Generating alpha with innovative ideas

Successful patents create a competitive advantage and added value for enterprises. The challenge in portfolio management is to determine the value of such innovations, and to apply this to single-issue share selection. Strategies that buy companies with high patent values and sell those with low patent values generate annual excess returns above 4%.

Dr Oliver Murschall
Executive Director, Research

Since the carrying amount of patents provides precious little support in this respect, Quoniam uses alternative data provided by IPR Strategies, a company specialising in performing patent valuations for enterprises. IPR Strategies has set up a database for traded patents which is based on information gained from M&A transactions, auctions, and manual valuations. The dataset is trained using a learning process and identifies indicators which are relevant to valuation – including market-based, legal or technical details of patents. To assess individual patents, they are compared to traded patents with similar characteristics.

IT sector holds the lion’s share of global patent value

At present, the global dataset contains usable data for around 10,000 enterprises covering more than 86% of the MSCI World market capitalisation. The five companies with the most valuable patent portfolios account for an aggregate value in excess of €267 billion. Aggregated by country (cf. figure 1), Japan and the US account for more than 70% of global patent values; these countries are the biggest innovators by a large margin, followed by Korea, Germany and France. Taking a sector view, the technology sector comes first – as one would expect – with just under 40% of global patent values.

Especially important: patent value and quality

Two aspects of patent portfolios are relevant for enterprises – the first one being the value of the patent portfolio, which impacts the enterprise’s valuation. Even though carrying amounts provide an indication of the value of patents, under US GAAP these must be recognised at cost – meaning that only purchased patents are reflected here. Under IFRS, patents may be carried at their development costs – which, however, frequently provide precious little information concerning economic potential.

Patent value distribution by country
Source: MSCI, IPR Strategies, own calculations
Patent value distribution by sector
Source: MSCI, IPR Strategies, own calculations
Indexed performance of long/short strategies
Indexed performance of long/short strategies

The importance of an enterprise’s innovation activity – relative to its competitors – differs by sector. For companies in the healthcare and technology sectors in particular, the power to innovate is highly important, whilst it is less relevant for businesses in the real estate and financial services sectors, or for utilities.ingegen weniger relevant.

The second relevant aspect covered by patent data refers to the quality of patents: high-quality patents secure a company’s future earnings during the term of the patent right and beyond, and the power to innovate sustainably secures an enterprise’s continued existence on the market – with market potential as well as legal and technical aspects all playing a role in this context.

Patent data offer significant added value

Quoniam has been incorporating patent information since 2018 in the form of factor composites in its alpha and risk models, based on univariate and multivariate analyses. Figure 2 shows the indexed performance of a long/short strategy between a top and a bottom quintile. Portfolios were structured on the basis of parameters for the factors of ‘patent value’ and ‘patent quality’ in a European equity universe. The strategy using the patent value factor generated an average annual return of 4.55% during the period running from December 2010 to May 2019. The same strategy using the quality factor generated an average annual
return of 4.89% since 2010.

The analyses have shown that using patent data generates significant added value for share selection. In this context, the low correlation of the patent value and patent quality factors with the existing alpha signal is worth highlighting. They comprise additional information that is not included in this form in existing alpha factors.