## Tuesday, August 27, 2019

### Policy uncertainty in Indian e-commerce

by Megha Patnaik.

Reduced investment in India by private persons is a key part of the present growth challenge. Investment is shaped by macroeconomic uncertainty, sectoral uncertainty, and regulatory risks that firms face. In this article, we think about the risks that an E-commerce firm such as Amazon perceives in India. These include the changing FDI rules, unresolved issues of data localisation and code disclosure, the multiple reports on technology-related activities that various government agencies are releasing, and the problems of rule of law in licensing and investigation.

#### Economic policy uncertainty

Shrinking investment in the Indian economy is a concern. Investments in new projects fell to a 15 year low in the last quarter according to the Centre for Monitoring the Indian Economy (CMIE) Capex data which tracks large investment projects.

Firms are deterred from investing by policy uncertainty (Bloom, 2009). When firms are unclear about the future economic environment, they hold back on investing till uncertainty declines. This delays the pickup of the investment cycle, where firms generates jobs and business for linked firms, fueling aggregate economic activity. Uncertainty particularly affects long-term investments that are irreversible in nature, and for which horizons for cost recovery run into years. These can be investments in new technologies or market segments, or investments in infrastructure. Such investments are particularly important, as they can benefit other firms in the economy, fueling productivity and long-term growth in addition to their business cycle effects.

Private sector investment is adversely affected by three kinds of policy uncertainty - macroeconomic uncertainty, sectoral policy uncertainty, and regulatory risk. The role of Economic Policy Uncertainty at the macroeconomic level has been measured globally (Baker et al, 2016). In the original measure, an index is created by quantifying newspaper coverage of policy-related economic uncertainty mentions in the national newspapers, through combinations of keywords related to policy and uncertainty. Macroeconomic policy uncertainty has been applied to understand global events. Brexit-driven policy uncertainty in the UK moved closely with the GBP Real Exchange Rate in recent times, and the uncertainty surrounding US trade policy affected importing firms. This measure of macroeconomic policy uncertainty correlates strongly with stock market volatility.

Firms face much more than macroeconomic uncertainty. They also face uncertainty at sectoral, geographical and individual levels. Sectoral-level policy uncertainty can be measured through surveying firms sampled across sectors, asking them about expectations about future growth and costs at various horizons (Altig et al, 2019). For example, firms can report not just their expectations about future profits, but the distribution across the possible profit outcomes that they can expect.

An additional source of uncertainty that firms operating in India face is regulatory risk. Even when regulations are formulated, there is a lack of predictability, and excessive executive discretion, in how a stated regulation will be enforced. For example, the licensing by the RBI of 11 payment banks from 41 applicants who wanted to start payment systems was a non-transparent process inconsistent with the rule of law (Roy and Shah, 2015). Another example is the Copyright Board order of 2010 on statutory licensing fees paid by Radio stations. This order arose out of nine one-on-one disputes between radio stations and music producers, but was applied as an in rem order rather than an in personem order. Thus, music producers who weren't part of the original disputes also became governed by the order, despite the appeals by T-Series and SIMCA against the Copyright Board order applying to them. Aggarwal and Zaveri (2019) show the uncertainty induced for private persons through executive discretion in enforcement at SEBI.

#### Drivers of uncertainty in the E-commerce sector

In the recent Q2 earnings announcement, the Amazon CFO Brian Olsavsky mentioned uncertainty in India's e-commerce policy. He expressed hope for stable' and predictable' policy, for the company to continue with its investments in technology and infrastructure in India. This explicit mention about policy uncertainty in India is a unusual moment, and requires attention by policy thinkers. What is the uncertainty associated with investing in India, as seen by Amazon?

1. India's draft e-commerce policy rules earlier this year preventing firms from influencing prices or selling products in which they hold stakes disrupted business plans for e-commerce companies. It bring companies back to the drawing board to ensure they can comply with the current regulations while limiting losses that rose from lack of clear direction from the start. The final e-commerce policy has been held back for another year, putting existing investments of firms in this sector at risk during the interim months, and deterring further investments.
2. The uncertainty around data localisation is another deterrent. The recent announcement by a high-level government panel to do away data localisation for non-critical data, and the upcoming announcement of the position of the Prime Minister's Office on data localisation are policy announcements that drive sentiments on this debate, though none are legal instruments. Under data localisation requirements, companies would need to redesign internal algorithms to access data locally, pay up for new servers, and face costs to protect data in less-secure environments. The predictive power of firms' algorithms would weaken with fewer data points to train models on. The due process of discussions with various government bodies and stakeholders on this issue is still in process. The RBI's requirement for financial data localisation despite existing provisions (Bailey and Parsheera, 2018) for access under the Payments and Settlements Act (2007) suggest that any Indian regulator can step in with special requirements at unforeseen times.
3. A related issue is the disclosure requirement of source code under the draft e-commerce policy. E-commerce firms depend on data-driven marketing and use of collaborative filtering for customer recommendations. A code submission requirement is a coercive technique aimed at achieving the transfer of technology and local needs' described under the proposed e-commerce policy. Technology transfers cannot and should not be coerced: they happen in an organic and legitimate manner through managers and employees developing skills and passing them onward in data communities or by workers moving across companies (Bloom et al, 2019). It is also doubtful how technological transfers can be achieved with segments of code without underlying data. Will code disclosure requirement be combined with data localisation to pass on core business value to competitors? Will companies need to invest in staff and technologies to find workarounds to be able to mask their key assets? Whether such a code disclosure requirement will come into effect remains unresolved. In mid-2020 the final e-commerce policy will describe the stand of the government on this issue, but this is not definitive either.

Multiple guidelines on the same subject can cause delays in the resolution of uncertainty. The RBI Report of the Working Group on FinTech and Digital Banking includes E-aggregators, Robo advisors and Big Data all under Fintech. E-commerce firms, which are data intensive and provide multiple services, will be included under this description. The fintech steering committee report of the Ministry of Finance is still pending. Each of these reports is a statement about how government agencies are likely to move in the future but these are not legal instruments. Government reports can only suggest but not surely state how future laws will change.

Infirmities of the regulatory processes in India also exacerbates uncertainty. As an example, data localisation requirements by RBI for payments firms were translated from an early idea into an enforced law within a matter of days. There was no due process surrounding how officials could change the law.

The last leg of the legal system -- how laws are enforced -- also suffers from concerns about non-equal application of law, as shown in the examples from RBI (Roy and Shah, 2015) and SEBI (Aggarwal and Zaveri, 2019). For a prospective investor, the risk of investing in India lies in how the law might change in the future through an undemocratic process, and in how the law will be applied to her.

#### Conclusion

For India to have a stable investment environment, we need to provide firms a stable and predictable policy environment. Investments from firms in various sectors will boost the investment cycle for India. Resolving policy uncertainty both at the macroeconomic level as well as in different sectors, and reducing regulatory risk through better rule of law is critical for India in the current investment scenario as well as for long term growth.

#### References

Aggarwal, Nidhi and Zaveri, Bhargavi. Problems with evidentiary standards for proving securities fraud in India, The Leap Blog, 23 August 2019.

