SESSIONE VII IMPRESE: TASSAZIONE, CREDITO E OCCUPAZIONE
Foreign ownership of business affected employment? Some evidence for Italian firms using a panel model
Matilde Bini, Leopoldo Nascia, Margherita Velucchi e Alessandro Zeli
GIORNATE DELLA RICERCA IN ISTAT | 10-11 NOVEMBRE 2014
Index
1. Introduction (1,2)
2. Effects of the occupational behaviour of fob
3. Indirect effects of fob
4. The panel data
5. Variables
6. Model (1,2,3)
7. Results (1,2,3)
8. Final remarks
GIORNATE DELLA RICERCA IN ISTAT | 10-11 NOVEMBRE 2014Foreign ownership of business affected employment? Alessandro Zeli
After the Euro introduction there had been an increasing foreign investment presence in Italy through acquisition of Italian firms and (at a lesser extent) greenfield investments
Introduction (1/2)
1
Stronger presence of foreign ownership of business (fob) in some economic sectors spillover effects foster the growth and the employment.
GIORNATE DELLA RICERCA IN ISTAT | 10-11 NOVEMBRE 2014Foreign ownership of business affected employment? Alessandro Zeli
2
Foreign ownership of business affected employment? Alessandro Zeli – Roma, 10-11 November 2014
We are focused on the following questions:
• Is there a relationship between fob and employment?
• Does the stronger presence of foreign ownership of business (fob) in some economic sectors foster the employment in the same sectors?
The aim of this paper is to understand the actual effects of increasing fob presence in Italy in terms of employment
changes during the 2002 – 2007 years
A first differencing panel model is applied and preliminary results are presented using a panel of Italian firms
Introduction (2/2)
3
The factor influencing the occupational behavior of fob (and the related industry) may have “substitution” effects:
• The fob carried out most aggressive competition policy by internal side towards domestic firms
• The local sub-contractors are being substituted by foreign ones (that may be belonging to the same fob Multinational group)
The fob may be more efficient: namely firms may adopt labor saving processes or they may broke the productive chain by delocalizing some processes to other foreign firms belonging or not to the same group
Effects of the occupational behaviour of fob
GIORNATE DELLA RICERCA IN ISTAT | 10-11 NOVEMBRE 2014Foreign ownership of business affected employment? Alessandro Zeli
4
• Positive spill-over may be originated by the exploitation of wider markets so the increased production may improve the firms growth and so increase the persons employed and, above all, skilled ones.
• Positive effect of the introduction of innovation and new technologies in acquired firms.
This may essentially improve productivity but it has a contradictory effects in terms of an increasing fate of the employment, if the introduction of new technologies permits the production of new goods and services (creating new markets).
Otherwise, if new technologies have impact on the production process of existing goods and services they can improve productivity but the employment rate decreases.
Indirect effects of fob
GIORNATE DELLA RICERCA IN ISTAT | 10-11 NOVEMBRE 2014Foreign ownership of business affected employment? Alessandro Zeli
4
It is important to underline the complex mutual link among these variables:
Domestic firms are not acquired randomly, foreign MNE acquires the most productive ones.
In general there is a relationship between size and productivity.
There is a fob tendency to employ a more skilled employees and to invest (or acquire from the other group members) in new technologies.
All these correlation imply some problems of endogeneity when testing econometric relationships.
