Recession Scars and the Growth Potential of Newborn Firms in General - - PowerPoint PPT Presentation

Recession Scars and the Growth Potential

• f Newborn Firms in General Equilibrium

Petr Sedl´ aˇ cek∗ and Vincent Sterk†

*Bonn University † University College London

Dutch National Bank October 18, 2013

Sedl´ aˇ cek, Sterk (Bonn, UCL) Recession Scars DNB, October 2013 1 / 56

Motivation

Weak firm entry during Great Recession job creation of entrants in 2006: 3.5 million jobs job creation of entrants in 2010: 2.3 million jobs

Sedl´ aˇ cek, Sterk (Bonn, UCL) Recession Scars DNB, October 2013 2 / 56

Motivation

Weak firm entry during Great Recession job creation of entrants in 2006: 3.5 million jobs job creation of entrants in 2010: 2.3 million jobs Does this have (persistent) macroeconomic effects?

Sedl´ aˇ cek, Sterk (Bonn, UCL) Recession Scars DNB, October 2013 2 / 56

This paper: 1. Empirical Analysis

US Business Dynamics Statistics (BDS) data, 1979-2010 follow job creation by cohorts of entrants as they age

◮ extensive margin (number of firms) ◮ intensive margin (average firm size)

document cyclical patterns quick & dirty counterfactuals for potential macro impact

Sedl´ aˇ cek, Sterk (Bonn, UCL) Recession Scars DNB, October 2013 3 / 56

This paper: 2. General Equilibrium Model

build heterogeneous firm model with aggregate shocks

◮ heterogeneity in technology types ◮ endogenous entry ◮ aggregate shocks ◮ general equilibrium

fit model to data redo counterfactuals, now accounting for GE effects

Sedl´ aˇ cek, Sterk (Bonn, UCL) Recession Scars DNB, October 2013 4 / 56

Empirical evidence

Sedl´ aˇ cek, Sterk (Bonn, UCL) Recession Scars DNB, October 2013 5 / 56

Data and methodology

BDS data, 1979-2010

◮ 98% of all US private employment ◮ annual information: number of firms, net job creation ◮ broken down according to age, size, sectors Sedl´ aˇ cek, Sterk (Bonn, UCL) Recession Scars DNB, October 2013 6 / 56

Data and methodology

BDS data, 1979-2010

◮ 98% of all US private employment ◮ annual information: number of firms, net job creation ◮ broken down according to age, size, sectors

employment and average firm size of entrants age breakdown → track them until 5 years old inspect patterns within and across cohorts

Sedl´ aˇ cek, Sterk (Bonn, UCL) Recession Scars DNB, October 2013 6 / 56

Three stylized facts

• 1. cohort-level employment is largely determined in year of birth
• 2. variation in cohort-level employment is mainly driven by intensive

margin

• 3. cohorts of small firms are born in times of low economic activity

Sedl´ aˇ cek, Sterk (Bonn, UCL) Recession Scars DNB, October 2013 7 / 56

Stylized facts - 1. cohort employment highly persistent

Cohort employment at t and t + 5

1980 1985 1990 1995 2000 2005 2010 1.5 2 2.5 3 3.5 4 year of birth employment (millions) age 0 age 5

Sedl´ aˇ cek, Sterk (Bonn, UCL) Recession Scars DNB, October 2013 8 / 56

Stylized facts - 1. cohort employment highly persistent

Correlation of employment at t and t + a

1 1.5 2 2.5 3 3.5 4 4.5 5

• 0.4
• 0.2

0.2 0.4 0.6 0.8 1 age correlation coefficient cohort

Sedl´ aˇ cek, Sterk (Bonn, UCL) Recession Scars DNB, October 2013 9 / 56

Stylized facts - 1. cohort employment highly persistent

Correlation of employment at t and t + a

1 1.5 2 2.5 3 3.5 4 4.5 5

• 0.4
• 0.2

0.2 0.4 0.6 0.8 1 age correlation coefficient cohort aggregate

Sedl´ aˇ cek, Sterk (Bonn, UCL) Recession Scars DNB, October 2013 10 / 56

Three stylized facts

• 1. cohort-level employment is largely determined in year of birth
• 2. variation in cohort-level employment is mainly driven by

intensive margin

• 3. cohorts of small firms are born in times of low economic activity

Sedl´ aˇ cek, Sterk (Bonn, UCL) Recession Scars DNB, October 2013 11 / 56

Stylized facts - 2. intensive margin dominates

decompose variation in cohort-level employment:

