Creative Destruction and Subjective Well-Being Empirics Data (2) To proxy for subjective well-being in the Gallup Healthways data, we use the Cantril ladder of life : Imagine a ladder with steps numbered from zero at the bottom to 10 at the top; the top of the ladder represents the best possible life for you and the bottom of the ladder represents the worst possible life for you. On which step of the ladder would you say you personally feel you stand at this time? And which level of the ladder do you anticipate to achieve in five years? We refer to answers to the first question as the current ladder and to the second question as the anticipated ladder Aghion (Harvard), Akcigit (UPenn), Deaton (Princeton), and Roulet (Harvard) 22
Creative Destruction and Subjective Well-Being Empirics Data (3) To proxy for subjective well-being in the BRFSS, we use the Life satisfaction question : ”In general how satisfied are you with your life?” The possible answers are ”Very satisfied” (1), ”Satisfied” (2) ”Dissatisfied” (3), ”Very dissatisfied” (4) We recoded them so that an increase in the variable means an increase in subjective well-being Aghion (Harvard), Akcigit (UPenn), Deaton (Princeton), and Roulet (Harvard) 23
Creative Destruction and Subjective Well-Being Empirics Empirical framework MSA-level regressions of creative destruction on subjective 1 well-being − → Across years averages to mirror the steady-state analysis of the model Individual level regressions 2 → Rich set of controls for individual determinants of well-being − Robustness checks 3 Heterogeneity: analysis of how the effect varies with the MSA’s 4 sectoral composition, the state-level unemployment insurance generosity and individuals’ age Aghion (Harvard), Akcigit (UPenn), Deaton (Princeton), and Roulet (Harvard) 24
Creative Destruction and Subjective Well-Being Empirics Metropolitan Statistical Area (MSA) Results 1/4 MSA-level averages; Gallup data (2008-2011) - (1) (2) (3) (4) (5) VARIABLES Current ladder Unemployment rate -2.303*** -2.970*** -1.972*** (0.731) (0.730) (0.757) Job turnover rate 0.652* 1.377*** (0.365) (0.379) Job creation rate 5.889*** 4.715*** (0.945) (0.925) Job destruction rate -3.851*** -1.947** (0.843) (0.983) Observations 363 363 363 363 363 R-squared 0.097 0.015 0.158 0.163 0.211 Aghion (Harvard), Akcigit (UPenn), Deaton (Princeton), and Roulet (Harvard) 25
Creative Destruction and Subjective Well-Being Empirics Metropolitan Statistical Area (MSA) Results 2/4 MSA-level averages; BRFSS data (2005-2010) - (1) (2) (3) (4) (5) VARIABLES ”How satisfied are you with your life?” Unemployment rate -1.504*** -1.702*** -1.524*** (0.270) (0.247) (0.270) Job turnover rate 0.121 0.348*** (0.0868) (0.0784) Job creation rate 1.652*** 0.859*** (0.241) (0.259) Job destruction rate -1.622*** -0.285 (0.259) (0.320) Observations 364 364 364 364 364 R-squared 0.257 0.007 0.311 0.133 0.323 Aghion (Harvard), Akcigit (UPenn), Deaton (Princeton), and Roulet (Harvard) 26
Creative Destruction and Subjective Well-Being Empirics Metropolitan Statistical Area (MSA) Results 3/4 MSA-level averages; Gallup data (2008-2011) - (1) (2) (3) (4) (5) VARIABLES Anticipated ladder Unemployment rate -0.560 -1.486*** -1.007** (0.465) (0.461) (0.498) Job turnover rate 1.549*** 1.911*** (0.319) (0.341) Job creation rate 4.109*** 3.509*** (0.851) (0.855) Job destruction rate -0.653 0.320 (0.760) (0.885) Observations 363 363 363 363 363 R-squared 0.006 0.087 0.122 0.122 0.134 Aghion (Harvard), Akcigit (UPenn), Deaton (Princeton), and Roulet (Harvard) 27
Creative Destruction and Subjective Well-Being Empirics Metropolitan Statistical Area (MSA) Results 4/4 MSA-level averages; Gallup data (2008-2011) - (1) (2) (3) (4) (5) VARIABLES Worry Unemployment rate 0.462*** 0.382*** 0.258*** (0.0812) (0.0854) (0.0952) Job turnover rate 0.257*** 0.163*** (0.0522) (0.0565) Job creation rate -0.405*** -0.251 (0.152) (0.158) Job destruction rate 0.826*** 0.576*** (0.116) (0.155) Observations 363 363 363 363 363 R-squared 0.119 0.073 0.145 0.144 0.170 Aghion (Harvard), Akcigit (UPenn), Deaton (Princeton), and Roulet (Harvard) 28