Indirect effects of fob
GIORNATE DELLA RICERCA IN ISTAT | 10-11 NOVEMBRE 2014Foreign ownership of business affected employment? Alessandro Zeli
5
ISTAT proposed a panel on microdata at a firm level
It is a catch-up prospective database for the period 1999-2007,containing 13,573 firms, representing the firms population with 20 persons employed and over, classified by sector of activity
It is mainly based on cross-sectional surveys’ microdata (at firm-level) with the integration of administrative sources (to ensure the matching of units over time
and cover possible non-respondents) We used a balanced panel (around 8,000 firms)
Four different sources: •Istat Business Register of Italian firms, •the so-called SCI survey, focusing on all firms with more than 100 employees, •the so-called PMI survey that includes firms with 20-100 employees •the annual report of incorporated firms collected by the Central Balance-Sheet Data Office of Italy (plus several other surveys on specific firms’ aspects)
The panel data
GIORNATE DELLA RICERCA IN ISTAT | 10-11 NOVEMBRE 2014Foreign ownership of business affected employment? Alessandro Zeli
Variables (in logarithms)
6
e log of employment
lsales log of total sales lclul log of labur cost for unit of labur
vad log of productivity
lwdip log of share of white collars on total employees
linvit log of total investment for person employed
qfobshare of foreign ownership of business on total value added of economic sector (2-digit NACE)
qadd log of NACE 2-digit persons employed share on total employment
qfadd interaction effect between qadd and fob
fob dummy for foreign owned firm
GIORNATE DELLA RICERCA IN ISTAT | 10-11 NOVEMBRE 2014Foreign ownership of business affected employment? Alessandro Zeli
Model (1/3)
7
Two indicators: 1)the firm belonging to a foreign group (fob dummy); 2)rate of presence of fob within an industry, for instance: the ratio of fob value added on total industry, the value added to estimate the influence of fob on reference markets.
Following this approach, we define two dummies fob and qfob that identify the foreign owners firms and the share of value added produced by fob in its own the two-digit NACE sector respectively. So fob and qfob are calculated as follows:
F 0
F 1)(x
i
;
iif
if
tix
xfob
n
ti
jti
m
ti
jti
tiVA
VAqfob
; ;
; ;
; )(x
p
ti
jtiVA
; ;
n
ti
jtiVA
; ;
F denotes the set of fob firms
is the value added produced by fob in the economic sector j
is the value added produced by all firms in the sector j
The aim is to distinguish the effects of foreign ownership of business on employment, at the firm level
GIORNATE DELLA RICERCA IN ISTAT | 10-11 NOVEMBRE 2014Foreign ownership of business affected employment? Alessandro Zeli
Under the assumption
We use the differencing transformation of the model, lagging the model one period and subtracting yields
Model (2/3)
Several dynamic models are estimated using a panel data; in particular we use the following model for T time periods:
A dynamic model is estimatedSome control variables that influence the firms’ employment growth are included
e = b0 + b1sales + b2clul + b3vad + b4 wdip + b5 invt + b6 qfob + b7 qadd + b8qfadd + b15 fob + uit t=2,….,T
8GIORNATE DELLA RICERCA IN ISTAT | 10-11 NOVEMBRE 2014Foreign ownership of business affected employment? Alessandro Zeli
As with FE transformation this first differencing transformation eliminates the unobserved effect ci. The strictly exogeneity holds under the condition:
As it is well known, this condition does not hold in the presence of endogenous variables, so another method has to be implemented in order to estimate correctly the model.
The analysis is carried out in three steps:•OLS regression estimation for pooled data •Fixed Effects and Random Effects models estimation•IV estimation (needed in case of the endogeneity problem)
Model (3/3)
9GIORNATE DELLA RICERCA IN ISTAT | 10-11 NOVEMBRE 2014Foreign ownership of business affected employment? Alessandro Zeli
Results (1/3)
10
Variables Coeff. Std. Err.
