- NPD regulations might force companies to share data, but data might not be usable or valuable to others
- Regulations might disincentivise companies from collecting data and innovating
- No other country has proposed mandatory data sharing
- NPD regulations will run into trouble with intellectual property laws
- Defining public value is difficult and needs to be explored on a case-by-case basis
The non-personal data (NPD) framework proposed by the expert committee set up by the government might not be the right approach to address antitrust concerns because it will change incentives for companies collecting data and the data shared will not really be useful to other companies, panellists speaking at MediaNama’s event on Regulating Non-Personal Data said on February 18, 2022.
The speakers in the panel on Competition, Antitrust, and NPD included Astha Kapoor from Aapti Institute, Shruti Aji Murali from Indian Competition Watch, Rahul Singh from Khaitan & Co, and Viswanath Pingali from IIM Ahmedabad. Abir Roy from Sarvada Legal and Nikhil Pahwa from MediaNama also shared their insights during the session.
This event was organised with support from Google, PhonePe, Amazon, Meta, and Microsoft. To support future MediaNama discussions, please let us know here.
Will NPD sharing be useful?
Firms might be forced to share data, but will that data be useful? Pingali explained that even though the NPD framework might mandate firms to share data, this data will not be really useful for the others because firms will be extremely tactical.
“I’ve been working a lot on some of the research projects going to be published shortly and the inspiration came from some other paper. So we wrote to the author of that paper saying, can you please share the data that you’ve used in your publication and then it turns out that the file we got from the chap, with due respect, is simply unusable. We had to completely recreate our own data. By the way, that person was obligated to share the data because the journal requires that the data has to be stored in a public repository.” – Viswanath Pingali, IIMA
What happens once the data is shared: Kapoor said that the expert committee’s report was not addressing questions like will it enhance customer value, increase market access, promote symbiotic collaborations, diversify the economy, and encourage innovation. The focus of the report was largely on the sharing of data and not on broader questions of what happens once the data is shared, Kapoor said.
Without more clarity, do not know if this will help small businesses: One of the features of a good law is consistency and clarity and without clarity on who’s going to share this data, who’s going to hold this data, what data is going to be shared, how is the data going to be shared, we do not know if it will be helpful for startups, Murali remarked. Just because you ask companies to share data, doesn’t mean there will be benefits for the digital economy, Murali said.
- Look at specific use-cases: Any kind of rules need to be evidence-based and with empirical backing to it, because you do not want to impose a data-sharing remedy that is either meaningless or effectively ends up actually strengthening our data monopolies, Murali said. She also added that we should start looking at specific contexts and use cases of data sharing.
“It needs to be considered in a specific context. I don’t think we would achieve our objective, our stated, legislative objective, or the report’s objective, if we were to sort of introduce very broad rules that have impacts that we don’t yet know. I think we start with use cases. We start with health for instance. All governments that have looked at harnessing AI, harnessing data, have kind of started with a particular problem, and then sort of brought in stakeholders, brought in technology to kind of address the problem. Simply mandating will create a whole new monster that we might not actually know what to do with it.” – Shruti Murali
- Take a broader ecosystem-led approach:
“What we suggest as an alternative is to take a much more broader ecosystem led approach that actually involves different stakeholders, creates incentives, builds possibilities of innovations coming up. It cannot just start and stop with acquiring data. You have to create the systems that allow you to share data meaningfully […] and then sort of from the create user benefit, societal benefit and stakeholder benefit.” – Astha Kapoor
Role of data in competition
Network effects are not the be-all and end-all of any business: “At one point in time, we all thought that network effects are be-all and end-all of any business. I think the ground reality out there is a little different. I think we know that there are gradations of network effects. And not all network effects are equal in every sense of the word. Data being an entry barrier will also fall into that category. Sooner than later we’ll start realising under what circumstances data is an entry barrier and under what circumstances data is not going to be an entry barrier,” Pingali said. “A recent Harvard Business Review article does make a distinction between what is a localised network effect and what is a global network effect. And how localized network effects need not be as strong an entry barrier as we always thought they were right. And therefore, the argument that data is an entry barrier is, once again, not black and white at all,” Pingali added.
“We all know that Uber has competition all over, we know that most of the platforms have significant amount of competition, third movers, fourth movers, fifth movers are also creating a niche for themselves.” – Viswanath Pingali
Is data really a barrier to competition? Murali wondered if data is really such a strong barrier to competition that you need to mandate sharing. I’m doubting that because data needs a particular context, data needs a particular infrastructure, out of which you derive value, Murali explained. Giving the example of how CCI gave approval to Facebook’s and Google’s investment in Jio despite concerns that two massive data businesses are investing in a third massive data business, Murali said that CCI does look at these things and judges them based on what are the specific datasets and what the companies will do.
CCI approach is better because of unintended consequences of NPD regulation: Taking the same example of Facebook and Google’s investment in Jio, Pahwa asked if the NPD approach is trying to address a limitation that the CCI has because CCI needs not only proof of dominance but also abuse of dominance before ruling something as anticompetitive. Responding to this, Murali said that she doesn’t think it’s a limitation and that the NPD approach is broadly a type of ex-ante regulation that is receiving pushback globally because of the unintended consequences that it might result in such as disincentivising people from collecting data in the first place.
