The Challenges of Using Beta to Measure Risk

by: Bennington Investment Ideas

The concept of beta to measure systematic risk is a key aspect of Modern Portfolio Theory. Beta is used in the capital asset pricing model to determine the hurdle rate for securities. A high beta is often used as shorthand for a risky or growth stock, while a low beta is traditionally associated with "safer" stocks, such as utilities. Gold is typically not correlated to the market as a whole and hence has a beta around 0. Betas can range from across the spectrum; however, it is rare to see one below -1 or above 4. The market, as measured by the S&P 500, has a beta of 1 by definition.

In mathematical terms, beta is the ratio of the covariance of the market's returns and security's returns to the variance of the market's return. However, this does not really provide any intuition around beta. Rearranging the mathematical terms gives beta to be the product of two numbers:

  1. The ratio of the security's volatility to that of the market's volatility
  2. The correlation of the security's returns to the market's returns

Thinking about beta in these terms provides more intuition. This article will look at the changes in Beta over time for several leading large capitalization stocks.

Leading Large Cap Stocks

Ticker Name Market Capitalization ($ Billions) Dividend Yield
XOM Exxon Mobil Corporation 404.5 2.3%
AAPL Apple Inc. 308.8 0.0%
GE General Electric Company 208.1 2.9%
CVX Chevron Corporation 205.9 3.0%
MSFT Microsoft Corporation 205.8 2.6%
WMT Wal-Mart Stores, Inc. 192.8 2.6%
PG Procter & Gamble Company (The) 188.7 3.1%
T AT&T Inc. 185.4 5.5%
C Citigroup Inc. 119.2 0.0%
MCD McDonald's Corporation 85.9 3.0%
UTX United Technologies Corporation 80.6 2.2%
F Ford Motor Credit Company 56.7 0.0%
DD E.I. du Pont de Nemours and Company 48.9 3.1%

Source: Data is provided by services and was downloaded on May 21, 2011.

I downloaded the price histories from Yahoo Finance for each of these securities and looked at the Beta over the available life, using rolling 30 month periods. The following table shows the minimum and maximum betas for each stock going back to January of 1980.

Historic Range of Betas

Ticker Minimum Beta Maximum Beta Range
F 0.0 4.0 3.9
AAPL (0.7) 2.9 3.6
C 0.3 3.3 3.1
MSFT 0.1 2.3 2.2
WMT (0.3) 1.8 2.2
T (0.1) 1.9 2.0
PG (0.4) 1.5 2.0
MCD 0.1 2.0 1.9
GE 0.3 2.2 1.9
CVX 0.2 1.6 1.4
XOM 0.1 1.5 1.4
UTX 0.5 1.9 1.4
DD 0.5 1.6 1.1
Source: Derived from monthly split and dividend adjusted prices from Yahoo Finance.

At first glance, one would think that beta is not a very useful concept to use given that fact that the value can change significantly over time as noted above. Only two stocks, have a beta range even close to 1 and those are DD and UTX. The following figure (click to enlarge images) shows a subset of these stocks and how their beta has changed over time.

Select Betas from 1980 to Present

Source: Created from Yahoo Finance data.

The interesting observation is that these stocks were initially clustered around a beta of 1 and have since shown much less stability. It should be noted that GE, C, and F were greatly impacted by the recent Great Recession. Furthermore, AAPL has gone through several "near death" experiences. However, simply looking at betas over time does not explain the full picture. It is even more illustrative to review the two components: correlation and volatility ratio. The following chart shows these factors over time for GE:

GE Beta Decomposition 1970 to Present

Source: Created from Yahoo Finance data

The above graph shows the GE beta over time. Yahoo Finance currently states it as 1.82. Using my models, this suggests they might be using a time frame around 36 months, since if I apply that it will give 1.8 as well. However, in any case it shows a substantial increase in beta suggesting that the stock has become more risky to own. Prior to the Great Recession, GE's correlation few to historic lows while its volatility ratio remained stable around 1.5. However, the onset of the Great Recession and concerns about toxic assets pushed GE's correlation back towards 90% while its volatility ratio climbed. Having both factors increase resulting in a drastic increase in beta from around .5x to about 2x.

