Presentation is loading. Please wait.

Presentation is loading. Please wait.

Modelling Cohesion Policy in a DSGE Model with Semi-Endogenous Growth: simulations with the QUEST III model Janos Varga and Jan in t Veld Research Directorate.

Similar presentations

Presentation on theme: "Modelling Cohesion Policy in a DSGE Model with Semi-Endogenous Growth: simulations with the QUEST III model Janos Varga and Jan in t Veld Research Directorate."— Presentation transcript:

1 Modelling Cohesion Policy in a DSGE Model with Semi-Endogenous Growth: simulations with the QUEST III model Janos Varga and Jan in t Veld Research Directorate DG ECFIN - Economic and Financial Affairs European Commission Evaluation Conference, Warsaw, 30 November 2009 The views expressed here are those of the authors and should not be attributed to the European Commission. EUROPEAN COMMISSION DIRECTORATE GENERAL ECONOMIC AND FINANCIAL AFFAIRS

2 2 Cohesion Policy EU Cohesion Policy one of the key pillars of the European Union. Designed to increase economic and social cohesion among member states, enhancing a faster catch-up process of the less developed member states in terms of income per capita. Structural and Cohesion Funds are now the second- largest item in the budget, receiving around one-third of the total EU budget. The resources are targeted on public and private investment in physical and human capital, Lisbon Strategy became more and more the leitmotiv of many EU policies and this was the momentum for a paradigm shift in Cohesion Policy.

3 3 Structure of the presentation 1.Dynamic Stochastic General Equilibrium (DSGE) models 2.QUEST III model with endogenous growth 3.Simulations of Cohesion spending programme period 2000-06 4.Sensitivity analysis 5.Conclusions

4 4 Empirical evaluations EU Cohesion policy: Boldrin and Canova (2001) : no evidence of any form of systematic catching-up with the rest of regional income distribution. Cappelen et al. (2003): EU regional policy has become more effective in generating growth and contribute to greater equality in productivity and income in Europe. However, growth in poorer regions is hampered by an unfavourable industrial structure (dominated by agriculture) and lack of R&D. => fiscal transfers should be accompanied by policies that facilitate structural change and increase R&D capabilities in poorer regions. Ederveen et al. (2002, 2006): Structural Funds are – on average - ineffective. However, can be effective when 'right' institutions (conditioning variables like openness, institutional quality, corruption and indicators of good governance) Checherita, Nickel and Rother (2009): Trade-off: Fiscal transfers contribute to reducing disparities in income but they also impede output growth, i.e. there is a negative impact of net transfers on growth in receiving regions and a negative impact of net taxes on growth in donor regions => immiserising convergence : output growth rates in receiving poor regions declining by less than in paying rich regions.

5 5 Why model-based evaluations of Cohesion policy? Take into account cross economy spillovers of policies - General equilibrium effects Models can provide coherent and internally consistent framework to analyse channels through which policies have effect Account for international spillovers Dynamic profile – adjustment costs But: Dependent on efficient use of funds: assuming no waste, no sub-optimal use Model simulations can only indicate the potential effects Reality may be different (absorption problems)

6 6 Absorption Problems (Herve and Holzmann, 1998) 1.Waste of transfers (projects with zero or negative economic return) 2.Administrative costs- extra resources for programming and monitoring, cannot be used for increasing the productive capacity of the economy. 3.Rent-seeking activities: incentive to invest resources in unproductive activities to catch a rent in the form of share of the transfers. Competition for resources absorbs resources that can no longer be used productively. 4.Diversion of funds to consumption: increase in future consumption possibilities will lead to a higher consumption on impact (consumption-smoothing) to the detriment of investment 5.Timing related problems (time lags before returns to investment materialise, opportunity costs are high and private investment decisions may be delayed), 6.Information disadvantage of the disbursing authority (leading to support of sub- optimal investment projects), 7.Public choice considerations (leading to intentional support of suboptimal projects). 8.Changes in relative prices could lead to Dutch disease type phenomena (rising factor demand non-tradable sector leading to decline in tradable sector), 9.Immiserising growth phenomena (industrial restructuring in favour of protected subsectors, with harmful consequences for long run growth ) 10.Worsening of negative effects of market failures ( polarisation effects of transfers due to increasing returns to scale and labour market distortions). => Transfers may be detrimental to economic growth and real convergence (most likely cause: rent seeking, protectionism and market rigidities) Absorption problems are likely to increase with the amount of transfers.

