# However the r 2 is only a descriptive statistic it

• Notes
• 29

This preview shows page 21 - 26 out of 29 pages.

However, the R 2 is only a descriptive statistic; it does not play a role in hypotheses testing, so that the unreliability of the R 2 in small samples is harmless. Notice the quite dramatic increase of the t-values. Recall that these t-values are the test statistics of the null hypotheses that the corresponding parameters are zero. Because the true parameters are equal to 1, what you see in Table 2 is the increase of the power of the t-test with the sample size.

Subscribe to view the full document.

22 APPENDIX Proof of (1): The first-order conditions for a minimum of are: Q α , ˆ β ) ' ' n j ' 1 ( Y j & ˆ α & ˆ β X j ) 2 dQ ( \$ " , \$ \$ )/ d \$ " ' 0 ] j n j ' 1 2( Y j & \$ " & \$ \$ X j )( & 1) ' 0 ] j n j ' 1 ( Y j & \$ " & \$ \$ X j ) ' 0 ] j n j ' 1 Y j & j n j ' 1 \$ " & j n j ' 1 ( \$ \$ X j ) ' 0 ] j n j ' 1 Y j ' n \$ " % \$ \$ j n j ' 1 X j ' 0 ] ¯ Y ' \$ " % \$ \$ . ¯ X , (42) and dQ ( \$ " , \$ \$ )/ d \$ \$ ' 0 ] j n j ' 1 2( Y j & \$ " & \$ \$ X j )( & X j ) ' 0 ] j n j ' 1 ( Y j X j & \$ " X j & \$ \$ X 2 j ) ' 0 ] j n j ' 1 X j Y j & \$ " j n j ' 1 X j & \$ \$ j n j ' 1 X 2 j ' 0 ] j n j ' 1 X j Y j ' \$ " j n j ' 1 X j % \$ \$ j n j ' 1 X 2 j ] 1 n j n j ' 1 X j Y j ' \$ " ¯ X % \$ \$ 1 n j n j ' 1 X 2 j (43) where are the sample means of the X j 's and Y j 's, ¯ X ' (1/ n ) ' n j ' 1 X j and ¯ Y ' (1/ n ) ' n j ' 1 Y j respectively. The last equations in (42) and (43) are called the normal equations : ¯ Y ' \$ " % \$ \$ . ¯ X , (44) 1 n j n j ' 1 X j Y j ' \$ " . ¯ X % \$ \$ 1 n j n j ' 1 X 2 j . (45) To solve these normal equations, substitute in (45). Then we get ˆ α ' ¯ Y & ˆ β . ¯ X
23 1 n j n j ' 1 X j Y j ' ( ¯ Y & ˆ β ¯ X ) 1 n j n j ' 1 X j % ˆ β 1 n j n j ' 1 X 2 j ' ¯ Y . ¯ X & ˆ β ¯ X 2 % ˆ β 1 n j n j ' 1 X 2 j ' ¯ X . ¯ Y % ˆ β 1 n j n j ' 1 X 2 j & ¯ X 2 hence 1 n j n j ' 1 X j Y j & ¯ X . ¯ Y ' \$ \$ 1 n j n j ' 1 X 2 j & ¯ X 2 . (46) Equation (46) can also be written as 1 n j n j ' 1 ( X j & ¯ X )( Y j & ¯ Y ) ' \$ \$ 1 n j n j ' 1 ( X j & ¯ X ) 2 , (47) because 1 n j n j ' 1 ( X j & ¯ X )( Y j & ¯ Y ) ' 1 n j n j ' 1 X j Y j & ¯ X . Y j & X j . ¯ Y % ¯ X . ¯ Y ' 1 n j n j ' 1 X j Y j & ¯ X . 1 n j n j ' 1 Y j & ¯ Y . 1 n j n j ' 1 X j % ¯ X . ¯ Y ' 1 n j n j ' 1 X j Y j & ¯ X . ¯ Y (48) and similarly 1 n j n j ' 1 ( X j & ¯ X ) 2 ' 1 n j n j ' 1 X 2 j & ¯ X 2 . (49) Moreover, 1 n j n j ' 1 ( X j & ¯ X )( Y j & ¯ Y ) ' 1 n j n j ' 1 ( X j & ¯ X ) Y j & 1 n j n j ' 1 ( X j & ¯ X ) ¯ Y ' 1 n j n j ' 1 ( X j & ¯ X ) Y j & ( ¯ X & ¯ X ) ¯ Y ' 1 n j n j ' 1 ( X j & ¯ X ) Y j (50)

Subscribe to view the full document.

24 The result (1) now follows from (44) and (46) through (50). Proof of Proposition 1. Recall from (1) that \$ \$ ' ' n j ' 1 ( X j & ¯ X ) Y j ' n j ' 1 ( X j & ¯ X ) 2 . (51) Substitute model (2) in (51). Then \$ \$ ' ' n j ' 1 ( X j & ¯ X )( " % \$ X j % U j ) ' n j ' 1 ( X j & ¯ X ) 2 ' " ' n j ' 1 ( X j & ¯ X ) % \$ ' n j ' 1 ( X j & ¯ X ) X j % ' n j ' 1 ( X j & ¯ X ) U j ' n j ' 1 ( X j & ¯ X ) 2 ' \$ . ' n j ' 1 ( X j & ¯ X ) X j ' n j ' 1 ( X j & ¯ X ) 2 % ' n j ' 1 ( X j & ¯ X ) U j ' n j ' 1 ( X j & ¯ X ) 2 ' \$ % ' n j ' 1 ( X j & ¯ X ) U j ' n j ' 1 ( X j & ¯ X ) 2 , (52) where the last step follows from the fact that similar to (50), j n j ' 1 ( X j & ¯ X ) 2 ' j n j ' 1 ( X j & ¯ X )( X j & ¯ X ) ' j n j ' 1 ( X j & ¯ X ) X j . (53) Now take the mathematical expectation at both sides of (52). Then, E [ \$ \$ ] ' \$ % E ' n j ' 1 ( X j & ¯ X ) U j ' n j ' 1 ( X j & ¯ X ) 2 ' \$ % ' n j ' 1 ( X j & ¯ X ) E ( U j ) ' n j ' 1 ( X j & ¯ X ) 2 ' \$ , (54) because taking the mathematical expectation of a constant ( β ) does not effect that constant, and taking the mathematical expectation of a linear function of random variables is equal to taking the linear function of the mathematical expectation of these random variables. The last conclusion in (54) follows from assumption II, and the second step in (54) can be taken because
25 we have assumed that the X j 's are non-random (assumption IV).

Subscribe to view the full document.

{[ snackBarMessage ]}

###### "Before using Course Hero my grade was at 78%. By the end of the semester my grade was at 90%. I could not have done it without all the class material I found."
— Christopher R., University of Rhode Island '15, Course Hero Intern

### What students are saying

• As a current student on this bumpy collegiate pathway, I stumbled upon Course Hero, where I can find study resources for nearly all my courses, get online help from tutors 24/7, and even share my old projects, papers, and lecture notes with other students.

Kiran Temple University Fox School of Business ‘17, Course Hero Intern

• I cannot even describe how much Course Hero helped me this summer. It’s truly become something I can always rely on and help me. In the end, I was not only able to survive summer classes, but I was able to thrive thanks to Course Hero.

Dana University of Pennsylvania ‘17, Course Hero Intern

• The ability to access any university’s resources through Course Hero proved invaluable in my case. I was behind on Tulane coursework and actually used UCLA’s materials to help me move forward and get everything together on time.

Jill Tulane University ‘16, Course Hero Intern