from population growth to firm demographics, slides · concentration average firm size exit rate...
TRANSCRIPT
From Population Growth to Firm Demographics:Implications for Concentration, Entrepreneurship and the
Labor Share
Hugo Hopenhayn Julian Neira Rish SinghaniaUCLA University of Exeter University of Exeter
May 17, 2019
1 / 43
Motivation
I Puzzling aggregate trends in the US since 1980s
I Decline in the firm entry rate (14% to 8%)
I Decline in firm exit rate (9.7% to 7.7%)
I Increase in average firm size (20 to 24 employees)
I Increase in (employment) concentration (51% to 58%)
I Decline in the (corporate) labor share (66% to 60%)
I What explains this?I We look at population growth + firm demographics
2 / 43
Motivation
I Puzzling aggregate trends in the US since 1980s
I Decline in the firm entry rate (14% to 8%)
I Decline in firm exit rate (9.7% to 7.7%)
I Increase in average firm size (20 to 24 employees)
I Increase in (employment) concentration (51% to 58%)
I Decline in the (corporate) labor share (66% to 60%)
I What explains this?I We look at population growth + firm demographics
2 / 43
Outline
I Why population growth + firm demographics?I Document new factsI Feedback effects
I Theory
I Accounting
I Calibration
I Results
3 / 43
Overview of Results
I Reallocation across firm-age groups accounts forI ConcentrationI Average firm sizeI Exit ratesI Labor share
I Declining entry rates generate the reallocation
I Declining population growth lowers entry rate
I Feedback from firm demographics to entry is 2/3 of the effect
4 / 43
Motivating Evidence
5 / 43
52%
54%
56%
58%
20
21
22
23
7%
8%
9%
10%
1980 1990 2000 2010 1980 1990 2000 2010 1980 1990 2000 2010
Concentration Average Firm Size Exit Rate
Source: BDS
0%
10%
20%
30%
40%
50%
1980 1990 2000 2010
Concentration
2.0
2.5
3.0
3.5
1980 1990 2000 2010
Log Average Firm Size
5%
10%
15%
20%
25%
1980 1990 2000 2010
Exit Rate
Firm agea) 0 b) 1 c) 2 d) 3 e) 4
f) 5 g) 6 to 10 h) 11 to 15 i) 16 to 20 j) 21 to 25
6 / 43
52%
54%
56%
58%
20
21
22
23
7%
8%
9%
10%
1980 1990 2000 2010 1980 1990 2000 2010 1980 1990 2000 2010
Concentration Average Firm Size Exit Rate
Source: BDS
0%
10%
20%
30%
40%
50%
1980 1990 2000 2010
Concentration
2.0
2.5
3.0
3.5
1980 1990 2000 2010
Log Average Firm Size
5%
10%
15%
20%
25%
1980 1990 2000 2010
Exit Rate
Firm agea) 0 b) 1 c) 2 d) 3 e) 4
f) 5 g) 6 to 10 h) 11 to 15 i) 16 to 20 j) 21 to 256 / 43
Firms are Aging
1990 1995 2000 2005 2010 201530%
35%
40%
45%
50%
55%
1975 1980 1985 1990 1995 2000 2005 2010 20157%
8%
9%
10%
11%
12%
13%
14%
15%
7 / 43
Firms are Aging
1990 1995 2000 2005 2010 201530%
35%
40%
45%
50%
55%
1975 1980 1985 1990 1995 2000 2005 2010 20157%
8%
9%
10%
11%
12%
13%
14%
15%
7 / 43
Firms are Aging
1990 1995 2000 2005 2010 201530%
35%
40%
45%
50%
55%
1975 1980 1985 1990 1995 2000 2005 2010 20157%
8%
9%
10%
11%
12%
13%
14%
15%
8 / 43
∆Population Growth → ∆Entry Rates
I Average firm size:et ≡ Nt/Mt
I Growth Decomposition:
M = N − e (1)
I Growth in the number of firms is entry rate minus exit rate
M = λ− ξ (2)
I Combining (1) and (2):
λ = N − e+ ξ
9 / 43
∆Population Growth → ∆Entry Rates
I Average firm size:et ≡ Nt/Mt
I Growth Decomposition:
M = N − e (1)
I Growth in the number of firms is entry rate minus exit rate
M = λ− ξ (2)
I Combining (1) and (2):
λ = N − e+ ξ
9 / 43
∆Population Growth → ∆Entry Rates
I Average firm size:et ≡ Nt/Mt
I Growth Decomposition:
M = N − e (1)
I Growth in the number of firms is entry rate minus exit rate
M = λ− ξ (2)
I Combining (1) and (2):
λ = N − e+ ξ
9 / 43
∆Population Growth → ∆Entry Rates
I Average firm size:et ≡ Nt/Mt
I Growth Decomposition:
M = N − e (1)
I Growth in the number of firms is entry rate minus exit rate
M = λ− ξ (2)
I Combining (1) and (2):
λ = N − e+ ξ
9 / 43
The Rise and Fall of Population Growth
1940s 1950s 1960s 1970s 1980s 1990s 2000s 2010s0.0%
0.5%
1.0%
1.5%
2.0%
2.5%
3.0%
10 / 43
Is This Driving Force Enough?
