**4. Empirical methodology**

*Outsourcing and Offshoring*

Petroleum Products &

6. Nonelectric Machinery, Office Equipment &

Computers, Electric Machinery, Electronic Mat.

7. Medical Apparel & Instruments,

**Table 2.**

Chemicals

and Publishing, Petroleum Products and Chemicals. On the other side, industries focusing more on horizontal FDI are Office Equipment and Computers, Electric Machinery, Electronic Materials and Transportation. Overall, firms in the whole sample seem to reduce the international activity in 2006 with respect to 2003. However, some industries increase their offshoring activity over time, such as Rubber and Plastics, Non-Metal Minerals, Metals, Metals Products and Furniture, while Food, Beverages, Textiles and Clothing increase their share in Horizontal FDI

**Category Year Distributi0n (%) N. firms**

1.Food & Beverages, 2003 0.0 0.25 67.75 32.0 400 Textiles, Clothing 2006 0.27 0.0 55.38 44.35 372 2. Leather, Wood, 2003 1.23 5.62 75.57 17.93 569 Paper products 2006 0.51 1.86 69.93 27.87 592 3.Printing & Publishing, 2003 0.68 2.05 58.56 39.38 292

4.Rubber & Plastics, 2003 0.96 0.48 76.50 22.06 417 Non-metal minerals, Metals 2006 0.46 0.70 65.20 33.64 431 5. Metal Products 2003 1.05 0.35 61.24 37.35 854

Vehicles 2006 0.53 0.80 60.90 38.03 376 8- Other Transportation 2003 2.25 1.12 68.54 28.03 89

9.Furniture 2003 1.21 1.21 83.0 14.98 247

**HFDI OFF EX D**

2006 0.0 0.24 38.59 61.17 412

2006 0.09 0.43 50.51 33.64 1,166

2003 22.45 1.51 85.69 10.73 531

2006 0.29 0.73 77.71 48.97 682

2003 2.11 3.17 73.94 21.13 284

2006 0.0 0.88 72.81 26.32 114

2006 0.34 2.68 72.15 25.17 298

**Figure 1** shows kernel densities of TFP for the four types of firms in our data set. The ordering of the firms' productivity seems to be the following: both in 2001-2003 and 2004-2006, firms producing abroad are more productive than those

As the figure illustrates there are productivity differentials among firm groups. The differences are more pronounced for the period 2004-2006. The distribution of the log of total factor productivity (TFP) for the four types of Italian firms are those serving only the domestic market (domestic firms), those engaging in export (pure exporters), those engaging in horizontal FDI, and those engaging in offshoring. concentrated over larger TFP values with respect to exporters. In turn, the latter are better performers in TFP than domestic counterparts. However, the ranking of distributions of firms that perform horizontal FDI with respect to offshorers is not

and, finally, Other Transportation raise their share in Exports.

*Distribution of different internationalization modes across industries.*

exporting, being the latter more productive than domestic firms.

clear-cut as they seem to be almost overlapping.

**76**

Since productivity is the key element of our study, in order to overcome simultaneity and andogeneity problems of parametric approximation of TFP, we use the semi-parametric method suggested by Levinshon and Petrin [39] and widely used in the literature.3 Specifically, this estimator permits to estimate production functions using firm-level data and solves the simultaneity bias of correlation of productivity shocks and input choices by using a composite index of materials (intermediates) to proxy unobserved productivity shocks.4

Consider the following Cobb-Douglas production function:

$$\mathbf{y}\_{\text{ir}} = \mathbf{a}\_0 + \mathbf{a}\_l \mathbf{l}\_{i,t} + \mathbf{a}\_k \mathbf{k}\_{i,t} + \mathbf{w}\_{i,t} + \mathbf{u}\_{i,t} \tag{2}$$

for the LP estimation it becomes:

$$\mathbf{y}\_{i,t} = \mathbf{a}\_0 + \mathbf{a}\_l l\_{i,t} + \mathbf{a}\_m m\_{i,t} + \mathbf{a}\_k k\_{i,t} + \mathbf{w}\_{i,t} + \mathbf{u}\_{i,t} \tag{3}$$

where y, l, k, m are respectively the log of output, employment, intermediate inputs, and capital stock for firm i at time t and ws,t is the productivity shock observable by firms. Although also this method of computation of TFP suffers some significant identification problem, it allows us to limit endogeneity issues. The regression implemented sector-by-sector on each wave's three-year panels uses materials from the balance sheet data as well as white and blue collars as labour

<sup>3</sup> The Levinshon and Petrin measure of TFP has been calculated by implementing the levpet routine available in Stata.