Creative Destruction and Subjective Well-Being Empirics Magnitude - MSA-level regressions A one standard deviation increase in job turnover has an effect on the current ladder of life equivalent to a 0.7 standard deviation decrease in the unemployment rate on the anticipated ladder of life equivalent to a 1.8 standard deviation decrease in the unemployment rate Aghion (Harvard), Akcigit (UPenn), Deaton (Princeton), and Roulet (Harvard) 29
Creative Destruction and Subjective Well-Being Empirics Individual Level Regressions The specification at the individual level is: SWB i , m , t = δ CD m , t + α U m , t + β X i , m , t + T t + ǫ i , s , t , Individual controls include : gender, ethnicity, detailed education and family status, age, age2 Year and Month Fixed effect Standard errors clustered at the MSA level We restrict attention to working-age individuals (18-60 years old) Aghion (Harvard), Akcigit (UPenn), Deaton (Princeton), and Roulet (Harvard) 30
Creative Destruction and Subjective Well-Being Empirics Individual level results 1/4 - Life satisfaction (BRFSS) (1) (2) (3) (4) (5) VARIABLES ”How satisfied are you with your life?” Unemployment rate -0.871*** -0.954*** -0.910*** (0.144) (0.143) (0.150) Job turnover rate 0.141** 0.212*** (0.0609) (0.0543) Job creation rate 0.384*** 0.294*** (0.0942) (0.0858) Job destruction rate -0.124 0.115 (0.0885) (0.0784) Year and Month F.E. x x x x x Observations 856,906 856,906 856,906 856,906 856,906 R-squared 0.064 0.063 0.064 0.063 0.064 Aghion (Harvard), Akcigit (UPenn), Deaton (Princeton), and Roulet (Harvard) 31
Creative Destruction and Subjective Well-Being Empirics Individual level results 2/4 - Current ladder (Gallup) (1) (2) (3) (4) (5) VARIABLES ”Current ladder” Unemployment rate -2.456*** -2.878*** -2.704*** (0.422) (0.431) (0.437) Job turnover rate 0.254 0.752*** (0.246) (0.230) Job creation rate 1.560*** 1.224*** (0.440) (0.352) Job destruction rate -0.764*** 0.331 (0.289) (0.267) Year and Month F.E. x x x x x Observations 502,334 502,334 502,334 502,334 502,334 R-squared 0.058 0.058 0.059 0.058 0.059 Aghion (Harvard), Akcigit (UPenn), Deaton (Princeton), and Roulet (Harvard) 32
Creative Destruction and Subjective Well-Being Empirics Individual level results 3/4 - Anticipated ladder (1) (2) (3) (4) (5) VARIABLES ”Anticipated ladder” Unemployment rate 0.108 -0.705** -0.677** (0.357) (0.307) (0.307) Job turnover rate 1.319*** 1.441*** (0.154) (0.151) Job creation rate 1.602*** 1.517*** (0.275) (0.259) Job destruction rate 1.099*** 1.373*** (0.230) (0.218) Year and Month F.E. x x x x x Observations 490,086 490,086 490,086 490,086 490,086 R-squared 0.077 0.077 0.077 0.077 0.077 Aghion (Harvard), Akcigit (UPenn), Deaton (Princeton), and Roulet (Harvard) 33
Creative Destruction and Subjective Well-Being Empirics Individual level results 4/4 - Worry (1) (2) (3) (4) (5) VARIABLES ”Worry” Unemployment rate 0.420*** 0.367*** 0.357*** (0.0715) (0.0759) (0.0784) Job turnover rate 0.159*** 0.0954** (0.0419) (0.0408) Job creation rate 0.0249 0.0693 (0.0747) (0.0672) Job destruction rate 0.263*** 0.119** (0.0554) (0.0569) Year and month F.E. x x x x x Observations 503,159 503,159 503,159 503,159 503,159 R-squared 0.014 0.013 0.014 0.014 0.014 Aghion (Harvard), Akcigit (UPenn), Deaton (Princeton), and Roulet (Harvard) 34
Creative Destruction and Subjective Well-Being Empirics Magnitude - Individual-level regressions A one standard deviation increase in job turnover has an effect on the current ladder of life equivalent to a 0.4 standard deviation decrease in the unemployment rate on the anticipated ladder of life equivalent to a 3.7 standard deviation decrease in the unemployment rate Aghion (Harvard), Akcigit (UPenn), Deaton (Princeton), and Roulet (Harvard) 35
Creative Destruction and Subjective Well-Being Empirics Robustness analysis We check whether results hold when restricting attention to 1 sub-periods We use an alternative database to measure creative destruction 2 We perform a panel analysis 3 We construct a predicted (Bartik-type) measure of job turnover to 4 neutralize variations of turnover driven by idiosyncratic local shocks that could have a direct effect on well-being Aghion (Harvard), Akcigit (UPenn), Deaton (Princeton), and Roulet (Harvard) 36
