Notional IP portfolio holdings must start with bottom-up valuation. Standards from USPAP, IVSC, FASB, and IFRS/IASB establish baselines for assessing IP value. WIPO's list of documents on IP valuation help build out the methodologies needed. Valuation matters for establishing an investment's entry point. I wouldn't pay a million bucks for some patent that's only worth a nickel. The "cost" method reminds me of equity valuation methods that use book value and Tobin's Q. The "income" approach is pretty much the discounted cash flow method that most equity analysts should recognize.
The illiquid nature of IP means a pure-play investment in a bunch of intangible assets would bear more resemblance to a private equity investment than a public stock. Liquid proxies for an illiquid patent portfolio do exist. These include ETFs based on the Ocean Tomo 300 Patent Index. The simplest way to measure the alpha of a pure-play IP portfolio is to compare it to that index's return over whatever holding period is relevant.
There are "unknown knowns" that pose risk management challenges to IP portfolio management. Patent quality matters. There must be some way to perform statistical analysis of an IP portfolio's quality beyond just number of filings by country or sector. WIPO's PATENTSCOPE database and the USPTO probably have enough data to make this analysis worthwhile. Managing an IP portfolio's risk should consider litigation trends within the taxonomy of patent classifications. The risk breakdown would assign different risk weights to electronics, hydraulics, or whatever to avoid overconcentration of ownership in some class subject to heavy litigation (like software). My Google searches for "patent portfolio theory" and "patent portfolio race" reveal a notable amount of theory addressing these issues. Finally, I wonder whether IP portfolio considerations should differentiate by type of IP: patent, trademark, copyright, etc. A diverse portfolio containing many IP types may achieve an optimal risk-return tradeoff, or it may be an encumbrance to the search for pure-play returns.
Investing in IP or IP-heavy companies reminds me of the "innovation premium" theory. I proposed my own metrics for an innovation premium in an Alfidi Capital blog article last year. Those metrics can be weighted differently to account for qualitative differences; i.e. "number of patents filed" can carry a smaller weight if patent quality declines. I feel like publishing a longer research report on generating alpha from an IP portfolio but I will only do so if it adds something not already covered in WIPO's literature. Good knowledge of IP valuation would make a huge difference in technology transfer from research laboratories to the marketplace. I would consider commercializing tech from a government or university lab if I understood how its valuation fit into a larger portfolio.
Full disclosure: No positions in any investment products mentioned at this time.