Priyabrata Pramanik
Addl. Commissioner,
Tax Policy Research Unit
Department Of Revenue, Ministry Of Finance

Sri Priyabrata Pramanikhas done his Masters in Physics with specialization in Electronics from Jadavpur University, Kolkata. He belongs to IRS (Income Tax) and currently is working as Additional Commissioner, Tax Policy Research Unit, in the Department of Revenue. His area of specialization and interest lies in Microsimulation Modelling in PIT, CIT, GST, Overlapping Generation Models in linking the Union Budget Constraints with common man’s House Hold Budget Constraints, Pricing of commodities,GDP growth, National Accounts, Big Data Analytics, Super Computer, etc.

He is a member of the Working Group headed by Dr. Ramesh Chand, Member, NitiAayog, formed by the Ministry of Commerce and Industry to revise the Wholesale Price Index and Producer Price Index of India. He is also a member of the Working Group on Economic Census of India formed by the Ministry of Statistics &Programme Implementation headed by the Secretary. In the Income Tax Department, he worked in various wings like I&CI, Investigation, Assessments, Audit etc. While working in the field, he has developed the first Microsimulation Model in India for calculating Sectoral Tax Gap in Direct and Indirect taxes.

Executive Summary

The world is now divided as to how to frame the rules of international taxation in digital economy. No consensus has been reached as yet. In this article opinion has been expressed as how one may go about framing new international taxation rules, taking into account global value creation through global value chains. One should consider value of capturing the behavioral pattern of users in algorithms, network effect creation, how critical mass of end users are reached for global sustenance and monetization of globalized data. A simultaneous approach through the laws of Physics being universal in nature as well as economic pattern unique to a country has been taken into account.


1.1 The top 20 economy countries of the world, forming the group G20, decided in summit meeting that taxation of Multi National Enterprises (MNEs) operating in digital economy platform should be made on the cardinal principle of ‘where economic activities occur and value is created’. OECD has been assigned the task of framing new International Taxation Rules keeping this principle in mind.

The report of OECD is to be submitted before next G20 summit meeting for ratification by the global leaders. This poses a serious challenge to determine how the ‘value is created’ in Digital Economy which is different from Brick and Mortar model. This also reveals the shortcomings of the existing International Rules in dealing with this aspect.

1.2 How India fits in this global scenario? 1 Amongst countries with market access in digital economy, India has emerged as the world’s largest global marketplace with its 462 million internet users (11 percent of world-wide users) in 2017, followed by the European Union with 433 million internet users (10 percent of worldwide users). The number of internet users in India is expected to increase rapidly from 462 million in 2017 to 650 million in 20202. With rising internet penetration and greater digital maturity of users in India, the e-commerce is likely to increase from $45 billion in 2017 to $100 billion by 2020g.


2.1 OECD calls for a paradigm shift4 in framing the rules in International Tax. It observes that ‘ the advent of e-commerce and the internet has had a dramatic impact on intellectual supplies. Intellectual content can now be readily digitized. Direct provision of intellectual supplies is available. Basic ingredients like speed, bandwidth, software and hardware applications have achieved appropriate and workable levels. It requires little infrastructure to operate from their original jurisdiction and easily shift to locations of low tax jurisdictions.’ It adds ‘Digitalization is pervasive, and hence does not pose unique issues, but exacerbates the challenges for international taxation. In particular, it has further increased the opportunities for multinational enterprises (MNEs) to exploit the fundamental flaw in international tax rules: the independent entity principle. This requires tax authorities to start from the accounts in each country of the various affiliates of MNEs; and, although they have powers (though without adequate resources) to adjust those accounts, they are expected to do so on the basis that such affiliates are independent entities, dealing at ‘arm’s length’ with each other. This independent entity fiction runs counter to the economic reality that MNEs operate as unitary firms under centralized control and direction. It also allows, indeed encourages, MNEs to create complex corporate groups, with often hundreds of affiliates, many located in tax havens, enabling them to achieve low overall effective tax rates on their global profits. Such strategies have become easier for all MNEs due to digitalization of business models, even those which involve supplying physical commodities (e.g. Apple, Amazon). This clearly requires a paradigm shift in international tax. It was implicit in the call from the G20 leaders for reform of the rules to ensure that MNEs could be taxed ‘where economic activities occur and value is created’.

2.2 In the said input BMG has also suggested to create a Digital PE for digital economy instead of the conventional PE for brick and mortar model. For the test of the digital PE the criteria of the following should be met.

  • Significant economic presence.
  • Withholding tax on digital transactions
  • Digital Equalising Levy

This can be measured through allocation keys involving users.

2.3 ‘Simplified Allocation Keys: For the combined profits

of this common business model, two equally weighted allocation keys are defined as follows:

USERS: Using users as an allocation key reflects the importance of each market and the value of users to the global business of MNE from fee-paying third-party customers seeking advertising services. The country is determined by the location of the user and not the legal terms of any contracts, licenses, or other documents with either users or the third-parties that pay the MNE for advertising, aggregate user data, etc.

Operating Expenses: This allocation key recognizes all operational inputs.

  • As such, it covers all research and development, website maintenance, sales, marketing, distribution, management, support functions, etc. This key would include categories of expenses such as: Salaries and bonuses of all operations personnel (allocated by location of personnel).
  • All other direct and allocated operating expenses (allocated by location of personnel or facility to which the expenses relate).
  • Commissions and service fees paid to other parties for all operational functions (allocated by location where the other party provides the services) (These payments economically include all personnel costs, office and

manufacturing costs, etc. of the legal entity performing the relevant operational functions for the taxpayer. Payments to any related parties whose profits are included in the combined profits for the profit split would of course be excluded.)’

