Three regressions have been carried out in order to test the first main hypothesis, being a basic augmented Gravity regression without the immigrant stock as explanatory variable (no. 1 and 4), a basic equation with the inclusion of immigrant stock and (no. 2 and 5) and an extended equation with three additional explanatory factors pertaining to institutional quality in the partner country (no. 3 and 6). Each of these regressions has been conducted for both export and import.40
Table 4 The effect of immigrants on Danish international trade – OLS estimation.
Export Import No. (1) (2) (3) (4) (5) (6) Immigrants 0,129*** 0,079*** 0,328*** 0,311*** (0,016) (0,021) (0,039) (0,052) GDP Partner 0,861*** 0,833*** 0,719*** 1,116*** 1,035*** 0,869*** (0,021) (0,021) (0,025) (0,052) (0,052) (0,079) Distance -0,919*** -0,790*** -0,752*** -0,164 0,187 -0,116 (0,055) (0,056) (0,060) (0,135) (0,140) (0,148) Population Partner 0,018 -0,052** 0,058* 0,237*** 0,077 0,291*** (0,025) (0,024) (0,034) (0,062) (0,064) (0,098) European Union 0,360*** 0,368*** 0,211** 0,678*** 0,670*** 0,221 (0,077) (0,077) (0,085) (0,155) (0,159) (0,178) WTO member -0,028 0,023 0,061 0,381** 0,490*** 1,004*** (0,065) (0,066) (0,078) (0,148) (0,148) (0,212) Exchange rate -0,006 -0,008 0,013 0,091*** 0,065*** 0,067*** (0,008) (0,008) (0,009) (0,019) (0,019) (0,022) Financial integration 0,280*** 0,274*** 0,394*** 0,235** 0,214* 0,162 (0,040) (0,039) (0,102) (0,111) (0,112) (0,228) Military conflict -0,291*** -0,316*** -0,297*** 0,136 0,071 0,071 (0,049) (0,046) (0,045) (0,126) (0,118) (0,111) RTA -0,221*** -0,371*** -0,305*** 0,373** -0,035 -0,448** (0,073) (0,075) (0,068) (0,168) (0,168) (0,169) Landlocked -0,538*** -0,386*** -0,636*** -0,174 0,209 0,069 (0,051) (0,054) (0,065) (0,131) (0,139) (0,171) Trade openness -0,001 0,025 -0,085 0,586*** 0,631*** 0,986*** (0,067) (0,067) (0,082) (0,169) (0,168) (0,183) Linguistic distance -0,289*** -0,206*** -0,111*** -0,350*** -0,136 0,042 (0,040) (0,040) (0,051) (0,101) (0,101) (0,121) Corruption -0,336*** -0,655*** (0,052) (0,108)
Business Freedom 0,196* 0,413 (0,117) (0,332) Trade policy 0,202 -0,999*** (0,129) (0,296) Constant 4,609*** 4,440*** 4,236*** -12,037*** -12,602*** -7,812*** (0,549) (0,527) (0,841) (1,373) (1,338) (2,067) Time fixed effects
Yes Yes Yes Yes Yes Yes
Region fixed effects
Yes Yes Yes Yes Yes Yes
N 1774 1739 1046 1711 1677 1044
R2 0,909 0,914 0,943 0,773 0,785 0,836
Note: Robust standard errors are placed in brackets underneath the estimated coefficient. ***/**/* signals the level of significance that is either 1/5/10%.
Overall the number of significant estimates is higher for export than for import, which is also reflected in the amount of the total variation explained in import and export – R2.
This might cohere with the methodical choice of not including explanatory variables regarding Denmark in the specification that are important in a foreign country point of view when exporting, e.g. Danish GDP and size of the Danish population. The regressions conducted unveil to a large extend the expected signs as argued in the data description but also some interesting and deviating outcomes have been estimated. The variables trading partner’s GDP, EU membership, and financial integration are significant and have a positive effect on the bilateral aggregated trade level. Comparing the estimates for GDP with the other coefficients in the model this variable has a very high magnitude. The opposite is found for linguistic distance, which has a significant and negative impact independent of whether import or export are regarded, this is not the case for other explanatory variables in the regressions. This establishes that the more distant the official language in the trading partner nation is from English, has a negative influence on the bilateral trade between Denmark and the specific country. The variables distance, military conflict, and landlocked are all significant and negative but only for export. These impacts are influenced by the Danish trade pattern and the type of goods that are mainly exported – differentiated goods. The fact that if the trade partner has a regional trade agreement (RTA) with Denmark beyond EU has a negative impact on export is closely related to the number and type of countries that Denmark have trade regional agreements with. These are mainly nations that are economically dissimilar compared to Denmark.41 The factors WTO membership, exchange rate, trade
41
In period concerned then in total Denmark has a RTA with 26 countries. Among them are countries as Syria, Lebanon, and Morocco.
openness are all significant and positive but only for Danish import. The insignificance and signs of some of the outcomes whether import or export is considered are more or less surprising. This especially concerns distance, WTO membership, military conflict,
RTA, landlocked, and trade openness.