Altig, David, Jose Maria Barrero, Nicholas Bloom, Steven J. Davis, Brent H. Meyer and Nicholas Parker. Surveying Business Uncertainty University of Chicago Working Paper (2009)

Bloom, Nicholas. The impact of uncertainty shocks. Econometrica (2009)

Bloom, Nicholas, Erik Brynjolfsson, Lucia Foster, Ron Jarmin, Megha Patnaik, Itay Saporta-Eksten, and John Van Reenen. What Drives Differences in Management Practices? American Economic Review (2019)

Baker, Scott R., Nicholas Bloom and Steven J. Davis. Measuring Economic Policy Uncertainty. The Quarterly Journal of Economics (2016)

Bailey, Rishab, and Smriti Parsheera. Data localisation in India: Questioning the means and ends, The Leap Blog, 22 February 2018.

Roy, Shubho, and Ajay Shah Payment bank entry process considered inconsistent with the rule of law, The Leap Blog, 1 September 2015.

Megha Patnaik is faculty at the Indian Statistical Institute, Delhi and Fellow at the Esya Centre. The author thanks Radhika Pandey and Ajay Shah for useful inputs.

## Friday, August 23, 2019

### Problems with evidentiary standards for proving securities fraud in India

by Nidhi Aggarwal and Bhargavi Zaveri.

### Introduction

Did O.J. Simpson kill his wife? A criminal jury said no, a civil jury said yes. The standard of proof applied by the two juries made all the difference to the outcome of the case (Vars 2010).

In India, the securities regulator adopts a very low standard of proof for cases involving wrongdoing in the securities market. For many people, standard of proof related questions are procedural and semantic exercises in the dispensation of justice. However, the standard of proof adopted by a judge has direct impact on the outcome of the case and over time, the quality of the investigation conducted by the investigative agency. When the standard of proof is low, there is a high chance that initiating an investigation will induce an adverse order. This creates substantial discretion in the hands of the investigator, to choose the persons against whom State power will be directed. This runs against a basic theme of liberal democracy, of containing executive discretion. The ability of the executive to direct punishment upon chosen ones is inconsistent with the rule of law. It creates policy risk for persons who may consider participating in the Indian financial markets, and creates a bias in favour of participation by politically connected persons.

There are three reasons why the standard of proof in securities fraud cases in India are low. First, the Supreme Court has held that the standard of proof which SEBI must meet to establish securities fraud is the 'preponderance of probability' standard. This is lower than the standard of proof required to establish a crime under criminal law. In civil proceedings, there are usually two versions of the facts. The court, on the basis of the evidence before it, chooses that version which it thinks is 'more probable', that is, it will accept a version which a prudent man will act upon the supposition that it exists. On the other hand, in criminal cases, the prosecutors must satisfy the court that the existence of a fact is not only probable, but that its existence is beyond reasonable doubt. Simply put, the prosecution must satisfy the court that 'a reasonable alternative version is not possible' (185th Report of the Law Commission). Courts have explicitly acknowledged that it is not possible to mathematically define the degree of probability for meeting a certain standard of proof and there is an inherent subjective element within each of these standards. (State of UP v. Krishna Gopal and Anr.)

The other two reasons are inter-connected. SEBI exercises regulation-making, executive and quasi-judicial powers in connection with the securities market. It defines what conduct would constitute fraud for the purpose of exercising its enforcement powers. The concept of fraud under the SEBI (Prohibition of Fraudulent and Unfair Trade Practices) Regulations, 2003 (PFUTP Regulations) - a regulation that defines fraudulent conduct in the securities market - is wider than what is understood as fraud in common law and the codified law applicable to fraud in India. It dispenses with critical elements such as intent, deceit and damages. A wider definition of fraud sets the bar very low for establishing securities fraud before a court or tribunal. SEBI is also responsible for conducting investigations of suspected fraud and makes decisions on whether the conduct investigated meets its definition of fraud. This violates the constitutional scheme of the separation of powers that applies to areas of public administration which are not governed by a technocratic regulatory agency. As an example, the Indian Penal Code defines what constitutes theft, the investigation is conducted by the executive agencies and the decision on whether the investigated conduct amounts to theft (as defined in Parliamentary law) is made by the judiciary.

The concentration of all three powers in a single body creates scope for bias towards a lower standard of proof. Empirical work done on the cases investigated by the Securities Exchange Commission, the securities regulator in the United States, is indicative of such bias. The SEC is empowered to choose whether to pursue a proceeding before one of its own internal administrative law judges or an independent federal court. Reportedly, while the SEC enjoyed a 90% success rate in its own hearings, it had only a 69% success rate against defendants in federal court. (here and a perma link here)

In February 2018, a judgement of the Supreme Court further diluted the standard of proof for securities fraud in India in a case involving synchronised and reverse trades executed on the exchange. The Supreme Court reversed the decision of the Securities Appellate Tribunal which had held that although the trades in question might have been synchronised, they were not manipulative and "market manipulation of whatever kind, must be in evidence before any charge of violating these Regulations could be upheld." The Supreme Court dispensed with the need to show manipulation and relied on the notion of "market integrity" as a standard for adjudging the conduct of market participants. SEBI has extensive powers to sanction wrongdoing in the securities markets, such as the power to bar access to the market, suspend professional licenses and impose hefty monetary penalties. The Supreme Court's ruling has serious implications for the manner in which these powers are exercised as it effectively introduces a new standard of proof of 'market integrity' to be met by persons accused of securities market fraud.

### What is market integrity?

The notion of market integrity is both un-defined and hard to measure. World over, there is considerable debate on its meaning in the context of financial regulation (Austin, 2016). Given the subjectivity of the concept, using market integrity as a standard of proof is equivalent to using 'public interest' or 'public good' as a standard for establishing wrongdoing. It leaves tremendous scope for discretion and creates the potential for differential standards of enforcement across a range of practices, depending on the adjudicator's view of whether or not a particular trading practice affects market integrity. This creates uncertainty in the manner in which the law will be applied and enforced and has adverse implications for the rule of law. Ambiguity in the grounds of enforcement and the standard of conduct that could invite legal sanctions, is detrimental to the development of the market as well.

In this article, we advocate the use of empirical approaches for establishing wrongful conduct in the securities markets. We do this by demonstrating the empirical evidence that should have been used to support a claim of fraud in the very same case where the Supreme Court lowered the standard of proof by relying on the vague and problematic ground of market integrity. In a world of electronic trading, empirical evidence of fraud and its impact on the market is not hard to collect and investigators and courts must rely on such evidence instead of holding market participants to a vague and subjective notion of market integrity and unfairness, which the world at large is struggling to define.

### Judgement of the Supreme Court in Rakhi Trading

In February 2018, the Supreme Court in the case of Securities and Exchange Board of India v. Rakhi Trading Private Ltd. upheld an order passed by the Securities and Exchange Board of India (SEBI) that levied a penalty on some traders who had synchronised their trades off the exchange before placing them on the exchange. The trades in question were a series of orders placed on the F&O segment of the Nifty index. The modus operandi was to place an order for Nifty options, which matched with a particular party and subsequently reverse the position taken by placing an opposite order, which also matched with the same party. SEBI penalised the party placing such orders on the ground that these "transactions were in the nature of fictitious transactions resulting in creation of misleading appearance of trading in these options." The SEBI order did not elaborate the manner in which the synchronised trades sought to manipulate the price of either the option itself or the underlying securities, which in this case, was the basket of securities included in the Nifty index.

