Flexibility – the key to optimized fixed income attribution
04 June 2018
When it comes to performance attribution, complexity is a fact of life – both in terms of the approach, and the volumes of data it needs to accommodate. The investment process for fixed income managers is typically multilayered, and investment managers need attribution models with the flexibility to match this process. Meanwhile, asset owners need to validate and understand performance – attribution analysis is critical in determining if the fixed income manager’s stated alpha sources are borne out in reality, and if the investment process is robust and repeatable. Asset managers therefore need to deliver reporting that helps the asset owner meet these objectives.
RiskFirst’s Charles Glah looks further at the issues facing asset managers and asset owners, and how they can best overcome them.
Historically, investment managers have adopted simplified models for client reporting purposes. Yet, while the middle office has opted for a standard and scalable approach to attribution that works across its client base, one-size-fits-all approaches rarely work for the front office. Instead, they require a model that explains in detail the sources of a manager’s active return. Every portfolio manager is different; with a multilayered and complex process, attribution can be challenging to implement and often plagued by large, unexplained residual returns.
With this in mind, it is no wonder that there is such a divergence in the tools used by the front and middle office. Inevitably, however, this leads to an added headache of reconciling the differences between the disparate approaches.
Yet, while it may seem that there is a conflict in the underlying needs of a portfolio manager and asset owner, in fact they are perfectly aligned – and therefore a single flexible attribution tool can (and should) deliver results to both audiences.
Why is fixed income attribution so complicated?
Firstly, let’s examine the attribution landscape itself and what makes attribution inherently complex. I’d argue there are four reasons: the number of investment strategies, the number of attribution models, the number of security types and the diverse target audience.
- The number of investment strategies
The first challenge is the sheer number of investment strategies available to a portfolio manager. For example, take a portfolio that is hedged against movements in FX rates. Hedging in this manner comes at a price, which is typically the difference in interest rates between the portfolio’s base currency and the currency to be hedged.
Should the return generated by these forward premiums be classed as interest holdings, or as carry return, or as curve return, or placed into a separate category entirely? There is no one correct answer; it depends on the investment process. Sometimes, hedging is run as part of a stand-alone portfolio; sometimes, it’s done over all non-domestic holdings, at a much later point in the investment process.
For this reason, it is important that returns from hedges be allocated to the right part of the investment process.
Consider another issue: should we include returns due to convexity in an attribution report? Convexity just measures non-linearity of return against yield change. All else being equal, a bond with high convexity will outperform a bond with identical modified duration when yields change.
If the strategy is simply to be long (or short) against a benchmark, then probably not. Convexity has played no part in the investment process, and to report on this source of return in isolation is at best misleading.
On the other hand, if the investor has followed a strategy that relies on convexity (such as taking a barbell allocation), then it is proper and correct to report on convexity returns, so that the investor can verify that the return made by convexity outweighs any losses made elsewhere. It all depends on the investment strategy.
- The number of attribution models
With so many possible investment strategies, the industry has responded by creating a number of fixed-income performance attribution models. These models generally fall into three categories: sector based, factor based, and hybrid models. Sector-based attribution, a variation of the Brinson model, decomposes active return into bond market sectors and security selection within each sector, with securities weighted by duration. Factor-based models explain excess returns based on systematic and non-systematic risk and return sources. Hybrid models, as the name suggests, combine sector-based and factor-based attribution models. In practice, fixed-income attribution vendors offer an array of model variations, and many managers have their own in-house, customized attribution models as well.
An asset owner will typically have numerous investments following a range of investment processes – and having each manager report attribution differently presents a significant challenge!
- The number of security types
The third challenge is the number of security types traded in the marketplace. Most fixed income portfolios are driven off changes in the yield curve, but even defining what a parallel change in the curve means is still an open question. Inflation-linked securities will be driven by the level of indexation to a reference index; mortgage-backed securities by the prepayment rate of the underlying security pool; and so on. There is no reason to expect that the pace and number of new security types will wane, and this requirement is unlikely to disappear.
There are certain instruments that have investment characteristics that can further complicate the analysis. For example, US bond futures have embedded options, including quality, timing and the switching of deliverable bonds after the future’s last trading date.
It is difficult to unpack the effects of these options, and their presence can mean that the price of the future changes even when the price of the underlying remains unchanged. This will translate into a significant residual where bond futures form a significant part of the portfolio.
- The target audience
Given the insight that attribution provides into both a particular portfolio and an institution’s strategy, wide interest should be expected in any attribution analysis.
Clearly not all audiences have the same level of expertise – a trustee of a pension plan may not want or need to drill into the carry effect in the same way as the portfolio manager. The market has often misinterpreted the different requirements of different audiences. Just because an asset owner does not need to see every attribution effect to the nth degree, does not necessarily mean that they want a “simple” attribution model.
Rather, more care should be taken in how the attribution results are presented to the various audiences. Drill down reports, enabling different users to stop at their desired level of transparency, are useful. Equally, reports that bring key information and trends to the fore, without the need to work through pages of data, can be valuable to all audiences.
Let’s now explore the different requirements of asset owners and managers in more detail.
What do asset owners want?
Asset owners need to understand, challenge and validate their fixed income managers’ performance over time. What is driving performance, and are these drivers consistent with the investment process prescribed? Is the process repeatable? Is the process delivering alpha based on the philosophy of the fixed income manager? This information all needs to be translated in a concise and understandable way.
In summary, asset owners need attribution analysis that is flexible enough to fully describe their managers’ investment process. The alpha should be explained by the investment process, and attribution. When done well, this will give the asset owner insight into this process and allow them to ensure that the manager is managing money in a manner consistent with why they were hired. For instance, if they hire a high-yield investment manager whose investment process is based on bottom-up fundamental analysis, the asset owner will want to see security selection being the key driver of manager performance.
What do fixed income managers want?
Managers should use attribution analysis in three ways: to provide feedback on the investment decisions they have made, for client reporting – demonstrating the investment process and explaining performance – and, finally, as a sales tool. Undoubtedly, attribution analysis that is tailored to the investment process and clearly illustrates consistent, repeatable outperformance aligned to the manager’s strategy, is the perfect sales aid.
To deliver these benefits it is essential that the attribution analysis matches the investment process. That is a challenge because fixed income managers make more decisions and contend with more risk factors than equity managers. Fixed income managers may take positions based on their views on interest rates, yield curves, credit, quality and securitized sector spreads, inflation, and idiosyncratic risk, among others. A representation of performance attribution taking all these factors into account requires a hugely flexible solution and the ability to show attribution results through many different “factor lenses”.
Flexible attribution, tailored to the investment process, is the solution
Asset owners and asset managers fundamentally need the same thing. Managers need to understand and sell their investment process. Asset owners need to hold managers (and their investment process) to account. Given the complexity of fixed income attribution, flexibility of attribution model to match the investment process is key.
Taking this one stage further, why not look at trends in attribution through time – how are my attribution effects evolving? Again, we see good alignment between the managers and asset owners in this area. Asset owners are typically concerned with performance over the long-term. Managers can use trend analysis to demonstrate their ability to consistently outperform, using a robust investment process. Trend analysis can capture drifts in investment philosophy and act as an early warning signal for all stakeholders.
All of this should be achievable without rivaling the complexity of the controls in an airliner’s cockpit (as some do). Albert Einstein famously said “Make it as simple as possible, but no simpler”. Attribution systems are going to remain complex. And so they should, for they will be worthless otherwise. As with all other areas, complexity can be tamed. The best systems will combine simple, elegant design with flexible and powerful functionality.
This article was featured in TSAM Insights.
Another article of interest on this topic can be found here.