0
Nationalities of board members and past M&A flow
A gravity model approach
MSc Thesis 2012-2013
Anne Spanjer S2184907 g.a.spanjer@student.rug.nl
University of Groningen Faculty of Economics and Business
Supervisor: Dr. P. Rao Sahib Co-assessor: Dr. R.K.J. Maseland
1 ABSTRACT
This research explores the impact past M&A flow in determining the share of foreign board members in a country. It is expected that a larger past M&A flow from a home country to a host country, increases the share of nationals from the host country on boards in the home country. To be able to answer this question the gravity model is used and applied in a new setting. The data on the nationalities of board members on the boards of MNCs in 15 European countries are used. The results suggest that the effect of past M&A flow on the nationalities of board members is positive. This finding confirms the predictions, as it implies that past M&A flow is a determinant of the share of foreign board members.
2 1. INTRODUCTION
The gravity model is a popular tool used to estimate and explain trade between countries. The smaller the distance between countries, the more trade there will be between these countries. The reason for this to happen is that if countries share the same cultural background and the same language, this reduces the costs of doing business (Anderson, 1979). Examples of such costs are time costs, transport costs, transaction costs, and etcetera. More recently, the gravity model has been applied to the good consumption over the internet (Blum & Goldfarb, 2006). Samy & Deheija (2011) use the gravity model to estimate the effect of labor standards on bilateral trade. This paper is using the gravity model to investigate the effect of past merger and acquisition (M&A) flow on the nationality diversity of board members within countries.
Research has found that nationality diversity of board members of multinational corporations (MNCs) has been increasing. The share of foreign board members differ per company and country. In Spain only 2.5 per cent of the board members are foreigners, while in Luxembourg 75 per cent of the board members are foreigners (van Veen & Marsman, 2008). The firms with foreign board members, are mainly MNCs. Since a few decades engaging in M&A has become a major strategic tool for growth of multinationals. MNCs use M&As to expand business to other countries. 78 per cent of all foreign direct investment (FDI) are cross-border M&As (Brakman et al., 2007).
If MNCs become more global, this is reflected in their boards. After an M&A, the board members of the acquired firm and the board members of the acquiring firm are expected to become a member on the new board. Thus, it is expected that the nationality diversity of the board increase after an M&A. Staples (2007) shows that nationality diversity of the board increases after cross-border M&A. With every M&A, the nationality diversity of the MNC should increase. The gravity model will be helpful to explain the effect of M&A history on the nationality diversity of boards.
3 foreign board members in a country. The following research question is investigated in this paper:
Do M&As between two countries increase the exchange of board members between the two countries?
4 2. LITERATURE REVIEW
2.1 The gravity model
The gravity model originates from the law of gravity developed by Newton. This law states that two objects in universe are attracted to each other. The force with which they are attracted to each other is proportional to the product of the masses of these objects and divided by the squared distance between these objects. The gravity model for trade was developed in a similar line of reasoning (Tinbergen, 1962). Just like in Newton’s law of gravity the volume of trade between two countries is proportional to the product of the GDP of these two countries. The volume of trade is inversely proportional to the distance between these two countries. The volume of trade increases as the joint income of the two countries increases (Baldwin & Taglioni, 2006). Leamer and Stern (1970) based this on their so called ‘potluck assumption’. According to this assumption all products produced by country A are thrown into a large pot. Country B takes its consumption out of this pot. The consumption of Country B is proportional to country B’s GDP. Country A’s production and country B’s consumption equals their joint income. The greater the distance between countries, the less trade there will be between these two countries and the shorter the distance between the two countries, the more trade there will be between these two countries (Anderson, 1979). The costs of doing business decreases as the distance decreases, and increases as distance increases. Examples of such costs are time costs, transport costs, transaction costs, and etcetera. Thus, when these costs decrease there are less barriers to trade.
The gravity model has become a popular tool to explain trade. Bergstrand (1985) is one of the first to empirically test the gravity equation for international trade. Nowadays, scholars who use the gravity model include more explanatory variables to find additional determinants of trade. Examples of this are Rose (2000), Head and Mayer (2000) and Glick and Rose (2002). Rose (2000) estimates the effect of multilateral trade agreements on trade using the gravity model. Head and Mayer (2000) evaluate the effect of borders across industries on trade within the European Union. Glick and Rose (2002) study the effect of currency unions on trade using time series data.
5 good consumption on the internet. The basic theory behind this argues that when consumers buy purely digital goods, the physical distance should matter less. However, results indicate that some taste-dependent products are sensitive for distance, while others, such as software, are not affected by the distance. This implies that trade cost cannot fully explain the effect of distance, but that distance also is a proxy for taste and culture. Picci (2010) uses a gravity model to estimate the internationalization of inventive activity using patent data. The inventive masses, which is the country’s patent portfolio, of countries and distance are used to explain the international collaboration with regard to inventive activity. Picci (2010) argues that the larger the product of the inventive masses of two countries, the more collaboration there would be between these two countries with regard to inventive activity. However, the larger the distance between the two countries, the less collaboration there would be between the two countries. The results show that distance has a negative effect on the collaboration with regard to inventive activity. The results for the inventive masses of the countries are inconclusive.