7 7 Use of Models in DG ECFIN Research tools to support policy making Help to determine priorities and guide + support policy making in the European Commission in general; Quantitative analysis of macro economic issues: What are effects of shocks and/or policies on rest of economy (spillovers/interactions) ? Analysis must to be conducted in a fully coherent, disciplined and internally consistent framework => microfounded DSGE model

8 8 Use of Models in DG ECFIN (2) Some recent model applications: 1.Impact of fall of dollar on EU economy 2.Impact of oil price shocks 3.Financial crisis: boom and bust house prices, rise in spreads 4.Monetary and fiscal response to crisis, fiscal stimulus measures 5.Structural reforms (Lisbon strategy/EU2020): labour market reform, product market reform, promoting R&D 6.Climate action: introduction carbon tax, directed technical change

9 9 DSGE models used in other policy-making institutions IMF: GEM ( Laxton and Pesenti, 2003), GIMF (Kumhof and Laxton, 2007) ECB: NAWM ( Coenen, McAdam and Straub,2008 ) Fed: SIGMA ( Erceg, Guerrieri and Gust, 2006 ) Bank of Canada: BoC-GEM ( Lalonde and Muir,2007 ) OECD: OECD-Fiscal ( Furceri and Mourougane, 2009 ) European Commission: QUEST III (Ratto, Roeger and in t Veld, 2009)

10 10 Dynamic Stochastic General Equilibrium (DSGE) models (1) : DSGE models are derived from micro principles in a consistent manner - fully coherent, internally consistent framework Decisions are based on intertemporal optimisation, subject to technological, budget and institutional constraints Consistent modelling of intertemporal budget constraints for households and government plus the current account

11 11 DSGE models (2) DSGE models include nominal rigidities that give rise to imperfections in labour and product markets Explicit modelling of structural rigidities (mark ups, entry barriers, financing restrictions (credit constrained households+firms)) Role for active monetary + fiscal policy (New-Keynesian) Explicit modelling capital flows (channel of adjustment) DSGE models allow for an analysis of transition to new steady state (short term adjustment costs, short term distributional consequences) Adjustment costs in labour and capital determine dynamic adjustment Anticipatory effects of (announced) future policies

12 12 Model-based evaluations of EU Cohesion Policy using DSGE models: Varga and in 't Veld (2009): QUEST III with endogenous growth. –Impact of cohesion spending in NMS for programme period 2007-2013. –NMS treated as one region Allard et al. (2008) : GIMF model (IMF) –Comparison EU transfers to households vs. public infrastructure investment –Stronger impact of government investment on long term growth –Similar DSGE model structure, but without endogenous growth

13 13 QUEST III model Dynamic Stochastic General Equilibrium (DSGE) model: –Microfounded: decisions based on dynamic optimisation subject to technological, budgetary and institutional constraints –Two types of households: non-constrained (Ricardian) and liquidity-constrained (backward-loooking) –Representative agent or overlapping generations framework –Nominal and real rigidities –Adjustment costs –Global model (flexible regional aggregation) –Different versions: R&D, skill disaggregation, energy, housing, multi-sector –Estimated (Bayesian methodology)

14 14 QUEST III model variants

15 15 QUEST III R&D model Economy populated by: Households Final goods producing firms Intermediate goods producing firms R&D sector Monetary and fiscal authorities Disaggregation of labour force: low-, medium, high skilled (employment rate, skill efficiencies) Technological change: increasing product variety (Dixit&Stiglitz)

16 16 QUEST III R&D model Knowledge investment is key to economic growth. Disaggregation of investment into tangibles and intangibles Physical capital: –rivalrous –constant returns to scale Knowledge capital : design for production of new good –non-rivalrous (Romer, 1990) ( and knowledge spillovers) –sunk cost for firm – increasing returns to scale What polices can induce firms to increase intangible investment ? Romer (1990), Jones (1995), Aghion and Howitt (1998)

17 17 Households Ricardian (non-liquidity constrained) and Liquidity-constrained households Habit persistence Non-liquidity constrained households buy new patents of designs produced by the R&D sector rent their total stock of design to intermediate goods producers pay income tax on the period return of intangibles receive subsidies after their investment in R&D products.

18 18 Final good firms Final output is produced using a labour aggregate, L Y and A t varieties of intermediate inputs (x i,t ) with an elasticity of substitution θ.

