Building Conditional Quantitative Foresight: the MASST-3 Model Roberto Camagni and Roberta Capello Politecnico di Milano, ABC Department

Transcription

1 Building Conditional Quantitative Foresight: the MASST-3 Model Roberto Camagni and Roberta Capello Politecnico di Milano, ABC Department ESPON Seminar Scenarios and modelling in the framework of exploring Territorial Cohesion, Brussels, 4 September 2014

2 Objectives of the presentation To present the tool for conditional quantitative foresight (not forecast), i.e. the MASST3 econometric model, on which the ESPON ET2050 Scenarios are based. To present the methodology for conditional quantitative foresights. To present the Baseline Scenario assumptions. To present the results for European regions up to

3 The foresight tool: the MASST-3 model (1) MASST is an econometric forecasting model. Previous versions of MASST developed for ESPON 3.2 (2004-5); ESPON SPAN ( ); DGRegio projects. The ET2050 project is based on a new version of MASST, considering the economic crisis (two periods of model estimation), public budget constraints (deficit and debt), innovation modes, the role of urban areas in regional growth. The model is able to simulate effects on regional growth of: - the economic crisis; - macroeconomic elements (public budget constraints, sovereign debt, spread in interest rates on public bonds, exchange rates); - territorial capital (innovativeness, trust, agglomeration economies, HC); - Spatial spillovers; - cohesion and infrastructure policies. 3

4 The foresight tool: the MASST3 model (2) The logics of the model is at the same time : -Top-down, i.e. distributive: national growth is distributed among regions (demand side), and - Bottom-up, i.e. generative: regions add a differential effect, determined by presence of territorial capital, feeding-back on national performance (supply side). Quantitative foresights are produced for all NUTS2 regions of all 27 EU countries (270 regions). 4

5 The MASST-3 model (3) Submodel 1: National component FINAL OUTCOME Submodel 2: Regional differential component internal consumption in national GDP Tax rates Macroeconomic elements internal consumption Territorial capital assets Innovativeness: - product/ process innovation Regional innovation Human capital R&D Urbanization economies (LUZ population) Traditional urban benefits: - quality of life - creativity Unemployment growth Policies: Structural funds investments in national GDP Δ interest rates Δ Unit Labour Costs Δ of FDI stock imports in national GDP investments National component of regional growth National growth Final economic effect Regional growth as a result of MAcroeconomic Social, Sectoral and Territorial components Regional differential component Differential shift Localization economies: Hirschman Herfindahl Index Urbanization economies: LUZ population Sectoral component: dynamics of sectoral structure Innovation patterns Dynamics of sectoral structure Regional specialization Traditional urban costs: - cost of the city - social conflict Nonconvention al urban benefits: - city networks - high level urban functions Accessibility: - infrastructure FDIs/Population Functions Δ of FDI stock Inflation imports Legend: Inter-sectoral productivity: - infrastructure - skills - energy cons. Functions MIX effects Nonconventional urban costs: - diffused urban form Sectoral component: Dynamics of sectoral structure Exchange rates exports Δ Unit Labour Costs exports Endogenous variables Exogenous variables Relations in estimations Social component: - trust FDIs/Population Policies: Structural funds Territorial structure: settlement Migration Flows Unemployment rate Spatial and territorial structure: - spatial spillovers Inflation Relations in simulation Settlement structure Δ Exchange rates GDP growth in USA, Japan and BRICs Spatial and territorial structure: - spatial spillovers Population growth Birth rate Mortality rate Regional differential GDP Δ public expenditure Δ public Migration flows in national GDP expenditure Public exp'enditure Interest payments Tax revenues All equations are differentiated between periods of crisis and of no crisis Permanent income and long-run relationships are assumed and estimated 5

6 Methodology for building conditional quantitative foresights The methodology for building conditional quantitative foresight is made of subsequent steps: starting from a seminal idea about the driving forces believed to characterize future economic-territorial development, the basic characteristics of a scenario are built, together with the relevant conditional elements, the most likely bifurcations in the driving forces (qualitative assumptions); these conditional elements are inserted into the econometric model, as hypotesized values of the independent variables of the model (quantitative assumptions: levers of the model); identifying the magnitude of the most likely effects on European regions through a simulation procedure (conditional quantitative foresights). 6

7 Conditional elements: integrated scenarios A Scenario is an integrated vision of the different driving forces that are expected to have effects on future trajectories. Therefore: - individual driving forces must be related to each other, and cross feed-back effects must be underlined, i.e. the assumptions have to be highly consistent: this overall coherence has to be reflected in the label given to the scenario; - one has to assume an if... then.. logic, keeping assumptions carefully separated from effects; - the assumptions on the driving forces should be as differentiated as possible, sometimes even opposite to each other, so as to yield differentiated images. 7

