Banque Nationale De Belgique Sa · EURONEXT:BNB
Banque Nationale De Belgique Sa — Investor Presentation (March 2018)
64 slides from Banque Nationale De Belgique Sa's investor presentation, Mar 14, 2018
Slide 1: Productivity and Misallocation in General Equilibrium
Productivity and Misallocation in General Equilibrium David Baqaee Emmanuel Farhi LSE Harvard March 14, 2018
Slide 2: Aggregation Theorems for Efficient Economies
Aggregation Theorems for Efficient Economies For efficient economy, Solow (1957): d log Y = d log TFP + ΛL d log L + ΛK d log K . For efficient economy, Hulten (1978): salesk d log TFP = ∑ λk d log Ak , where λk = . GDP Ex-ante (structural counterfactuals) and ex-post (growth accounting) content.
Slide 3: What We Do
What We Do Extend these results to inefficient economies and other shocks. General reduced-form, non-parametric formula. Mapping from micro to macro using a general structural model. micro wedges. structural micro elasticites of substitution. returns to scale. factor market reallocation. network lin
Slide 4: Research Agenda
Research Agenda “The macroeconomic Impact of Microeconomic Shocks”, Baqaee and Farhi (2017): nonlinearities in efficient economies. “Propagation and Aggregation in Production Networks” , Baqaee and Farhi (2018): propagation of microeconomic shocks and “industry” aggregation. “The Microeconomic Found
Slide 5: Research Agenda
Research Agenda Work in progress: structural models of frictions (IO, financing constraints, search and matching, nominal rigidities, etc.), fixed costs, entry and exit, dynamics, non-homotheticities, endogenous innovation, other models of network formation, etc.
Slide 6: Related Literature
Related Literature Efficient Network Production Economies: Long and Plosser (1983), Gabaix (2011), Acemoglu et al. (2012), Foerster et al. (2011), Acemoglu et al. (2016), Baqaee and Farhi (2017). Inefficient Network Production Economies: Basu and Fernald (2001), Fernald and Neiman (2011), Jones (201
Slide 7: Related Literature
Related Literature Falling Labor Share, Increasing Markups, Productivity Slowdown: Davis et al. (2007), Gordon (2012), Neiman and Karabarbounis (2014), Elsby et al. (2013), Piketty and Zucman (2014), Baqaee (2015), Barkai (2016), Rognlie (2016), Koh et al. (2016), Gutiérrez and Philippon (2016), De
Slide 8: Agenda
Agenda General Non-parametric Result General Parametric Result Applications Growth Accounting Quantitative Model Extensions (see paper) Conclusion
Slide 9: Agenda
Agenda General Non-parametric Result General Parametric Result Applications Growth Accounting Quantitative Model Extensions (see paper) Conclusion
Slide 10: General Framework
General Framework Final demand as maximizer of homothetic aggregator: Y = D (c1 , . . . , cN ) , with ck final consumption of good k . Budget constraint: ∑(1 + τkc )pk ck = ∑ wf Ff + ∑ πk + τ, k f k with pk prices, πk profits, τkc consumption wedges, wf wages, Ff factors, τ lump-sum rebate.
Slide 11: General Framework
General Framework Good k produced with constant-returns cost function: yk Ck (1 + τk1 )p1 , . . . , (1 + τkN )pN , (1 + τkf 1 )w1 , . . . , (1 + τkF f )wF , Ak with yk total output, Ak Hicks-neutral productivity shock, τkl input-specific wedge, τkif factor-specific wedge. Markup µk over marginal cos
Slide 12: Generality
Generality Captures factor augmenting productivity shocks with relabeling. Captures demand shocks as mix of productivity shocks. Captures decreasing returns with fixed quasi-factors. Can capture “technical” adjustment costs and capacity utilization. See later for increasing returns. Can be applied t
Slide 13: Notation and Accounting Convention
Notation and Accounting Convention Represent all wedges as markups with relabeling. Assume that in data, expenditures by i on j and revenues of i recorded gross of wedges and markups. If not, for ex. with implicit wedges (e.g. credit constraints), re-write expenditures gross of these wedges.
