1. This document provides solutions to problem set questions about econometric theory. It summarizes properties of sample means and variances, the method of moments estimator, and limiting distributions of estimators.
2. Key results shown include that the sample mean is an unbiased estimator of the population mean, the variance of the sample mean decreases to 0 as the sample size increases, and method of moments estimators are consistent and asymptotically normally distributed under regularity conditions.
3. The document also proves that if a sequence of random variables converges in probability to a constant, then the distribution of that sequence converges to a degenerate distribution concentrated at the constant.
1. This document provides solutions to problem set questions about econometric theory. It summarizes properties of sample means and variances, the method of moments estimator, and limiting distributions of estimators.
2. Key results shown include that the sample mean is an unbiased estimator of the population mean, the variance of the sample mean decreases to 0 as the sample size increases, and method of moments estimators are consistent and asymptotically normally distributed under regularity conditions.
3. The document also proves that if a sequence of random variables converges in probability to a constant, then the distribution of that sequence converges to a degenerate distribution concentrated at the constant.
1. This document provides solutions to problem set questions about econometric theory. It summarizes properties of sample means and variances, the method of moments estimator, and limiting distributions of estimators.
2. Key results shown include that the sample mean is an unbiased estimator of the population mean, the variance of the sample mean decreases to 0 as the sample size increases, and method of moments estimators are consistent and asymptotically normally distributed under regularity conditions.
3. The document also proves that if a sequence of random variables converges in probability to a constant, then the distribution of that sequence converges to a degenerate distribution concentrated at the constant.
1. This document provides solutions to problem set questions about econometric theory. It summarizes properties of sample means and variances, the method of moments estimator, and limiting distributions of estimators.
2. Key results shown include that the sample mean is an unbiased estimator of the population mean, the variance of the sample mean decreases to 0 as the sample size increases, and method of moments estimators are consistent and asymptotically normally distributed under regularity conditions.
3. The document also proves that if a sequence of random variables converges in probability to a constant, then the distribution of that sequence converges to a degenerate distribution concentrated at the constant.
where 2 = V ar(yi2 ). P P 3. When n12 i wi2 0, V ar(y ) 0, which implies that y 0. P P P 4. Let n := maxin wi . Then n12 i wi2 n12 i wi n = nn n1 i wi = nn , since P n 1 i wi = 1. We are given that n = o(n), which means n 0. Therefore, n P 1 2 i wi 0. n2 2. (5.4) Method of moments replace population moments with sample P estimators 1 3 versions, so that 3 = n i yi .Then if Ey 6 < and the data are i.i.d. we have
d n( 3 3 ) N (0, 2 ), where 2 = V ar(yi3 E(y 3 )) = E(y 6 ) [E(y 3 )]2 . 3. (5.6) 1.
n( ) =
n( 2 2 ) = ( + ) n( ). P
By the last question in this problem set we know that
, so that + 2. Then by Slutsky (or the continuous mapping theorem),
P n( ) N (0, 42 2 ).
Intro. Econometric Theory (Fall 06/07)
1-2
Nese Yildiz
2. When = 0 variance of the limiting distribution of n( ) is 0, which
means that this limiting distribution is degenerate, or alternatively, n(
P ) 0. 2
d 3. n = n 2 = [ n ] = [ n( )]2 = 2 [ n( )/]2 2 Q, where Q 21 , by continuous mapping theorem
2 P P 4. When = 0, n 0, and hence n 0. To get a nice non-degenerate
limiting distribution, we had to multiply this quantity by n.
4. We know that > 0, P (|Xn c| ) = P (c Xn c + ) 1. If we let the probability distribution of the constantbe denoted by F , then F (t) = 0 for . each t < c and F (t) = 1 for each t c. Next, consider t < c and let = ct 2 Then c > t, and P (Xn t) P (Xn < c ) 0 = F (t), since P (c Xn c + ) 1. Next, consider t > c and let = tc , so that c + < t. We 2 have P (c Xn c + ) P (Xn c + ) P (Xn t) 1. Moreover, P (c Xn c + ) 1. Thus, P (Xn t) F (t), for t > c and for t < c. Since c is not a continuity point of F this proves the claim. 5. We know that for each t R, P (an (Xn b) t) FZ (t), where FZ denotes the distribution of Z. Let > 0 P (|Xn b| ) = P ( Xn b ) = P (Xn b ) P (Xn b < ) = P (an (Xn b) an ) P (an (Xn b) < an ). Note an and an . Since P (an (Xn b) t) FZ (t), P (an (Xn b) an ) FZ (an t) 1 and P (an (Xn b) < an ) P (an (Xn b) an ) FZ (an ) 0.
