from good to bad concentration? u.s. industries over the ... · introduction theory measurement...

Introduction Theory Measurement Predictions Tests Conclusion Appendix References

From Good to Bad Concentration?

U.S. Industries over the past 30 years

Matias Covarrubias, Germán Gutiérrez

and Thomas Philippon

NYU, NBER, CEPR

April 2019, NBER Macro Annual


Stylized Fact: Concentration

02

46

8

1980 1990 2000 2010

Manufacturing Non−Manufacturing

Panel A. Cumulative Change in CR8 (%)

Source: U.S. Economic Census for all Businesses. Dashed lines because of changes in industry classification fromSIC to NAICS.


Stylized Fact: After-Tax Net Profit Margins

.04

.06

.08

.1.1

2

1970q1 1980q1 1990q1 2000q1 2010q1 2020q1

Panel B. Profits/VA

Source: FRED Non Financial Corporates.


Stylized Fact: Labor Share

.56

.58

.6.6

2.6

4.6

6

1970q1 1980q1 1990q1 2000q1 2010q1 2020q1

Panel C. Labor Share

Source: FRED Non Financial Corporates.


US vs EU: Concentration

.25

.3.3

5.4

.45

2000 2005 2010 2015

NA EU

CR8 Level (OECD 2019)

Source: Bajgar et al. (2019), OECD STAN


US vs EU & Asia: Profit Rates

.17

.18

.19

.2.2

1.2

2

1990 1995 2000 2005 2010 2015

USA EU28 EU15 JPN+KOR

GOS/PROD

Source: OECD STAN. Pre-Tax Gross Operating Surplus over Revenues.


US vs EU: Labor Shares

Source: EU KLEMS 2017. See also Cette et al. (2019).


Theory: Good vs Bad Concentration

• Ex-post profits: π = µ1+µ

(

aA

)σ−1 PYN −φ

• Free entry:E[π]r+δ ≤ κ

• Number of firms (symmetric)

N =µ

1+µ

PY

(r +δ )κ +φ

- Bad concentration: barriers to entry κ , regulatory capture

- Good concentration: efficient response to TFP, competition

• Selection effect a∗ (σ)ր

(r +δ )κ = (1−F (a∗))×E [π | a > a∗] .


China Shock: N, Import-Adj. CR8.6

.81

1.2

1.4

# o

f firm

s (

1991=

1)

1985 1990 1995 2000 2005 2010 2015

Low IE High IE

Number of firms by Chinese Exposure (1991 = 1)

02

46

8W

td. A

vg. C

hange in C

R8 fro

m 1

997 (

%)

1995 2000 2005 2010 2015

Raw Import adjusted

Raw and Import−Adjusted Concentration

Notes: Compustat, NBER-CES and Peter Schott’s data.


China Shock: Evaluating Measures−

4−

3−

2−

10

12

1992 1994 1996 1998 2000 2002 2004 2006

Response of log(OIADP) to China Shock

−.5

0.5

11.5

21992 1994 1996 1998 2000 2002 2004 2006

SALE/COGS

Notes: Compustat, NBER-CES and Peter Schott’s import data. Plots show βt from regressions

yi ,j ,t = βt ×NTR Gapj +δi + γt + εi ,j ,t .


SALE/COGS vs Gross Profit Rates

.8.9

11

.11

.21

.3

1980 1990 2000 2010 2020

USA

.8.9

11

.11

.2

1980 1990 2000 2010 2020

EU15.8

11

.21

.4

1980 1990 2000 2010 2020

GBR

.91

1.1

1.2

1.3

1980 1990 2000 2010 2020

JPN

SALE/COGS GOS/PROD

Sources: OECD Stan and Compustat.


Testable Prediction: Dynamics of Market Shares

• Market shares

si ,j ,t =hi ,j ,t

Nj

(

(

1+µj

)

ai ,j ,t(

1+µi ,j

)

Aj ,t

)σj−1

• Prop. All else equal, an increase in σ leads to an increase

in the volatility of market shares.