Karemera et al. (2000), Clark et al. (2007) and Lewer and van den Berg (2008) have used a gravity model for immigration. In the gravity model for immigration, immigration is positively determined by the sizes of the population of the two countries, because the larger the population the larger the chance that somebody will migrate and the larger the labor market for immigrants. Immigration is also positively influenced by the ratio of destination to source country per capita income, because if the ratio of per capita income increases, the labor market of the destination country becomes more attractive for immigrants. Immigration is negatively influenced by the distance between the two countries (Lewer and van den Berg, 2008).
2.2 The decision to migrate
6 Migrants are connected to networks in their country of origin and in the country of residency (Gaston & Nelson, 2013). These networks exists in all sort of forms. These networks provide the migrants with information and resources. Gaston and Nelson (2013) argue that networks play a crucial role in the decision for individuals to migrate, and in the location decision for migrants. This is because these networks decrease the cost of migrating from one country to another.
Javorcik et al. (2011) mention that FDI can promote migration. Since largest part of FDI consists of M&A (Brakman et al., 2007), it can be assumed that cross-border M&A activity is a good predictor for a firm’s internationalization. The underlying reason for FDI, and thus also cross-border M&A activity, promoting migration is that working for a foreign firm makes it easier for employees to be transferred to the firm’s subsidiaries in other countries or to the firm’s headquarter. Next to this employees gain experience by working for an MNC. This increases the employee’s network and, therefore, lowers the barriers to migrate. These effects are likely to be higher for high skilled employees than for low skilled employees.
It is therefore expected that a larger past M&A flow from the home country to the host country increases the share of nationals of the host country on boards in the home country.
2.3 Research from management
7 has international experience and the access to board member’s networks helps the firm to overcome information barriers when the firm decides to go abroad.
It is suggested by Gillies and Dickinson (1999) and by Heiltjes et al. (2003) that more internationalized firms have more internationalized boards. Heiltjes et al. (2003) found that the number of foreign board members has increased over time in Sweden and the Netherlands. They address several underlying reasons for having foreign board members. If there are more foreign members working in the firm, the chance of having foreign board members increases. Another factor influencing the number of foreigners on boards is the internalization of ownership. If a firm has many investors in a particular country, the likelihood of having board members of the same country increases (Heiltjes et al., 2003).
Gillies and Dickinson (1999) investigates whether the globalization of firms drives the globalization of boards. Their results show that about 36 per cent of the MNCs had one or more foreign board members. Staples (2007) has done a follow up study of Gillies and Dickinson (1999). The study of Gillies and Dickinson (1999) is used as a benchmark by Staples (2007). His results indicate that the share of firms with one or more foreign board members has increased with 75% in the period 1993-2005. Cross-border M&As increase the nationality diversity on boards, because the members of the acquired firm and the acquiring firm become members of the new board. Many MNCs have foreign board members. However, on only 9 per cent of the boards of MNCs, more than half of the board members are foreigners.
8 This suggestion is confirmed by Staples (2007). He argues that the most common cause of an increase in the share of foreign board members is cross-border M&A.
9 3. METHODOLOGY AND DATA
3.1 The model
The gravity model for trade uses the product of the GDP of the two countries that are trading, the distance between these two countries to estimate the volume of trade between these two countries. The gravity model I am using predicts the share of nationals from the host country on board in the home country. The home country is defined as the country in which the headquarter of the acquirer is located. The host country is defined as the country of origin of the foreigner who is on the board in the home country. It is expected that a larger past M&As flow from the home country to the host country increases the board member flow from the host country to the home country. Board members include both the board members of the supervisory board and the board members of the board of directors. Past M&A flow is measured as the sum of the value of all M&A of the home country in the host country. A merger occurs when the businesses of two or more firms are brought together as one, neither of the firms are the acquiring firm or the acquired firm. An acquisition occurs when one firm acquires from another firm either a controlling interest in the firm’s stocks or a business operation and its assets. This involves an acquiring firm and an acquired firm. (Coyle, 2000: 2,4).
In my specification of the gravity model I will use the following independent variables, population of the home country, population of the host country, the ratio of host to home country per capita income, distance, past M&A flow. The control variables used in my specification of the gravity model are three dummy variables that capture the economical, geographical and historical proximity of the host and home country, and a dummy variable for whether the home country has a coordinated market economy.