Creative Destruction and Subjective Well-Being Empirics Robustness check 1 - Restricting to 2005-2007 MSA-level 2005-2007 averages - (1) (2) (3) (4) (5) VARIABLES ”How satisfied are you with your life?” Unemployment rate -1.741*** -1.825*** -1.602*** (0.315) (0.295) (0.267) Job turnover rate 0.0938 0.185*** (0.0826) (0.0686) Job creation rate 0.935*** 0.718*** (0.169) (0.169) Job destruction rate -1.116*** -0.607*** (0.247) (0.223) Observations 364 364 364 364 364 R-squared 0.146 0.004 0.161 0.077 0.189 Aghion (Harvard), Akcigit (UPenn), Deaton (Princeton), and Roulet (Harvard) 37
Creative Destruction and Subjective Well-Being Empirics Robustness check 1 ct’ed -Restricting to 2008-2010 MSA-level 2008-2010 averages - (1) (2) (3) (4) (5) VARIABLES ”How satisfied are you with your life?” Unemployment rate -1.191*** -1.359*** -1.465*** (0.204) (0.203) (0.248) Job turnover rate 0.0441 0.336*** (0.120) (0.118) Job creation rate 0.780*** 0.0505 (0.274) (0.265) Job destruction rate -0.538** 0.602** (0.235) (0.278) Observations 364 364 364 364 364 R-squared 0.192 0.001 0.224 0.032 0.228 Aghion (Harvard), Akcigit (UPenn), Deaton (Princeton), and Roulet (Harvard) 38
Creative Destruction and Subjective Well-Being Empirics Panel analysis with a ”predicted” measure of creative destruction 1/2 We now construct a predicted (Bartik-type) measure of job turnover to neutralize variations of turnover driven by idiosyncratic local shocks that could have a direct effect on well-being CD m , t = ∑ � ω j , m ,2004 × CD j , USA , t j ω j , m ,2004 : share of sector j in total employment of MSA m in 2004 CD j , USA , t : national measure of creative destruction in sector j Aghion (Harvard), Akcigit (UPenn), Deaton (Princeton), and Roulet (Harvard) 42
Creative Destruction and Subjective Well-Being Empirics Panel analysis with a ”predicted” measure of creative destruction 2/2 (1) (2) (3) VARIABLES Current ladder (Gallup) Quarterly MSA-level averages Predicted job turnover 0.586*** 1.537*** 1.542*** (Quarterly) (0.200) (0.225) (0.516) Unemployment rate x x MSA F.E. x Year and quarter F.E. x x x Observations 5,600 5,600 5,600 R-squared 0.137 0.166 0.292 Robust standard errors in parentheses *** p < 0.01, ** p < 0.05, * p < 0.1 Aghion (Harvard), Akcigit (UPenn), Deaton (Princeton), and Roulet (Harvard) 43
Creative Destruction and Subjective Well-Being Empirics Heterogeneity analysis Interaction with type of sectors in the MSA 1 Interaction with state level UI generosity 2 Interaction with individuals’ age 3 Aghion (Harvard), Akcigit (UPenn), Deaton (Princeton), and Roulet (Harvard) 44
Creative Destruction and Subjective Well-Being Empirics Interaction with sectoral composition at MSA level We check whether the effect depends on the type of sectors in the MSA : SWB i , m , t = δ CD m , t + γ CD m , t ∗ Abovemedian m , t + θ Abovemedian m , t + α U m , t + β X i , m , t + T t + ǫ i , s , t Above median is either in terms of predicted productivity growth or in terms of predicted outsourcing threat using the same Bartik-type approach as before Individual controls include : gender, ethnicity, detailed education and family status, age, age2 Year and month fixed effects Standard errors clustered at the MSA level Aghion (Harvard), Akcigit (UPenn), Deaton (Princeton), and Roulet (Harvard) 45
Creative Destruction and Subjective Well-Being Empirics Measure of the type of sectors in the MSA The measure of productivity comes from the NBER-CES Manufacturing database: for each sector, we average annual 5-factors TFP growth over 2005-2009 (the data stops in 2009) Productivity m = ∑ � ω j , m × TFPgrowth j , USA j Following Autor et al. (2013) we proxy outsourcing by growth of imports in a given sector between 1991 and 2007 Outsourcing m = ∑ � ω j , m × Importgrowth j , USA j Aghion (Harvard), Akcigit (UPenn), Deaton (Princeton), and Roulet (Harvard) 46
Creative Destruction and Subjective Well-Being Empirics Interaction with the type of sectors- Productivity growth (1) (2) (3) (4) VARIABLES Life satisfaction (BRFSS) Above median * Job turnover 0.154** 0.146** (0.0711) (0.0711) Job turnover rate 0.0786 0.157*** (0.0599) (0.0604) Above median * Job destruction 0.190* 0.145 (0.107) (0.108) Job destruction rate -0.201** 0.0815 (0.0937) (0.0969) Above median * Job creation 0.145 0.151 (0.0986) (0.0986) Job creation rate 0.315*** 0.216** (0.0896) (0.0898) Unemployment rate x x Observations 825,298 825,298 825,298 825,298 R-squared 0.065 0.066 0.065 0.066 Aghion (Harvard), Akcigit (UPenn), Deaton (Princeton), and Roulet (Harvard) 47