<sup>4</sup> The method relies on a a function in which intermediate inputs are used to control for productivity and this has an advantage over the Olley and Pakes [40] method which uses investment to proxy for productivity. In our data set (as well as other firm-level datasets) this variable was not available.

inputs. As previously, also these measures at the firm level were re-scaled by the macro-sector level mean<sup>5</sup> of TFP. Finally, we averaged the values over the three years wave.

With this measure we provide results from premia estimates in the Table below. In more detail, we seek to measure the difference in performance among firms in overseas markets according to different strategies. Thus, as standard in the literature, we run OLS regressions to estimate the relationship between firms 'performances and various internationalization strategies. The procedure follows the Bernard and Jensen (1995) paper extended to include our strategies as follows:

(ISi,t = [HFDIi,t, OFFi,t, EXPi,t, Di,t])6

The regression implemented is:

$$\text{Y}\_{i,t} = \alpha + \beta \text{IS}\_{i,t} + \gamma \text{Empolyment}\_{i,t} + \sum\_{i} \gamma\_i \text{INDUSTRY}\_i + \sum\_{j} \delta\_j \text{AREA}\_j + \varepsilon\_{i,t}, \tag{4}$$

where i is the index of the firm and t is the time indicator ISi, t is a dummy variable for the international status of the firm, that takes on value of 1 if the firm internationalizes in year t, and 0 otherwise. y represents the measure of firm performance. We consider as firm performance measures not only TFP and labour productivity (Value added/L) but also the capital/labour ratio and gross sales perworker. As usual we control for industry, region dummies and firm size measured by the number of employees. Productivity premia calculated by the β coefficient are reported in the **Table 3**.

In the second part of the Table we have divided our firms by country destinations of their internationalization activities. The geographical areas of internationalization of Italian firms in our dataset have been distinguished in the North in which we have included all high income countries (EU15), USA, Japan, Canada, Australia) and the South in which we have included less developed countries (East Asian countries and 8 Central and Eastern European countries (see Appendix).

The analysis of the simplest strategy considered (i.e. purely exporters), EXP yields the clearest outcome: exporters perform better than domestic firms in terms of TFP7 sales, capital/labour ratio and labour productivity. Distinguishing by export destination does not affect what just assessed. The main implication of this result is that the importance of distance should have diminished over time in the sense that advances in technology have contributed to reducing the costs of trade. Therefore, the well-established-finding that bilateral trade diminishes with distance should be rethought. Indeed, in some recent papers this puzzle has been explored and some explanations have been advanced, which are based on the concept of "geographic neutrality" (see [41]).

Firms doing both export and offshoring turn out to have significantly larger sales with respect to only exporters. Moreover, they also show larger labour productivity. In terms of TFP, offshorers seem to be better performers than exporters only when the destination country is located in the North. Finally, companies performing both export and horizontal FDI have significantly larger sales with respect to both only exporters and offshorers. Results in terms of labour productivity are not statistically significant, differently from results on capital/labour ratio, that turns out to be larger for foreign investors in the South. Hereby, our investigation shows that FDI

**79**

*\**

**Table 3.**

*\*\*At 5% significance. \*\*\*At 1% significance.*

*At 10% significance. Robust standard errors are calculated.*

*Productivity premia based on regression estimates.1*

*Entry-Mode Selection and Firm's Productivity across Market Destinations: An Empirical…*