2.4 OECD Reports4 : The Interim Report identifies three characteristics that are frequently observed in highly digitalized business models (HDBM):

2.4.1 Cross jurisdictional scale without mass

  • Digitalisation has made business functions more mobile;
  • digital businesses can interact with global customer bases without substantial increases in physical presence

2.4.2 Importance of intangible assets and intellectual property rights (IPRs)

  • Algorithms, patents and other intangibles are becoming more relevant in production processes
  • These are thus affecting many different sectors; crucial for some business models (e.g., platform-based businesses) which would not exist in their current form without intense use of intangibles

2.4.3 Role of data and user participation, including network effects

  • Data collection increased in quality, quantity and speed
  • Data and user-generated content (UGC) are used in all phases of value creation
  • There are important synergies between knowledge creation from data and IPRs
  • Reliance on intangibles also reinforces the contribution of data to value creation
  • Data often a key driver of firms’ profitability and competitive advantage
  • Economic gains can be leveraged through data value cycles:
  • Collection big databases analytics knowledge bases data-driven decisions
  • Businesses are able to improve products, operations, marketing and decision making through increased reliance on data value cycles
  • Productivity increase imply gains in market share, competitiveness, profitability
  • Data and user-generated content is a significant source of value creation

Member countries agreed to undertake a coherent and concurrent review of the rules governing

  • nexus
  • profit allocation

with respect to the fundamental concepts relating to the allocation of taxing rights between jurisdictions and the determination of the relevant share of the multinational enterprise’s profits that will be subject to taxation in a given jurisdiction. However, while giving their decisions the opinions varied among countries and different stake holders.

2.5 The European Union (European Commission) in its communication on 5“Proposal for a Council Directive laying down rules relating to the corporate taxation of a significant digital presence” dated 21.03.2018 held that,

“The fundamental principle for profit allocation should remain that taxation takes place in the jurisdiction where value is created. Considering that in the digital economy, a significant part of the value of a business is created where the users are based and data is collected and processed, the directive would set out criteria specifically targeted at attributing profit to a digital permanent establishment.”

For example, these criteria could be based on criteria such as:

  • the users’ engagements and contributions to the development of a platform;
  • the data collected from users in a Member State through a digital platform;
  • number of users; and/or
  • user-generated content.

3. Stakeholders’ view – Digital Economy Group (DEG)

3.1 Digital Economy Group (DEG) consisting of 10 top companies of the world operating in the digital platform viz., Amazon, Expedia, Google, Facebook, Netflix, Microsoft, Twitter, etc. submitted before OECD6 in their Request for Input (RFI) dated October, 2017 that ‘the digital economy is becoming the economy itself, and that it is not possible to “ring fence” the “digital economy”. ‘Any proposal which applies only to a certain segment of the digitalized economy would constitute a clear violation of the neutrality principle. This would include any proposal which applies only to “highly digitalized business models”. Digitalization derives innovation and growth and has enhanced the business efficiency of virtually every sector of the economy.’

3.2.1 Value-creation: DEG has stated that this topic has not been discussed transparently earlier. They have strongly objected that value is being created simply by the existence of market and hence under international tax policy market should attract some part of the tax. These arguments would be a substantial departure from existing policy with possibly far reaching consequences for all business engaged in cross-border trade.

3.2.2 Secondly, DEG states ‘we believe that an enterprise creates its success through its deployment of personnel and capital resources. Innovation and production create value, consumption does not. A commercial transaction between a supplier and a purchaser is an exchange of value for value. (whether for goods or services) but that transaction creates no new value. Digital companies compete with each other worldwide to attract the best talent, in order to design the business models, features, platforms, and analytical tools that drive their businesses these businesses, the capital investment necessary to fund them, and the risks innovators take on them with no guarantee of success. These businesses certainly utilize the internet, but using this means of connectivity does not discount the value of the personal skill needed to create. The capital investment necessary to fund or attract the best talent, in order to design the business model, features, platforms, and analytical tools drive their businesses. It depends on the risks innovators take with no guaranty of success. Endorsing the policy notion that value is created by consumption rather than production would be a fundamental shift of rights to a tax base from the country of production to the country of consumption. This conclusion cannot be limited in application to the digitalized economy, or indeed some subsect of the digitalized economy. If the OECD were to endorse this shift for some subset of ordinary business transactions, it is hard to see a principled reason how to limit the expansion of this theory to apply to other export sales into a market.’

3.2.3 The DEG in its RFI on the role of the data emphasized that: “Data itself does not create value. Rather, the value is created through those processes which structure, aggregate, and analyze, and present the data in a manner responsive to the users’ objectives. Data has always existed; what is new is the ability of many enterprises (not just the “highly digitalized” ones) to structure data in ways that allow the application of analytical tools against that aggregated data. That value is created through application of the enterprise’s investments in advanced computer processing and software tools. Users and consumers of digital services are in no way involved in these structuring and analysis functions. We understand that arguments have been made that a direct income tax is appropriate for those enterprises which obtain access to a market through digital means. This market access theory is essentially a tariff on imports. We would be surprised if the OECD were to endorse this theory of taxation.”

3.3 The DEG further argued that “We also understand that arguments have been made that tax should be imposed on a theory that users constitute a natural resource that is mined by enterprises. This is a novel analogy but it certainly fails to justify an extraordinary allocation of taxation rights to the jurisdiction where users reside, any more than would be the case for purchasers of luxury goods, high performance automobiles, or any other item. We also understand that arguments have been made that sales made by digitalized economy enterprises benefit from the infrastructure of the consumer’s jurisdiction. Again, this theory does not distinguish the cases of luxury goods, which are bought by educated and sophisticated consumers, or luxury cars, which are driven on roads financed by residents of the market jurisdiction. In fact, the capital investment in the hardware which supports digital connectivity has been financed, built and operated in large part by the enterprises which provide the digital services.”