Slide 14: Revenue-Based vs. Cost-Based
Revenue-Based vs. Cost-Based Definition Ω and Ω̃ are N × N input-output matrices with ijth element: pj xij pj xij Ωij = , Ω̃ij = . pi yi ∑k pk xik + ∑f wf Fif Ψ and Ψ̃ are N × N Leontief inverse matrices: Ψ = (I − Ω)−1 , Ψ̃ = (I − Ω̃)−1 . b is N × 1 consumption-shares vector with ith element: pi ci
Slide 15: Revenue-Based vs. Cost-Based
Revenue-Based vs. Cost-Based Cost-based definitions capture correct notion of exposure: Ω̃ij is direct exposure of i to j. Ψ̃ij is direct and indirect exposure of i to j. ˜ k is direct and indirect exposure of household to k . λ
Slide 16: Macro Impact of Micro Shocks
Macro Impact of Micro Shocks Y (A, X) : output Y given productivities A and shares Xij = xij /yj . Change in equilibrium in response to shocks: ∂ log Y ∂ log Y d log Y = d log A + dX . | ∂ log A {z } | ∂ X{z } ∆Technology ∆Allocative Efficiency For efficient economies, macro-envelope implies Hulten:
Slide 17: Macro Impact of Micro Productivity Shocks
Macro Impact of Micro Productivity Shocks Theorem d log Y ˜k d log Λf = λ − ∑ Λ̃f . d log Ak f d log Ak |{z} | {z } ∆Technology ∆Allocative Efficiency Yields Hulten’s theorem for efficient economies: ˜ k = λk d log Λf λ and − ∑ Λ̃f = 0. f d log Ak See later for structural formula for − ∑f Λ̃f d log
Slide 18: Macro Impact of Micro Markup Shocks
Macro Impact of Micro Markup Shocks Theorem d log Y = −λ˜ k − Λ̃f d log Λf . d log µk ∑ d log µk f | {z } ∆Allocative Efficiency Also applies to shocks to other wedges. Can be applied to endogenous wedges via chain rule. See later for structural formula for − ∑f Λ̃f d log Λf / d log µk .
Slide 19: Ex. Simple Vertical Economy
Ex. Simple Vertical Economy Example of multiple marginalization taken from Baqaee (2016): HH 1 ··· N L ˜ k = 1 6= λk = ∏k −1 µ −1 and ΛL = ∏N µ −1 6= 1. λ i =1 i i =1 i Productivity shocks: d log Y ˜ k − d log ΛL = 1 =λ d log Ak d log Ak Markups/wedges shocks: d log Y ˜ k − d log ΛL = 0 = −λ d log µ
Slide 20: Ex. Simple Horizontal Economy
Ex. Simple Horizontal Economy L 1 ··· N HH ˜ k = λk and ΛL = ∑ λj µ −1 6= 1. λ j j Productivity shocks: ! d log Y ˜ k − d log ΛL = λk − (θ0 − 1) µk−1 =λ − 1 λk . d log Ak d log Ak ∑j λj µj−1 Markup/wedge shocks: ! d log Y ˜ k − d log ΛL = θ0 µk−1 = −λ − 1 λk . d log µk d log Ak ∑j λj µj−1
Slide 21: Ex. Cobb-Douglas Economies
Ex. Cobb-Douglas Economies Productivity shocks: d log Y ˜ k − Λ̃f d log Λf = λ d log Ak =λ ∑ d log Ak ˜k . f Markup/wedge shocks: d log Y ˜ k − Λ̃f d log Λf = −λ = −λ ˜ k + λk Λ̃f Ψkf /Λf . d log µk ∑ d log Ak ∑ f f Cobb-Douglas functional forms very popular in literature. For an efficient economy,
Slide 22: Ex. US Economy
Ex. US Economy Productivity shocks: d log Y ˜ k − Λ̃L d log ΛL − Λ̃K d log ΛK . =λ d log Ak d log Ak d log Ak Markup/wedge shocks: d log Y ˜ k − Λ̃L d log ΛL − Λ̃K d log ΛK . = −λ d log µk d log µk d log µk