◮ intensive (firm size) vs. extensive (number of firms) margin ◮ according to age Sedl´ aˇ cek, Sterk (Bonn, UCL) Recession Scars DNB, October 2013 12 / 56

Stylized facts - 2. intensive margin dominates

decompose variation in cohort-level employment:

◮ intensive (firm size) vs. extensive (number of firms) margin ◮ according to age

ln Ea,t = ln S0,t−a + ln N0,t−a +

a

• j=1

ln γj,t−a+j +

a

• j=1

δj,t−a+j γa,t =

Sa,t Sa−1,t−1

δa,t =

Na,t Na−1,t−1

Sedl´ aˇ cek, Sterk (Bonn, UCL) Recession Scars DNB, October 2013 12 / 56

Stylized facts - 2. intensive margin dominates

Variance decomposition of E5,t

‐10 10 20 30 40 50 60 70 80 Average firm size Number of firms 5th year 4th year 3rd year 2nd year 1st year Entrants Sedl´ aˇ cek, Sterk (Bonn, UCL) Recession Scars DNB, October 2013 13 / 56

Three stylized facts

• 1. cohort-level employment is largely determined in year of birth
• 2. variation in cohort-level employment is mainly driven by intensive

margin

• 3. cohorts of small firms are born in times of low economic

activity

Sedl´ aˇ cek, Sterk (Bonn, UCL) Recession Scars DNB, October 2013 14 / 56

Stylized facts - 3. recession cohorts are small

Cohort employment at t and t + 5

1980 1985 1990 1995 2000 2005 2010 1.5 2 2.5 3 3.5 4 year of birth employment (millions) age 0 age 5

Sedl´ aˇ cek, Sterk (Bonn, UCL) Recession Scars DNB, October 2013 15 / 56

Stylized facts - 3. recession cohorts are small

Cohort-level average size; weak and strong cohorts

0.5 1 1.5 2 2.5 3 3.5 4 4.5 5

• 1.5
• 1
• 0.5

0.5 1 1.5 age deviation from trend/standard deviation age 0 cohort (weak) cohort (strong)

Sedl´ aˇ cek, Sterk (Bonn, UCL) Recession Scars DNB, October 2013 16 / 56

Stylized facts - 3. recession cohorts are small

Cohort-level and aggregate average size; weak and strong cohorts

0.5 1 1.5 2 2.5 3 3.5 4 4.5 5

• 1.5
• 1
• 0.5

0.5 1 1.5 employment age deviation from trend/standard deviation age 0 cohort (weak) aggregate (weak) cohort (strong) aggregate (strong)

Sedl´ aˇ cek, Sterk (Bonn, UCL) Recession Scars DNB, October 2013 17 / 56

Stylized facts - 3. recession cohorts are small

Table: Correlations of average size with BC indicators in year t

age Levels linear trend CF filter(6,12) E/L E/L GDP E/L GDP cohort-level a = 0 0.50 0.36 0.33 0.74 0.61 a = 5 0.44 0.28 0.10 0.74 0.74

Sedl´ aˇ cek, Sterk (Bonn, UCL) Recession Scars DNB, October 2013 18 / 56

Stylized facts - 3. recession cohorts are small

Table: Correlations of average size with BC indicators in year t

age Levels linear trend CF filter(6,12) E/L E/L GDP E/L GDP cohort-level a = 0 0.50 0.36 0.33 0.74 0.61 a = 5 0.44 0.28 0.10 0.74 0.74 aggregate-level a = 0 0.75 0.74 0.76 0.72 a = 5 −0.17 −0.37 −0.73 −0.65