19 19 Infrastructure investment Productivity enhancing effect of public capital K G : Investment in infrastructure I G raises total factor productivity By how much depends on α G !! (default : α G = 0.10)

20 20 Intermediate good firms The intermediate sector consists of monopolistically competitive firms enter the market by licencing a design from domestic households make an initial payment (FC A ) to overcome administrative entry barriers (tangible) capital inputs are also rented from the household sector firms which have acquired a design can transform each unit of capital into a single unit of an intermediate input entry occurs until the PDV of profits (where the discount factor contains the risk premium for intangible capital) is equal to the price of the patent (intangible) and a fixed entry cost

21 21 R&D sector Innovation corresponds to the discovery of new designs. The R&D sector hires high-skilled labour L A,t and generates new designs ΔA t according to a following knowledge production function: International R&D spillovers (Bottazzi and Peri, 2007): ω and Φ measure the foreign and domestic spillover effects from the aggregate international A* and domestic A t-1 stock of knowledge

22 22 R&D promoting policies R&D sector sells new patents of designs to households who rent them out to intermediate goods producers at rental rate i A. Households pay income tax at rate t K on the period return of intangibles and they receive tax subsidies at rate τ A i A : households require a rate of return on intangible capital which is equal to the nominal interest rate minus the rate of change of the value of intangible assets and also covers the cost of economic depreciation plus a risk premium. Government policies to promote R&D: –tax incentives in the form of tax credits/depreciation allowances or –lowering the tax on the return from patents.

23 Final Goods Intermediate Goods Entrants Household Government Mark up Credit frictions Tangibles Credit frictions Intangibles Competition Research Subsidies QUEST III RD Model Admin. Entry Barriers

24 24 Human capital accumulation Labour-aggregate composed of three skill-types: s s population share of group s L s employment rate of group s h s accumulated human capital group s Accumulated human capital h s is produced by participating in education: where Λ s is amount of time spent accumulating human capital (years of schooling- ψ Mincerian return to schooling) Additional training:

25 25 Cohesion Policy programme 2000-6 Objective 1 : to promote the development and structural adjustment of regions whose development is lagging behind;Objective 1 Objective 2 : to support the economic and social conversion of areas experiencing structural difficulties;Objective 2 Objective 3: to support the adaptation and modernisation of education, training and employment policies and systems in regions not eligible under Objective 1.

26 26 Cohesion Policy programme 2000-6 European Regional Development Fund (ERDF European Social Fund (ESF) European Agricultural Guidance and Guarantee Fund (EAGGF) Financial Instrument for Fisheries Guidance (FIFG) Total Structural Funds Cohesion Fund (CF) Total Cohesion Policy expenditure Czech Republic 0.980.400.170.001.550.812.37 Cyprus Estonia Hungary 1.230.450.310.002.000.822.82 Lithuania 0.580. Latvia 0.380. Malta Poland 4.952. Slovenia Slovakia 0.600.320. Germany 15.4711.733.650.1431.000.0031.00 Italy 17.397.803.220.3428.760.0028.76 Ireland 1.931. Portugal 13.014.882.240.2120.332.1722.50 Greece 14.364.772.690.1922.001.7923.80 Spain 26.2711.725.681.7845.448.8654.30 Total 97.6045.6319.832.95166.0020.18186.18 Bln. euros

27 27 Table 2.a Yearly payment profile 2000-2009 (m. Euros) 2000200120022003200420052006200720082009Total Czech Rep.0102753229205506610694312365 Cyprus000059152230181 Estonia081016611001591621460663 Hungary0274146239401745852423442820 Latvia010182388167175386263101140 Lithuania0123028109186226388483271490 Malta000065163321082 Poland046157173106499521273367314729211369 Slovakia04243115320129843150701649 Slovenia02119325997781068403 Germany10183053334733854032430442263998351312631002 Greece02238144614082547243134344678545316423798 Ireland208458614552537431407245300113763 Italy1512609157034733842412943734355434654628755 Portugal13401657280231073195267823732122293529122500 Spain3035327792982158100762555065460543939854303

28 28 Table 2.b Yearly payment profile 2000-2009 (% of GDP) Country2000200120022003200420052006200720082009 Czech Rep. Cyprus0.00 Estonia0. Hungary0.000.050.06 0.290.450.830.840.400.05 Latvia0. Lithuania0. Malta0.00 Poland0. Slovakia0. Slovenia0. Germany0.050.140.16 Greece0.001.530.920.821.371.231.612.052.240.07 Ireland0.200.390.470.400.360. Italy0. 0.28 0.04 Portugal1. Spain0.050.781.091.050.960.840.560.520.500.04