Slide 23: Sources of Growth and Solow Residual
Sources of Growth and Solow Residual Easy extension to changing factor supplies: ˜ 0 d log A −λ d log Y − Λ̃0 d log L = λ ˜ 0 d log µ + Λ̃0 d log Λ . | {z } | {z } ∆Technology ∆Allocative Efficiency Solow residual: d log Y − Λ̂0 d log L = λ ˜ 0 d log µ + Λ̃0 d log Λ + ˜ 0 d log A −λ (Λ̃ − Λ̂)0 d log
Slide 24: Alternative Decompositions
Alternative Decompositions Alternative decompositions of Basu-Fernald (2002) and Petrin-Levinsohn (2012). Do not use input-output information. Revealing example of acyclic economies: F1 ··· FK ··· ··· 1 ··· N HH These decompositions detect changes in allocative efficiency, even though allocation is
Slide 25: Measuring Allocative Efficiency
Measuring Allocative Efficiency Measure of change in allocative efficiency along equilibrium path. Different from change of distance to frontier a la Restuccia and Rogerson (2008) or Hsieh and Klenow (2009). Relation between the two concepts: Z 1 Y (A, 1) d log Y (A, µ̂(t )) d log µ̂(t ) log =− dt Y
Slide 26: Agenda
General Non-parametric Result General Parametric Result Applications Growth Accounting Quantitative Model Extensions (see paper) Conclusion
Slide 27: Parametric Model
Parametric Model Final demand: θ θ−0 1 θ0θ−1 0 Y ck 0 = ∑ bk . Y k ck Production of good k : θkθ−1 θkθ−1 θkθ−1 k yk lk k Xk k = Ak ak + (1 − ak ) . yk lk Xk Xk composite intermediate input given by ! εkε−1 εk −1 k Xk εk = ∑ ωkl xlk , Xk l where xkl intermediate inputs from industry l u
Slide 28: Parametric Model
Parametric Model Relabel network so that each node corresponds to one CES nest. Structure can actually represent any nested CES economy with arbitrary pattern of nests and wedges. Definition ! ! CovΩ̃(j ) Ψ̃(k ) , Ψ(L) = ∑ Ω̃ji Ψ̃ik ΨiL − ∑ Ω̃ji Ψ̃ik ∑ Ω̃ji ΨiL . i i i
Slide 29: Macro Impact of Micro Productivity Shocks: One Factor
Macro Impact of Micro Productivity Shocks: One Factor Proposition Suppose there is only one factor (with index L). Then d log Y ˜k d log ΛL = λ − , d log Ak d log A |{z} | {z k} ∆Technology ∆Allocative Efficiency where Ψ(L) d log ΛL −1 = (θj − 1)µj λj CovΩ̃(j ) Ψ̃(k ) , . d log Ak ∑ j ΛL Centrality
Slide 30: Explaining Covariance Operator
Explaining Covariance Operator ˜ k − (θj − 1)µ −1 λj Cov (j ) Ψ̃ , Ψ(L) . d log Y =λ ∑ j Ω̃ (k ) d log Ak j ΛL | {z } Ψ̃1k , Ψ1L /ΛL Ψ̃2k , Ψ2L /ΛL ··· Ψ̃N −1,k , ΨN −1,L /ΛL Ψ̃Nk , ΨNL /ΛL Ωj2 Ωj ,N −1 Ωj1 ΩjN j High Ψ̃ik : i’s highly exposed to k. High ΨiL /ΛL : most of i’s revenues are ultimately
Slide 31: Ex. Back to Simple Horizontal Economy
Ex. Back to Simple Horizontal Economy L 1 ··· N HH Change in technology and change in allocative efficiency: ! d log Y µk−1 = λk − (θ0 − 1) − 1 λk . d log Ak ∑j λj µj−1 Key: markup vs. average and elasticity minus one.