Sedl´ aˇ cek, Sterk (Bonn, UCL) Recession Scars DNB, October 2013 18 / 56

Stylized facts - 3. recession cohorts are small

Table: Correlations of employment with BC indicators in year t

age Levels linear trend CF filter(6,12) E/L E/L GDP E/L GDP cohort-level a = 0 0.62 0.41 0.43 0.76 0.72 a = 5 0.59 0.35 0.23 0.84 0.88 aggregate-level a = 0 0.91 0.88 0.96 0.98 a = 5 −0.07 −0.26 −0.67 −0.55

Sedl´ aˇ cek, Sterk (Bonn, UCL) Recession Scars DNB, October 2013 19 / 56

What are the aggregate implications?

Sedl´ aˇ cek, Sterk (Bonn, UCL) Recession Scars DNB, October 2013 20 / 56

What are the aggregate implications?

2 counterfactual series for aggregate employment:

◮ extensive margin: hold the number of firms aged 0 to 5 fixed at average Sedl´ aˇ cek, Sterk (Bonn, UCL) Recession Scars DNB, October 2013 20 / 56

What are the aggregate implications?

2 counterfactual series for aggregate employment:

◮ extensive margin: hold the number of firms aged 0 to 5 fixed at average ◮ both margins: hold the number and average size of firms aged 0 to 5

fixed at average

Sedl´ aˇ cek, Sterk (Bonn, UCL) Recession Scars DNB, October 2013 20 / 56

What are the aggregate implications?

2 counterfactual series for aggregate employment:

◮ extensive margin: hold the number of firms aged 0 to 5 fixed at average ◮ both margins: hold the number and average size of firms aged 0 to 5

fixed at average

plot the differential from aggregate employment

Sedl´ aˇ cek, Sterk (Bonn, UCL) Recession Scars DNB, October 2013 20 / 56

What are the aggregate implications?

Differential employment ( Et−Ecount,t

Et

100)

1980 1985 1990 1995 2000 2005 2010

• 5
• 4
• 3
• 2
• 1

1 2 percent of aggregate employment extensive margin fixed

Sedl´ aˇ cek, Sterk (Bonn, UCL) Recession Scars DNB, October 2013 21 / 56

What are the aggregate implications?

Differential employment ( Et−Ecount,t

Et

100)

1980 1985 1990 1995 2000 2005 2010

• 5
• 4
• 3
• 2
• 1

1 2 percent of aggregate employment extensive margin fixed extensive and intensive margin fixed

Sedl´ aˇ cek, Sterk (Bonn, UCL) Recession Scars DNB, October 2013 22 / 56

What next?

we observe 3 new stylized facts

explanations

ultimately interested in macroeconomic implications counterfactuals cannot account for GE effects! → build a GE model that can explain the above facts investigate scarring effects of recessions in model

Sedl´ aˇ cek, Sterk (Bonn, UCL) Recession Scars DNB, October 2013 23 / 56

General equilibrium model

Sedl´ aˇ cek, Sterk (Bonn, UCL) Recession Scars DNB, October 2013 24 / 56

Related literature

Hopenhayn (1992), Hopenhayn and Rogerson (1993), Cooley and Quadrini (2001), Melitz (2005) Lee and Mukoyama (2012), Clementi and Palazzo (2010), Siemer (2012) Kaas and Kircher (2011), Schaal (2012), Sedl´ aˇ cek (2012)

Sedl´ aˇ cek, Sterk (Bonn, UCL) Recession Scars DNB, October 2013 25 / 56

Model features

Neoclassical general equilibrium model with heterogeneous firms heterogeneity in returns to scale

◮ BDS data, many old small firms ◮ many startups do not want to grow: Campbell and de Nardi (2009),

Hurst and Pugsley (2012)

◮ direct evidence: Basu and Fernald (1997), Holmes and Stevens (2012) Sedl´ aˇ cek, Sterk (Bonn, UCL) Recession Scars DNB, October 2013 26 / 56

