This is a good question and I think it is an important one since the title implies a different situation than what was actually tested.
So the 11% reduction was due to an effort to increase mask wearing compliance by means of...
> Providing free masks, informing people about the importance of covering both the mouth and nose, reminding people in-person when they were unmasked in public, and role-modeling by community leaders
It's probably not that masks are less effective on people under 50, but that the information campaign was less effective at influencing that group to wear their masks.
Also, the results are interesting:
> The observers found that just over 13% of people in the villages that received no interventions wore a mask properly, compared with more than 42% of people in the villages where each household received free masks and in-person reminders to wear them.
It seems that there is much more progress that can be made in getting people to wear masks. And we still don't know how effective 95%+ compliance will be at reducing the spread.
Finally, the report does note that villages provided with cloth masks had worse outcomes than those who were provided with surgical masks.
There could be any number of reason why the result of the data look like it is. If those villages are mostly small farming communities it might be that young people spend most of their time outside on the fields or herding animals while the older population spend more time inside. It might be that the older generations after a long hard working life need more active help from younger family members and can't do distancing (like in our elder care facilities). Mask has a rather good track record when people are working inside in close physical proximity to risk groups.
Since the study did not explore those question we can only speculate.
Why aren't masks effective to reduce the spread for that age group? The discussion part of the study don't explore it. Is the distancing behavior of people aged 50 or higher different than those below?
I would like to see further studies being done on the younger demographic in order to explain why the data is looking like it is.