3.4 “The main feature at the core of all three options is the ability for an enterprise to access a market remotely. That commercial reach is possible due to the widely available internet and cloud infrastructure, not necessarily due to the IP of the enterprise selling remotely. Those infrastructure resources are essential to all remote sales models. They also are available to all enterprises. Thus, the technology which provides the highways for remote sales and the other benefits of digitalization are available to all enterprises, large and small, without distinction of geographic origin. Every innovator may capitalize on these resources at relatively low cost, enabling start-ups and SMEs in developing markets to compete in the global market. IP does not arise spontaneously; it is the result of substantial human effort and investment risk. Statutorily protected IP such as patents, copyrights, and trademarks protect the innovative work of employees or individual entrepreneurs. To the extent that IP is viewed as especially significant to value creation in the digitalized economy, its role should be seen as emphasizing the value contributed by the human innovators whose work the law protects.”

3.5 The DEG concluded that “Despite the review of business models conducted for the 2015 Action 1 Final Report, we believe that in some quarters there is still an assumption that highly digitalized enterprises “operate in the cloud”. Enterprises which have managed to grow in these highly competitive sectors have done so only through heavy investment in very substantial workforces, and multi-billion-dollar yearly investment in infrastructure to enable the cloud’s operation (e.g., data centers, services, network architecture, etc.). “


4.1 The DET-3 in contrast has the following views, “This

brings a need to evaluate impact of A.I. and Data analytics role in the significant people functions BEPS analysis. But the key game changes in our view is that in the new world of client centricity, a relevant part of the interaction with client or potential client has moved to the digital space and brands in any segment/industry and market level, are re-designing their customer touch-points strategies to adapt to the new digital channels, that will co-exist with other channels. Omni-channel management is what companies are forced to do to play in most markets, especially in some industries. That has a significant cost, and many times the cost of related resources managing that budget and activity is away of the market country where consumer is based.

4.2 A new intangible asset class is in our opinion strongly emerging and required to be identified due to its relevance: the “data rights”. Unprecedented and scaling at a pace whose growing curve was not seen before in human history. This fact makes it necessary for most companies to invest efforts and resources (functions) in organizing and curating that vast amount of data collected as a first step, and second, to develop a culture that enable the business to think about extracting value from that data, which takes time. Some digital native companies created only in the last 15 years are already mastering the use of algorithm driven business models working predominantly with data and being completely relying on it for every turn of the key in their business model (see the case of some Live Stream content value propositions or E-Market places). But most other companies, in any industry, will be soon reaping the benefits of the last part of the data journey, which is the exploitation and use of the data to produce value.

4.3 Data is used for almost anything in HDBMs (Highly Digitalized Business Models), but this is not just a digital companies pattern, it is changing now the status quo in most industries. Global value chains depend on seamless, dynamic, continuous information/data flows across the legal entities and the different functions/departments. We go to a multinational company type where data is used real time everywhere across the value chain. That is why we talk about knowledge base capital, that is not in the balance sheet but is indeed a key element of any Transfer Pricing analysis and likely a unique and valuable tax asset. Data ownership across the value chain needs to be mapped, but at the right level of effort/detail.

4.4 In our view, trying to list the types of data that will be generated is an endless exercise because the data touch points that most companies will generate across their business model is vast. But we can nonetheless make 3 general categories to simplify any analysis:

1. Class C: Data that enhance MNE operations

2. Class B: Data that enhance customer relations

3. Class A: Data enabling New Products/Business models

While there could be some unavoidable intra-categories overlap, this type of general conceptualization, would allow at least an initial approximation to the subsequent topic of how much data could be worth. Let us also have a look at how raw data is creating a value chain for the experiments in physics of global importance.”

4.5 The OECD in the Interim report in March 2018 summed up all the countries’ views on User participation as 2 broadly differing views as follows:

  • Role of user participation is a unique and important driver of value creation – Allows HDBMs to collect a great deal of information and monetise it in various ways (e.g., pricing, advertisement)
  • The action by HDBMs to source data from users is NOT an activity to which profit should be attributed – User’s contributed data, content and other information similar to any other input sourced from an independent, third party.


5.1 In view of these contrasting views from economic

angles, let us take the help of the Laws of Physics, which is more universal in nature cutting across countries, jurisdictions, administrations etc. to ascertain the role of user participation and value of user’s data. The branch of Physics which explains economic phenomenon is known as Econophysics.


5.2.1 The World Science Festival and Annals of Physics join forces at an event at Columbia University and has been opined in that conference that: “In the era of big data, every scientific discipline must find a way to tackle challenges in storing, handling and interpreting large amounts of raw information. Earlier this month, experts examined how to address that issue in the physical sciences. The panel, titled Big Data and the Future of Physics, was part of the World Science Festival in New York City”.

8 How Big Data Advances Physics –The World Science Festival & Annals of Physics- Columbia University- Marc Chahin- June 27,2017- Elsevier

5.2.2 “As scholarly research is becoming increasingly digitized and data science is taking over many domains, the importance of managing and sharing data is being felt throughout the scientific community”, explained panelist Anita de Waard, Elsevier’s VP Research Data Collaborations, who develops cross-disciplinary frameworks to store, share and search experimental outputs, in collaboration with academic and government groups. “As data, software and ideas become available to everyone, science can take advantage of the network effect to radically accelerate.” Prof. Michael Tuts, the current Chair of the Columbia University Physics Department and an experimental particle physicist, shared his experiences with the ATLAS Experiment at the Large Hadron Collider (LHC) accelerator at CERN in Geneva5. At the LHC, two beams of 100 billion subatomic particles are collided at high speeds in the hopes of finding evidence of new physics. In 2012, the Higgs boson was found at the LHC, leading a year later to the Nobel Prize.

