ANIRUDH BURMAN, UPASANA SHARMA
Data localization refers to various policy measures that restrict data flows by limiting the physical storage and processing of data within a given jurisdiction’s boundaries. Multiple countries have adopted localization policies to combat multiple concerns over the free flow of data. A vital question, then, is whether any particular variant of data localization would help the Indian government meet its multiple stated objectives for considering such a policy course.
There are four key types of localization variants. These include (a) conditional localization that entails a local storage requirement, (b) unconditional local storage requirements (for all personal data), (c) unconditional mirroring requirements (for all personal data), and (d) the unconditional free flow of data with bilateral/ multilateral agreements for data access and transfers. This paper breaks up these four variants further into a total of nine specific designs and evaluates which would best serve India’s objectives.
THE CASE FOR DATA LOCALIZATION IN INDIA
Data localization has become a significant policy issue in India in the last decade. This is primarily due to the perceived economic benefits of processing Indian consumer data, and difficulties accessing personal data for national security and law enforcement purposes. In 2019, the Indian government introduced a data protection bill in the Indian parliament, which is still being debated and considered. This bill proposes the country’s first economy-wide data localization framework. That said, more tailored, sector-specific data localization measures have already been implemented in many parts of the Indian economy. For example, the telecommunications sector already requires the local storage and local processing of subscriber information and prohibits the transferring of subscribers’ account information overseas. Most recently, India’s central bank, the Reserve Bank of India, has mandated that all payment data be stored in India, though it can be processed abroad.
Anirudh Burman is an associate fellow at Carnegie India. He works on key issues relating to public institutions, public administration, the administrative and regulatory state, and state capacity.
The Indian government’s motivations for enacting data localization have been articulated in multiple government documents. One especially noteworthy document is a report produced by the Committee of Experts under the chairmanship of Justice B. N. Srikrishna. The report has provided detailed reasons for proposing the localization of personal data. The same committee then formulated a legislative proposal based on its findings in the form of a 2018 draft known as the Personal Data Protection Bill. The Indian government introduced the 2019 bill in the Indian parliament based on this draft.
The Indian government has given four stated objectives for pursuing data localization:
securing faster and better access to personal data for law enforcement
increasing economic growth and boosting employment
preventing foreign surveillance and
better enforcing data protection laws
A central question is what kind of data localization measure, if any, is best suited to meet the government’s stated objectives? This paper sought to answer this question by studying each of the state’s objectives to understand whether any specific design of data localization measure would be most suitable.
FINDINGS
At the outset, this analysis revealed that data localization measures do not further two of the stated objectives: preventing foreign surveillance and better enforcing data protection laws. This is because localization does not enable the Indian government to further legal claims for accessing data. The best way to make such legal claims is to do so directly by establishing regulatory jurisdiction. The paper therefore focused on assessing the government’s other two objectives—securing data access for law enforcement agencies and increasing economic growth.
First, Indian law enforcement agencies face major constraints in getting timely access to data. This is usually because the data that law enforcement personnel is seeking is collected in India and stored in another jurisdiction, leading to a conflict of legal systems. Currently, governments go through legal instruments known as mutual legal assistance treaties (MLATs) to access this data. However, this process is considered cumbersome, as on average it takes ten months for governments to gain data access in this way.
To measure which type of localization measure would best help law enforcement personnel get improved access to personal data, four criteria were used: (a) the scope of data accessible after localization (31 percent weight), (b) the speed at which data would be accessible to law enforcement (50 percent), (c) the countervailing risk of losing foreign businesses that provide data-related services in India due to localization requirements (5 percent), and (d) the countervailing risk of retaliatory action against Indian firms by foreign governments (14 percent). Each localization measure was scored on these criteria to tabulate aggregate scores for ranking the different kinds of localization measures.
The scores highlight that restrictive localization options do not actually increase the speed of access or the scope of access to different types of data because foreign businesses are still required to comply with the laws of their home countries while serving consumers in India.
Localization measures that impose minimal restrictions on the flow of data, combined with bilateral or multilateral arrangements for access to data, scored the highest. The current status quo is the most suboptimal—the policy option of allowing free-flowing data while keeping the cumbersome MLAT process scored the lowest. Furthermore, the 2019 bill before the Indian parliament is not well-designed to meet the needs of India’s law enforcement agencies. (See the full paper, especially appendix 2, for a full explanation.)
Upasana Sharma is a program coordinator and research assistant at the Political Economy program at Carnegie India.
Second, it is commonly asserted that localizing the data of Indian consumers within India would likely spur economic growth and support innovation in India more than allowing data to flow freely. To gauge which localization measure would best help increase economic growth by benefiting Indian producers and the larger economy (compared to the status quo of free-flowing data), four criteria were used. These include (a) increasing demand for local goods and services (36 percent weight), (b) giving Indian firms a competitive advantage (38 percent), (c) the countervailing risk of loss of businesses related to Indian data (18 percent) and (d) the countervailing risk of retaliatory action against Indian firms from foreign governments (8 percent). Again, each localization measure was scored on these criteria to tabulate aggregate scores so that the various localization options could be ranked.
The localization measures that impose a local storage requirement and permit the global processing of data scored the highest. On the other hand, the policy status quo is suboptimal—policy options featuring the free flow of data with cumbersome MLATs again scored low. Based on these findings, the 2019 bill before the parliament is not well-designed to best meet the government’s stated objective of spurring economic growth and innovation. (The full paper, especially appendix 3, contains a full explanation.)
CONCLUSION
It turns out that data localization and data access cannot be equated. Localization policies premised on the mistaken assumption that they can be are not likely to serve their purpose and may result in unintended costs. Specifically, law enforcement personnel must gain the data access they need by either establishing an explicit claim of jurisdiction or by solving jurisdictional conflicts through bilateral or multilateral arrangements. The best alternative for India to enable higher economic growth and innovation involves a localization framework that mandates a local storage requirement but allows global processing of data.
Policymakers must work toward reconciling the differences and tradeoffs between different localization frameworks. While these findings did not seek to exhaustively replicate all the localization proposals in the Indian data protection bill, this paper examines stylized policy alternatives designed to give stakeholders a clear idea of the relative advantages of various localization options. Another caveat is that underlying facts are likely to change rapidly due to the nature of technological progress and regulatory developments in domestic and international settings. This paper offers an in-depth explanation of these findings as a substantive contribution to ongoing policy discussions about the merits and drawbacks of different localization approaches.
INTRODUCTION
Data localization has become a significant policy issue in India, as it has in many countries. Data localization refers to various kinds of policy measures that restrict the free flow of data across geographic boundaries. Countries limit such data flows in multiple ways. In some cases, firms are required to store a local copy of all data, even though this data can be taken and analyzed outside the country. In other cases, countries do not allow any data to be taken outside their territorial jurisdiction.
Over the years, the Indian government has passed sector-specific data localization measures, but now it is contemplating whether to pass a more expansive, economy-wide proposal. The Indian government has released multiple official reports and documents over the last five years that have articulated New Delhi’s objectives for pursuing further data localization. Advocates for localization in India have highlighted the perceived economic benefits of processing Indian consumer data within the country, asserting that greater data localization would enable greater innovation and a larger producer surplus in the Indian economy. They also have noted the difficulties Indian law enforcement faces in accessing Indians’ personal data stored outside the country to prevent crimes and pursue investigations. One especially noteworthy document, the Report of the Committee of Experts under the chairmanship of Justice B. N. Srikrishna, provided detailed reasons for proposing the localization of personal data in India.1
While some claim that keeping local data within a country’s borders may result in economic and security-related benefits, this is not always the case. Multiple considerations determine whether a data localization measure has a net benefit on the localizing country. These considerations include the precise objectives that localization is meant to achieve, the specific localization measures implemented, the underlying economic context, and the country’s national security apparatus.