Model features

Neoclassical general equilibrium model with heterogeneous firms heterogeneity in returns to scale

◮ BDS data, many old small firms ◮ many startups do not want to grow: Campbell and de Nardi (2009),

Hurst and Pugsley (2012)

◮ direct evidence: Basu and Fernald (1997), Holmes and Stevens (2012)

costly labor adjustment

◮ firms grow gradually as they age Sedl´ aˇ cek, Sterk (Bonn, UCL) Recession Scars DNB, October 2013 26 / 56

Model features

endogenous entry

◮ number and composition of entrants endogenous

aggregate uncertainty estimated on BDS data

Sedl´ aˇ cek, Sterk (Bonn, UCL) Recession Scars DNB, October 2013 27 / 56

Heterogeneous firms

Sedl´ aˇ cek, Sterk (Bonn, UCL) Recession Scars DNB, October 2013 28 / 56

Existing firms

endogenous measure, owned by household produce a homogeneous good using only labor finite number of technology types i = 1, ..., I. production function y(nt, At; i) = yi,t = ziAtnαi

t

Sedl´ aˇ cek, Sterk (Bonn, UCL) Recession Scars DNB, October 2013 29 / 56

Existing firms

Firms maximize expected discounted profits: Vi,a(ni,a−1,t−1, St) = max

ni,a,t

ziAtnαi

i,a,t − Wtni,a,t − Qtζa(ni,a,t, ni,a−1,t−1)

+ (1 − ρa) EtΛt,t+1Vi,a+1 (ni,a,t, St+1)

• Sedl´

aˇ cek, Sterk (Bonn, UCL) Recession Scars DNB, October 2013 30 / 56

Firm entry

free entry pay cost χ to choose business opportunity of any type

Sedl´ aˇ cek, Sterk (Bonn, UCL) Recession Scars DNB, October 2013 31 / 56

Firm entry

free entry pay cost χ to choose business opportunity of any type there is a time-invariant mass of opportunities per type: Ψ =

i ψi

Sedl´ aˇ cek, Sterk (Bonn, UCL) Recession Scars DNB, October 2013 31 / 56

Firm entry

free entry pay cost χ to choose business opportunity of any type there is a time-invariant mass of opportunities per type: Ψ =

i ψi

some startup attempts fail due to a coordination friction

◮ matching function Sedl´ aˇ cek, Sterk (Bonn, UCL) Recession Scars DNB, October 2013 31 / 56

Firm entry

mass of entrants in technology type i mi,0,t = xφ

i,tψ1−φ i

, for i = 1, 2, .., I,

Sedl´ aˇ cek, Sterk (Bonn, UCL) Recession Scars DNB, October 2013 32 / 56

Firm entry

mass of entrants in technology type i mi,0,t = xφ

i,tψ1−φ i

, for i = 1, 2, .., I, probability of starting up a technology type i given payment of entry cost Pi,t = mi,0,t/xi,t

Sedl´ aˇ cek, Sterk (Bonn, UCL) Recession Scars DNB, October 2013 32 / 56

Firm entry

mass of entrants in technology type i mi,0,t = xφ

i,tψ1−φ i

, for i = 1, 2, .., I, probability of starting up a technology type i given payment of entry cost Pi,t = mi,0,t/xi,t free entry condition χ = Pi,tVi,0,t (0, St) , for i = 1, 2, .., I,

Sedl´ aˇ cek, Sterk (Bonn, UCL) Recession Scars DNB, October 2013 32 / 56

Firm entry decisions

technology type is a choice more attractive technologies are tougher to startup entry happens in all technology types

Sedl´ aˇ cek, Sterk (Bonn, UCL) Recession Scars DNB, October 2013 33 / 56

Representative household, market clearing and shocks

Sedl´ aˇ cek, Sterk (Bonn, UCL) Recession Scars DNB, October 2013 34 / 56

Households

representative household with continuum of members. Choose consumption and labor: max

{Ct,Nt}∞

t=0

E0

∞

• t=0

βt C1−σ

t

1 − σ + ZtN1+κ

t

1 + κ

• s.t.