The parties against whom this order was passed appealed against the order before the Securities Appellate Tribunal (SAT). SAT confined its review to whether the synchronised trades in the F&O segment of the Nifty index artificially manipulated the underlying cash segment, which in this case is the Nifty index itself. It observed that:

"To say that some manipulative trades in Nifty options in the F&O segment could influence the Nifty index is too farfetched to be accepted. The only way Nifty index could be influenced is through manipulation of the prices of all or majority of the scrips in the cash segment that constitute Nifty."

This in line with precedent case-law laid down by SAT on the requirement to show manipulative conduct to demonstrate that the synchronised trades constituted fraud under the PFUTP Regulations. On an appeal by SEBI against this order of the SAT, the Supreme Court reversed the order of the SAT without explaining how the synchronised trades in question affected the price discovery system or created a misleading impression of volumes, but emphasised the notion of market integrity as under:

"According to SAT, only if there is market impact on account of sham transactions, could there be violation of the PFUTP regulations. We find it extremely difficult to agree with the proposition...SAT has missed the crucial factors affecting the market integrity, which may be direct or indirect (emphasis supplied) ...By synchronization and rapid reverse trade, as has been carried out by the traders in the instant case, the price discovery system itself is affected."

### The problem with synchronised trades

Synchronised trading involves pre-negotiating the trade off the exchange and subsequently placing the order on the exchange such that it matches with the counterparty with whom the trade was pre-negotiated. To synchronise trades on an exchange platform, the buyer and seller of the pre-negotiated trade will enter their respective orders at the same time (with same price and quantity) to maximise the chance of matching their orders against each other. To ensure that the order does not match against another counterparty, the first order may be placed away from the touch, that is, at a price significantly different from the ongoing bid / ask price.

By itself, synchronised trading is not a harmful practice. In fact, block trades for which exchanges have a block trading window are synchronised trades. It is difficult to synchronise trades on liquid securities on the exchange because such orders run the risk of matching against other counterparty(ies). However, even on relatively illiquid securities, synchronised trading is not risk-free. The probability of getting hit by another order (especially a market order) on the opposite side of the book is low, but not zero.

How can synchronised orders be used to manipulate the market? Synchronised trades could create misleading, artificial trading interest in a security. High volumes and significant price changes on an otherwise illiquid security may cause participants to believe that there is some news on the security. This may induce them to buy those securities based on unexplained changes in prices and volumes. This is especially possible in an illiquid scrip, which may be perceived to be suddenly liquid if a series of transactions are executed on such a scrip. While the SEBI order levying the penalty does not specify exactly what was manipulated, evidence of manipulation could be demonstrated in either or all of the following ways:

1. Manipulation of Nifty index: One way of determining manipulation in the Rakhi Trading case is by analysing the changes in the value of the underlying security, the Nifty index. The SAT order almost exclusively focused on the possibility of the synchronised trades having manipulated the value of the Nifty index. The order rightly concludes that price manipulation on the index can happen only if an equivalent position is taken on the spot market on all the 50 constituent stocks of the index. Neither the SEBI order nor the Supreme Court order show any evidence of a position by Rakhi Trading on the spot market. This leaves the scope of manipulation on the Nifty index value to the price transmission from derivatives market to the spot market. Empirical evidence on price transmission from derivatives market to spot market suggests that such transmission is subject to the liquidity of the derivatives instrument (see Fleming et al, 1996, Aggarwal and Thomas, 2019). We examine the liquidity of the relevant Nifty options and also examine if the volumes traded by Rakhi Trading in the options segment were significant enough to impact the value of the underlying Nifty index.
2. Volatility in option premium: Manipulation of stock price increases the volatility in returns on such stocks (Aggarwal and Wu, 2006). If the synchronised trades on the Nifty options were manipulative, we would expect that the real price discovery process on option premium would be hampered, resulting in higher volatility in the premium of the Nifty options in question. This manipulation can be established by examining the volatility of the premium on the Nifty options on which the synchronised trades were executed.

### Size of the market for the relevant Nifty options

Rakhi Trading executed synchronised trades on 13 specific Nifty option contracts (hereafter referred to as "the relevant Nifty options") on four days of the year 2007, namely, on March 21, March 22, March 23 and March 30 (hereafter,"the event days").

We begin by a simple comparison of the liquidity and volatility of the relevant Nifty options on the event days and compare it with other days between March 15, 2007 and March 31, 2007 (hereafter, "non-event days"). Such a comparison should be the beginning of the court's enquiry when dealing with an order punishing a market participant for securities market fraud. Table 1 provides basic summary statistics on the traded volumes (a measure of liquidity) and volatility of the relevant Nifty options.

 Event days Non Event days Total Volumes Volatility Rakhi Trading Volumes Total Volumes Volatility (%) (%) Min 9,550 0.97 2,000 50 0.10 Mean 26,142 6.19 10,015 7,712 2.25 Median 25,750 5.08 10,700 3,925 0.37 Max 38,600 14.85 11,900 50,050 17.62 SD 7,752 3.86 2,477 10,718 4.39

The key observations from Table 1 are as follows:

1. First, the traded volumes of the relevant Nifty options fluctuated significantly on a daily basis on the event and non-event days, ranging from 9,550 to 38,600 on the event days, and 50 to 50,050 on the non-event days.
2. Second, in comparison to the average daily traded volumes on a liquid Nifty option, Table 1 shows that the relevant Nifty options were relatively illiquid. For a frame of reference, the maximum traded volumes on one single Nifty option in March 2007 was 9.2 million.
3. Third, the volatility (measured by the Parkinson's range measure) of the relevant Nifty options was in the range of 1-15% on the event days, while it was in the range of 0-18% on the non-event days.

### Did the synchronised trades manipulate the Nifty index?

Table 2 gives a picture of the size of the overall Nifty options market on the event days and compares it for non-event days based on traded volumes. It also shows the traded volumes of the stocks which constitute the Nifty index for these dates. The last column of the table shows the volumes and contracts traded by Rakhi Trading on the event days to provide a perspective on the possible influence of its trades on the overall options and underlying spot market.

Table 2: Size of overall Nifty options market and proportion of synchronised trades
Non-event
days
All Nifty Options Traded volumes (in
Rs. millions)
49,978 61,041 277
Contracts
(in 000s)
262
316
1.35
volumes (in Rs. millions)
36,228 37,932 NA

The key observation from Table 2 is that on the event days, the volumes traded by Rakhi Trading on the Nifty options were less than one percent of the average traded volumes on the stocks that constitute the Nifty index on the spot market. The minuscule proportion of the volumes traded by Rakhi Trading in the Nifty options market relative to the total traded volumes on the stocks constituting the Nifty index on the spot market, re-affirms the finding of the SAT that the synchronised trades executed by Rakhi Trading could not have possibly manipulated the underlying Nifty index.

However, Table 2 also shows that the volumes and number of contracts on the Nifty options segment on the event days were higher than on non-event days. This might or might not have been due to the synchronised trades executed by Rakhi Trading. In the next few paragraphs, we zoom in our analysis on the specific Nifty options which were involved in the synchronised trades executed by Rakhi Trading and examine if those trades did manipulate the individual options traded by Rakhi Trading.

To test the claim of manipulation of option premium, we examine the volatility of the premium of the relevant Nifty options, and compare it with the volatility of the premium of other Nifty options with similar liquidity. We call the former as the treated set, and latter as the control set. We identify the control set as the options on which the traded volumes were in the same range as that on the treated" set, to ensure comparability across the two sets. If there was indeed manipulation on the treated options, we expect the volatility of the treated set to be higher than that of the control set.

