Mirroring Hypothesis and Intergrality in the Electric Vehicle Industry: Evidence from Tesla Motors

Mots-clés:

Résumé:

Although auto industry and its original equipment manufacturers (OEMs) have pursued vertical integration to an extreme level in its early history, the strategy seems to have lost its traction since late 1980s. One main contributor to this change, according to research in organizational economics and strategic management, is outsourcing and its consequence on the evolution toward more modular product architectures (Frigant, 2011; MacDuffie, 2013). This relationship between the degree of vertical integration in an industry and product architecture comprising integrality and modularity is presented as the “mirroring” hypothesis in the literature (Argyres & Bigelow, 2010; Baldwin, 2008; Colfer & Baldwin 2010; Sanchez & Mahoney, 1996). It suggest that vertical integration of component products in the upstream and downstream operations of the firm is associated with integrality of product architecture (“where the product results from the assembly of physically interdependent or multi-functional subassemblies, and/or whose interfaces have been decoupled” (Frigant, 2011: 326)). Conversely, the outsourcing strategy of firms is associated with the modularity degree of products.

Electric vehicles (EV) is considered as a radical innovation in today’s automobile industry. EV has challenged current automobile product architecture, as it involves new components inside vehicle (e.g. battery and electric motor) and sets a new ecosystem that require the commitment from electricity and recharging system. The growth in EV market and changes in EV architecture prompted us to examine two interrelated questions regarding the mirroring hypothesis: 1) have architectural product changes in the EV industry shifted towards higher levels of integrality?; 2) does the product architecture of the EV mirror its OEMs’ vertical integration strategy? This proposal is intended to answer these questions.

In the search for a dominant design, two approaches are evident in the EV industry. On the one hand, most incumbent auto manufacturers define the architecture of their EVs by minimizing changes in their existing production infrastructure, capabilities, and network of external partners. In its most basic form, automakers refit existing conventional car models with electric powertrains. The alternative is a purpose-designed EV. This radical strategic choice involves a completely new design with an idiosyncratic architecture dedicated to the battery and range anxiety issues. This architecture leads to the production of specific modules and components. It also pushes the boundaries of the automotive industry by mandating an alignment of the interests of various stakeholders in the EV ecosystem. 

We focus on purpose-designed EV to highlight the mirroring hypothesis. We analyze a single in-depth case study on Tesla and its flagship Model S. Two considerations contributed to Tesla’s selection: 1) its dedication to purpose-designed EV allows us to study their architecture-organization relationships; and 2) its influence on the EV industry as a relatively successful player in the EV ecosystem. Following the literature of modularity and integrality in product architecture (Ulrich 1995; Ulrich & Eppinger, 2000; Fixson, 2005), we study the integrality level of Model S according to i) the mapping from function elements to physical blocks; and ii) dependencies on interfaces between blocks. Key blocks and their interfaces in EV system are identified in Figure 1. We capture the coupling characteristics of interfaces by the product Design Structure Matrix (DSM) map (Baldwin & Clark, 2006). Our approach generates theoretical insights from a single in-depth case (Yin, 2013) based on qualitative data on the supply chain relationships of Tesla and its product architecture. Data collection covers the period from June 2011 (Tesla went public) to June 2016.

Our initial findings shows 1) Tesla designs its Model S, its battery pack, the charging interface (plugs and IT) and a fast-charging Supercharger simultaneously for coupling interfaces and ‘one to many’ mapping which, in turn, increases the integrality of the product. It calls for the right level of vertical integration as a result of more transaction costs caused by the opportunism risk, asset specificity, and potential hold-up situation. 2) Tesla’s focus on the density (defined in terms of easy availability and quantity) of the charging network has no direct impact on the integrality of the Model S. However it calls for the right degree of vertical integration. This alignment between the components of the Model S and vertical integration demonstrates the mirroring hypothesis of integrality in the case of Tesla.

Our work contributes to the body of knowledge on the mirroring hypothesis of integrality, which was hitherto unexplored in the literature. To the best of our knowledge, only one study (e.g., Fixson & Park, 2008) explored this issue and called for other research on other industries for additional insight. Our research will contribute to this end. Our research calls attention of OEMs to a new meaning of product architecture and make-or-buy question in EV industry: a) our research argues with previous studies (e.g., Cabigiosu, 2013; Christensen, 2011) on EV architecture, which are focusing on refitted EV and conclude that EV is becoming more modular. We argues that the refitted EV industry may indeed become more modular, mainly because the rest of the architecture is already in place. This conclusion might not be true for purpose-designed EV. Contrariwise, the case of Tesla shows that a radical change in the architecture of the BEV leads to a higher level of integrality. b) As EV is still in the emerging phase of the industry lifecycle, the strategy of vertical integration may play a key role in the success or failure of the incumbents or the new entrants. The mirroring hypothesis of integrality may make innovation of Tesla’s architecture novel and preemptive so that it can shape the structure of the EV industry. c) This research highlights the dynamism of the integrality-modularity sequence in the EV industry. At the beginning of the industry, the integrality of the product resulted in a high level of vertical integration. Over time, the drive for outsourcing may lead to modularity in the product architecture. This allows us to offer a new proposition for further research: When reduction of costs becomes an imperative for the OEMs in a mature EV industry, more modularity may appear.