By Hamlin Lovell, NordicInvestor 

BNY Mellon operates a multi-boutique model of affiliated asset managers including Insight Investment which managed £683 billion in assets as of 30 September 2022. This interview with Portfolio Manager, Paul Benson, from Insight focuses on the ‘efficient’ beta range, which comprises of around USD 6.5 billion in six credit strategies and USD 1.5 billion in a sovereign bond strategy.

In some credit sub-asset classes, such as US high yield, many pure passive products and active managers have historically lagged the leading benchmarks used by institutions. A happier medium could be an efficient beta approach with tight tracking and a fundamental factor tilt approach mitigating some biases that can hamper long-term returns for both index-tracking and active products.

Insight’s largest and oldest efficient beta product is the US High Yield Corporate Beta strategy, which has matched its benchmark – the Bloomberg US Corporate High Yield Index – over seven years. This is no mean feat, especially when some ETF providers still fail to keep pace with more liquid benchmarks – such as the iBoxx USD Liquid High Yield Index and Bloomberg US HY Very Liquid – that have historically lagged the mainstream Bloomberg index.

Index pioneers

The team managing the strategy has a long heritage in passive investment products, having pioneered equity index funds in the 1980s and fixed income index funds in the early 1990s, as well as becoming a large sub-advisor to fixed income ETFs. Efficient beta credit came later: “In the late 2000s, we started to receive enquiries from investors who decided that high yield and emerging market debt might not be ideally suited to a fundamental active approach. Institutions had significant allocations to high yield but active managers making up 95% of the high yield universe nearly always struggled to match the broad Bloomberg US corporate high yield benchmark, net of fees,” says Benson.

Transaction costs

One challenge of high yield is the transaction costs, such as bid/offer spreads as high as 50-80 basis points for individual bonds, compared with pennies or fractions of pennies for equities. This cost is material since even a bond index strategy typically has portfolio turnover of close to 50% – and active managers trade more often. The efficient beta strategies have turnover close to the bond indices, but minimise costs through predominantly trading baskets somewhat similar to those used by many ETFs, with bid/offer spreads of 0.25 to 0.35%. But whereas most ETFs trade more standardized baskets, the efficient beta range customize its baskets to some degree. The historic transaction costs of the efficient beta US HY Composite has been 0.13%

Instruments: baskets, indices and timing

Around 95% of the book is typically traded through customized baskets. Index CDS, such as CDX and ITRAXX, is used as a balancing item for the remainder to reduce cash weightings and help keep beta close to one. Very little uninvested cash is held, though a small amount of coupon income in cash is immaterial so long as the beta target is met.

The process is substantially systematic, though some discretion is allowed over execution: “If liquidity disappears in some areas of the credit markets, then more credit indices can be used,” points out Benson. The baskets however worked well around the COVID-crisis in March 2020, when some single name credits and funds investing in them ran into liquidity problems.

As standardized credit derivative indices such as CDX are the cheapest products to trade, with a bid/offer spreads of 4-6 basis points even tighter than on the customized baskets, they are ideally suited to the efficient beta range. “However, they entail complexities around basis, skew and rolls, and therefore need to be paid close attention to as they are not a pure proxy for broad high yield,” says Benson.

Though indices can be sliced and diced into various tranches, they would not permit the degree of granularity and subtlety that Insight pursues in its factor tilts. The efficient beta range typically own around 75% of issues and issuers once mandates are fully ramped up, but for a new SMA or fund it can take a few months to build the portfolio up to this level. Assets are roughly evenly split between comingled funds and separately managed accounts (there is also one ETF: BNY Mellon High Yield Beta ETF, ticker BKHY, launched in April 2020). The high yield strategy recently won a EUR 240 million mandate from Danish pension fund LD, who have also been interviewed by Nordic Investor.

Insight does not publish estimates of market impact, which is difficult to estimate, but timing trading decisions may help to reduce both bid/offer spreads and (any) market impact. “Index reconstitution dates are often the most liquid time to trade, with more volume and tighter bid-offer spreads, and could be a good time for a large allocation. In contrast, US holidays can see less liquid and thinner markets,” explains Benson.

Beta One, CCC credits and defaults

Though it can sometimes make sense to take a view on the timing of basket trade execution, the efficient beta range do not attempt to add value through market timing. To reiterate, they target a strict beta of one, with relatively small tilts in terms of industries and credit ratings. This also implies maintaining close to benchmark weights in the lowest credit rating CCC bonds, which is one policy that distinguishes efficient beta from both active managers and some ETFs.

