Beta by itself is not very useful, other than to filter for higher or lower average volatility. While higher beta overall deotes higher risk, it also denotes higher returns usually. It is the correlation of beta to the markets that shifts over time - high beta stocks can go through periods of low volatility, low beta stocks can do the opposite.
It is when this beta gets correlated to market moves that the risk comes into play - the market is up 1% but the stock is up 4% type of thing. The correlation depends on several factors - sector money flows, economic outlook for that sector, state of the economy etc.
Beta is an averaging tool that takes into accounts large periods of time which kind of "pools" all price action together and does not always reflect risks on a shorter timeframe.
It is when this beta gets correlated to market moves that the risk comes into play - the market is up 1% but the stock is up 4% type of thing. The correlation depends on several factors - sector money flows, economic outlook for that sector, state of the economy etc.
Beta is an averaging tool that takes into accounts large periods of time which kind of "pools" all price action together and does not always reflect risks on a shorter timeframe.
The statistics of beta....it's not a one man pony
Also - one of the most common errors in Beta - often overlooked by less mathematically inclined people is that: By construction Beta is just the product of whatever particular time period you choose...what i mean is that lots of people will look at the beta of a stock, let's say it's 0.5. And they'll assume that if the market's up 1% the stock will be +0.5%. Problem is that typically beta is calculated on fornightly movements for two years. Clearly that has little relevance for the movement of a stock throughout the day. Remember that if you're going to look at past data to infer indeas and risk, then the data set that you look at must be the same periodicity as that which you're trading. E.g. if you want to look at daily stocks, look at a daily beta; hourly trading, look at hourly betas. This idea that "BETA" is tjust this static, immovable risk measure is a false assumption...it's like looking at a monthly indicatior to show you hourly behaviour. Good article, and yes obviously beta has nothing to do with the fundamentals of a stock. But remember to get the statistics right guys if you're going to look at things like this.
Also - one of the most common errors in Beta - often overlooked by less mathematically inclined people is that: By construction Beta is just the product of whatever particular time period you choose...what i mean is that lots of people will look at the beta of a stock, let's say it's 0.5. And they'll assume that if the market's up 1% the stock will be +0.5%. Problem is that typically beta is calculated on fornightly movements for two years. Clearly that has little relevance for the movement of a stock throughout the day. Remember that if you're going to look at past data to infer indeas and risk, then the data set that you look at must be the same periodicity as that which you're trading. E.g. if you want to look at daily stocks, look at a daily beta; hourly trading, look at hourly betas. This idea that "BETA" is tjust this static, immovable risk measure is a false assumption...it's like looking at a monthly indicatior to show you hourly behaviour. Good article, and yes obviously beta has nothing to do with the fundamentals of a stock. But remember to get the statistics right guys if you're going to look at things like this.
Errr....that doesn't make sense...at all. Correlation of beta to the market. Beta is not a constant. It doesn't correlate to the market. There is no secondary derderivative of beta? The reason you're seeing your beta as being a bad predictor of the your stock is due to the exact reason i first stated. You're looking at a beta calculated from the wrong periodicity. It's nothing to do with it's correlation to the market. Also you're confusing stats with fundamentals. Clearly beta does not account for all extreme moves in a stock...as like the initial article rightly mentions, you can correctly calculate beta but if the stock declares itself bankrupt tomorrow it doesn't matter how good your beta is it's not going to predict it. Beta is an EX-POST calculation.
