My Problem with Market Beta

[samdaman]

So I've been doing a bit of research on the ole beta metric. I've been looking at it in respect to TTK.NZ and well the results are puzzling. I started by checking a few sites for beta values and they returned varying results. As an engineer I get really unsettled by something like this when mathematically this value should constant. But is it really a constant?

I then tried calculating my own by getting historical prices from yahoo of the nzx50 and TTK for the last give or take 5 years. I used my handy dandy programming skills to take values out that coincide with yahoos data missing pieces so I had a perfect representation of the changes between the index and the stock. I was really looking forward to getting my a mathematically stable beta I could use to model this stock (big fan of models, engineer showing again).

And to my dissapointment i get the trivial value of 0.16 that makes sense though a correlation isn't going to last 4 years. but how long will a certain beta last then? well a few different calcs only returned non trivial results of periods of about 30 - 60 days.

My question is then why would I want to use a beta associated valuation on a stock that I intend to hold for longer than that? It makes sense a company can do good or bad in a market travelling the other way. I'm currently looking at beta as more of an indicator for chart patterns than a valid metric at the moment.

In a nutshell. Whats so important about beta?

Cheers Guys
Sam

[blackcap]

Hi Sam,

I have always been puzzled by the whole CAPM theory as it contrasts the EMH which states that you cannot predict the future of stock prices from past (historic) info. Why would Beta be any different? I mean all the correlations and math used to formulate Beta is taken from past pricing. So why should this Beta apply to a stock in the future?

[samdaman]

Exactly my point. Ive been quoted it by old timers and a friend studying finance currently as the holy grail of portfolio building but it seems like such a sensitive number based on past correlations. Sounds like the TA guys should love it yet a lot of FA guys won't go near a stock if the betas not in the right range

[blackcap]

Exactly my point. Ive been quoted it by old timers and a friend studying finance currently as the holy grail of portfolio building but it seems like such a sensitive number based on past correlations. Sounds like the TA guys should love it yet a lot of FA guys won't go near a stock if the betas not in the right range

If you like reading I can advocate "Taleb" a Lebanese author ("The Black Swan" and "Fooled by Randomness") who disputes the merits of financial theory and is very scathing of Merton, Scholes and Black. He also argues that the gaussian bell curve is fatally flawed (a lot of finance is based on this). The problem with finance in this country is that I believe our universities are 10-15 years behind and still religiously quote Fama et al.

[samdaman]

I'll give that a look. I'm a big fan of models (DCF, DDM) as they follow cash flows and a companies performance opposed to how a company "should" react to a market. I get uneasy putting numbers into a model which I can't physically substantiate. I'll have a squizz of those books he sounds synical of financial theory like myself :D

[catbert]

I'm not sure what you are expecting, especially when you say "should [be] constant"? Beta is just a statistical construct, based on past market prices. Just like, say, a moving price average. Would you expect a moving price average to be "constant"? Not at all. It changes as prices change, it changes as you vary the amount of history in your sample, it changes as the frequency (period) of your observations changes. There is no "correct" or "constant" value for a stock's beta, just as there is no "correct" or "constant" value for an average price.

[samdaman]

I'm not sure what you are expecting, especially when you say "should [be] constant"? Beta is just a statistical construct, based on past market prices. Just like, say, a moving price average. Would you expect a moving price average to be "constant"? Not at all. It changes as prices change, it changes as you vary the amount of history in your sample, it changes as the frequency (period) of your observations changes. There is no "correct" or "constant" value for a stock's beta, just as there is no "correct" or "constant" value for an average price.

I understand that catbert my quarrel is with its use. People use it to value a stock and to find cost of equities for discounting. If its changing so rapidly in terms of investment times then I don't see why people put so much faith in it. I just can't see it as a reliable input for finding an intrinsic value. If you agree with me then do you still use it in your valuations?

[catbert]

Ok, good. Actually, I didn't know that beta was ever used to value stocks (and I can't imagine how it would be used for that), I thought it's main use was in portfolio construction, e.g. "give me a portfolio of liquid stocks which will be less volatile than the market".

So, no, I can't see it as a reliable input for finding value, and I certainly don't use it myself.

Cheers, catbert :-)

[samdaman]

That makes a lot of sense and was what I had thought about beta. I haven't really analysed many companies from the overseas markets but I might look at beta a little more when I look at them. At least I know I can get an accurate beta now :)

[samdaman]

spot on this clears up a lot about the whole beta conundrum. haha cheers cuban