Σ2logs =Σ2

logh +(

σj −1)2

Σ2loga


Testable Predictions

Theories

Good σ Bad κ

Turnover / Exit + –Investment / K Growth + –

Corr(∆CR,∆TFP) + –Corr(∆CR,∆P) – +


Declining Turnover

.06

.08

.1.1

2.1

4A

vg

of

with

in in

du

str

y r

esh

uff

ling

1980 1990 2000 2010

By Sale By MV

Source: Compustat NA, following BEA industries. Measure is 1- 5-year ahead correlation of ranks within industry.


Investment: Growth in K vs. Q−

.06

−.0

4−

.02

0.0

2

1985 1995 2005 2015

All

−.0

4−

.02

0.0

2

1985 1995 2005 2015

Equipment−

.04

−.0

20

.02

1985 1995 2005 2015

Structures

−.1

−.0

50

.05

1985 1995 2005 2015

Intellectual Property

Cumulative gap Residual

Notes: Q for Non-Financial corporate sector from FRED; Capital stock from BEA


TFP, Prices and Markups

Source: BLS multifactor and Compustat.


σ vs κ? Principal Component Loadings

−.5

0.5

loadin

gs

cr4_

cen

cr4_

cp

inta

n_ks

hare

_bea

inta

n_ks

hare

_pt

impo

rt_sh

are

prof

it_m

argin_

bea

prof

it_m

argin_

med

_cp

ls_k

l

dtfp

_kl

Dlogu

lc_k

l

Dlogp

_kl

Dlogm

u_kl

lead

_tur

nove

r_m

v

ikga

p

PC 1: "Intangibles" PC 2: "Barriers to Entry"

Source: “cen” census; “cp” Compustat; “pt” Peters & Taylor; “kl” Klems.


Components over Time−

.4−

.20

.2.4

.6

1995 2000 2005 2010 2015year

PC1: "Intangibles" PC2: "Barriers to Entry"

Average scores for PC1 and PC2


Lobbying?

11.5

22.5

33.5

Tota

l Lobbyin

g E

xpenditure

s (

$ B

N U

SD

)

2000 2005 2010 2015Year

US − Total US − Business

EU − Total EU − Business

Source: Gutiérrez and Philippon (2018)


Four prominent explanations

• Rising Capital Share (α).

- Alexander and Eberly (2016), Crouzet and Eberly (2018),

Acemoglu and Restrepo (2017)

• Rising Elasticity (σ ).

- Autor et al. (2017)

• Increasing Returns to Scale (γ).

- Aghion et al. (2018)

• Rising Barriers to Competition (κ).

- Gutiérrez and Philippon (2018), Jones et al. (2018) ,Gutiérrez and Philippon (2019)


Summary of Theories and Measures

Theories

Data α σ γ κ

1. Intern.

Evidence

Common global

trends

– + + + –

2. Entry, Exit

and Turnover

Leader turnover – ? + – –

Elast. of Entry

to Q

– ? + + –

Exit Rate – ? + + –

3. Corr. of CR,

TFP and prices

Corr(∆CR,∆TFP) +..- ? + + –

Corr(∆CR,∆P) -..+ ? – – +

4. Inv. & profits

by leaders

Agg. inv. rate – + ? ? –

Leader inv. rate – – + + –

Leader profits + ? ? + +

5. Ret. to

ScaleEstimated RS* 0+ 0 – + ?


Airlines and Telecom

−.2

−.1

0.1

.2.3

CR

8

−.2

0.2

.4.6

Mark

−up

1990 2000 2010year

Mark−up CR8

Telecom

−.1

0.1

.2C

R8

−.4

−.2

0.2

.4M

ark

−up

1990 2000 2010year

Mark−up CR8

Airlines

Source: BLS multifactor and Compustat.


Turnover of Leaders

.2.2

5.3

.35

.4A

vg. of in

dustr

y−

level exit p

robabili

ties

1980 1990 2000 2010

By Sale By MV

Source: Compustat NA, following BEA industries.


Change in Profits - Within or Between Firms?