λ = N − e︸︷︷︸0
+ ξ︸︷︷︸12%
1940 1950 1960 1970 1980 1990 2000 2010 20202%
4%
6%
8%
10%
12%
14%
16%
18%
20%
11 / 43
Is This Driving Force Enough?
I Qualitatively yes, quantitatively no.
I Cannot explain movements in exit rate
I Cannot explain increase in average size
I In the data
∆λ︸︷︷︸6%
= ∆N︸︷︷︸2%
− ∆e︸︷︷︸2%
+ ∆ξ︸︷︷︸2%
12 / 43
Is This Driving Force Enough?
I Qualitatively yes, quantitatively no.
I Cannot explain movements in exit rate
I Cannot explain increase in average size
I In the data
∆λ︸︷︷︸6%
= ∆N︸︷︷︸2%
− ∆e︸︷︷︸2%
+ ∆ξ︸︷︷︸2%
12 / 43
Firm Demographics
13 / 43
∆Population Growth + Firm Demographics
I Need to account for firm demographics
I Important feedback effects
λ = N − e+ ξ
I ∆ entry rates → ∆ age distributionI This affects average firm sizeI Also affects average exit rates
I Decrease in population growth implies:I Decline in entry rateI Aging of firms
14 / 43
∆Population Growth + Firm Demographics
I Need to account for firm demographics
I Important feedback effects
λ = N − e+ ξ
I ∆ entry rates → ∆ age distributionI This affects average firm sizeI Also affects average exit rates
I Decrease in population growth implies:I Decline in entry rateI Aging of firms
14 / 43
Theory
15 / 43
Environment
I Common discount factor β
I Fixed endowment of a resource (labor) Nt inelasticallysupplied. Numeraire.
I Firm’s idiosyncratic state s
I st ∼ F (st+1|st). Persistence.
I Revenue function R (s, n, Z)
I Aggregate summary state Z
I Employment function n (s, Z)
I Profit function π (s, Z)
I Both strictly increasing
I Accomodates perfect competition and variable markups
16 / 43
Environment
I Common discount factor β
I Fixed endowment of a resource (labor) Nt inelasticallysupplied. Numeraire.
I Firm’s idiosyncratic state s
I st ∼ F (st+1|st). Persistence.
I Revenue function R (s, n, Z)
I Aggregate summary state Z
I Employment function n (s, Z)
I Profit function π (s, Z)
I Both strictly increasing
I Accomodates perfect competition and variable markups
16 / 43
Environment
I Common discount factor β
I Fixed endowment of a resource (labor) Nt inelasticallysupplied. Numeraire.
I Firm’s idiosyncratic state s
I st ∼ F (st+1|st). Persistence.
I Revenue function R (s, n, Z)
I Aggregate summary state Z
I Employment function n (s, Z)
I Profit function π (s, Z)
I Both strictly increasing
I Accomodates perfect competition and variable markups
16 / 43
Equilibrium: Definition
An equilibrium for a given sequence {Nt} and given initial measureµ0 are sequences {s∗t ,mt, µt, Zt}
1. Exit: Optimal exit condition.
2. Entry: No rents for entrants
3. Resource constraint holds
17 / 43
Equilibrium: Analysis
I Guess Zt = Z∗ for all t, where ve (Z∗) = 0
I Exit rates, average firm size, and size distributions by cohortsare time invariant.
