Trading jobs in finance: Should you study machine learning or economics?

If you want a trading job in financial services you might be wondering whether to study econometrics or the growing field of machine learning. A new paper offers some pointers. 

Click here to sign up to our technology newsletter 🔧

Joseph Simonian, founder of AI quant consultancy Autonomous Investment Technologies, published a research paper this week which looks at the key differences between econometrics and machine learning. Ultimately, it concludes that the former is best at explaining the past, while the latter is best at predicting the future.

The paper says a key benefit of econometrics is that it allows researchers “to understand the relationship between variables and make meaningful economic inferences.” Econometrics has more of a ‘data modelling culture’, in which data is used to provide “information about the relationships that exist between input and response variables.”

This approach can be an issue when predicting the future because causal relationships are rarely linear. External stimuli, both visible and invisible, will constantly be acting on whichever variables you focus on. Variables are manually selected, and thus potentially important variables can be overlooked.

Machine learning models, meanwhile, are “highly flexible”, and can understand the complex causal relationships hidden in data. They are also robust enough to handle missing data, whereas econometric models would need “explicit treatment” for it, potentially introducing biases. The key issue with machine learning models is that it’s not often clear how they reach a solution, even when that solution is right.

If you want a job at the smaller, high paying hedge funds and trading firms, you’d likely want to prioritize machine learning. These companies tend to be private, and thus less susceptible to public scrutiny and regulation. Because of this, they prioritize results, which machine learning does better.

Trading jobs in banks and asset managers, companies with clients, are different. Simonian previously said on a panel for the CFA Institute that these companies have a “very strong need to explain their products to their client.” In those cases, using econometrics in your models will make them much easier to explain than a ‘magic’ AI algorithm.

Have a confidential story, tip, or comment you’d like to share? Contact: +44 7537 182250 (SMS, WhatsApp or voicemail). Telegram: @SarahButcher. Click here to fill in our anonymous form, or email editortips@efinancialcareers.com. Signal also available.

Bear with us if you leave a comment at the bottom of this article: all our comments are moderated by human beings. Sometimes these humans might be asleep, or away from their desks, so it may take a while for your comment to appear. Eventually it will – unless it’s offensive or libelous (in which case it won’t.)