Julien Turc, head of the QIS Lab at BNP Paribas, talked to us about systematic indices and his team's new research paper on factor investing for government bonds.
An increasing number of investors are moving towards systematic strategies, including in fixed income.
Hi Julien, and thanks for taking the time to talk to us. First off, what is QIS and what does it do?In a nutshell, what QIS – quantitative investment strategies – does is to set up investment strategies and make them accessible to a broad range of investor clients.
We are part of BNP Paribas CIB, and we launch and manage investable indices which track the performance of trading algorithms in all the different asset classes.
What are systematic indices?A systematic index includes decision rules set up for an investment strategy. While previously these rules had to be enforced manually, we now have systems in place that update the performance of these strategies automatically, following transparent, predefined rules, and which include replication fees - fees for making the index investable. The task of following a systematic strategy is now run on a computer system, which reduces risk of human error around execution and gives the client peace of mind.
How different are these products from what came before?The trend towards QIS is part of the movement away from active investment management, where a manager might have full discretion to make the investment decision he or she wants. At the other end of the spectrum you have for example exchange-traded funds, where a large investor might replicate the market portfolio and the only way to outperform is by reducing costs – so you are entirely passive.
There's a whole gamut of possibilities between one extreme and the other, and you could say QIS sits right in the middle. It's passive in the sense that no human being is directly involved once a trading rule is agreed on. But trading decisions arguably involve a lot of intelligence regarding the data they use and how the algorithm was built, and they can involve fairly active rebalancing of positions. In this way, QIS is a very active form of investing.
What might be the benefits for investor clients?The first thing is cost. Institutional investors feel that by using algorithms and a trading platform, they can directly access strategies that used to be accessible only via expensive intermediaries.
" A revolution is taking place in the fixed income market. "Another question is transparency: if you entrust your money to an asset manager, you can never be entirely sure what decisions they will take. Moreover, some fund managers come and go, others retire. With indices driven by algorithms, you're 100% sure they are going to follow what's been agreed.
A third point concerns rationalisation: once you know exactly what's in a strategy, you can try to anticipate the future behaviour of a fund and to understand and quantify its risk and return features, which some portfolio managers find useful in the construction of their global portfolio.
There are other issues too: for example, our strategies are most directly accessed via swaps. In some countries, investors find it easier to use this kind of derivative contract than to put money directly into a fund.
How do you design the indices, how are they tested? Are there situations where you would override the algorithm?As far as possible, we try to launch strategies which have both academic and economic rationale and to quantify the expected outcome, doing our utmost to ensure we have the best possible understanding of how indices are likely to perform in the future.
More specifically, before launch, all indices must go through an investment committee involving internal experts – including Legal, Compliance and investment specialists – who investigate whether a strategy makes sense. A stress test is also carried out, which simulates changes based on financial market movements and tests the various parameters involved in the strategy.
All the testing is carried out beforehand: once an index is launched, there is no human intervention involved and we would never override an algorithm – though we might decide to launch a new, improved version.
Do clients just use the indices as they are, or do they adjust them?Some of our clients take off-the-shelf strategies, but many want to discuss how to tailor them to meet their particular needs – whether for their global portfolios, or specific objectives in terms of risk and return or macroeconomic views.
Turning to your job now, you head up the QIS Lab. What's that?The QIS Lab is a research team BNP Paribas set up within QIS. My colleague Jerome Gava and I are also visiting researchers at the economics department of France's Ecole Polytechnique. We hired a PhD student who is supervised by Huyen Pham, a professor at Paris 7 University.
We divide our time between, on the one hand, our clients and our colleagues, and on the other, conducting academic research.
We have recently published our first academic paper, and we hope it will serve to reach out to final investors – mainly pension funds and some sovereign wealth funds – which tend to rely heavily on academic research for their investment decisions.
Tell us more about your research paper, 'Beyond carry and momentum in government bonds'.We've been looking at sources of returns in government bonds. The "factors" we have identified as potentially driving returns can be achieved by buying and selling futures across different countries in such a way that portfolios would not be impacted overall by global interest rate movements. In classical microeconomic theory, you wouldn't expect such factors to be achievable, as market returns are either country-specific and idiosyncratic or driven by global market changes. Factor investing tries to identify something that sits in between the two.
Factor-based investing has generated a lot of interest in the stock market, where it's now commonly accepted. It made its way into the fixed income market a couple of years ago: a number of large asset managers published research studies on the topic, claiming that many of the factors used in the stock market could also be applied to government bonds.
Our paper revisits that question. We generally share the view that there are factors that sit in between global market changes and idiosyncratic movements, and we've tried to examine the rationale behind these factors along with the empirical evidence.