Alan Shaw of Harbeth will tell you it’s bogus and a phycological effect,
There's a nice article on driver break-in at the GR Research site
complete with measurements of various drivers as a function of time.
I have no reason to question Danny's numbers and I know my colleague has had some discussions with Danny over the years, I think they even defended each other on this topic on a forum once.
It goes back to the ops questions,
"So have these new speakers and been told they need a hundred hours to be broken in, and then sound will improve."
... whoever told him this was out to lunch. They don't magically improve after 100 hours. The major compliance "issues" work themselves out quickly, i.e. the distortion will drop quickly. After that, the driver has predominantly settled in 5-10 hours. After that, it will be subtle changes and more so on the bass. The changes in the other speakers, i.e. crossover points, etc. would be within the tolerance of manufacturing.
w.r.t capacitor break-in, I would challenge anyone to show any measurements of a film capacitor that shows parameter changing greater than the part tolerance ran at the typical volts/current in a speaker. Don't say "well we can't measure" ... We can't interpret how humans will respond to the measurement, that does not mean we cannot measure changes. Some capacitors are not well made and susceptible to absorbing moisture, and could change with humidity / temperature. Well most will change with temperature to some degree any way.