TRAI’s fresh thoughts on net neutrality

BY SMARIKA KUMAR| IN Law and Policy | 28/05/2016

In its series of consultations and policymaking efforts around net neutrality in India, TRAI has issued yet another Consultation Paper last week, on May 19, 2016. This time, the paper focuses on the construction of models for the provision, distribution and use of free data (in other words, zero-rated data or zero rating) in digitally connected India.

This consultation comes in the wake of the Prohibition of Discriminatory Tariffs for Data Services Regulations, 2016 in February. That TRAI-developed Regulation had in essence, issued a ban on the offer or charging of data services “on the basis of content” through “discriminatory tariffs” (see Section 3.(1) of the Regulation). This move effectively put an end to projects like Internet.org/Free Basics and Airtel Zero in the country. And for a while, it looked as if a regulation designed with this wording would spell the death knell for zero-rated services.

So it is quite exciting that this new Consultation Paper from TRAI is seeking to think about working models whereby zero-rated services, can, in fact be offered, without of course violating the said prohibition on “discriminatory tariffs.”

Broadly, the Consultation Paper seeks responses to the question of how models incorporating  zero-rating services or free data for Indian citizens can be designed while also being wary of giving unlimited internet content gatekeeping power to Telecom Service Providers (TSPs). In other words, as the Paper puts it: “Is there a need to have TSP agnostic platform to provide free data or suitable reimbursement to users, without violating the principles of Differential Pricing for Data laid down in TRAI Regulation? Please suggest the most suitable model to achieve the objective.” (see page 7, Question 1, Consultation Paper.)

This need to imagine different models for the provision of zero-rated services, while circumventing a scenario of “discriminatory tariffs”, constitutes the central concern of the Consultation Paper. The paper sees such a provision of zero-rated services as crucial to giving “consumers more choices for accessing the internet” (see para 9, Consultation Paper.)

In continuing this discussion on pushing the boundaries of zero-rating imagination, the Consultation Paper suggests three models that offer alternatives to “discriminatory tariffs”:

Model 1: The Incentives-to-Consumer Model

The first model consists of rewarding the users of a particular application or website with a recharge for data or voice usage, irrespective of the TSP they get their internet services from. The paper quotes some examples of such Incentives-to-Consumer models in India as mCent, Gigato, Taskbucks, Ladoo, EarnTalktime, and Pokkt. It outlines a second kind of consumer-rewards models in internet platforms which tie up with other apps to offer mobile data rewards in exchange for activities like paying one’s electricity bill on time or for checking out, say, out of a hotel on time (see para 12, Consultation Paper.)

Significance of Model 1:       

• This model is useful for enabling “smaller entrepreneurs to flourish without permitting gate keeping function in the hands of the TSPs” (see para 9, Consultation Paper.) This is because it manages to offer incentives to citizens to use particular websites or apps which offer such rewards, thus giving internet businesses the freedom to employ this technique to attract more consumers and render themselves more competitive in the market.
  
  

• By suggesting the Incentives-to-Consumer model as one of the alternatives to Internet.org/Free Basics to provide free data to citizens, TRAI does manage to clarify that businesses which use the Incentives-to-Consumer model will not be rendered illegal as per its February Regulation, as this kind of arrangement will not fall within its category of “discriminatory tariffs.” The scope of what constitutes “discriminatory tariffs” is thus also elaborated upon by this new Consultation Paper. The policy implication of this is that zero-rating will still live yet in India, just not in its “discriminatory tariff” form.

Limitations of Model 1:

• The issue with an Incentives-to-Consumer Model, however, is that it enhances internet accessibility for only those citizens who already have at least some access to the internet. That’s because to avail themselves of the “rewards” of free data or voice under this model, consumer-citizens first need access to the relevant apps or websites, viz. internet. The paper itself recognizes this limitation when it says that the Incentives-to-Consumer model “requires an action before the reward thereby excluding the mobile users who have low or zero balance” (see para 13, Consultation Paper.) However it does not cater to the citizen who does not at all have internet access either due to reasons of affordability or the lack of quality infrastructure, or because of the lack of infrastructure altogether. The larger internet accessibility issue in the country thus still remains unaddressed under this model.

