Picture: ISTOCK
Picture: ISTOCK

Two groundbreaking research papers on corporate tax avoidance in SA have found that the authorities have grossly underestimated the scale of unlawful profit shifting out of the country.

It conservatively amounts to R7bn a year in foregone taxes, and some of the largest multinationals are the biggest culprits.

The working papers, both authored by Ludvig Wier, a doctoral fellow at the University of Copenhagen, are among the first to emerge from a new research programme, SA-Tied. It is a tie-up of the National Treasury and other economic cluster ministries with several international research institutions, including the UN University World Institute for Development Economics Research.

The crux of the outcome has been the creation of a high-quality, anonymised tax information database on companies and individuals operating in SA that is enabling the authorities to gain fresh insights into the workings of the economy — and the way to close tax loopholes.

The two studies are the first to exploit actual tax returns to estimate profit shifting in a developing country. Until now only Germany, Denmark, Norway, Sweden and the US have granted researchers access to multinationals’ tax returns.

The issue of profit shifting — when multinationals lower their global tax bill by shifting their earnings from affiliates in high-tax countries to those in low-tax countries — is particularly relevant for SA, given its fiscal squeeze, reliance on corporate tax receipts and growing exposure to foreign-owned firms.

The top decile (10%) of foreign-owned firms account for 98% of the total estimated tax loss

Since 1994, there has been an explosion of foreign activity in SA. As a share of GDP, the earnings of foreign-owned corporations have doubled in the past 25 years.

A clear incentive exists for these firms to shift profits out of SA, where the corporate tax rate has been 28% since 2008 — four percentage points above the world average and 13 percentage points above nearby tax haven Mauritius.

Even though there are only about 2,000 foreign-owned firms out of a total of 1.2-million businesses in SA, these are large compared with domestic operations, accounting for more than 30% of the sales of all firms operating in the country. Roughly one-fifth of these foreign-owned companies are owned directly through tax havens.

In the first tax paper, "Big and Unprofitable", Wier and the Treasury’s Hayley Reynolds estimate that on average haven-owned businesses avoid taxation on as much as 80% of their true income by understating it in SA. The assessment was made by comparing the wage, asset and turnover data of local firms with that of foreign-owned companies.

Where the study breaks new ground is in finding that this aggregate tax loss conceals large differences across firms in accordance with their size. Most firms shift little or no income to tax havens, while a few large firms shift a lot.

In SA’s case, the authors estimate that a combination of high profits and more aggressive profit shifting results in the top 10% of foreign-owned firms accounting for 98% of the total estimated tax loss.

Wier says he was "very surprised" to detect this stark difference among firms.

The implications are globally significant. What it means is that all countries, including SA, that have relied on existing global research to estimate the extent of profit shifting affecting their economies — such as the OECD’s 2015 "Base Erosion and Profit Shifting" (Beps) report — have underestimated the scourge by up to 80%.

In SA’s case, Wier estimates that the SA Revenue Service (Sars) is being short-changed by about R7bn a year due to profit shifting. This is equal to about 4% of total current corporate income tax receipts, but it appears to be in line with that experienced in other countries.

However, this is a conservative estimate, given that the data does not allow the researchers to study profit shifting to sister firms rather than the parent, or from SA parents to foreign subsidiaries.

The study finds that the worst offenders are mining companies, accounting for almost 30% of profit shifting, followed by financial services firms, accounting for 19%. Wier considers this "alarming", given that resource extraction constitutes a large share of economic activity in developing countries.

Switzerland is ostensibly the main destination for siphoned-off profits, as Swiss-owned firms account for roughly half of the profit gap between haven-owned and non-haven-owned firms. Other havens that host the parents of firms that report near zero or negative profits in SA, despite having large wage bills, include Bermuda, Liechtenstein, the Cayman Islands, Mauritius, Singapore and Cyprus.

Wier’s second paper is on transfer mispricing — a form of profit shifting that occurs when firms apply a high price to items flowing from affiliates in low-tax countries (like Switzerland) to affiliates in high-tax countries (like SA) and vice versa. This erodes the profits in the high-tax affiliate, which is paying the high price, but equally increases the profits in the low-tax affiliate, which is receiving the high price.

Legally, firms are supposed to use arm’s-length pricing when transacting internally. That is, they should set prices internally as if they were trading with an external party.

The disaggregated customs data set that Wier used covers all imported goods to SA from 2011 to 2015, allowing him to estimate precisely what the arm’s-length price of each transaction should be. When comparing unit prices on intra-firm transactions to these arm’s-length prices, Wier found that imports from low-tax countries into SA are overpriced by at least 8%. This provides strong evidence to suggest that firms are engaging in tax-motivated transfer mispricing.

He estimates that SA is losing R1bn every year in this way alone — equivalent to 2% of foreign-owned firms’ tax payments.

SA’s legislation on transfer mispricing was amended in 2012 in line with the international standards laid out in the OECD’s Beps programme. Though there is evidence that transfer mispricing decreased after the amendments, by 2015 it was back up to former levels. Wier argues that this is because the new legislation was not matched by greater enforcement, so as soon as firms realised they were not being audited they went back to their old ways.

Fortunately, there is a cost-effective way to curb transfer mispricing. All the tax authorities need to do is apply the econometric method used in Wier’s paper to act as an automated digital flagging system. It took him just two weeks to set up the data in SA to flag companies with systematic deviations from estimated arm’s-length pricing automatically.

What it means

The biggest multinationals avoid 80% of tax

To his knowledge, no other country has set up a similar system. At a minimum, the tax authorities could send flagged firms an e-mail cautioning them that they have been flagged and to desist from such behaviour. "This seems to be a very low-hanging fruit for tax authorities globally to pursue," says Wier. "In many cases the data is already there, stored in a raw format on a server and used in the calculation of import statistics." The cost of this monitoring would be in the thousands of rands, while the potential tax gain runs into the billions.

According to Wier, three things are needed to curb profit shifting. First, countries should implement all the OECD’s Beps recommendations, as Sars is already doing. Second, they should invest in their tax authorities in terms of headcount, skills and technology, including the application of his algorithm. Third, there should be a push for global reform of the corporate tax system to prevent firms from shifting income.

The EU is considering a new system called "formula apportionment". It would require an international firm to allocate profits to its SA operation based on the proportion of activity carried out in the country. So, if 10% of a multinational’s sales, assets and employees are in SA, Sars should tax 10% of its global profits.

Chris Axelson, chief director of economic tax analysis at the Treasury, says it takes Beps "very seriously". The Treasury, with Sars, has adopted many of the recommendations emanating from the OECD’s Beps project, including making tax administration data available for research.

SA has also implemented the Beps minimum standards, in terms of which large multinationals are now required to report information about their global sales, assets and workforce to Sars.

Axelson declined to be drawn on how much of the R7bn in foregone tax the Treasury might be able to rake back through better enforcement, saying only that SA-Tied is likely to undertake more studies and that "the minimum standards that have been implemented, along with measures implemented in future, will bear fruit".

*This is an edited version of this story. We previously reported that the figure was R8bn, when it should be R7bn.