It makes sense that data localization proposals should be tailored to the country in question and that the costs and benefits should be considered contextually. After all, localization is a significant departure from the existing design principles of the internet, which are premised on the free flow of data across borders. Historically, consumers across the globe have gained access to innovative digital products premised on free flows of data. While some argue that this arrangement increasingly does not benefit local producers, local consumers have benefited from the proliferation of services created outside their countries. Given all these considerations, localization in some cases might be disadvantageous to a host country overall, even if it meets its stated objectives.
In addition to articulating the government’s reasons for considering more robust localization, the Srikrishna Committee also formulated a data localization legislative proposal in the form of the 2018 Draft Personal Data Protection Bill.2 Based on this draft, the Indian government introduced its own 2019 Personal Data Protection Bill in parliament. This bill contains a framework for implementing data localization requirements across the entire Indian economy.3 In parallel, other government departments and bodies have also articulated reasons for data localization.4
The Indian government’s proposal to localize data must be evaluated on how well it would meet the government’s multiple stated objectives. While there may be other reasons for and against data localization, understanding whether such a policy change would meet these democratically articulated objectives is an important starting point.
Crucially, it is important to understand what specific variant of data localization (if any) would best achieve the government’s objectives. Rather than pose the broader question (as some studies have) of whether data localization would have a desirable overall impact on India’s economy, trade, or citizens’ privacy, the aim here is to examine what kind of data localization measure, if any, is best suited to meet the Indian government’s objectives.5 These include the aforementioned concerns over national security, domestic law enforcement, and economic growth.
This paper is structured to contextualize, measure, and analyze the likely effects of various data localization measures the Indian government could pass. First it gives the historical context of data localization globally and in India particularly. Next it provides an overview of the rationale for localization, its prevalence across jurisdictions, and the specific designs adopted in major jurisdictions. Third, the paper explains the multicriteria decisionmaking (MCDM) methodology adopted to understand what localization measures, if any, are best suited for the Indian context. The analysis of these localization alternatives is presented in the paper’s fourth and fifth sections.
The research methods impose certain constraints on the precision of these findings. The research approach combines both qualitative and quantitative information on different aspects of data localization and seeks to create a ranking of localization alternatives. The scores provided in this analysis are therefore indicative of the relative viability of a localization measure, rather than an absolute judgment on its feasibility. In addition, while the suitability of data localization alternatives is analyzed with regard to two objectives (data access for law enforcement and economic growth), the paper does not provide an overall composite score or ranking for the data localization measures it considers.
This analysis contributes to the debate on data localization in multiple ways. First, this study analyzes the viability of localization alternatives from the perspective of the state. The study is based on the rationales the Indian government has given for pursuing such a policy. Second, the paper reaches definitive conclusions on the viability of data localization with regard to certain officially stated objectives. Data localization, for example, would not enable access to data in cases when the data sought by Indian law enforcement is stored in another country and subject to foreign laws. This fact has important implications. The best way to establish jurisdiction over data seems to be to do so directly by establishing legal jurisdiction over firms that conduct business in India or to enter into international arrangements that allow hassle-free access to data. Data localization is unlikely to help India achieve objectives that actually require access to data. Law enforcement and national security objectives may instead be best served by a combination of light-touch localization requirements (such as mirroring requirements that mandate the storage of a local copy in India, while the data can be processed and stored globally) and bilateral and multilateral frameworks that enable India’s access to data stored outside its jurisdiction.
Lastly, this paper finds that local storage requirements could promote India’s stated objectives for spurring economic growth. Such requirements could drive up demand for goods and services in India, while also giving Indian firms a slight competitive advantage over their foreign peers. This calculus depends on various contingent factors, such as whether the resultant demand for data centers would be met by indigenous firms or through imports (which would not add to Indian gross domestic product, GDP), the adaptability of Indian firms to the costs of implementing localization, and the likelihood and severity of retaliatory measures by other countries on Indian service-sector exporters. However, data localization is not required to give Indian producers greater access to personal data for innovation. This is because, as stated before, localization does not by itself advance jurisdictional claims.
HISTORY OF DATA LOCALIZATION
The introduction of the internet transformed the global economy, altered how businesses are organized, and changed how trade is conducted. It also led to a significant increase in technology-driven productivity. From 1992 to 2017, worldwide internet networks went from carrying 100 gigabytes (GB) of traffic per day to over 46.6 terabytes (or 46,600 GB) per second.6 It is estimated that, by 2022, global online traffic will reach 150.7 terabytes per second.7 The growth of cross-border data flows has benefited many small and medium-sized enterprises around the world. There is evidence that businesses that utilize the internet to trade globally have a higher survival rate compared to those that do not.8 Simultaneously, the free flow of data via the internet has also allowed multinational companies to process large volumes of data across national boundaries. Free flows of data have spurred significant innovation and productivity gains globally for both small and large firms.
THE CASE FOR LOCALIZATION OF VARIOUS STRIPES
However, as the digital economy expands, countries have expressed four key concerns over the free flow of data. These are: (1) storage of data on foreign servers, which has impeded data access for domestic national security agencies, (2) the loss of economic benefits due to exploitation of data by foreign firms, (3) concerns about foreign surveillance, and (4) misuse of personal data in violation of privacy rights.
These risks have led many countries to conclude that data flows must be regulated. Data localization is one such measure, and it is not a novel idea. Over the past two decades, many countries have implemented restrictions on the free flow of data. Many countries have adopted localization requirements in selected sectors or industries, such as for critical infrastructure or national defense.
There are different ways data localization can restrict data flows by limiting the physical storage and processing of data within a given jurisdiction’s boundaries.9 These measures can be broadly divided into two categories: hard and soft localization. Hard localization requires local storage and local processing of data. This form of localization does not allow for cross-border data transfers. Soft localization requires some form of local storage but allows data to be transferred and processed outside domestic borders if certain conditions are met. These two broad categories can be subcategorized more granularly. These include sector-specific localization, conditional localization, and general localization measures that can be exempted by bilateral or multilateral arrangements. Table 1 provides definitions of the main variants.
Different countries (and blocs) have employed data localization of varying designs, sometimes across their entire economies and sometimes within specific economic sectors. One of the most significant pieces of legislation on data flows is the European Union’s (EU) 2018 General Data Protection Regulation (GDPR), which imposes conditions on the free flow of data affecting all EU member states.10 China, meanwhile, requires that all “important data” concerning “critical information infrastructure” be localized.11 Similarly, Russia requires all personal data of its citizens to be locally stored.12 Other countries have taken other approaches. The United States requires that all defense-related data be locally stored. For its part, Indonesia requires that all information related to public services be localized.13 Table 2 provides a summary of major countries that have employed localization measures.
Countries have tried to simultaneously regulate the flow of data across territorial boundaries by way of international frameworks. As table 3 shows, many bilateral and multilateral frameworks involve countries promising to adhere to similar standards on personal data protection. In addition, they also agree to facilitate the exchange of information needed for critical national security or law enforcement purposes. Such agreements allow signatory countries to lower data localization requirements for each other.