8 Assumptions of the Baseline Scenario (2030) - the socio-economic and demographic trends of the past will continue, and no major change (beyond the crisis) will alter the EU economy; - economic policies will remain the present ones (stable budget for SFs); - a general slow economic recovery will start in 2016; - a slight increase in competitiveness of European countries is assumed in 2030; - interest rates on bonds will return back to lower, pre-crisis values, thanks to the end of strong financial speculation; - the stability pact remains the same, imposing highly restrictive fiscal policies. n.b. These assumptions were agreed in December 2011, and revisited after Draghi s interventions in July/September 2012, which stopped the speculations on the euro. The final conditional quantitative foresights have been run in early

9 RESULTS FOR THE BASELINE SCENARIO (2030)

10 Aggregate results of the Baseline scenario Average annual GDP growth rate Average annual population growth rate Average annual employment growth rate Average annual manufacturing employment growth rate Average annual service employment growth rate EU Old New The New12 countries grow a little more than the Western countries. 2. New12 countries increase employment in services more than in manufacturing, entering a new stage of development. 3. Western countries show a balanced growth between manufacturing and services. 10

11 Baseline: annual average GDP growth rate Reykjavik Baseline March Average regional GDP growth rate < > 2.95 Dublin Oslo København Stockholm Helsinki Tallinn Riga Vilnius Minsk Canarias Guadeloupe Guyane Madeira Martinique Réunion Two speed Europe; Southern peripheral countries grow less than Northern countries. Southern European countries discount the difficult present conditions on their future evolutionary trajectories. London Paris Amsterdam Bruxelles/Brussel Luxembourg Bern Vaduz Berlin Praha Ljubljana Zagreb WienBratislava Budapest Sarajevo Warszawa Beograd Bucuresti Kyiv Kishinev Acores Eastern European countries still grow more than the EU 15, but this is not enough to catch up the GDP per capita levels of the Western countries in Lisboa Ar Ribat Madrid El-Jazair Tounis Politecnico di Milano, Project ET2050, 2013 Roma Valletta Podgorica Tirana Skopje Sofiya Athinai Ankara Nicosia km Regional level: NUTS2 Source: Politecnico di Milano, 2013 Origin of data: - MASST3 model EuroGeographics Association for administrative boundaries Intra-national regional disparities increase. For the first time, inter-national ones do not shrink significantly and do not lead to decreasing total disparities. 11

12 Theil index in the Baseline scenario (2030) Total regional disparities Inter-national disparities Intra-national disparities Total Theil Index Between Countries Theil Index Within Countries Theil Index 12

13 EXPLORATORY SCENARIOS (2030)

14 Summary of assumptions for the exploratory scenarios Megas scenario Market driven scenario; welfare system fully privatized; financial debt repaid in 2030; budget reduced for cohesion policies; concentration of investments in European large cities. Cities scenario Public policies mostly at national level; actual welfare system reinforced through increased taxation; financial debt not fully repaid in 2050; budget maintained for cohesion policies; concentration of investments in second rank cities. Regions scenario Social policies; strong public welfare system; financial debt repaid in 2050; budget significantly increased for cohesion policies; concentration of investments in rural and cohesion areas. 14

15 Aggregate GDP growth results for the exploratory scenarios Aggregates Baseline Megas Cities Regions Megas vs. baseline Cities vs. Baseline Regions vs. Baseline EU old new The Cities scenario is the most expansionary: territorial capital and the urban system are better exploited than in the other scenarios. 2. This holds also for New 12 countries. 3. New 12 countries are those that gain in the regions scenario with respect to the baseline. 15

16 Theil Index by Scenario: Total Regional Disparities Baseline scenario Megas scenario Cities scenario Regions scenario The Cities scenario is also the most cohesive (disparities increase less)

17 Theil Index by Scenario: Between Country Disparities Baseline scenario Megas scenario Cities scenario Regions Scenario The Cities scenario is the most cohesive (disparities shrink more)

18 Theil Index by Scenario: Inside Country Disparities Baseline scenario Megas scenario Cities scenario Regions scenario In this case, the Regions scenario is the most cohesive (et pour cause)

19 Conclusions 1. Difficult times ahead for inter-regional disparities (due to centralization trends, macroeconomic constraints), 2. A scenario supporting cities (first, second and possibly third rank cities) looks the most desirable: in terms of growth potential (best use of concentrated diffusion of territorial capital) and in terms of territorial cohesion 3. Relevant policy tasks for CEECs: - keep wage and price increases in line with productivity growth - launch a wave of endogenous capital accumulation - find a new innovation pattern, based on a smart specialization and on selected inter-regional cooperations. 19

20 THANK YOU VERY MUCH FOR YOUR ATTENTION 20