18 / 43
Equilibrium: Analysis
I Guess Zt = Z∗ for all t, where ve (Z∗) = 0
I Exit rates, average firm size, and size distributions by cohortsare time invariant.
18 / 43
Resource Constraint
I Define firm demographic variables:I Sa : Probability an entrant survives at least a periodsI ea : Average size of cohort of age a
I Resource constraint
Nt = mte0 + EIt
I Mass of entrants:
mt =Nt − EIt
e0
19 / 43
Resource Constraint
I Define firm demographic variables:I Sa : Probability an entrant survives at least a periodsI ea : Average size of cohort of age a
I Resource constraint
Nt = mte0 + EIt
I Mass of entrants:
mt =Nt − EIt
e0
19 / 43
Resource Constraint
I Define firm demographic variables:I Sa : Probability an entrant survives at least a periodsI ea : Average size of cohort of age a
I Resource constraint
Nt = mte0 + EIt
I Mass of entrants:
mt =Nt − EIt
e0
19 / 43
Dynamic Entry Equation
mt =Nt − EIt
e0
I Employment by incumbents depends on firm demographics
EIt =
∞∑a=1
mt−aSaea
I History dependence: Current entry depends on past entry
mt =Nt −
∑∞a=1mt−aSaeae0
I Feedback of firm demographics on entry
20 / 43
From Population Growth to Entry
1. Long run effects
I e = 0I Population growth g affects share of age cohorts:
Lower growth implies lower exit rates
2. Adjustment Path
I Change in g implies changes in average size
I e 6= 0 in the transition
21 / 43
From Population Growth to Entry
1. Long run effects
I e = 0I Population growth g affects share of age cohorts:
Lower growth implies lower exit rates
2. Adjustment Path
I Change in g implies changes in average size
I e 6= 0 in the transition
21 / 43
Accounting
22 / 43
Accounting Exercise
I Composition effect due to changes in Nt
I Take firm demographics Sa and ea from data
I Feed Nt into dynamic entry equation
mt =Nt −
∑∞a=1mt−aSaeae0
I Remain agnostic about the underlying model
23 / 43
1980 1985 1990 1995 2000 2005 20100%
2%
4%
6%
8%
10%
12%
14%
16%
18%
1980 1985 1990 1995 2000 2005 201018
20
22
24
26
28
30
1980 1985 1990 1995 2000 2005 20106%
7%
8%
9%
10%
11%
12%
1980 1985 1990 1995 2000 2005 201050%
51%
52%
53%
54%
55%
56%
57%
58%
24 / 43
1980 1985 1990 1995 2000 2005 20100%
2%
4%
6%
8%
10%
12%
14%
16%
18%
1980 1985 1990 1995 2000 2005 201018
20
22
24
26
28
30
1980 1985 1990 1995 2000 2005 20106%
7%
8%
9%
10%
11%
12%
1980 1985 1990 1995 2000 2005 201050%
51%
52%
53%
54%
55%
56%
57%
58%
24 / 43
I Extrapolation necessary due to data limitations Extrapolation
I Do not observe ea and Sa for older firms (age > 25)
I Match time-series of average firm size and exit rates ofleft-censored firms Match
I Do not observe µ0 (age distribution in 1940)
I Match time-series of employment weight of left-censoredfirms Match
I Alternative to extrapolation: calibrate a structural model
25 / 43
I Extrapolation necessary due to data limitations Extrapolation
I Do not observe ea and Sa for older firms (age > 25)
I Match time-series of average firm size and exit rates ofleft-censored firms Match
I Do not observe µ0 (age distribution in 1940)
I Match time-series of employment weight of left-censoredfirms Match
I Alternative to extrapolation: calibrate a structural model
25 / 43
I Extrapolation necessary due to data limitations Extrapolation
I Do not observe ea and Sa for older firms (age > 25)
I Match time-series of average firm size and exit rates ofleft-censored firms Match
I Do not observe µ0 (age distribution in 1940)
I Match time-series of employment weight of left-censoredfirms Match
I Alternative to extrapolation: calibrate a structural model
25 / 43
Calibration
26 / 43
Calibration strategy
I Pick a model: perfect competitionI Homogeneous goodI Aggregate state Z equals market price p
I Assume economy in balanced growth path in 1939
I Feed labor force growth rate from 1940 to 2014
I Calibrate to (mostly) 1978 moments
I Look at non-targeted moments
27 / 43
Functional Forms
I Production function
f(s, n) = snα; α < 1
I Log-productivity follows AR(1)
log(st+1) = µs + ρ log(st) + εt+1; εt+1 ∼ N (0, σ2ε)
I Startups draw productivity from
G ∼ logN(s0, σ
20
)I Overhead labor is increasing in firm size
cf = cfa + cfb × h(s)
28 / 43
Parameter Values
Non-targeted moments on Firm Dynamics?