- Experiment cautiously: Pingali suggested that we cautiously experiment with minor bits since our understanding of network effects and data monopolies is very limited at the moment.
- Review Competition Law: Even if CCI or the current regime has shortcomings, NPD regulation is not the solution here and Competition Law needs review, Pingali said.
Why mandatory NPD sharing is not the right approach
No other country proposes a mandatory sharing mechanism: Kapoor gave examples of EU, Japan, Australia, and said that data sharing proposals in these jurisdictions do not advocate for mandatory sharing in a blanket manner like India’s framework. ” There are limits in terms of certain sectors, there are limits in terms of certain solutions, there are limits in terms of certain sizes of companies, etc,” Kapoor said.
“Making things mandatory from day one, and not creating an external ecosystem to enable it will also just mean that people will start to scuttle the rules and there will be delinquency on many fronts.” – Astha Kapoor
There isn’t enough data sharing at the moment because of lack of incentives: The first thing that we diagnosed about why there isn’t enough sharing at the moment is because of the lack of incentives in terms of infrastructure to share such as standards and interoperability, mechanisms like data stewardship that helps share data meaningfully, financial incentives, public value incentives, Kapoor said.
“There isn’t an easy way of sharing data, there isn’t a reliable way of sharing data, there isn’t the public value reason to share data, the government doesn’t reciprocate the sharing of data from the private companies so all of this means that the broader ecosystem for data sharing in a voluntary way is broken.” – Astha Kapoor
Cannot be as general as public interest: Compulsory sharing is a huge issue because you do not want to scuttle innovation and to create datasets is a mammoth task, Roy said. For the government or anybody to mandate sharing of this data with a third party, there have to be certain parameters and it cannot be as general as public interest, Roy said.
- Explore how mandatory sharing will work: We need to step back and explore if we can mandatorily require firms to share data because firms will strategically react to government restrictions, Pingali said.
- Build a system on incentives instead: We need a voluntary system built on incentives rather than mandatory sharing forced by the government, Kapoor suggested.
“Data sharing for the sake of data sharing is not a reasonable solution. So let’s build trust, let’s build reciprocity, let’s build a system where private companies that share data also receive data from the government and other stakeholders, and what that looks like as a broader ecosystem and in that case voluntary will occur.” – Astha Kapoor
How incentives for companies collecting data will change
Will there be an incentive for data collection in the first place? Companies collect data and derive insights from it to grow their business. But if the company has to give the same competitive advantage to others by sharing this data, will there be an incentive for them to collect data in the first place? Pingali asked.
“What I collect right now is for me, and for my clients, so that the platform becomes stronger, the chain in the platform becomes stronger in every sense of the word. There may be lots of elements that I wouldn’t even bother collecting if it turns out that I’m asked to share with everybody. So the question therefore is, has it helped any business, as a result of that.” – Viswanath Pingali
What will happen if Bloomberg data is democratised: When asked if Bloomberg data should be democratised and made public and accessible because there is value in the data they collect, Pingali said the company would simply stop collecting data then because there is no incentive for them.
“Today, I’m able to buy the data for a price. And therefore a market gets created for data. Bloomberg is a very good example when data need was there, somebody came up and established this particular portal. […] The moment you start going down that route of democratizing Bloomberg, will Bloomberg remain what it is? No it won’t. What would you rather have: a good quality Bloomberg available for a price or a Bloomberg which you can access anytime you want, but [of a lower quality].” – Viswanath Pingali
How will NPD regulations interface with the intellectual property laws
Information is different from data and data is intellectual property: Pingali distinguished between information and data, defining information as what is collected and data as what is processed from the information by making sense of it and getting it into a format that is usable. Data is intellectual property according to me, Pingali said.
Law will face challenges with the IPR regime: If the NPD regulations are formed on the basis of the expert committee report as it is, the first challenge will be legal conflicts with the intellectual property rights regime because you’re asking a company to do compulsory licensing, Roy said.
Data gives competitive advantage: From a competition perspective, there is no specific protection for data, but there is some effort that goes into collecting data and transforming it, which gives a company its competitive edge, Murali said. Data that usually gives a company a competitive advantage would be stuff that is more granular rather than less, Murali added.
Mandatory sharing of data is only practical in some circumstances because you are asking a company to share its property: When you ask a company to share NPD, you’re asking it to share its property, Roy remarked. This is why many countries have found it challenging to find remedies for this and the EU proposed that sharing data can only be mandated provided that sharing is indispensable for the creation of a new product or a new market, Roy explained.
