Reading 40 Monitoring and Rebalancing really, provide, higher, classes

bodhisattva

Page 88 in the book:
Correlations also shoud play a role in the optimal corridor.  So, the book says, the higher the correlation of two asset classes, the wider the optimal corridor should be.
Why?  I don’t really get this concept.  Can anyone provide some intuition?
Thanks

giorgio10

correlation is only meaningful when measured over longer intervals. If correlation is high , it doesn’t mean prices always move in lock step over  every interval. There are always  short term  divergence happening . Still we can use correlation to judge early if this is long term divergence ( correction ) or merely short term fluctutaion which will revert to the mean once the event has passed.
The trick is to balance costs of rebalalncing with negative alpha from straying too far from the  design. If you react to every little event , yes you will have a good record tracking the design weights , but cost will be too high . You could use correlation as a guide to recommend when it is desirable and when it is not worthwhile.

NakedPuts2011

Thanks, that helps, however the only thing i am not comfortable with is that you mentioned that asset prices move away from each other.  What i don’t understand is that shouldn’t asset prices move in the same direction if the correlation is high?  
So let’s say you have equities and bonds; equities have 50% +/- 5% corridor and they have a high correlation with bonds.  So if equities move up so that they have an allocation of 55% then bonds would also move up and the allocation would remain the same. ahhh!!! i got it now.
Thanks

andytrader

If correlations are high , and there is a short term divergence from the allocation weights  because asset prices move away from each other ,  there is a greater confidence that they will soon swing back towards the set weights.
The corridor can be left  wider which simply means the investor does not have to react too quickly to the divergence. There is an automatic tendency to converge back if correlations are high.