Model 2: The Toll-Free API Model

The second model suggested in the Consultation Paper is that of a “toll-free” API (Application Programme Interface, see para 13, Consultation Paper.) The paper seems to suggest that this would operate on similar lines to the Incentives-to-Consumer model, except instead of “rewarding” the user of the app/website, the model would enable no charging for data while the consumer is using the website in question. And of course, it also outlines that such an arrangement would be functional irrespective of the TSP service the internet user-citizen chooses to use to access the app/website.

Significance of Model 2:       

•    Like the Incentives-to-Consumer model, the Toll-Free API model allows private businesses, (especially the smaller entrepreneurs and homegrown startups which TRAI seems to be concerned about) the requisite freedom to add competitive advantage features to their services and offers. If a toll-free or no-data-charge feature for an app works as an incentive for consumers to use that app, then it is a clear win for the said app enterprise.

•  The Toll-Free API model has the added advantage that it has the ability to reach the consumer-citizen “without impacting the mobile bill of the consumers” (see para 13, Consultation Paper.) This means that it can also enable the relevant app/website to reach the citizen who cannot afford to pay for internet or data charges in the first place.

•  It is easier to figure out the “true cost” of data used for a particular website/app under this model, as compared to the Incentives-to-Consumer model. As the Consultation paper explains: “Suppose Person A is on 1GB datapack costing Rs. 250, and Person B is on the tariff of 10 paise per 10 Kb. If the data consumed in one session of Internet surfing is 50 MB, Person A incurs Rs. 12.20 whereas Person B incurs Rs. 512. This is nearly 42 times the cost incurred by Person A. In the toll-free model, given the data used is not getting billed to the mobile subscriber, true cost of data is covered without any confusion” (see para 13, Consultation Paper.)

Limitations of Model 2:

• It is unclear how payments might be figured under this model. Even if the so-called “true cost” of the data is calculated, the question of what kind of arrangement might exist between an app service (or, over-the-top service, or OTT) and between a TSP or between several TSPs to settle this “true cost” remains a mystery. The issue of how to structure compensatory payments between different TSPs (or paid peering) and between TSPs and OTTs lies at the heart of the entire debate around net neutrality, zero-rating and (the ban on) Free Basics (see the work of Visal Misra on this co-operative settlement, and discussion around paid peering in Jean Tirole, Competition in Telecommunications. See also, discussion on page 19, Report on Consultation on Net Neutrality in the Indian Context: Reality v. Rhetoric). Given this, it is unclear how Model 2 hopes to address this core issue. 
  
  

• In the absence of clarity on how the Toll-Free API model proposes to process payments between TSPs and OTTs, the most obvious market solution (certainly, the most popular yet) might, in fact, be the facilitation of payments for the “true cost” of data used by a consumer by an OTT to the relevant TSP. This would presumably require the OTT to enter into an agreement with the TSP of a nature similar to what was seen in the case of Reliance and Free Basics/Internet.org last year. And that, in turn, would violate Section 3(2) of the February Regulation issued by TRAI which states: “No service provider shall enter into any arrangement, agreement or contract, by whatever name called, with any person, natural or legal, that has the effect of discriminatory tariffs for data services being offered or charged to the consumer on the basis of content.” And eventually, such arrangements will also lead to gatekeeping of content and apps by TSPs, which is what TRAI is looking to avoid in the first place (see para 8, Consultation Paper.)

• It is difficult to see how Model 2 will enable the coverage of the cost of the “toll-free” data used by the consumer without resorting to scrutinizing and separating what a consumer is browsing on the internet on the basis of “content,” and then charging for it to the relevant OTT/app. For example, if consumer C browses a proposed toll-free app P for 13 minutes and toll-free app Q for 20 minutes on her TSP/ISP network R, with the true cost of data she uses in each case being Rs x and Rs y respectively, then R would need to be compensated by someone (most obviously, as noted in the previous point, by P and Q) to the tune of Rs x and Rs y each. Which means that in order to calculate what it is owed, R would need to scrutinize what exactly C is browsing on its network and how much time C is spending on each OTT/website she uses. And since a private entity, R, in this case will be scrutinizing the internet browsing habits of every consumer, such a scenario raises serious concerns over the privacy of citizens.