Creative Destruction and Subjective Well-Being Empirics Interaction with the type of sectors - Outsourcing threat (1) (2) (3) (4) VARIABLES Life satisfaction (BRFSS) Above median * Job turnover -0.160** -0.181*** (0.0657) (0.0656) Job turnover rate 0.235*** 0.316*** (0.0483) (0.0485) Above median * Job destruction -0.290*** -0.323*** (0.105) (0.104) Job destruction rate 0.0725 0.332*** (0.0902) (0.0922) Above median * Job creation -0.0347 -0.0554 (0.0906) (0.0906) Job creation rate 0.388*** 0.309*** (0.0737) (0.0735) Unemployment rate x x Observations 852,783 852,783 852,783 852,783 R-squared 0.074 0.074 0.074 0.074 Aghion (Harvard), Akcigit (UPenn), Deaton (Princeton), and Roulet (Harvard) 48
Creative Destruction and Subjective Well-Being Empirics Interaction with state level UI generosity We check whether the effect depends on the generosity of UI : SWB m = δ CD m + γ CD m ∗ Abovemedian m + θ Abovemedian m + α U m + ǫ m Abovemedian m is a dummy equal to 1 if MSA m is located in a state above median in terms of UI generosity where UI generosity is proxied by the maximum weekly benefit amount We also split the sample of MSAs according to whether they are above median or not and run separately the baseline regression SWB m = δ CD m (+ α U m ) + ǫ m on the 2 sub-samples Aghion (Harvard), Akcigit (UPenn), Deaton (Princeton), and Roulet (Harvard) 49
Creative Destruction and Subjective Well-Being Empirics Interaction with UI generosity - 1/3 VARIABLES ”Current ladder” (Gallup) Above median * Job turnover 1.411** 1.784*** (0.637) (0.680) Job turnover rate -0.118 0.433 (0.529) (0.512) Above median * Job destruction 3.708*** 4.526*** (1.395) (1.507) Job destruction rate -5.380*** -3.524*** (0.946) (1.077) Above median * Job creation -1.484 -1.734 (1.689) (1.497) Job creation rate 6.186*** 4.900*** (0.993) (1.057) Above median UI -0.277* -0.382** -0.256 -0.330** (0.166) (0.175) (0.159) (0.165) Unemployment rate x x Observations 363 363 363 363 R-squared 0.093 0.237 0.216 0.281 Aghion (Harvard), Akcigit (UPenn), Deaton (Princeton), and Roulet (Harvard) 50
Creative Destruction and Subjective Well-Being Empirics Interaction with UI generosity - 2/3 VARIABLES Current ladder (Gallup) Panel A: States above median in terms of UI generosity Unemployment rate -1.357** -2.390*** (0.678) (0.654) Job turnover rate 1.299*** 2.039*** (0.356) (0.410) Observations 173 173 173 R-squared 0.048 0.072 0.198 Panel B: States below median Unemployment rate -3.644** -3.915*** (1.402) (1.463) Job turnover rate -0.115 0.597 (0.529) (0.511) Observations 190 190 190 R-squared 0.187 0.000 0.199 Aghion (Harvard), Akcigit (UPenn), Deaton (Princeton), and Roulet (Harvard) 51
Creative Destruction and Subjective Well-Being Empirics Interaction with UI generosity - 3/3 VARIABLES Current ladder (Gallup) Panel A: States above median in terms of UI generosity Job creation rate 4.701*** 3.415*** (1.371) (1.303) Job destruction rate -1.670 0.576 (1.036) (1.328) Observations 173 173 R-squared 0.150 0.209 Panel B: States below median Job creation rate 6.203*** 4.579*** (0.985) (1.374) Job destruction rate -5.375*** -3.070** (0.934) (1.487) Observations 190 190 R-squared 0.175 0.261 Aghion (Harvard), Akcigit (UPenn), Deaton (Princeton), and Roulet (Harvard) 52
Creative Destruction and Subjective Well-Being Empirics Interaction with individual’s age VARIABLES Current ladder (Gallup) Panel A: Age below median (Median age is 40) Unemployment rate -1.683** -2.358*** (0.827) (0.834) Job turnover rate 0.819** 1.394*** (0.409) (0.430) Observations 363 363 363 R-squared 0.043 0.020 0.095 Panel B: Age above median Unemployment rate -2.815*** -3.374*** (0.606) (0.612) ) Job turnover rate 0.333 1.156*** (0.363) (0.370) Observations 363 363 363 R-squared 0.122 0.003 0.158 Aghion (Harvard), Akcigit (UPenn), Deaton (Princeton), and Roulet (Harvard) 53