We obtain a total of 50 unique options in the control set which we compare with data on the treated set on the event days. Table 3 presents summary statistics on the volatility of the premium for the options in the treated set and control set for our period of analysis.

Table 3: Summary statistics on volatility of options premium on the treated and control sets (in %)
Treated Control
Min 0.97 0.25
Mean 6.19 4.78
Median 5.08 2.91
Max 14.85 35.17
SD 3.86 5.09

We observe that the average volatility of the treated set was slightly higher than that of the control set. However, a simple t-test of comparison of means of the treated and control set volatility shows that the difference between the volatility of the two sets is not statistically significant. In a regression analysis (not shown here), we also control for other factors that affect volatility of the option premium of the treated and control sets. We do not find any evidence of significant difference across the two sets even after controlling for other factors such as strike price, days to expiry, value and volatility of the underlying. The analysis finds that the price range in which the option premium varied for the treated set was similar to that of the control set. Thus, we find that the option premium on the Nifty options that were traded by Rakhi Trading was not manipulated.

### Did the synchronised trades manipulate the volumes?

We also examine the question whether the synchronised trades in the relevant Nifty options led to higher volumes in these options thereby creating a possibly misleading impression of volumes. For this, we analyse the traded volumes on the relevant Nifty options, after excluding the volumes arising out of synchronised trades themselves. We compare the traded volumes of the relevant Nifty options on the event and non-event days (Table 4). If the synchronised trades did result in higher trading activity from other market participants, we expect to see significant difference in the traded volumes on the event and non-event days.

Event days Non-event days
Min 0 50
Mean 5,342 5,170
Median 4,500 2,450
Max 15,400 21,350
SD 5,736 5,762

Table 4 shows that the traded volumes on the Nifty options involved in the Rakhi Trading case were, on an average and on a median scale, slightly higher on the event days compared to the non-event days. However, a statistical test (t-test) of the comparison of means finds no significant difference across the two sets. A regression analysis on the event and non-event set confirms this finding. This indicates that Rakhi trading trades did not lead to any jump in volumes in the options so traded.

### Trading for tax evasion or tax planning

An ancillary concern expressed by the Supreme Court was that Rakhi Trading conducted the trades in question for tax planning or avoidance. While this may or may not be true, the securities markets regulatory framework should not be used for punishing tax evasion. Cases of trades that SEBI has reason to believe were meant for tax avoidance, must be reported to the tax authorities, which is the appropriate forum for addressing questions of tax evasion. The objective of the securities market regulatory regime is not to deal with tax evasion, but to protect investors and develop the securities markets. More importantly, a regulator has scarce resources and dedicating investigative and adjudicatory capacity for dealing with tax evasion cases is not the best use of these resources.

### Conclusion

Our empirical analysis finds that the synchronised trades did not manipulate the underlying Nifty index, the premium on the relevant Nifty options or the traded volumes of the relevant Nifty options. However, we recognise that our analysis is limited to daily data. An analysis of this kind must be underpinned by examining the intra-day data around the time of synchronised trades. By using such data, the regulator can further make a case for whether the trades in question were indeed manipulative.

The objective of this article is not to establish the guilt or innocence of any specific market participant. By using publicly available inter-day trading data on the security involved in the Rakhi Trading case, we make a case for using empirical strategies to establish fraudulent conduct under the Indian securities regulatory regime. As demonstrated above, the advanced nature of the securities market infrastructure in India and the availability of data ensures that this is not difficult. The use of empirical evidence to substantiate charges of fraudulent conduct will ensure that enforcement orders pass the muster of the appellate forums without having to compromise on evidentiary standards for establishing guilty conduct. More importantly, backing enforcement with robust underlying evidence will help the regulator build the trust of the regulated and testify to the high standards of proof that our society should place for the deprivation of liberty.

### References:

Stock market manipulations by Aggarwal R and Wu G, The Journal of Business, 2006.

When stock futures dominate price discovery by Aggarwal N and Thomas S, Journal of Futures Markets, 2019.

What exactly is market integrity? An Analysis of One of the Core Objectives of Securities Regulation by Austin J, William and Mary Business Law Review, Vol.8(2), 2017.

Trading costs and the relative rates of price discovery in stock, futures, and option markets by Fleming J, Ostdiek B and Whaley R, Journal of Futures Markets, 1996.

State Of U.P vs Krishna Gopal & Anr 1988 AIR 2154.

185th Report of the Law Commission of India on a Review of the Indian Evidence Act, 1872.

Toward a General Theory of Standards of Proof, Frederick E. Vars, Catholic University Law Review, Vol. 60(1) (Fall 2010).

Nidhi is faculty at IIM-Udaipur and Bhargavi is a researcher at IGIDR.

## Wednesday, August 07, 2019

### IBC (Amendment) Bill, 2019: Implications for judicial review of resolution plans

by Pratik Datta and Varun Marwah.

Shroff and Misha (2019) had earlier highlighted that excessive judicial discretion in corporate insolvency resolution is contrary to the express provisions of the Insolvency and Bankruptcy Code, 2016 (“IBC”). Indian policymakers have now recognized this problem. Accordingly, the IBC (Amendment) Bill, 2019 (“2019 Bill”) has been passed by the Parliament.

The immediate trigger for this reform was the National Company Law Appellate Tribunal (“NCLAT”) judgment dated July 4, 2019, in the resolution of Essar Steel Ltd. (“Essar judgment”). In this case, the NCLAT found ArcelorMittal’s resolution plan to be discriminatory. Accordingly, it went on to modify the plan such that the financial and operational creditors enjoy the same recovery rate. In the process, it obliterated the distinction between secured and unsecured creditors. Moreover, the NCLAT also held that the statutory waterfall under section 53 of IBC does not apply to distribution under a resolution plan. By disrupting the basic fundamentals of banking, this decision caused much concern in the Indian financial sector. Among others, the 2019 Bill seeks to address these concerns as well.

In this backdrop, we contextualize the amendments proposed by the 2019 Bill within a hypothetical theoretical framework to analyse their implications on the scope of judicial review of resolution plans under IBC.

### Theoretical framework

An insolvency law should have two broad objectives. First, it must achieve the most efficient economic outcome for the insolvent company. This outcome could be going concern sale, restructuring, liquidation or a combination of these outcomes. Second, the value (cash/non-cash) received from the outcome must be distributed among the claimants of the company according to the waterfall provided in the insolvency law.

In a going concern sale, the business of the insolvent company is marketed for price discovery, that is, auctioned. Potential buyers submit their price bids in their respective resolution plans. Once the best price is discovered, the price signifies the precise value of the business. The successful buyer deposits this value in cash or its equivalent in an escrow account and takes over the business, immediately putting the assets to their best use. The resolution plan need not deal with how the deposited value would be distributed among the insolvent company's claimants.

The value deposited in the escrow could be separately distributed by the resolution professional among the claimants of the insolvent company as per the statutory waterfall. This scheme of distribution cannot be altered unless a claimant(s) affected by such alteration specifically consent to it. However, the resolution plan need not deal with any of these distribution issues. Therefore, the resolution plan cannot in any way unfairly discriminate against any claimant. Consequently, there is no need for judicial review of resolution plans to ensure fair and equitable distribution in a going concern sale for cash or its equivalent. The same logic applies to distribution in liquidation involving sale of assets for cash or its equivalent.