10 be higher as well. The distance between the host and home country is included because the larger the distance the higher the barriers to move. And therefore the less likely it will be that a national of the host country will be on a board in the home country (Lewer and Van Den Berg, 2008). A larger past M&A flow from the home country to the host country is expected to increase the share of nationals of the host country on boards in the home country. This is because, employees that work for a subsidiary of a foreign MNC enlarge their networks and these employees gain experience. The enlarged network and the experience gained by working for a foreign MNC, decrease the barriers of migrating to the home country (Javorcik et al., 2011; Gaston & Nelson, 2013). Van Veen and Elbertsen (2008) have suggested that the type of market economy plays an important part in the ease with which a foreigner can get a board position in the home country. In Liberalized Market Economies (LMEs) the share of foreigners on boards is higher than in Coordinated Market Economies (CMEs). This is because for outsiders it is easier to get a board position in LMEs, while for outsiders it is more difficult to get a board position in CMEs (van Veen & Elbertsen, 2008). Three dummy variables are included to measure the host to home proximity, namely a dummy variable for if the home and host country have a contingent border, a dummy variable for if the countries have a common language and a dummy variable for if the home and host countries have a colonial link. If the home and host countries have a contingent border, a common language, and a colonial link, this increases the cultural, geographical and economical proximity of the home and host country (Cerutti et al., 2007). This increases the probability of a national of the host country to be on a board in the home country.
The following model is estimated;
( ) ( ) ( )
(1)
In this model country i is considered to be the home country and country j is considered to be the host country, and i≠j.
Yijt is the share of nationals of country j on boards located in county i in year t. POPi is the population of country i in year t
POPj is the population of country j in year t
11 distij is the distance between country i and country j
MAij, t-5 is the value of M&As going from county i to country j in year t-5
Market economyi is a dummy variable for the type of market economy of the host country
Home-host proximityij is a matrix consisting of variables that capture the economical, geographical and historical proximity of the home and host country
According to Mátyás (1997) the home country effect, the host country effect and the time effect should be taken into account. A restricted model assumes the home country, host country and time effects to be equal to zero. Therefore, an unrestricted model should be used, because this model considers the home country effect, the host country effect and the time effect. Since I am not completely sure what the appropriate model is to fit the data, I have conducted the appropriate tests to check whether the random effect model is the right model. An Breusch-Pagan Lagrange Multiplier test showed that random effects are present. The Hausman test indicated that the random effects model is the appropriate model to fit the data.
3.2 Selection issues
12 0.004. Adding a small constant to the dependent variable, would have a huge impact. This is because, this small constant is relatively large in comparison to the values of the dependent variable.
Other scholars have chosen not to use the natural logarithm of the dependent variable, but to use a Heckit model (Helpman et al., 2008), a Tobit model (Rose, 2000) or a Poisson model (Siliverstovs & Schumacher, 2009) to estimate the gravity model. When estimating the Poisson model the assumption of equidispersion should hold. This assumption holds when the conditional mean and conditional variance are equal (Kohl, 2012). However, the assumption of equidispersion is often violated. Santos Silva and Tenreyro (2006) argue that the estimates of the gravity model when using the Tobit model will be inconsistent. The degree of the inconsistency depends on the sample used. The inconsistency of the estimates is more sever if the assumptions of homoskedaticity and normality are violated (Santos Silva & Tenreyro, 2006). This study uses the Heckit model to deal with the sample selection bias. Next to the Heckit model, the Tobit model is estimated as well. This is done to check for the robustness of the estimates. The Heckit model is chosen, because it determines the probability that a board member will migrate and the size with which board members migrate separately. When estimating equation 1, the observations in which there are no nationals of the host country on boards in the home country are omitted from the sample. Ignoring these ‘zeros’ would mean that the estimated model only considers board members that have already migrated. The board members that have not migrated are ignored, while these board members might have explicitly chosen not to migrate. Motives for not migrating could be language barriers, large distances between countries, and etcetera. Therefore, a sample selection bias could exist as the sample is not a random sample. The solution to the sample selection bias consists of two stages developed by Heckman. The first stage involves a probit model to estimate whether there are nationals of country j on boards country i. Equation (2), reflects the conditions for which the POBijt is one and for which POBijt is zero. POB is the presence of a national of country i on the board in country j.
{
13 ( ) (
) (3)
Equation (3) reflects the probit model for the presence of foreign board members. Beine et al. (2011) argue that the first stage equation should only contain variables that increase the likelihood of a national of the host country to be on a board in the home country. Besides this, it should not contain variables that affect the number of nationals of the host country being on boards in the home country. The included variables all increase the probability that a national of the host country will be on a board in the home country.
After equation 3 is estimated the inverse Mills ratio is calculated. The inverse Mills ratio corrects for the fact that the share of nationals of host country on boards in home country is not random (Hill et al., 2012).
The second stage estimates the model described above in equation (1) plus one extra independent variable, namely the ‘inverse Mills ratio’, which is created in the first stage. This is reflected in equation (4).
( ) ( ) ( )
( )
In which IMR is the inverse Mills ratio.
3.3 Data
3.3.1 Dependent variable
14 in the sample when there is a least one board member with the nationality of the host country on a board in the home country. For example, if there is one Malaysian board member on a board in the United Kingdom, then Malaysia is added to the sample as a host country even though the
share of Malaysian board members on the boards in the other home countries is zero. Hence, there is at least one board member with the nationality of the included host countries on the boards in at least one of the home countries. There are 855 (=57x15) country pairs in total and 2565 (=855x3) observations in total. All board members of the MNCs are included in the sample.