5.2.3 One of the main challenges in running the LHC is handling the vast amounts of data it produces. Prof. Tuts compared the LHC to a 100-megapixel digital camera that takes 40 million electronic “pictures” of the colliding proton bunches per second. The raw data has to be turned into data that can be used for physics analysis, resulting in various separate sets of data that need to be saved on disks and tapes for posterity. All these data are being put on the Worldwide LHC Computing Grid (WLCG), consisting of 167 computing sites located in 42 countries and holding over 200PB (200,000TB) in 1 billion files.

5.2.4 How the “Network Effect” is accelerating science

Anita de Ward pointed out different ways that sharing is taking place, leading to networked knowledge. De Ward summarized the three presentations as describing shared data (on particles and stars), shared software (preserved and “dockized” or wrapped up in preservable containers) and shared ideas. As science thus becomes “deconstructed” in its component parts, it allows the “network effect,” meaning that many more connections become possible between nodes in a network than in the traditional linear stream, where a scientist creates his or her own data, software and ideas in relative isolation. This enables scientific progress to accelerate at an exponential rate: not only can data created by one team be used by the whole world, but new parties can contribute to software and ideas.”


Let us examine some of the arguments for and against these views critically.

5.3.1 In the very beginning of this chapter through an example of the value creation of an orange hanging in the tree in California, it has been clearly determined through the basic principle of Economic Allegiance Doctrine (EAD) that the value is not created until it is consumed by the consumer. This proposition laid down the basis of international taxation. Hence, the argument put forward by Digital Economic Group (DEG), the largest group to be affected when OECD calls for a paradigm shift in approach in international tax, that “consumption does not create value” is in complete contrast to the fundamentals of EAD. This also fundamentally opposes the summary of the G 20 summit call “where economic activities occur and value is created”.

5.3.2 Regarding the issues raised by the Digital Economy Group (DEG) that the value is created not from the data itself, but the value is created through those processes which structure, aggregate, and analyse, and present the data in a manner responsive to the users’ objectives has to be looked deeper from a different angle to understand the entire process. One has to look into the processes by which the Digital Economy Multinational Enterprises (MNE) function-

  • Internet of things, (IoT)
  • Machine Learning (ML)
  • Artificial Intelligence (AI)
  • Algorithm(Alg)

5.3.3 Internet of things (IoT): It is the network of physical devices, vehicles, home appliances and other items , software, sensors, actuators, and connectivity which enables these objects to connect and exchange data. Each thing is uniquely identifiable through its embedded computing system but is able to inter-operate within the existing Internet infrastructure. The figure of online capable devices increased to 8.4 billion in 2017. Experts estimate that the IoT will consist of about 30 billion objects by 2020. It is also estimated that the global market value of IoT will reach $7.1 trillion by 2020.

5.3.4 Machine Learning (ML): It is a field of computer science that gives computer systems the ability to “learn” (i.e., progressively improve performance on a specific task) with data, without being explicitly programmed. The name Machine learning was coined in 1959 by Arthur Samuel. Evolved from the study of pattern recognition and computational learning theory in artificial intelligence, machine learning explores the study and construction of algorithms that can learn from and make predictions on data , such algorithms overcome strictly static program instructions by making data-driven predictions or decisions, through building a model from sample inputs. The computers are fed millions of user’s data to understand the behavioural attribute of a particular user from its experience of learning from the other millions of data and build the process itself.

5.3.5 Artificial Intelligence (AI): Wherein the machines which becomes accustomed on being fed with millions of data can build any user preference of choice likings, dis-likings and can suitably suggest how the particular user will respond to a particular type of situation. Businesses today view AI as systems that can learn instead of systems that are programmed. The pace of The pace of deep neural network innovation means companies can now solve an entirely new set of problems. As of February 2018, there are 2,117 AI start-ups, segmented into 13 categories that collectively raised $29 billion in funding, according to Venture Scanner.

5.3.6 Algorithms(Alg): The characteristics and all the tell-tale details of the users are captured through millions of algorithms done simultaneously. So when 2 users chat among themselves in a social networking site viz. Facebook or WhatsApp, they provide a huge amount of unstructured data to the machines which captures the personal preferences of the user and suggest how the user be approached from the point of customised advertising. The power to evaluate, and influence the choices of the captive user base, is the USP of the big digital companies. These often leads to monopolistic trade practices and even can break the barrier of ethical trade practices. The European Union has levied a penalty of $ 2.99 billion on Google for showing only their own products at the top of the search list bypassing all products of other enterprises.

5.3.7 Whatever be the investment made in Human Resources and otherwise in developing new artificial intelligence, algorithm, neuron networks, this will have no real value if these are not based on the real time data of millions and millions of users to know their preference or choice. Thus, capturing the behavioral attributes in the algorithm is the actual value that determines its value creation or value base.

5.3.8 Value Chains, Value Networks and Value Shops: The OECD report highlights the three concepts of creating of value or profit as above. In the Global Value Chains apt for the digital economies wherein value creation begins with raw materials and proceeds to a finished product as per Stabell and Fjeldesad (1998), two alternative models that of value networks and value shops have been described. The concept of the value network model depends on where value is created by linking users, suppliers or customers i.e. creating network relationship using a mediating technology. This covers all types of multisided platforms. The value shops models are where value is created to resolve customers’ problems, grievances by marshalling of resources viz. Cloud Computing. (pg.19)


6.1 Network Effect: This issue is also highlighted in the

interim report of the OECD dated 11-13, December, 2017. This report illustrates that the success of the business models depends on “network” only whether being a direct network or indirect network. In direct network the value increases as the number of users increases. The Direct network uses the digital platform and benefits from the group of end users. (pg.8). The OECD report also highlights that the digital markets are not competitive and single firms become large enough to influence market prices and it becomes difficult for new firms to gain significant market shares. Here digital products and services disseminate faster, markets clear faster, idea circulate faster and it becomes much easier for businesses to identify, engage and develop their customer base. (pg.10)

6.2 Critical Mass: ‘Critical Mass’ term is derived from Nuclear Physics and it is to measure the minimum amount of a given fissile material necessary to achieve a self-

sustaining fission chain reaction under stated conditions. Critical mass refers to the level of users that are required to help create a set of network effects that are so strong, that they build a moat for that particular business. Sometimes there is no second place in a market when network effects are that strong. In Digital Economy this term can be used to measure the number of users to start the chain reaction of multiplying the user reach which has a multiplier effect and is the USP for Digital MNEs in maximizing profit.

