Econ 452 Comparative Statics Comparative Static Analysis 1. What? - compare equilibria - model gives equilibrium conditions - first order conditions - frequently in implicit form - if exog. variables change, how does new eq'm compare to original? - check changes in endog. variables wrt exog. variables through use of eq'm conditions . 2. Why? - comparative dynamics are hard - belief that economic system is in equilibrium Example: basic consumer theory: - 2 goods, (x,y); - budget set given by { p x , p y , M } - choose x,y to max u(x,y) s.t. px x + p y y ≤ M - properties of demand functions? Page 1 of 10 Econ 452 Comparative Statics 1. set up Lagrangean, obtain FOC's ux ( x, y) − λ px = 0 u y ( x, y) − λ py = 0 px x + p y y − M = 0 These 3 equations implicitly define Marshallian/Walrasian demand functions x ( p x , p y , M ), y ( p x , p y , M ), and the marginal utility of income, λ ( p x , p y , M ). 2. Totally differentiate equations wrt exog variables ( px , p y , M ) and endog variables ( x, y,λ ) ; write in matrix form. Page 2 of 10 Econ 452 Comparative Statics u xx u yx − px u xy u yy − py − px dx λ dpx − p y dy = λ dp y 0 d λ xdpxx + ydp y − dM Determinant of LHS matrix can be signed: SOC for constrained maximization > 0 3. Can find slopes of demand curves and engel curves using Cramer's rule. Alternative (if only two choice variables): - use constraint-as-equality (from theory) to eliminate one variable here: y = (Μ − px x)/ py so u( x, y) = u( x, y( x)) = u( x, M −p px x ) y - treat as unconstrained max'm Page 3 of 10 Econ 452 Comparative Statics problem: FOC is du dy = u x ( x, y ) + u y ( x, y ) dx dx px = u x ( x, y ) + u y ( x, y )(− ) = 0 py ux px = This yields FOC for max: u py y - once again, can derive properties of demand functions, recognizing that FOC defines x ( p x , p y , M ) , and x ( p x , p y , M ) = ( M − px x ( px , p y , M )) / p y Sometimes, need to do more…: Demand for insurance - with no asymmetric information: Contingent commodities / state space Consider the standard insurance model - accident occurs with prob. 0 < p < 1 - loss of L Page 4 of 10 Econ 452 Comparative Statics - Individual's endowment point, E, (if buys no insurance): - if accident (state 1), ind'l has y = y − L; 1 0 - if no accident(state 2), ind'l has y2 = y0 - If individual buys insurance paying out I in state 1, at cost δ per dollar: - state 1: has y1 = y0 − L −δ I + I = y0 − L + (1−δ )I - state 2: has y = y −δ I 2 0 EU without insurance? EU = pU ( y0 − L) + (1− p)U ( y0) EU with insurance? EU = pU ( y0 − L + (1−δ )I ) + (1− p)U ( y0 −δ I ) Optimal insurance? Choose I to max Page 5 of 10 Econ 452 Comparative Statics ∂EU = pU '( y )(1−δ ) −δ (1− p)U '( y ) = 0 1 2 ∂I I* satisfies pU '( y1) = δ (1− p)U '( y2) 1−δ What is value of delta? Cost of insurance: Assume: 1. identical individuals 2. risk neutral firms 3. competitive insurance markets 4. no administrative costs (1) + (2) implies firm max's expected profit on representative contract; p(δ I − I ) + (1− p)δ I = I (− p(1−δ ) + (1− p)δ ) Eprofit = (3) implies Eprofit = 0 Therefore: in competitive insurance market, Page 6 of 10 Econ 452 Comparative Statics p = δ (1− p) (1−δ ) Sub'g this back into FOC for consumer, tells us I* satisfies U '( y1) =1 : RA individual U '( y2) will buy full insurance if "fair insurance" is available. Suppose insurance is unfair, so δ > p . Then Eprofit > 0 on the average contract (assuming no fixed costs). Will consumer still want full insurance? Comparative statics of insurance: From above, FOC for optimal insurance is Page 7 of 10 Econ 452 Comparative Statics pU '( y0 − L + (1−δ )I *)(1−δ ) −δ (1− p)U '( y −δ I *) = 0 0 Rearranging: (1 − δ )U '( y0 − L + (1 − δ ) I *) (1 − p ) = p δ U '( y0 − δ I *) This defines demand for insurance as the implicit function I * = f ( y , L,δ , p) . 0 Properties of this function? 1. How much insurance? i) if δ = p , ind'l buys full insurance at all income levels. ii) if δ > p , so insurance is unfair (and ins. firm makes strictly positive profit), then ind'l buys less than full insurance, so y < y . 1 2 2. How does insurance vary with y? - totally differentiate FOC wrt I* and endogenous & exogenous variables - look at partial wrt initial income: Page 8 of 10 Econ 452 Comparative Statics dI * = δ (1− p)U "( y2) − p(1−δ )U "( y1) dy0 (1−δ )2 pU "( y1) + δ 2(1− p)U "( y2) Sign? denominator <0, numerator….? - sign depends on relative sizes of U", evaluated at different income levels. - coeff. of ARA useful - to use, need to bring in U'( ) - how? i) rearrange FOC to solve for p(1−δ )U '( y1) δ (1− p) = U '( y2) (dI */ dy0) to obtain U "( y2) U "( y1) p(1−δ )U '( y1)[ ] − U '( y2) U '( y1) ii) sub. into Page 9 of 10 Econ 452 Comparative Statics = p(1−δ )U '( y1)[ A( y1) − A( y2)] iii) sign depends on term in [ ], which is coeff. of ARA at two values; - have y > y , so… 2 1 - CARA - no change in insurance - DARA - I* down as y up - IARA - I* up as y up Page 10 of 10
© Copyright 2024 Paperzz