Making the alternative a reality

If alternative data is to transform investing, it looks likely to be a long, slow process.

Quant investors say the data, from social media, satellites and digital transactions, suffers from having gaps or too short a history to be useful. Often it fails to yield insights different from those they can find elsewhere, in more conventional information sources.

Detractors joke that data sellers’ pitch books always contain the same case studies. “Everybody talks about counting cars in parking lots of Chipotle restaurants and it just makes me think: don’t you guys have anything else to talk about?” says one. Eight in nine datasets fail to meet the mark.

And yet there are cases where alternative data is bringing about the transformation in investing its fans imagine. A few managers are finding ways to put it to work – just not quite as expected.

Some are using it to augment existing quant strategies even as its standalone value is questioned. Firms talk, for example, of using sentiment data collected from online text sources to filter systematic value strategies for stocks in companies with flagging reputations.

Elsewhere, quants are experimenting with using new data to time trades in established strategies, or to weight the values of more conventional signals
in models.

Others point out the new data can be more timely than existing data, and when combined has the potential to replace existing macroeconomic releases. Banks are today compiling macro indicators using hundreds of data sources, including alternative data, updated with far greater frequency than official numbers.

The idea of small bands of investors mining obscure proprietary datasets for alpha may be wrong.

Quant firms will face a choice. Much of the new data is hard to apply to the types of strategies they run, which typically trade thousands of securities and are based on lasting market patterns

Instead, a picture is emerging in which alternative data becomes commoditised. But critically, it reaches investors more quickly and gives a richer view of what’s happening in markets than they’re used to now.

As this unfolds, quant firms will face a choice. Much of the new data is hard to apply to the types of strategies they run, which typically trade thousands of securities and are based on lasting market patterns.

Investment opportunities from alternative data might be transient or relate to just a few securities. Quant firms should have an edge in identifying such openings, but many are reluctant to move from relying on statistically robust signals to something less concrete. Some commentators think they will have to adapt to prosper.

Fundamental investors will face upheaval too. They will need quant skills to process the huge volume of data newly available to them. Some firms – such as Point72, BlackRock and Schroders – have already been recruiting data scientists as they look ahead to competing in the new environment.

As for the legion of alternative data sellers that has grown up in just a few years, a wave of consolidation seems likely. Incumbent data vendors would be the natural buyers, acting as aggregators, standardising the new data and mapping it to individual securities, as they do now with existing data.

It’s a picture in which much changes, but some things stay the same. Investors of all types will continue to consume whatever information they think can help them forecast prices. More timely information will be more valuable, but analytical capabilities – rather than getting hold of unique data – will likely hold the real key to investment success. In time, alternative data will cease to be seen as alternative.

It’s a picture, in other words, much like today only different.