HFDI β 0.738 1.563\*\*\* -0.634 0.419

OFF β 0.411 0.858\*\*\* -0.032 0.482\*\*\*

EXP β 0.142\*\*\* 0.582\*\*\* 0.068\* 0.134\*\*\*\*

HFDI (North) β 1.975 1.628\*\*\* -0.029 0.386

OFF(North) β 1.119\* 1.295\*\*\* 0.278 0.364

EXP (North) β 0.123\*\*\* 0.536\*\*\* 0.089\*\* 0.114\*\*\*

HFDI (South) β -0.030 0.666\*\*\* 0.388\*\*\* -0.019

OFF (South) β 0.917 0.468\* 0.047 0.467\*\*

EXP (South) β 0.142\*\*\* 0.358\*\*\* 0.068\* 0.134\*\*\*

**TFP (2006) Sales (2006) K/L (2006) VA/L (2006)**

s.e 0.795 0.302 0.528 0.289 n.obs. 2605 2671 2670 2671

s.e 0.259 0.177 0.225 0.140 n.obs. 2605 2671 2670 2671

s.e 0.048 0.036 0.042 0.028 n.obs. 4165 4264 4261 4264

s.e 1.272 0.542 0.554 0.449 n.obs. 2605 2671 2670 2671

s.e 0.664 0.377 0.633 0.256 n.obs. 2605 2671 2670 2671

s.e 0.047 0.035 0.040 0.026 n.obs. 4165 4264 4261 4264

s.e 0.175 0.104 0.122 0.091 n.obs. 2605 2671 2670 2671

s.e 0.596 0.257 0.528 0.229 n.obs. 2605 2671 2670 2671

s.e 0.048 0.031 0.042 0.028 n.obs. 4165 4264 4261 4264

and offshoring are riskier strategies. To minimize risk during the process of complex strategies of internationalization it is better to enter countries with similar institutional environments which facilitate coordination need. Thus, internationalization performance is better when FDI and offshoring firms choose markets that have preferences and norms similar to those of the home market. Many studies show that institutional distance is important for internationalization choices and FDI flows. Among the dimensions of institutional distance it should be considered legal rules [42], protectionist policies, credit market regulations as well as legal constraints in the labour market [43]. More recently, s such concepts come out in Cezar and Escobar [44], that set up a heterogeneous firm theoretical framework, also empirically validated, about the effect of institutional distance on both the location and the

*DOI: http://dx.doi.org/10.5772/intechopen.95288*

<sup>5</sup> Because of data constraints, we aggregated ATECO 1991 2-DIGIT manufacturing sectors into nine broader categories that are defined in appendix C.

<sup>6</sup> Tests on H FDI and Offshoring are run over a sample of firms all doing also export (domestic firms are dropped). Tests on EX (only Export) are run over a sample of firms that do not engage neither in FDI nor offshoring.

<sup>7</sup> As TFP measure, we use LP estimates, scaled by the macro-sector level mean.


*Entry-Mode Selection and Firm's Productivity across Market Destinations: An Empirical… DOI: http://dx.doi.org/10.5772/intechopen.95288*

*\* At 10% significance. Robust standard errors are calculated.*

*\*\*At 5% significance.*

*\*\*\*At 1% significance.*

#### **Table 3.**

*Outsourcing and Offshoring*

macro-sector level mean<sup>5</sup>

(ISi,t = [HFDIi,t, OFFi,t, EXPi,t, Di,t])6 The regression implemented is:

> γ

αβ

reported in the **Table 3**.

neutrality" (see [41]).

broader categories that are defined in appendix C.

of TFP7

years wave.

inputs. As previously, also these measures at the firm level were re-scaled by the

With this measure we provide results from premia estimates in the Table below. In more detail, we seek to measure the difference in performance among firms in overseas markets according to different strategies. Thus, as standard in the literature, we run OLS regressions to estimate the relationship between firms 'performances and various internationalization strategies. The procedure follows the Bernard and Jensen (1995) paper extended to include our strategies as follows:

*i t i t it i i j j it*

In the second part of the Table we have divided our firms by country destinations of their internationalization activities. The geographical areas of internationalization of Italian firms in our dataset have been distinguished in the North in which we have included all high income countries (EU15), USA, Japan, Canada, Australia) and the South in which we have included less developed countries (East Asian countries and 8 Central and Eastern European countries (see Appendix). The analysis of the simplest strategy considered (i.e. purely exporters), EXP yields the clearest outcome: exporters perform better than domestic firms in terms

sales, capital/labour ratio and labour productivity. Distinguishing by export

destination does not affect what just assessed. The main implication of this result is that the importance of distance should have diminished over time in the sense that advances in technology have contributed to reducing the costs of trade. Therefore, the well-established-finding that bilateral trade diminishes with distance should be rethought. Indeed, in some recent papers this puzzle has been explored and some explanations have been advanced, which are based on the concept of "geographic

Firms doing both export and offshoring turn out to have significantly larger sales with respect to only exporters. Moreover, they also show larger labour productivity. In terms of TFP, offshorers seem to be better performers than exporters only when the destination country is located in the North. Finally, companies performing both export and horizontal FDI have significantly larger sales with respect to both only exporters and offshorers. Results in terms of labour productivity are not statistically significant, differently from results on capital/labour ratio, that turns out to be larger for foreign investors in the South. Hereby, our investigation shows that FDI

<sup>5</sup> Because of data constraints, we aggregated ATECO 1991 2-DIGIT manufacturing sectors into nine

<sup>7</sup> As TFP measure, we use LP estimates, scaled by the macro-sector level mean.