“Active managers generally underperform over a full cycle because, as fundamental credit investors worry about default, they are naturally going to be structurally underweight CCC credit ratings. This may contribute to outperformance in a year like 2022, but in general running lower beta is a drag on performance in high yield,” says Benson. Indeed, high yield credit has clearly not exhibited the ’low beta anomaly‘ observed in equities.

Some credit ETFs can also be somewhat underweight smaller CCC-rated issues sized below around 300 million of issuance. “These are an important part of the high yield universe where risk is well compensated by higher returns,” observes Benson.

The efficient beta strategy is close to the benchmark in its CCC weighting, but does not indiscriminately own every CCC credit. Its fundamental factor models have historically avoided about one third of defaults, before the event. A further nuance is that when the portfolio does run into defaults, they are not sold immediately. “We do not mind holding defaults because we are typically fairly compensated for that risk. The worst time to sell is when a default is announced or when the name disappears from the index, when other investors are selling for window-dressing reasons. We may hold for a few weeks or a month or two and figure out recovery rates, though we do not go through a full workout process,” says Benson.

Factor tilts

The fundamental factor tilts build on the firm’s many years of experience developing quantitative factor investing expertise on the GTAA (Global Tactical Asset Allocation) side, and the approach has been adapted to different asset classes.

“Whereas hundreds of factors can be involved with equities, in fixed income there is widespread market agreement that just a handful are meaningful,” says Benson. Moreover, the efficient beta range only uses three of the five key recognized factors in Insight’s fundamental credit models: quality, momentum and value (QMV). Carry is not used because it would tend to imply higher beta, and the small size factor is not used because it is harder to locate and more expensive to trade smaller names and issues.

The QMV exposures should probably be seen as ’tilts‘ rather than ’bets‘ because they are optimized subject to ratings and tracking error constraints. “For instance, quality, value and momentum factors are applied to selection within the CCC rating bucket, rather than making a big wager on CCC versus B. We aim to mitigate exposure to left tails, such as the highest default risks, or the highest mis-pricings. If such a bond is a small part of the benchmark, we might not own it. If it has a 1.5% weight in the benchmark, we might hold an underweight position and allocate part of the risk budget to a more healthy cohort,” explains Benson.

This filters through to outperformance targets, which are calibrated to the tracking error. “We have confidence in generating an information ratio of 0.6, which means that we expect gross alpha of 0.30% from a tracking error of 0.50% on the efficient beta range”. This basically means that the gross alpha covers fees and the target is to match the benchmark.

Efficient Beta Plus: Fallen Angels and Investment Grade

The strategies within the efficient beta plus range have higher tracking error targets of 100-200 basis points, and correspondingly higher alpha targets of 50 to 100 basis points.

Insight’s second largest Efficient Beta product, US Fallen Angel Beta Plus, is part of the Efficient Beta Plus range, partly because allocators have more latitude to accommodate deviations from the relevant index: “Though the fallen angels benchmark has historically beaten the high yield benchmark by 2-3% per year, it is not widely tracked as a benchmark, which is one reason to allow more tracking error,” says Benson.

Insight also decided to launch global investment grade as an efficient beta plus for other reasons: “both index products and active managers have historically done better versus the benchmark”.

The Global IG Corporate Beta Plus strategy needs some headroom to invest in non-investment grade credits, where it has a link with the fallen angels product. One of the largest anomalies in credit investing is how robustly ’fallen angels‘ bonds downgraded from BBB to BB recover. “It is a terrible time to sell fallen angels in the event of a downgrade, when many holders may be forced sellers. Therefore, the IG product allows some non-investment grade exposure and expects to earn better risk-adjusted returns from shorter maturity, lower quality paper, than a typical longer-maturity, higher-quality portfolio,” says Benson.


All products are currently making disclosures under SFDR article 6, except for Fallen Angels which makes disclosures under SFDR article 8.

Avoiding tobacco, coal and oil tar sands for ESG reasons has had a marginal impact on efficient beta performance to date. Going beyond those exclusions to screen out oil and gas or gambling may result in a material tracking error versus conventional US high yield indices.

However, the approach is benchmark-agnostic and customized mandates could use ESG credit benchmarks.