“Just to give an example, in one of the cases what had happened was, you have those directories of TV guides. Where in say suppose 6:00, this show will come, 6:30 that will come. So this used to come on a daily basis. So somebody wanted to come up with a weekly guide kind of thing. So they approached the commission saying that these people are not giving me that information. So there the EU Commission actually said there is a demand for a weekly guide to be created. So in that very niche of circumstances, there was a compulsory licensing which was done, saying that you have to license.” – Abir Roy
Trade secret law is not a solution: Any kind of IP law such as patents and copyrights will provide a limited monopoly for a few years, which is around 50 to 60 years for copyright and 14 to 20 years for patents. This is why companies tend to keep some aspects of their businesses as secrets and not copyright or patent them. “The moment you want to get that limited monopoly, you will have to disclose certain things and therein lies the conundrum for non-personal data. If you say that tomorrow, some trade secret law will come in, presumably on similar lines, there’ll be some limited monopoly idea, but in order to get that limited monopoly, you will have to share something. But by fundamental definition, if you want to keep it a secret you don’t want to share with anybody else, why will you share and so it looks like a circle,” Singh said.
Defining public value
According to the NPD framework, organisations can request high-value data sets only if they are building something that brings public value, but the expert committee doesn’t really define the term public value or layout how to determine if some initiative is going to increase public value apart from giving few examples like education, healthcare, etc. While the panellists concurred that defining the term is difficult, they still attempted to do so:
- Consumer welfare: From a competition perspective we can draw parallels between public value and consumer welfare but there cannot be a static definition for the term, Murali said. And consumer welfare itself has taken several twists and turns over the years in different jurisdictions and it’s not merely low prices or more products, Murali added.
“I would go back to identifying, to at least narrowing it down to a particular sector, to a particular problem, and trying to at least identify the stakeholders there. And seeing what the problems there are that can be solved through data sharing. Data sharing might not actually be the solution in many cases.” – Shruti Murali
- Public interest: We can draw analogies and see whether we can make sense of this idea of public value from other concepts like “public interest,” Singh said. But the jurisprudence around the term is not clear as it doesn’t mean whatever the public is interested in, and it might be something counterintuitive to whatever the public might be interested in. “The public in India might be interested in lynching. But that can’t be public interest, that can’t be public value either,” Singh explained.
“For people who have done law, there are all sorts of judges who would say public interest is an unruly horse. It’s quite possibly it’s very difficult to get on maybe, and it’s also difficult to get off it.” – Rahul Singh
- Participation, protection, agency, transparency: We sort of hazarded a guess as to what it could be and we thought it’s like participation, which is the ability of individuals and communities to exercise rights in decision making, Kapoor said. In addition to that, protection, agency, transparency, and then equitable benefit, were some of the characteristics of public value that we define, Kapoor added.
- The betterment of society: I can think of public value as the betterment of society from the current status quo where the information sharing happens in a much more smooth way than you have thought businesses are capable of, Pingali said. But I would leave it to businesses to create mechanisms for public welfare to go up, rather than force them to, Pingali added.
“I’ll refer to another economist called Ronald Coase, who has talked about how when the transaction costs are limited, who owns property rights does not matter. What is socially optimum will always happen. That’s the whole crux behind the modern day carbon trading and all that stuff. If you go by that particular Coasean view of the world, you know that if there is a betterment of public value, the aggregate value that the society incurs as a result of certain activity, the market will create a mechanism. So therefore, the argument comes back to, the regulation ought to reduce the transaction costs between various people, various entities.” – Viswanath Pingali
- Intangible concept: Public interest is a very intangible concept and you cannot put a straight jacket to it, Roy said. For example, you might have to share health data because it brings health benefits but this data to a pharmaceutical company is the be-all and end-all of their business, Roy explained.
- Look at it on a case by case basis: We can leave it to the courts to look at this on a case by case basis, but this will have other problems because if the law is not sufficiently specific then society will pay a lot of variable costs down the line, Singh said.
Government access to NPD
How far will government access to NPD go? One of the examples of how NPD could be used by the government was that Zomato data could be made available to the health department for health interventions. For example, which area is consuming more sugar, targeted interventions can be carried out. Commenting on this, Pingali asked where would this stop. “So tomorrow, maybe every one of our car tires needs to have sensors in it, that will tell the government exactly where you go to every time. You know, that saves a lot of things, there is a public value, if you think about it, from that way,” Pingali remarked. “Where do we start prioritizing individual rights over public welfare,” Pingali added.
Paternalistic approach: Speaking on the same Zomato example, Singh commented that healthcare interventions by the state based on health data from Zomato is a paternalistic approach and paternalism is never really a good solution in law. “At least in a democratic context, whichever society believes in the autonomy of individuals, they also believe that health issues would be a personal concern,”
Government is the biggest data monopolist: “When somebody talks about data concentration, the biggest concentrator and data monopolist is the government. Why don’t they start with sharing their data mandatorily before they are asking somebody else to do it?” Singh asked.
- A Guide To Non Personal Data Regulation In India
- Summary: Revised Draft Report On Non Personal Data
- MP Amar Patnaik On Non-Personal Data: Different DPAs Would Impede Protection Of Citizens’ Rights #NAMA
- Data Protection Bill: Financial Data, Non-Personal Data And Algorithmic Transparency Should Be Regulated Separately #NAMA
- Data Protection bill 2021: How the JPC wants to deal with non-personal data
Have something to add? Subscribe to MediaNama here and post your comment.