Discussion of “Productivity and Taxes as Drivers of FDI”

by A. Razin and E. Sadkain Brookings Trade Forum, 2007.

Assaf Razin and Efram Sadka have synthesized some of their recent work on productivity, taxes and FDI in this welcome contribution. Their theoretical work emphasizes how set-up costs at home and in the host country can drive a wedge between marginal and total profit conditions. This insight combines with a description of FDI decision-making that separates the decision to invest and how much to invest. Their theory leads to some intriguing results on the relationship between productivity, taxes and FDI. Their empirical work provides some support for their theoretical results.

In the Razin-Sadka setup, setup costs at home and in the host country are critical and lead to the curious results they deliver. For example, they argue that a positive productivity shock in the host country can deter entry of multinational firms. The productivity shock bids up wages and, because host country setup costs are entirely wage costs, the total profit conditions that dictate entry now are tilted toward not entering. Setup costs, and their deductible nature, also influence the role of taxes. In particular, tax rate increases at home can lead to increased outbound FDI because the value of the deductions for setup costs borne at home goes up. 
Setup costs are surely important to the spread of multinational firms. But such counterintuitive results demand some motivation for the underlying mechanism. For example, what are these sizable wage-driven setup costs that are so critical to the model? Little motivation is provided for these costs and, while the results are all derived nicely, it is hard to know what to make of their importance without such motivation. I confess to having a hard time deriving any intuition or examples for such costs other than the fees paid to investment bankers and consultants upon entry in a country. I find it hard to believe that such costs are meaningful relative to the scale of the projects under consideration.

It is also not clear that such setup costs in the host country, and particularly at home, are the same every time a multinational enters a new country. When General Electric sets up operations in the nth country that they operate in, do we believe they bear considerable setup costs at home or, even for that matter, that the costs they bear in the nth country are the same as the costs borne when they first became a multinational firm? These costs would also seem to bear some relationship to market size that are not developed in the model. Similarly, it is also not clear why firms are choosing amongst projects in an exclusive, presumably reflecting some hidden  financing constraints. Finally, the treatment of taxes, particularly home-country taxes, is somewhat primitive as the rich interactions of home and host country regimes that have been shown to be so important to patterns of multinational firm activity are neglected in the model.

The paper takes these intuitions from the theory and attempts to apply them to bilateral FDI flows within the OECD. Given the interest in the separability of the decision to invest and how much to invest, it would seem that it would be useful to look for data sources outside of the OECD where “zeros” (bilateral pairs where there is no FDI) are more likely to be prevalent. I was also puzzled by the treatment of Europe in some of the tables as the authors appear to not examine any FDI flows within Europe. Nonetheless, the authors report results on productivity shocks at home and taxes at home that conform to the theory’s predictions.

Of course, it is always possible to come up with alternative explanations for such results. For example, the result that increases in the source country tax rate leads to more outbound FDI because of the increased value of the deductions of the setup costs could have a simpler interpretation. The more naive interpretation of that result is that when the source country tax rate goes up, you have an incentive to go abroad with real activity to facilitate profit relocations. This alternative explanation and others like it are hard to disentangle in this empirical setting. It would be nice for the authors to attempt to take their intuitions to micro data on multinational firms as this is where their predictions would seem to have the most purchase and where most of the recent literature on multinational firms is now.

This fine paper pushes the scholarly community to take setup costs seriously in their consideration of FDI. This is surely a welcome direction for scholarship as the recent work on firm heterogeneity and patterns of trade and FDI have suggested. I look forward to further work by these scholars as they elaborate these mechanisms and provide more empirical evidence of their relevance.