{[ promptMessage ]}

Bookmark it

{[ promptMessage ]}

Econometrics-I-23

Econometrics-I-23 - Applied Econometrics William Greene...

Info iconThis preview shows pages 1–9. Sign up to view the full content.

View Full Document Right Arrow Icon
Applied Econometrics William Greene Department of Economics Stern School of Business
Background image of page 1

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

View Full Document Right Arrow Icon
Applied Econometrics 23. Sample Selection
Background image of page 2
Samples and Populations Consistent estimation The sample is randomly drawn from the population Sample statistics converge to their population  counterparts A presumption:  The ‘population’ is the  population of interest. Implication: If the sample is randomly drawn  from a specific subpopulation, statistics  converge to the characteristics of that  subpopulation
Background image of page 3

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

View Full Document Right Arrow Icon
Nonrandom Sampling Simple nonrandom samples:  Average incomes of  airport travelers   mean income in the population as a  whole? Survivorship: Time series of returns on business  performance.  Mutual fund performance.  (Past  performance is no guarantee of future success.   ) Attrition:  Drug trials.  Effect of erythropoetin on quality  of life survey. Self-selection:   Labor supply models Shere Hite’s (1976) “The Hite Report” ‘survey’ of sexual habits  of Americans. “While her books are ground-breaking and  important, they are based on flawed statistical methods and  one must view their results with skepticism.”
Background image of page 4
Heckman’s Canonical Model A behavioral model: Offered wage       = o*  =  'x+ v    (x age,experience,educ...) Reservation wage = r*   =   'z +  u  (z = age, kids, family stuff) Labor force participation:                       LFP = 1  β = δ 2 2 v u if o*    r* , 0 otherwise                    Prob(LFP=1) =  ( 'x- 'z)/ Desired Hours      = H*  =  'w  +  Actual Hours        = H*   if LFP =  1                              unobserved if LFP = 0   Φ β δ σ + σ γ ε ε and u are correlated.    and v might be correlated. What is E[H*  | w,LFP = 1]?  Not  'w. ε γ
Background image of page 5

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

View Full Document Right Arrow Icon
Standard Sample Selection Model i i i i i i i i i i i 2 i i i i i i i i d * ' z u d   =  1(d *  >  0) y *  =  'x + y    = y *  when d  =  1, unobserved otherwise (u ,v ) ~  Bivariate Normal[(0,0),(1, , )] E[y | y  is observed] =  E[y|d= 1]                             =   'x + E[ | = α + β ε ρσ σ β ε i i i i i i i i i d 1]                             =   'x + E[ | u ' z ] ( ' z )                             =   'x + ( ) ( ' z )                             =   'x+ = β ε φ α β ρσ Φ α β θλ
Background image of page 6
Incidental Truncation u1,u2~N[(0,0),(1,.71,1) Conditional distribution of u2|u1 > 0. No longer ~ N[0,1] Unconditional distribution of u2 ~ N[0,1]
Background image of page 7

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

View Full Document Right Arrow Icon
Selection as a Specification Error E[y i |x i ,y i  observed]  =  β ’x i  +  θ   λ i Regression of y i  on x i  omits  λ i .   λ i
Background image of page 8
Image of page 9
This is the end of the preview. Sign up to access the rest of the document.

{[ snackBarMessage ]}