Creative Destruction and Subjective Well-Being Empirics Conclusion (1): Summary We have analyzed the relationship between turnover-driven growth and subjective well-being, using MSA level turnover data from the Longitudinal Business Database and subjective well-being data from Gallup-Healthways and from the BRFSS Our main results are consistent with a simple Schumpeterian model of growth and unemployment, namely: The overall effect of turnover (creative destruction) on subjective 1 well-being is unambiguously positive when we control for MSA-level unemployment, less so if we do not Creative destruction has a more positive effect on anticipated life 2 satisfaction than on current life satisfaction Creative destruction increases ”worry”, but less so if control for 3 unemployment Creative destruction has a more positive effect on subjective 4 well-being in MSAs dominated by sectors that are faster-growing or outsource less Aghion (Harvard), Akcigit (UPenn), Deaton (Princeton), and Roulet (Harvard) 54
Creative Destruction and Subjective Well-Being Empirics Conclusion (2): Extensions Compare more systematically the determinants of (per capita) 1 GDP growth with the determinants of life satisfaction Look at other individual and/or labor market characteristics 2 (training systems, availability of vocational education,..) which might have an impact on the effect of turnover on subjective well-being Aghion (Harvard), Akcigit (UPenn), Deaton (Princeton), and Roulet (Harvard) 55
Creative Destruction and Subjective Well-Being Empirics Solving the model (1): Equilibrium wage and profits Logarithmic technology for final good production implies that y jt = Y t / p jt . Then equilibrium wage is β w jt = 1 + β Y t = βπ Y t whereas equilibrium profit is 1 π jt = p jt y jt − w jt = 1 + β Y t = π Y t with 1 π = 1 + β . Aghion (Harvard), Akcigit (UPenn), Deaton (Princeton), and Roulet (Harvard) 56
Creative Destruction and Subjective Well-Being Empirics Extension: Risk aversion (1) We now consider risk averse individuals with U = ln C . Now well-being can be shown to be equal to: � � � g � W u ( c )= ln c = 1 x 1 + 1 1 + x ln ( b ) + 1 + x ln ( βπ ) ρ + ln Y ρ ρ Aghion (Harvard), Akcigit (UPenn), Deaton (Princeton), and Roulet (Harvard) 57
Creative Destruction and Subjective Well-Being Empirics Extension: Risk aversion (2) Proposition A higher turnover rate has a less positive effect on life satisfaction of agents that are risk-averse with U = ln C than on risk-neutral agents : ∂ W u ( c )= ln c < 0. ∂ x Moving continuously from the baseline case where individuals are risk-neutral towards the risk-averse case where individuals have log preferences, makes the effect of creative destruction on life satisfaction become increasingly less positive (or increasingly more negative) Aghion (Harvard), Akcigit (UPenn), Deaton (Princeton), and Roulet (Harvard) 58
Creative Destruction and Subjective Well-Being Empirics Extension: Exogenous job destruction (1) In our baseline model, the only source of job destruction, as well as job creation, was new entry. Now assume instead that each job can also be destroyed at the rate φ . Upon this shock, worker joins the unemployment pool and the 1 product line becomes idle. When a new entrant comes into this product line at the rate x , it 2 first posts a vacancy in which case then the same product line moves from ”idle” into ”vacant” state. When a vacant product line finds a suitable worker, the product line 3 enter into ”production state”. Similarly, if a new entrant enters into a actively producing line, then the worker joins the unemployment pool and the new firm posts a vacancy as in the previous model. Aghion (Harvard), Akcigit (UPenn), Deaton (Princeton), and Roulet (Harvard) 59
Creative Destruction and Subjective Well-Being Empirics Extension: Exogenous job destruction (2) A product line j ∈ [ 0, 1 ] can be in one of three states: production µ 1 vacant v 2 idle i 3 We have the steady-state flow equations: production : m = µφ + µ x µ x + ix = m vacant : idle : µφ = ix Aghion (Harvard), Akcigit (UPenn), Deaton (Princeton), and Roulet (Harvard) 60
Creative Destruction and Subjective Well-Being Empirics Extension: Exogenous job destruction (3) For analytical tractability, assume α = 0.5. Then the unemployment rate is simply � ( Ψ + 1 ) 2 − 4 [ Ψ − Ψ 2 x 2 ] ( Ψ + 1 ) − u = 1 − 2 [ Ψ − Ψ 2 x 2 ] where Ψ ≡ 1 + φ / x . Aghion (Harvard), Akcigit (UPenn), Deaton (Princeton), and Roulet (Harvard) 61