6.3 Research by European School of Management and Technology(ESMT), Berlin, Germany9 shows that the common definition of critical mass in business would be when “diffusion becomes self-sustaining” (Rogers 2003: 243). Critical mass phenomena rely on a rapidly evolving endogenous process over time, e.g. installed base effects driving diffusion even in the absence of price decreases. The link between technology diffusion and installed base effects is well-established (Cabral 1990, 2006; Rohlfs 1974; Kretschmer 2008; Granovetter 1978; Markus 1987), and research identifying multiple stable equilibria separated by an unstable one (Evans and Schmalensee 2010; Economides and Himmelberg 1995; Katz and Shapiro 1985) characterizes the transition from one equilibrium to the other as critical mass. By using the logic of multiple equilibria and endogenous diffusion shows that critical mass – a self-sustaining diffusion process – will only emerge if boundary conditions on the strength of installed base effects, the size of the installed base, and the current market price, are met. The research found that these three parameters are substitutes in terms of reaching critical mass.


6.4.1 Suppose that at each time, t, consumers decide, depending on the net benefit, whether or not to subscribe to a service. The service displays network effects such that the installed base of adopters (subscribers) affects consumer willingness to pay. There is a measure one of infinitely lived consumers with unit demand for the service influenced by consumer type and the installed base of users. Consumer v’s preferences are represented by the willingness-to-pay function u(v,xt-δ), where v is the individual preference parameter, xt- δ is lagged network size at time t, and the

6.4.2 Short-Run and Long-Run Subscription Demand of Critical Mass:

At time t consumer v decides whether to subscribe by considering the net utility from joining

where δ is lag length & u is willingness to-pay function.

There is one market price, If (1) is non-negative, the consumer will join, otherwise not. The consumer indifferent between joining or staying out at time t ( v t )is given by the following equation.

All consumers with v will join. Define

such that H(.) equals the number of consumers in the network at time t. The state equation describing network size at time t, that is, short-run demand, is given by:

In steady state, no consumer can increase utility by joining or leaving; the network stays constant over time, which gives the following long-run demand condition:

Long-run demand is reached when the market is saturated and there are no more consumers to fuel further diffusion. However, long-run demand can also fall short of full saturation depending on prices and consumer preferences.


(1990) analysis of the equilibrium network size path shows that for sufficiently strong network effects and lag length δ approaching zero, the equilibrium adoption path is unique and discontinuous as described by equation (4). Because H(.) maps the change in network size from time t – δ to t, it is convenient to think of it as of a function of lagged network size To see how network externalities and price affect diffusion, we calculate the derivatives of H(.) with respect to lagged network size and price p as below. The slope of H(.) increases in the strength of network effects measured by , as shown in figure 1 below:

Figure 1. Stable vs. unstable equilibria

(I) H is non-decreasing if and only if network effects are non-negative,

(ii) the slope of H equals zero if there are no network effects, and

(iii) the slope of H increases with network effects whenever the density of types is strictly positive.

Consumer v’s willingness-to-pay function is specified as follows:

where c and d are parameters that determine the extent of installed base effects (this is called Network Effects parameters), with the square term capturing possible non-linearities. Installed base is defined as the number of subscribers normalized by population size in a given geographic market (i.e. country) in period t-1

Thus, specification (6) that network are a function of relative rather than absolute number of subscribers facilitate analysis of multiple differently-sized markets. It is assumed that the preference parameter v to be uniformly distributed over with density

The highest consumer type in the population depends on a country’s GDP per head and the unobserved heterogeneity across countries; the extent to which demand reacts to price changes depends on the country’s overall population.

Diffusion Equation (4) becomes:

The structural parameters of this model can be recovered from the following estimation equation:

where denotes the error term, which is heteroscedastic and correlated across time t, but not across markets i. The error term captures the effects of variables that affect subscriptions, but are not observed in data set, e.g. marketing effort of operators, or the degree of non-price competition more generally, in each geographic market. Equation (10) simplifies to a multi-market version of the original Bass model if 3=0 (i.e., price does not matter for network diffusion). To see this, rearrange the terms to obtain:

The left-hand side of (11) corresponds to subscription sales at t and the right-hand side is a square function of cumulative sales through period t-1 with a market-specific intercept, which is a straightforward extension of the Bass model to a multi-market context. In fact, a single market version of (11) matches exactly the discrete analog of the Bass (1969) diffusion equation, as shown below:

The estimates of coefficient in equation (10) identify the structural model parameters as follows: the highest consumer type in the population is identified via two parameters, and the density
of the distribution of types is given by the price parameter,  The network effects parameters c and d are identified via-γ1/3and-γ2/3, respectively. Thus, the network effects in this model are identified by separating the impact of installed base on current subscriptions from the impact of price. More generally, the installed base effect could also be due to other social contagion effects including social learning under uncertainty and social-normative pressures, as has been pointed out for aggregate diffusion models (Van den Bulte and Lilien, 2001; Van den Bulte and Stremersch, 2004). The focus, however, is less on the source of the installed base effect than on its implications for critical mass, and the installed base parameters are estimated to gain insights about critical mass by simulating counterfactual steady states of the diffusion process.

