Job creation, job destruction and firm-level international trade involvement

1
Job creation, job destruction and firm-level
international trade involvement

• Mauro Pisu

Research Department
2
Introduction

• How the globalisation process may affect
• employment growth
• job reallocation
• Possible to investigate these issues using
  firm-level data
• Advantage
• micro explanations of macro phenomena
• analysis of topics macro data cannot deal with
  (e.g. job creation and destruction)
• Firm-level data evidence are important since
• academic literature has shown that firms are
  greatly heterogeneous
• net changes in employment at country- or
  industry-level result from expansion of some
  firms and contraction in others
• there is contemporaneous job creation and
  destruction

3
Heterogeneous firms trade theory

• Main features Melitz (2003), Bernard, Redding,
  Schott (2007)
• Heterogeneous firms (in productivity levels)
• Sunk and variable costs
• Results
• Only the most productive firms self-select into
  export markets
• Trade causes reallocation of jobs from non
  exporting to exporting firms
• Simultaneous job creation and destruction (Davis
  and Haltiwanger 1992) within industry because of
  international trade
• Models have been extended to FDI and imports
  decisions

4
Belgian firm-level data

• Merging of three data sets from the NBB
• Annual accounts of Belgian companies (Central
  Balance Sheet Office) information on the number
  of employees in FTE
• External trade statistics (Intra- and Extrastat
  inquiries)
• Survey on foreign direct investment (balance of
  payments statistics)
• Time span 1997 - 2004
• Population in 2004
• 136,945 firms1 that reported at least one
  employee (1,827,786 workers in FTE), of which
• 2,643 firms involved in direct investment
  relations (either as direct investor or as
  direct investment company or both at the same
  time)
• 26,744 firms involved in external trade
• Not entirely comparable with the population
  considered in the national accounts

1 Out of 309,486 firms that filed a balance sheet
in 2004.
5
Distribution by type of firm in 2004
(percentages of the total)
Total
Manufacturing
Wholesale and retail
Services
2
4
1
1
2
1
2
2
8
14
7
4
21
Shares in the number of firms
14
79
59
65
87
8
6
7
3
50
26
11
35
4
27
43
24
21
Shares in employment
8
15
28
6
11
12
24
6
16
3
8
11
6
2
Source NBB.
6
Sectoral distribution of the external trade
(percentages of the total for the whole
population)
  Exports1 value Exports1 value Imports1 value Imports1 value
  1997 2004 1997 2004
Agriculture, fishing and mining 0.5 0.5 0.2 0.2
Manufacturing 69.2 69.3 46.5 46.9
Recycling, utilities and construction 0.7 1.4 2.9 3.4
Services 29.6 28.8 50.4 49.5
Wholesale and retail 28.1 26.6 48.0 46.3
Other services 1.5 2.3 2.4 3.1
of which Coordination centres 0.1 0.2 0.1 0.2
Total 100.0 100.0 100.0 100.0
Source NBB. 1 Exports and imports of goods only.
7
Contribution to net changes in employment
(manufacturing and services, selected industries,
percentage changes between 1997 and 2004)
Source NBB. Note only those industries that
account for 2 p.c. or more of total employment
are represented here.
8
Job creation and destruction
(Cumulated annual changes in the number of jobs
1997 - 2004)

• Net changes in employment hide a great deal of
  the contemporaneous job creation and destruction
• These figures imply that on average 224,491 jobs
  were created or destroyed each year in these two
  broad sectors

(1) Creation (2) Destruction (1) - (2) Net creation (1) (2) Total turnover
Manufacturing 207,730 251,805 -44,075 459,535
Services 629,312 482,593 146,719 1,111,905
Total 837,042 734,398 102,644 1,571,440
Source NBB.
9
Share of jobs created and destroyed due to
entries and exits, 1997-2004
Source NBB. Note These graphs show the
percentage of the jobs created and destroyed by
each type of firms due to entries and exits,
respectively.
10
Job creation and destruction rates across trading
status
(contemporaneous creation and destruction rates
Davis and Haltiwanger (1992) methodology)
NT Non-traders, EXP Exporters, IMP Importers,
TWT Two-way traders, FOR Foreign affiliates,
MNE Belgian Multinationals.
Source NBB Note 1 p.c. 0.01
11
Regression analysis

• Employment can also be affected by other factors
  (e.g. firms' size, productivity, profitability,
  business cycle and industry-specific effects)
• The respective impacts can be disentangled using
  regression analysis.
• Specification
• gijt a ßTSijt dSijt ?Pijt sPRijt
  ID YD eit
• Where gijt employment growth rate of firm i in
  industry j at time t TSijt trading status
  Sijt size Pijt productivity PRijt
  profitability ID industry (four digits) YD
  time

