110 likes | 214 Views
Growth and Volatility in EU Regions: Does Space Matter?. Author : Vicente Rios Universidad Pública de Navarra. OUTLINE. Scientific Problem Theoretical Framework Quantitative Analaysis Contribution. 1. The Scientific Problem.
E N D
Growth and Volatility in EU Regions: Does Space Matter? Author: Vicente Rios Universidad Pública de Navarra
OUTLINE ScientificProblem Theoretical Framework QuantitativeAnalaysis Contribution
1. TheScientificProblem Whatistherelationshipbetweenvolatility and growth in Europeanregions? Empirical studies in this topic at the regional level: i) are few ii) do not account for unobserved spatial heterogeneity iii) reach diverging conclusions Focus: i) Re-Examine the relationship between volatility and growth (spatial panel) ii) Explore the role played by spatial spillovers and interdependences
2. Theoretical Framework There are many reasons to believe that: i) volatility and growth are connected Positively: Schumpeter (1939); Mirman (1971), Black (1987); Bean (1990); Hall, (1991); Saint-Paul (1993); Helpman and Trajtenberg(1998), Negatively: Pindyck (1982);Bernanke(1983) Ambiguously: De Hek (1999), (2002); Blackburn and Galindev (2003); Manuelli and Jones (2005); Galindev (2007) ii) space might be a channel of difussion(with some frictions) Spatially Augmented Growth Models: López-Bazo et. al (2004), Erturand Koch (2007), M. Fisher (2009) iii) Work in progress derive a SDM equation from • stochastic endogenous growth model with • spatial diffussion, iid productivity shocks, CES function
3. QuantitativeAnalysis Sample: 198 NUTS 2 Regions (EU 13+ Norway + Switzerland) Time:1980-2010 Model Y: GDP pc growth rate Key variable: Volatility (Std GDP pc growth rate) Controls: Investment, Initial GDP pc, Industry Mix, Pop, Agglomeration Methodology • Estimation and Selection of Spatial Panel Data • Simulation: Internal to region + Neighbor’s volatility effect • Robustness Analysis
3. QuantitativeAnalysis MixedApproach: recommendedbyElhorst(2010) - Estimate non spatialmodels • Use RobustLagrandeMultiplierTeststocheckthespatialdependenceform (Spatial Error, SpatialLag). • W Matrix: distancematrix, exogeneity Result 1: Rejectthenull of no spatialdependence in all cases Result 2: Time Effects are jointlysignificant (notshown)
3. QuantitativeAnalysis CheckifSpatialDurbin can be simplifiedtoSpatialLagorSpatial Error versions ModelSelectionTests Result 4: SelectedModelisSpatialDurbin
3. QuantitativeAnalysis Key Result 1: Volatility and Growth are positivelyrelated Key Resul 2: Spatialspilloversaccountforhalf of theimpact
3. QuantitativeAnalysis: Robustness Results are robusttodifferentW’s
4. Contributions Europeanregionswithhighvolatilitytendtogrowfaster Spatialspillovers are a key factor reinforcingtheeffect of internalvolatilityongrowth Theresultisrobusttospatialweightmatrixes Expected (Link Theory-EmpiricalModel)