Over the past two decades, governments have introduced measures to curb multinational tax avoidance, yet evidence from low- and middle-income countries remains scarce. This paper presents new findings from South Africa, linking administrative trade and corporate tax data to assess the effects of key anti-avoidance reforms. We examine two OECD BEPS initiatives, Country-by-Country Reporting and enhanced transfer pricing documentation, alongside the domestic implementation of an earnings-stripping rule that restricts interest deductibility.Using difference-in-differences estimators, we find that BEPS reforms significantly reduced transfer mispricing, lowering intra-firm prices and traded quantities with tax havens. Similarly, the debt-related reform resulted in a measurable reduction in the use of debt financing, while we do not observe any negative real effects for treated firms. Our findings offer novel evidence on the effectiveness of international tax reforms in reducing profit shifting in a developing country context.

Using a newly available exhaustive and granular dataset on the worldwide activity of multinational firms matched with oil, gas & mining production data at the country and firm level, we find evidence that windfall profits of multinational firms are excessively booked in low-tax countries. To identify the effect of price shocks on worldwide profit allocation, we exploit the heterogeneity in commodity price changes across mining and oil & gas products and use the fact that extractive firms specialize in different commodities. We find that a one percent increase in commodity prices leads to a 0.7pp excess increase in profits of subsidiaries located in tax havens.

How can tax authorities in low-income countries identify multinational enterprises (MNEs) engaged in profit shifting? In this project, we introduce a low-cost, data-driven methodology to identify profit-shifting MNEs using administrative data readily available to tax authorities in South Africa. By applying panel regression analysis and probabilistic detection methods, the method generates a variety of "red flags" for firms involved in suspicious activities, enabling tax authorities to prioritize high-risk cases for further auditing within their existing resource constraints. This approach empowers resource-limited tax authorities to target high-risk profit-shifting cases, supporting South Africa's broader revenue mobilization and fiscal sustainability goals. We contribute to the literature on profit shifting by presenting a novel method for identifying profit-shifting behaviours, which can be adapted by other low-income countries facing similar challenges.

Aggressive profit shifting by MNEs is a growing concern for domestic resource mobilization in developing economies. This paper evaluates the revenue and welfare consequences of a flagship anti-avoidance rule that has been implemented in morethan45countries to prevent profit shifting by MNEs through the debt channel. Our focus is Uganda, a representative developing country which implemented the rule in 2018. Exploiting admin data comprising the universe of corporate tax returns, we find that the rule does not significantly increase profits reported by MNEs in Uganda or tax remitted by them in Uganda. As an unintended consequence, however, the implementation of the rule leads to a contraction in real economic activity, reducing the turnover, employment, and trade of treated MNEs. We highlight the limited targeting efficiency of the rule, questioning its overall effects on welfare.