29 29 Additionality and co-financing Additionality: Structural Funds are additional to domestically-financed expenditure and are not used as a substitute for it. Co-financing: EU provides only matching funds to individual projects - EU funds are matched to a certain extent by domestic expenditure. How to define proper benchmark for simulations? In practice this principle of additionality is hard to verify and rarely binding. M.S. not required to create new budgetary expenditure to co-finance cohesion policy support. Existing national resources can be 'earmarked' to co-finance SF transfers. Total spending increases only by the amount of SF transfers. As spending on infrastructure and education typically exceeds the co-financing requirements, this exercise takes domestically-financed expenditure in the counterfactual situation (without structural and cohesion funds) as the benchmark and only examines the impact of the fiscal transfer received from the EU cohesion funds.

30 30 Table 3. Fields of interventions Structural Funds (% of total spending 2000-2009) Category Agri.&Ind.&Serv. Hum. Res. RTDIInfrastructure Techn. Ass. CZ31.726.51.337.13.3 CY40. EE30.419.48.937.34.0 HU30.521.74.938.54.5 LV41.421. LT35. MT21.413.80.359.35.2 PL27.923.42.744.02.0 SK24.328.80.937.78.3 SI42.527. DE30.737. GR22.320.41.852.03.6 IE20.327.56.545.10.6 IT35.527.13.629.54.4 PT30.221.74.541.91.7 ES25.

31 31 Matching fields of interventions and model variables (p.1)

32 32 Table 4. Matching fields of interventions and model variables GBC: Cohesion Policy COH financed from EU budget Contributions to EU budget (EU15 ): proportional to countrys GDP financed by labour taxes

33 33 Figure 5.1. Simulated GDP impacts EU27 Member States

34 34 Example Table: Portugal

35 35 Effects Cohesion Spending Consumption increases: –Ricardian consumers anticipate higher permanent income and raise their consumption –Liquidity-constrained consumption also higher (employment and wage developments) Wages grow in long run in line with productivity Donor countries: higher contributions to the EU budget - increase in labour taxes - negative impact on employment growth. Recipient countries: higher growth boosts tax revenues. => For the largest net recipients this second effect outweighs the first and there is room to lower labour taxes, giving rise to positive employment effects. Corporate investment is generally crowded out in the short run. In the medium run productivity enhancing effects come to dominate and investment spending increases. Upward pressure on inflation as the demand effects dominate in the short run, But in the medium term, as potential output increases, inflationary pressures subside. Imports are boosted by the increase in demand Sizeable real appreciation in the largest recipient countries reduces exports growth. Trade balances deteriorate and current account deficits become larger.

36 36 Effects Cohesion Spending (2) Support Agriculture, Industry&Services plus Technical assistance: Reductions in fixed costs (lowering startup costs and increasing entry of new firms) Lower capital costs for tangible capital (increasing investment and capital accumulation). Government consumption (unproductive government spending), (only growth boosting effect in the short run) Infrastructure spending : i.e. Transport, telecommunication, energy, environmental, social infrastructure Government investment (productive) Government consumption (unproductive) (social infrastructure) Both lead to higher aggregate demand but are partly crowded out by lowering private consumption and private investment and some of the demand impulse leaks abroad through higher imports. However, in the medium term government investment raises productivity

37 37 Effects Cohesion Spending (3) Support to R&D reductions in fixed costs reductions in intangible capital costs for the intermediate sector By reducing costs, new start-ups enter the market (new products). By supporting innovation, high skilled workers are reallocated in the model from the production sector to the R&D sector. Initially, this reallocation reduces final goods production and has a negative impact on growth, but over time the positive output effects dominate as productivity increases, and this also stimulates physical investment (endogenous growth ) Investment Human Capital non-productive government spending direct transfers to households, Improvement skill efficiencies. The effects on average skill efficiencies take time to build up (cohort effects)

