trickle rebalance?

[rickb]

Do folks with substantial portfolios pick a day and rebalance - or is mechanically rebalancing over some fixed interval (e.g. one transaction per day over 2 weeks) with smaller transactions a reasonable idea?  As the portfolio size increases the absolute dollar amount involved in rebalancing becomes quite large.  If one asset exceeds 35%, to rebalance that asset back to 25% you're selling 10% of your total portfolio amount - which for a $1M portfolio is $100,000 and for a $10M portfolio is $1M.  If you happen to do this on a day where the price of the asset is (for whatever reason) 2% higher or lower than the price the next day this is a substantial amount of money (in absolute terms).  Instead of a single set of transactions on one day involving 10% of the portfolio, doing this with sets of transactions on 10 consecutive days each involving 1% of the portfolio seems like a way to make sure the price at the point of rebalancing is not some daily fluke (granted there's 10x the transaction costs, but at the transaction sizes I'm talking about the transactions costs are relatively insignificant).

Any thoughts or opinions about this?

[MediumTex]

I would be inclined to eyeball it and decide.

If the asset I was needing to sell had hit the rebalancing band because its price had surged, I would probably be more inclined to sell it all at once than I woud if the rebalancing band had been hit due to another asset falling in value.

Ultimately, I doubt if it matters that much, in part because rebalancing events are so infrequent (historically every 2-3 years for a portfolio not receiving new money).