S value of amenities a jtis s i a jt β a zx a jt z

This preview shows page 52 - 59 out of 61 pages.

’s value of amenitiesAjtis:si(Ajt)=βAzxAjt|{z}ExogenousAmenities+βeduzajt|{z}EndogenousAmenities+βstzstixstj+βdivzdivixdivj|{z}Taste for Birth Location+σzεijt|{z}IdiosyncraticTasteεijtType I Extreme Value.Izindexed three groups of workers (β= fct of demographics)Econ 280D. Spring 2019. C. GaubertLecture 1Agglomeration: Evidence50 / 59
Increasingly Skilled MSAs: Some Amenity ImprovementsEating and DrinkingPlaces per 1000ResidentsPropery Crimes per1000 ResidentsEPA Air QualityIndexGov K-12 Spendingper StudentΔCollege Emp Ratio0.182***-0.231*-0.539***0.129**[0.0539][0.122][0.171][0.0639]MSA College Ratio Changes on Amenity Changes: 1980-2000Diamond(Stanford GSB)Geographic Skill Sorting14 /Econ 280D. Spring 2019. C. GaubertLecture 1Agglomeration: Evidence51 / 59
Demand for Cities & Amenity SupplyDemand for Cities & Amenity SupplyNon-CollegeCollegeNon-CollegeCollegeNon-CollegeCollegeWage4.155***5.523***3.757***-1.783***4.026***2.116***[0.603][1.797][0.561][0.682][0.727][1.146]Rent-2.418***-1.404-2.329***1.105***-2.496***-1.312***[0.349][0.833][0.348][0.423][0.451][0.711]ExpenditureShare0.582***0.254**0.620.620.620.62[0.0678][0.078]----Amenity Index----0.274*1.012***[0.147][0.115]College EmpRatio[3]A. Worker Preferences for CitiesD. Amenity Supply2.60**[2][1]Birth State EstimatesDiamond(Stanford GSB)Geographic Skill Sorting35 / 48IIncluding amenities is necessary to get reasonable estimates (sign, pricesensitivity of college vs non college)ICollege workers value amenities moreEcon 280D. Spring 2019. C. GaubertLecture 1Agglomeration: Evidence52 / 59
What are the determinants of the ”Great Divergence”?ProductivityIBack out initial/final unobserved productivity/amenity levels that rationalizeinitial/final distribution, given estimated elasticitiesIModel based decomposition:1.Compute counterfactual wage change and population change if all thatchanged wasexogenousproductivity levels between 1980-2000(rents/amenities held constant):2.Allow wages and rents to change from their 1980 levels to those observed in2000, but hold amenities (both endogenous and exogenous) fixed to their1980 level.3.Allow wages, rents, and endogenous amenities to change from their 1980levels to those observed in 2000, hold exogenous amenities fixedICompare expected migration to actual.Econ 280D. Spring 2019. C. GaubertLecture 1Agglomeration: Evidence53 / 59
Productivity Changes Closely Predict Migration:1980-2000-1.5-1-0.500.511.52-0.4-0.200.20.40.60.81Predicted Ln(College Ratio) ChangeActual Ln College Ratio Changeρ=0.80Figure 2.A: Predicted change in ln college ratio due onlyto productivity changesstrong predictive power of exogenous productivity changesEcon 280D. Spring 2019. C. GaubertLecture 1Agglomeration: Evidence54 / 59
Predicted College Ratio Changes From Observed Wage &Rent Changes-0.4-0.200.20.40.60.811.2-0.4-0.200.20.40.60.81Predicted Ln(College Ratio) ChangeActual Ln College Ratio Changeρ=0.32Figure 2.B: Predicted change in ln college ratio due toobserved wage & rent changesproductivity + rent changes alone are not enough ; need amenitiesEcon 280D. Spring 2019. C. GaubertLecture 1Agglomeration: Evidence55 / 59
Predicted College Ratio Changes from Wage, Rent, &Endogenous Amenities Changes-1.5-1-0.500.511.52-0.4-0.200.20.40.60.81Predicted Ln(College Ratio) ChangeActual Ln College Ratio Changeρ=0.86

  • Left Quote Icon

    Student Picture

  • Left Quote Icon

    Student Picture

  • Left Quote Icon

    Student Picture