Outsourcing versus integration at home and abroad

1
Outsourcing versus integration at home and abroad

• Stefano Federico
• (Bank of Italy)

December 2008 Vienna
2
Motivation

• Bernard, Jensen, Redding and Schott (2007, JEP)
• Further progress () will require explicit
  consideration of the boundaries of the firm,
  including the decisions about whether to insource
  or outsource stages of production, and whether
  such insourcing or outsourcing takes place within
  or across national boundaries (p. 128, italics
  added)

3
Framework

• Firms in need of inputs choose both location and
  organizational form

Integration Outsourcing
At home Domestic integration (DI) Domestic outsourcing (DO)
Abroad Foreign integration (FI) Foreign outsourcing (FO)
4
Related literature

• Theory
• Antràs (2003)
• Antràs and Helpman (2004, 2008)
• Grossman and Helpman (2004, 2005)
• Empirics
• U.S. trade data Antràs (2003), Yeaple (2006),
  Nunn and Trefler (2008), Bernard et al. (2008)
• Firm-level data Tomiura (2007), Defever and
  Toubal (2007)

5
Our contribution (1)

• For each firm, we observe inputs acquired from
  
• domestic affiliates (DI)
• domestic non-affiliates (DO)
• foreign affiliates (FI)
• foreign non-affiliates (FO)
• Improvement on previous literature
• Studies based on trade data do not observe
  domestic inputs

6
Our contribution (2)

• Only subcontracting inputs are considered
• Intermediate goods and services, in conformity
  with the acquiring companys projects,
  technologies or prototypes
• Fully consistent with theory (relationship-specifi
  c investments)
• To us, outsourcing means more than just the
  purchase of raw materials and standardized goods.
  It means finding a partner with which a firm can
  establish a bilateral relationship and having the
  partner undertake relationship-specific
  investments () (Grossman and Helpman 2005,
  italics added)

7
Productivity ordering
AH(04) fi gt fo AH(04) fi lt fo GH(04)
Foreign integration 1 2 3
Foreign outsourcing 2 1 1, 4
Domestic integration 3 4 2
Domestic outsourcing 4 3
Source adapted from Spencer (2005).
8
Headquarter intensity

• Integration will be preferred to outsourcing when
  the contribution of final-good producer is larger
• i.e. in industries with high intensity in
  headquarter services (capital intensity, skill
  intensity, RD intensity, advertising intensity,
  etc.)

9
Data and sample

• Survey on Italian manufacturing firms
  (Mediocredito Capitalia), 7th ed.
• Period 1997 (1989-1997 for balance-sheet data)
• Our sample 3,819 firms (4 of universe)
• Biased in favour of medium-large firms

10
Sourcing strategies
of firms (on sample)
FI FO DI DO No
1.2 7.0 4.7 30.9 65.7

• Note Sourcing strategies are not mutually
  exclusive

11
Econometric analysis 1
A given characteristic of firm i (VA, L, TFP,
etc.)
Set of controls industry, area, export status
Dummies for each sourcing strategy (relative to
the baseline group DO)
12
Size and productivity premia
Va L VA/L TFPOLS TFPFE TFPLP
FI 2.02 1.73 .29 .17 .42 .76
FO .39 .31 .08 .06 .10 .12
DI .99 .81 .18 .11 .22 .38

P-value tests P-value tests P-value tests P-value tests P-value tests P-value tests P-value tests
FIFO 0.00 0.00 0.00 0.00 0.00 0.00
FIDI 0.00 0.00 0.21 0.49 0.04 0.00
FODI 0.00 0.00 0.08 0.24 0.02 0.00
N. Obs. 1,316 1,316 1,316 1,316 1,316 1,316

• All coefficients are significant at 10.
  Coefficients for industry, area and export status
  dummies are not reported.

13
Econometric analysis 2
Inputs from foreign affiliates on foreign inputs
Inputs from domestic affiliates on domestic inputs
TFP fixed effects
Indicator of headquarter intensity (? next slide)
14
Headquarter intensity
Industry-level Industry-level Firm-level
Capital intensity Kj/Lj SCALEj Ki/Li
Skill intensity Hj/Lj Wj/Lj Hi/Li
RD intensity RDj RDi
15
Foreign integration
(1) (2) (3)
TFPi,FE .174 .165 .147
Kj/Lj .058
Hj/Lj .213
RDj -.178
SCALEj .044
Wj/Lj .066
Ki/Li .038
Hi/Li .024
RDi -.466
R-sq. .083 .085 .075
N.Obs. 298 298 298
16
Domestic integration
(1) (2) (3)
TFPi,FE .084 .083 .076
Kj/Lj .040
Hj/Lj .060
RDj -.084
SCALEj .027
Wj/Lj .034
Ki/Li .036
Hi/Li -.065
RDi .873
R-sq. .025 .027 .037
N.Obs. 1,283 1,283 1,283
17
Economic significance

• Foreign integration
• A one standard deviation increase in TFP results
  in a .21/.25 standard deviation increase in the
  share of foreign integration
• A one standard deviation increase in capital
  intensity results in a .11/.12 standard deviation
  increase in the share of foreign integration
• Domestic integration
• Smaller values (.12/.14 for TFP, .07/.13 for
  capital intensity)

18
Robustness

• Additional explanatory variables
• Firms wage costs
• Firm age
• Demand cyclicality
• Demand seasonality
• Area dummies
• U.S. headquarter intensity measures (NBER)
• Alternative methods
• Tobit
• Probit
• SURE

19
Conclusions

• Productivity ordering FIgtDIgtFOgtDO
• Consistent with ordering assumed by Antràs and
  Helpman (2004), except DIgtFO in our case
• To our knowledge, first estimation for all four
  organizational forms
• Integration is more likely in capital-intensive
  industries
• Controlling for both skill intensity and RD
  intensity
• Consistent with theory (Antràs 2003) and previous
  evidence