29 / 43
Parameter Values
Non-targeted moments on Firm Dynamics?29 / 43
Exit, Size, and Concentration by Firm Age
30 / 43
Entry Rate
1940 1950 1960 1970 1980 1990 2000 2010 20202%
4%
6%
8%
10%
12%
14%
16%
18%
20%
31 / 43
Non-targeted moments
1980 2000 202018
20
22
24
26
28
30
1980 2000 20207.0%
7.5%
8.0%
8.5%
9.0%
9.5%
10.0%
10.5%
1980 2000 202050%
51%
52%
53%
54%
55%
56%
57%
58%
59%
32 / 43
Counterfactuals
1940 1950 1960 1970 1980 1990 2000 2010 20204%
6%
8%
10%
12%
14%
16%
33 / 43
Counterfactuals
1940 1950 1960 1970 1980 1990 2000 2010 2020-10%
-8%
-6%
-4%
-2%
0%
2%
34 / 43
Projections
1940 1960 1980 2000 2020 2040 2060
0%
2%
4%
6%
8%
10%
12%
14%
16%
35 / 43
Labor share: Autor et al / Kehrig (2019) + Aging
Labor sharei︸ ︷︷ ︸Declining in firm size
=Comp. to production labori
Value addedi︸ ︷︷ ︸Same across firms
+Comp. to overhead labori
Value addedi︸ ︷︷ ︸Declining in firm size
1940 1950 1960 1970 1980 1990 2000 2010 2020-9%
-8%
-7%
-6%
-5%
-4%
-3%
-2%
-1%
0%
1%
Karabarbounis and Neiman (2014)Koh, Santaeulalia-Llopis and Zheng (2018)Model
36 / 43
Labor share: Autor et al / Kehrig (2019) + Aging
Labor sharei
︸ ︷︷ ︸Declining in firm size
=Comp. to production labori
Value addedi
︸ ︷︷ ︸Same across firms
+Comp. to overhead labori
Value addedi
︸ ︷︷ ︸Declining in firm size
1940 1950 1960 1970 1980 1990 2000 2010 2020-9%
-8%
-7%
-6%
-5%
-4%
-3%
-2%
-1%
0%
1%
Karabarbounis and Neiman (2014)Koh, Santaeulalia-Llopis and Zheng (2018)Model
36 / 43
Labor share: Autor et al / Kehrig (2019) + Aging
Labor sharei
︸ ︷︷ ︸Declining in firm size
=Comp. to production labori
Value addedi︸ ︷︷ ︸Same across firms
+Comp. to overhead labori
Value addedi
︸ ︷︷ ︸Declining in firm size
1940 1950 1960 1970 1980 1990 2000 2010 2020-9%
-8%
-7%
-6%
-5%
-4%
-3%
-2%
-1%
0%
1%
Karabarbounis and Neiman (2014)Koh, Santaeulalia-Llopis and Zheng (2018)Model
36 / 43
Labor share: Autor et al / Kehrig (2019) + Aging
Labor sharei
︸ ︷︷ ︸Declining in firm size
=Comp. to production labori
Value addedi︸ ︷︷ ︸Same across firms
+Comp. to overhead labori
Value addedi︸ ︷︷ ︸Declining in firm size
1940 1950 1960 1970 1980 1990 2000 2010 2020-9%
-8%
-7%
-6%
-5%
-4%
-3%
-2%
-1%
0%
1%
Karabarbounis and Neiman (2014)Koh, Santaeulalia-Llopis and Zheng (2018)Model
36 / 43
Labor share: Autor et al / Kehrig (2019) + Aging
Labor sharei︸ ︷︷ ︸Declining in firm size
=Comp. to production labori
Value addedi︸ ︷︷ ︸Same across firms
+Comp. to overhead labori