MSFT Beta Decomposition IPO to Present

Source: Created from Yahoo Finance data

Looking at the above graph for MSFT, one can see that its correlation to the market has been relatively stable over time - hovering around 50% with a notable dip in late 2005. Again noting that I use a 30 month time frame. A longer time frame would smooth out the variations even more; however, the same trends would be present. In the beginning, MSFT was a more volatile technology play with a volatility around 2.5 to 3x the market. However, as the company has grown and the business is largely driven by two established franchises, Windows and Office, the beta was converged lower due to a reduced volatility ratio and a slight increase in correlation.

DD 30 Month Beta Decomposition 1970 to Present

Source: Created from Yahoo Finance data

DD shows a reasonable amount of stability for such a long history. Like many securities during the Great Recession, its correlation converged towards 100% as investors exited equities on a wholesale basis. The correlation has typically been above 50% and below 90%. With the exception of a few spikes, the volatility ration has been right around 1.5x for the last 40 years. This stability is further noted when using a 60 month beta as shown below:

DD 60 Month Beta Decomposition 1970 to Present

Source: Created from Yahoo Finance data


Whenever looking at beta calculation or any metric that is based on historical data, it is important to understand how it is calculated and over what time frame the data is taken. I used 30 months as my default. The following table shows various betas sourced from different data provides as well as my own analytic calculations using different time frames.

Comparing Betas from Data Providers

Ticker 30 Month Beta 36 Month Beta 60 Month Beta Yahoo Finance Beta Beta
MSFT 1.09 0.96 1.03 0.97 1.05
MCD 0.39 0.36 0.47 0.39 0.50
GE 2.16 1.78 1.63 1.82 1.64
WMT 0.45 0.35 0.32 0.36 0.31
XOM 0.62 0.42 0.48 0.41 0.49
PG 0.62 0.51 0.50 0.52 0.53
DD 1.53 1.49 1.40 1.50 1.39
F 2.58 2.40 2.40 2.40 2.37
T 0.64 0.59 0.65 0.59 0.66
UTX 1.13 1.04 1.01 1.05 1.00
CVX 0.91 0.70 0.74 0.70 0.75
C 3.21 2.57 2.50 2.62 2.52
AAPL 0.94 1.14 1.32 1.14 1.36

Source: Yahoo Finance, Beta provided by services downloaded on May 22, 2011 and author calculations. The table uses two decimal places for better comparison between my calculations and the betas from the data providers.

Just glancing at this chart it is possible to see that my calculation for a 36 month beta aligns closely to the Yahoo Finance figure, while my calculation for a 60 month beta is close to the beta. In a few cases, where the beta has been relatively stable and Yahoo Finance are aligned; however, that is not consistent.


The casual investor can take a few things away from this analysis.

  1. Betas change over time and so it is important to use them with skepticism. Would the appropriate beta for GE be the current value around 2.0x or would it be better to use an earlier figure? A large change in beta probably suggests that significant new information about a company has been revealed and it would necessitate a reassessment if one has a position in that company. The problem here is that the beta is a historical measure that looks back after the fact. The next question in looking at a beta is asking the question whether it makes sense looking forward by making a fundamental assessment of what the company is doing.
  2. It is always important to understand how a derived metric is calculated so you can make your own assessment as to how you should use it when making an investment decision.
  3. Let me repeat, simply looking at beta is probably not a great way to make an investment decision.
  4. Beta does not necessarily provide any insight into whether a security is overvalued or undervalued. Significant increases in beta might suggest that the security has been previously incorrectly valued and the market it is now correcting itself. GE might be an example of this as a nice stable industrial company. During the Great Recession there was a broad realization that GE had substantial exposure to financing activities and potentially was at risk.

Because beta is a central part of Modern Portfolio Theory, a stronger adherent to it can probably look for different investment opportunities based on the trend of the beta. As beta increases, the hurdle rate for a stock increases as well. If there is no change in the expected free cash flows than the present value of those free cash flows will decline, resulting in a declining price. However, this is making the assumption that the market is moving in the correct direction. If you think the market is wrong in assessing the risk, the opportunity is taking the opposing position - going long a stock that has an increasing beta.

A dividend investor could look at this analysis and use it as a final check on stocks. If I were a dividend investor, I would seek to minimize risk. First, I would want to have a relatively low beta. Second, I would want to have a stock with a relatively stable beta. A DD, UTX, and CVX might look a little better than GE. MSFT, and WMT.

Again it is always valuable to do some fundamental analysis prior to making any investment decision.

Disclosure: I am long SPY.