THE CASE FOR LOCALIZATION IN INDIA
Before considering the recent localization measures the Indian government is contemplating, it is helpful to give a snapshot of what the country has done so far. Like some other countries, India has already adopted a patchwork of data localization measures in some economic sectors. For example, the Reserve Bank of India requires all payment data to be stored in India, though it can be taken out of the country for processing.30 Another sector-specific measure in the telecommunications sector requires local storage and local processing of subscriber information and prohibits the transfer of accounting information related to subscriber or user information.31
Table 4 chronologically lists the major sector-specific localization measures already in effect in India. As the table shows, many sectors of the Indian economy already require data localization, mostly by way of local storage of consumer data.
The analysis in the rest of this paper assesses whether the additional data localization measures in the proposed 2019 bill are suitable for achieving the government’s stated objectives. Now, as the Indian government considers a more robust, all-encompassing bill on data localization, it is important to assess the likely effects of such a policy in terms of the objectives it would be designed to meet. The Justice Srikrishna Committee proposed data localization in the 2018 Draft Personal Data Protection Bill it prepared as a part of its report. It provided a cross-sectoral data localization requirement, with escalating levels of restrictions on the flow of personal data, sensitive personal data, and critical personal data, respectively. The Indian government modified this proposal in the 2019 Personal Data Protection Bill, which was introduced in the Indian parliament. The most significant change was that a larger volume of data would be allowed to flow freely across Indian borders.
The bill proposes that localization measures be adopted across all sectors of the Indian economy for specific categories of personal data. If the bill is passed, all data not deemed “sensitive” personal data or “critical” personal data could be moved out of India freely. Categories of data defined as sensitive personal data could only be taken out of the country if certain conditions were met, and this data would have to be stored in India once processed. Critical personal data could not be taken out of India except under very limited circumstances. The bill does not define critical personal data and leaves it to the central government to define such data categories, but the terms have not been publicly defined yet.
Any analysis of India’s data localization proposals must begin with the government’s four stated objectives: (1) securing speedier and better access to personal data for law enforcement, (2) increasing economic growth, (3) preventing foreign surveillance, and (4) helping enforce data protection laws.
On the national security front, Indian law enforcement agencies face difficulties when the personal data of Indian residents is stored outside the country. The Justice Srikrishna Committee report states that access to personal data for law enforcement agencies is one of the primary considerations for requiring data localization. For example, it explicitly cites the local storage of data as a means of facilitating speedier access to data for law enforcement agencies.48 The document also highlights the importance of speedy access to data for national security purposes.49 This thorny issue came to the forefront when fake news circulating on WhatsApp in various parts of India resulted in mob violence against innocent individuals.50 In many such cases, police needed information on the origin and content of the messages, which were (and continue to be) encrypted. The Indian government wants to exercise greater regulatory supervision over social media services like WhatsApp. Some observers have posited that data localization would enable Indian authorities to do so.
In various government documents, New Delhi also makes it clear that fostering greater economic growth is a key consideration, one that must be evaluated on how well it increases innovation, and also the demand for data-related goods and services within India.51 To this end, it has also been claimed that data localization will enable local IT businesses to innovate and compete with large technology firms. We therefore have to assess whether data localization measures enable innovation and provide competitive advantages to such local businesses.
In addition, the Justice Srikrishna Committee’s report states that data localization is essential to protect Indians from foreign surveillance.52 Similarly, the draft National E-Commerce Policy Report also anticipated the role of localization in preventing foreign surveillance.53 In 2013, Edward Snowden, a former U.S. contractor for the Central Intelligence Agency, disclosed that the United States’ National Security Agency was surveilling the communications of foreign governments and citizens. This revelation highlighted the extent to which digital surveillance could be conducted.54 Simultaneously, the internet has also facilitated a steady increase in the scale and scope of cyber attacks.55 In 2020, a World Economic Forum survey ranked cyber crime as the “second most concerning risk for doing business globally” over a ten-year horizon.56
Finally, the committee’s report argues that data localization would enable India to maximize the economic potential of the vast troves of personal data the country has generated.57
But the report does not provide any evidence for any of these four claims, based on which it advocates data localization. Lastly, the committee’s report also discussed efforts to better enforce data protection laws through the creation of a Data Protection Authority.58
These are the four government-stated objectives that this analysis examined to gauge the potential impact of an economy-wide data localization law in India.
RESEARCH METHODOLOGY FOR EVALUATING DATA LOCALIZATION PROPOSALS IN INDIA
To restate the paper’s central inquiry, the specific design of the localization measure in question must be evaluated based on how well it would help the Indian government meet its stated objectives. To do so, this analysis employs MCDM methods. These methods are useful for considering multiple alternatives by ranking them on selected criteria. This method has been used previously for evaluating policy alternatives in a diverse range of fields such as energy, the environment, sustainability, supply chain management, infrastructure, and others.59
To apply MCDM methods to the case of Indian data localization, it is necessary to follow these seven steps:
clearly define the relevant independent variables (the Indian government’s desired objectives for pursuing data localization)
clearly define the range of relevant dependent variables (in this case, the different data localization measures the Indian government could adopt)
identify clear, objective criteria for evaluating the relative merits of India’s various localization options in terms of the government’s stated objectives
design a suitable scale to measure the goal-based criteria and weight them appropriately so that the effects of the localization policy variants can be compared to a standardized scale
set the baseline scenario based on the current status quo and state the study’s operative hypothesis
assign each policy option’s scores for each defined criterion by assigning a value from the ten-point scale and properly weighting its value
tabulate each policy option’s total score to evaluate its overall merits in a holistic way
The authors hypothesized that each objective articulated for localization would be achieved best depending on what specific kind of localization measure is adopted. They expected that an assessment of each alternative based on specific relevant criterion would help identify which kinds of localization measures, if any, are best suited for achieving India’s objectives. At this stage of the analysis, the running hypothesis was that more free flow of data would be better for economic growth, while law enforcement objectives could be met by complementing the free flow of data with bilateral and/or multilateral agreements that enabled data access for law enforcement agencies.
IDENTIFYING THE INDIAN GOVERNMENT’S STATED GOALS
The first step is to identify and define the Indian government’s objectives precisely. As explained in the previous section, the following four objectives have been articulated in multiple Indian government documents.
securing faster and better access to personal data for law enforcement
spurring increased economic growth and employment
preventing foreign surveillance, and
better enforcing data protection laws
The authors studied each objective to analyze whether any form of data localization would help meet these stated objectives and realized that data localization does not enable governments to achieve the last two objectives.
The prevention of foreign surveillance is a legitimate objective of every sovereign state. Nation-states are interested in preventing the surveillance of certain categories of individuals such as senior government officials, defense personnel, and sensitive scientific research personnel. In India, data localization measures are already in place with respect to government data and communications, so further measures to protect their data would be redundant.60
Therefore, the scope of the debate on the efficacy of data localization for preventing foreign surveillance pertains to personal data that is not already subject to the regulatory requirements and other steps mentioned above. This personal data could possibly include the personal data of government officials generated while they act as private citizens outside the scope of their employment. However, their personal data on social media platforms and other online businesses may be accessed legally by foreign governments (as per the law of the foreign government) if such data is stored within the jurisdiction of foreign governments.