Ct = WtNt + Πt

Sedl´ aˇ cek, Sterk (Bonn, UCL) Recession Scars DNB, October 2013 35 / 56

Market clearing

We impose maximum age K (ρK = 1). Aggregate resource constraint:

I

• i=1

K

• a=0

mi,j,t (yi,a,t − Qtζi,a,t) −

I

• i=1

xi,tχ = Ct Labor market clearing:

I

• i=1

K

• a=0

mi,a,tni,a,t = Nt Aggregate state: St = {mi,a,t, ni,a,t−1, At, Qt, Zt}a=0,...,K

i=1,...,I

⇒ large but finite-dimensional object

Sedl´ aˇ cek, Sterk (Bonn, UCL) Recession Scars DNB, October 2013 36 / 56

Aggregate shocks

yi,t = ziAtnαi

t

WtC−σ

t

= ZtNκ

t

Wt = αiziAtnαi−1

i,a,t

• 1 − Qtζ′

i,a,t) + (1 − ρa)βEtΛt,t+1Qt+1ζ′ i,a,t+1

• stationary processes with continuous support

estimated and used for counterfactuals

Sedl´ aˇ cek, Sterk (Bonn, UCL) Recession Scars DNB, October 2013 37 / 56

Quantitative implementation

Sedl´ aˇ cek, Sterk (Bonn, UCL) Recession Scars DNB, October 2013 38 / 56

Parametrization

Parameter values obtained using hybrid of: matching long-run targets

◮ average size age 0 ◮ average size age 1 ◮ size distribution of firms aged 16-20 (use BDS size brackets)

matching key moments

◮ volatility number of entrants ◮ volatility avg. size age 5 / volatility avg. size age 0

maximum likelihood estimation (aggregate shock processes)

◮ time series used: output, employment rate, average entrant size ◮ obtain estimated shocks as by-product Sedl´ aˇ cek, Sterk (Bonn, UCL) Recession Scars DNB, October 2013 39 / 56

Parametrization

Adjustment cost assumed to be quadratic: ζa(ni,a,t, ni,a−1,t−1) = ζa 2 (ni,a,t − ni,a−1,t−1)2 ζ0 ≥ ζ1 = ζ2 = ... = ζK. ζ1 calibrated to match growth rate of average size young cohorts ζ0 calibrated to match relative volatility of avg. size at age 5 initial level ni,−1 calibrated

Sedl´ aˇ cek, Sterk (Bonn, UCL) Recession Scars DNB, October 2013 40 / 56

Parametrization

Adjustment cost assumed to be quadratic: ζa(ni,a,t, ni,a−1,t−1) = ζa 2 (ni,a,t − ni,a−1,t−1)2 ζ0 ≥ ζ1 = ζ2 = ... = ζK. ζ1 calibrated to match growth rate of average size young cohorts ζ0 calibrated to match relative volatility of avg. size at age 5 initial level ni,−1 calibrated Exit rates age-dependent ρa = ξ0 + ξ1/a, ξ0, ξ1 > 0 parameters ξ0 and ξ1 fitted to exit rates observed in BDS

Sedl´ aˇ cek, Sterk (Bonn, UCL) Recession Scars DNB, October 2013 40 / 56

Parameter values

U]^ha 8? E]he^n]pa` l]n]iapano l]n]iapan r]hqa p]ncap3aopei]pa " `eo_kqjp b]_pkn 42>; ]jjq]h ejpanaop n]pa 8+ , nah]pera neog ]ranoekj _ka'_eajp 5 hkc1qpehepu ( qpehepu kb haeoqna l]n]iapan 5 qjep Hneo_d ah]opee_pu & ]`fqopiajp _kop0 ]ca 5194 4244<

• eva kb 5 ua]n kh` %nio

&" ]`fqopiajp _kop0 ajpn]jpo 42485

• eva kb ajpn]jpo

)" atep n]pa _ka'_eajp 42494 atep n]pao ^u ]ca0 DFT `]p] )# atep n]pa _ka'_eajp 425<4 atep n]pao ^u ]ca0 DFT `]p] . ajpnu _kop 42>74 ajpnu _kopo A 424<7 IFR # ia]oqna kb ^qoejaoo kllknpqjepeao 424>4 > A 50 jkni]hev]pekj

• ah]ope_epu ej ajpnu bqj_pekj

42944

• p`-ajpnu.3op`-u.