Value addedi︸ ︷︷ ︸Declining in firm size
1940 1950 1960 1970 1980 1990 2000 2010 2020-9%
-8%
-7%
-6%
-5%
-4%
-3%
-2%
-1%
0%
1%
Karabarbounis and Neiman (2014)Koh, Santaeulalia-Llopis and Zheng (2018)Model
36 / 43
Labor share: Autor et al / Kehrig (2019) + Aging
Labor sharei︸ ︷︷ ︸Declining in firm size
=Comp. to production labori
Value addedi︸ ︷︷ ︸Same across firms
+Comp. to overhead labori
Value addedi︸ ︷︷ ︸Declining in firm size
1940 1950 1960 1970 1980 1990 2000 2010 2020-9%
-8%
-7%
-6%
-5%
-4%
-3%
-2%
-1%
0%
1%
Karabarbounis and Neiman (2014)Koh, Santaeulalia-Llopis and Zheng (2018)Model
36 / 43
Discussion
Labor force and labor supply
Job Creation and Destruction
TFP
CONCLUSION
37 / 43
Alternative Measures of Labor Supply Back
1947-54 1955-64 1965-74 1975-84 1985-94 1995-04 2005-140.0%
0.5%
1.0%
1.5%
2.0%
2.5%
3.0%
3.5%Labor ForceTotal EmploymentPrivate Sector EmploymentNon-Manufacturing Employment
38 / 43
Labor Force Growth Decomposition Back
LFt = Civilian Noninstitutional Population Age 16 And Overt× Participation Ratet
LF Growth Ratet = CNP16 Growth Ratet + PR Growth Ratet
CNP16 Growth Ratet = Birth Ratet−16 + Other(Migration, Death, Instit)t
39 / 43
Labor Force Growth Decomposition Back
LFt = Civilian Noninstitutional Population Age 16 And Overt× Participation Ratet
LF Growth Ratet = CNP16 Growth Ratet + PR Growth Ratet
CNP16 Growth Ratet = Birth Ratet−16 + Other(Migration, Death, Instit)t
39 / 43
Labor Force Growth Decomposition Back
LFt = Civilian Noninstitutional Population Age 16 And Overt× Participation Ratet
LF Growth Ratet = CNP16 Growth Ratet + PR Growth Ratet
CNP16 Growth Ratet = Birth Ratet−16 + Other(Migration, Death, Instit)t
39 / 43
Labor Force Growth Decomposition Back
LFt = Civilian Noninstitutional Population Age 16 And Overt× Participation Ratet
LF Growth Ratet = CNP16 Growth Ratet + PR Growth Ratet
CNP16 Growth Ratet = Birth Ratet−16 + Other(Migration, Death, Instit)t
39 / 43
Participation Rate By Gender Back
1940 1950 1960 1970 1980 1990 2000 2010 202030%
40%
50%
60%
70%
80%
90%
40 / 43
Job Reallocation: Accounting Approach Back
1980 2000 202022%
24%
26%
28%
30%
32%
34%
36%DataComposition
1980 2000 202011%
12%
13%
14%
15%
16%
17%
18%
19%
20%
21%DataComposition
1980 2000 202011%
12%
13%
14%
15%
16%
17%
18%
19%DataComposition
41 / 43
TFP Growth by Decade Back
I Aggregate production function is Y = AM1−αNα.
I TFP is A =[∫s1/(1−α)dµ(s)
]1−αI Measured TFP is AM1−α.
