Digital Assets Report


Like this article?

Sign up to our free newsletter

The case for alternative data

Related Topics

One of the major trends to emerge in the global hedge fund industry in recent years has been the growing sophistication and proliferation of quantitative strategies, some of which are pushing the edge of what is possible with machine learning algorithms, in an attempt to uncover novel sources of alpha. 

Machine learning strategies are developing rapidly thanks to improved computer processing power. To be effective, however, managers need to feed their models with good data sets. 

Speaking on the panel A Data Future, moderated by Stewart Jardine, Director, Market Technology & Data Services, CME, Rani Piputri, Head of Automated Intelligence Investing, NNIP (a Netherlands-base investment manager), said that how one defines a good alternative data set depends on one’s investment process. “We’ve found some alternative data providers who can provide us with very good equity analyst estimates,” said Piputri. 

She added that fund managers who use alternative data sets have to think carefully about the cost of alternative data. If an equities data set costs USD500K, how much alpha would one need to generate to cover that cost, based on the size of the portfolio? 

It is enticing to think that these data sets could provide a Holy Grail of finding alpha in the market for lower risk, but the reality is quite different. Costs could spiral quickly if managers are not disciplined in how they use alternative data and understand how additive they are to the portfolio.

“Some alternative data is good at giving you short signals, but again, it depends how much short selling you do in your strategy,” added Piputri.
Professor Christophe Boucher is Head of Quantitative Research & Strategy, ABN AMRO Investment Solutions. He cautioned that machine learning models are useful in some but not all areas of finance, such as stock picking, or risk management to classify markets into different risk regimes. 

He cited shadow rating of bonds, when there is no publicly available data from ratings agencies, and developing ESG scoring models for corporates, as two areas that the bank has been looking at recently. 

There are four key factors that have contributed to the effectiveness of machine learning in the last few years. As the panel explained, these include: 

• Research and funding;
• Enhanced cloud computing; 
• Libraries of code developed by a fast growing online community;
• Volume of data. 

Strategy first or data set first?
For investors looking to allocate to a quant fund, one of the key questions should be how the manager thinks about data. Do they collect alternative data in the hope that their algorithms will find something? “If so,” said Professor Boucher, “this is a very poor process. It can create highly spurious results. When you use alternative data, you need to have an established hypothesis to explore.”

This is important because data mining can easily lead to the risk of overfitting – where the machine learning model finds correlations in data that do not exist. Alternative data sets are useful, and will continue to be, but they should be used for idea generation, rather than forming the basis of a strategy. 

Indeed, as the panel explained, alpha decay is also a risk, alongside overfitting.

Although an alternative data provider might have generated good alpha when back testing it for the previous decade, by the time it is sold to fund managers, and then applied to the current marketplace, they might find that there simply isn’t much inherent alpha left in the data set. One has to therefore be careful there is still a signal in the alternative data sets being purchase. 

Moreover, having data that is too precise is not necessarily going to improve one’s trading strategy. After all, if the rest of the market does not share the same view, it will not be possible to generate a positive return. 

There are huge advantages to using alternative data sets, but there are certainly risks, which investors should be mindful of. 

Investor insights on hedge fund performance 

One of the key features of the Amsterdam Investor Forum is to gauge how institutional allocators think about alternatives in the current climate. 
Last year was a tough one for hedge funds. Overall performance, on average, was down -4.1 per cent on a fund-weighted basis, according to Hedge Fund Research Inc. When asked to vote, 69 per cent of the AIF audience said they were disappointed by hedge fund results in 2018.

Still, on the back of market volatility in December 2018, 2019 could shape up to be a year for the active trader and give hedge funds a chance to reinforce their value proposition to investors. 

As Paul Tudor Jones, founder and CIO of Tudor Investment Corp, told Bloomberg recently, 2019 might be a better time to be a trader than to just hold. “I don’t know if we’re going to have a huge amount of trends. It could just be an enormously volatile period with a lot of back and forth,” he said. 
Dampened volatility has really hurt hedge funds since the Global Financial Crash. The central banks have not helped at all. 

Per Ivarsson is Executive Vice President and Head of the Investment Management Team, RPM, a Swedish asset manager which focuses on directional investment strategies, specifically managed futures and global macro.

Speaking on the Investor Insights panel, moderated by Andrea Marmolejo, Founder and Managing Partner, Blue Topaz, Ivarsson, said that “if the trends managers we invest with to make money are positive, they tend to get broken by an expectation of reduced stimulus and if they are negative, they get broken by the expectation of increased stimulus.”

This has made performance very challenging because the markets have not tended to operate on the basis of fundamentals. They have been propped up by central bank intervention. Rather than improve the economy, they have inflated asset prices. The panel said that the longer this goes on, the worse the crash might be, when it happens. 

