ESG point-in-time (PiT) content helps quants looking to integrate sustainable investing into their traditional financial strategies. Without ESG PiT you will end up with lookahead bias or overly conservative lags, and either scenario will misrepresent your alpha.
- Combining ESG factors with traditional financial factors is a potential source of alpha.
- Including point-in-time ESG data in your quant studies more accurately represents real-world conditions and can yield better backtest results.
- Existing smart money quant signals should start incorporating ESG factors because institutional funds are putting more emphasis on sustainable investing.
For more data-driven insights in your Inbox, subscribe to the Refinitiv Perspectives weekly newsletter.
PiT data includes all reported values in the order they were published, timestamped to the moment they were made available to the market. Original values are not overwritten by reclassifications and restatements, and the data gives a more realistic view of history, and we covered its utility in a previous blog.
It would be an invaluable tool in the fast-growing world of sustainable finance, but there’s a problem: many investors only use ESG data in a rudimentary fashion, taking the form of negative screening or tilting. Point-in-time data is much less commonplace.
At StarMine Research, we are concluding research on our first quantitative finance model that marries traditional financial factors with ESG factors in a sophisticated algorithmic way.
When our research began, we didn’t have access to ESG PiT data, so our quant studies suffered from a lack of restatements. We also used conservative lags, an entire year in this case, when simulating the availability of ESG data to avoid lookahead bias (the horrible sin in quant modelling of using data before it was actually available).
When ESG PiT was released over RDP Bulk, we integrated it into our research and saw cumulative spread performance improve by over 7 percent now that we were more accurately modelling real-world conditions.
ESG with traditional financial factors and behavioural finance
ESG Smart Holdings builds upon the behavioural finance-based StarMine Smart Holdings model, which has been live for over a decade.
The heart of the model comes from creating a factor preference profile for every active mutual fund and hedge fund, over a certain minimum AUM threshold.
Looking at 25 traditional financial factors across several categories (value, growth, profitability, leverage, price momentum, analyst revisions, volume), we determine which factors matter most to each smart money owner based on their recent purchases over the past three months.
Then, for all publicly traded stocks, each owner profile will ‘vote’ on how attractive that stock is based on that stock’s current factor attributes and recent changes to those factor values.
The stocks with the most ‘votes’ will get the highest ranks and will represent which stocks the smart money will likely buy in the near future.
General market participants tend to follow the smart money trends. Hence, StarMine Smart Holdings presents an opportunity to be right on the cusp of changing smart money factor preferences.
ESG Smart Holdings adds to the mix 13 sustainable finance factors such as CO2/Revenue, % Women Managers and Management Category Score.
All 13 factors were chosen from Refinitiv’s ESG content set through a rigorous quantitative screening process overlayed with hypothesis-driven factor selection, then narrowed down to the most impactful ESG factors after several rounds of experiments and backtests.
The result is an adaptive behavioural finance model that picks up on the hottest sustainable finance trends among the smart money investors, while still keeping tabs on the hottest traditional financial factors.
Never skimp on point-in-time data
The importance of point-in-time content to the quantitative research process should, hopefully, be well understood by quant investors.
For example, thinking about fundamental financial data, a company’s revenue is immensely important to the calculations of ratios used in value, growth and profitability factors.
A company reports their revenue after each fiscal period ends, but may later revise that revenue number in a subsequent filing. If you’re not using point-in-time data, you may just have one value for that fiscal period, which would be quite inaccurate. If you use just the first reported value, you never get the benefit of the restated revenue value.
And if you only use the most recently reported value, you’ll have a lookahead bias in your revenue as your backtests consider data which wasn’t known until later on.
The same importance of point-in-time content is true for sustainable finance data since the Refinitiv ESG content is also sourced from company-reported filings.
As an illustration of the importance of point-in-time, let’s consider CO2 Equivalents Emissions, which is probably one of the most used sustainable financial factors today.
Pfizer reported 2,005,730 CO2 emissions on 8/27/2021, about nine months after the PeriodEndDate (not a year after).
However, later on, 10/1/2021, the company restated its CO2 emissions to 1,350,000. These types of restatement are commonplace.
This as-was history of how its CO2 changed is crucially important to capture, as it will affect the relative attractiveness of the stock to smart money investors who care about CO2 emissions.
Whereas without PiT, you only see the most recently restated value.
Similarly, the Total CO2 Emissions / Revenue analytic also adjusts after the change to CO2.
Code examples for integrating ESG PiT into AWS
As any data engineer or data-savvy quant can tell, point-in-time content is a bit more challenging to work with because of all the as-was date considerations.
However, the effort is worth it when you gain confidence that the alpha in your backtests comes from more accurately modelled as-was real-world conditions.
Having recently gone through the process of integrating ESG and then integrating ESG PiT into the ESG Smart Holdings research, I’m happy to share with you my approach and actual code sample used.
Please dive into this Refinitiv Developer Community to learn step-by-step how I integrated the files downloaded from RDP Bulk into the AWS ecosystem, where our quant research is conducted at scale.