Creative Destruction and Subjective Well-Being Empirics Extension: Exogenous job destruction (4) Aghion (Harvard), Akcigit (UPenn), Deaton (Princeton), and Roulet (Harvard) 62
Creative Destruction and Subjective Well-Being Empirics Extension: Exogenous job destruction (5) Aghion (Harvard), Akcigit (UPenn), Deaton (Princeton), and Roulet (Harvard) 63
Creative Destruction and Subjective Well-Being Empirics Summary statistics - subjective well-being Mean Standard deviation Min Max Current ladder (Gallup) 6.78 1.95 0 10 MSA-level averages 6.74 0.187 6.15 7.51 Anticipated ladder (Gallup) 8.05 1.99 0 10 MSA-level averages 7.97 0.187 7.42 8.48 Worry (Gallup) 0.35 0.48 0 1 MSA-level averages 0.35 0.034 0.24 0.46 Life satisfaction (BRFSS) 3.37 0.63 1 4 MSA-level averages 3.37 0.046 3.14 3.58 Aghion (Harvard), Akcigit (UPenn), Deaton (Princeton), and Roulet (Harvard) 64
Creative Destruction and Subjective Well-Being Empirics Summary statistics -creative destruction (2005-2010 averages) Mean Standard deviation Min Max JJob turnover rate (BDS) 0.29 0.035 0.18 0.43 Job creation rate (BDS) 0.15 0.015 0.08 0.22 Job destruction rate (BDS) 0.14 0.017 0.09 0.22 Unemployment rate (BLS) 0.07 0.015 0.03 0.24 Aghion (Harvard), Akcigit (UPenn), Deaton (Princeton), and Roulet (Harvard) 65
Creative Destruction and Subjective Well-Being Empirics Unemployment and job turnover rates - 2005-2010 Unemployment and job turnover rates MSA-level 2005-2010 averages 45 40 Job turnover rate (in %) 35 30 25 20 5 10 15 20 25 Unemployment rate (in %) beta=0.57, R2=0.07; correlation=0.27 Aghion (Harvard), Akcigit (UPenn), Deaton (Princeton), and Roulet (Harvard) 66
Creative Destruction and Subjective Well-Being Empirics Unemployment and job turnover rates - 2005-2007 Unemployment and job turnover rates MSA-level 2005-2007 averages 50 Job turnover rate (in %) 40 30 20 0 5 10 15 20 Unemployment rate (in %) beta=0.46, R2=0.02; correlation=0.15 Aghion (Harvard), Akcigit (UPenn), Deaton (Princeton), and Roulet (Harvard) 67
Creative Destruction and Subjective Well-Being Empirics Unemployment and job turnover rates - 2008-2011 Unemployment and turnover rates 2008-2011 averages, MSA-level 40 35 Job turnover rate (in %) 30 25 20 15 5 10 15 20 25 30 Unemployment rate (in %) beta=0.48, R2=0.11 Aghion (Harvard), Akcigit (UPenn), Deaton (Princeton), and Roulet (Harvard) 68
Royal Economic Society
Growth through Heterogeneous Innovations 1 Ufuk Akcigit University of Pennsylvania & NBER Royal Economic Society Conference - April 1st, 2015 1 joint work with Willliam Kerr (Harvard University) Akcigit (UPenn) and Kerr (Harvard) Growth through Heterogeneous Innovations April 1st, 2015 1
Introduction Introduction Innovations come in different sizes and shapes: internal vs external product vs process radical vs incremental Small vs large firms have different incentives to innovate. Spillovers generated by different innovations and different-sized firms are likely to be different. How is firm size related to the innovation size? Which type of firms generate bigger spillovers? The answer is important for many reasons, particularly for policy. Akcigit (UPenn) and Kerr (Harvard) Growth through Heterogeneous Innovations April 1st, 2015 2
Introduction This Paper: To answer these questions, we bridge a tight link between general equilibrium firm dynamics theory and micro data of firms, innovations and patent citations. We proceed in 3 steps: Part 1. Theory New theory of firm dynamics and innovation where: firms come in different sizes, firms compete in the market for leadership, firms produce different type and size innovations, hence generate different spillovers, introduce an explicit model of citations. Akcigit (UPenn) and Kerr (Harvard) Growth through Heterogeneous Innovations April 1st, 2015 3
Introduction This Paper: To answer these questions, we bridge a tight link between general equilibrium firm dynamics theory and micro data of firms, innovations and patent citations. We proceed in 3 steps: Part 2. Empirics Using Census and patent data, study the empirical link between Firm size vs firm growth, Firm size vs innovation intensity, Firm size and innovation quality. Akcigit (UPenn) and Kerr (Harvard) Growth through Heterogeneous Innovations April 1st, 2015 3