Slide 32: Macro Impact of Micro Markup Shocks: One Factor
Macro Impact of Micro Markup Shocks: One Factor Proposition Suppose there is only one factor indexed by L. Then d log Y = −λ˜ k − d log ΛL , d log µk d log Ak | {z } ∆Allocative Efficiency where Ψ(L) d log ΛL −1 ΨkL = ∑(1 − θj )µj λj CovΩ̃j Ψ̃(k ) , − λk . d log µk j ΛL ΛL Positive markup shock like
Slide 33: Ex. Simple Horizontal Economy
Ex. Simple Horizontal Economy L 1 ··· N HH Change in allocative efficiency: µk−1 λk µk−1 d log Y ˜ k − (1 − θ0 )λk = −λ −1 + , d log µk ΛL ΛL ! µk−1 = θ0 − 1 λk . ∑j λj µj−1 Key: markup vs. average and elasticity.
Slide 34: Macro Impact of Micro Productivity Shocks: Multiple Factors
Macro Impact of Micro Productivity Shocks: Multiple Factors Proposition The following linear system describes the elasticities of factor shares: d log Λ d log Λ =Γ + δ(k ) , d log Ak d log Ak with Ψ(g ) Γf ,g = − ∑(θj − 1)λj µj−1 CovΩ̃(j ) Ψ̃(f ) , , j Λg and Ψ(f ) −1 δfk = ∑(θj − 1)λj µj CovΩ̃(j )
Slide 35: Multiple Factors
Multiple Factors Extends Hulten (1978), Jones (2011), Oberfield and Raval (2014), Caliendo et al. (2017), , and Baqaee and Farhi (2017) in one place. Can derive similar formula for markups/wedge shocks.
Slide 36: Ex. Multiple Factors
Ex. Multiple Factors K L 2 4 1 HH 3
Slide 37: Ex. Multiple Factors
Ex. Multiple Factors Change in technology and change in allocative efficiency: ! d log Y µk−1 = λk + λk (θ0 − 1) 1 − −1 −1 (k = 1, 2), d log Ak λ1 λ2 λ1 +λ2 µ1 + λ1 +λ2 µ2 d log Y = λ3 , d log A3 d log Y ˜ 4 = µ3 λ4 . =λ d log A4 No change in allocative efficiency between (1+2) and (3+4).
Slide 38: Agenda
Agenda General Non-parametric Result General Parametric Result Applications Growth Accounting Quantitative Model Extensions (see paper) Conclusion
Slide 39: Sources of Growth and Solow Residual
Sources of Growth and Solow Residual Suppose markups are only distortions. Use annual IO tables from BEA from 1997-2015. Use markups from Gutiérrez and Philippon (2016), De Loecker and Eeckhout (2017), and Lerner Indices for firms in Compustat. All measures show large increases in markups, from com
Slide 40: Sources of Growth and Solow Residual
Sources of Growth and Solow Residual 0.14 Solow Residual Allocative Efficiency 0.12 Factor Undercounting Technology 0.10 0.08 0.06 0.04 0.02 0.00 −0.02 1996 1998 2000 2002 2004 2006 2008 2010 2012 2014 Using the Gutiérrez and Philippon (2016) markup data. Similar with De Loecker and Eeckhout (2017)
Slide 41: Sources of Growth and Solow Residual
Sources of Growth and Solow Residual 0.2 Solow Residual Allocative Efficiency Factor Undercounting 0.15 Technology 0.1 0.05 0 -0.05 -0.1 1996 1998 2000 2002 2004 2006 2008 2010 2012 2014 Using the De Loecker and Eeckhout (2017) markup data.