Home countries
Austria France The Netherlands Belgium Greece Portugal
Denmark Ireland Spain
Germany Italy Sweden
Finland Luxembourg United Kingdom
Table 1. Home countries
Host countries
Afghanistan Germany Ireland Norway
United Arab Emirates Denmark Iran New Zealand
Argentina Algeria Israel Poland
Australia Egypt Italy Portugal
Austria Spain Japan Russia
Azerbaijan Estonia South Korea Singapore
Belgium Ethiopia Lebanon Sweden
Bosnia and Herzegovina Finland Libya Tunisia
Brazil France Lithuania Turkey
Canada United Kingdom Luxembourg Uganda
Switzerland Greece Morocco USA
China Hong Kong Mexico South Africa
Colombia Croatia Malaysia Zimbabwe
Cyprus Hungary Namibia
Czech Republic India The Netherlands
15 In case of a two-tier board, board members of both the supervisory board and the executive board are included in the sample. The data on the nationalities of the board members are retrieved from the annual statements and their websites. If this was not available the nationalities of the board members are retrieved from other websites such as forbes.com and Zoom information. In case the annual reports, the firm’s websites, or other websites do not give the nationalities of the board members, the nationality is retrieved using in which country the board member was raised and its early education.
The dependent variable is measured by the share of nationals of the host country on boards located in the home country. This is calculated by the number of nationals of the host country on boards in the home country divided by the total number of board members in the home country. To calculate the total number of board members in the home country only the board members of the companies included in the sample are used.
3.3.2 Independent variables
Population. The population of a country is measured as the total population of a country in the
period 2005-2007. The data on population is retrieved from the World Development Indicators database (Worldbank, 2013).
Ratio of host to home country per capita income. The ratio of host to home country per capita
income is measured using the GDP per capita income in the period 2005-2007 that is corrected for the purchasing power parity. The data on GDP per capita is retrieved from the World Development Indicators database (Worldbank, 2013).
Distance. The distance is measured in kilometers. The data on distance is retrieved from the
geodesic distances database of the CEPII. Mayer and Zignago (2011) have calculated this using the great circle formula.
Past M&A flow. Past M&A flow is measured as the sum of the value of all M&A of the home
16 M&As occurring in the same year with the same host and home country are summed up. This is measured as the values in US dollars. A timespan of five years between the past M&A flow and the share of foreign board members in the home country is chosen, because it is expected that it will take some time before a national of the host country settles on a board in the home country (Staples, 2007; Harvey & Buckley, 1997). The data on past M&A flow is retrieved from Zephyr.
3.3.3 Control variables
Market economy These include a dummy variable for Liberal Market Economies (LMEs) and
Coordinated Market Economies (CMEs). The dummy variable is one in case of an LME and 0 in case of a CME. The control variables have been chosen following the study of van Veen & Elbertsen (2008). It is suggested by van Veen & Elbertsen (2008) that in LMEs the share of foreigners on boards is higher than in CMEs. For outsiders it is easier to get on a board in LMEs, while for outsiders it is more difficult to get on a board in CMEs. The scores of Hall and Gingerich (2004) on LMEs and CMEs are used to measure whether a country has an LME or a CME.
Home-host country proximity. This includes three dummy variables that capture whether the
culturally, and geographically close to each other and whether the two countries have colonial links. The first is a dummy variable for if the home and host country share the same language. It is 1 if the countries share the same language. The second is a dummy variable for if the home country and host country share a contingent border. It is 1 if the countries share a contingent border. The third is a dummy variable for if the home and host country have a colonial link. It is 1 if the countries have a colonial link. The data for these variables is retrieved from the geodesic distances database of the CEPII.
3.4 Descriptive statistics
17 on boards of home country. The data on population of the home country is very spread. On average the population of host countries is larger than the population of home countries. The data on population of the host country is also dispersed. This is caused by the countries, such as India, China and the United States which have huge populations that are much larger than the populations of other host countries. Of the home countries 47% have a coordinated market economy and 53% have a liberalized market economy. Only 8.6% of the country pairs share a common language. The average past M&A flow between the country pairs is 462 million US $.
Variable Obs Mean Std. Dev. Min Max
Yijt 2,565 0.004 0.014 0 0.160
Populationi (x1000) 2,565 25,632 26,209 465 82,689
Population j (x1000) 2,565 786,230 223,462 465 1,321,851
Relyij 2,565 0.670 0.520 0 3.350
Distanceij 2,565 4,537.84 4,138.04 80.98 19,586.18
Coordinated Market Economy 2,565 0.470 0.500 0 1
Common language 2,565 0.086 0.280 0 1
Contingent border 2,565 0.051 0.220 0 1
Colonial link 2,565 0.050 0.220 0 1
Deal value (x1000) 2,565 462,187.70 5,166,166 0 207,000,000
18 4. DATA ANALYSIS
4.1 Estimation results
Table 4 shows the regression results. Panel 1 shows the estimates for the random effects model without taking the selection bias into account. Panel 2 and 3 show the estimates for the two step Heckman regression, in which panel 2 shows the selection equation and panel 3 the random effects model including the inverse Mills ratio. Panel 4 shows the estimates of the model estimated with the Tobit model.
