Much has been written on transfer pricing in the automobile sector, but ever so often, new business models emerge or regulations change, requiring a relook at the approaches used to determine transfer prices in this dynamic sector.
India has witnessed large foreign direct investments by automobile companies, and Indian companies such as Tata Motors and Mahindra & Mahindra have also made outbound acquisitions in the form of JLR and Ssangyong, respectively. These investments have led to considerable import and export of tangible and intangible property among parent and affiliated entities in India. Transactions range from purchase of completely built units (CBU), semi-knocked down units (SKD), completely knocked down units (CKD) and components, to payment of royalties and trademark fees. Given the multitude of transactions and their diversity, transfer pricing for this sector tends to go beyond comparability analysis and profit margin determination.
A function-asset-risk analysis is fundamental to transfer pricing and, in this context, the activities of an automobile manufacturer can vary even from one vehicle model to another. Take, for instance, a company that assembles a vehicle (say AP3) from an imported CKD or SKD kit, while another vehicle (say GK5) is produced on a localised platform, with only the engine and transmission imported. On the surface, it would appear that the company is manufacturing automobiles with a certain import component. The profitability ought to be measured in order to ascertain if the transfer price from the affiliate from which the kits and components are imported is at arm’s length. However, a detailed view of the manufacturing process shows that the production of AP3 is an assembly operation, with far lower local value-addition, as opposed to GK5, which is manufactured from start to finish with a much smaller, yet essential import content.
This leads to a radically different comparability analysis. Just as not all vehicles are similarly manufactured, not all comparable companies operate under similar business models. For that matter, the profitability of AP3 accruing to the Indian company may not even be the right parameter to judge whether the import transaction is priced at arm’s length. Rather, the return on value added may be a more relevant criterion, and comparable companies may simply be assemblers of automobile components and not auto manufacturers.
Now contrast this with the determination of the transfer price of the engine and transmission imported for GK5. Measuring the profitability of the vehicle need not, by itself, reveal whether the import price is at arm’s length. It is produced locally and there could be generic strategies at work in relation to import substitution, positioning, pricing and distribution by the manufacturing company, which could make an operating profit-level comparison quite misleading. Rather, determining the transfer pricing policy of the overseas entity that supplied the engine and transmission could be more relevant to ascertaining profit level retained by the supplier-entity. Ultimately, only the components that are imported need to be transfer priced. Hence, the segmental profitability of the overseas entity with respect to supplies sent to India would have to be benchmarked against comparable uncontrolled transactions.
Similarly, with technology transfer being intrinsic to automobile manufacture, and brand names holding the key to customer perceptions and demand, both these transactions that entail royalty payouts need to be evaluated carefully. In conclusion, transfer pricing in the automotive sector is constantly evolving, with the introduction of several models, variants and ongoing transition to higher levels of the value chain. This calls for refined approaches to transfer pricing — such as SKU-level analysis, contribution analysis, profit split and activity-based costing — in order to arrive at appropriate measures of arm’s length price.
Transfer pricing in the automotive sector is constantly evolving as new business models emerge or regulations change.