of linking expected default frequencies to each loan to come up with a fundamental value.

Q: Data and its usage has changed dramatically since the financial crisis. Which data points are commonly used today and where are the remaining gaps?
Amador: Compared to ABS, trustees tend to report more information on CLOs, including for example the trades that managers are making. The difficulty is knowing what to do with all of this data.

Trustees typically report the seniority of an asset, its industry classification and lien status, as well as any triple-C bucket activity. Moody's Analytics then normalises the data and renders it usable.

We also create reports that track average purchase prices and sale prices, so clients can compare how one manager trades a loan versus another. In addition, they can analyse loan price distributions to gain greater insight into trading habits.

Boatwright: Given the ability to drill down into the performance of each underlying asset, it's incumbent on the industry to access that data as efficiently as possible and provide an effective means of reporting it. The industry is making important strides towards this goal.

Q: How significant is the impact of reinvestment and post-reinvestment language and assumptions to the performance of a deal?
Boatwright: Reinvestment language can have a significant impact on performance - especially for 1.0 transactions. The information isn't standardised, so it's a question of going through each indenture and interpreting it. There are many ways that reinvestment can be constrained, so it's incumbent upon investors to analyse the language carefully.

Reinvestment language is more standardised in 2.0 deals. But it's still possible to input unhelpful assumptions - for example, WAL caps - which could restrict reinvestment activity.

There also seems to be some misunderstanding about when a deal moves to post-reinvestment trading conditions: managers can continue to invest on a discretionary basis, providing they maintain their WAL tests. There is a disparity with deals that continue to invest for a long time, yet there isn't much collateral available to maintain the test - if there is, it has a low spread. These cases erode equity and shorten the life of the deal.

Q: Which advances have you seen when projecting reinvestment cashflows in CLOs?
Amador: We're seeing the use of more factors when setting up reinvestment models. Models tend to reinvest into loans and bonds with specific terms and conditions. Allowing for different types of assets over time has become more important, as well as looking at the price of an asset over time.

Q: Has the use of loan prices impacted the way you analyse transactions today?
Boatwright: Absolutely. Some loans may be being restructured and so a portion of the price is a function of the coupon. But residual value may not be reflected.

We're seeing a drive towards manager barbelling, whereby they add certain loans to the pool to make the arbitrage work or include discounted loans at the back of the ramp to meet target par.

Amador: Certainly the use of loan prices in metrics has increased. There has also been a shift in focus towards running market value OC, market value attachment and detachment points, thickness and collateral NAV metrics.

Q: How important is modelling single-name risk to the outcome of a deal?
Amador: Single-name analysis is extremely important. Moody's Analytics has developed modules that allow clients to flag individual names and keep track of their movements. If a client has an opinion on whether a loan is likely to default or pay-down, they can incorporate this opinion into the standard assumptions.