However, going concern sales are not always possible or desirable. For instance, during a recession, there could be no buyers in the market. Or even if there are buyers, there could be an oversupply of similar assets in the market due to industry wide factors, pushing down the price for such assets. In such circumstances, instead of a going concern sale to a new buyer, the claimants of the insolvent business may be better off by “selling” the business to some or all of the existing claimants themselves. Such “hypothetical sale” is commonly referred to as restructuring.

In such a restructuring, there is no auctioning of the insolvent business to potential outside buyers. As a result there is no price discovery and the precise value of the insolvent business in cash or its equivalent is not evident. Therefore, the notional distribution of rights (that is, securities in the restructured company) among the claimants is conceptually very different from the distribution of value in cash or its equivalent in a going concern sale.

Unlike a going concern sale, the notional distribution of rights in a restructuring has to be determined by the resolution plan. This creates two unique problems. First, given the uncertainty regarding the value of the restructured business itself, there could be ambiguity about how much value should each claimant receive under the waterfall. Second, even if the value payable to each claimant is agreed upon, there could be ambiguity regarding the value of the rights (that is, securities of the restructured company) allocated to each claimant by the resolution plan. These unique problems create a peculiar risk in restructuring - the rights distributed by the resolution plan may give one or more classes of claimants a value lesser than what they are entitled to under the statutory waterfall. This is often referred to as “unfair discrimination”. To prevent such unfair discrimination, judicial review of resolution plans in non-cash restructuring transactions is necessary.

This conceptual distinction between sale (cash) and restructuring (non-cash) transactions is critical. Minority creditor protection mechanisms including judicial review is necessary primarily for restructuring (non-cash) transactions; not so much for sale (cash) transactions. For example, Chapter 11 of the US Bankruptcy Code deals with restructuring. Section 1129 in this chapter requires a restructuring plan to provide every creditor at least the liquidation value. This safeguard is not applicable to sale transactions (for cash) under Section 363 of Chapter 3 of the US Bankruptcy Code. Evidently, the US law clearly recognises the difference between sale (cash) and restructuring (non-cash) transactions.

So far we have only dealt with neat categories of going concern sale and restructuring. However, in reality, resolution plans could propose transactions that may not squarely fall within these simple categories. For instance, some new investors may pay up cash to buy some equity or take on some debt of the restructured company, while existing lenders may swap some of their debt into equity of the restructured company. Such transactions would then be a hybrid
exhibiting features of both a going concern sale for cash as well as that of restructuring for non-cash considerations. It is important to recognise that such transactions effectively help preserve the going concern value of the insolvent business. Therefore, in such transactions, each claimant of the insolvent company must get at least the amount of value (cash, non-cash or both) that it would have received had the company been sold off as a going concern for cash or its equivalent. In other words, even in such transactions, the distribution of value must follow the statutory waterfall. And this distribution function has to be performed by the resolution plan itself. Consequently, the risk of unfair discrimination may arise even in such transactions, rendering judicial review of resolution plans necessary for minority creditor protection.

### Shortcomings in IBC and the proposed solutions

The IBC, as initially envisaged, differed from this ideal theoretical framework on two counts.

First, the statutory waterfall under section 53 of the IBC was initially envisaged only for distribution in liquidation. The Essar judgment held that section 53 cannot apply to distribution proposed by a Resolution Applicant. Consequently, in sale (cash) or restructuring (non-cash) transactions, a resolution plan would have to follow a scheme of distribution different from the statutory waterfall, opening up flood-gates of unfair discrimination allegations.

The 2019 Bill attempts to address this problem. Section 30(2)(b) is proposed to be amended to make the statutory waterfall under section 53 applicable to distribution in resolution as well. This clarity should reduce allegations of unfair discrimination in resolutions and help cut down on unnecessary judicial review in going concern sales.

Second, the IBC does not clearly recognise the conceptual difference between cash and non-cash transactions. Unlike US Bankruptcy Code, section 30(2) of IBC applies minority creditor protection mechanisms, meant primarily for restructuring (non-cash) transactions, to even sale (cash or its equivalent) transactions. This provision has been liberally interpreted by the NCLAT in Binani Industries to hold that every resolution plan must be fair and equitable – a vague judicially determined standard. Consequently, every resolution plan is now open to judicial review without any clear purpose. This has unnecessarily delayed many corporate insolvency resolutions. For example, the Essar resolution has been ongoing for almost 2 years, even after approval of a resolution plan by the CoC.

The 2019 Bill attempts to address this problem too. The proposed Explanation 1 to section 30(2)(b) explicitly clarifies that if a resolution plan follows the order of priority under section 53, it will satisfy the ‘fair and equitable’ standard. This explanation should narrow down the focus of judicial review to ensure distribution as per the statutory waterfall.

### Conclusion

Judicial review under IBC should be focused to ensure that distribution of value is as per the statutory waterfall under section 53. In going concern sales for cash or its equivalent, distribution of value among claimants based on the waterfall would be relatively unambiguous and should not require judicial review of resolution plans. In restructuring or other hybrid transactions involving non-cash consideration, notional distribution of non-cash value in form of rights to the restructured business (that is, its securities) is likely to be contentious, given that distribution will be provided by the resolution plan and/or issues relating to price discovery and valuation. This could create opportunities for value extraction from minority to majority claimants through resolution plans in such restructuring or hybrid transactions. Judicial review of resolution plans may be necessary only to prevent such unfair value extractions in violation of the statutory waterfall. But once distribution among claimants as per the statutory waterfall is achieved, there is nothing further to be gained by subjecting each and every resolution plan to judicial review. The 2019 Bill has now modified the law to achieve this optimal outcome. Whether this outcome is achieved in practice will depend largely on how the judiciary interprets and implements these new provisions.

### References

Shroff, S. and Misha, A question of balance, via the code or otherwise, Asia Business Law Journal, May 13, 2019.

Datta, P, Value Destruction and Wealth Transfer under the Insolvency and Bankruptcy Code, 2016, NIPFP Working Paper No. 247, Dec. 2018.

Pratik Datta is a Senior Research Fellow, and Varun Marwah is a Research Fellow at Shardul Amarchand Mangaldas & Co. We thank Mr. Shardul S. Shroff and four anonymous referees for useful discussions and comments.

## Monday, July 15, 2019

### How land laws create dead capital: A case study of Maharashtra

by Diya Uday.

Land is an important form of capital. In India, for most households, it is the dominant element of the household portfolio. About eighty per cent of all household assets in India, are in the form of real estate (land, buildings and other constructions owned by households, for residential, commercial or vacation purposes) (Ramadorai Committee Report, 2017, pg. 12). In the CMIE 'Consumer Pyramids' household surveys, almost all households own some land or real estate.

As with other factors of production, an efficient economy requires full utilisation of the resources in the country, efficient mechanisms for discovering its price, and frictionless transactions. This is not taking place well in India. For example, unsecured debt still accounts for two-thirds of the total liabilities for the very poor and one-third of the rich in India (Ramadorai Committee Report, 2017, pg. 6). Similarly, despite landlessness in rural areas, only fourteen per cent of all households reported leasing-in land (NSSO Report, 2013, pg. 30). In many states such as Maharashtra, the proportion of households leasing-out land is well below ten per cent (NSSO Report, 2013, pg. 30).

