As a CAPM exercise, you can take a different tack and look at the beta for the government bond index against the share market. Theory suggests that at least for a pure discount bond, there should be such a beta. You could certainly try to compute a beta for bonds overall, against the stock market, using the data from the web site. Thus, get the excess return by subtracting the CD returns off the bond returns and regress this on the market excess return. Is the bond beta greater or less than one, and how might this tie in with your prior expectations?