Even if such data is stored locally, and if foreign governments are determined to access such information, they would do so through mechanisms that make data localization redundant as a security measure. Data security and confidentiality are increasingly guaranteed through data security features other than data localization.61
With regard to the enforcement of data protection laws, as Basu et al. note, alternative approaches to localization for improving the enforcement of IT law are already in place in India.62 Enforcement of Indian law is ensured through local incorporation and establishment requirements, rather than requirements that businesses locate physical infrastructure in India. Under the proposed data protection law, the enforcement of data protection laws would be contingent on foreign businesses establishing a business presence in India, not on data localization. The bill requires significant data-related businesses (significant data fiduciaries) to register in India. The enforcement of data protection law against such businesses would therefore be a product of their legal registration in India rather than data localization stipulations.
For these reasons, the paper’s assessment of localization measures was confined to the achievement of the first two objectives: securing faster and better access to personal data for law enforcement and spurring increased economic growth and employment.
Once these objectives were clearly defined, the requirements for achieving them were disaggregated and defined based on secondary research and discussions with stakeholders.63
For law enforcement’s access to data, the following considerations came into play.
Crime prevention:
Accessing encrypted data could help law enforcement prevent crime by monitoring possible offenders
Monitoring of suspects’ social media accounts and financial activities using digital data can help law enforcement track financing for terrorism and prevent attacks
Investigation of crimes:
Gaining faster access to data would help law enforcement investigate crimes more rapidly
Getting access to GPS data could help law enforcement locate suspects and criminal offenders
Securing broader access to financial data could aid law enforcement in money laundering investigations.
Economic objectives
Gaining access to data sets already available with foreign businesses could help give Indian firms a competitive advantage in fields like artificial intelligence (AI)
Making data more available could help lower barriers to scientific innovation for smaller Indian companies
Enacting data localization stipulations would likely produce economic benefits and create jobs for more people in India, including both temporary and permanent positions (in data center management, AI, and other industrial applications, for example)
IDENTIFYING THE DIFFERENT LOCALIZATION VARIANTS
After defining the Indian government’s objectives, the authors identified four different, mutually exclusive variants of data localization measures that the Indian government could adopt (see table 5). These include: (1) some form of conditional localization through data mirroring or local storage requirements for critical personal data, (2) unconditional local storage requirements for all personal data, (3) unconditional mirroring requirements for all personal data, or (4) the unconditional free flow of data accompanied by bilateral and multilateral agreements for managing data access and transfers. These four broad alternatives are categorized by the stringency of local storage requirements and the scope of personal data that must be localized.
The authors then further disaggregated these variants into stylized models of more specific forms their localization measures could take. While many more variations are theoretically possible, the focus here is on the ones deemed most feasible given the current legislation that the Indian parliament is considering.64
IDENTIFYING CRITERIA FOR EVALUATING INDIA’S VARIOUS LOCALIZATION OPTIONS
After laying out these localization alternatives, the authors identified relevant criteria by which to assess their relative efficacy in meeting the government’s objectives. This task was done by using the aforementioned information to disaggregate the government’s two key objectives: improving law enforcement’s data access and spurring economic growth.
Four criteria were identified for assessing which alternative would best meet the objective of improving law enforcement’s access to data: (1) the scope of access, (2) the speed of access, (3) the risk of foreign retaliation against Indian firms abroad, and (4) the risk of data loss due to foreign firms exiting India amid heightened regulations. Similarly, four separate criteria were isolated to gauge how readily each localization variant would spur economic growth: (1) demand for goods and services, (2) competitive advantages for domestic producers and competitors, (3) the risk of data loss due to foreign firms exiting India amid heightened regulations, and (4) the risk of foreign retaliation against Indian firms abroad.
A more detailed explanation for each of these criteria is given in the next section of the paper.
MEASURING AND WEIGHTING THE GOAL-BASED CRITERIA
To score each of the potential localization alternatives, the authors first used a ten-point scale for each of the evaluating criteria (with ten being the scale’s highest value and one being the lowest). When gauging both law enforcement access and economic growth, two of the criteria—the risks of retaliatory action and of data loss—the scoring method was flipped. More specifically, the localization alternative posing the highest risk received the lowest score (zero) and the alternative posing the lowest risk scored the highest (ten). The authors chose to make this adjustment so that the scoring of the positive variables (data access and heightened demand, for instance) moved in the same direction as the negative variables (retaliatory and data loss risks) rather than at cross purposes.
Once the scales were established, the next step was to assign a weight to each criterion. To do so, the authors created a pair-wise comparison matrix where each criterion was measured against itself and the other criteria. This exercise was conducted for all the criteria separately for assessing law enforcement’s data access and the prospects for spurring economic growth. A detailed rationale for the usage of this scale is given in appendix 1 based on the methodology developed by Saaty.67
In the end, the various criteria were weighted in the following way. When gauging law enforcement’s access to data, the scope of data access was weighted at 31 percent, the speed of access was assigned 50 percent, the risk of retaliatory action against Indian firms abroad was listed at 5 percent, and the risk of data loss came in at 14 percent. On the question of spurring economic growth, heightened demand for goods and services was weighted at 36 percent, any competitive advantage accrued by domestic producers or competitors came in at 38 percent, the risk of data loss stood at 18 percent, and the risk of retaliatory action amounted to 18 percent. See appendix 1 for a full explanation of how these weightings were determined. Based on this ten-point scale and weighting system, the authors ranked all the criteria relative to each other and weighted them accordingly.
SETTING THE BASELINE AND STATING THE OPERATIVE HYPOTHESIS
Before the various policy options were scored, we decided to use India’s existing legal framework on localization as a baseline case to measure the potential alternatives against. To review briefly, the legal status quo does not impose localization requirements (except in specific, limited sectors) on the flow of data, and law enforcement agencies use MLATs to request access to data stored abroad. Other alternatives were scored in relation to this baseline scenario. The baseline scenario is therefore listed as alternative A1 in both appendixes 2 and 3 (which provide the detailed scoring and rationale for each policy alternative’s scoring).
SCORING EACH LOCALIZATION POLICY OPTION
Scores for each localization alternative were assigned according to a simple calculation: each policy’s assigned value on the ten-point scale times the assigned weight of that criteria weight. We provide an example of this below using the alternative “No localization: global data storage and processing with MLATs” from the law-enforcement objective. In each box below, the first number multiplied is the policy option’s score on the ten-point scale multiplied by a second number representing its assigned weight. On scope of data access, for instance, the policy option of no additional localization was assigned a score of six out of ten, and this value was multiplied by 31 (for its 31 percent weighting) to get a total of 186. The aggregate score in the final column on the right represents the tabulated total of these four composite scores. (See appendix 2 for a detailed rundown of how each of the policy options were scored in terms of granting law enforcement data access and appendix 3 for a rundown of how they were scored in terms of spurring economic growth in India.) Other alternative localization measures would score higher or lower than six out of ten on scope of data access, three out of ten on speed of data access, nine out of ten on the risk of retaliatory action, and nine out of ten on the risk of data loss, all numbers that affect the aggregate weighted scores.
TABULATING EACH LOCALIZATION POLICY’S TOTAL SCORE
For each policy alternative, the authors added up the weighted scores in each criteria column to assign an aggregate score (the farthest-right column in the example above). This process was done twice for each policy alternative, once to measure its efficacy in terms of giving law enforcement data access and again to measure its effectiveness in spurring economic growth. The alternative with the highest aggregate score compares most favorably to other localization alternatives and is the most likely to achieve the stated policy objectives: improving law enforcement access to data or fostering economic growth and innovation, while balancing countervailing risks.
