The United States maintains a clear position opposing a DST, arguing that it solely targets American tech giants. Nevertheless, many countries are ready to implement DSTs. Moreover, the most recent OECD/G20 document on Tax Challenges Arising from Digitalization offers three proposals, including the user participation proposal which is considered as endorsing a DST, stating that the proposals are reviewed without prejudice.
Given this reality, one might wonder whether there are merits to a DST, at least at a theoretical level. Wei Cui's paper, The Digital Services Tax: A Conceptual Defense, offers a worthy defense of DST from a novel perspective.
The first contribution of this paper is to show the implication of the narrow scope of digital businesses subject to DSTs. For example, in the UK DST proposal, digital business models subject to DST only include "a social media platform, search engines or online marketplace." Digital interfaces providing digital contents, communication services, or payment services to users are excluded. Thus, Facebook, Twitter, YouTube, Google, Amazon Marketplace, Kayak, Priceline, Uber, and Airbnb are in scope, whereas PayPal, Netflix, Spotify, and Ubisoft are excluded. However, there are certain platforms that need further clarification–e.g., it is still puzzling for me whether LinkedIn or YouTube are considered a social media platform or a digital interface providing digital content, especially when it relates to premium service. On the narrow scope of DST, the author remarks that, practically, it may be good thing for any real-world DST to start with only particular sectors first, because it would help governments understand the impact of the DST before implemented in a larger scale. However, the author recommends that the boundaries of the DST exemptions will likely come to receive greater scrutiny after DSTs are enacted in real practice.
The second, and most interesting contribution of the paper, is the justification of the DST because it would allow location-specific rent (LSR) earned by digital platforms to be captured by the countries in which such rent arises. Existing literature has explained that DST is another effort to allocate taxing rights to destination or market jurisdictions. On the other hand, the author argues that DST proposals do not embrace the allocation of taxing rights to either destination or source countries. Instead, the idea is to allow the country in which the rent is located to tax the rent. In two-sided business models, such as Amazon Marketplace or Airbnb, there exist "indirect network effects" when one type of user cares about what the other type of user does. Suppose there is a user-seller and a user-buyer in Amazon Marketplace. The platform (Amazon Marketplace) manipulates the structure of prices charged to each side and offers price below marginal cost to user-buyer, while making up for that loss by charging user-seller. Suppose further that user-buyer is in the UK while the user-seller is in Germany. User-buyer in the UK is crucial for the platform company (Amazon Marketplace)'s ability to profit from user-seller in Germany, yet such "user value creation" in the UK may be accompanied by little or no payment from the UK. The author explains that the economic rent from this business model is LSR arising in the UK, although Amazon Marketplace makes a profit from the fees charged to the user-seller in Germany.
However, traditional international income taxation fails to catch such hidden LSR arising in the UK and allocated to its proper origin. Instead, it attributes the source of income or economic rent mainly by tracking sources of payment (here, Germany). In order to allow the country where the rent is located to tax such hidden LSR, the author introduces a basic intuition about how rent accruing to mobile intangible assets should be assigned: when the deployment of a technology is non-rival with respect to multiple locations—its deployment in one country has no opportunity cost in terms of its deployment in another country—, it is both efficient and fair to assign rent earned from the technology's deployment with respect to a given location to that location.
It is possible for Airbnb to tell a story similar to that of Amazon Marketplace’s pricing, except with the seller and buyer's roles reversed. Many platforms providing connections between sellers and buyers of goods and services subsidize the seller side, as Airbnb does, because the participation of property owners is the key to success for Airbnb. Thus, Airbnb earns a profit mainly from service fees charged to the consumer side (i.e. user-buyers), while the origin of that very profit may have more to do with the supplier side (i.e., user-seller). Here, the hidden LSR earned by digital platforms arises in a supplier jurisdiction, rather than a consumer jurisdiction. Therefore, "user value creation" on digital platforms should not be conflated with participation by the consumer alone.
The paper then disputes some common claims against DST. First, it argues that the criticism on DST's gross-revenue based taxation is ill-conceived because digital platform firms operate with very little, often negligible, marginal costs. Once the platform is running, the placement of each additional service is largely automated with little additional labor or other input from the platform company, so that the revenue of the platform company from each additional transaction is essentially identical to its marginal profit from the transaction. Furthermore, it is arbitrary to argue that tax should not be imposed until the platform firm turns a profit because many governments often use a gross revenue-based royalty regime in taxation of natural resource extraction, which is another rent taxation.
Second, any claims arguing that the cost of the DST will simply be passed onto the consumers under a DST model are incorrect. For example, Google charges advertisers while subsidizing individual users of its search functions. Passing the DST cost of advertising onto individual users would not be a wise response to the introduction of the DST, since it would reduce its overall usage which would then also reduce the appetite of the advertisers. Instead, Google would be better off passing the DST cost to its advertisers. Since this additional cost represents a fixed rather than marginal cost, it will reduce advertisers' profit but not raise product prices. In these cases, it is implausible to argue that the DST cost of advertising would be passed onto final consumers.
The paper provides many rich arguments, all of which could not be introduced in this review. It provides many arguments that seek to clarify various analytical misconceptions held by both proponents and opponents of the DST, while also making the novel arguments introduced above. Cui's paper is a great addition to the literature of DST, and as such, this review recommends it to policymakers, practitioners, and commentators, regardless of their respective views of the DST.