8 Discussion
Results from our benchmark specification strongly suggest that the more migrants New Zealand receives from a given country, the more likely it is that New Zealand exports to that country, controlling for plausible confounding factors such as GDP, language, population, and distance. Similarly, the results strongly suggest that if New Zealand does export to a country, then an increase in migrants from that country is associated with an increase in exports. Results for imports are slightly different. There is some support for the idea that more migrants are associated with a greater probability of importing, and strong support for the idea that more migration is associated with higher imports. Taken at face value the results imply that if New Zealand does import from a country, a 1% increase in migrants is associated with a 0.15% increase in imports.
Following previous studies, we attribute any relationship between migration and exports to the ability of migrants to reduce transaction costs. We attribute the relationship between migration and imports to some unknown mix of transaction cost effects and preferences by migrants for goods from their country of origin. In most of our specifications it appears as though the relationship between migration and trade is strongest for imports. This suggests that the preferences of migrants may well be important, but could also be due to the goods that New Zealand imports being more differentiated in general than those New Zealand exports.
Using international visitor numbers as a proxy for earnings, we use the same specification to test for a relationship between migration and tourism “exports”. The relationship appears to be strong. However, some of this apparent strength is probably an artefact of some short-term visitors being recorded as long-term migrants.
Some variants on the benchmark case for merchandise exports and imports give results that are consistent with the hypothesis that migration stimulates trade. We expected that migration would have less effect on transaction costs and hence trade for agricultural exports and oil imports. It turns out that excluding agriculture and oil do indeed yield stronger relationships. A possible objection to our benchmark specification is that it does not include a real exchange rate variable. We applied our benchmark specification on the sub-sample for which the necessary data were available, and then added in the exchange rate variable. Adding the exchange rate variable made little difference to the results.
Experiments with interaction terms yielded some results that were inconsistent with the transaction cost interpretation. We had expected that the relationship between migration and trade would be stronger for migrants from non-English speaking countries than from English-speaking countries. When there are language differences and associated cultural and institutional differences, migrants’ potential contribution to facilitating trade is presumably greatest. The regression results suggest that the relationship between migration and the probability of trading is stronger for non-English speaking countries than for English speaking countries, we obtain the opposite result for our trade equations. We had also expected that the relationship between migration and trade would be stronger for migrants from low income countries than for high income countries (controlling for the independent effect of economic size), on the grounds that low income proxies for differences in culture and institutions. Again, the regression results were contrary to our expectations.
Other results make it difficult to be confident about the magnitude of the association between migration and trade. The experiment with the real exchange rate variable indicates that the results are sensitive to the choice of sample. The experiment with the square of the Migrant Stock term suggests that the elasticity of trade with respect to migration varies with the size of the migrant stock. We are not, however, satisfied that the squared term captures this relationship adequately.
It should be possible, in future work, to reduce some of the remaining uncertainties. We aim to disaggregate imports and exports by commodity type. As well as potentially generating further insights into the effects of migration, results from a disaggregated analysis can be used to test the prediction that migrants stimulate trade in differentiated goods more than homogenous goods. We will also experiment with specifications that allow elasticities to vary with the number of migrants.
In the meantime, our judgement is that migration to New Zealand does increase New Zealand’s trade with the migrants’ origin counties. We base this judgement on the fact that positive and significant associations occur in all our specifications, and in all overseas studies, and because the underlying theory is plausible. We are, however, uncertain about the strength of the relationship.
If migration does boost trade, what are the implications? When imports increase because of immigrant preference effects, standard welfare economics can say very little on whether this is good or bad. The situation is essentially the same as a change in the composition of demand due to a shift in preferences. In contrast, an increase in imports or exports brought about by a reduction in transaction costs must be welfare-enhancing. It is analogous to an increase in imports or exports brought about by a fall in shipping costs. A fall in costs allows New Zealanders to realise more gains from international trade. Any reduction in transaction costs can alternatively be viewed as a reduction in the effective distance between New Zealand and the rest of the world. New Zealand’s remoteness, combined with its small size, is often argued to be a serious handicap to the economy, so such effects are potentially important.
It would be easier, and safer, to draw policy implications when results of the analysis by commodity group are available. More satisfactory modelling of the relationship between numbers of migrants and the elasticity of trade with respect to migration would also reduce uncertainty about the strength of the effect. A provisional conclusion, however, is that immigration policies may need to be judged by their implications for trade, in addition to their implications for labour supply and human capital.