38 38 Figure 5.2. Simulated GDP impacts EU27 Member States DE IT IR PO

39 39 Figure 5.2. Simulated GDP impacts EU27 Member States EL ES CZ CY

40 40 Figure 5.2. Simulated GDP impacts EU27 Member States EE HU LT LV

41 41 Figure 5.2. Simulated GDP impacts EU27 Member States MT PL SI SK

42 42 Table 6: Cumulative output gains and multipliers of Cohesion Policy spending End of programming period 2009 Long term 2020 Cumulative GDP (% diff. from baseline) (1) Cumulative cohesion receipts (% of GDP) (2) Cumulative Multiplier (1)/(2) Cumulative GDP (% diff. from baseline) (1) Cumulative cohesion receipts (% of GDP) (2) Cumulative Multiplier (1)/(2) Germany0.611.370.443.641.372.65 Italy1.132.010.563.942.011.96 Ireland1.952.610.758.132.613.12 Portugal15.6915.471.0149.6815.473.21 Greece12.9911.851.1042.8711.853.62 Spain9.496.381.4929.816.384.67 Czech Republic1.391.990.705.961.992.99 Cyprus0.140.520.271.240.522.37 Estonia3.515.160.6812.005.162.33 Hungary3.083.031.0212.503.034.12 Lithuania7.675.851.3128.755.854.91 Latvia11.656.701.7441.106.706.13 Malta0.681.540.444.391.542.85 Poland4.983.961.2623.113.965.84 Slovenia0.841.260.663.391.262.69 Slovakia2.323.420.689.323.422.72

43 43 Sensitivity analysis (1) α G = output elasticity of public capital (infrastructure) Large literature on infrastructure investment and economic growth (since Aschauer 1989, 1990 ) Extremely wide range of estimates found in the literature Econometric problems relating to common trends, missing variables, simultaneity bias and reverse causation hamper a proper identification of this elasticity from macro- economic timeseries. Gramlich (1994) makes a case for an identical rate of return on private and public capital. Assumption adopted in the model : output elasticity of public capital is set such that the marginal product of public capital is identical to that of private capital (α G =0.10) => Sensitivity analysis 1 : α G = 0.15

44 44 Sensitivity analysis (2) slc = the share of liquidity constrained consumers The share of liquidity-constrained households is generally an important parameter as it determines the degree of so-called non- Ricardian behaviour in the model for non-productive government spending shocks. The lower the share of liquidity-constrained households, the higher the degree of crowding out of government spending shocks due to an offsetting response of Ricardian households who raise their precautionary savings in anticipation of higher future tax liabilities. The share of liquidity constrained households in the euro area is typically estimated to lie in the range between 0.2 and 0.4 (e.g. Ratto et al., 2009, Coenen et al., 2008). The assumption in the model version used here, that this share is equal to the share of low skilled workers. This implies substantial differences across countries. Labour force data on skill groups shows a large dispersion in the share of low skilled workers across countries and our model assumption implies a similar dispersion in the share of liquidity constrained households. sensitivity analysis 2 : slc=0.5

45 45 Sensitivity analysis (3) Classification of interventions Mapping of cohesion policy spending onto specific model variables lack of detailed information on some spending items alternative classifications are always possible. => Sensitivity analysis 3 We illustrate one variant in which we classify social infrastructure spending as productive

46 46 Figure 6.1 Portugal: sensitivity analysis: GDP impact (% diff.) and funds received (% of GDP)

47 47 Sensitivity analysis Output elasticity public capital (α G = 0.15) : As infrastructure spending amounts for a large share of overall spending (between 30-40 per cent) this has a significant impact on the results (eg. Portugal long term GDP effect from 3.1 per cent to 3.7 per cent ) Share liquidity-constrained consumption (slc = 0.5 ): The impact of this assumption is not particularly large. 1.Cohesion spending is financed by fiscal transfers from the EU budget. This spending does not give rise to proportionally higher tax liabilities in the future but is a pure fiscal transfer from donor counties to recipient countries. 2.Consumption by Ricardian households is also positively affected as most spending is productive and leads to a rise in permanent incomes

48 48 Conclusions Microfounded DSGE model with semi-endogenous growth In the short run: –spending could lead to crowding out of productive private investment and could give rise to real appreciations which lower export growth –R&D promoting policies could drive up wages of researchers and crowd out high skilled employment in other sectors. –little benefit one can expect in the short run from training and other human capital investments. In the medium term: –the productivity enhancing effects of infrastructure investment, R&D promoting policies, and human capital investments become gradually stronger Long run: –endogenous growth effects : positive benefits become stronger in the medium and long run


Download ppt "Modelling Cohesion Policy in a DSGE Model with Semi-Endogenous Growth: simulations with the QUEST III model Janos Varga and Jan in t Veld Research Directorate."

Similar presentations

Ads by Google