You have voted on this request. View results below.
There have been several compelling academic studies in recent years focusing on the use of "dual beta" systems, or calculating both an upside beta and a downside beta. The intuitive logic is straightforward: traditional beta measures imply that the extent to which a security moves with the market is symmetrical, but that is often not the case. Dual beta calculations help disaggregate the systematic risk into components for more precision.

I've found myself wanting to incorporate these beta formulae into both my simulations and my post hoc evaluations. Doing so provides another tool to select lower-risk stocks for long screens as well as stocks that actually are superior hedges for short screens. (That's actually the origin of this request. My shorts weren't hedging my market exposure very well in down markets, despite the overall betas for the longs and shorts being comparable, and some spreadsheet calculations revealed that using higher downside betas in my shorts would have done a vastly better job.)

Marco, I saw some of your discussions in the forums from the past year when you built the flexible beta function. What are the chances you might be able to extend that work to create upside and downside beta functions? Since the only difference from the regular beta calculation is subsetting the data into two partitions?"up" market periods and "down" market periods?and then calculating the numbers for each, I'm crossing my fingers this wouldn't be too bad!

Thanks for your consideration.
Results: Total score: 35 , # of Votes: 10 , Average: 3.5
20 (5)
15 (5)
Scores are calculated as (importance) × (# of votes), where importance ranges from 1 to 4.
Requested by: rthaler
On date: 10/08/13
Category: Factors and Functions