12
Regression results
(OLS estimates coefficients in percentage point)
Manufacturing Manufacturing Manufacturing Services Services Services
Net job creation Job creation Job destruction Net job creation Job creation Job destruction
Trade status (benchmark non-trader) Domestic exporter 1.5 -6.7 -13.3 1.3 -6.7 -12.4
Trade status (benchmark non-trader) Domestic importer 5.8 -4.5 -16.3 3.2 -5.5 -14.6
Trade status (benchmark non-trader) Domestic two-ways trader 5.2 -5.8 -18.4 3.3 -5.6 -14.8
Trade status (benchmark non-trader) Belgian multinational 4.5 -8.6 -21.9 -2.3 -9.8 -13.9
Trade status (benchmark non-trader) Foreign Affiliate 2.5 -11.5 -22.4 -2.0 -8.6 -13.4
Size (benchmark small) Medium-small -1.8 -12.1 -22.6 -6.0 -39.0 -6.8
Size (benchmark small) Medium-large -4.1 -17.2 -29.1 -5.1 -40.5 -21.4
Size (benchmark small) Large -7.5 -20.3 -27.6 -6.1 -45.4 -30.9
Productivity (benchmark low) Medium-low 1.8 -13.8 -15.4 -3.8 -17.7 -16.2
Productivity (benchmark low) Medium-high 4.7 -13.1 -18.2 -2.7 -19.2 -20.2
Productivity (benchmark low) High 5.7 -12.0 -18.4 -3.8 -21.4 -24.7
Profitability (benchmark low) Medium-low 1.3 -8.4 -8.2 -3.0 -10.2 -7.8
Profitability (benchmark low) Medium-high 3.6 -6.6 -8.6 n.s.s. -6.2 -6.6
Profitability (benchmark low) High 3.6 n.s.s. n.s.s. n.s.s. 1.1 4.7
Source NBB. n.s.s. not statistically
significant. Notes Specification is gijt a
ßTSijt dSijt ?Pijt sPRijt ID YD eit,
where gijt employment growth rate of firm i in
industry j at time t TSijt trading
status Sijt size Pijt productivity PRijt
profitability ID industry (four digits) YD
time. Time and industry dummies not reported.
13
Regression results, without entries and exits
(OLS estimates coefficients in percentage)
Manufacturing Manufacturing Manufacturing Services Services Services
Net job creation Job creation Job destruction Net job creation Job creation Job destruction
Trade status (benchmark non-trader) Domestic exporter 6.2 n.s.s. -15.0 6.2 n.s.s. -14.0
Trade status (benchmark non-trader) Domestic importer 9.4 n.s.s. -17.7 7.6 0.5 -15.8
Trade status (benchmark non-trader) Domestic two-ways trader 9.1 -0.7 -20.3 7.6 n.s.s. -17.0
Trade status (benchmark non-trader) Belgian multinational 7.5 -3.9 -22.4 1.8 -3.2 -12.9
Trade status (benchmark non-trader) Foreign Affiliate 6.5 -5.1 -23.0 1.4 -3.3 -13.8
Size (benchmark small) Medium-small 5.0 -2.8 -20.7 9.4 -14.8 -5.9
Size (benchmark small) Medium-large 3.4 -6.8 -26.3 11.4 -15.0 -19.2
Size (benchmark small) Large n.s.s. -10.5 -24.3 11.0 -18.5 -27.4
Productivity (benchmark low) Medium-low 6.9 -1.1 -10.8 6.9 -0.7 -11.4
Productivity (benchmark low) Medium-high 9.0 n.s.s. -11.9 8.7 -0.7 -14.0
Productivity (benchmark low) High 10.5 n.s.s. -11.6 10.6 0.4 -17.5
Profitability (benchmark low) Medium-low 4.2 -1.2 -6.1 3.7 -0.9 -6.5
Profitability (benchmark low) Medium-high 6.0 n.s.s. -6.5 5.6 1.0 -5.9
Profitability (benchmark low) High 5.2 3.8 n.s.s. 3.5 4.5 3.1
Source NBB. n.s.s. not statistically
significant. Notes Specification is gijt a
ßTSijt dSijt ?Pijt sPRijt ID YD eit,
where gijt employment growth rate of firm i in
industry j at time t TSijt trading
status Sijt size Pijt productivity PRijt
profitability ID industry (four digits) YD
time. Time and industry dummies not reported.
14
Regression results for wholesale and retail and
others services
(OLS estimates coefficients in percentage)
With Entries and Exits With Entries and Exits With Entries and Exits With Entries and Exits With Entries and Exits With Entries and Exits With Entries and Exits With Entries and Exits
Net job creation Net job creation Job creation Job creation Job destruction Job destruction
WR Services WR Services WR Services
Trade status (benchmark non-trader) Domestic exporter 1.5 n.s.s. -7.1 -6.9 -13.1 -11.4
Trade status (benchmark non-trader) Domestic importer 3.0 2.5 -6.4 -4.4 -16.0 -10.0
Trade status (benchmark non-trader) Domestic two-ways trader 2.8 1.6 -6.8 -6.2 -16.4 -10.9
Trade status (benchmark non-trader) Belgian multinational n.s.s. -4.6 -8.4 -10.9 -15.5 -12.9
Trade status (benchmark non-trader) Foreign Affiliate n.s.s. -4.6 -8.8 -9.0 -18.4 -9.5
Without Entries and Exits Without Entries and Exits Without Entries and Exits Without Entries and Exits Without Entries and Exits Without Entries and Exits Without Entries and Exits Without Entries and Exits
Trade status (benchmark non-trader) WR Services WR Services WR Services
Trade status (benchmark non-trader) Domestic exporter 6.7 5.3 -0.8 n.s.s. -15.1 -12.4
Trade status (benchmark non-trader) Domestic importer 8.1 7.2 n.s.s. 1.6 -17.6 -11.3
Trade status (benchmark non-trader) Domestic two-ways trader 8.3 7.3 n.s.s. 1.2 -19.2 -12.5
Trade status (benchmark non-trader) Belgian multinational 3.8 n.s.s. -3.7 -3.0 -17.4 -10.4
Trade status (benchmark non-trader) Foreign Affiliate 4.9 -1.3 -2.4 -4.1 -19.6 -9.4
Source NBB. n.s.s. not statistically
significant. Notes Specification is gijt a
ßTSijt dSijt ?Pijt sPRijt ID YD eit,
where gijt employment growth rate of firm i in
industry j at time t TSijt trading
status Sijt size Pijt productivity PRijt
profitability ID industry (four digits) YD
time. Time and industry dummies not reported.
15
International trade and job reallocation

