Standard Deviation - a Dangerous Metrics

**blast_investor** Tue Jul 03, 2007 1:27 pm

Standard Deviation - a Dangerous Metrics
Common misleading formula for risk-adjusted performance measurement:

risk adjusted return = performance / standard deviation

It is pretty clear that Standard Deviation can not measure risk. It can only measure volatility, which is not risk itself.

When economists or financial magazines tout volatitlity (Standard deviation) adjusted return, fund managers in mutual fund or hedge fund worlds will game the system to harm regular investors. For example, using extremely high leveraged (like Long Term Capital Management Fund) method to achieve low standard deviation, but highly risky fund product.

What is good for investors if the risk is very high but with low volatitlity (low standard deviation)?

How to use use standard deviation properly in investment domain?
I don't know. Before we know what it is, it is better not to touch it.

**blast_investor** Tue Jul 03, 2007 1:37 pm

On LTCM risk profile not captured by standard deviation:

http://en.wikipedia.org/wiki/Long-Term_Capital_Management

LTCM fund was a famous fund managed by Nobel Prize economist winners. In above link, it says:

"At the beginning of 1998, the firm had equity of $4.72 billion and had borrowed over $124.5 billion with assets of around $129 billion.
"
This is leverage ratio (debt equity ratio) of 129/4.72 = 27 times.
This is highly risky operation just on this leverage ratio. Now, of course, LTCM managers uses diversification and standard deviation proved before 1998 that the risk of LTCM was low with low deviation. But we know the answer on the risk. Standard deviation is a misleading indicator on risk.

BIRTP or other value investors mainly uses our own domain of method to measure the risk: leverage, price to liquidation value (or replacement cost), owner's earning etc. These are mainly economic indicators, not a mathematical standard deviations on the past stock price charts. The risk assessment can be applied on a stock or a mixture of picks in a portfolio.

Standard deviation or volatility on risk is very misleading theory to start with. As economic theory, it probably has its usage in its domain. But for investment, it is not that useful theory for investors to make risk-adjusted returns.

**blast_investor** Tue Jul 03, 2007 1:39 pm

When investors use standard deviation of past performance, they would not know how risky LTCM was before 1998.

But value investors should know the risk. 27x leverage, up to 100x leverage (rumored) with diversified portfolio was insanely risky with value investing risk assessment.

Those Nobel Prize winners in LTCM was insane.

**blast_investor** Tue Jul 03, 2007 2:29 pm

Another case study against Standard Deviation as measure of risk:

NetBank (NTBK)
http://finance.yahoo.com/q/bc?s=NTBK&t=5y

This stock had pretty reasonable standard deviation before 2005, at least no more volatitlity than many technology stocks. But now it is in danger of bankcruptcy wipeout. The stock was as high as $60 a share, now it is $0.30 a share

However, common sense would know that NetBank is highly leveraged financial bank stock. In my opinion, any leveraged financial stocks are risky, probably as risky or more riskier than tech stocks. Financial banks are well known to leverage the money 10x to 100x to boost return and use risky method to loan out the money in risky manner.

This is the case where standard deviation is not much usefull to tell the risk behind the stock. A portfolio is a mixture of stocks, if Standard deviation can not tell the risk on each stocks, it can not tell on portfolio of stocks either.

Value investing with economic indicators can tell. Financial stocks should be avoided due to its derivative risk or leverage, this should be the default decision. To invest into financial stocks with value investing, there got to be extra reason to make it pass the margin of safety test.

**forestforever** Thu Jul 05, 2007 3:04 pm

I have watched NTBK for some time. A broker Paul Balling recently set the target price to $0 after NETBANK sold a part of its business to everbank...
Here are some numbers from Yahoo:
book value: around $4/share
EV: around $2/share
EPS: around-$4/share

the accounting charge for selling the business at a discount to Everbank:
around $1.2/share.
Paul Balling mentioned the net asset in his calculation is around $0.6/share.

It looks like nothing will be left for us to pick up, but there were some gurus including Sherman Bruce bought NTBK at $4 range. Amazing!