−.0

4−

.02

0.0

2.0

4W

ithin

−F

irm

Change in P

rofit M

arg

in

1980 1990 2000 2010

Average Lead

Source: Compustat NA. Regression: πi ,j ,t = βt ×Leadi ,j ,t +δi + γt + εjt . Average: γt . Leader: γt +βt .


Increasing Harm at High CR?

Source: U.S. Economic Census, NBER-CES and BEA


Why After 2000? Returns to Scale?

• Why is bad concentration more prevalent after 2000?

• Perhaps returns to scale have increased?

- Using the approach of Basu and Fernald (1997), we find amoderate increase from 0.78 to 0.80.

- De-Loecker et al. (2019) estimate increase from 1.03 to1.08 in US.Diez et al. (2018) reach similar conclusions

globally.

- There is no evidence that returns to scale have increasedsignificantly in recent years


Estimates0

.51

1.5

Non

dur_

pape

r

Non

dur_

appa

rel

Dur

_prim

_met

al

Non

dur_

printin

g

Non

dur_

plas

tic

Dur

_electric

al

Dur

_woo

d

Dur

_misc

Non

dur_

chem

ical

Dur

_mac

hine

ry

Dur

_fab

_met

al

Non

dur_

food

Dur

_non

met

al

Dur

_fur

nitu

re

Non

dur_

textile

Dur

_com

pute

r

Dur

_tra

nsp

RTS estimates, Manufacturing

0.5

11

.5

Hea

lth_a

mbu

lato

ry

Hea

lth_h

ospita

ls

Arts_r

ecre

ation

Was

te_m

gmt

Lega

l_se

rv

Educa

tiona

l

Arts_p

erfo

rming

Adm_s

uppo

rt

Misc_

serv

Com

pute

r_se

rv

Hea

lth_s

ocial

Man

agem

ent

Oth

er_e

x_go

v

Acc_a

ccom

odat

ion

Acc_f

ood

RTS estimates, Services

0.5

11

.5

Inf_

data

Trans

p_air

Inf_

mot

ion

Inf_

teleco

m

Utilities

Trans

p_ot

her

Insu

ranc

e

Inf_

publish

Finan

ce_s

ecur

ities

Finan

ce_b

anks

Finan

ce_f

unds

Trans

p_pa

ssen

ger

Trans

p_stor

age

Con

stru

ction

Ren

tal_leas

ing

Trans

p_ra

il

Ret

ail_tra

de

Who

lesa

le_t

rade

Trans

p_tru

ck

Rea

l_Est

ate

Trans

p_wat

er

RTS estimates, Others

Source: BLS multifactor


First Principal Component: Intangibles

−4

−2

02

46

Score

(2012)

Tra

nsp

_p

ipe

line

Utilit

ies

Acc_

foo

dT

ran

sp

_tr

uck

Wa

ste

_m

gm

tA

cc_

acco

mo

da

tio

nF

ina

nce

_b

an

ks

He

alth

_h

osp

ita

lsT

ran

sp

_w

ate

rIn

su

ran

ce

Tra

nsp

_o

the

rA

rts_

recre

atio

nA

dm

_su

pp

ort

No

nd

ur_

foo

dD

ur_

wo

od

No

nd

ur_

pa

pe

rN

on

du

r_p

rin

tin

gD

ur_

fab

_m

eta

lD

ur_

no

nm

eta

lR

eta

il_tr

ad

eN

on

du

r_p

lastic

Re

nta

l_le

asin

gW

ho

lesa

le_

tra

de

Tra

nsp

_a

irD

ur_

prim

_m

eta

lN

on

du

r_te

xtile

Du

r_fu

rnitu

reF

ina

nce

_se

cu

ritie

sIn

f_m

otio

nIn

f_te

leco

mD

ur_

mis

cD

ur_

ele

ctr

ica

lD

ur_

ma

ch

ine

ryIn

f_p

ub

lish

Du

r_tr

an

sp

Co

mp

ute

r_se

rvN

on

du

r_a

pp

are

lN

on

du

r_ch

em

ica

lIn

f_d

ata

Du

r_co

mp

ute

r

PC1: "Intangibles"


