How one stock picker uses AI for investment management

Picture of Darren Boey
Posted by Darren Boey

How do you choose 50 stocks to invest in, from a universe of more than 3,000?

This is the challenge Richard Firth has wrangled with since starting his Asia excluding Japan smaller companies strategy in 2008. His solution is a due-diligence questionnaire that underpins how his Firth Investment Management decides whether a company is a worthwhile investment or not.

The questionnaire contains 120 questions that focus on the critical areas of a company’s corporate structure, its business model and its balance sheet, and involve a thorough analysis of financial data, corporate reports and disclosures.

Recently, Firth upgraded the questionnaire with an artificial intelligence filter: Transparently.AI’s Manipulation Risk Analyzer. The MRA is an AI-powered tool that scores how aggressively a company is manipulating its accounting on a 0-100 scale. 

AI for investment management

Now, it’s one of the first things that Firth IM analysts check in their due-diligence questionnaire. Any stock with a risky score - typically over 50 - is a warning signal to Firth and his team to leave it out of their investment portfolio.

“Transparently.AI is now integrated within our process,” Firth said in a recent interview from his Singapore office. “It’s a good starting point to filter for stocks that might be interesting for us to spend time on. We don’t have to hold anything that doesn’t pass our checks.”

Firth considers the MRA a “fantastic tool” to help him find the 50 companies that Firth IM runs in its portfolio out of the 3,500 stocks in his investment universe. 

Transparently.AI’s MRA puts an innovative and thoroughly technological twist on the craft of stock picking, something that Richard Firth has been doing for more than 35 years.

Firth, who is English, started in funds management in 1987. He’s been investing in Asian equities since 1993 and lived in Asia since 1998. He worked for Schroders for 18 years out of the firm’s London, Hong Kong, Seoul and Singapore offices. 

Image of Richard Firth, founder of Firth Investment Management

In that time, he managed a South Korean equity country fund, a corporate restructuring fund sponsored by the South Korean government, served as chief investment officer of Schroders Korea, and launched the Schroders Asian Smaller Companies Fund in 2005. 

Inefficient markets

It was during this last foray that Firth noticed material inefficiencies present in the pricing of Asia ex-Japan small-cap stocks. This class of equities is typically less well-researched: Sell-side analysts cover less than half of the listed Asia ex-Japan names with a market capitalization under US$1 billion, according to Firth estimates.

Firth quickly realised the opportunities for a manager of an active fund to exploit, offering investors the potential reward of better returns than regional equity index benchmarks. He left Schroders and started Firth Investment Management in 2007 to focus exclusively on Asia ex-Japan small caps.
Firth Investment Management logo

Firth’s due-diligence questionnaire emerged out of necessity. Small-cap stocks are riskier, there is less attention from the market, less information is available, and often there is less transparency.

“The aim is to identify and understand the risks of a company that we are researching,” Firth said. “This will inform our assessment of a company’s valuation and our investment decision. In over 15 years of this strategy, we have had no loss due to a proven fraud."

“Having said that, we recognise the limitations of due diligence conducted by humans, and the obvious component of subjectivity and bias,” he said.

Back to top

AI for due diligence

This is where Transparently.AI and its AI solution are adding value.

Taking learnings from forensic accounting, automated signal detection approaches for financial crime and machine learning, the MRA hunts for the digital DNA likely present in the manipulation of accounts and financial crime.

The company has built an array of 150 financial models into the MRA that are designed to replicate the investigative activities of forensic accountants, activist short sellers and auditors. In addition, the system is designed to identify patterns of behaviour that would never typically be considered by these industry participants.

That’s what the AI is really good at doing. It’s using its analysis of the data to join the dots between a combination of risk factors.

The MRA sifts through millions of publicly available corporate financial data points to deliver a 0-100 percentage score that ranks the probability of a company collapsing as a result of its accounting malfeasance. Transparently.AI then delivers a custom report with information on what to do next, what to investigate, what questions to ask and identifies where in the accounts things don’t look quite right.

“That’s what the AI is really good at doing,” said Firth. “It’s using its analysis of the data to join the dots between a combination of risk factors and make comments in an objective manner in a way that a human finds difficult. This is a very useful check to support our investment decisions.”

Using AI to monitor portfolio risk

Besides helping with stock decisions, the MRA is a useful tool to monitor risk in investment portfolios. Periodic checks of portfolio risk scores can alert fund managers to changes in how aggressively a company is managing its accounts. Rising scores provide a warning signal of deteriorating fortunes - and a signal to revisit an investment.

Firth has applied the MRA to his existing holdings.

“For stocks that we already held before we had access to the Transparently scores, some had high scores and that was a cue to look into those stocks again and question if we needed to  change our exposure.”

There was one holding where the risk score and comments supported a decision to sell it completely.

“There was one holding where the risk score and comments supported a decision to sell it completely.”

He offers one caution about relying too heavily on the MRA score: “It’s important to remember that the risk score isn’t an investment decision tool on its own. Rather, it’s supporting a decision by quantifying an important source of risk in individual stocks.”

Specifically, the MRA is a cue for further action: It’s an ethical application of AI that provides a useful co-pilot to fund managers to help them make investment decisions and to monitor their portfolio risk.

Richard Firth is a convert. His message to other portfolio managers:

“You need to look at this product.”

Back to top

Disclosure: Firth Investment Management is a shareholder in Transparently.AI’s parent company.

Subscribe to our blog

Sign up for more AI-powered insights