It is common for hedge funds to use alternative data sources to get an edge over competitors, but as technology moves forward, retail asset managers who do not address ‘blind spots’ in intelligence risk losing out.
That is what Mark Ainsworth, the ex-F1 strategist running Schroders’ data unit, says his team does: ‘We are augmenting their [fund managers’] capabilities.’
Ainsworth said that there are two ways the team can help fund managers: one is to answer questions, and the other is by building new capabilities. These two come together to ‘fill in the blind spots’ to let managers act with greater conviction. He gives the example of the work he did to answer a question from the emerging markets (EM) team, which was looking to invest in a grocery chain that was about to float.
‘They had 5,000 stores and management said they would double that to 10,000 in five years. But it occurred to them [the EM team] that they had no real evidence that it is a valid claim.
‘That was the question they put to us. We got all the facts of the demographics, towns, population… Having assembled all that data, we then used an algorithm to map out the capacity for this chain to open more stores. We saw very clearly they could build a couple of thousand, but going to 5,000 was going to be really hard.’
Ainsworth (pictured above) is quick to point out that the analysis does not make any recommendation relating to whether the fund manager should buy or sell a stock, but instead arms them with enough information to have a unique perspective when having further discussions with management.
Aside from answering a single question, the data team also builds new technologies and acquires data sets. In total, Ainsworth said they have more than 40 ‘data assets’, one of which is brand perception. This was used to answer a question about consumer perceptions in China regarding Burberry after the company posted ‘wobbly results’.
Another example is looking at bodies of text such as patent filings, and utilising AI to process the data. This allowed the team to come up with a view on the fate of car makers and how they will be affected by disruptive changes.
Private polling data
It emerged earlier this month that the Financial Conduct Authority could be looking to introduce guidelines around hedge funds’ use of private polling data ahead of big political events to make currency plays. But for David Wright, EMEA head of product strategy for BlackRock’s Systematic Active Equity team, events such as elections are quite difficult to predict and not very frequent.
‘For us, quant testing those is not an efficient use of our time. We want to build ideas that are repeatable. We have looked at using alternative data to understand potential surprise outcomes, but we use that as a guide to reduce risk. We have looked at Twitter, which has proven to be quite effective in calling a number of elections, Trump and the Brexit referendum.’
Other data that BlackRock uses include smart phone location, aggregate transaction data and satellite imagery, which the team has had success with in its China onshore strategy to understand regional growth expectations by tracking economic activity.
Social media is old hat
There has definitely been an increase in the use of alternative data, but it has not yet reached maturity.
A survey conducted by consultancy firm Element22 on 20 firms in Europe and North America with combined assets under management of $14.8 trillion (£11.4 trillion), found 55% of firms are still in the early stages of developing mature advanced analytics and alternative data capabilities. It said that only 10% of firms are breaking new ground.
‘What we have seen is an extensive increase in the interest in alternative data and the use of it. The dominant area is in investments and alpha generation. The other areas are client acquisition and retention and operations,’ founder Predrag Dizdarevic explained.
‘[But] it is not [so] anymore that you just have access to some data while others do not. It is more about combining different sources to get insight. You can use foot traffic data, for example, that can be captured through smart phones. There are also scanners in shops to track where people spend most of their time. The other is agriculture, where you can have a lot of aerial data, capturing information around the soil and other things related to it.’
He added that social media tracking is now so common that it is not even considered as new alternative data anymore.
The way the industry is moving, he pointed out, filtering of data will become crucial, because the quantity of data that is available is growing exponentially.
‘Even with the increasing capacities of systems and the quantity of the data, the power of computers, you need to filter it properly. The sources, the digitalisation of our lives, are progressing at a much faster rate than ever. One of the impacts of this is that things will be more factual and less based on estimates.
‘You need to now be extremely intelligent in how you combine some sources and how you use tools like machine learning.’