Modelling #LuxLeaks

Republished from the Tax Justice Research Bulletin – find it all at TJN, with added Afrobeat.

The industrial economics literature on FDI has tended to separate two elements: where multinational enterprises (MNEs) decide to locate, and where profit is subsequently declared. On the location question, the 1999 model of Haufler & Wootton has been influential. This shows that ‘competition’ between states to attract FDI, using tax breaks and subsidies, is likely to result in most or all of the benefits of the investment being captured by the MNE; but the market size advantage of a larger state ‘may even’ allow it to charge a positive tax rate and still attract the investment.

A new paper by Ma & Raimondos-Moller challenges this finding, by including the potential for profit-shifting as a factor in the location decision. Unsurprisingly, perhaps, this changes the set of possible equilibria in favour of a smaller state which is willing and able to apply transfer pricing rules more leniently – that is, to make it easier to shift in profits arising from economic activity located elsewhere.

Figures 1 and 2 show the position facing the MNE, in terms of the tax rates in the small country (t_A) and in the large country (t_B). The dotted black 45° line shows where these are equal. The blue line represents the iso-profit curve in the Haufler & Wootton model. Because this is above the 45° line, the MNE is only indifferent between investing in country A or country B if the former has a lower tax rate – so the large country can have a positive tax rate and still be more attractive.

The new (red) isoprofit line in Figure 1 shows that allowing for profit-shifting reduces the large country’s advantage – the red line is closer to 45° line in the cases of interest (tax below 50%). But the striking finding is the shift of the red line in figure 2, once ‘competition’ on transfer pricing laxness is included. Now for most likely tax rates, the isoprofit line has flipped to lie below the 45° line – so that the MNE will be indifferent about location only when the large country offers a lower tax rate than the small country, because the small country’s profit-shifting advantages now outweigh the market size advantage of the large country.

Of course a model is only a model, and inevitably lacks complexity. In this case, there is obvious scope for improvement. It is assumed that the location decision is ‘real’, so that locating in the small country (say Luxembourg) will imply transport costs in order to service the large country market (say the UK); and it is also treated as an either/or decision. In practice the structure in Luxembourg is unlikely to have any contact with physical product, and the decision is likely to be both/and. (TJN’s work shows the extent of the divorce, most strongly in Luxembourg, between profits declared and real economic activity.)

Nonetheless, the paper is important. As Ma and Raimondos-Moller point out, it is the first time that profit-shifting has been incorporated in MNE location decision modelling. We can surely look forward to further developments on similar lines, including empirical testing.