Motivations The K+S Model Model Validation Policy Experiments Concluding Remarks On the Complementarities between Innovation Policies and Demand Management in the Process of Development and Growth G. Dosi1 Email: A. Roventini1,2,3 andrea.roventini@univr.it 1 Sant’Anna School of Advanced Studies, Pisa (Italy) 2 University 3 OFCE, of Verona (Italy) Sophia-Antipolis (France) Motivations The K+S Model Model Validation Policy Experiments Concluding Remarks Recalling some Ingredients of Development Policies ... The “great transformation” from rural to industrial and advanced service economies involves a major process of accumulation of knowledge and capabilities at the levels of both individuals and organizations Successful episodes of industrialization / knowledge accumulation have always involved a rich array of public policies (from German and the U.S. to China and India) Some key domains of policy interventions: technological opportunities technological capabilities innovation appropriability industrial policies selection mechanisms “Keynesian” demand-management Motivations The K+S Model Model Validation Policy Experiments Concluding Remarks Recalling some Ingredients of Development Policies ... The “great transformation” from rural to industrial and advanced service economies involves a major process of accumulation of knowledge and capabilities at the levels of both individuals and organizations Successful episodes of industrialization / knowledge accumulation have always involved a rich array of public policies (from German and the U.S. to China and India) Some key domains of policy interventions: technological opportunities technological capabilities innovation appropriability industrial policies selection mechanisms “Keynesian” demand-management Motivations The K+S Model Model Validation Policy Experiments Concluding Remarks ... and the Significant Role of Inequality Inequality impacts strongly on aggregate demand dynamics and can contribute to the formation of bubbles in financial markets (Fitoussi and Saraceno, 2010; Stiglitz, 2011) The role of inequality in triggering crises: inequality reached its maximum just before the Great Depression and the current Great Recession The impact of policies may be conditioned by inequality levels Motivations The K+S Model Model Validation Policy Experiments Concluding Remarks Our Goals These intuitions and pieces of historical evidence have not been formalised We do it in the family of Keynes+Schumpeter agent-based models (Dosi et al. JEDC 2010, 2013) Study the long- and short-run impact of technology, industrial and demand-management policies Assess how different income distribution regimes affect growth, business-cycle fluctuations and the effects of policies Motivations The K+S Model Model Validation Policy Experiments Concluding Remarks Why Agent-Based Modeling (ABM)? ABMs model complex-system dynamics using boundedly-rational, heterogeneous, locally-interacting agents Capturing relevant features of innovation and finance “Strong” (Knightian) uncertainty Complex innovation and credit networks Risky, trial-and-error nature of innovation Policy design and implications Flexibility and modularity of ABMs allow to easily implement different scenarios Testing the short- and long-run effects of policies in a natural way Motivations The K+S Model Model Validation The Structure of the K+S Model Central Bank - Sets the baseline interest rate - Sets the max credit supply of banks - Stocks reserves of the banks Policy Experiments Concluding Remarks Motivations The K+S Model Model Validation Policy Experiments Concluding Remarks The Sequence of Microeconomic Decisions 1 The maximum credit supply is determined by the bank 2 Capital-good firms perform R&D. The productivity of new machines is determined. 3 Consumption-good firms decide how much to produce, choose their supplier and order machines 4 Firms hire workers (wages are anticipated and proportional to labor productivity). Firms pay machines using internal funds and credit provided by the banking sector 5 Production in both sectors begins 6 The consumption-good market opens, workers fully consume their wage, firms’ profits are determined 7 Entry and exit take place 8 Consumption-good firms receive the ordered machines Motivations The K+S Model Model Validation Policy Experiments Concluding Remarks Model Validation ABMs are much more complex than standard, e.g. DSGE, macroeconomic models The model is able to match many more macroeconomic stylized facts than standard models Empirical Discipline on Microfoundations: the model is also able to match the largest number of microeconomic (cross-sectional) stylized facts Motivations The K+S Model Model Validation Policy Experiments Concluding Remarks Macroeconomic Stylized Facts (1) Self-sustained growth with endogenous business cycles 25 20 Logs 15 10 (2) BP-filtered standard deviations 5 GDP Inv. Cons. 0 Investment more volatile than GDP; consumption less volatile than GDP 0 50 100 150 200 250 Time 300 350 400 450 Logs of GDP, C, and I 1 0.8 0.6 0.4 (4) Time-series GDP growth-rate distrbutions Quasi-Laplace fat-tailed distributions Percent 0.2 0 −0.2 −0.4 −0.6 GDP Inv. Cons. −0.8 −1 0 50 100 150 200 250 Time 300 350 400 450 BP-Filtered GDP, C, and I 4 10 3 10 Density (Logs) (3) BP-filtered cross-correlation with output Consumption, net investment and change in inventories