Stochastic_SL

Stochastic_SL - Stochastic Optimization Prof Lutz Hendricks...

This preview shows pages 1–12. 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 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: Stochastic Optimization Prof. Lutz Hendricks August 10, 2009 L. Hendricks () Stochastic Optimization August 10, 2009 1 / 58 Stochastic Optimization We add shocks to the growth model. Recursive methods are needed. The resulting model is used to study business cycles asset pricing L. Hendricks () Stochastic Optimization August 10, 2009 2 / 58 Model The household (or planner) maximizes max E ∞ ∑ t = β t u ( c t ) (1) subject to k t + 1 = f ( k t , θ t ) & c t (2) Productivity shocks take on N discrete values θ t 2 f θ 1 , ..., θ N g (3) and follow a Markov process Pr ( θ t + 1 = θ j j θ t = θ i ) = Ω ij E denotes the expectation given information at date 0. L. Hendricks () Stochastic Optimization August 10, 2009 3 / 58 What are the choice variables? The planner cannot choose sequences f c t , k t g because we don&t know the realizations of θ t . He must choose state contingent plans . For every history s t = ( s , ..., s t ) (4) where s t = ( k t , θ t ) , choose c t = c ( s t ) (5) and k t + 1 = κ ( s t ) (6) L. Hendricks () Stochastic Optimization August 10, 2009 4 / 58 Two period example L. Hendricks () Stochastic Optimization August 10, 2009 5 / 58 Two period example The problem is max E 2 ∑ t = 1 β t & 1 u ( c t ) subject to k 2 = f ( k 1 ) & c 1 c 2 = f ( k 2 , θ 2 ) θ 1 known L. Hendricks () Stochastic Optimization August 10, 2009 6 / 58 Two period example Write out the expectation explicitly: max β u ( c 1 ) + N ∑ j = 1 Pr & θ 2 = θ j ¡ β u & c 2 & θ j ¡¡ + λ 1 [ f ( k 1 , θ 1 ) & c 1 & k 2 ] + N ∑ j = 1 Pr & θ 2 = θ j ¡ λ 2 & θ j ¡ ¢ f & k 2 , θ j ¡ & c 2 & θ j ¡£ The household chooses c 1 , k 2 and c 2 & θ j ¡ . L. Hendricks () Stochastic Optimization August 10, 2009 7 / 58 Budget constraints Note the constraint terms: Pr & θ 2 = θ j ¡ λ 2 & θ j ¡ ¢ f & k 2 , θ j ¡ & c 2 & θ j ¡£ (7) This looks like the household only needs to satisfy the constraint with some probability. Not true - the constraint must hold in every history. But this is just notation. De&ne ˜ λ 2 & θ j ¡ = Pr & θ 2 = θ j ¡ λ 2 & θ j ¡ to see this. L. Hendricks () Stochastic Optimization August 10, 2009 8 / 58 First-order conditions c 1 : u ( c 1 ) = λ 1 k 2 : λ 1 = ∑ Pr & θ 2 = θ j ¡ λ 2 & θ j ¡ f k & k 2 , θ j ¡ c 2 & θ j ¡ : β u & c 2 & θ j ¡¡ = λ 2 & θ j ¡ Euler equation: u ( c 1 ) = β ∑ Pr & θ 2 = θ j ¡ u & c 2 ( θ j ) ¡ f k & k 2 , θ j ¡ = β E ¢ u ( c 2 ) f k ( k 2 , θ 2 ) j θ 1 £ L. Hendricks () Stochastic Optimization August 10, 2009 9 / 58 Many periods L. Hendricks () Stochastic Optimization August 10, 2009 10 / 58 Many periods A history of length t is s t . The household chooses c & s t ¡ and k & s t ¡ to maximize ∑ s t p & s t ¡ β t u & c & s t ¡¡ subject to x & s t ¡ + c & s t ¡ = f & k & s t ¡ , θ & s t ¡¡ , 8 s t k & s t + 1 , s t ¡ = x & s t ¡ , 8 s t , s t + 1 The last constraint ensures that k & s t + 1 ¡ is the same for all s t + 1 .....
View Full Document

{[ snackBarMessage ]}

Page1 / 58

Stochastic_SL - Stochastic Optimization Prof Lutz Hendricks...

This preview shows document pages 1 - 12. Sign up to view the full document.

View Full Document
Ask a homework question - tutors are online