Systematic, unsystematic risk and Beta
...thor’s chosen source (yahoo) calculates its beta in a different way. According to the source, beta is determined as follows: “Regress excess returns of stock y on excess returns of the market. The slope coefficient is beta.” Beta = [(n) (sum of [xy]) ]-[ (sum of x) (sum of y)]/ [(n) (sum of [xx]) ]-[ (sum of x) (sum of x)] where n = # of observations (usually 36 to 60 months) x = rate of return for the S&P 500 index y = rate of return for the security Two factors might account for the differences among the author’s betas and the published beta: 1. The source calculates its beta using an equation while the author estimates his betas by determining the best fit line (where beta is the slope of that line). 2. The source calculates its beta using a period of 36 to 60 months while the author estimates his betas using a period of 12 months for each. The author prefers the published beta since it covers a bigger period, which reflects more the systematic risk of NSM’s stock ( due to general market conditions that cannot be diversified away) during the last 3 to 5 years. Furthermore, the published beta of 1.39 shows a less risky stock relative to the market (S&P500) as compared to the author’s estimated betas of 1.82 and 2.11. This would look more attractive to stockholders who are risk averse. This small analysis makes the author question whether he had a large enough sample size of what he was doing. Maybe the answer is a lack of large sample size, a fact already mentioned above. Also, it is mentioned in the chosen source (yahoo) that beta values are not calculated if less than 24 months of pricing is available. The author has calculated his betas using a period of 12 months for each. Further research and analysis would be needed to answer the above questions more accurately. Actual and predicted company returns The actual return is calculated for three days prior to the low point and for three days after the low point. Also, the predicted return for those same days is calculated using SML (Security Market Line). SML equation Required return on stock (predicted) = KRF + (KM – KRF) â The author assumes KRF to be negligible in the market risk premium factor. The used equation would become K = KRF + KM â The actual and predicted returns are presented in table2 AppendixB. To illustrate the comparison more clearly, the author compares the three sets of returns in fig.3 AppendixB. It could be noticed that there are more fluctuations in the actual return than in the predicted returns. Why are those returns different? Were there any important announcements for National during this period? And if so, did those announcements have any effect on the results? Important announcements during the period of 11-28-00 to 12-04-00 November 28, 2000 - National Semiconductor Corporation (NYSE:NSM) today announced a CMOS Low Drop-Out (LDO) regulator designed to improve talk and standby times in GSM, TDMA and CDMA cell phones. Dec. 4, 2000 - National Semiconductor Corporation (NYSE:NSM) today reported net earnings of $106.7 million, or 56 cents per share, on revenues of $595.0 million for the second quarter of fiscal 2001, which ended November 26, 2000. In the comparable quarter last year, which ended November 28, 1999, the company reported net earnings of $70.4 million, or 37 cents per share, on sales of $513.9 million. Year 2000 • Introduced industry's first 10/100/1000 Megabit per second Ethernet transceiver • First commercial Geode™ WebPAD™ devices introduced • Entered digital imaging market and announced high-performance CMOS image sensors • Formed technology partnership with TSMC (Taiwan Semiconductor Manufacturing Company) • Acquired Vivid Semiconductor, a leading supplier of flat panel display drivers The announcement by National about the new designed regulator on November 28th should have given a good signal to the stockholders. This might explain the difference between the predicted returns and the actual return National had. Yet, National experienced one of its lowest stock prices on December 1st ($18). In the author’s opinion, this announcement did not have any effect on the results; even if this announcement was considered to be important, it would take time to have the desired effect on National’s return. This information might be considered good news since it might have affected National’s return positively in the few weeks or even the few months to come. Other important announcements such as the formed technology partnership with TSMC in year 2000 should be good news to the stockholders eager for better results in year 2001. The author also notices a sharp decline in the market price index (S&P500) on November 30th. Since National’s stock was found to be twice as risky as the market, the decline in the stock’s price on December 1st (though the return went up) could be explained by the decline in the market price index. That explains the systematic, or market, risk that the company’s stock faces according to the CAPM approach. This risk could not have been predicted by the analysts, which explains the difference between the predicted and actual return. Problems faced during the analysis During the analysis, several problems were faced by the author. First of all, it was hard to conceive whether the announcements made by national during the low period stock were important or not. This would give an unclear indication of whether those announcements had any real effect on the results of National’s stock return. Second, the author found some difficulties interpreting the effects that those announcements might have had on National’s stock return during that same period. This made the author question the effectiveness of the CAPM approach. How does this approach compare with the “constant dividend growth” model? CAPM approach vs. “Constant Dividend Growth” model The CAPM models attempts to describe how the market values investments with expected returns and provides some additional insight to the nature of risks. There are several ways to classify investment risks, however, under the CAPM approach there are only two types of risks. The first is diversifiable risk, which is also known as nonsy...