In this blog post, I am going to look at how Panel Regression can be used in the music business to analyze the effect of live touring on album sales and digital streams in different regions.
Panel data refers to data that are collected on the same set of units over time. Panel regression estimates the relationship between a dependent variable and independent variable(s) while controlling for individual and time-specific effects.
Panel regression models can be either fixed effects (FE) or random
effects (RE).
Fixed effects models assume there is a time-invariant unobserved effect
for each individual that is correlated with the independent variables,
but the effect does not vary over time. Random effects models assume
that the individual-specific effect is uncorrelated with the independent
variables and can vary over time. Both control for individual and
time-specific effects that are not captured by the independent
variables.
A panel regression model can be used on time-series data that includes an artist’s live touring activity/ticket sales, album sales, and digital streaming numbers across different markets. Additional relevant data such as marketing spend and social media engagement, and radio airplay could also be included as predictors. To determine whether to use a FE or RE model, we can use a Hausman test to compare the two estimators. The Hausman test compares the difference between the two estimators and the variance of the difference, and tests whether the difference is statistically significant. If the difference is significant, fixed effect should be used. Otherwise, random effects are preferred. Separate models can be built using album sales and digital streaming as dependent variables. Conversely, the live touring/ticket sales data could also be modeled as a dependent variable, to estimate the effect of these same factors on ticket sales.
Overall, panel regression offers a powerful tool for exploring the factors that contribute to the success of a particular artist over time in different markets.