WHICH LOCALIZATION MEASURE WOULD BEST SECURE DATA ACCESS FOR LAW ENFORCEMENT?
Lacking timely access to data held overseas can be a major constraint on Indian law enforcement personnel when conducting investigations. This is usually because the data law enforcement is seeking was collected in India and stored in another jurisdiction, leading to a conflict of legal systems. While MLATs can offer access to personal data in such situations, this cumbersome process often takes around ten months on average in all cases.68 Mandating local data storage and processing through localization would not necessarily solve this problem since businesses can still be bound by the laws of their home country even if they store their data in India.69
Indian law enforcement also faces issues with getting access to different kinds of data. Laws in other countries prohibit their businesses from sharing certain kinds of data with investigative authorities in other jurisdictions.70 One argument for data localization is that it will resolve both issues for law enforcement agencies by enabling faster access to different kinds of data.
To measure which localization alternative would best help law enforcement get improved access to personal data, the authors assessed the scope and speed of data access and the risks of foreign retaliation against Indian firms abroad and of data loss stemming from foreign firms ceasing operations in India to escape heightened regulations.
Two criteria (scope of data access and speed of data access) were clearly more important for the fulfillment of this objective and therefore received a significantly higher weighting (31 and 50 percent, respectively) relative to the others. Of these, the speed of data access has been clearly highlighted to be more important in multiple documents.71
The countervailing risk of data loss due to foreign companies leaving India receives a much lower weighting (5 percent). While India does face a risk of losing foreign businesses, India is also a large market, and foreign investments into technology have continued to flow into the country despite localization already mandated in some sectors, and the impending prospect of localization under the Personal Data Protection Bill. While it is logical to assume that localization would increase compliance costs and lead to firms leaving India, this has not been borne out by evidence. Technology FDI has continued to pour into India despite the increased threat of localization. In 2020, India received its highest ever FDI inflows, mostly into the technology sector.72 Therefore, while this analysis considers the possibility of losing business due to localization, recent evidence shows that the likelihood of this occurring is extremely low. This criterion is therefore accorded a very low weight compared to the two previous criteria.
Similarly, while there is a real threat of retaliatory action by foreign governments, this would be a minor consideration in terms of Indian law enforcement’s overall stated objective of receiving greater data access, so that criterion is weighted accordingly (14 percent). While there is a real and possible threat of retaliatory action, the only recent such measure has been the measures the United States adopted in response to China’s new national security law. There have been no other occurrences of retaliatory action in response to a localization measure. Therefore, this criterion is assigned a low weight based on the low probability of such actions materializing.
SCOPE OF DATA ACCESS
Since Indian law enforcement’s limited access to certain kinds of data (specifically content data) held overseas is a key constraint on law enforcement personnel’s ability to perform their duties, the authors measured which localization variant would grant them access to the largest variety of personal data. At present, the status quo (see alternative A1 in appendix 2) means that foreign firms are not allowed to share content data, which is data that is identifiable to a specific user.73 U.S. companies currently hold most of the personal data in the world, and therefore the main counterparts on questions of access to such data are U.S. businesses.74 Because of an exception under the U.S. Electronics Communications Privacy Act, U.S. businesses voluntarily provide noncontent data to Indian law enforcement.75 This practice would continue regardless of which option India chooses.
Consequently, Indian law enforcement also has been increasingly requesting access to encrypted data.76 Yet mere localization is not likely to significantly increase the scope of data available to Indian law enforcement mainly because of a conflict between Indian and U.S. law.77 A U.S. law known as the Stored Communications Act (SCA) prohibits businesses from sharing personal data with any foreign government unless certain legally mandated procedures are followed. Only certain kinds of data (noncontent data or subscriber information) can be shared with foreign law enforcement without undergoing this process. That means that, with respect to Indian data captured by U.S. businesses, the scope of accessible data for Indian law enforcement does not and will not vary significantly based on which localization measures are enacted. The scope of India’s data access may, however, vary in relation to other countries (apart from the United States) that do not have similar blocking statutes.
Because of this, Indian law enforcement’s access to personal content data would increase the most if India enters into a bilateral or multilateral framework because these frameworks would specifically enable law enforcement’s access to an increased scope of data (see alternatives A2 and A4 in appendix 2). The status quo baseline scenario is clearly suboptimal because of the inefficiency of MLATs (see alternative A1 in appendix 2). India could also make modest gains by imposing restrictions on the flow of sensitive or critical personal data (see alternatives A7, A8, and A9 in appendix 2). This is because, while data currently located in non-U.S. jurisdictions is theoretically easier to access if such restrictions are imposed, data stored in the United States would continue to be as inaccessible as it is right now. These limited gains could also be significantly impaired if non-U.S. jurisdictions implement laws similar to prevailing U.S. law.
In short, Indian law enforcement will not be able to access foreign-controlled encrypted data by requiring localization. Doing so would require a legal mandate for decrypting data. Foreign businesses who wish to continue to provide services to consumers in India would comply with this legal mandate irrespective of whether they are asked to localize data.
SPEED OF DATA ACCESS
Since speedy access to personal data is such a key requirement for law enforcement investigations, speed of access outweighs other considerations by a significant margin in this analysis.78 Both preventing crime and investigating past crimes require quick access to relevant information, and delays can drastically reduce the likelihood of success.
Under India’s legal status quo (see alternative A1 in appendix 2), Indian law enforcement must first ask a business for information about an individual. A business that is subject to U.S. law must then consider whether it is has the information sought and crucially whether it is permitted to share such information with Indian law enforcement. If the information is content data and therefore subject to U.S. law, Indian law enforcement is asked to file an MLAT request.79 If the MLAT request is directed to the U.S. government, the U.S. Department of Justice reviews the request and then a prosecuting attorney presents the request before a U.S. federal judge. If the judge agrees that the request meets U.S. judicial standards for sharing that information, the court issues an order for the business to share such information. As previously noted, on average this process takes a minimum of ten months.
While some form of local storage mandate might increase the speed with which Indian law enforcement gains access to data collected by non-U.S. businesses, it would not affect the aforementioned legal process for accessing content data from the United States (see alternatives A5–A9 in appendix 2). When it comes to data controlled by non-U.S. businesses, the speed of access for Indian law enforcement under all potential local storage alternatives is essentially the same, but that does not hold true for data controlled by U.S. firms. EU law, for example, does not contain provisions like those in the U.S. SCA.80 Therefore, data collected by a business that is subject to the EU’s GDPR could be accessed faster if the Indian government were to impose a localization mandate. This is subject to the EU’s existing legal framework remaining static. However, the best alternative available for securing Indian law enforcement speedy access to content data is to enter into multilateral/bilateral agreements that overcome jurisdictional conflicts over data (alternatives A2 and A4 in appendix 2).
Data localization is likely to be insufficient for speeding up Indian law enforcement’s access to personal content data. Other than entering into bilateral or multilateral frameworks, Indian policymakers can consider establishing local licensing requirements that would ensure that foreign businesses incorporate in India as subsidiaries or foreign branches and that the legal liability for collecting and storing Indian data rests with such Indian entities. For example, the Reserve Bank of India’s localization requirement on payment-processing businesses and other financial firms is enforceable primarily because these entities cannot provide services in India without the Reserve Bank’s prior authorization.81 Securing such authorization in turn requires them to register as Indian entities.82
This is also how the United States exercises jurisdiction over data stored by its companies in other countries. The SCA enables the U.S. government to access data stored outside the United States by U.S. companies.83 By making a claim of legal jurisdiction and regulating U.S.-based businesses, the U.S. government can require access to data if such data is in the company’s “possession, custody or control, regardless of whether such . . . information is located within or outside of the United States.”84
The Indian government may also therefore consider establishing legal jurisdiction to access data from businesses that provide services in India. The 2019 Personal Data Protection Bill provides for the registration of “significant data fiduciaries” in India.85 However, it is unclear whether this registration will be a form of licensure and authorization to conduct business in India, or a registration of the service. The former would enable the Indian government to exercise greater jurisdictional control over the businesses that collect and store data on Indian citizens, in a manner similar to U.S. law.