When comparing panel 1 with panel 3 there are some clear differences. The sign of the estimate of the population of the home country changed from negative to positive. The coefficients of almost all variables increased. There is one exception, namely the M&A flow, this coefficient decreased from 0.053 to 0.049. In panel 1 these zero observations for the share of nationals of host country on board in home country are excluded from the analysis. Thus panel 1 only considers board members from the host country that have already migrated to the home country. By ignoring the zero observations in important information is lost about the circumstances for which board members decide not to immigrate to the home country. In panel 3 an extra variable is added to control for the fact that the sample only includes board members from the host country that have already migrated to the home country, namely the inverse Mills ratio. The inverse Mills ratio measures the sample selection effect and accounts for the non-randomness of the sample. The inverse Mills ratio is statistically significant and has a positive sign. This indicates that there is indeed a selection bias. The significance of the inverse Mills ratio indicates that if there are no nationals of a host country on the boards of a home country, this happens for a reason. As mentioned earlier possible causes could be, language barriers, large distances between the two countries or any other barriers that makes the cost of migrating higher than the personal gains (Clark et al., 2007).
The model without taking the selection bias into account explains 24% of the relative variance in the share of nationals from host country on boards in home country. The model including the inverse Mills ratio explains the relative variance in the share of nationals from host country on boards in home country somewhat better.
19
Heckman Tobit
(1) (2) (3) (4)
Yijt Pr(POBijt) Yijt Yijt
Populationi -0.002*** 0.349** 0.001* -0.001 (-4.38) (2.71) (2.46) (-1.85) Populationj 0.003*** 1.455*** 0.012*** 0.011*** (9.49) (8.30) (9.79) (18.60) Distanceij -0.002*** -1.243*** -0.010*** -0.009*** (-4.64) (-6.77) (-9.08) (-10.44) Relyij 0.004*** 2.969*** 0.024*** 0.022*** (5.16) (7.77) (9.09) (13.02) M&Aij,t-5 0.053** 0.049** 0.408*** (3.05) (2.87) (3.99)
Coordinated Market Economy -0.000 -0.000 0.004**
(-0.41) (-0.37) (2.89) Common language 0.005** 1.545 0.015*** 0.014*** (3.24) (1.74) (7.33) (5.47) Contingent border 0.013*** 6.305*** 0.020*** 0.016*** (6.14) (8.41) (8.71) (5.20) Colonial link 0.005* 3.249 0.021*** 0.018*** (2.28) (1.68) (7.34) (5.80) IMR 0.007*** (7.86) Constant 0.004 -14.570*** -0.096*** -0.071*** (0.81) (-5.71) (-7.06) (-7.56) Observations 2565 2565 2565 2565 R2 0.24 0.29 LR X2 866.76*** t statistics in parentheses * p<0.05, ** p<0.01, *** p<0.001
Table 4. Regression results
20 The size of the population of the home country has a positive effect on the share of nationals of the host country on boards in the home country. If the host countries have a larger population, they have more nationals on boards of the home country. However, host countries that are further away from the home countries, have less nationals on boards of the home country. The ratio of host to home per capita income has a positive effect. This is the opposite of what was expected. The coefficient for the dummy variable for whether the home country has a coordinated market economy is not statistically significant. When the home country and the host country share a common language the share of the nationals from the home countries on boards in the home country increases. When the home country and the host country share a contingent border, the share increases as well. So, when the home and host country are geographically and culturally close to each other, the share of nationals of the host country on boards in the home country increases. A colonial link between the host country and the home country has a positive effect on the share. The M&A flow has a significant positive effect on the share of the nationals from the home countries on boards in the home country. For every one trillion US $ increase in the M&A flow of five years before, the share of nationals of the host country on boards in the home country increases by 4.9%.
21 4.2 Robustness checks
4.2.1. Multicollinearity
It could be argued that population and distance influence the annual value of M&A deals between two countries. A country with a larger population attracts more FDI, and therefore also more M&As (Buch & Delong, 2004). The population of a country changes over time. Thus, it could be argued the population of today does not affect the M&A flow of five years before. On the other hand, it is highly unlikely that the population of 2005-2007 is hugely different from the population of five years before. Foccareli and Pozzolo (2001) test the effect of population on the size of M&A flow. Their results showed that population has an insignificant effect on the size of M&A flow. Distance, however, does not change over time. If the distance between the host and home country is smaller, the host country receives more M&As (Rose, 2000; Erel et al., 2013). Thus distance and past M&A flow can be correlated and population and past M&A flow can be correlated. Hence there might be multicollinearity. The pairwise correlations between the independent variables are given in table 5. None of the variables have a correlation of more than 0.500. It can therefore be assumed that there is no multicollinearity (Tabachnick, & Fidell, 1996).
