Harry Gloom;326013 wrote:1) Annual rebalancing back to 75/25, so sell stocks when they are up, use cash/MMF to buy stocks when they are down at the end of each year.
2) Keep the 25% cash/MMF and only use it where there is a significant downturn.
I see the logic in both but Option 2 requires a determination (market timing) of when to turn off drawdown from stocks and use the cash/MMF and a 2nd determination when to revert back.
I think option 2 has too many inconsistencies with how it would play out.
The main reason for regular rebalancing is to avoid *style drift*. So if stocks were up 1,000%, your cash wouldn't be 25% anymore. It would be closer to 2.5%. So it wouldn't be much use to buy the dip with.
It also wouldn't be logically consistent, as the point you get into the market is arbitrary – stocks have already gone up >1,000% before you started investing ... So if 25% cash made sense when you started, it should still make sense 20 years later.
There have been lots of attempts to improve periodic rebalancing – like rebalance bands, etc. But they're usually just the result of back-fitting. There's really nothing you can do to improve a simple, periodic approach unless you're adding information to the market.
My favourite approaches:
1) Volatility harvesting – this is what the Yale model does. They'd have more asset classes with low correlations, so they'll rebalance between these sometimes multiple times a day. The theory is that short-term share price fluctuations are an anomaly – they're larger than they should be. IT discounts may be an opportunity. You make very small daily profits which add to the absolute return nature of a portfolio.
2) Timed rebalancing – you can use any algorithmic, macro or trend strategy to try and time when markets are about to turn, and rebalance then. The advantage is that it doesn't really matter if you get it right or not – all you're doing is maintaining the risk profile of your portfolio.
In long-term backtests, rebalancing every 19 months tends to be about optimal.