Tap into an alternative source of returns – with low correlations to traditional and alternative asset classes

Theo, after successful implementation in client mandates, the Quoniam Global Data Sentiment Strategy (GDS) has been launched as a mutual fund^*. How does the underlying strategy work?

The Global Data Sentiment strategy analyses news fully automatically to capture sentiment in over 30 capital markets, derives investment signals from this using innovative AI-supported methods and takes long or short positions in the relevant market instruments. More than 1.5 million news items are processed daily and grouped into overarching themes. Both commercial and non-commercial data providers are used in order to achieve better diversification and a broader data base when calculating signals.

GDS adapts more quickly to a changing market environment than most traditional portfolio managers who rely only on fundamental and traditional economic indicators published quarterly or monthly.

Global Data Sentiment has a performance of approx. 19.0 % p.a. with a volatility of 14.7 % p.a. and a Sharpe ratio of 1.3. The strategy’s first mandate was launched in February 2021 with a target volatility of 5 %. On October 29th 2024, a sub-fund of the QFS SICAV with 15 % volatility and leverage of approx. 400 % gross and approx. 150% net has been launched by Union Investment Luxembourg S.A. This analysis refers to the strategy with 15 % target volatility over the period 01/2015-12/2024.

What makes the strategy so attractive?

We have now been running the strategy for almost four years with a volatility target of 5 %, and since 2015 we have been running model calculations for both the 5 % volatility target and the 15 % volatility target, the latter being in line with the recently launched mutual fund^*. Both show attractive results. However, we also see uncorrelated performance versus traditional asset classes and traditional hedge fund strategies. This means that Global Data Sentiment is an attractive option for institutional investors not only from a return perspective, but also from a diversification perspective.

How does the strategy fit into the hedge fund universe?

Hedge fund strategies fall into five main categories: Relative value, event-driven and directional, plus CTA/managed futures and multi-strategy. In addition to extraordinary events, technical indicators, company figures and macroeconomic publications are often analysed in order to participate in both rising and falling markets.

Global Data Sentiment is a directional managed futures strategy. In terms of the investment universe, it is similar to global macro strategies and managed futures. However, there are significant differences in the way signals are generated: The strategy is based solely on news sentiment. This results in a low correlation with most hedge fund strategies. Unlike managed futures strategies, which are based on price momentum, i.e. pursue a trend-following strategy on asset prices.

Does this mean that the different performance drivers lead to lower correlations?

Yes, we have compared the Eurekahedge hedge-fund indices with the Global Data Sentiment strategy. There are 20 indices, which can be divided into the five categories described above.

Table 1 illustrates that the correlations range from around -14 % to approx. 26 %. The clearest positive correlations can be seen for CTA/managed futures (26.5 %) and the (global) macro hedge fund index (17.4 %), which also pursue directional strategies and utilise a similar investment universe. The overall level of correlation is moderate.

The largest negative correlations exist with strategies based on specific events rather than market trends, such as distressed debt (-13.8 %) and event-driven (-9.6 %). However, even here the levels are low.

Fixed income hedge fund strategies are also negatively correlated, as the asset class tends to be negatively correlated with equities.

The lowest correlations are with market-neutral or non-directional strategies (relative value: -1.2 %, long-short strategies: 2.9 %), mixed strategies (multi-strategy: -2.7 %) and the arbitrage hedge fund index (2.6 %). As GDS does not invest in commodities, the correlation to commodity hedge funds is also negligible at 2 %.

When analysing hedge funds as a whole, there is little correlation with Global Data Sentiment: The correlation with the equally weighted total hedge fund index, which comprises over 3,000 hedge funds, is only around 6 %.

How do the correlations compare to the liquid asset classes of bonds and equities?

The Global Data Sentiment strategy is designed to participate in the longer-term uptrend in global equity and bond markets, as evidenced by its low positive correlation with equity and bond markets (see Table 2). The FX model, on the other hand, is designed to participate in both falling and rising longer-term trends in currency markets, with an overall correlation of zero over the longer term.

Investors are often interested in correlations in falling markets. How does the strategy perform in the event of a crisis? Is it a good crisis indicator?

The main reason for the low correlation between Global Data Sentiment and many hedge fund strategies is the use of alternative data, which reflects information and sentiment faster than it is priced into the market. As a result, GDS shows rapid short positioning during crises. Many CTA/managed futures strategies use trend-following strategies based on price data and as a result react slower.

This can be illustrated by two crisis events from recent years: When the stock market crashed during the coronavirus pandemic in March 2020, the Global Data Sentiment strategy entered into a short position and exited from equities in good time. In the bond market, the Russian invasion of Ukraine initially led to a flight into high-quality securities from March 2022 and subsequently to increased inflation concerns. Again, GDS correctly anticipated the correlation and went short after being long in the bond market.

Conclusion

The correlation of Global Data Sentiment to hedge fund strategies is low to moderate. It shows that GDS can match CTA/managed futures in certain market phases, but still offers sufficient independence to create diversification benefits for investors. In addition, Global Data Sentiment can be used to participate in long-term rising equity and bond markets with a very good risk-return ratio. The main differentiator is the text-based trend identification, which differs from the price-driven models of many CTAs.

Read more about the strategy

^* Sub-funds of the Quoniam Funds Selection SICAV
Capital management company: Union Investment Luxembourg S.A.
Portfolio management: Quoniam Asset Management GmbH