There’s not been enough money going in and lubricating the real economy, from a demand perspective, they said.

Marcus Storr is Head of Alternative Investments, FERI Trust. Two elements in particular, concern him at present. The first, unsurprisingly, is ongoing Brexit uncertainty, and the second is liquidity in the underlying capital markets. 

In Q4, some corporate bonds were trading at 30, 40 basis points, “which is huge”, remarked Storr. “If we allocate to managers and they are forced through redemptions to dispose of assets at a time when the bid:ask spread widens, that is nothing but pure cost. And that’s what I’m worried about.”

The audience was asked: When trading markets, what is the predominant feature you are looking for?

The results were as follows:

• Value – 25 per cent
• Market neutral – 35 per cent
• Low volatility – 7 per cent
• Inflation protection – 4 per cent
• Growth – 29 per cent

Growth opportunities featured prominently but with equity markets getting toppish, and uncertainty surrounding global economic growth in 2019, market neutral came out as the preferred option, as investors look to reduce directionality and hedge their portfolios against an expected downturn. 

Speaking on a second Investor Insights panel, Lars Dijkstra, CIO, Kempen Capital Management, which focuses on specialist hedge fund strategies in niche markets, said that equity long/short managers did not fare well last year. “We like managers with a high level of complexity,” he said. “Structured credit is one example where we’ve selected a strong group of managers, as well as those running insurance-linked securities and long volatility strategies.”

One of the issues that investors have to contend with, when thinking about allocating in to new strategies is how to confidently identify managers who will turn out to be top decile performers. As Marianne Dernies, Head of Business Development, ABN AMRO Investment Solutions, pointed out, the difference between 2017 and 2018 was that dispersion of returns occurred not just broadly, from strategy to strategy, but within strategies; up to 70 per cent dispersion in UCITS-compliant equity funds. 

Susanna Wallis, Investment Manager, GAM, explained that there are three main factors GAM considers during manager selection: How long have they been running the strategy? Is the alpha sustainable over time? And thirdly, is the company robust? 

She said having specialists in one’s team is important. “We actually like dispersion (of returns). August 2015, for example, was good for volatility strategies, 2011 was good for structured credit, etc. You’ve got to know the personalities of different strategies to win out,” said Wallis. 

One of the points made on the panel, which was moderated by Jack Inglis, CEO of AIMA, was the ongoing importance of developing bespoke investment solutions, to try to generate the optimal risk/return profile for end investors. This has led to carve-outs, with managers selected to run a sub-set of their strategy, such as Asian equities for example. According to Wallis, 75 per cent of GAM’s asset base is now allocating in to customised managed accounts.

ETFs remain resilient in volatile markets

This desire among asset allocators to develop bespoke solutions comes at a time when the debate between active versus passive investing remains polemical. One of the consequences of central bank intervention over the last decade has been to send equities on a high velocity trajectory. Volatility in Q4 last year did contribute to the S&P 500 ending the year down -4.38 per cent but it has already rebounded strongly. Between 2009 and 2017, the index gave investors annualised returns of 15.65 per cent. 

Why pay 2/20 for an active equity long/short manager when all one needed to do was take a long position on various indices using a selection of low cost, highly liquid ETFs?

In a compelling panel debate entitled Trends and challenges driving ETF market growth, Marlene Hassine Konqui, Head of ETF Research, Lyxor, highlighted how investor sentiment towards ETFs has changed in recent times, going against an ingrained market perception that ETFs are the first instrument to be jettisoned when markets get stressed. 

Hassine Konqui explained that performance in 2018 for both active equity and fixed income managers was poor and that for the first time in Europe, “we saw passive flows exceed active flows. Not only in equity, which was already the case for the last few years, but in fixed income.”

From July 2018 onwards, alpha generation among equity hedge funds went into reverse, culminating in substantial losses in December. Fixed income strategies, on average, have returned -1.23 per cent over the past 12 months according to Hedge Fund Research. 

Research has shown that if actively managed fixed income funds do not outperform over three to six months, one starts to see outflows. 

“2018 was the first time we observed this phenomenon in fixed income markets,” said Hassine Konqui. “I think some progress has now been made in using fixed income ETFs for portfolio allocations.

“We saw another interesting development during the market correction in Q4 2018. Normally, one might expect people to sell their highly liquid ETFs first in a difficult market. In Q4, however, there was EUR100 million of net outflows in active funds compared to EUR6 billion of inflows into passive funds.”

This goes against the belief that passive funds add to a trend when markets experience a downturn. Rather, in Q4, passive funds showed their resilience. This is a new phenomenon and one that ought to reassure investors of the role of ETFs in global financial markets, suggested Hassine Konqui.

Like this article? Sign up to our free newsletter

Most Popular

Further Reading