Image courtesy of Behnam Norouzi.
Editor's note: Sasha Stoikov is a Senior Research Associate at Cornell Financial Engineering Manhattan, and he's also the founder of Piki, a music tech company. He works on high-frequency trading algorithms and music recommendation algorithms.
Decentralized finance, royalty exchanges, and global investment funds are changing the landscape of the music business. Record labels and publishers, the traditional investors in music rights, are seeing some new competition from fintech.
The emergence of financial products linked to music rights — also known as the financialization of music — has progressed on three main fronts. First, investment funds (Hipgnosis, Blackstone, KKR, Primary Wave, Mills Music Trust) are acquiring the steady cash flows of top musicians. Second, royalty exchange platforms (Royalty Exchange, Songvest, Anote Music) are facilitating the trading of music rights. Third, decentralized finance marketplaces (Foundation, OpenSea, Royal, sound.xyz) are proposing NFTs as a new investment vehicle for musicians.
All these developments aim to address a fundamental problem in the music business: How can a musician get paid now, allowing investors to benefit later? The main pitch to investors is that royalties are uncorrelated with the markets. There are also some tax benefits.
Let’s compare music rights to stocks. If companies can raise money and allow investors to benefit later, why can’t musicians do the same? Understanding this analogy between corporate stocks and music rights may shed some light on the development of a market for allocating music resources more efficiently.
The dataset: Using Monthly Listeners data for 1.1M artists on Spotify going back to 2019, we split the data into a training set (12 months of 2019) and a testing set (24 months of 2020-2021). We chose to focus on Monthly Listeners because this metric is strongly related to revenue. On their Loud and Clear website, Spotify has published data on average Monthly Listeners and royalties generated by various classes of artists. The relationship between the two variables is close to linear, with chart toppers generating slightly more revenue per Monthly Listener than emerging artists, most likely due to their deeper catalogs.
Diversifying Your Music Portfolio
The portfolio theory developed by Harry Markowitz in 1952 is based on the idea that a rational investor will minimize risk for a target level of expected return. In 1964, William Sharpe took this theory to its logical conclusion: If all investors are rational and invest in efficient portfolios, then they should hold investments proportional to the market capitalization of investments.
In the same spirit, we look at the performance of a portfolio of Top 500 streaming artists of 2019, weighted by Monthly Listeners. We compare this portfolio to the growth in Monthly Listeners of individual artists: Taylor Swift, Drake, Ariana Grande, and Kanye West. Despite the well-established nature of these artists, their streaming revenues can be volatile. When spreading this risk across the Top 500 artists, we see that diversification greatly reduces the volatility of revenues.
Smaller Artists Grow Faster
In the 1990s, finance researchers started uncovering new factors that impact the future returns of stocks. For instance, Fama and French found that size, i.e., the market capitalization of a stock, matters. In particular, they found that small stocks tended to outperform large stocks. In recent years, the opposite has been true, as large companies have dominated smaller ones.
Similarly, we split artists into four tiers:
- Tier 1: Top 500
- Tier 2: Next 2.5K
- Tier 3: Next 10K
- Tier 4: Next 20K
What we find is that the Tier 2 artists have grown more than twice as fast as Tier 1 artists on average.
Past Performance Is No Indication of Future Results
In 1993, Jagdeech and Titman found that stocks that performed well in the past year tended to outperform in the next year. This momentum factor has not persisted in more recent decades, but the recipe is very simple: Just buy the stocks that performed best in the last year.
In the case of Monthly Listeners, we can see that the opposite is true. Among Tier 1 artists, the Trending 100 (100 artists that performed best in 2019) tend to perform worse in the next two years. This seems to support the idea that listenership decays faster among artists who have performed unusually well in the year prior. This may be a natural reversion for an artist who is off-cycle, after having trended due to an album release in the previous year.
Artists With Steady Monthly Listeners Do Better
In a majority of markets, researchers have observed that stocks with a lower volatility tend to outperform those that are highly volatile. This is referred to as the low-volatility anomaly, since lower risks in finance are usually associated with lower rewards.
We found a similar effect among Tier 1 artists when selecting the 100 artists with the lowest volatility, shown in the graph below as the Stable 100. These artists represent the evergreen or legacy artists who have a very consistent following. This may justify why artists such as Bob Dylan, Bruce Springsteen, and Neil Young were recently able to sell their rights at high multiples of their recent revenue.
Alternative Music Data
In recent years, investment funds have gone beyond the standard stock features that are publicly available (size, momentum, volatility, etc.) and have pursued other sources of data made available by new data vendors. Some examples include web traffic, social media sentiment, satellite imagery, and geolocation.
When using data from Piki, an app that collects like and dislike feedback from its users on a variety of music videos, we find that out of the Tier 1 artists, those with the highest like-to-dislike ratios — the Piki 100 — have outperformed the Tier 1 portfolio. This group represents high-value artists whose streaming revenue is most likely to grow. As a result, these artists may be valued at higher multiples of their yearly revenues.
Top artists have recently sold rights to their catalogs for multiples of 15 to 20 times the yearly revenue. Royalty exchanges typically sell rights at multiples in the 5 to 15 range. The next waves of music rights transactions will depend on reliable sources of data that can help predict future music revenues for lesser known artists.
Ultimately, successful financial products need to be linked to cashflows. This is why we explored technical factors, constructed from historical data, to help predict future Monthly Listeners. For example, we showed that past growth, size, and volatility of Monthly Listeners can help identify groups of artists whose streams are likely to grow faster than average in the future.
Additionally, buyers of financial instruments linked to music rights will want to consider alternative music data sets that can predict the long term viability of music, particularly in our era of exploding music supply. By combining such fundamental analysis to technical indicators based on historical data, prospective music investors will be able to make more informed investment decisions.