Because Ameritrade's Initial public offering was in March of 1997, the time series is to short for estimating beta (August 1997), as the provided data will give inaccurate results. That's why in order to estimate beta for Ameritrade correctly we can refer to the finance data of the comparable companies. It means that we should choose companies, which have cash flows with similar risk indexes, as these companies will have same asset beta indexes. For running such test 14 different firms in 4 industries (investment services, Internet services, discount brokerage, internet) were chosen: A.G. Edwards, Bear Stearns, Lehman Brothers, MSDW, Paine Webber, R. James, Merrill Lynch, Mecklemedia, Netscape, Yahoo, Charles Schwab, E*Trade, Quick & Reilly, Waterhouse Investor Services. We can use the following formula to calculate beta for the firms above:

In order to estimate reliable results we should choose firms, which specialize in brokerage services. Discount brokerage firms get considerable share of their revenues from brokerage commission activities (about 80%) and net interest revenues. Understandably, these sources of income depend upon stock market activity. Discount brokerage firms, which will be our main focus, have activity similar to the activities of Ameritrade. Merrill Lynch on the hand with discount brokerage provides such services as asset management, investment banking and trading. We should not mainly focus on Internet service firms, as even though Ameritrade is planning to start it's own Internet services it can be classified as a real Internet firm. The data of E*trade is not very reliable as its stocks have a very short history, so it can not be accurate enough for conducting a regression analysis.

Stock prices and return data for Charles Schwab, E*Trade, Quick & Reilly, and Waterhouse Investor Services were used in order to make a regression analysis for beta calculations. Because too long periods may influence the stability of beta, we used data for the period of 4 years in order to run regression,...

Using equity betas data and data from exhibit 4 about capital structure we can evaluate average assets betas. After estimations we get the value of 2.1 for asset beta for Ameritrade (which represents estimated average value of betas for comparable companies of the discount-brokerage industry) as asset betas for companies used in regression analysis lie in the interval 1.98-2.2
We can conclude that asset beta, which we have evaluated for Ameritrade should be equal to equity beta as Ameritrade's debt equals to 0. If to apply capital assets pricing model we can also estimate Ameritrade's cost of equity (as we know market risk premium for Ameritrade).

Because Ameritrade has a zero debt, we can easily calculate opportunity cost of capital, which will be also equal to cost of company's equity or 21.5%

Making a conclusion we can say that the estimated cost of capital for Ameritrade is 21.5%. It means that any project, which will be done by the company, has to have a return higher than 21.5% in order to be ready. That's why the project proposed by Joe Ricketts, to invest in advertising and technological equipment for providing higher quality services should give return higher than 21.5%.

The forecast of financial specialists who analyzed performance of the company and performance of the industry in general was not as optimistic as predictions and expectations of Joe Ricketts. Even though that prediction of Ricketts is very optimistic (about 30-50% of return on investment), the verdict of analysts is only 10-15%. Because investments demand considerable funds Ameritrade will have serious issues with its stock prices (either interest rate may become lower or price of stocks may fall).

References

Mitchell, Mark Stafford, Erik Cost of capital at Ameritrade, Harvard Business School April 26, 2001

Ameritrade, History

http://www.tdameritrade.com/history.html

Federal Government vs. Private Accounting