By continuing on this site you have agreed to cookies being placed and accessed by this website. More information and adjusting cookie settings.

Robeco uses cookies to analyze your visit to this site, to share information via social media and to personalize the site and advertisements in line with your own preferences. By clicking on agree or by continuing on this site, you agree to the above. More information and adjusting cookie settings.

AGREE

Robeco uses cookies to analyze your visit to this site, to share information via social media and to personalize the site and advertisements in line with your own preferences. By clicking on agree or by continuing on this site, you agree to the above. More information and adjusting cookie settings.

AGREE

By continuing on this site you have agreed to cookies being placed and accessed by this website. More information and adjusting cookie settings.

Making better credit risk assessments

06-01-2014 | Interview | Patrick Houweling, PhD

Ground-breaking research by Robeco that changed the way the riskiness of corporate bonds can be evaluated has celebrated its 10th anniversary. This riskiness needs to be carefully calculated as bonds issued by companies have a greater chance of defaulting than government bonds. Their returns can also be more volatile, as they are linked to the underlying performance of the company that issues the bond.

In December 2003, Robeco’s credit and quantitative strategies teams developed a model to better evaluate both risks and returns of corporate bonds. The model links the return volatility directly to current credit spreads – the difference between the bond’s yield and that of a super-safe sovereign bond - rather than to historical returns. The model can therefore react instantly to market conditions rather than having to wait years to collect data.
 
In this question and answer session, Patrick Houweling, quantitative researcher and portfolio manager, explains how it works.
 
What was the aim of the research on credit risks?
Ten years ago we wanted to find a better way to evaluate the risk that is inherent in corporate bonds. A traditional method of measuring risk is to use ‘moving windows’ that assess risk and return periods retrospectively, and so they react much more slowly - sometimes even one or two years later. Subsequently, they underestimate risk in times of crisis. Also, when markets rebound, traditional models lag behind and subsequently overestimate risk by using data that is related to the previous crisis rather than the recovery. Our aim was to find a way to more accurately and timely measure the volatility of corporate bonds.
 
How did the research add value?
We were the first to discover that credit spreads move in a relative way rather than an absolute way. This means that a bond with a higher credit spread is not only more prone to default, as was already known, but also that its price reacts more strongly to changes in market sentiment. Investors were used to measuring the risks of corporate bonds like they did for government bonds – the longer the maturity, the longer the risk. We add the credit spread to this picture – the bigger the spread, the bigger the risk. 
 
How does the Robeco risk model work?
The risk model gives more accurate and timelier risk estimates. When a crisis such as the credit crunch of 2008 hits, our risk model uses real-time data and reacts instantaneously. Likewise, our risk model responds quickly to the rebound. This is illustrated in the chart below.

making-better-credit-risk-assessments.jpg

How do we use the model at Robeco?
The research was implemented at Robeco soon after it was discovered in 2003 and has been used ever since. Initially it was only used as a risk model, but gradually it was applied to all aspects of portfolio management. In every aspect, the model gives us a competitive edge, not only in risk measurement, but also in portfolio construction, relative valuation and performance attribution. With the concept of relative credit spread changes, we can make and implement superior company selection and market timing decisions.
 
Was the impact of the Robeco research for the industry?
A few years after our discovery, we shared our findings with researchers from Lehman Brothers, which later became part of Barclays. They validated our research and distributed it among their clients. In 2007 we published a joint article in the Journal of Portfolio Management. The research was soon put to good use. In 2008, Barclays implemented the methodology in its POINT software. Currently, about 40% of Barclays’ worldwide client base uses the relative spreads methodology originally developed at Robeco.
 
So how does it work within a portfolio?
We can calculate the risk that individual bond positions contribute to a portfolio by using what we call ‘risk points’. These focus on the credit spread of the bond and its duration. By multiplying the credit spread and duration, we can calculate the bond’s likely levels of volatility. It does not matter how the result of that calculation comes about, as long as the multiplication of credit spread and duration gives the same number. This makes the model very easy to use in practice.
 
Can you give an example of the usage of these risk points?
A bond with 1-year duration and a credit spread of 500 basis points has the same expected volatility as a bond with 5-year duration and a credit spread of 100 basis points or a bond with 10-year duration and a 50 basis point credit spread. They all have 500 risk points and are thus equally risky. This is quite remarkable, since the former bond is a short-maturity bond of a risky company, whereas the latter one is a long-maturity bond of a safe company. Nonetheless, these bonds have the same risk.
 
Does the use of risk points uncover unseen problems?
Yes, because unless you use risk points, the risk of bonds with a high credit spread is underestimated, while the risk of bonds with a lower credit spread is overestimated. This can be seen when a credit portfolio is either underweight or overweight relative to the benchmark. Being underweight means the portfolio owns less of a bond than the benchmark. If this calculation is not made using risk points, it would be possible to be underweight in terms of market value, but overweight in terms of risk contribution. This ‘unseen’ riskiness is not brought out by looking at traditional market values, but is revealed by risk points. 
 
How has this worked in practice since the credit crunch?
The use of risk points is particularly useful when transiting from crisis to calm periods, as was seen when we came out of the credit crisis into a more stable period. By using risk points rather than relying on historical data, a position of 1,000 risk points in the midst of a crisis is comparable to the same position in a calm period. This is a clear advantage of risk points over traditional models, as it avoids overly risky portfolios when crisis periods turn into calm ones, or vice versa.
 
Can it also find market anomalies, or bargains?
Yes, because relative valuation is all about relating returns to risk. Investors want to be compensated with higher returns for taking a higher risk. Robeco’s risk model allows us to identify sweet spots in the market, and to properly demand higher return for riskier bonds. By comparing realized returns to expected volatilities, we can monitor in real time which bonds, sectors, ratings, regions, etc. have outperformed and which ones have lagged the market. This helps us to optimally position the portfolio for future performance.
 
How are risk points used to evaluate performance?
Risk points tell you where your risk comes from, but for performance attribution, the main goal is to evaluate portfolio management decisions by measuring the returns that they generated. As positions are measured in risk points, our performance attribution is also based on the risk point framework. This allows us to attribute the entire performance of the portfolio to individual decisions. This method provides greater transparency for investors, as it shows how the portfolio management team has added the value.

Share this page:

Author

Patrick Houweling, PhD
Portfolio Manager, Senior Quantitative Researcher