Jos Gijsbers: Yes, but it is statistical evidence, so it isn’t an answer with 100 per cent certainty.
We don’t do this research ourselves, but we do see a lot of research coming to the market from academics, asset managers or other studies which show that certain factors do add value in a risk-adjusted way in building an investment portfolio.
You can add value by getting higher returns using factor investments, but typically your risk profile also increases, so the return per unit of risk should improve.
You need a long time series of data to analyse and see how certain factors have behaved and how this performance would be better or worse than in a standard investment portfolio.
"We do see a lot of research coming to the market from academics, asset managers or other studies which show that certain factors do add value"
However, not all factors will contribute to improving the risk/return profile of a portfolio at the same moment. It also needs a lot of data mining over a long time series in order to be analysed.
There is enough data available on a number of traditional factors, but it is a more difficult to get enough data on the newer factors to do the right statistical analysis that leads to scientific evidence.
I have also seen a lot of new studies available on ESG factors, which is a trend that has emerged in the past 10 years. Although this period limits the availability of data to analyse, 90 per cent of these studies show that ESG factors can add value.
Jos: You will always need to have a risk management framework in place with specific budgets that fit with the decisions and tilts that you have within your investment portfolio.
For factor investing, most of the factors will behave differently in different market conditions. So, in a multi-factor investment portfolio, you should add or reduce risk on certain factors depending on these market conditions.
"These budgets should be defined in a way that they cover all
situations in the market for the overall portfolio."
This could also imply that in certain situations you will have more exposure to some factors and zero exposure to others. This is where your risk budgeting comes into play for these separate factors.
You will need an overall risk budget that fits within the guidelines of the portfolio, because risk is not a constant in the market and can also be volatile.
So, these budgets should be defined in a way that they cover all situations in the market for the overall portfolio but also in separate risk budgets for factor exposure.
Jos: Combining factors gives you the possibility to enhance returns or to mitigate risks in different market circumstances. You can also see this in time series, when certain factors have performed very well or have generated negative returns in other periods.
When you have the capabilities in your own team to take a tactical decision on what factor exposure will do best, then you can maximise returns by tilting your investment portfolio to this factor exposure and reducing or eliminating the other factors that won’t deliver.
"Combining factors gives you the possibility to enhance returns
or to mitigate risks in different market circumstances."
There is also a debate about whether you should just let the multi-factors do their work, as over the long run they will add value and that is why the strategy is defined by certain factor exposures.
However, using multi-factors means that you will be exposed to factors at certain times when they will not deliver.
If you can do an active overlay when taking technical decisions on the right factor exposure at any given moment, then you can profit more from using a multifactor model in your investment portfolio.
Jos: Depending on the market conditions you will put more money on certain factors and less on others.
Unfortunately, how you measure the results isn’t straightforward because factors are correlated to some extent and it may not always be clear as to where your return is coming from.
A very thorough attribution of your factors is difficult, and models are still being developed for investing according to these factors within a multifactor model as well as having ex-post analysis.
"How you measure the results isn’t straightforward because
factors are correlated to some extent"
I expect that techniques such as AI (Artifical Intelligence) will help us to get a better understanding in this area.
Of course, this is difficult to explain to an investment committee or board because they prefer to see exactly what decisions have been taken and what the outcome is.
It is also more of a statistical outcome, so you can explain it to a certain extent but not in black and white. So, you will need a longer time series to show the evidence that it has added to the portfolio.
Jos: Risk budgeting is applied by portfolio managers to create the best portfolio with the risk factors that are being used in the portfolio construction process. This deployment of the risk budget needs to be explained to the rest of the organisation.
On the risk management side, risk budgeting follows the investment process and we need a budget to hire people who can understand this and also to have systems in place to track it.
These systems are not off-the-shelf products, so a lot of work needs to be done. We have carried this work out ourselves with software building using Python, which helps us to analyse the data, optimise portfolios and also get the right risk analysis in place.
"It isn’t just a number that you are pulling out of Bloomberg,
you need to do a lot of work yourself."
It isn’t just a number that you are pulling out of Bloomberg, you need to do a lot of work yourself.
Factor investing comes with specific risk budgets within the internal risk framework and guidelines. It also requires the involvement of people who can understand the systems and to use them for analysis and reporting, including management reporting.
Python software is continuously under construction, because we are still learning how to use data and how to get feedback on the outcomes in order to improve or optimise the investment process. We hope that we can develop more AI techniques, which will help improve this process.