Model 3: The Consumer-Subsidies Model

The third model discussed by TRAI in the Consultation Paper lays out a scenario where a citizen pays for getting a basic internet connection, and then he/she is offered subsidies for data through relevant sources for his browsing needs. This is envisioned as similar to the subsidies provided for domestic LPG connections currently to eligible citizens: “The direct money transfer approach could be similar to the subsidy payment, for the domestic LPG connections, wherein the user pays for the connection like any other normal connection, and then the Oil Company/Government pays the subsidy directly into his/her bank account” (see para 14, Consultation Paper.) In other words, this Model 3 seeks to offer reimbursements to citizen-internet users for data already consumed by them.

Significance of Model 3:

• Like Model 2, the Consumer-Subsidies Model has the advantage that it allows the calculation of “true cost” of the data used by a citizen-consumer, while also seeking to be TSP-agnostic. As TRAI states, “However, just like in Toll Free Model, the Platform owner not only measures the real time data consumption but also the tariff that is being applied to each individual user and reimburse/recharge actual amount incurred by the user in the form of a recharge for data usage or for voice usage to the user” (see para 14, Consultation Paper.)
  
  

• What is most interesting about the Consumer-Subsidies model is that it seems that TRAI is envisioning access to the internet almost as a basic necessity, or a public good for citizens, on the lines of domestic cooking gas. This could possibly imply that access to the internet, or more simply, data charges, can be subsidized, like LPG, by the State for its citizens.
  
  

Limitations of Model 3:

• When discussing any subsidies model, the immediate question which jumps to one’s mind is who should bear the burden of such subsidies? This same question also becomes relevant for this Model 3. Should it be the State which should reimburse citizen-consumers for their internet data charges? Or should such subsidies be achieved through some sort of free-market mechanism and the development of innovative business models? Or should it be a model that is a combination of both State-based regulatory interventions and free market mechanisms to incentivize private players (either TSPs or OTTs) to subsidize or reimburse data uses by citizens? If so, what should be the design of such a model? And wouldn’t designing such models run into the same issues around compensatory payments and paid peering as encountered under Model 2? These questions remain unanswered, and frankly, these are the questions TRAI is seeking an answer to, through the Consultation Paper.
  
  

• If Model 3 is to work like the LPG subsidies model, then one seriously wonders where the State is looking to raise resources from to cover this. It seems that the burden will ultimately fall upon the taxpayer. Plus, the fact that the State is already urging “well-off” citizens to give up their LPG subsidies to the benefit of the “poor,” is a clear reminder that State resources to subsidize basic citizen needs are hardly infinite. It is the precise reason why the other “LPG” (Liberalisation, Privatisation, Globalisation) policy had to be adopted in the 1990s.
  
  

• Even if it is decided that the State will only subsidise or reimburse the data charges of the “most needy,” the issue of how the State defines this category remains an open question. And this dilemma of classification in law when technological conundrums arise, is an old, persisting one, that does not look as though it will have sustainable solutions anytime soon.
  
  

• The Consumer-Subsidies Model additionally, like Model 1, only benefits those who can already afford a basic internet connection, or access, as it focuses around reimbursements, rather than “no charge.” Therefore, for citizens who no access to internet infrastructure, or who just cannot afford to pay for the internet even the first time, Model 3 holds an empty promise.
  
  

The new Consultation Paper certainly indicates a desire to create nuances in the design of the regulatory framework for internet access, infrastructure and net neutrality. In this, TRAI has been ambitious. However, the three models suggested in the paper fail to live up to this ambition. There is a lot of work and a lot of gaps still to be filled if any of these models - improved versions of them - are to cater to the public interest in this country comprehensively (see point 5/last, here.)

TRAI is asking us for our help for reimagining this design for the internet. How can we help? The deadline for comments is 16 June 2016 and comments should be sent to broadbandtrai@gmail.com.

Smarika Kumar is an independent legal researcher and was formerly with the Alternative Law Forum, Bangalore.