返回列表 发帖

Sneaky Little CFAI TRICKS

Not sure if there is a thread on this already, but I figured it would be useful to start a thread with examples of questions that had little tricks to confuse us ...

CFA Mock 2010

Quant vignette - a couple of the questions require the 'standard error' term but it is nowhere to be found (directly). However, in the regression output, we have the following:

SPREAD (one of the independent variables)
Coefficient = 1.0264
t-statistic = 4.28

In the vignette, it says the analyst hypothesizes that the coefficient for SPREAD is actually equal to 1. Then, Q40 asks to calculate the t-statistic relating to the hypothesis that the SPREAD coefficient is 1.

So, how do we calculate the t-statistic for the analyst's hypothesis on SPREAD without the standard error for SPREAD? Remember, the t-statistic in the regression output is calculated assuming the null hypothesis that the coefficient = 0. So, from the regression output, we know:

t-statistic = (coefficient - 0) / Standard Error (SPREAD)
4.28 = 1.0264 / Standard Error (SPREAD)
Standard Error (SPREAD) = 0.24

Then, to answer Q40:

t-statistic = (1.0264 - 1) / 0.24

...I know that was lengthy, but I will add some shorter ones that I come across. Just wanted to see if other people noticed these tricky questions ...

返回列表