In the study of land economics in India, a key question that has to be pursued is: Why, despite the seizable presence of land as an asset in household balance sheets, is there poor capitalisation of land in India? What obstructs harnessing the full productive capacity of land in India?

The idea of land as dead capital was made prominent by Hernando de Soto, who used the term to refer to something that could not be easily bought, sold or used for an investment. He showed that the poor possess far more capital than is evident, but institutional failures hinder utilisation of the land as wealth. For example, he found that in Egypt a person who wants to acquire and legally register a lot on either state-owned desert land or former agricultural land has to navigate a plethora of bureaucratic procedures, estimated to take anywhere between five to fourteen years. As a consequence, a large number of people chose to build dwellings illegally (de Soto 2000, pg. 20). This means that these properties cannot be used to access formal credit or be legally sold or rented.

There is a need for a comparable literature on India. How does the land market work? What are the impediments to translating land into value added? How can wealth in the form of land impact upon the life of the owner to a greater extent than is presently the case?

In this article, three questions are studied, treating Maharashtra as the environment under examination:

1. Do regulations on land hinder the effective utilisation of land as an asset in India?
2. What are the restrictions imposed?
3. What are the less visible effects of these restrictions?

### Three pathways to harnessing land wealth

The value of land is unlocked in three ways:

1. A mortgage, whereby land is used as a security to access credit.
2. A sale, where the owner of the land transfers it to a buyer.
3. A license or lease on the land in exchange for regular payments.

There has been a considerable focus on mortgage transfers as a means of capitalising the value of land. Land titles and access to credit are now intricately connected in policy discourse on financial inclusion and access to finance. For example, the RBI has recognised the role of land titles in access to credit and consequently to financial inclusion (Mohanty Committee Report, 2015, pg. 26). Some court interventions too, have given creditor rights precedence over transfer restrictions on land.

The other two methods -- sale and lease -- have received less attention, but are no less important. The owner of an asset must be given the freedom to choose the method by which the asset is to be capitalised. Directing policy attention only towards land as a means of accessing credit, and ignoring reforms in sale and rental markets, reduces the choices available for a land owner.

### A case study of land laws in Maharashtra

We now turn to identifying provisions of law that affect the transferability of land in Maharashtra.

Methodology. A list of laws was obtained from the website of the Bombay High Court. This list was examined to identify the laws that potentially affect transfers of both agricultural and non-agricultural land and real estate, by reading the names of the laws. The long titles of these short-listed laws were examined to assess the applicability of the law for this case study. This yielded the following list of laws that impact upon land transfers:

1. Maharashtra Land Revenue Code, 1966
2. Maharashtra Tenancy and Agricultural Lands Act, 1948
3. Maharashtra (Prevention of Fragmentation and Consolidation of Holdings) Act, 1947
4. Maharashtra Stamp Act, 1958
5. Registration Act, 1908
6. Maharashtra Rent Control Act, 1999

The analysis of these laws yields the following results:

• Restrictions on the transfer of agricultural land: There are two kinds of restrictions on the transfer of agricultural land. The first kind of restriction is that under the Maharashtra Tenancy and Agricultural Lands Act, 1948, agricultural land can only be transferred to a resident agriculturist. An agriculturist is defined as a person who cultivates land personally. A non-agriculturist can only buy agricultural land after obtaining the permission of the Collector, unless the property is specifically allocated to residential, commercial or industrial uses or is to be used for a bona fide industrial purpose. In all other cases, the restrictions on transfers continue to exist. These restrictions apply to subsequent transfers as well. Similarly, mortgages to lenders other than co-operative societies, also require permission of the Collector. In each case, the Collector may grant permission subject to conditions.

These restrictions induce three problems. First, they increase the cost of transacting on such land. Second, the law does not prescribe a time limit for granting such approvals. Neither are these permissions covered under the Maharashtra Right to Public Services Act, 2015. Without statutory timelines, the procedure for transfer could be time-consuming and tedious. Discretion in delay creates the possibility of corruption. Third, since these restrictions continue to apply even after the land the purchased, they also handicap future purchasers.

A second class of restrictions kicks in after the sale is completed. Where the law has done away with the requirement for permission, the land must be put to the intended and permitted use within five years from the date of transfer. Failure to do so incurs a penalty of two per cent of the market value and even forfeiture. Further, when a purchaser of this land wants to sell it without utilising the land for a non-agricultural purpose, she can do so only with the permission of the Collector and after payment of a transfer fee of twenty five per cent of the market value of the property. If sold within ten years, these restrictions continue to apply to the transferee as well. There are further restrictions placed on certain classes of agricultural land, such as the payment of fifty per cent of the purchase price to the Collector. In case of a delay in the payment of price, this amount increases to seventy five per cent of the purchase price.

• Restrictions on the transfer of tribal land: The Maharashtra Land Revenue Code, 1966, places three types of restrictions on the transfer of tribal land. First, for sale of tribal land to a non-tribal, the permission of the Collector with the previous approval of the State Government has to be obtained. Before the grant of this approval, the Collector has to first offer the land to tribal persons residing in the village of the transferor or within five kilometres of the land. In scheduled areas, the additional sanction of the Gram Sabha has to be taken, unless it is for a 'vital government project' such as for highways, canals, etc. In all cases, the Collector is permitted to grant approval for transfer, with conditions. These restrictions would apply upon a lender, who might repossess land, also.

The second type of restriction relates to the mortgage of such land. In case the mortgage is below five years, the permission of the Collector has to be taken for transfer. In the event that the mortgage is above five years, the permission of the Collector with the previous approval of the State Government has to be obtained. In case of a mortgage to a non-tribal, the same procedure of offering it first to a tribal person within the village or within a five kilometre distance from the land is undertaken. Further, the law permits the Collector to restore possession of the land to the tribal person at the end of the mortgage period, regardless of any court order or law. This means that if a tribal person defaults on a loan where land is taken as collateral, at the end of the term of loan, regardless of a court order to the contrary, the land can be restored to the mortgagee at the discretion of the Collector. These restrictions hamper lending against such land. In a field study conducted across twenty villages in Maharashtra, respondents were unanimous in stating that if they defaulted on a loan for which land was collateral, nothing would happen and that when land is used as collateral it was rarely enforceable if there was a default (Narayanan et. al, 2019). Court rulings on this subject have been conflicting, sending mixed signals to lenders (Zaveri, 2017). Poor transferability also hampers the establishment of a credit history and thus access to credit.

The third type of restriction relates to lease of such land. The provisions requiring permission of the Collector and State Government and the requirement of offering the land to a tribal person within the village or within a five kilometre distance from the land, also apply to lease transactions. These provisions make leasing of such land to anyone but tribals, a lengthy and expensive procedure. Where there are no takers from among the tribal community, the owner of such land is effectively left with one less tool for capitalising the value of her asset.

• Restrictions on the transfer of notified fragments: A fragment of land is defined as a plot of land which measures less than the notified standard area. The Maharashtra (Prevention of Fragmentation and Consolidation of Holdings) Act, 1947 imposes two types of prohibitions and restrictions on such land. First, any land which is notified as a fragment, can only be sold to the owner of a contiguous parcel of land. In addition to limiting the owner's access to the land market, as in the case of tribal land, these sale restrictions also inhibit recoveries of lenders.

