Music playlists can predict market returns
17:28, 28 January 2022
'September' has the highest valence score in 20 of 40 countries - Photo: ShutterstockA study being published in a forthcoming Journal of Financial Economics suggests near-term stock market performance can be predicted by the types of songs people listen to over the music streaming service Spotify.
Specifically, by analysing a statistic Spotify tracks called “valence” or positivity to assess the overall mood of a population, four professors theorise increases in music sentiment translate to a 1.2 basis point increase in returns, followed five days later by a 1 basis point decline.
On an annualised basis, this equates to a 3% boost in equity returns on days when people are listening to upbeat music followed by a 2.40% decline the following week.
The basis for the study, titled “Music Sentiment and Stock Returns Around the World” – co-authored by Alex Edmans of the London Business School, Adrian Fernandez-Perez of Auckland University of Technology, Alexandre Garel of Audencia Business School and Ivan Indriawan of the Auckland University of Technology’s Department of Finance – is that numerous variables affect human sentiment, and as stock traders are human, that sentiment has an impact on equity returns.
The song with the top overall valence rating in the US, Canada, UK and much of the EU is Earth Wind & Fire's 1978 Billboard #1 hit September, with a 0.982 valence score, according to data maintined by Spotify. Conversely, Legion Inoculant by TOOL ranks last with a 0.026 valence rating. For just the 2010s decade, the aptly named Happy by Pharrell Williams scored the highest at 0.962, while Post Malone's Rockstar, at 0.129, had the lowest valence score (see chart).
Highest valence ratings of songs on Billboard's Top 100 Songs of the 2010s - Photo: Edmans, AlexSeptember by Earth, Wind & Fire makes stocks go up
What is your sentiment on US500?
Music as proxy for mood
The many variables that can influence our collective mood include weather, sporting event results, geopolitical unrest and Covid-19 restrictions. While mood can be difficult to quantify, the professors believe it is reflected in the choice of music we consume.
“We find that daily index returns are positively correlated with contemporaneous music sentiment and negatively correlated with sentiment five days prior, suggestive of a reversal,” the group wrote. “This result is consistent with the pattern we observe at the weekly frequency and suggests that mood swings, as reflected in music sentiment, lead to changes in stock prices.”
Additionally, they find that in the US – where stock trading is less restricted than in most other markets – the turnaround happens more rapidly, after three days of a positive sentiment gain.
Covid-19 restrictions influence mood
This is evident in Covid-19-related short-sale bans and trade volume restrictions implemented in some countries but not others in March of 2020, exacerbating the effect. Two countries introduced such stock trading restrictions soon after Covid-19 was declared a global pandemic and stock returns experienced a greater than average correlation to music sentiment.
In France, where a short-sale ban was instituted on 17 March 2020, and when Australia capped trading volume, a one standard deviation increase in music sentiment was followed by a 33.6 basis point increase in contemporaneous returns, followed by an 89.2 basis point decrease in future returns.
“In sum, the effect of music sentiment on market returns is markedly stronger when a country’s stock market is subject to limits to arbitrage,” the professors surmised.
Lowest valence ratings of songs on Billboard's Top 100 Songs of the 2010s - Photo: Edmans, AlexMutual fund flows affected
Mutual fund capital flows are also correlated to music sentiment, the professors found. For example, based on an average $976m (£727.7m) fund size, a one standard deviation increase in music sentiment saw a weekly $29,000 capital inflow, or $1.5m on an annualised basis. Conversely, the same size fund experienced a $31,000 capital outflow the following week, or $1.6m on an annualised basis.
For flight-to-safety government bonds, the results were limited. While the study did see a one standard deviation increase in music sentiment led to a 0.01 basis point decrease in bond-indexed fund returns or a 0.5% annual decline – significant given the mean bond-fund return is 2.2% per year – the study found no inverse relationship for future returns for bond funds.
In conclusion, the professors believe that music sentiment, as measured by Spotify’s valence metric, can be used as a proxy for a population’s overall mood, and therefore is predictive of its near-term asset prices.