procyclical and coincident Unemployment, prices and mark-ups are countercyclical. Productivity and inflation are pro-cyclical Total credit is procyclical and coincident Bankruptcy rates are pro-cyclical 2 10 1 10 0 10 −6 −4 −2 0 Growth Rate 2 4 6 GDP Growth-Rate Distribution Motivations The K+S Model Model Validation Policy Experiments Concluding Remarks Microeconomic Stylized Facts The model is able to account for a rich ensemble of micro (firm-level) cross-section stylized facts (Dosi, 2007) 0 10 Simulation Data −1 10 Tail Distribution (1) Productivity dispersion among firms is large (2) Inter-firm productivity differentials are persistent over time (3) Firm size distributions are right-skewed (and even more skewed than log-normal distributions) (4) Firm growth rates can be proxied by fat-tailed quasi-Laplace densities (5) Investment lumpiness (coexistence of firms investing a lot and investing almost-zero, see Gourio & Kayshap, J. Mon. Econ., 2007) (6) Bankruptcy rates can be proxied by power-law densities (see Fujiwara, 2004, Di Guilmi et al. 2003) −2 10 −3 10 0 10 Bad Debt Motivations The K+S Model Model Validation Policy Experiments Concluding Remarks Policy Combinations Schumpeterian innovation policies affecting opportunities (e.g. expected value of innovation draws) firm search capabilities (e.g. R&D productivity) appropriability conditions (e.g. patents; imitation) Entry and competition policies affecting market structure: competition policies (e.g. antitrust policy) entry and exit (e.g. barrier to entry and/or exit) Keynesian demand macro-management policies: fiscal policies Motivations The K+S Model Model Validation Policy Experiments Concluding Remarks Schumpeterian Innovation Policies Vary opportunities of technological innovation change the productivity of new technological draws GDP growth rises unemployment fall with increasing technological opportunities Vary firm search capabilities change the capital-good firm R&D productivity GDP growth rises, GDP volatility and unemployment fall as the R&D productivity increases Vary appropriability conditions: Patent system patent length w. or w/o patent breadth patents reduce GDP growth rate and raise unemployment, the effects are stronger in presence of patent breadth Motivations The K+S Model Model Validation Policy Experiments Concluding Remarks Entry and Competition Policies Vary entrants’ expected productivity different technological entry barriers GDP growth rises, GDP volatility and unemployment fall as the expected productivity of entrants increases Antitrust Policy (Capital-good Industry) capital-good firms with a market share higher than a fixed threshold cannot add new customers antitrust policy spurs GDP growth and it reduces both unemployment rate and output volatility Motivations The K+S Model Model Validation Policy Experiments Concluding Remarks Are Schumpeterian Technology Policies Enough? So far we have found that Schumpeterian policies has both long-run and short-run effects However, such results are conditional on a “Keynesian machine” well in place What happen if we switch that off? More generally, do Keynesian fiscal policies have also long-run effects? Motivations The K+S Model Model Validation Policy Experiments Concluding Remarks Keynesian Policies Eliminate the public sector we eschew the public sector from our model and we then “pump up” Schumpeterian policies threshold effect: Keynesian policies are necessary to support sustained long-run economic growth Schumpeterian policies are not enough to push the economy away from low growth trajectories Changing taxes and unemployment benefits tuning up fiscal policies does delock the economy from the low growth trajectory and brings it to the high growth one avg. GDP growth almost the same, but Keynesian policies have strong countercyclcial effects More generally, strong complementarity between “Keynesian” policies affecting demand and “Schumpeterian” policies affecting innovation Motivations The K+S Model Model Validation Policy Experiments Concluding Remarks Inequality and Macroeconomic Dynamics Distributions How does inequality affect the the aggregate behavior of the economy? We tune the mark-up rate to study different income distribution regimes We then consider the impact of fiscal policies under different income distribution scenarios Notice that the mark-up rate determines: the distribution of income between profits and wages (distributive effect) the flow of internal funds of firms (financial dependence effect) Motivations The K+S Model Model Validation Policy Experiments Concluding Remarks Changing Income Distribution Frequency of full employment At low mark-up rates the economy spends much more time in full employment... Motivations The K+S Model Model Validation Policy Experiments Changing Income Distribution GDP growth volatility ... business fluctuations are smoother ... Concluding Remarks Motivations The K+S Model Model Validation Policy Experiments Concluding Remarks Changing Income Distribution Unemployment ...unemployment rates are lower; average growth is unaffected. Motivations The K+S Model Model Validation Policy Experiments Fiscal Policy and Income Distribution until now there is a positive level of fiscal policy intervention we repeat the same experiment on mark-up rates by setting to zero fiscal policies Results without fiscal policy, the effects of