Slide 42: Sources of Growth and Solow Residual
Sources of Growth and Solow Residual 0.15 Solow Residual Allocative Efficiency Factor Undercounting Technology 0.10 0.05 0.00 −0.05 −0.10 1996 1998 2000 2002 2004 2006 2008 2010 2012 2014 Using Lerner indices.
Slide 43: Agenda
General Non-parametric Result General Parametric Result Applications Growth Accounting Quantitative Model Extensions (see paper) Conclusion
Slide 44: Quantitative Results
Quantitative Results Calibrate parametric model. Benchmark elasticities of substitution: consumption 0.4; value and intermediates 0.3; across intermediates be 0.01; between labor and capital 1; within industries 8. Use IO table from BEA from 2015. Robustness checks: role of elasticities and input-ou
Slide 45: Gains from Eliminating Markups
Gains from Eliminating Markups Gutierrez-Philippon Lerner Index De Loecker-Eeckhout 2014 20% 17% 35% 1997 3% 5% 21% Measures show big increase between 1997 and 2014. Contrast with 0.1% estimate of Harberger (1954) triangles! “It takes a heap of Harberger triangles to fill an Okun gap.” — Tobin
Slide 46: Gains from Shrinking Markups: Robustness
Gains from Shrinking Markups: Robustness Benchmark CD+CES CD+CD VA Benchmark VA CD + CES VA CD + CD GP 20% 21% 4% 8% 8% 1% LI 17% 18 % 4% 7% 7% 1% DE 35% 38% 7% 18 % 18% 3% Elasticities matter. Input-output structure matters. Value-added production functions misleading!
Slide 47: Macro-Volatility from Micro Shocks
Macro-Volatility from Micro Shocks Var (log Y ) ≈ kDlog A log Y k2 Var (d log A) + kDlog µ log Y k2 Var (d log µ). Thought experiment: i.i.d. shocks to Compustat firms, not others. Tabluate diversification factors std (log Y )/std (d log A) and std (log Y )/std (d log µ). Benchmark Competitive Cobb-
Slide 48: Macro Impact of Micro Shocks
Macro Impact of Micro Shocks (a) (b) 200 number number 100 100 0 0 −5.00 0.00 5.00 −5.00 0.00 5.00 (c) (d) 20 number number 10 * * 0 0 0.80 1.00 1.20 −0.10 −0.05 0.00 0.05 Output elasticity productivity and markup shocks relative to size. For firm shocks and for sectoral shocks. Distortions matter!
Slide 49: Agenda
General Non-parametric Result General Parametric Result Applications Growth Accounting Quantitative Model Extensions (see paper) Conclusion
Slide 50: Extensions (see paper)
Extensions (see paper) Endogenous markups/wedges. Elastic Factors. Fixed costs and entry. Nonlinearities.