Introduction This Paper: To answer these questions, we bridge a tight link between general equilibrium firm dynamics theory and micro data of firms, innovations and patent citations. We proceed in 3 steps: Part 3. Quantitative Analysis Using indirect inference we find: External innovation generates larger spillovers. External innovation does not scale with firm size, = ⇒ Small firms generate larger spillovers on average. Akcigit (UPenn) and Kerr (Harvard) Growth through Heterogeneous Innovations April 1st, 2015 3
Model Part 1. Model Akcigit (UPenn) and Kerr (Harvard) Growth through Heterogeneous Innovations April 1st, 2015 4
Model The Model Economy quality level q sector j 0 1 US Economy Akcigit (UPenn) and Kerr (Harvard) Growth through Heterogeneous Innovations April 1st, 2015 5
Model The Model Economy quality level q sector j 0 1 GDP = Sectors combined Akcigit (UPenn) and Kerr (Harvard) Growth through Heterogeneous Innovations April 1st, 2015 6
Model Sector-specific Productivities quality level q sector j 0 1 Akcigit (UPenn) and Kerr (Harvard) Growth through Heterogeneous Innovations April 1st, 2015 7
Model Example of a Firm and Firm Size Heterogeneity quality level q sector j 0 1 Akcigit (UPenn) and Kerr (Harvard) Growth through Heterogeneous Innovations April 1st, 2015 8
Model Example of another Firm quality level q sector j 0 1 Akcigit (UPenn) and Kerr (Harvard) Growth through Heterogeneous Innovations April 1st, 2015 9
Model Productivity Growth: Internal R&D quality level q internal R&D sector j 0 1 Akcigit (UPenn) and Kerr (Harvard) Growth through Heterogeneous Innovations April 1st, 2015 10
Model Productivity Growth: Internal R&D quality level q internal R&D sector j 0 1 Akcigit (UPenn) and Kerr (Harvard) Growth through Heterogeneous Innovations April 1st, 2015 11
Model Productivity Growth: External R&D quality level q sector j 0 1 External R&D Akcigit (UPenn) and Kerr (Harvard) Growth through Heterogeneous Innovations April 1st, 2015 12
Model Productivity Growth: External R&D quality level q sector j 0 1 External R&D Akcigit (UPenn) and Kerr (Harvard) Growth through Heterogeneous Innovations April 1st, 2015 13
Model Reallocation is Taking Place... quality level q sector j 0 1 External R&D Akcigit (UPenn) and Kerr (Harvard) Growth through Heterogeneous Innovations April 1st, 2015 14
Model Competition Creates Selection quality level q sector j 0 1 External R&D Akcigit (UPenn) and Kerr (Harvard) Growth through Heterogeneous Innovations April 1st, 2015 15
Model Eventually Some Firms Exit quality level q sector j 0 1 External R&D Exit Akcigit (UPenn) and Kerr (Harvard) Growth through Heterogeneous Innovations April 1st, 2015 16
Model In the Meantime... quality level q sector j 0 1 Akcigit (UPenn) and Kerr (Harvard) Growth through Heterogeneous Innovations April 1st, 2015 17
Model Some New Entrants Show Up quality level q new entrants sector j 0 1 Akcigit (UPenn) and Kerr (Harvard) Growth through Heterogeneous Innovations April 1st, 2015 18
Model And New Entrants Replace Incumbents quality level q sector j 0 1 Akcigit (UPenn) and Kerr (Harvard) Growth through Heterogeneous Innovations April 1st, 2015 19
Model R&D and Innovation Profits in each line j is linear in quality q j : π j = π q j Innovation in product line j : q new = ( 1 + s ) q old j , j s depends on R&D type Firms invest in two types of R&D: Internal R&D ( z ) 1 External R&D ( x ) 2 Akcigit (UPenn) and Kerr (Harvard) Growth through Heterogeneous Innovations April 1st, 2015 20
Model External and Heterogeneous Innovations External innovations can be of different qualities: Akcigit (UPenn) and Kerr (Harvard) Growth through Heterogeneous Innovations April 1st, 2015 21
Model External and Heterogeneous Innovations External innovations can be of different qualities: follow-on innovation, 1 Akcigit (UPenn) and Kerr (Harvard) Growth through Heterogeneous Innovations April 1st, 2015 21
Model External and Heterogeneous Innovations External innovations can be of different qualities: follow-on innovation, 1 major innovation. 2 Akcigit (UPenn) and Kerr (Harvard) Growth through Heterogeneous Innovations April 1st, 2015 21
