If you look at the study, this is based on an n of 6 and there was significant variation across altitude, so reporting this as an average or a definitive result seems a little weird.
Moreover, the degradation rate was highest in the low-altitude installations, which experienced up to 1.5% efficiency loss per year. Whether your system operating at 80% efficiency after 20 years is material to you depends on the original installation size and power consumption (obviously), but I imagine it would be a replacement signal for most.
On top of all that, given changes in solar panel tech over the years, it's not like current installations really resemble the ones in the study. Apparently thin-film solar panels have a shorter service life, but modern monocrystalline panels can be rated up to 40 years with estimated 0.3-0.5% efficiency loss per year.
It's a fun study but I'm not sure why science reporting keeps insisting on trying to generalize or why this information would be considered groundbreaking given that it seems in line with manufacturers' current understanding of how their products perform.
