Numerous backtests have shown that periodic rebalancing increases portfolio return...

I would be hesitant to draw that conclusion.

It depends *very* much on what's in the portfolio, how it was chosen, and what the long term prospects of those thigns are.

If you don't know much about them, then rebalancing makes some sense.
i.e., you don't have an idea about differences in their long term prospects, and you're not likely to learn that either.

But for a portfolio of hand picked equities, I would be more hesitant.
Imagine you built a portfolio of a few blue chips in 1990: Berkshire, Kodak, and GE.
Would annual rebalancing have been a good idea?
Would you at some point have had an opinion about the ongoing prospects of the underlying businesses?

The only things worth lightening up on (if such is ever worthwhile) are
* those things which are temporarily overvalued, meaning you have an idea of what they're worth, or
* those about which you have no idea about their being overvalued, so you might as well assume each price rise is a rise in multiples.

Jim