• Issue effect of firm-level parcitipation in
  int. markets on job reallocation shifts of jobs
  from non-trading to trading firms
• Methodology of Davis and Haltiwanger (1992)
  applied to trading and non trading firms (within
  sectors)
• Excess job reallocation
• s are different types of firms (trading and
  non-trading firms)
• If between component high and within component
  low
• then firm-level trade is an important cause of
  job reallocation (consistent with theoretical
  models)

16
Between share of excess job reallocation(average
of three-digit industry level figures)
Year Groups International status Groups International status Groups International status Groups International status
Simple classification Simple classification Detailed classification Detailed classification
Without entries and exitsa With entries and exitsb Without entries and exitsa With entries and exitb
1998 11.30 20.17 24.99 32.17
1999 2.93 13.24 20.64 29.20
2000 5.22 16.67 13.79 27.95
2001 5.88 11.59 15.69 19.91
2002 3.44 9.25 13.16 17.46
2003 6.14 13.03 25.66 30.02
2004 10.28 16.68 23.77 26.13
Mean 6.46 14.38 19.67 26.12
Source NBB. Notes Yearly figures are weighted
averages of the between shares computed for each
three-digit Nace sector the weights used are the
number of firms in each three-digit industry-year
cell a figures computed considering only
surviving firms (i.e. dropping entries and
exits) b figures computed allocating births and
deaths to specific types of firms according to
their trading status the year they entered and
the year before they exited.
17
Summary and conclusions (1/2)

• Firm level data allow to study how international
  involvement affects the process of job creation
  and destruction rates, not only net changes
• Belgian data controlling for other factors,
  international trade involvement is associated
  with
• higher net employment growth rates
• lower job creation and destruction rates
• ? Workers face a lower uncertainty in trading
  firms

18
Summary and conclusions (2/2)

• Job reallocation rates (Davis and Haltiwanger
  1992) between trading and nontrading firms are
  relatively low
• 15-25 percent at three-digit level
• All types of firms (traders, non-traders, MNEs
  and foreign affiliates) create and destroy jobs
• participation in international markets is not the
  only explanation of job creation and destruction
  process
• Open questions
• impact of international trade involvement on
  workers having different skills
• how international trade compares with other
  shocks