Model External and Heterogeneous Innovations External innovations can be of different qualities: follow-on innovation, 1 major innovation. 2 Follow-on innovations build on the previous technology and have declining impacts. Akcigit (UPenn) and Kerr (Harvard) Growth through Heterogeneous Innovations April 1st, 2015 21
Model External and Heterogeneous Innovations External innovations can be of different qualities: follow-on innovation, 1 major innovation. 2 Follow-on innovations build on the previous technology and have declining impacts. After k follow-on improvements, the step size becomes s = ηα k such that: q new = ( 1 + ηα k ) q old j j Akcigit (UPenn) and Kerr (Harvard) Growth through Heterogeneous Innovations April 1st, 2015 21
Model External and Heterogeneous Innovations External innovations can be of different qualities: follow-on innovation, 1 major innovation. 2 Follow-on innovations build on the previous technology and have declining impacts. After k follow-on improvements, the step size becomes s = ηα k such that: q new = ( 1 + ηα k ) q old j j Major innovations create a new technology wave and therefore k = 0: q new = ( 1 + η ) q old j j Akcigit (UPenn) and Kerr (Harvard) Growth through Heterogeneous Innovations April 1st, 2015 21
Model Evolution of Step Size “ s ” Evolution of Innovation Quality (step size) 1 0.9 0.8 0.7 Innovation step size, s 0.6 0.5 0.4 0.3 0.2 0.1 0 0 1 2 3 4 5 6 7 0 1 2 0 1 2 3 4 5 6 Number of times the latest technology is improved Akcigit (UPenn) and Kerr (Harvard) Growth through Heterogeneous Innovations April 1st, 2015 22
Model Evolution of Step Size “ s ” Evolution of Innovation Quality (step size) 1 0.9 0.8 0.7 Innovation step size, s 0.6 0.5 0.4 0.3 0.2 0.1 0 0 1 2 3 4 5 6 7 0 1 2 0 1 2 3 4 5 6 Number of times the latest technology is improved Akcigit (UPenn) and Kerr (Harvard) Growth through Heterogeneous Innovations April 1st, 2015 22
Model Evolution of Step Size “ s ” Evolution of Innovation Quality (step size) 1 0.9 0.8 0.7 Innovation step size, s 0.6 0.5 0.4 0.3 0.2 0.1 0 0 1 2 3 4 5 6 7 0 1 2 0 1 2 3 4 5 6 Number of times the latest technology is improved Akcigit (UPenn) and Kerr (Harvard) Growth through Heterogeneous Innovations April 1st, 2015 22
Model Evolution of Step Size “ s ” Evolution of Innovation Quality (step size) 1 0.9 0.8 0.7 Innovation step size, s 0.6 0.5 0.4 0.3 0.2 0.1 0 0 1 2 3 4 5 6 7 0 1 2 0 1 2 3 4 5 6 Number of times the latest technology is improved Akcigit (UPenn) and Kerr (Harvard) Growth through Heterogeneous Innovations April 1st, 2015 22
Model Evolution of Step Size “ s ” Evolution of Innovation Quality (step size) 1 Major Breakthroughs Follow ‐ on Innovations 0.9 0.8 0.7 Innovation step size, s 0.6 0.5 0.4 0.3 0.2 0.1 0 0 1 2 3 4 5 6 7 0 1 2 0 1 2 3 4 5 6 Number of times the latest technology is improved Akcigit (UPenn) and Kerr (Harvard) Growth through Heterogeneous Innovations April 1st, 2015 22
Model Internal Innovation ( z j ) Internal innovations improve the quality by s = λ such that: q new = ( 1 + λ ) q old j j Akcigit (UPenn) and Kerr (Harvard) Growth through Heterogeneous Innovations April 1st, 2015 23
Model Sequence of Innovations in line j and Citations Example: | | ηα 2 ηα 3 η ηα λ λ η λ ηα | | ... P 1, f 1 P 2, f 2 P 3, f 3 P 4, f 3 P 5, f 3 P 6, f 4 P 7, f 5 P 8, f 5 P 9, f 6 | | � �� � � �� � Tech Cluster 1 Tech Cluster 2 A N EXAMPLE OF A SEQUENCE OF INNOVATIONS IN A PRODUCT LINE Akcigit (UPenn) and Kerr (Harvard) Growth through Heterogeneous Innovations April 1st, 2015 24
Model Sequence of Innovations in line j and Citations Example: | | ηα 2 ηα 3 η ηα λ λ η λ ηα | | ... P 1, f 1 P 2, f 2 P 3, f 3 P 4, f 3 P 5, f 3 P 6, f 4 P 7, f 5 P 8, f 5 P 9, f 6 | | � �� � � �� � Tech Cluster 1 Tech Cluster 2 A N EXAMPLE OF A SEQUENCE OF INNOVATIONS IN A PRODUCT LINE Cited w/p Citing Cited w/p Citing ηα 3 γ P 1 : P 2 , P 3 , P 4 , P 5 , P 6 P 6 : none ηγ P 2 : ηαγ P 3 , P 4 , P 5 , P 6 P 7 : ηγ P 8 , P 9 , ... ηα 2 γ P 3 : P 4 , P 5 , P 6 P 8 : P 9 , ... λγ ηα 2 γ P 4 : λγ P 5 , P 6 P 9 : ... P 5 : P 6 λγ Akcigit (UPenn) and Kerr (Harvard) Growth through Heterogeneous Innovations April 1st, 2015 24
Recommend
More recommend