RISK OF DATA-HOLDING FOREIGN BUSINESSES EXITING INDIA
Foreign businesses could react negatively to any further localization requirements India may institute for two reasons: escalating compliance costs and perceived privacy concerns. Any decision such firms may make to stop providing services in India due to compliance costs would involve a careful consideration of the benefits businesses would lose as well. India’s digital economy is growing rapidly. Moreover, India is the world’s largest accessible data market, and an extremely competitive one at that. These benefits would have to be weighed against any compliance costs to be incurred by businesses due to localization requirements.
Privacy concerns could also pose another serious reason for foreign businesses. U.S. and EU businesses are affected not just by shareholder concerns over consumer privacy but also by laws in these jurisdictions that place a great deal of weight on consumer privacy even in foreign jurisdictions.86 For example, the United States has been ramping up pressure on its businesses providing services in China after a new Chinese national security law was passed in 2017.87 These contextual and political considerations will also depend on the quality of bilateral relationships between India and countries like the United States.
If foreign businesses stop providing services in India for either of these reasons, the Indian economy would lose the benefits of these services, especially platform services that enable buyers and sellers to conduct e-commerce. More significantly from the perspective of meeting this objective, Indian law enforcement would lose access to the personal data of consumers already collected by such businesses since they would no longer be operating in Indian territory.
It is likely that, in some cases, an Indian business may crop up to provide similar services to Indian consumers, and the loss of the data and services could thus be mitigated. However, this process would take time and may not occur in all circumstances. For example, when the Indian government banned the Chinese social media app TikTok, user engagement on alternative apps remained much lower: As one Indian observer put it, “although TikTok users shifted to other alternatives, user engagement indicators such as app open rates and average session times are yet to catch up with TikTok’s engagement levels.”88 This could be either because Indian substitutes do not exist, or because Indian competitors may be unable to replicate the same quality of service.89
This analysis therefore assigns the highest scores to localization alternatives that pose the least risk of driving away foreign businesses. Alternatives that do not impose a local storage restriction would pose fewer risks (alternatives A1–A4 in appendix 2) than ones that do. Therefore, even alternatives that impose mirroring requirements but do not otherwise restrict the flow of data score high on this criterion. Conditional localization requirements (alternatives A7–A9 in appendix 2) score the lowest since, by definition, they restrict the flow of data that is considered more sensitive. In addition, the compliance requirements for segregating classes of data depending on their sensitivity is also higher than those of unconditional localization.
RISK OF FOREIGN RETALIATION AGAINST INDIAN FIRMS ABROAD
The analysis also measures the risk of other countries retaliating against localization measures taken by India. Such actions might lead other countries to impose localization restrictions of their own on Indian firms that export services and could negatively affect Indian data-related businesses that serve consumers in these jurisdictions. For example, the U.S. government issued an executive order banning TikTok on the grounds that “TikTok automatically captures vast swaths of information from its users, including Internet and other network activity information such as location data and browsing and search histories. This data collection threatens to allow the Chinese Communist Party access to Americans’ personal and proprietary information.”90
Retaliatory measures could also involve measures that are not directly related to data. Countries that may adopt such measures are those with significant exporters of data-related services to India, including the United States.91 Respondents in a November 2019 survey of Indian businesses by the Indian Council for Research on International Economic Relations said that there is a fear of retaliatory action against Indian IT companies that use “off-shore models using citizen data of other countries.”92 Such retaliatory measures would likely be stronger if India adopts stringent data localization measures.
More stringent and comprehensive localization measures therefore carry higher risks of retaliatory action. Unsurprisingly, alternatives that allow for the free flow of data score the highest. These options pose the lowest risks of retaliatory action (alternatives A1 and A2 in appendix 2). Unconditional hard localization poses the highest risk of retaliatory action (alternative A6 in appendix 2).
OVERALL ASSESSMENT
The aggregated scores on these individual criteria provide a final ranking of the data localization alternatives India is considering.93 The current status quo whereby Indian law enforcement seeks and receives access to data through MLATs is clearly suboptimal (alternative A1). Localization measures that allow the highest degree of free-flowing data through a bilateral/multilateral framework for data access receive the highest aggregate scores (see alternatives A2 and A4 in appendix 2). A regime for free-flowing data supported by bilateral/multilateral agreements would enable Indian law enforcement to resolve the issue of conflicting legal regimes and expedite access to data. In addition, it would also mitigate Indian law enforcement’s current inability to access different kinds of data due to legal restrictions in foreign jurisdictions.94 The model of multilateral/bilateral arrangements considered here does not reflect actual international agreements, which are either limited in the scope of data covered or allow free-flowing data for specific purposes.95 This analysis provides a framework for thinking about the tradeoffs between free-flowing data versus localization restrictions.
As the scores highlight, restrictive localization options do not actually increase the speed or scope of data access that Indian law enforcement enjoys. This is because foreign businesses are still required to comply with the laws of their home country while serving consumers in India. This point is especially relevant given the existing market conditions whereby U.S. companies hold most of the world’s data and U.S. privacy laws create a significant conflict with Indian laws.
The perceived benefits of data localization for supervisory purposes are therefore a function of local incorporation and authorization requirements. While the Reserve Bank of India has implemented a localization requirement, it is important to note that its localization requirement for payment service providers has been coupled with the fact that the Reserve Bank specifically authorizes these businesses to provide services, a step that requires them to register as Indian businesses.96 As Indian service providers, such businesses are then regulated primarily by Indian law and the Reserve Bank of India.
WHICH LOCALIZATION MEASURE WOULD BEST SPUR ECONOMIC GROWTH?
Many have asserted that localizing the data of Indian consumers within India would likely boost India’s economic growth and support innovation more than the free flow of data.97 While there have been some studies on the impact of localization requirements in India,98 no study has analyzed costs and benefits of data localization across the economy comprehensively.
This analysis highlights the relative efficacy of different localization alternatives instead of providing an absolute quantification of costs and benefits. It analyzes which localization alternative (if any) would benefit Indian producers and the larger economy more than the status quo framework of free-flowing data. This analysis looks at how different data localization alternatives fare on four criteria: increasing demand for local goods and services, providing Indian firms a competitive advantage, the countervailing risk of India losing access to data held by foreign companies that opt to pull out of India rather than face greater regulations, and the countervailing risk of foreign governments retaliating against Indian firms abroad.
Of these, the most weight is given to the first two criteria due to their direct relevance to economic growth. Both these criteria receive approximately equal importance (36 and 38 percent, respectively). Localization has the potential to increase local demand for data storage services, and this is an important factor that the impact of localization measures should be measured against. Similarly, many have argued that localization will be beneficial for India because it would lead to competitive advantages for local industry, and this has become another important reason articulated for requiring localization in India. The countervailing risks of lost businesses (18 percent) and retaliatory action (8 percent) are given slightly higher weights here than they were in the law enforcement case. This is because, by comparison, lost businesses and retaliatory actions by other governments against Indian firms are more likely to have a direct impact on economic performance than on the data-access capabilities of law enforcement. Specific reasons for the relatively low weightage for the countervailing risks have been provided in the previous section. Those reasons remain the same for the purposes of weighting criteria for this objective.