lsales 0.2458 0.0025 ***lclul -0.4411 0.0041 ***
vad -0.0719 0.0020 ***lwdip -0.0022 0.0008 ***linvit -0.0009 0.0004 **qfob -0.0057 0.0029 **qadd 0.023*10-4 0.002*10-4 ***qfadd 0.003*10-4 0.001*10-4 ***fob -0.0256 0.0105 **cons 0.0025 0.0006 ***
First differences pooled OLS estimation
Adj R-squared = 0.3506F( 9, 42663) = 2560.70 Prob > F = 0.0000 1 = -0.0527734 (0.0052456) ***
*** significant at 1% level, , ** significant at 5% level, , * significant at 10% level
We considered the period from 2002 to 2007
Serial correlation of residual (1 = -0.053) indicates a problem of endogeneity
GIORNATE DELLA RICERCA IN ISTAT | 10-11 NOVEMBRE 2014Foreign ownership of business affected employment? Alessandro Zeli
11
FE REVariables Coeff. Std. Err. Coeff. Std. Err.Dlsales 0.2108 0.0027 *** 0.2458 0.0025 ***Dlclul -0.4460 0.0042 *** -0.4411 0.0041 ***Dvad -0.0667 0.0020 *** -0.0719 0.0020 ***Dlwdip -0.0022 0.0008 *** -0.0022 0.0008 ***Dlinvit -0.0009 0.0004 ** -0.0009 0.0004 **Dqfob -0.0094 0.0033 *** -0.0057 0.0029 **Dqadd 0.026*10-4 0.003*10-4 *** 0.023*10-4 0.002*10-4 ***Dqfadd 0.003*10-4 0.001*10-4 ** 0.003*10-4 0.001*10-4 ***Dfob -0.0190 0.0114 * -0.0256 0.0105 **cons 0.0028 0.0007 *** 0.0025 0.0006 ***R2: within 0.3554 R2 between 0.3502 R2 overall 0.3483
F test that allF(7572, 35091) = 1.05 =
u_i=0 Prob > F 0.0030
Hausman test: chi2(7) =1263.75Prob>chi2 = 0.0000
Results (2/3)
First differences FE and RE estimations
The Hausman test rejects the null, so we can correctly use FE model given the presence of fixed effects and having excluded the presence of firm specific effects on the rate of growth.
GIORNATE DELLA RICERCA IN ISTAT | 10-11 NOVEMBRE 2014Foreign ownership of business affected employment? Alessandro Zeli
IV (2SLS) IV Fixed Effects
Coeff. Std. Err. Coeff. Std. Err.lsales 0.353312 0.040434 *** 0.326256 0.046436 ***lclul -0.20604 0.091818 ** -0.19013 0.106831 *vad -0.46867 0.151052 *** -0.49353 0.173836 ***qfob -0.01088 0.004168 *** -0.01606 0.005177 ***qadd 1.73*10-6 3.17*10-7 *** 1.83*10-6 5.15*10-7 ***qfadd 3.91*10-7 1.54*10-7 ** 3.44*10-7 1.78*10-7 *fob -0.04266 0.015751 *** -0.03064 0.017512 *cons 0.008402 0.002811 ***Underidentification test 14 10.783Chi-sq(2) P-val 0.0009 0.0046Weak identification test (Cragg-
Donald Wald F statistic):7 5.391
Stock-Yogo weak ID test critical values: 10% maximal IV size
19.93 19.93
15% maximal IV size 11.59 11.59
20% maximal IV size 8.75 8.75
25% maximal IV size 7.25 7.25Sargan statistic 0.005 0.066Chi-sq(1) P-val 0.9431 0.7974
Results (3/3)
IV estimations
12
Tests confirm the correct identification of the IV model, Dvad was instrumented by Dwdip and Dinvt
GIORNATE DELLA RICERCA IN ISTAT | 10-11 NOVEMBRE 2014Foreign ownership of business affected employment? Alessandro Zeli
Final remarks
13
This paper deals with the change in the Italian employment rate, depending on both the presence of international business and its effects on the Italian economic system.
The results confirm the recent literature on the relationship between employment and foreign ownership of business.
Many factors influence the employment rate of growth in presence of firms belongings to a foreign group: innovation, access to wider markets, etc.
GIORNATE DELLA RICERCA IN ISTAT | 10-11 NOVEMBRE 2014Foreign ownership of business affected employment? Alessandro Zeli
14
The analysis shows a prevalence of negative effects that overcame the positive ones:
there is a negative effect of qfob (penetration of fob in a specific industry) and to a lesser extent of fob on employment
the negative qfob effects implies a negative spillover of fob on the employment rate of growth.
These results are still preliminary but they suggested to investigate more on the complex relationships involving employment and fob, especially including in the models innovation role, international diversification or trade openness.
Final remarks
GIORNATE DELLA RICERCA IN ISTAT | 10-11 NOVEMBRE 2014Foreign ownership of business affected employment? Alessandro Zeli
Thank you for your attention
GIORNATE DELLA RICERCA IN ISTAT | 10-11 NOVEMBRE 2014Foreign ownership of business affected employment? Alessandro Zeli