<sup>6</sup> Tests on H FDI and Offshoring are run over a sample of firms all doing also export (domestic firms are dropped). Tests on EX (only Export) are run over a sample of firms that do not engage neither in FDI nor

γ

*y IS Empolyment INDUSTRY AREA* ,, , =+ + + + + ∑ ∑ ,

where i is the index of the firm and t is the time indicator ISi, t is a dummy variable for the international status of the firm, that takes on value of 1 if the firm internationalizes in year t, and 0 otherwise. y represents the measure of firm performance. We consider as firm performance measures not only TFP and labour productivity (Value added/L) but also the capital/labour ratio and gross sales perworker. As usual we control for industry, region dummies and firm size measured by the number of employees. Productivity premia calculated by the β coefficient are

of TFP. Finally, we averaged the values over the three

*i j*

δ

 ε, (4)

**78**

offshoring.

*Productivity premia based on regression estimates.<sup>1</sup>*

and offshoring are riskier strategies. To minimize risk during the process of complex strategies of internationalization it is better to enter countries with similar institutional environments which facilitate coordination need. Thus, internationalization performance is better when FDI and offshoring firms choose markets that have preferences and norms similar to those of the home market. Many studies show that institutional distance is important for internationalization choices and FDI flows. Among the dimensions of institutional distance it should be considered legal rules [42], protectionist policies, credit market regulations as well as legal constraints in the labour market [43]. More recently, s such concepts come out in Cezar and Escobar [44], that set up a heterogeneous firm theoretical framework, also empirically validated, about the effect of institutional distance on both the location and the volume of FDI. In particular, they show that the larger the institutional distance, the larger the adaptation costs multinational have to overcome in order to access foreign markets. In turn, large adaptation costs due to institutional gap reduce both the number of firms able to undertake FDI and the profitability of FDI themselves.

Indeed, the inefficiency in FDI and offshoring in the South evidenced in our work may be due to additional operational costs related to extended supply chains. While some costs are expected, such as those of carrying higher inventories due to longer delivery chain, higher costs of inventory obsolescence, higher insurance costs, higher management operational requirements, there are many additional costs that are unexpected and labelled "hidden costs of offshoring" recently investigated by the international business literature [45]. There can also be higher local legal and administrative burdens, country trade disputes resulting in punitive fines and instances of intellectual property theft. It is also felt that more successful products can be better designed and improved by having the relevant functions (design, research and development, production, and sales) close to each other.

### **5. Conclusions**

Based on simple regression tests and using a panel data set of about 7300 Italian manufacturing firms, we have explored in this work to what extent the ordering of the productivity distributions of firms differently engaged in overseas markets conforms to the predictions of the literature. We categorized our firms into four groups according to whether they perform FDI of horizontal type, offshoring activities motivated by comparative advantages of the host country, purely exporters as well as firms that serve only domestic consumers.

Our results suggest that exporters outperform firms serving only the domestic market and outperforms also firms engaging in H-FDI in terms of productivity. Even when we include offshoring firms the productivity of this type of firms is not higher than exporting firms. Hence, our simple analysis shows that firms that perform FDI, either horizontal or vertical do not show higher productivities. The possible explanation of no difference in productivity between these two forms of foreign investments is that they are strictly interrelated and firms engaged in both activities perform equally in terms of productivity. Another reason is that increasing productivity from FDI and offshoring is not a short run phenomenon but it takes time to be conducive to high productivity (see [46]) On the contrary, exporting firms are exposed to new knowledge, technology and greater competitiveness in the global market and take advantage from this exposure through substantial learning processes that may improve their performances. The learning effect of exporting, as the literature shows, does not require a long time spin. On this ground, there is a large body of empirical evidence - known as "the microeconomics of international firm activity [16, 47, 48] that shows a positive correlation between firm productivity and export propensity just after two or three years. This evidence follows key theoretical contributions that points to the existence of large fixed cost of horizontal FDI and offshoring. To these contributions adds the ones that comes from the recent literature on the hidden costs of offshoring. Many offshored activities are strictly linked with domestic processes, which require complex coordination costs and unanticipated organizational need as well as other hidden costs that can disrupt in-house learning processes [49–53].