A detailed explanation of these considerations and the findings is given below.
INCREASED DEMAND FOR GOODS AND SERVICES
This analysis considers whether data localization would increase the demand for data storage services and data storage infrastructure in India. Data storage services in the country are already growing, but it is hard to determine how much of this growth is driven by the anticipation of data localization requirements being imposed.99 The relative increase in demand for such goods and services is therefore assessed depending on the localization measure imposed (see column C1 in appendix 3).
Data localization is likely to spur demand for the establishment of data centers in India. India is one of the largest consumers of data in the world, and the size of its market is likely to double in the next five years.100 At least some of the demand for storing this data is likely to be met locally, and localization measures can give a fillip to such demand.
Data centers can have significant positive benefits for local economies. A 2014 impact assessment of Facebook’s data center in Sweden done by the Boston Consulting Group found significant direct spending in the local area resulting from the establishment of the data center, and the study also estimated that a total of 4,500 jobs would be directly and indirectly created during the life cycle of the center. It also found that the presence of the company and its data center contributed to “the emergence of a new ecosystem of information and communication technology companies . . . public and private investments in infrastructure and utilities.”101
A report by the U.S. Chamber of Commerce on the benefits of data centers in the United States stated that data centers create close to an average of 1,700 direct and indirect jobs when they are being built and produce $243.5 million in output within a state.102 Other reports on the impacts of data centers have noted similar direct and indirect benefits.103 One report states that data centers contributed to the creation of more than 45,000 jobs in the U.S. state of Virginia in 2018.104 Another report studying the impact of data centers in the state of Washington estimates that the job multiplier effect of data centers ranges from 2 to 3.54.105 In other words, for every job created within a data center, there are 2 to 3.54 jobs created in the local economy.
It is likely that the establishment of data centers could have similar potential benefits for the Indian economy. This possibility merits giving a significant weight to the effects that localization requirements may have on the demand for such goods and services in India. In addition, the discussion above highlights that stricter data localization requirements would presumably lead to higher demand for data centers in India.
However, it must be noted that domestic demand could be affected based on the value of imported equipment required for data centers. A recent study by an Indian institution states that, while India has comparative advantages in the production of certain items required for building data centers, in the past few years, imports of equipment for data centers have grown at a much faster rate than exports (with compound annual growth rates of 13.8 percent and 7.4 percent, respectively).106 In addition, India’s overall trade balance has been worsening over the years, and there is very little added domestic value for India on such imported products. So, even though demand for data-related infrastructure can lead to GDP growth, the overall impact on GDP growth would also depend on whether this demand is met with domestic production or imports.
On the whole, while localization measures would likely lead to an increase in demand for data centers and indirect benefits to the economy, the overall effect of such demand on India’s GDP could be negated to some extent by the trade imbalance alluded to above. This would be especially true in the short term when the existing trend of increasing trade imbalances would be hard to reverse.
This analysis gives a modest increase in scoring to localization measures that promote the increased usage of local data centers. A stricter localization mandate would result in the highest increase in demand for data storage infrastructure and services because it would provide the highest incentives for increasing local creation and production of data storage infrastructure within India. That likely means hard, unconditional localization would lead to the highest increase in demand (alternative A6 in appendix 3), followed closely by requiring local data storage with global processing (alternative A5 in appendix 3).
The current baseline scenario of mostly unrestricted data flows receives the lowest score (alternative A1), and data mirroring requirements also would likely create very limited additional demand (alternatives A2 and A3) because data would continue to be stored in the most cost-effective locations for firms and would therefore not necessarily incentivize further creation of data storage infrastructure in India.
COMPETITIVE ADVANTAGE FOR INDIAN DOMESTIC FIRMS
Mandating data localization in India might put foreign firms at a disadvantage for two reasons. First, these firms would have to incur the costs of data storage and processing capabilities in India as a capital investment. Second, the recurring costs of renting or operating data-related infrastructure in India would be higher than foreign firms’ existing costs. Conversely, restricting access to data storage facilities in India may lead to higher prices for such services than at present.107
But these consequences would also affect Indian firms that currently store their consumer data outside India.108 In addition, while some have argued that the storage of Indian data within India would give a boost to domestic innovation, it is unclear how this would work in practice.109 While the increased availability of data may have some economic benefits,110 the sharing of proprietary personal data would require additional policy measures, and even if such measures are contemplated, it is unclear if data localization is necessary to implement them. In addition, as mentioned before in the law enforcement section above, data localization does not advance jurisdictional claims for data access. Jurisdictional claims for data access must be advanced by creating licensure or authorization requirements for the operation of the relevant foreign businesses in India, while making data sharing a condition of such authorization.
The earlier discussion on the risk of loss of data in the law-enforcement objective details the issues with automatically assuming that Indian businesses would step in to replace any foreign service providers that may decide not to provide services in India. As previously discussed, in the wake of India’s ban on Chinese apps in India in 2020, Indian businesses found it hard to replace successful Chinese apps. While this assumption of substitution by Indian businesses may hold true in the medium to long term, it would be hard to attribute any such substitution solely to data localization, especially over a long time period.
Additionally, barriers to the free flow of data could also hurt Indian businesses by increasing delays and costs in terms of innovation if such businesses have collaborative ties with research or business partners outside India.111 Indian businesses would also need to use multiple data storage facilities if they serve consumers outside India. Another study finds a positive correlation between innovation and exports of digital services.112 Lastly, some research implies that additional costs in technologically dynamic sectors seem to lead to greater concentration and further consolidates the market dominance of “superstar firms”:
“Growth of concentration is disproportionately apparent in industries experiencing faster technical change as measured by the growth of patent intensity or total factor productivity, suggesting that technological dynamism, rather than simply anticompetitive forces, is an important driver—though likely not the only one—of this trend.”113
The costs of localization may therefore be internalized better by large multinational technology firms than Indian ones, negating significant competitive advantages.
Consequently, while local storage requirements, including unconditional hard localization (alternatives A5 and A6 in appendix 3), provide a slight competitive advantage to local firms by forcing foreign firms to invest in data storage services in India, this benefit is not significantly higher than the baseline scenario (alternative A1) because a significant proportion of data center equipment is imported and does not add to Indian GDP. This could, however, change if the growing data center business in India were able to use Indian-manufactured equipment, or if other benefits such as job creation significantly outweighed any loss of benefits. Moreover, these benefits have to be weighed against the other disadvantages discussed above. Mirroring and conditional local storage requirements (alternatives A3, A7, A8, and A9 in appendix 3) also provide some advantages compared to the baseline scenario, but these benefits are lower than local storage alternatives since the costs of these requirements on foreign firms would be lower.
RISK OF DATA-HOLDING FOREIGN BUSINESSES EXITING INDIA
The risk of lost business related to foreign-controlled Indian data could be consequential for India’s economy if foreign service providers opt to stop providing existing services or stop innovating for the Indian market in the face of heightened regulations.114 Businesses may be inhibited from providing data-related services in India due to localization requirements that increase compliance costs, privacy concerns, or both.