42 / 43
Conclusions
I Unified quantitative explanation for long-term changes inI Entry rateI Exit ratesI Average firm sizeI ConcentrationI Labor Share
I Population growth as driving force
I Importance of firm demographics
I Interplay of population and firm demographics explains a largepart of these facts
43 / 43
25%
30%
35%
12.5%
15%
17.5%
20%
11%
13%
15%
17%
1980 1990 2000 2010 1980 1990 2000 2010 1980 1990 2000 2010
Reallocation Rate Job Creation Rate Job Destruction Rate
20%
30%
40%
50%
60%
1980 1990 2000 2010
Reallocation Rate
10%
20%
30%
1980 1990 2000 2010
Job Creation Rate
10%
20%
30%
1980 1990 2000 2010
Job Destruction Rate
Firm ageb) 1 c) 2 d) 3 e) 4 f) 5
g) 6 to 10 h) 11 to 15 i) 16 to 20 j) 21 to 25
Back
Implied interest rate with log utility Back
1980 1985 1990 1995 2000 2005 20103.0%
3.5%
4.0%
4.5%
5.0%
5.5%
6.0%
Regression of reallocation rate on firm age Back
Regression of job creation rate on firm age Back
Regression of job destruction rate on firm age Back
Regression of log average firm size on firm age Back
Regression of exit rate on firm age Back
Regression of concentration on firm age Back
Extrapolation Back
0 10 20 30 40 50 600
50
100
150
0 10 20 30 40 50 600%
5%
10%
15%
20%
25%
0 10 20 30 40 50 600%
20%
40%
60%
80%
0 20 40 60 80 1000%
20%
40%
60%
80%
100%
Accounting Exercise: Left-Censored Match Back
1975 1980 1985 1990 1995 2000 2005 2010 201520
40
60
80
100
120
140DataAccounting
1975 1980 1985 1990 1995 2000 2005 2010 20150.03
0.04
0.05
0.06
0.07
0.08
0.09
0.1
0.11
0.12
0.13DataAccounting
1975 1980 1985 1990 1995 2000 2005 2010 201555%
60%
65%
70%
75%
80%
85%DataAccounting
1975 1980 1985 1990 1995 2000 2005 2010 201540%
50%
60%
70%
80%
90%
100%DataAccounting
Accounting Exercise: Robustness checks Back
0 50 1000%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
BGPFitted
0 50 1000%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
BGPFitted
1980 1990 2000 20100.75
0.8
0.85
0.9
0.95
1
1.05DataAccounting
Distributional Moments Match: Competitive Model
1990 2000 2010 202030%
35%
40%
45%
50%
55%
1990 2000 2010 202066%
68%
70%
72%
74%
76%
78%
80%
82%
84%
86%
1980 2000 20201.0%
1.5%
2.0%
2.5%
3.0%
3.5%
4.0%
4.5%
Age-Size Distribution Match: Competitive Model
1990 1995 2000 2005 2010 20150
0.05
0.1
0.15
0.2
0.25
0.3
1990 1995 2000 2005 2010 20150.2
0.25
0.3
0.35
0.4
0.45
0.5
0.55
0.6
I Model: Dashed Line
I Data: Solid Line
Establishments
1980 1990 2000 20107%
8%
9%
10%
11%
12%
13%
14%
1990 1995 2000 2005 2010 201530%
32%
34%
36%
38%
40%
42%
44%
46%
48%
Establishments
16.0
16.5
17.0
17.5
18.0
8%
10%
12%
14%
1980 1990 2000 2010 1980 1990 2000 2010
Average establishment size Exit rate
2.0
2.4
2.8
1980 1990 2000 2010
Log Avg. Estab. Size
5%
10%
15%
20%
25%
1980 1990 2000 2010
Exit Rate
Establishment agea) 0 b) 1 c) 2 d) 3 e) 4
f) 5 g) 6 to 10 h) 11 to 15 i) 16 to 20 j) 21 to 25
0
50,000
100,000
150,000
200,000
250,000
300,000
350,000
400,000
450,000
b) 1 c) 2 d) 3 e) 4 f) 5 g) 6-10 h) 11-15 i) 16-20 j) 21-25
Employment Destroyed by Exit(adjusted per year)
Statistics including firms age > 25
0%
10%
20%
30%
40%
50%
60%
70%
80%
1980 1990 2000 2010
Concentration
2.0
2.5
3.0
3.5
4.0
4.5
1980 1990 2000 2010
Log Average Firm Size
5%
10%
15%
20%
25%
1980 1990 2000 2010
Exit Rate
Firm agea) 0 b) 1 c) 2 d) 3 e) 4 f) 5
g) 6 to 10 h) 11 to 15 i) 16 to 20 j) 21 to 25 k) Above 25