Paris-based asset manager Ossiam has entered the rarefied market for smart beta in corporate bonds with a new strategy drawing on Moody’s Analytics credit-risk measures.

The affiliate of Natixis Global Asset Management, best-known for its minimum variance ETF products, joins the likes of index provider Research Affiliates, which recently launched the Citi RAFI World Corporate Investment-Grade Bond index, and Frankfurt’s Quoniam Asset Management, with its ‘MinRisk Credit’ products.

Portfolio manager Ksenya Rulik told IPE the move from minimum-variance equities to smart-beta bonds was not a natural one.

“In equities, the risk is mostly about price variability, and this is something you can estimate quite reliably,” she said.

“Credit risk, by contrast, is all about what happens in the tail of return distributions: looking at historical variability in bond markets offers no clue to the level of credit risk in a bond, and no insight into future variability. We took our minimum variance strategy from equities and applied to it to bonds, and it really didn’t work at all.”

Not surprisingly, credit analysis is critical for managing the risk of a corporate bond portfolio.

Whereas the Citi RAFI index considers issuers’ long-term assets and cash flow when weighting their bonds, Ossiam relies on Moody’s Analytics Expected Default Frequency (EDF) measure, one of many commoditised, model-based risk products available in the fixed income world.

Rulik characterises this as a way to combine the analytical expertise of the credit-risk product manufacturers with the portfolio construction expertise of an asset manager.

Ossiam’s process begins by selecting, from the more liquid bonds, those with the lowest EDF – a model of publicly listed companies ‘distance to default’ derived from Moody’s Analyitics’ proprietary database of historical defaults and credit-strength trends.

Applied to the broad market, the measure has tended to predict realised default rate increases with a lead of about one year.

Following this step, the process then selects again based on estimates of the bonds’ fair value spread (FVS), calculated by combining their EDFs with a variety of risk factors.

Comparing this FVS with the option-adjusted spread (OAS) observable in the market offers an indication of which bonds are undervalued or overvalued.

Finally, individual bonds and issuers are capped in the portfolio, and duration is adjusted to match that of the broad market.

Using EDF alone to construct a portfolio that ran between 2007 and 2013 improved returns marginally relative to the broad market, but the biggest impact was reduced drawdowns and volatility.

Using the difference between OAS and FVS alone dramatically improved returns – 71.3% versus 43.7% – but made the drawdowns slightly worse and did not improve volatility.

Combining the two appears to select the best portfolio to protect against downside credit risk, before selecting from within that portfolio the most under-valued bonds that present the highest expected return.

A representative portfolio for the strategy saw 125 basis points of annualised outperformance of the broad universe between 2007 and 2013, with similar volatility and a maximum drawdown 175bps smaller.

Turnover is 14.89% versus 4.54% for the investment benchmark.

At the moment, Ossiam is offering this as a systematic active strategy, but head of business development Isabelle Bourcier told IPE its entirely rules-based nature made it possible to adapt it for index construction.

While recent years have seen the launch of a host of smart-beta equity products, the indexing and asset management industry has so far been less inclined to develop fixed income equivalents, despite interest from institutional investors.