++ UHR sa`ca0 lanoeopaj_a 42=59 ,+ UHR sa`ca0 op]j`]n` `are]pekj 42455 +2 ]`fqopiajp _kop sa`ca0 lanoeopaj_a 42977 ,2 ]`fqopiajp _kop sa`ca0 op]j`]n` `are]pekj 524== +6 h]^kn sa`ca0 lanoeopaj_a 429>9 ,6 h]^kn sa`ca0 op]j`]n` `are]pekj 42466 !9 napqnjo pk o_]ha ]ran]ca oeva ej DFT oeva _h]ooao 42>5; 42>8= 42>9> 42>;< 42><6 42><; 42><> 42>=6 42>>> @9 A #

'$ ?$

$#!& lnk^]^ehepu kb op]npejc ql ] pula K %ni %ni od]nao ej DFT oeva _h]ooao 42<>> 42895 426<6 42597 424=< 42495 42474 4245= 42445

Sedl´ aˇ cek, Sterk (Bonn, UCL) Recession Scars DNB, October 2013 41 / 56

Results

Sedl´ aˇ cek, Sterk (Bonn, UCL) Recession Scars DNB, October 2013 42 / 56

Steady state: Firm size by type and age

10 20 30 40 50 20 40 60 80 100 120 140 160 age type 1 (alpha=0.9305) type 2 (alpha=0.9573) type 3 (alpha=0.9664) type 4 (alpha=0.9728) type 5 (alpha=0.9770) type 6 (alpha=0.9801) 10 20 30 40 50 1000 2000 3000 4000 5000 6000 7000 8000 age type 7 (alpha=0.9827) type 8 (alpha=0.9849) type 9 (alpha=0.9993)

Sedl´ aˇ cek, Sterk (Bonn, UCL) Recession Scars DNB, October 2013 43 / 56

Steady state: Fraction of cohort-level employment by type and age

10 20 30 40 50 10 20 30 40 50 60 70 80 90 100 age type 1 (=0.9305) type 2 (=0.9573) type 3 (=0.9664) type 4 (=0.9728) type 5 (=0.9770) type 6 (=0.9801) type 7 (=0.9827) type 8 (=0.9849) type 9 (=0.9993)

Sedl´ aˇ cek, Sterk (Bonn, UCL) Recession Scars DNB, October 2013 44 / 56

Shock estimation: Historical decomposition

1980 1985 1990 1995 2000 2005 2010

• 0.1
• 0.05

0.05 0.1

• utput

TFP wedge TFP + Labor wedge TFP + Labor wedge +Adjustment cost wedge (= data) 1980 1985 1990 1995 2000 2005 2010

• 0.04
• 0.02

0.02 0.04 employment 1980 1985 1990 1995 2000 2005 2010

• 0.1

0.1 0.2 0.3 entrant size

Sedl´ aˇ cek, Sterk (Bonn, UCL) Recession Scars DNB, October 2013 45 / 56

Counterfactuals

model matches observed aggregate output and employment by construction take estimated shocks run them through a model in which we fix the type-composition of entrants general equilibrium effects are preserved

Sedl´ aˇ cek, Sterk (Bonn, UCL) Recession Scars DNB, October 2013 46 / 56

Counterfactuals

Figure: Output and employment differentials

1980 1985 1990 1995 2000 2005 2010

• 0.4
• 0.3
• 0.2
• 0.1

0.1 0.2 0.3 differentials deviation from data (p.p.)

