Transfer pricing is what happens when goods and services are traded between companies that are part of the same multinational group. It is often stated that developing countries are “haemorrhaging billions of dollars” of tax revenues through companies abusing this mechanism, in particular by mispricing commodities. Numbers racking up in the tens and hundreds of billions have influenced international debates on the subject (see for example: here, here, here). These big numbers—based on gaps and mismatches in trade data—are unreliable and should be taken with a pinch of salt. So too should narratives which offer extraordinary examples, such as the accusations being made in the ongoing case of Acacia Mining in Tanzania, as a guide to general expectations.

There is no doubt that customs inspection and auditing of transfer prices are significant technical challenges, and that companies can take advantage of weak regulations and enforcement. There are real gains to be made by developing countries strengthening transfer pricing rules and auditing capacity, and from developed countries sharing information with them to reduce information asymmetries between international businesses and national authorities.

However, the scale of revenues that might be recovered is unlikely to match up to heightened popular expectations. South Africa developed new transfer pricing rules in 2012, and in the following four years the revenue service successfully finalised 35 transfer pricing cases which raised around $650 million (see annual reports from 2013-2016). Early results from the “Tax Inspectors Without Borders” programme include increasing revenue from transfer pricing audits in Colombia from $6 million in 2012 to $33 million in 2014, in Kenya from $52 million in 2012 to $107 million, and in Vietnam from $3.9 million in 2013 to $40 million in 2014.

If the popular “billions and trillions” studies are not a good guide to sizing the prize from improving international corporate tax compliance, what is? A number of revenue authorities including the UK’s HMRC, the South African Revenue Service, the Finish tax authority, and the IRS allow carefully vetted researchers to analyse anonymised tax records under controlled conditions. Studies based on microdata from such “datalabs” may yield a clearer picture of the revenues which might be at stake.

Some recent studies

Hayley Reynolds and Ludvig Wier used firm-level tax returns to estimate profit shifting in South Africa. They find “a semi-elasticity of taxable income with respect to the parent tax rate of 1.7.” That means that a South African subsidiary of a foreign company whose headquarters is based in a country where the tax rate is 10 percent lower than South Africa’s (i.e., something like the UK or Singapore) would tend to have a 17 percent lower taxable income than one whose parent company is in a country matching South Africa’s rate. They make a ballpark estimate of revenues lost in this way; finding that it amounts to 7 percent of subsidiary income or 1 percent of the total corporate tax base. This implies that profit shifting removes 0.2 percent of the total tax base in South Africa, reducing government revenues by 0.05 percent of GDP (around $147 million or $2.60 per person per year).

Li Liu, Tim Schmidt-Eisenlohr, and Dongxian Guo look at UK company tax returns and transaction level trade data to explore transfer pricing of goods exports from the UK. They find that transfer prices of exports to lower tax countries such as Ireland, Turkey, Denmark, Russia, and Netherlands are most sensitive to changes in relative tax rates, while there is little mispricing of exports via small economy “tax havens.” The price differential is more pronounced for R&D intensive firms (i.e. where the product exported are specific rather than generic). However again the revenue forgone was small in absolute terms—amounting to £168 million in 2010 (0.01 percent of GDP, again $2.60 per person per year).

There are many caveats to these studies. Both rely on statutory tax rates for their calculations, which does not allow for the impact of special incentives. The UK study only looks at goods trade, so does not include profit shifting via services, interest and royalties. On the other hand, it is not at all clear that the price differences it identifies as “mispricing” would necessarily fall outside of a defensible arms length range. The South African study only considers revenue losses associated with foreign multinationals, not profit shifting by South African headquartered companies.

Nevertheless, what is clear is that the figures they show, which approximate to the price of a hamburger per person per year, are far removed from the great expectations of transformative amounts of public revenue. Both studies suggest that companies are undertaking a significant degree of profit shifting, but this has a relatively small impact on overall public revenues, simply because multinational companies are a limited portion of the corporate tax base, and this in turn is a limited portion of the overall tax base.

Are these findings surprising?

These findings are similar to other studies of corporate tax elasticities. It is notable, however, that they are at least an order of magnitude smaller than those that were generated by the ‘spillovers’ study conducted by Ernesto Crivelli, Ruud De Mooij and Michael Keen of the IMF in 2015. This much quoted study gave a speculative figure that losses to developing countries from tax avoidance were in the order of 1 percent of GDP ($200 billion overall).

The amount of corporate tax revenues that countries might be losing fundamentally depends on the amount of profit that is generated by companies in the jurisdiction, the extent to which those companies are able to access international profit shifting opportunities (are they part of a multinational group?), the nature of their business (for example does it involve differentiated products and intangible assets?) and the effectiveness of the tax administration.

The IMF study did not include any of these factors but instead looked how fast a subset of countries (where data was available) were “broadening the base and lowering the rate” of overall corporate taxes, to try to identify a general relationship and isolate the impact of “base erosion” via tax havens. They then applied this tentative relationship more generally to a wider set of countries. A follow-up study by Alex Cobham and Petr Janský provides a breakdown by country, and notes that the methodology leads to some hard-to-believe findings. For example, Chad is said to be losing tax revenues worth some 8 percent of its GDP. Pakistan is said to be similarly losing tax revenues worth 5 percent of GDP. For this to be true untaxed profits related to internationally connected formal sector business would account for 20 percent of GDP in Chad, and 14 percent in Pakistan—suggesting the corporate sector might be more prominent in these economies than in countries such as the UK and Denmark where the corporate tax base is around 11 percent of GDP.

One place where findings from bottom-up studies, top-down studies, and popular expectations are closer together: the United States.

Kimberley Clausing uses microdata from confidential surveys carried out by the US Bureau of Economic Analysis. She estimates that losses to profit shifting from the US to countries such as the Netherlands, Ireland, Luxembourg, and Singapore was around $100 billion in 2012; that is around 45 percent of actual corporate taxes collected, or around 0.7 percent of GDP. (The lost revenue is around half of the amount suggested for the US by the IMF methodology, according to Cobham and Jansky’s disaggregation or more than 125 hamburgers per person per year.) This aligns with what we know from specific cases, which is that many US multinationals have very low effective tax rates on profits from revenues generated in other countries (including global household names such as Google, Starbucks and Microsoft) and US companies have permanently reinvested around $2.4 trillion of earnings offshore in order to defer US tax liabilities.

The United States is also exceptional—both as a major originator of globally-used intangible assets, and as a country with an unusual tax system that encourages retaining earnings offshore. Further studies using microdata will help to understand the nature of corporate profit shifting globally in different countries, but we should not be surprised if they don’t look like the United States.