We asked Chairman of the Board of Inquire Europe Hans Fahlin to provide insight into some of the most pressing themes within quant investing today.
What are the factors that are making the current market environment so challenging for investors?
Uncertainty is very high. In a way, every forecast on an economic variable or the return on a financial variable will probably have the largest distribution of outcomes since the Great Financial Crisis (GFC).
If investors structure their portfolios taking the available history of financial data into account, there will be periods that are similar to the current environment; at least for developed markets there is a lot of data, dating very far back. Superficially some may say that there are similarities to the 1970’s, but if you look closer at central bank policy and energy dependence, the current market environment is very different.
When using long term data to construct a portfolio, there will be situations like this in your data set, and the likelihood of getting acceptable outcomes is high. Investors must not base their decisions on short-term data sets, particularly because the time period since the GFC has been exceptional and therefore not representative.
I had expected inflation to rise – but had not anticipated it would happen with this speed or to this extent. But inflation before the pandemic was unusually low. Higher interest rates should have been part of the assumption built into portfolio design the last few years – or at least it should have been taken into account as a possibility.
There is speculation that the value factor is finally back in favour. What is your position on this?
The value factor had not been doing well for an unusually long time. When considering a very long history of value this period was an unusually long period for the value factor to underperform, versus the growth factor. It became noticeably worse after the spring of 2020.
A lot of it was influenced by the valuation of growth stocks. Value stocks were normally priced, but growth stocks were not – they were historically very expensive. The valuation discrepancy was abnormally high. People had to make an assessment as to whether the value factor was dead, or simply dormant. Looking into the historical properties of the value factor, my conclusion was that the period of underperformance was odd, but not unreasonable, and certainly not to the point of discarding the legitimacy of the historical observations that prove that over time, the value factor does outperform.
What is now apparent is that we have the confirmation that it was indeed an exceptional situation and things are normalizing. The Gordon dividend discount model which says that when real interest rates rise, stocks that have more of their dividends long term in the future should cheapen more than stocks that have a more uniform distribution of dividends. From that point of view, you could claim that the rise of interest rates is the driver of the decline of growth stocks and that would explain the resurgence of value.
The relationship between nominal interest rates and the return of growth compared to value is not really easy to piece out from the data. Event though we have a dividend model that says that this relationship should be there, it’s hard to pin it down. The work I have been involved with hasn’t been very successful in establishing a strong relationship between interest rates and value factor returns. It seems intuitive that the reason growth stocks are now underperforming is that interest rates are higher and investors are now discounting their future growth rates. But this isn’t particularly easy to prove econometrically.
How important will alternative data be for quant investors in the future?
There is great demand for alternative data in quant investing. We have had a technological change that has made a lot of data accessible, that we didn’t have before, and it is more abundant. Even if methodologies are not very new the computing power that everyone has now makes it possible to use methods that were unavailable previously. Methodologies must now adapt to include computational methods that can leverage this input to its maximum potential. Techniques that were invented in the past can be refined to harness the power of massive amounts of data.
There are two developments that have been happening in the past couple of years: new techniques and new datasets. It’s very tempting. Everyone who is doing quant investing is very aware that there are lots of other people doing it too. And they are torturing the same data sets, so we welcome these developments. There is much to gain.