Construction of Thematic Portfolios


The last decade has seen a huge pick-up in demand for thematic indices and ETFs. Since 2018, the assets under management in thematic ETFs have more than doubled.

For index providers like Solactive AG, this continues to offer an attractive playing field to provide the underlying indices for these investment products. Given that most themes are new, there are of course no established market benchmarks and whoever comes up with the best concept first has a good chance to win over the market.

There are two distinct steps in creating the next hot thematic index. Firstly, we need to identify the next topic or theme that will capture investors’ interest, attention, and ultimately their assets. Once the theme is clear, the next step is to build the index rules that will ultimately construct the portfolio to track the theme in a transparent, efficient, and rules-based manner.


Step 1: Identification of Investment Themes

Many of our new thematic indices are triggered by conversations we have with our customers about interesting investment themes. For the ideal thematic idea, we look for sectors, technologies, or trends that meet the following criteria:

  • The theme is sufficiently mature to have a sufficient number of securities that are listed and have reached a certain level of market capitalization and liquidity in order to construct a diversified and investable index basket; in other words, it only makes sense if the theme is actually investable.
  • The theme has significant long-term growth opportunities and upside potential for the valuations of the securities and those growth expectations have not (yet) been fully incorporated into the current price of the security; in other words, there needs to be room for growth or outperformance not just in backtesting but also going forward.
  • There is a fundamental, easy-to-explain-and-understand rationale for the growth of the industry or industries comprising the theme, ranging from upcoming regulations to disruptive shifts in various industries; in other words, there needs to be a good investment case.

We mostly try to identify future growth topics and we publish an annual report – The Solactive Future Trends – outlining some of the future trends that we specifically like, as a basis for the ongoing discussions with our clients. The 2022 edition covers 35 potential investment themes, from Autonomous Driving to Robotic Elderly Care. However, sometimes a change in the overall economic environment may also push an already-established theme back to everyone’s attention. During a period of high inflation, as we are experiencing now in 2022, taking a look at old-fashioned consumer staples companies that tend to be more resistant to inflationary pressures could be just as attractive as the next future technology trend.


Step 2: Construction of the baskets

Generally, we use two very different approaches to construct thematic indices and it very much depends on the theme and more specifically on the type of companies active in this field.

The first approach uses an existing sector classification scheme in order to identify companies. In this approach, we select a number of (sub-) sectors that together form the theme. For example, our Electric Vehicle Charging Infrastructure Index combines the sectors of Battery Charging Equipment Manufacturing and  Electric Vehicle Charging Stations using the RBICS sector classification system maintained by Factset. As companies can only be associated with one sector this approach ensures that all companies in the index do the majority of their business in the selected theme. On the other hand, however, it fails to identify companies that may be leaders in the theme but have other business lines that are (currently) still dominant. In some cases, for very new themes, it can also be the case that the sectors simply haven’t been created yet as the classification systems are somewhat lagging in including new trends.

These issues have triggered the emergence of the second approach, using big data, machine learning, and natural language processing (NLP) to create a multi-dimensional approach to thematic classification. By screening hundreds of thousands of documents, such as company reports, news items, and official filing documents, as well as using NLP to identify references to both companies and themes and their associated keywords, our ARTIS® algorithm assigns a thematic relevance score for each company respective to any given theme.

With this approach, we can identify which companies are most relevant to our theme, regardless of their other business lines. In our Cloud Technology Index, for example, we can use this approach to include Amazon – the leading cloud provider worldwide with a 33% market share – despite its one-dimensional sector classification as an internet retail company.

Having both approaches at our disposal allows us to construct the right index tailored to the respective theme. In the case of the ETC Group Digital Assets and Blockchain index, we even use a combination of the two approaches to ensure we capture all aspects of the theme.


This article was originally published in Han ETF Thematic Review in October 2022