Figure 2 sensitivity of long-run demand to changes in the net effect parameters


The research as above56 shows the importance of the critical mass of end users and the value of network effects based on social networking irrespective of prices for a globally operating digital MNE to be profitable in the global perspective. The installed base of interactive users plays a pivotal role in garnering profit for the MNE after it reaches a stage of diffusion attaining critical mass of users similar to happenings in a nuclear chain reaction. a slight variation of 10% in network effect parameter c results in the long run demand to appreciate by more than 60%, thus creating a multiplier effect after reaching diffusion stage.


7.1 Research by G20 Young Entrepreneurs Alliance10, published by Accenture shows that to Create a dynamic platform ecosystem that enables Digital Businesses to achieve critical mass, the following five distinctive capabilities has to be achieved. One may call it five Ps:

  • a differentiated value proposition,
  • service personalization,
  • market responsive pricing,
  • effective cyber protection,
  • scalability power of ecosystem partners.

7.2 This depends on fostering a supportive enabling environment: There are factors and conditions within the broader economy that are required for platforms to emerge and grow and include

  • digital user size and savviness,
  • public policies.

7.3 It has been observed by OECD that the economic success of the digital business model relies on a critical mass of end users, wherein many of the end users benefit from “free” access to a specific service and the operators of digital platform compensate for this by leveraging data on users and transactions e.g. by selling targeted advertisements to customers on the other side of the markets, advertisers. (pg.12)

7.3.1 Customer data and user generated content contribute hugely to value creation. Without user participation and User Generated Content(UGC) the digital business as it is, cease to exist. A user whether active or passive is contributing in the value creation since the digital platform is capturing his preference and behaviors which can directly be monetized. The target of a digital MNE is to reach a critical mass of users, thus giving it the major competitive advantage and hence, the source of profitability. (pg.45-46)

7.3.2 Conclusion on Critical Mass So, for instance, when a continuously loss making WhatsApp is purchased by Facebook for a whopping price of $ 14 Billion , this amount is paid not only for the machines on which WhatsApp works or the Human Resource of software engineers acquired, or for the tangible assets reflected in the balance sheet, but for capturing the behavioral attributes of billions of WhatsApp users over a long period of time , whose value is not mentioned anywhere, which can be used effectively for future revenue generations by Facebook since WhatsApp possessed huge critical mass of users on real time basis through capturing the behavioral attributes of end users in algorithms.

7.4 Data Monetization “Rather than just being an e-Commerce platform, Alibaba is an infrastructure and data company—which is our strength and also our future. By focusing on big data capabilities, we gain a full, clear picture of buyers and sellers on the Alibaba platforms, and we are able to offer additional services to them.” – Cheng Ouyang, Executive Senior Advisor – Alibaba Group, China

7.4.1 The research by G20 Young Entrepreneur’s Alliance57 shows that Data generated by platforms proliferates—whether from analysis of user experience, behaviors, service consumption or productivity measures. In turn, this creates a multiplier loop, where the value of data multiplies with the number of users and partners in the ecosystem. Platforms enable the gathering of data and the generation of real-time insights on customers, market trends and operations. Indeed, the rich volumes of data and the speed of intelligent service enhancement that is feasible on platforms are possibly beyond reach of the traditional business model.

7.4.2 The largest data-driven opportunity is the ability of a platform to capture value by creating new products and services, improving user experiences, managing risk and increasing productivity. These are avenues of internal monetization where data-driven enhancements are generated within the company. The impact is difficult to measure and the opportunity is often not maximized. Data monetization can also be achieved by providing data-based services to third-parties—which can be a high-margin business for platform players.

7.4.3 Although the largest opportunity will be internal monetization, the potential for external monetization is high if the platform holds unique data and has the capability to package innovative services around that data for third parties While some platforms are primarily transaction oriented and others are strongly data oriented, the opportunity from data monetization is undeniable for all platforms.

7.4.4 Alibaba’s asset-light model means China’s biggest online e-Commerce company can invest in next-generation technologies and services, such as cloud computing and big data, to maintain its competitive edge. Data—and better understanding of it—is integral to the company’s operations. More than 37 percent of Alibaba’s workforce is science, technology, engineering and mathematics (STEM) talent, mainly employed in database management, machine learning and artificial intelligence. The data insights gained are being monetized in a number of different ways. For instance, Alibaba uses data to derive 49 percent of its group revenue from advertising services, including third-party advertisers. “Super-ID” under the Dharma Sword initiative tracks the preferences of 630 million users, the vast majority of China’s internet population. Sellers pay a monthly fee for Alibaba software that they use to analyse relevant data and personalize services for customers. Smart logistics data predicts supply and demand and guides platform sellers to pre-transfer merchandise to designated warehouses where there is strong demand. Finally, Alibaba is developing a unique enterprise credit system by bringing together data on sellers’ financial records, including affiliate Ant Financials’ records, past transactions and information from partners, such as banks. Unique data can be of immense value to businesses and societies outside the platform business.”

7.4.5 The OECD Report highlights that Consumers provide data when they interact with the company’s website or app. The interaction may be active, such as when users create a profile, save items of interest for future reference or make a purchase. It may also be passive, such as when users browse the website or authorize the company to access their browser histories or geolocation data. It may also be possible for the company to access information via other websites or apps open at the same time. Such data collection, and the value that can be extracted from it through data analysis, is an important aspect of the reseller business model.

7.4.6 A reseller extracts value from consumer data in two main ways. First, it may use personalized data such as demographic information as well as data about the consumers’ behavior and product use to understand consumer preferences and, based on these preferences, improve their products and target their marketing at the individual user level. Online stores or apps may be tailored to each individual consumer. Second, a reseller may also use data to engage in differential pricing, charging consumers different prices based on their personal information.

7.4.7 Little is publicly known about resellers’ potential differential pricing strategies. While some companies have denied that they change prices based on personal information, however, anecdotal evidence is nevertheless plentiful (Mohammed, 2017). In a summary of evidence of price differentiation by US retailers, a 2015 Council of Economic Advisors report also outlined three categories of price differentiation strategies:

(i) exploring the demand curve, i.e. conducting online experiments to learn about demand elasticities.