Any land which is notified as a fragment can only be leased to the cultivator of a contiguous parcel of land. This provision is problematic in that it operates within an already restrictive lease market, where incentives to lease are few. Further in the event that the cultivator of a contagious parcel is not interested in leasing-in the land, the owner is either forced to cultivate the land herself or to let it lie fallow, effectively making it a dead asset.

The following laws do not place direct restrictions on sale, lease or mortgage transfers, but have provisions that affect these transactions (Category 2 provisions):

• Rental market restrictions: There are two main laws that govern rental markets in Maharashtra, one for agricultural land and one for constructed property. While these laws, unlike those in some other states, do not prohibit leasing of land, they do impose other restrictions. First, the Maharashtra Tenancy and Agricultural Lands Act, 1948, which applies to agricultural land, prescribes rent ceilings. The prescribed formula for determining the maximum amount of rent payable is that the rent must not exceed five times the assessment or twenty rupees per acre, whichever is lower. Similarly, the Maharashtra Rent Control Act, 1999 controls rents in specified properties by imposing a statutory maximum rent which is below the equilibrium rent (determined on the basis of the market value of the property). The law also limits the percentage of yearly escalation in rent chargeable to tenants.

Rent ceilings reduce the rental revenues of owners. In addition to prescribing the value of the agreements, these laws also prescribe other terms such as the grounds of termination exhaustively. This means that the parties have little or no freedom to contract grounds of termination beyond those prescribed in the law. Further, the recovery of possession of the premises is a difficult process which will most likely require administrative or court intervention, which means additional costs to the owner. These features coupled with rent ceilings, leave no incentive to owners who wish to capitalise their asset by means of a lease. In fact, anecdotal data from land-owner farmers in Palghar and Mulshi Districts in Maharashtra suggests that the fear of non-recovery of leased out land is a key reason for not leasing out of land.

• Registration of transfers and stamp duties: The Registration Act, 1908 requires certain deeds used to effect transactions in immovable property, to be registered with the office of the registrar or sub-registrar. A document has to be registered by payment of a registration fee. In Mumbai, this amount is thirty thousand rupees. The biggest problem with this system is that while registering this deed, the registrar is not under any obligation to confirm the veracity of the contents of the deed or the marketability of title. The registration of fraudulent documents of transfer thus is possible. For example, in a recent episode, an examination of property documents revealed forged signatures and non-existent parties to the transaction. This system therefore adds to the cost of the transaction, without any real benefit to the parties. A purchaser, borrower or lessee is incurring the cost of registration without actually having the assurance of marketability of the land or identity of the parties.

Stamp duty is incurred before the registration. Like other taxes upon transactions, stamp duty is a `bad tax' in public finance parlance. The Maharashtra Stamp Act, 1958 imposes a stamp duty of five per cent on the market value of the property in a sale transaction. Recently, a surcharge was introduced which has further increased the applicable stamp duty.

• Presumptive titles: The Maharashtra Land Revenue Code, 1966 requires the updation of land records each time a transfer takes place. For this, the transferee has to make an application for updation of the record. The concerned officer will invite objections. If there are no objections, the record is updated. If there are objections, the dispute will be adjudicated before updating the record. Despite this lengthy procedure, land records do not have any value in terms of proof of title. The Supreme Court recently reiterated that entries of transactions in revenue records, do not create title to land.

Purchasers demand the issue of a certificate of marketability of title from the seller. This involves a process of legal due diligence or examination of title by legal experts. In most cases, the purchaser also conducts their own diligence in addition to demanding this certificate, adding further costs to the transaction.

• Form of land records: There are two types of textual records in the state of Maharashtra: the 7/12 extract for agricultural/rural land, and the Property Rights Card in urban areas. A study on urban records in Mumbai reveals that the fields of information required to be recorded under the rules of Maharashtra Land Revenue Code, 1966, only serve the purpose of collection of revenue (Sheikh et. al., 2018). These records are therefore fiscal cadastres and information contained in these records is limited. Vital information such as easementary rights, restrictive covenants on the land, litigation and encumbrances are not recorded.

Insufficient information in land records increases the cost to and risk borne by the buyer. In case of a mortgage, the cost of lending is also likely to increase as it will take into account these risks, making borrowing more expensive for land owners.

Thus, we have a depiction of how the landscape of laws in Maharashtra interferes with the translation of land into value added. There are three limitations of this work. First, the list of laws is not exhaustive. There are other laws, regulations, government orders, which affect the transferability of the property, which do not find mention in this case study. For example, there are a number of transfer restrictions in urban areas, such as restrictions on change of land use and development under the Maharashtra Regional and Town Planning Act, 1966, the Development Control Regulations, the imposition of transfer fees payable by apartment owners to co-operative societies and transfer fees paid on Collector's land. Second, at present, there is no empirical evidence, that links these provisions to the impact on land economics. Third, this study is theoretical and does not analyse how these provisions will play out on-ground. Depending on the administrative processes, these provisions may have either a large or small impact in obstructing transactions.

### Conclusion

The aim of this study was to examine the regulatory restrictions imposed on land transfers and their potential role in creating dead capital. This study documents two classes of restrictions: Category 1 provisions, which directly restrict transfers in the land market and Category 2 provisions, which affect the ease of doing transactions. Both these categories of provisions create a complex web of conditions for transfer, which may contribute to create dead capital in land markets, with varying effects. For example, a land owner may not be able to capitalise land by transfers on account of Category 1 provisions for two reasons. First she may be outright prohibited from undertaking a transaction, or second, the provision generates an unseen restriction of some kind, which operates to effectively disallow her from capitalising her asset. Category 2 provisions, again may affect the effective capitalisation of land by transfers in two ways. First, by creating disincentives to capitalisation in some manner; rent ceilings are a classic example of this. Second, these restrictions impose costs which reduce the benefits from capitalisation.

Furthermore, this study highlights the unseen consequences of economic policies and the regulations made to implement them. There is a need for policy makers to think about the secondary consequences of any regulation. Each instance listed above is a case of the proverbial coin with two sides, one of which has been overlooked. In the area of land markets in particular, it appears that policies have been unidirectional; predominantly discounting the impact of these policies on capitalisation of land. Recent amendments to the law too reflect this fallacy. For example, the 2016 amendment to the Maharashtra Land Revenue, 1966, introduced the requirement of additional permission Gram Sabha for transfers in scheduled areas. While the argument for this was that it would would lend more accountability to the transfer process, one must also recognise that this increases the complexity and cost of the transfer process which might reduce demand for such land, leading to a situation where even a willing owner, is left with no market for capitalisations.

The effects of these restrictions will be greater in states which have more restrictive regimes. In solving these problems, therefore, the first step is a comprehensive study, at the level of each state, which documents these restrictions. As an example, for Maharashtra, this would be a more complete version of this article. The impeding provisions must be categorised in the manner described by this case study. The reason for this is that different categories of restrictions will require different action points. The second step is to determine empirically if each category of these restrictions affect the capitalisation of land as an asset. Field studies are required which determine the on-ground effects of these provisions on aspects such as (i) the costs of lending, (ii) the cost to the owners of the land and (iii) the transaction frequency.

### References

Narayanan et. al., 2019, Land as collateral in India, Sudha Narayanan and Judhajit Chakraborty, Indira Gandhi Institute of Development Research, February 2019.

Sheikh et. al., 2018, Rethinking urban land records: A case study of Mumbai, Gausia Sheikh and Diya Uday, The Leap Blog, November 1, 2018.

