European Integration and Regional Specialization

European Integration and Regional Specialization

Carlo MenonCa’ Foscari Venice

LSE

Background

• EU (Krugman’s) paradox: is integration fostering more diversification?

• OCA and asymmetric shocks• Models of endogenous growth and NEG• Intra-industry trade• Empirical issues: Krugman and Locational

Gini indexes, geographical units, tradable/not tradable, service sector, data

Distributive approach• First proposed by D. Quah (1993) as an alternative

to the “β-convergence” test• Based on the representation of the stochastic kernel

operator Mu*, which codify the relation between the probability functions associated to the distribution F at time T and t :

• It can be represented by a Markov-Chain process or by a 3D graph.

• Points of strength: focus on distributional dynamics and multimodality, possibility of applying conditional schemes (i.e. spatial dependency).

)(* tuT fMf

Distributive approach to regional specialization

• For a given industry, analysis of the evolution of the distribution at time t and T of the share of employment in each region, indexed for each period by the average value of all regions

• Stochastic kernel: 3D plot of P(fT|ft)• Possibility of applying different conditional

schemes (national values, spatial contiguity, trade volumes...)

Example: Textile 1990-2001,EU-15, NUTS2

Regional sectoral labour productivity

• Distributive analysis of dynamics of GVA/L for sub-national units, inside each sector (surprisingly rare!)

• Speculating that manufacturing and service sector have different spatial properties, the composition effect may hide relevant results

• First (raw) outcomes: divergence for manufacturing, persistence/divergence

for service sector After applying a “contiguity” conditioning scheme:

clear convergence for manufacturing, persistence for service sector

Next steps and research problems:specialisation

• Locational dynamics of each manufacturing branch

• Spatial properties of service sector• Work on data in order to overcome the NUTS2

classification (toward FUR or continuous space)• Regional Intra-industry trade • Accounting for functional urban specialization

and demanufaturing• Links to other relevant issue of economic

geography (cities hierarchy, convergence...)

Next steps and research problems:sectoral convergence

• Improve the empirical (distributional) approach to regional sectoral labour productivity convergence

• Explore and explain the differences between service and manufacturing

• Shadow effect (negative autocorrelation?)• Overcome the NUTS2 classification (toward

FUR or continuous space)

Thanks

CONVERGENCE

PERSISTENCE

DIVERGENCE