22 1 2 3 4 5 6 7 8 9 10 1. Yijt 2. Populationi -0.086 (0.000) 3. Population j 0.200 -0.015 (0.000) (0.461) 4. Distanceij -0.208 0.001 0.284 (0.000) (0.978) (0.000) 5. Relyij 0.117 0.157 -0.355 -0.203 (0.000) (0.000) (0.000) (0.000) 6. Deal value 0.208 0.081 0.079 -0.049 0.072 (0.000) (0.000) (0.000) (0.013) (0.000)
7. Coordinated Market Economy -0.007 -0.082 0.001 -0.056 -0.054 -0.021
(0.719) (0.000) (0.952) (0.004) (0.006) (0.283) 8. Common language 0.228 -0.030 -0.038 -0.115 0.027 0.055 -0.105 (0.000) (0.133) (0.057) (0.000) (0.167) (0.006) (0.000) 9. Contingent border 0.337 0.015 -0.015 -0.431 0.158 0.012 0.058 0.343 (0.000) (0.002) (0.444) (0.000) (0.000) (0.546) (0.003) (0.000) 10. Colonial link 0.134 0.141 -0.008 -0.028 0.001 0.098 -0.097 0.405 0.142 (0.000) (0.000) (0.687) (0.154) (0.958) (0.000) (0.000) (0.000) (0.000) 11. IMR -0.356 -0.277 -0.476 0.400 -0.401 -0.144 0.045 -0.279 -0.438 .0.295 (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.021) (0.000) (0.000) (0.000) p-values in parentheses
23
4.2.2. The model with GDP instead of population
Trade between countries can be in goods, but countries can trade in services as well. Board member flow between countries can be seen as trade in services, because one country ‘sends’ board members to another country. Migration of board members is not the same as regular migration. The motives for board members to migrate differ from the motives of regular immigrants. This is because board members are often send to another country by the firm they are working for (Findlay et al., 1996), while a fair share of the regular immigrants migrates in the search of a better life (Javorcik et al., 2011). Board member migration happens on a smaller scale, because the labor market for board members is much smaller than the labor market for regular immigrants. Since board members migrate for different reasons, population may not be the best predictor for board member flow. When board member flow is seen as trade in services, then the home and host country’s GDP would be a better predictor of the board member flow. The data on GDP is retrieved from the World Development Indicators (Worldbank, 2013). Table 6 presents the regression results when GDP is used instead of population to estimate the share of nationals from the host country on boards in the home country. Panel 1 shows the estimates for the random effects model without taking the selection bias into account. Panel 2 and 3 show the estimates for the two step Heckman regression, in which panel 2 shows the selection equation and panel 3 the random effects model including the inverse Mills ratio.
The estimates of the probability of a national from the host country being on a board in the home country carry the same signs and are larger than in the model with population. The estimates of sharing a colonial link and sharing a common language now have positive and significant effects on the probability of a national from the host country being on a board in the home country. The ratio of host to home per capita income has a statistically insignificant effect on the share of nationals of the host country on boards in the home country. The estimate for whether the home country has a coordinated market economy is statistically not significant. All the estimates in panel 3 carry the expected signs. For every one trillion US $ increase in the M&A flow of five years before, the share of nationals of the host country on boards in the home country increases by 4.4%.
24 boards in the home country, while the model with population explains 29% of the relative variance.
Heckman
(1) (2) (3)
Yijt Pr(POBijt) Yijt
GDPi -0.001** 1.274*** 0.003*** (-2.77) (5.76) (7.16) GDPj 0.003*** 3.303*** 0.013*** (11.09) (15.19) (16.38) Distanceij -0.002*** -2.668*** -0.011*** (-4.06) (-10.20) (-13.75) Relyij -0.002* 0.173 0.000 (-1.98) (0.32) (-0.57) M&Aij,t-5 0.046** 0.044** (3.18) (3.06)
Coordinated Market Economy 0.000 0.000
(-0.47) (-0.25) Common language 0.006** 2.543* 0.012*** (3.26) (2.24) (7.49) Contingent border 0.013*** 6.762*** 0.007*** (5.72) (3.88) (3.42) Colonial link 0.004* 5.747** 0.019*** (2.03) (3.07) (8.39) IMR 0.004*** (13.55) Constant -0.026* -104.7*** -0.37*** (-2.28) (-11.14) (-13.55) Observations 2505 2505 2505 R2 0.26 0.39 t statistics in parentheses
25
5. DISCUSSION
Over the past 2 decades the share of foreign board members has been increasing on European boards. Both Staples (2007) and van Veen and Marsman (2008) argue that cross-border M&A is the most common cause for the increase in the share of foreign board members. An M&A of the home country in host country would decrease the costs and the barriers to migrate for a board member. The lowered barrier to migrate is caused by the experience a board member gains from working for an international firm. Besides this, the network of the board member working for a firm in the host country enlarges after this firm is acquired by a firm from the home country (Javorcik et al., 2011). The result found in this study partly supports the findings of other studies. The effect of past M&A flow is positive, but small. The result does confirm Javorcik et al.’s (2011) arguments, as they argue that past M&A flow increases the migration of board members from the host country to the home country. However, the effect of past M&A flow is not as large as Staples (2007) and van Veen and Marsman (2008) argued. A possible cause for this is that it takes longer than five years for a national from the host country to get on a board in the home country. Another explanation could be that nationals from the host country get high management positions in the home country after an M&A has occurred, but that these managers do not obtain a board position in the home country.
