slides_5_conddist

# Now y − y 2 y − x x − y y − x 2 2 y − x x

This preview shows pages 59–65. Sign up to view the full content.

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: Now Y − Y 2 Y − X X − Y Y − X 2 2 Y − X X − Y X − Y 2 Taking the expected value conditional on X E Y − Y 2 | X E Y − X 2 | X 2 X − Y X − Y 2 Var Y | X E Y | X − Y 2 59 Now, Y E E Y | X by LIE. Take expectations with respect to X , use LIE again, and the definition of a variance: E Y − Y 2 E Var Y | X E E Y | X − Y 2 E Var Y | X Var E Y | X 60 EXAMPLE : Suppose Y | Z ~ Poisson Z where Z 0 is a continuous random variable. Then E Y | Z Z Var Y | Z Z The unconditional mean and variance are E Y E Z Var Y E Z Var Z Without completely specifying D Z , we cannot find D Y . But we know the first two moments of Y . 61 ( cv4 ) We can write a variance conditional on a smaller set of variables in terms of conditional means and variances on a larger set of variables: Var Y | X E Var Y | X , Z | X Var E Y | X , Z | X ∙ This feature is very useful for generating overdispersion in conditional models. Often Z is an unobserved random variable (or variables) that need to be “integrated out” to obtain usable conditional mean and variance functions. 62 EXAMPLE : Suppose D Y | X , U Poisson U exp X , where U 0 is independent of X with E U 1 and Var U U 2 . Such models arise often in econometrics, where X are observable factors (for a firm, say, spending on research and development, sales, profitability) and U are unobserved attributes (managerial “talent,” workplace environment). We want to find features of D Y | X . 63 ∙ We can obtain the conditional mean and conditional variance: E Y | X E E Y | X , U | X E U | X exp X exp X Var Y | X E Var Y | X , U | X Var E Y | X , U | X E U exp X | X Var U exp X | X E U | X exp X Var U | X exp X 2 exp X U 2 exp X 2 E Y | X U 2 E Y | X 2 ....
View Full Document

{[ snackBarMessage ]}

### Page59 / 67

Now Y − Y 2 Y − X X − Y Y − X 2 2 Y − X X − Y X...

This preview shows document pages 59 - 65. Sign up to view the full document.

View Full Document
Ask a homework question - tutors are online