Getting three different value from Beta from Reuters, Bloomberg and Google Finance… Are there different methodologies for calculating this? Why is there a discrepancy?
Time frames?
Hey this is CFA stuff you should know it! There is 60 monthly observations and 260 weekly observations beta and they differ. Also Bloomberg offers adj beta although I forget what exactly the math is here as they took my terminal away.
“I prefer a lumpy 15% return over a steady 10% return any day”
beta adjusted = (2/3)*beta calculated + (1/3)*1 Bloomberg thinks that beta has mean reversion to 1 over the long-term.
fxguy1234 Wrote: ------------------------------------------------------- > Getting three different value from Beta from > Reuters, Bloomberg and Google Finance… > > Are there different methodologies for calculating > this? Why is there a discrepancy? You could calculate a beta using daily, weekly, monthly or quarterly returns. You could use 1 year of weekly returns or 3 years of monthly returns, or 2 years of daily returns, or 10 different variations on that. I’ve seen a stock beta with weekly betas calculated for a 2,3,4,5 and 6 years period, and they ranged from negative to over 5, on the same stock.
gotcha… didn’t know I could adjust time frames. makes sense now. Thanks all
everything that was said is right, but most importantly, beta is the covariance of the stock returns with the market returns. so you need some benchmark for the market returns. Bloomberg uses S&P 500, I’ve seen people use all kinds of other indices like Wilshire 3000, the Dow, NYSE indices, and what not. as you can imagine it makes a huge difference
Switching benchmarks (S&P500?, Russel3000?, DJIA?) is a factor that can make a big difference. Running a regression to get beta on 5 years of data vs 10 years can make a difference. Running a regression on daily data vs weekly or monthly data can make a difference. Deciding to tweak beta and assume that it tends toward 1 over the long term can make a difference.
Interesting discussion. In a related matter, I wonder how to treat positive betas but negative expected returns (without going short on the expectation), e.g. to find out which stock fared less bad than it’s peers?
This discussion kind of suggests Beta is a slightly ridiculous concept. Beta is whatever you want it to be.
In the equity level 2 book, and my limited experience in equity valuation, beta adjustments in 0.5 increments have a huge effect on instrinsic price calculation. Try 40 dollars up and down on a 50 dollar stock. I can make it do anything I want.
no, the discussion suggests that beta is a relative concept. it is what it is in a certain context, not in absolute terms
Beta is a tool for evaluating something. Just because there are variationsbon how the tool is used doesn’t make it useless. Just because the big screwdriver can’t be used to undo a tiny screw doesn’t mean that screwdrivers are useless. However, I do think the bit about adjusting beta closer to 1 seems awfully arbitrary. It would help if there would be some criterion for determining what the weight of the correction should be.
i like the reference to screwdrives, which implies in a subtle way that the beta is a tool that you use to screw things up. on the mean reversion adjustments - i couldn’t agree more. as someone mentioned before, bloomberg reports raw beta, as well as “adjusted beta” which is simply 2/3*raw beta + 1/3*1, essentially implying beta has a tendency to revert to the market beta of 1.00. nobody in bloomberg has a freakin’ clue where the 2/3, 1/3 weightings come from. no reference to any research publication or anything, they just have no idea - it’s been there forever and that’s all. call and ask, it is funny 
Mobius Striptease Wrote: > on the mean reversion adjustments - i couldn’t > agree more. as someone mentioned before, bloomberg > reports raw beta, as well as “adjusted beta” which > is simply 2/3*raw beta + 1/3*1, essentially > implying beta has a tendency to revert to the > market beta of 1.00. nobody in bloomberg has a > freakin’ clue where the 2/3, 1/3 weightings come > from. no reference to any research publication or > anything, they just have no idea - it’s been there > forever and that’s all. call and ask, it is funny > I second that. I was at a conference where Aswath Damodaran was speaking. He said he finally met someone at Bloomberg who’s in charge of this stuff and when he asked him where the 2/3 and 1/3 thing comes from, the answer he got was, “I don’t know–it was like that when I got here.”
What do you guys use Beta for anyway?
Carson Wrote: ------------------------------------------------------- > What do you guys use Beta for anyway? 1.2. Sometimes 1.4. Sometimes I add a size premium to the market risk premium. So basically, whatever I want.
goblue06 Wrote: ------------------------------------------------------- > > I second that. I was at a conference where Aswath > Damodaran was speaking. He said he finally met > someone at Bloomberg who’s in charge of this stuff > and when he asked him where the 2/3 and 1/3 thing > comes from, the answer he got was, “I don’t > know–it was like that when I got here.” was this in the Bellagio? :)) i went to that conference too, where he said that. i also tried to investigate it myself once before that with the same results
NakedPuts Wrote: > > > 1.2. Sometimes 1.4. Sometimes I add a size > premium to the market risk premium. So basically, > whatever I want. that is just weird to say that it is always bigger than one…