Alternative Data May Soon Become Mainstream Investing Tool, Says S&P Global's Aditya Sharma

Alternative data, once seen as a niche resource used mainly by quantitative investors, could become a mainstream investment tool over the next decade as artificial intelligence (AI) makes it easier to analyse large volumes of unstructured information, according to Aditya Sharma, head of product management for textual and professional datasets and quant signals at S&P Global.

Speaking at a panel discussion during the Lambda Quantitative Strategies Association's (LAQSA) sixth edition of the Indian Institutional Quant Conference (IIQC) on July 17, Sharma said advances in AI and natural language processing (NLP) are changing how institutional investors extract investment signals from data.

"Thirty years ago, estimates used to be alternative data. Now it's become fundamental... I think we will start seeing the transition of these datasets into fundamental over the next decade or so," Sharma said.

According to Sharma, alternative data should not be viewed merely as information coming from new sources such as satellite imagery or credit card transactions. Instead, it includes any dataset that has not historically been used to invest in a particular asset class.

"The data that you were not using for investing in that asset class historically is alternative to that asset class. It does not have to come from a new source... There is a lot of rich data content already out there which we can transform and turn into these quantitative strategies," he said.

He pointed to earnings call transcripts, broker research reports, news and social media as examples of datasets that have become significantly more valuable because AI can now analyse them at scale.

"Companies have been hosting earnings calls for a long time... But I think the advancements in technology have allowed us to process data at scale. I think that becomes a very important aspect," Sharma said.

He added that in today's markets, the speed of processing information has become just as important as access to the information itself.

"Even if you are not a high-frequency trading firm, you still need the data fast. You want it before your competition gets it, so that your models can run and make decisions. Otherwise, the market prices these things very quickly, and at that point, you start losing your edge in terms of trading," he said.

India Requires A Different Approach

Sharma said firms developing AI-driven datasets for India cannot simply replicate global models because the country's market structure and data availability differ from developed markets.

"We need a global lens for all the datasets that we create. Now, that does not mean that we take a cookie-cutter approach and say that, 'Okay, here's the data for the US, apply the same approach for India,'" he said.

He noted that while India offers abundant social media data, certain datasets, such as short-interest data, are unavailable, forcing firms to develop alternative indicators using futures and other market data.

"There are limitations, there are regional discrepancies, but what we are able to do is get creative about how we can provide the same information," Sharma said.

Defining Alternative Data

Alternative data has become a growing area of interest among quantitative investors, hedge funds and asset managers as AI-powered tools make it possible to quickly analyse large volumes of text, images and other unstructured data to uncover investment opportunities ahead of the broader market.

Balakrishnan Ilango, head of innovation and analytics, APAC at LSEG, said alternative data broadly refers to non-conventional datasets that go beyond traditional financial information, such as company fundamentals and analyst estimates.

"Alternative data is... all non-conventional data sets... apart from what we use as structured data, like fundamentals and your estimates," Ilango said during the panel discussion.

According to Ilango, the most widely used alternative datasets in India include news, social media, earnings call transcripts, corporate filings and sell-side research reports. He added that newer datasets, such as vessel-tracking information through satellite imagery, are increasingly finding applications in commodity investing.

"One of it which we use, even for India, is vessel tracking... especially when you are trying to model the commodities... to see the flow. And given the market situation now, where crude is driving the market, you want to try to use those elements in your ecosystem," he said.

According to one of LSEG’s white papers, examples of satellite imagery in investing include tracking the number of vehicles in retail parking lots to estimate sales, monitoring ships entering and leaving ports to gauge trade activity, and analysing shadows at construction sites to assess the pace of real estate projects.