This investor uses search data, patents, and earnings calls to unearth the world's best small caps

There are lots of reasons why investors love digging around in the small end of the stock market. But one of the biggest ones, especially for professional investors, is the advantages you can glean in an often mispriced area of the stock market. There are numerous fundamental and technical ways to harness that advantage - but this wire is about the third way to outperform as an investor: through quantitative analysis.

Recently, I had the privilege of sitting down with someone who knows quantitative investing inside and out - Dennis Walsh, Global Co-Head of Quantitative Investment Strategies (QIS) at Goldman Sachs Asset Management.

In this interview, Walsh shares how 'big data' investing has changed over the last few years, the data points he and his team use to identify great opportunities, as well as a high conviction opportunity in which the team has invested based on the data they have collated.

Big data, big reward potential

Walsh joined Goldman Sachs in 2005 and has worked on the QIS team since 2009. When I asked him why big data was so important to his investing process, he said it all comes down to efficiency.

"It's a model-driven approach ultimately governed by hundreds of investment signals that are ingesting over a trillion data points and ultimately creating forecasts for over 15,000 stocks globally. You can't effectively cover that breadth of an investment universe without being highly dependent on data," argued Walsh.

He also says big data has allowed the team to break down the analysis of thousands of stocks into bearable chunks.

"I think if you're able to decompose various parts of the market into specific buckets that will allow you to understand what are the trends driving stock returns at any given time and, how those trends are related," Walsh said.

For his part, Walsh and his team use proprietary and publicly available tools to narrow down the investment opportunity among 7,500 global small caps into less than 750 actual holdings. Traditionally, this process would have involved hundreds of analysts being on the ground and stationed all over the world. Now, all the key information on companies can be process in real-time with less resources.

Additionally, the quant "toolkit" (as Walsh describes it) also helps add a layer of research in a part of the market that itself is already under researched.

"The average analyst coverage for a small-cap stock is only seven or so compared to 25 for a large or mid-cap company that you'd find in a benchmark like the MSCI World," Walsh pointed out. 

"So one of the things we're able to do is glean insights from the large-cap segment of the market and have an understanding of how that may impact small-cap pricing," Walsh added.

A deep dive into the changing face of data

In the 15 years since Walsh joined Goldman Sachs, the biggest change he has noticed is the type of data that is being analysed. At first, all the data was structured and mirrored the kind of information that can be crunched by all investors.

"Financial statements for a company or market data points- generally use very simplistic tabular or panel data that didn't require a level of sophistication to synthesise information from that data. If you wanted to know the book-to-price value of a company, you would simply take its book value from its financial statement and divide it by the current market cap of the company. It was very straightforward," Walsh recalled.

Now, the data is getting far more complex, and even non-numerical data like management sentiment and audio transcripts are being added to the investment process.

"You're consuming textual data, you're consuming news articles, regulatory filings, [Form] 10-Ks, and earnings call transcripts," he said.

Another set of data points that the team uses include consumer behaviour trends, seen in things like website and search traffic (for online retail firms), and geolocation data (particularly in brick-and-mortar stores).

Even more obscurely, the team analyses the number and quality of patents coming through as a tool to determine which ones can genuinely be earnings-accretive.

"95% of patents are worthless and so, it's a really highly skewed distribution where a few patents in the grand scheme of things will dictate value," he said.

Walsh expects this trend to continue and become even more sophisticated as technology evolves, the data becomes more granular, and the pressure to find novel and tangible insights grows ever more pressing.

The influence of AI

With artificial intelligence vastly changing the way all of us work, I wanted to find out how AI is used by a quant equity team. Walsh described to me an example of how AI can help analyse how management really feels about future earnings prospects.

“Think about the CFO or CEO of a company. Company management has a much better sense of the prospects of their company than the market does. So, if you can properly infer whether management is bullish or bearish, you can outperform the market," Walsh said.

"It's a very straightforward intuition, but in practice, it's difficult to actually get a true sense of management's actual view on any particular time," he added.

How all of this data become investable ideas

One sector where the team have turned data into tangible stock picks is the GLP-1s space.

"Some of the data sets have helped us get an understanding of where that global trend is going and who the relative winners and losers potentially will be. That's an area where we've seen a lot of negativity when you read between the lines, from companies that otherwise are trying to pretend like it's not a going to have a big impact on their particular market," he said.

Another sector where the team has used data to their advantage stems directly from the AI trend. While most of the major players like data centre companies, and chip manufacturers are known, lesser-known electrical equipment companies also hold a lot of the IP to make this megatrend possible.

The one question they have yet to solve...

I closed the interview by asking Walsh for the one question he gets most from investors and clients. He offered the following:

"There are a lot of elections this year and everyone's thinking about how we manage and think about election risk. We always do a lot of work of modelling potential outcomes. We think about what and how industries will be impacted by various policy outcomes,” he said.

For more information on the Yarra Global Small Companies Fund, please click below or visit the Yarra website: 

Managed Fund

Yarra Global Small Companies Fund

Global Shares

View