Slide 51: Agenda
General Non-parametric Result General Parametric Result Applications Growth Accounting Quantitative Model Extensions (see paper) Conclusion
Slide 52: Conclusion
Conclusion Reduced-form aggregation theorem for economies with frictions. Structural aggregation theorems. Wide range of applications in different contexts. Work in progress: structural models of frictions (IO, financing constraints, search and matching, nominal rigidities, etc.), fixed costs, entry
Slide 53: Ex. Cost-Based vs. Revenue-Based Domar Weights
Ex. Cost-Based vs. Revenue-Based Domar Weights Example of multiple marginalization taken from Baqaee (2016): HH 1 ··· N L Cost-based vs. revenue-based Domar weights: k −1 ˜k = 1 b0 Ψ̃ = λ b0 Ψ = λk = and ∏ µi−1 < 1. i =1
Slide 54: Ex. Back to Simple Vertical Economy
Ex. Back to Simple Vertical Economy Example of multiple marginalization taken from Baqaee (2016): HH 1 ··· N L Change in technology, no change in allocative efficiency: k −1 d log Y ˜ k = 1 > λk = d log Ak =λ ∏ µi−1 . i =1 In accounting sense, Hulten’s theorem fails. In economic sense, Hulten’s theo
Slide 55: Ex. Back to Simple Horizontal Economy
Ex. Back to Simple Horizontal Economy L 1 ··· N HH Shares and factor shares: ˜ k = λk , λ Λ̃L = 1 > ΛL = ∑ λj µj−1 . j Change in technology and change in allocative efficiency: d log Y ˜ k + d H (Λ̃, Λ) = λk − d log ΛL . =λ d log Ak d log Ak d log Ak
Slide 56: References
Acemoglu, D., V. M. Carvalho, A. Ozdaglar, and A. Tahbaz-Salehi (2012). The network origins of aggregate fluctuations. Econometrica 80(5), 1977–2016. Acemoglu, D., A. Ozdaglar, and A. Tahbaz-Salehi (2016). Networks, shocks, and systemic risk. In The Oxford Handbook of the Economics of Networks. Alti
Slide 57: References
Baqaee, D. R. and E. Farhi (2017). The macroeconomic impact of microeconomic shocks: Beyond Hulten’s Theorem. Barkai, S. (2016). Declining labor and capital shares. Bartelme, D. and Y. Gorodnichenko (2015). Linkages and economic development. Technical report, National Bureau of Economic Research. Ba
Slide 58: References
transmission of monetary policy in a multisector economy. International Economic Review 50(4), 1243–1266. Buera, F. J., J. P. Kaboski, and Y. Shin (2011, August). Finance and Development: A Tale of Two Sectors. American Economic Review 101(5), 1964–2002. Buera, F. J. and B. Moll (2012, January). Agg
Slide 59: References
De Loecker, J. and J. Eeckhout (2017). The rise of market power and the macroeconomic implications. Technical report, National Bureau of Economic Research. D’Erasmo, P. N. and H. J. Moscoso Boedo (2012). Financial structure, informality and development. Journal of Monetary Economics 59(3), 286–302.
Slide 60: References
Gabaix, X. (2011). The granular origins of aggregate fluctuations. Econometrica 79(3), 733–772. Giné, X. and R. M. Townsend (2004). Evaluation of financial liberalization: a general equilibrium model with constrained occupation choice. Journal of development Economics 74(2), 269–307. Gopinath, G.,
Slide 61: References
Gutiérrez, G. and T. Philippon (2016). Investment-less growth: An empirical investigation. Technical report, National Bureau of Economic Research. Hopenhayn, H. and R. Rogerson (1993, October). Job Turnover and Policy Evaluation: A General Equilibrium Analysis. Journal of Political Economy 101(5),
Slide 62: References
economic development. American Economic Journal: Macroeconomics, 1–28. Jones, C. I. (2013). Input-Output economics. In Advances in Economics and Econometrics: Tenth World Congress, Volume 2, pp. 419. Cambridge University Press. Koh, D., R. Santaeulàlia-Llopis, and Y. Zheng (2016). Labor share decli
Slide 63: References
multisector menu cost model. The Quarterly journal of economics 125(3), 961–1013. Neiman, B. and L. Karabarbounis (2014). The global decline of the labor share. The Quarterly Journal of Economics 129(1), 61–103. Oberfield, E. (2013, January). Productivity and Misallocation During a Crisis: Evidence
Slide 64: References
Reis, R. (2013). The Portugese Slump and Crash and the Euro Crisis. Brookings Papers on Economic Activity 44(1 (Spring), 143–210. Restuccia, D. and R. Rogerson (2008). Policy distortions and aggregate productivity with heterogeneous establishments. Review of Economic dynamics 11(4), 707–720. Rognlie
Banque Nationale De Belgique — Investor Presentation (March 2018) on 6ix