What you describe is called "active" or "redistributive" balancing. While widely discussed on academical (and hobbyist) level, very rarely used in the real world. Simply because: li-ion cells have almost negligible self-discharge rates, and the differences between self-discharges between the cells (which determines the balancing need) are even lower.
Think about this: a typical cell self-discharges around 2%/year. Let's say it's neighbor self-discharges by 2.1%/year; the difference to balance out is 0.1%/year. At, say, 10Ah capacity, this is 10mAh per year. Assuming an example pack of 10 series cells at 10Ah, this would be around 0.37Wh wasted in heat - in a year! Approximating at $0.15/kWh, the cost for this electricity would be around $0.00005, per year.
Now, passive balancing BMS cost per cell is about $0.50, and well designed redistributive is about $2.00 - assuming both types are cost optimized for mass market. (This is based on my own designs of both types, so I have a general idea of what it costs. Sadly, the differences are actually even bigger, since redistributive balancers are expensive niches.)
Saving $0.00005 per year, the payback in a 10s pack (10 * $1.50) would happen in 300000 years. Of course, assuming that everything else, such as quiescent current, stays the same, which is a ridiculous assumption.
In larger packs, the numbers are less ridiculous. Yet, it's highly improbable that an active system ever makes sense, even in large systems. They are used in very specialized conditions. Another chemistry in the future might require the re-evaluation of redistributive balancing; li-ion is just too good.
Of course, if this pack is in any electric vehicle, the extra consumption of carrying a few dozen grams of extra electronics with you makes the bottom line negative.
Factor in any cost for developing the more complex system, and it falls apart completely.