The greatest risk arises from conditional hard localization measures (alternative A6 in appendix 3). The risks in that scenario may likely be higher than the risk from an unconditional hard localization measure depending on how critical personal data is defined (alternatives A7, A8, and A9 in appendix 3). If the way the Indian government defines critical personal data raises privacy concerns in jurisdictions like the United States, this could lead to a higher risk of lost business than even an unconditional hard localization measure. Predictably, the lowest risk of lost business related to Indian data emanates from alternatives that impose few or no restrictions on free data flows (alternatives A1–A5 in appendix 3).
RISK OF FOREIGN RETALIATION AGAINST INDIAN FIRMS ABROAD
The countervailing risk of retaliation by foreign governments against Indian firms abroad remains the same under this objective as it was when assessing law enforcement’s data access. Stringent localization alternatives carry a higher risk of retaliatory action than others (alternatives A6 and A7 in appendix 3).
The scores were aggregated to understand which alternatives would best meet the stated objective.115 A regime that requires local data storage but permits global processing best meets the objectives of promoting economic growth (alternative A5 in appendix 3). Following very closely in second are regimes that require mirroring of Indian data within India (alternatives A3 and A4). The hard localization requirement alternative is also a close second-best alternative (alternative A6). If hard localization mandates are adopted, an unconditional localization measure scores higher than conditional localization measures (alternatives A7, A8, and A9). This is because unconditional localization scores higher on the effect it has on stimulating demand for domestic goods and services compared to conditional localization. This means that the localization measure proposed in India’s data protection bill is not likely to be the design that can best achieve the government’s stated objectives for economic growth and innovation.
CONCLUSION
Some key findings from the previous two sections bear repetition. Most significantly, access to data stored outside India will require either or both of the following: a regulatory mechanism that would allow India to advance a jurisdictional claim on the entities that store foreign-controlled content data of Indian citizens and bilateral/multilateral agreements that reduce or remove conflicts between Indian law and foreign (especially U.S.) law. Localization does not advance jurisdictional claims or reduce conflicts in jurisdiction.
Ultimately, the best localization alternative for Indian law enforcement purposes is to enter into bilateral agreements with countries that restrict access to such data. Localization measures do not improve access to data significantly with respect to the United States because of domestic U.S. laws, though they may help improve data access with respect to Indian service providers incorporated in other countries. This is because, unlike the United States, many other jurisdictions do not prevent access to personal data of nonresidents stored in their country. Even so, since a large proportion of Indian consumer data is collected and stored by U.S. entities, localization is unlikely to be the best overall strategy for improving law enforcement’s data access.
Meanwhile, a localization framework involving local data storage and global processing appears to be the best alternative for enabling higher economic growth in India. This option is closely followed by measures requiring mirroring of data, and alternatively, an unconditional, hard localization requirement. The key driver of these alternatives is the anticipated increase in the demand for goods and services pursuant to forced localization. That fact notwithstanding, India has an increasingly negative trade balance in the trade of equipment for data centers. The degree of impact this would have on India’s net GDP growth depends on the continuation of this pattern of trade imbalance.
Notably, localization itself does not ensure data access that would drive innovation in India. Such access is contingent on India exercising legal jurisdiction over entities that collect and control this data, not on the data itself.
These analytical findings are contingent on the stylized design of the localization alternatives and an assessment of existing facts and risks. It is likely that domestic policy measures or geopolitical changes may affect the validity of this analysis. Still, these findings highlight the relative benefits of different localization measures in the specific context of a developing country like India. India’s 2019 Personal Data Protection Bill imposes a conditional, local storage requirement on personal data. This alternative is not the most beneficial alternative in terms of meeting law enforcement’s needs or spurring India’s economic growth. The Indian government and other governments around the world can use this framework to balance different considerations while deciding on the feasibility of data localization requirements and other related policies.
APPENDIX 1: THE THEORETICAL UNDERPINNINGS OF THE CRITERIA WEIGHTING
Saaty’s scale allowed the authors to rank the criteria introduced in the methodology section of the paper in relation to each other based on relative importance. Tables 6 and 7 provide the relative comparison matrices for both the objectives: data access for law enforcement and improving economic growth. The example below in table 6 provides a breakdown of how the criterion scope of access was weighted by relative importance compared to the other criteria and itself.
By way of example, to fill up the first row of numbers in table 6, the authors answered four questions.
How important is scope of access in relation to itself? Here, the criterion was simply measured against itself, giving a value of 1, which signifies equal importance.
How important is scope of access to speed of access? Getting access to all types of data is a very important requirement for law enforcement. Yet while law enforcement agencies can usually access metadata and subscriber data easily, they cannot access content data easily. In addition, getting speedy access to data allows law enforcement to determine what other kinds of data are required and seek access to this data too. Therefore, the authors concluded that localization is less of an imperative for scope of access than it is for speed of access. This is why a value of 1/2 was assigned. This value is employed for inverse comparison and shows that scope of access is less important than speed of access.
How important is scope of access to the risk of retaliatory action against Indian firms abroad? Scope of data access is significantly more important than the risk of retaliation against Indian firms abroad. This is because the probability of such retaliatory action is low. In addition, these issues were examined from a law enforcement perspective, the gains from access significantly outweigh the risks of retaliation. This is why a value of 5 was assigned, signaling strong relative importance.
How important is scope of access in terms of the risk of lost data stemming from foreign firms leaving India? Scope of access is more important to Indian law enforcement than the potential loss of data. While Indian law enforcement will lose access to a foreign firm’s data altogether if the firm chooses to leave India, the scope of access is still significantly more important to Indian law enforcement. This is why an intermediate value of 4 was assigned, signaling a strong relative importance for the criterion of scope of data compared to loss of data.
A similar exercise was conducted for all the other criteria in this matrix. But this was just the initial step in the process as the final weighting computation required more steps.
Similarly, below in table 7 is a breakdown of how the criterion demand for goods and services was weighted by relative importance compared to the other criteria and itself.
Weighing the criteria for the second objective (spurring economic growth) entailed conducting the same exercise again (see table 7) by answering the following four questions.
How important is demand for goods and services in relation to itself? Here, the criterion is simply measured against itself, giving a value of 1, which signifies equal importance.
How important is demand for goods and services in relation to a boost to Indian firms’ competitive advantage? Competitive advantage is equally important to demand for goods and services because, while competitive advantage for domestic producers is important from the government’s perspective, the demand created is equally important. This is why it is assigned a value of 1, signifying equal importance.
How important is demand for goods and services in terms of the risk of lost data stemming from foreign firms leaving India? Demand for goods and services is slightly more important than the risk of data loss. The risk of loss of data has low impact on economic growth and innovation because data is nonrival and can be recollected/generated. In addition, the actual probability of the risk materializing is low based on existing evidence of the behavior of foreign firms and investments coming into India. Compared to this, the benefits from heightened demand for goods and services is of larger economic importance to India’s domestic economy. This is why it was assigned a value of 3, signaling moderate importance.
How important is demand for goods and services in relation to the risk of foreign retaliation against Indian firms abroad? Admittedly, the risk of such retaliation would also affect Indian firms operating abroad and thus directly impact India’s economic growth. But heightened demand for goods and services is still more important for India’s economy than the risk of foreign retaliation. This is why it has been assigned a value of 3 signaling moderate importance.
Again, a similar exercise was conducted for all the other criteria in this matrix. But this was just the initial step in the process as the final weighting computation required more steps.
No comments:
Post a Comment