The estimated results from immigration on export and import respectively show a positive and statistically significant influence on bilateral trade relationships between Denmark and the immigrant origin country. For both export and import the coefficient of the immigrant stock variable is positive and significant with estimated elasticities on
0,129% and 0,328% for export and import, respectively. In total, the outcomes clearly suggest and confirm the hypothesis that there is a positive relationship between the number of immigrants and the level of Danish bilateral trade. When comparing the estimates from immigration with the effect of traditional trade policy, here represented by membership of WTO and EU, it is reasonable to infer that immigration is an important aspect when explaining the level of Danish international trade. Although, there is evidence of a statistical significant influence from immigration the explanatory power that the inclusion of the immigrant stock brings to the specified model is not of greater magnitude. Comparing the total explained variance in the models – R2 - before
and after the inclusion of the immigrant stock as an explanatory variable only displays a minor rise. With that being said, the specifications show their relevance by explaining a very significant share of the complete variance both within export and import.
The third and extended estimation includes three variables to control for institutional quality in the trading partner countries; corruption, business freedom and trade policy. In these estimations the strong effect from immigration consists with a high level of significance. When comparing the coefficient with the previous estimation the effect from immigration is diminished a little. Considering the three extra variables corruption is significant for both import and export whereas an advantageous trade policy in a partner nation only concerns import. A favorable business climate is positive and significant for export. The magnitudes of the estimates exhibit their large impact on trade, which particular regards import. When comparing the significant measures of institutional quality and coefficients with the ditto WTO- and EU membership their relevance and importance in describing trade relations are underlined. With WTO membership and trade policy being insignificant, which is almost the case for business freedom in relation to Danish export as well; a natural deduction is that Danish export is
independent of the business conditions. The pivotal point is the level of corruption in a trading partner country and how it functions. Surprisingly, an open trade policy has a negative influence on import, which intuitively does not make sense. The factor should be viewed from a foreign point of view that is a more favorable trade policy in the origin countries is thought to have a positive influence of the nation’s level of export. Nevertheless, this cannot be confirmed in the case of Denmark. A comparable study, (Hatzigeorgiou, 2010b), uses the same three variables of institutional quality but in the case of Sweden and also finds a negative estimate though being insignificant.
8.1.1 Magnitude and importance of immigrants
It is noticed that import is influenced more by the immigration than export is. The coefficient for import is approximately 150 % greater than for export, which can be explained by several reasons.42 First, the explanation often given and presented in the theory part is that immigrants have a preference for familiar goods from their country of origin. The unavailability of these products or satisfactory substitutes can potentially increase the import from the source country. The magnitude of the transplanted bias effect or inherent preference effect among immigrants is however dependent on the size of the immigrant share of the total population in Denmark. Second, it might be the case that foreign-born people residing in Denmark are in a better position to lower or eliminate information barriers that are more important for import than for export. Moreover, the distinct resources and knowledge that immigrants possess with regards to their origin country might benefit import companies more than exporting firms. Third, the influence on the number of traded varieties – extensive margin – of import is greater than of the corresponding impact on export. In other words this intimates that the presence of foreign-born people in Denmark increases the likelihood of establishing more importing firms relative to exporting. Finally, it is reasonable to assume that in a country like Denmark it is easier to start of up an import business than to do export businesses to foreign markets where immigrants have their contacts and competences. The business environment and conditions in Denmark are simply better to operate in.43
42
The number is based on: ((0,328/0,129)-1)*100 = 154%.
43
The numbers regarding institutional quality in Denmark in 2008 are the as follow. Corruption: 9,3 - lowest level in the world together with Sweden and New Zealand. Business freedom: 99,9 –highest score together with New Zealand. Trade policy: 86,0 - top-10 score.
In total, this confirms the stated hypothesis that the effect from immigration is stronger regarding import than export.