When viewing the board member flow as migration, and thus estimating the gravity model with population, the model explains only a small part of the relative variance in the share of nationals of the host country on board in the home country. If the sample is reduced to only the observation in which the past M&A flow is higher than zero, then the magnitudes of the estimates remain the same. However, the R2 is about 0.10 higher than the R2 of the model including all observations. Nonetheless, the goodness-of-fit is still lower than gravity models of immigration estimated by Lewer and van den Berg (2008), Clark et al. (2007) and Karemera et al. (2000).
26 host to home per capita income is positive, while a negative sign was expected. This indicates that per capita income is not an important determinant for board members to decide to migrate to a certain country. This contradicts with other findings in the literature. Lewer and van den Berg (2008) and Clark et al. (2007) find a negative effect for the ratio host to home per capita income. However, the gravity models estimated by Lewer and van den Berg (2008) and Clark et al. (2007) estimates regular immigration, whereas the model estimated in this study estimates board member flow. The wages of regular immigrants are much lower than the wages of board members. Therefore the higher wages in the home country may not be a motive for board members to migrate to another country, whereas the higher wages are a motive for regular immigrants to migrate.
27 6. CONCLUSION
The purpose of this paper is to examine the role of past M&A flow in determining the share of foreign board members in a country. It is expected that a larger past M&A flow from a home country to a host country, increases the share of nationals from the host country on boards in the home country. To be able to answer this question gravity model was used. The gravity model is applied in a new setting. The data on the nationalities of board members on boards of MNCs in 15 European countries are used. The Heckit estimator is used to take the selection bias into account.
The results suggest that the effect of past M&A flow on the nationalities of board members is positive, but small. This implies that past M&A flow is a determinant of the share of foreign board members. However, the size of the past M&A flow must be very large for the share of nationals of the host country on boards in home country increases. For every one trillion US $ increase in the M&A flow of five years before, the share of nationals of the host country on boards in the home country increases by 4.9%. Possible causes for this are that it takes more than five years before a national of the host country becomes a board member in the home country. Javorcik et al. (2011) suggests that after an M&A it is easier for high skilled employees to migrate from the host country to the home country, because of the enlarged networks and newly gained knowledge from working at an MNC. However it might be possible that these high skilled employees never become board members in the home countries.
29 7. REFERENCES
Anderson, J. E., 1979. A theoretical foundation for the Gravity Equation. American Economic Review 69, 106–16.
Athanassiou, N., & Nigh, D., 2002. The impact of the top management team's
international business experience on the firm's internationalization: Social networks at work. MIR: Management International Review 42(2), 157-181.
Ashby, W. R., 1956. An Introduction to Cybernetics. Champan & Hall, London. Baldwin, R., & Taglioni, D., 2006. Gravity for dummies and dummies for gravity
equations. National Bureau of Economic Research Working Paper 12516. Beine, M., Docquier, F., & Özden, Ç., 2011. Diasporas. Journal of Development
Economics 95(1), 30-41.
Bergstrand, J. H., 1985. The Gravity Equation in international trade: Some
microeconomic foundations and empirical evidence. The Review of Economics and Statistics 67, 474–481.
Blum, B.S., & Goldfarb, A., 2006. Does the internet defy the law of gravity? Journal of International Economics 70, 384-405.
Bosker, E. M., 2008. The empirical relevance of geographical economics. Utrecht University, Utrecht.
Brakman, S., Garretsen , H., & van Marrewijk, C., 2007. Cross-border Mergers and Acquisitions: The Facts as a Guide for International Economics. In Gregoriou, G., & Renneboog, L. (eds.). International Mergers and Acquisitions Activity Since 1990: Recent Research and Quantative Analysis. Academic Press/Elsevier, Cambridge MA, 23-49.
Buch, C. M., & DeLong, G., 2004. Cross-border bank mergers: What lures the rare animal? Journal of Banking & Finance 28(9), 2077-2102.
CEPII. Geodesic Distances. Retrieved from
http://www.cepii.fr/anglaisgraph/bdd/distances.htm. Accessed on April 6th, 2013. Cerutti, E., Dell’Ariccia, G., & Martinez Peria, M. S., 2007. How banks go abroad:
Branches or subsidiaries? Journal of Banking & Finance 31(6), 1669-1692.
30 Coyle, B., 2000. Mergers & acquisitions: corporate finance. Lessons Professional
Publishing, Chicago IL.
Erel, I., Liao, R.C., & Weisbach, M.S., 2012. Determinants of Cross-Border Mergers & Acquisitions. Journal of Finance 67(3), 1045-1082.
Evenett, S., 2004. The cross-border mergers and acquisitions wave of the late 1990s. In Balwin, R. E., & Winters, L.A. (eds.). Challenges to Globalization: Analyzing the Economics. University of Chicago Press, Chigago IL, 411-470.
Findlay, A. M., Li, F. L. N., Jowett, A. J., & Skeldon, R., 1996. Skilled international migration and the global city: a study of expatriates in Hong Kong. Transactions of the Institute of British Geographers 21(1), 49-61.