(ii) steering and differential pricing based on demographics, and

(iii) behavioral targeting and personalized pricing (CEA, 2015). To the extent that such strategies are employed, resellers are able to capture consumer surplus for themselves by using data, thereby maximizing profits. (pg. 52-53)

7.4.8 In another example in the ride for hire companies like Uber, Bla Bla, wherein the core of the company’s operations is connecting passengers and drivers using internet connectivity and running the algorithms by which, the company sets prices. Whereas it has computer hardware, servers, knowledge based capital, and intellectual property components but the service depends upon the data describing the precise locations of drivers and passengers. The company collects and stores user data including ride history, user profile details and payment and bank information which leads to product developments and ability to maintain the quality. User data is the input to the sophisticated price setting algorithm. Unlike, the traditional taxi business model where fares are fixed, here fares are fixed by the digital MNEs. This also relies on surge pricing depending upon the demand of the user at a given time on real time basis. “(pg.63)

7.4.9 This point has been highlighted by submission of DET3, wherein they have mentioned that global value chains depend on seamless, dynamic, continuous, information data flow, across the legal entities. So, real time data is the life line of any digital operation and data used for almost everything in Highly Digitalized Business Models (HDBM). Three kinds of data can be used where-

  • Data that enhance MNE operations
  • Data that enhance customer relations
  • Data enabling New Products/Business models

7.4.10 Even for an experiment of world-wide ramification in Large Hadron Collider (LHC) which ultimately detected the existence of Higgs- Boson particle, popularly known as God’s particle, the success depended on collection of billions of raw data and analyzing them across countries, through network effect, as duly acknowledged in the World Science Festival. This once again proves the fundamental principle of the chain of value creation in Science through

  • Experiment,
  • Observation,
  • Inference.

This fundamental principle of science is ruling the thinking of the humanity over ages, cutting across all countries and responsible for human progress. So, it follows logically that, without experiment there can be no observation and without observation there can be no inference. Taking a corollary in the present instance of digital economy it can be inferred that:

  • without experiment, i.e. collection of real time raw data,
  • there can be no observation, i.e. processing the data in Artificial Intelligence and
  • no inference i.e. drawing of algorithm depicting the specific behavioral attributes of a particular user on real time basis.

7.4.11 Any denial to the access of the data source in any country tantamount to denial of experiment, that renders the process of observation and inference i.e. building artificial Intelligence and algorithm, useless. This happened in the case of China which blocked the entry of non-resident MNEs in Digital Economy. So, the non-resident digital MNEs could not draw any algorithm based on data of Chinese population irrespective of their vast access to high end technology. Thus, the argument of the Digital Economy Group that raw data has got no value and only the process of creation with that data has got value, is not acceptable as per global standard of science as well as in economic reality. So, there can be no doubt that data has got enormous value.

7.4.12 Thus, the value is created wherein real time user data is collected, processed and used by various digital companies providing different kind of services with value additions. In this regard the comments made by the BEPS monitoring group on the need for a paradigm shift in approach from the brick and mortar model to the digital model in taxing the digital economy cannot be over stated that the country of taxation is determined by the location of users and not by other standards. Taking the stand taken by BEPS group it can be further stated that the maximum value is created by the collection of raw data or to the access of the raw data source which is consonant with a successful experiment in science and has the maximum value in creating the value chain as observation and inference follows subsequently. Thus, the arguments put forward by the Digital Economy Group and the countries supporting this view cannot be accepted in view of the Principles of Science laid down over ages.

8. Profit Attribution on the basis of Fractional Apportionment

8.1 The concept of fractional apportionment has originally been proposed by the four economists in their report to the League of Nations. Wherein it is difficult to quantify taxable profits correctly within a specific jurisdiction due to economic complexity, then this approach would be very useful.

8.2 Methods based on Fractional Apportionment Another approach considered would be to apportion the profits of the whole enterprise to the digital presence either on the basis of a predetermined formula, or on the basis of variable allocation factors determined on a case-by-case basis. In the context of a significant economic presence, the implementation of a method based on fractional apportionment would require the performance of three successive steps:

(1) the definition of the tax base to be divided,

(2) the determination of the allocation keys to divide that tax base, and

(3) the weighting of these allocation keys. (Para 287)

It is important to note that the domestic laws of most countries use profit attribution methods based on the separate accounts of the PE, rather than fractional apportionment. In addition, fractional apportionment methods would be a departure from current international standards of taxation.

8.3 In view of the discussions on fractional apportionment as above one may think of the Massachusetts formula wherein equal weightage were given to the following 3 factors for profit attribution –

  • Property factor
  • Pay role factor
  • Sales factor

8.4 It is observed that European Union (EU) in their communique dated 25.10.2016 on Common Consolidated Corporate Tax Base (CCCTB) the same apportionment formula with equal weightage but slightly different nomenclature:

  • Assets- Company has in an EU Member Country
  • Labour- (number of employees and employment costs): – the company has in that EU Country
  • Sale – The company made in that EU Country.

Since, this fractional apportionment formula finds wider acceptance in international circles, one may think about the same in Indian perspective also.

8.5 In this respect it may be considered as to why the costs on labour expenditure in the case of an MNE doing business in a digital economy is much lesser than the labour cost incurred by a brick and mortar model MNE of same proportion, the reason being that the users of the digital economy functions as employees in a real sense. So, the difference in the labour cost of a brick and mortar MNE and the MNE doing digital business gives one a fair idea of the user’s contribution by way of reduced cost and increase in profit in the hands of the digital company. So, while attributing profits, a portion of the 1/3rd attributable to labour cost in a brick and mortar MNE would actually come for user contribution.

