{[ promptMessage ]}

Bookmark it

{[ promptMessage ]}

# 1 include more explanatory variables in the

This preview shows pages 9–18. Sign up to view the full content.

l Desire to make ceteris paribus interpretations of 1  include more explanatory variables in the regression

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

View Full Document
10 Engel curve for food l Sample of 40 households with 3 family members l Data on l Y = weekly expenditure on food (\$) l X = weekly income (\$) l Cross-sectional data l Engel curve (demand function holding prices constant) Y i = β 0 + 1 X i + ε i i =1, … , n 0 10 20 30 40 50 60 0 20 40 60 80 100 120 140 Income Expenditure
11 Engel curve for food… l For these data l b 0 = 7.38 & b 1 = 0.23 l How do you interpret b 1? l What is predicted food expenditure for a household with weekly income of \$100? l Why restrict to households with 3 people?

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

View Full Document
12 Assumptions of Classical Linear Regression Model l A1: Linear model l Y i = β 0 + 1 X i + ε i , i = 1, … , n l A2: Random sampling l Have a random sample ( Y i , X i ) governed by model in A1 l A3: Sample variation in X l Values of X 1, X 2, … , X n are not all equal l A4: Zero conditional mean l E( i | X i ) = 0 l Key assumption because implies i & X i are uncorrelated l A2 & A4 mean we can think of X as fixed rather than random
13 Assumptions of Classical Linear Regression Model… l A5: Homoskedasticity l Var( ε i ) = σ 2 l A6: Disturbances uncorrelated l Cov( i , j ) = 0, ( i not equal j ) l A7: Disturbances are normal (used for inference) l i ~ N (0, 2 ) l In some circumstances these assumptions will not be realistic l Further study of econometrics (beyond BES) l Would cover detection of violations of assumptions l Subsequent remedial methods

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

View Full Document
14 Classical Linear Regression Model (CLRM) l Given A1-A7 l Y i ~ N ( β 0 + 1 X i , σ 2 ) l The (conditional) mean of Y depends on X l If 1=0 what is 0? l Given A1-A4 can show that OLS estimators are unbiased & consistent l Given A1-A6 OLS ‘best’ in a particular sense but beyond scope of BES
15 Explaining variation in the dependent variable w w w w Y X X Y Y Y ˆ i - i i Y ˆ Y - X b b Y 1 0 ˆ + =

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

View Full Document
16 Decomposition of variance ( 29 ( 29 ( 29 ( 29 ( 29 SSE SSR SST e Y Y Y Y X e Y Y Y Y Y Y Y Y i i i i i i i i i + = + = + - = - + = + - = - + - = - squares of sum error squares of sum regression squares of sum Total ˆ that show Can d unexplaine part by explained part deviation Total ˆ ˆ ˆ 2 2 2
17 Standard error of estimate l Population variance σ 2 measures the spread around the population regression line l Standard error of estimate ( SEE ) is an estimator of l It measures fit of regression model l Low SEE  good fit estimator unbiased an ensure to estimated) be to need that parameters 2 the (for adjustment freedom of degrees a is 2 of Divisor 2 2 where 1 2 2 β - - = - = = = n n SSE n e s s SEE n i i

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.

{[ snackBarMessage ]}

### Page9 / 31

1 include more explanatory variables in the regression l...

This preview shows document pages 9 - 18. Sign up to view the full document.

View Full Document
Ask a homework question - tutors are online