Bristol Gate Capital Partners is a Canadian-based company that has machine learning in its DNA. Hamm and his team have concluded there is a strong correlation with returns if the focus is on dividend growth, and this focus inherently lowers the risk of the portfolio. To accomplish this with a certain level of predictability, Bristol Capital uses machine learning to identify the companies that pay (or will pay) the highest dividends.

After applying machine learning algorithms, humans perform classic fundamental analysis to handpick from shortlisted investment options. As Hamm explained, they’ve used this method since the founding of Bristol Gate, and the reason is that man and machine work better together.

“This has been proven in the case of when Garry Kasparov [former world chess champion] lost to the IBM machine, but Kasparov and the machine [together] beat the machine,” Hamm says.

Today, we can’t imagine life without machines; the only problem is, machines don’t think. They can do the work of a thousand people, and, in the case of Bristol Gate machine learning can sift through the thousand features that go into their model to pinpoint the ones that are relevant for a particular strategy. Bristol Gate’s US Equity Strategy is dependent on seven or eight dominant factors and 200 features.

In an interview, Leyla Imanirad, senior research associate at Bristol Gate, said the following:

“In many quant (or smart-beta) strategies, there is very little human involvement. As a portfolio manager, you decide on a factor driving a strategy and the machine makes the decisions based on each stock’s factor exposure. For us, predicting dividend growth is a first step in our man and machine approach, meaning, we simply use this step to build a focus list of [the] top 50–60 securities that we want our fundamental team to analyze. This allows them to focus their efforts on a smaller set of securities, saving time in the analysis step.”

To prove that their marriage of data science and fundamental analysis works in emerging markets, Hamm looked no further than his home country, Canada. Despite Canada being in the top 10 countries by gross domestic product, he explained that it displays the characteristic signs of an emerging market: it is narrow, focused on resources and banks, and lacking depth.

“This is where machine learning and thinking about what you’re trying to get out of a market index, what’s in that market index that’s reliably better than an index as a whole, really works,” Hamm says.

The Bristol Gate approach to machine learning

To provide human analysts with the best possible starting point for choosing investment options, Bristol Gate uses gradient-boosting machine technology. The objective of this technology is tree-based algorithms. Essentially, it iteratively evaluates thousands of factors to arrive at a final prediction, and then averages the collective predictions from many trees. The machine then decides on the factors that matter the most without human interference or bias.