8.6 Similarly, in the case of assets which in a conventional brick and mortar model relates to only physical and tangible assets but in the case of digital economy MNE, the real value of these assets forms a larger part in intangible assets viz intellectual property rights. The assets of the company also consist of using these assets in an unfettered manner in a country like India which guaranty by law the user of this intellectual property rights and prevents it from getting copied or unlawful uses. These guaranty by the Government of India carries a lot of value which would be understood from the benefit doctrine discussed earlier. So, while understanding the scenario of doing the fractional apportionment, while taking users contribution as a sizeable portion of the 1/3 weightage allocated to Labour costs, the spending done in public domain to build up the infrastructure and the law guaranteeing unfettered use of the IP rights should form part of the 1/3 weightage to Asset.


Is consensus at OECD the only way, even at the cost of sacrificing national interest? It will be important to understand the view of the Chief of International Tax in UN in this regard.

10 “Is a consensus the holy grail?”11

10.1 It seems that a full debate on this issue of limiting consideration to the supply side would probably not lead to a consensus by 2020. And, in fact, to revert to Schön’s analogy, the siren song for developing countries (among others) might be the push for consensus by 2020. Caution and a degree of scepticism are warranted, and developing countries may need to show a willingness to prefer no deal in 2020 over a bad deal that may effectively constrain policy space for decades. Moving from the more readily accepted activities excluded by the term “value creation” to the more contested issues, it should be borne in mind that attempts to achieve a consensus in so large a body of countries as the Inclusive Framework, especially one composed of developing, emerging and developed economies, are unlikely to exhibit complete coherence in either legal or economic terms.”

10.2 IMF in its press release12 observed in this regard, “Directors welcomed the discussion on tax challenges associated with digitalization. They recognized that this is a difficult issue, technically and politically, and that views on whether special treatment is needed, and if so in what form, continue to differ widely.

They noted that the Platform for Collaboration on Tax provides a useful framework for bringing together the IMF, OECD, UN, and World Bank, and could continue to play an active role in supporting international tax coordination.”

10.3 Since India has been the pioneer in championing the cause of the developing world and the entire developing world looks up to what India is deciding, time has come for India to assert its rightful place in the global scenario to get a justifiable deal in the Platform for Collaboration on Tax, not only for itself but for the entire developing world.


1. The team TRPU,

2. Indira Iyer , Chief Director , TRPU

3. Sri Amit Govil , Commissioner, TRPU

4. Vinay Kumar , ex- Director, FT & TR Division ,CBDT

Disclaimer: The opinions expressed in this article is that of the author only. This should not be construed as representing the view of the Ministry of Finance, Government of India.


1 – Harvard Business Review. 2017. Digital Evolution Index. In December 2017, China accounted for almost a fifth of the world’s internet users. However, the internet space is tightly controlled in China and the domestic market is dominated by Chinese tech giants Alibaba and Tencent. Global players in China are disproportionately low. For instance, Facebook users were just 1.8 million or 0.08% of the global share.

2 – BCG. 2016. Digital Influence Study. This study forecast that India with its 650 million internet users, will cross the expected combined populations of the G7 countries (Canada, France, Germany, Italy, Japan, the United Kingdom, and the United States)

3 – Boston Consulting Group. February 2018. Digital Consumer Spending in India: a $100 billion Opportunity. This is a conservative estimate compared to the estimates made by the Bank of America/ Merill Lynch which project the digital economy in India to touch $120 billion in 2020.

4 – OECD -Chapter-II,Digitalisation, Business models & Value creation, Note by the Secretariat,11-13 December, 2017

5 – Proposal for the council directive laying down rules relating to the corporate taxation of a significant digital presence Brussels, dated 21.03.2018 COM(2018) 147 final.

6 – OECD- Tax challenges of the digitalization- comments received on request for input, PART I pages 137 to 150, dated 25.10.2017

7 – OECD – Tax challenges of the digitalisation- comments received on request for input– PART I pages117 to 136, dated 25/10/2017

8 – How Big Data Advances Physics –The World Science Festival & Annals of Physics- Columbia University- Marc Chahin- June 27,2017- Elsevier.

9 – Identifying Critical Mass in the Global Cellular Telephony Market, Michał Grajek , European School of Management and Technology, , Schlossplatz 1, 10178 Berlin,

1 Chief, International Tax Cooperation Unit, financing for Sustainable Development Office, United Nations. This paper was written while on sabbatical at the Oxford University Centre for Business Taxation. The comments of Professor Michael Devereux of the Centre on an earlier version of this paper are especially acknowledged; however, the paper reflects only the personal views of the author.

12 The IMF Press Release No. 19/69 ; March 10, 2019;

IMF Executive Board Reviews Corporate Taxation in the Global Economy

Source- CBDT Taxalogue Magazine Jul – Oct 19 | Volume 1 | Issue 1

Join Taxguru’s Network for Latest updates on Income Tax, GST, Company Law, Corporate Laws and other related subjects.

Join us on Whatsapp

taxguru on whatsapp GROUP LINK

Join us on Whatsapp

taxguru on whatsapp GROUP LINK

Join us on Whatsapp

taxguru on whatsapp GROUP LINK

Join us on Whatsapp

taxguru on whatsapp GROUP LINK

Join us on Whatsapp

taxguru on whatsapp GROUP LINK

Join us on Whatsapp

taxguru on whatsapp GROUP LINK

Join us on Whatsapp

taxguru on whatsapp GROUP LINK

Join us on Whatsapp

taxguru on whatsapp GROUP LINK

Join us on Whatsapp

taxguru on whatsapp GROUP LINK

Join us on Whatsapp

taxguru on whatsapp GROUP LINK

Join us on Telegram

taxguru on telegram GROUP LINK

Download our App


More Under Income Tax

Leave a Comment

Your email address will not be published. Required fields are marked *

Search Posts by Date

November 2023