• utput

employment

Sedl´ aˇ cek, Sterk (Bonn, UCL) Recession Scars DNB, October 2013 47 / 56

Counterfactuals

redo the same exercise now also fix adjustment cost shock to 1 for young firms

◮ fix composition of startups at steady state, but let the number of

entrants adjust

◮ free young firms from adjustment cost fluctuations, but let growth

rates respond to aggregate productivity and labor-leisure shocks

◮ i.e. a less restrictive version of empirical counterfactuals Sedl´ aˇ cek, Sterk (Bonn, UCL) Recession Scars DNB, October 2013 48 / 56

Counterfactuals

Figure: Output and employment differentials

• ‐

‐

‐

• time to talk quickly?

1 of 1 10/11/2013 5:47 PM

Sedl´ aˇ cek, Sterk (Bonn, UCL) Recession Scars DNB, October 2013 49 / 56

Persistence - “recession scar”

Sedl´ aˇ cek, Sterk (Bonn, UCL) Recession Scars DNB, October 2013 50 / 56

Persistence - “recession scar”

2005 2010 2015 2020 7.7 7.8 7.9 8 8.1 8.2 8.3 8.4 8.5 8.6 x 10

• 3 entrant share (high/low alpha)

benchmark fixed composition 2005 2010 2015 2020

• 1.5
• 1
• 0.5

%

• utput

Q 2005 2010 2015 2020

• 0.8
• 0.75
• 0.7
• 0.65
• 0.6
• 0.55
• 0.5
• 0.45
• 0.4

% labor productivity 2005 2010 2015 2020

• 1
• 0.8
• 0.6
• 0.4
• 0.2

0.2 0.4 % employment

Sedl´ aˇ cek, Sterk (Bonn, UCL) Recession Scars DNB, October 2013 50 / 56

Conclusions

fluctuations in composition of firm entrant cohorts important for aggregate outcomes smaller firms born in recessions, effects on output very persistent future work:

◮ analyze micro data underlying BDS ◮ endogenize wedges; more detailed explanation of drivers behind

• bserved cyclical patterns

Sedl´ aˇ cek, Sterk (Bonn, UCL) Recession Scars DNB, October 2013 51 / 56

Thanks

Sedl´ aˇ cek, Sterk (Bonn, UCL) Recession Scars DNB, October 2013 52 / 56

Possible explanations: sectoral composition?

sectoral composition of entrants?

Sedl´ aˇ cek, Sterk (Bonn, UCL) Recession Scars DNB, October 2013 53 / 56

Possible explanations: sectoral composition?

sectoral composition of entrants? manufacturing firms are on average larger → if also more sensitive to the BC → relatively less manufacturing firms in recessions

Sedl´ aˇ cek, Sterk (Bonn, UCL) Recession Scars DNB, October 2013 53 / 56

Possible explanations: sectoral composition?

1980 1985 1990 1995 2000 2005 5.2 5.4 5.6 5.8 6 6.2 6.4 6.6 6.8 7 year employment (millions) Sectoral composition actual within sector variation only between sector variation only

Sedl´ aˇ cek, Sterk (Bonn, UCL) Recession Scars DNB, October 2013 54 / 56

Possible explanations: necessity entrepreneurs?

“necessity entrepreneurs”: no ambitions to create jobs

Sedl´ aˇ cek, Sterk (Bonn, UCL) Recession Scars DNB, October 2013 55 / 56

Possible explanations: necessity entrepreneurs?

“necessity entrepreneurs”: no ambitions to create jobs if entry of necessity entrepreneurs is counter-cyclical → relatively more small firms in recessions

Sedl´ aˇ cek, Sterk (Bonn, UCL) Recession Scars DNB, October 2013 55 / 56

Possible explanations: necessity entrepreneurs?

1980 1985 1990 1995 2000 2005 1.8 2 2.2 2.4 2.6 2.8 3 3.2 3.4 3.6 year employment(millions) Necessity entrepreneurs actual variation in large firms only variation in small firms only

back Sedl´ aˇ cek, Sterk (Bonn, UCL) Recession Scars DNB, October 2013 56 / 56