income distribution on real variables are strengthened long-run growth effects: high levels of the mark-up rate lock the economy into a low-growth trajectory Concluding Remarks Motivations The K+S Model Model Validation Policy Experiments Concluding Remarks Fiscal Policy and Income Distribution Volatility furthermore, fiscal policy is more effective in reducing volatility and unemployment when mark-up rate is high. Motivations The K+S Model Model Validation Policy Experiments Concluding Remarks Globalization, Inequality, Policies and Growth: Some Reflections by Way of Conclusions How do our policy experiment results relates to the challenges posed by globalization and increasing inequality? Let me conclude this talk showing further empirical evidence... ... And possibly providing some policy responses Motivations The K+S Model Model Validation Policy Experiments Concluding Remarks Globalization Financial globalization is weakening the policy setting powers of States Globalization is weakening the manufacturing base in the US and Europe What is so special about manufacturing? innovation opportunities dynamic increasing returns most R&D is in manufacturing large base of decent jobs crucial contribution to foreign accounts Motivations The K+S Model Model Validation Policy Experiments Concluding Remarks Gross International Financial Assets of Advanced Countries (millions of current US dollars) 160000000 140000000 120000000 DEBT FDI Equity DerivaFves Total GDP 100000000 80000000 60000000 40000000 20000000 0 1970 1975 1980 1985 1990 1995 Source: updated and extended version of the database developed by Lane and Milesi-­‐FerreQ (2007). 2000 2005 Motivations The K+S Model Model Validation Policy Experiments Concluding Remarks Manufacturing Value Added (% of World Value Added in Manufacturing, Current Prices) (% of world v.a. in manufacturing, current prices) EU 27 Japan France 35 30 China Germany India United States Italy United Kingdom 25 20 15 10 5 0 1980 1985 1990 1995 2000 2005 2010 Motivations The K+S Model Model Validation Policy Experiments Concluding Remarks Growing Inequality Possible causes: globalization skills (tasks) destroying technical change institutions (stagnant or declining minimum wages, weaker unions, tax-cuts to the super-rich) More consequences: families forced to borrow beyond their means lower aggregate demand (via differential consumption propensities) wage differential hinders work motivation and productivity At the same time public expenditures necessary to finance a universal welfare model are increasing Motivations The K+S Model Model Validation Policy Experiments Concluding Remarks Inequality: Mostly about Top Incomes (% of top-­‐incomes) 1975 1985 1995 2005 Top 10% Top 5% top 1% Top 10% Top 5% top 1% Top 10% Top 5% top 1% Top 10% Top 5% top 1% France 33,41 22,06 8,48 31,05 19,96 7,2 32,41 20,93 7,7 32,89 21,88 8,73 Germany 30,8 21,6 10,1 31,37 21,24 9,64 31,4 20,84 8,84 . . . Italy 31,2 20,04 7,24 26,83 17,5 6,81 30,57 20,58 8,13 33,19 22,78 9,35 Japan 30,52 19,58 7,08 31,92 20,25 7,03 34,02 21,47 7,3 40,56 25,96 9,42 United Kingdom 27,82 17,4 6,1 32,65 20,75 7,4 38,51 25,8 10,75 41,62 29,57 14,25 United States 32,62 21,03 8,01 34,25 22,38 9,09 40,54 28,46 13,53 44,94 33,12 17,68 Source: the World top-­‐incomes database Motivations The K+S Model Model Validation Policy Experiments Concluding Remarks Public Expenditures on Health Programs (% on GDP) 10 9 France Germany Italy Japan United Kingdom United States 8 7 6 5 4 3 1980 Source: OECD 1985 1990 1995 2000 2007 2009 Motivations The K+S Model Model Validation US Marginal Tax Rates Policy Experiments Concluding Remarks Motivations The K+S Model Model Validation Policy Experiments Concluding Remarks Scenario and Policy Options Business as usual further weakening of the manufacturing base low rate of growth increasing inequalities Managing globalization: industrial policy mission oriented programs (equivalent to Apollo/military space programs) pragmatic use of competition policy strengthening European ventures such as EADS/Airbus fiscal policies heavy taxation on financial rents (e.g. Tobin tax) heavy progressive taxation in general stop a race-to-the-bottom in European fiscal polices Motivations The K+S Model Model Validation Policy Experiments Concluding Remarks Scenario and Policy Options Business as usual further weakening of the manufacturing base low rate of growth increasing inequalities Managing globalization: industrial policy mission oriented programs (equivalent to Apollo/military space programs) pragmatic use of competition policy strengthening European ventures such as EADS/Airbus fiscal policies heavy taxation on financial rents (e.g. Tobin tax) heavy progressive taxation in general stop a race-to-the-bottom in European fiscal polices Motivations The K+S Model Model Validation The Wal-Mart Archetype Policy Experiments Concluding Remarks Motivations The K+S Model Model Validation Policy Experiments Concluding Remarks The Wal-Mart Archetype 1.3 mln employees (largest retailer in the US) over 15% of U.S. imports of consumer goods from China political involvement to reduce trade barriers each Wal-Mart worker replaces 1.4 retail workers (-2.7% reduction in retail employment) Wal-Mart store openings lead to declines in county-level retail earnings of about 1.5% NB: all this, net of the effects on US manufacturing employment