Focarelli, D., & Pozzolo, A. F., 2001. The patterns of cross-border bank mergers and shareholdings in OECD countries. Journal of Banking & Finance 25(12), 2305-2337. Gaston, N., & Nelson, D. R., 2013. Bridging trade theory and labour economics: the
effect of international migration. Journal of Economic Surveys 27(1), 98-139. Gillies, J. and Dickinson, M., 1999. The Governance of Transnational Firms: Some
preliminary hypotheses. Corporate Governance: An International Journal 7, 237–247. Glick, R., & Rose , A., 2002. Does a Currency Union Affect Trade? The Time Series
Evidence. European Economic Review 469 (6), 1125-1151.
Greve, P., Nielsen, S., & Ruigrok, W., 2009. Transcending borders with international top management teams: A study of European financial multinational corporations. European Management Journal 27(3), 213-224.
Hall, P., & Gingerich, D. W., 2004. Varieties of Capitalism and institutional complementarities in the Macroeconomy. Max-Plank-Institute fuer
Gesellschaftsforschung Discussion Paper 04/05.
Harvey, M. G., & Buckley, M. R., 1997. Managing inpatriates: Building a global core competency. Journal of World Business 32(1), 35-52.
Head, K., & Mayer, T., 2000. Non-Europe: the magnitude and causes of market fragmentation in the EU. Weltwirtschaftliches Archiv 136(2), 284-314.
31 Helpman, E., Melitz, M., & Rubinstein, Y., 2008. Estimating trade flows: Trading
partners and trading volumes. The Quarterly Journal of Economics 123(2), 441-487. Hill, R.C., Griffiths, W.E., & Lim, G.C., 2012. Principles of econometrics. 4th
edition. Wiley, Hoboken NJ.
IMF. Directions of Trade Statistics. Retrieved from:
http://www.imf.org/external/data.htm. Accessed on April 20th, 2013.
Javorcik, B. S., Özden, Ç., Spatareanu, M., & Neagu, C., 2011. Migrant networks and foreign direct investment. Journal of Development Economics 94(2), 231-241.
Kaczmarek, S., Kimino, S., & Pye, A., 2012. Antecedents of Board Composition: the Role of Nomination Committees. Corporate Governance: An International Review 20(5), 474-489.
Karemera, D., Oguledo, V. I., & Davis, B., 2000. A gravity model analysis of international migration to North America. Applied Economics 32(13), 1745-1755. Kohl, T., 2012. Trade agreements galore: Who, what, when, where, why, how and how
much? University of Groningen, Groningen.
Leamer, E. E., & Stern, R. M., 1970. Quantitative international economics. Allyn and Bacon, Boston.
Lewer, J. J., & Van den Berg, H., 2008. A gravity model of immigration. Economics Letters 99(1), 164-167.
Linders, G., & De Groot, H. L. F., 2006. Estimation of the Gravity Equation in the Presence of Zero Flows. Tinbergen Institute Discussion Paper No. 06-072/3. Available at SSRN: http://ssrn.com/abstract=924160. Accessed on February 19th, 2013.
Mahroum, S., 2000. Highly skilled globetrotters: mapping the international migration of human capital. R&D Management 30(1), 23-32.
Mátyás, L., 1997. Proper econometric specification of the Gravity Model. The World Economy 20, 363–369.
Mayer, T., & Zignago, S., 2011. Notes on CEPII’s distances measures (GeoDist). CEPII Working Paper 2011-25
32 Rose, A. K., 2000. One Money, One Market: Estimating the Effect of Common
Currencies on Trade. Economic Policy 30, 9-45.
Rose, A., 2004. Do we really know that the WTO increases trade? American Economic Review 94, 98–114.
Samy, Y., & Dehehejia V.H., 2011. Trade and labor standards in the European Union. The International Trade Journal 25(5), 581-618.
Silva, J. S., & Tenreyro, S., 2006. The log of gravity. The Review of Economics and Statistics 88(4), 641-658.
Siliverstovs, B., & Schumacher, D., 2009. Estimating gravity equations: to log or not to log? Empirical Economics 36(3), 645-669.
Staples, C.L., 2007. Board globalization in the world’s largest TNCs 1993-2005. Coporate Governance 15(2), 311-321.
Tabachnick, B.G., & Fidel, L.S., 1996. Using Multivariate Statistics. HarperCollins, New York.
Tinbergen, J., 1962. Shaping the world economy; suggestions for an international economic policy. Twentieth Century Fund, New York.
Van Veen, K. & Elbertsen, J., 2008. Governance regimes and nationality diversity in corporate boards. Corporate Governance 16(5), 386-399.
Van Veen, K. & Marsman, I., 2008. How international are executive boards of European MNCs? European Management Journal 26, 188-198.
Wooldridge, J. M., 2002. Econometric Analysis Cross Section Panel. MIT press, Cambridge MA.
World Bank. World Development Indicators. Retrieved from:
http://data.worldbank.org/data-catalog/world-development-indicators. Accessed on April 6th, 2013