Second Principal Component: Barriers to Entry

−2

02

4S

core

(2012)

No

nd

ur_

ap

pa

rel

Acc_

foo

dD

ur_

wo

od

Du

r_fu

rnitu

reT

ran

sp

_tr

uck

Fin

an

ce

_se

cu

ritie

sN

on

du

r_te

xtile

Ad

m_

su

pp

ort

No

nd

ur_

prin

tin

gD

ur_

fab

_m

eta

lA

rts_

recre

atio

nD

ur_

no

nm

eta

lD

ur_

mis

cD

ur_

co

mp

ute

rC

om

pu

ter_

se

rvN

on

du

r_p

lastic

Du

r_e

lectr

ica

lA

cc_

acco

mo

da

tio

nD

ur_

ma

ch

ine

ryH

ea

lth

_h

osp

ita

lsW

ho

lesa

le_

tra

de

Du

r_p

rim

_m

eta

lR

eta

il_tr

ad

eIn

su

ran

ce

Utilit

ies

No

nd

ur_

foo

dN

on

du

r_p

ap

er

Du

r_tr

an

sp

Wa

ste

_m

gm

tT

ran

sp

_w

ate

rT

ran

sp

_p

ipe

line

Tra

nsp

_o

the

rIn

f_p

ub

lish

No

nd

ur_

ch

em

ica

lT

ran

sp

_a

irIn

f_d

ata

Re

nta

l_le

asin

gF

ina

nce

_b

an

ks

Inf_

mo

tio

nIn

f_te

leco

m

PC2: "Barriers to Entry"


PC2 scores (“Barriers to Entry”) vs Import Shares

Dur_computerDur_electrical

Dur_fab_metal

Dur_furniture

Dur_machinery

Dur_miscDur_nonmetal

Dur_prim_metal

Dur_transp

Dur_wood

Nondur_apparel

Nondur_chemical

Nondur_paper

Nondur_plastic

Nondur_textile

−2

−1

01

2P

C2

Sco

re (

20

12

)

0 1 2 3 4 5Import Share

Sources: BEA. Imports from Peter Schott’s data. Notes: Only manufacturing.


References I

Acemoglu, D. and P. Restrepo (2017). The race between man

and machine: Implications of technology for growth, factor

shares and employment. American Economic Review,

forthcoming.

Aghion, P., A. Bergeaud, T. Boppart, P. J. Klenow, and H. Li

(2018). A theory of falling growth and rising rents.

Alexander, L. and J. Eberly (2016). Investment hollowing out.

Working Paper .

Autor, D., D. Dorn, L. Katz, C. Patterson, and J. V. Reenen

(2017, jan). Concentrating on the fall of the labor share.

Technical report, mimeo.

Bajgar, M., G. Berlingieri, S. Calligaris, C. Criscuolo, and

J. Timmis (2019). Industry concentration in europe and north

america. Technical report, OECD.


References IIBasu, S. and J. G. Fernald (1997). Returns to scale in u.s.

production: Estimates and implications. Journal of Political

Economy 105(2), 249–283.

Cette, G., L. Koehl, and T. Philippon (2019). Labor shares.

Working Paper.

Crouzet, N. and J. Eberly (2018). Intangibles, investment, and

efficiency. AEA Papers and Proceedings.

De-Loecker, J., J. Eeckhout, and G. Unger (2019). The rise of

market power and the macroeconomic implications.

Diez, F. J., J. Fan, and C. Villegas-Sanchez (2018). Global

declining competition.

Gutiérrez, G. and T. Philippon (2018). How eu markets became

more competitive than us markets: A study of institutional

drift.

Gutiérrez, G. and T. Philippon (2019). The failure of free entry.


References III

Jones, C., G. Gutiérrez, and T. Philippon (2018). Declining

competition and investment in the us. Working Paper.

from good to bad concentration? u.s. industries over the ... · introduction theory measurement...

Documents