More work is necessary to demonstrate how these costs arise and quantify their impact especially when the distance between countries and fragmentation of various stages of production in different countries are taken into account. Indeed, when we differentiate our firms by geographical location of FDI and export destinations

**81**

*Entry-Mode Selection and Firm's Productivity across Market Destinations: An Empirical…*

we find support to the HMY ranking only for FDI decisions in the high-income countries of the North but not when activities are located in Southern countries. For firms operating in low-income countries of the South the more productive firms are purely exporting firms. This means that more distant markets either in physical terms or in technological and institutional characteristics entail diversities in terms of costs and risks. Therefore, only firms with higher productivity may serve these countries. Overall, the productivity premia of FDI firms are higher for firms operating in high income countries, exporting firms are the best performers across

The results of this work is likely to be helpful in the formulation of market entry strategies. Before proceeding with complex entry mode, managers need to evaluate costs and benefits of their moves as well as country risks relative to the home country. In terms of policy implications, the evidence of this work suggests that exporting brings with it potentially positive effects. When evaluating more complex forms of entry-modes managers should consider that they seem to be favorable only for locations in the North where firms have previous experience, the cultural distance is low and where they can find market similarities such as favorable conditions to increase their performance (knowledge infrastructure and availability of qualified personnel). Then, the indication is that for Italy, export-enhancing public policy should promote exporting to all destinations especially considering small businesses, which are the majority in the industrial structure of the country.

By concluding, some caution must be exercised in generalizing the outcomes of this work. A limitation of this work is the small number of firms that perform FDI and offshoring with respect to the number of firms that perform exporting. While it is possible to isolate pure exporters, this cannot be done for the other entry-modes: companies that perform FDI and offshoring are simultaneously also exporters. This status is common to many internationalized firms, especially if the process of internationalization is a sequential one which starts with exporting and then evolves in

Further work is necessary to understand the differences in productivity, if ever

The source of our data set are the 9th and 10th waves of Capitalia surveys covering the periods 2001-2003 and 2004-2006. The survey sample contains all Italian manufacturing firms with more than 500 employees and small and medium sized firms are selected through a stratified sample. In addition to the detailed qualitative information, the sample is complemented by annual balance sheets data for all the

K = fixed capital stock at the end of the period as the accounting value of net

VA = the balance sheet value added of firm deflated with the corresponding

L = total employment has been split between white and blue collars. The number of white collars is obtained by the difference between total employment and the

Below is the description of the variables used in the analysis

any, between FDI and vertical forms of sourcing abroad in the Italian context. Therefore, we expect our analysis act as a guidance to identify more precise impact

of different entry-modes on firm level productivity.

immobilization as reported in the balance sheet.

*DOI: http://dx.doi.org/10.5772/intechopen.95288*

the majority of geographical destinations.

more complex forms.

**A. Appendix**

**A.1 Description of variables**

firms included in the sample.

producer price index.

number of hand workers.

#### *Entry-Mode Selection and Firm's Productivity across Market Destinations: An Empirical… DOI: http://dx.doi.org/10.5772/intechopen.95288*

we find support to the HMY ranking only for FDI decisions in the high-income countries of the North but not when activities are located in Southern countries. For firms operating in low-income countries of the South the more productive firms are purely exporting firms. This means that more distant markets either in physical terms or in technological and institutional characteristics entail diversities in terms of costs and risks. Therefore, only firms with higher productivity may serve these countries. Overall, the productivity premia of FDI firms are higher for firms operating in high income countries, exporting firms are the best performers across the majority of geographical destinations.

The results of this work is likely to be helpful in the formulation of market entry strategies. Before proceeding with complex entry mode, managers need to evaluate costs and benefits of their moves as well as country risks relative to the home country. In terms of policy implications, the evidence of this work suggests that exporting brings with it potentially positive effects. When evaluating more complex forms of entry-modes managers should consider that they seem to be favorable only for locations in the North where firms have previous experience, the cultural distance is low and where they can find market similarities such as favorable conditions to increase their performance (knowledge infrastructure and availability of qualified personnel). Then, the indication is that for Italy, export-enhancing public policy should promote exporting to all destinations especially considering small businesses, which are the majority in the industrial structure of the country.

By concluding, some caution must be exercised in generalizing the outcomes of this work. A limitation of this work is the small number of firms that perform FDI and offshoring with respect to the number of firms that perform exporting. While it is possible to isolate pure exporters, this cannot be done for the other entry-modes: companies that perform FDI and offshoring are simultaneously also exporters. This status is common to many internationalized firms, especially if the process of internationalization is a sequential one which starts with exporting and then evolves in more complex forms.

Further work is necessary to understand the differences in productivity, if ever any, between FDI and vertical forms of sourcing abroad in the Italian context. Therefore, we expect our analysis act as a guidance to identify more precise impact of different entry-modes on firm level productivity.