Zaveri, 2017, Distortions in the Indian land collateral market, Bhargavi Zaveri, The Leap Blog, February 1, 2017.

Ramadorai Committee Report, 2017, Report of the Household Finance Committee, Reserve Bank of India, July 2017.

Mohanty Committee Report, 2015, Report of the Committee on Medium-term Path on Financial Inclusion, Reserve Bank of India, December 2015.

NSSO Report, 2013, Household Ownership and Operational Holdings in India, National Sample Survey Office, December-January 2013.

de Soto, 2000, The Mystery of Capital: Why Capitalism Triumphs in the West and Fails Everywhere Else, Hernando de Soto, Basic Books, 2000.

Diya Uday is a researcher at the Indira Gandhi Institute of Development Research, Mumbai and visiting faculty at the Tata Institute of Social Science, Mumbai. The author would like to thank the three anonymous referees for their comments and suggestions.

## Friday, June 28, 2019

### Announcements

#### R conference as part of IISA conference at IIT Bombay, 26 December 2019

We are pleased to announce the first ever R conference under the auspices of the main conference of the International Indian Statistical Association, to held on December 26, 2019 at IIT Mumbai.

Registration is now open and the registration fee is intentionally minimal for students and academics from India. We welcome submission of abstracts on R related topics. We plan to award 40 to 50 scholarships to deserving students and academics around India who are interested in R. While submission of an abstract is not a requirement for a scholarship award, it will, of course, improve the chances of being selected. The scholarship awards may cover the cost of lodging, board and travel.

## Friday, May 24, 2019

### Household debt over time

by Subhamoy Chakraborty and Renuka Sane.

In a previous article, Estimates of household debt in India, we presented some basic facts about household borrowings in the months of May - August 2018. We found that 46% of households in India had debt outstanding. There was a significant reliance on informal sources for the purpose of borrowing. Consumption expenditure was the most important purpose for which households borrowed followed by housing and business.

In this article, we use the same dataset, the Consumer Pyramids Household Survey (CPHS) to understand household borrowing across time. We present metrics on household indebtedness across 3 waves (January-April 2016, January-April 2017, January-April 2018). We have chosen the first wave of each year so that comparisons are made at the same point of time each year. We use household weights for each wave provided by CPHS to get population estimates for share of households having debt and the distribution of debt across different sources and purposes. CPHS does not provide information on the value of debt outstanding. Our analysis, therefore, is restricted to understanding the proportion of households borrowing from various sources for different purposes.

### Overall level of Borrowing

Table 1 presents percentage of households having debt outstanding in Wave 1 in each of the three years. This has increased dramatically over the last three years. In January - April 2016, only 12% of the population had debt outstanding. This increased to 39% by January - April 2018. There has been an increase in household indebtedness in urban regions - by 2018, 40% of urban households had debt outstanding relative to 37% of rural households.

 WAVE NATIONAL RURAL URBAN Jan16 - Apr16 33 million (11.6%) 21.0 million (11.1%) 11.4 million (12.6%) Jan17 - Apr17 67 million (23%) 45.3 million (23.1%) 20.8 million (22.4%) Jan18 - Apr18 114 million (38.8%) 79.8 million (40.0%) 34.3 million (36.5%)

### Share of sources in the population

Figure 1 presents the percentage of households in the population who have borrowed from the different sources. This throws up several interesting facts. First, the share of households who have debt outstanding from a bank has been steadily rising, consistent with rise in personal loans from banking data. The same is true for debt from relatives and friends. Second, there has been a reversal of borrowing from money-lenders. In 2016, less than 3% of households claimed to have debt outstanding from money lenders. In 2017, this rose to a little over 7.5%, and since then has fallen to 4.5% in 2018. Third, there has been a dramatic increase in borrowing from shops. In 2016, less than 1% of households had debt outstanding from shops. In 2018, this number was about 11%.

Figure 1:Share of borrowing in the population

Figure 2 presents the percentage of households who have borrowed from shops in each income decile. Borrowing from shops increased remarkably in 2018 across all income deciles. The increase was highest for the lowest income decile from 2% in 2017 to 15.5% in 2018. The corresponding increase in highest income decile was from 3% to 7.5%.

### Reasons for borrowing

Figure 3 presents the percentage of households who have borrowed for various purposes. Consumption expenditure remains the single largest reason behind household borrowing. Although in 2016 borrowing for consumption and housing were at the same level, housing grew slowly compared to consumption. Borrowings for business and repaying outstanding debt saw a sudden jump in overall share in 2018.

Figure 4 presents the the percentage of household who have borrowed for consumption across income deciles. In 2016 the share of borrowing for consumption was almost equal across income deciles, around 2-3 %. However, by 2018, there was a huge difference between the deciles. The share of borrowing in the lowest income decile increased from 2% in 2016 to 21% in 2018. The corresponding rise in the highest income decile was from 2.5% to 10%. In 2018 21% of households in the lowest decile had borrowed for consumption expenditure as compared to 10% in the highest income decile.

### Conclusion

In this article we have presented some facts about the change in household borrowings in India between 2016 and 2018. These are:

• The percentage of households who have debt outstanding has increased from 12% of the population in 2016 to 39% of the population in 2018. The increase has been slightly higher in urban India relative to rural India.
• There has been an increase in borrowing from banks, relative and friends, and shops. In fact, shops have seen the largest increase as the source of borrowing.
• The incidence of borrowing from shops has been higher for the lower income deciles.
• The biggest jump in the reasons for borrowing has been on consumption expenditure.
• Borrowing for consumption expenditure increased more for lower income deciles.

The authors are researchers with the National Institute of Public Finance and Policy.

### Developing Public Policy Leaders for a better tomorrow

India has completed its phases of elections, with results on the horizon; policy discourses have been in focus during the entire election campaigning. In fact, if one looks at the incumbent Government's records, a number of policy frameworks have emerged: National Policy on Biofuels 2018, National Health Policy 2017, National Energy Policy 2017, National Steel Policy 2017, National Civil Aviation Policy 2016, National Offshore Wind Policy 2015, National Policy of Skill Development and Entrepreneurship 2015, National Agroforestry Policy 2014 are just some of the examples at the national level. States have their own policies, since many of the key areas are State subjects under the Constitution. As the Government moves from being a service provider to a service facilitator, one sees the importance of policies and regulatory bodies on the rise.

Further to the above, the Government has begun exhibiting interest in hiring people from outside the bureaucratic ranks, in many of their policy and development work, as consultants. NITI Aayog, the in-house government think tank, which replaced the Planning Commission, is staffed by young graduates, from top universities in the world, as policy consultants. The Prime Minister Fellowship Scheme is an initiative to attract young people in policymaking. A range of Government departments and ministries are housed by professionals, from various disciplines, engaged in research and advisory services. In fact, as a marked departure from tradition, the Indian government recently recruited 9 individuals working in the private sector at a joint secretary level (senior bureaucrats) as lateral hires.

Interestingly, however, despite the interest and willingness, there stands a lack of quality public policy professionals in the country. This is attributed to a deficiency of good Schools teaching public policy as a discipline. Of those that exist, almost all of them are embedded in a university set-up with regulations that make them less agile in adapting to the rapidly changing world of public policy. More importantly, they offer two-year degree programmes, since India does not recognize a one-year, post-graduate degree.

Signs of change are emerging however; these are being adapted to by educational institutions imparting training in the field of public policy - be it in the duration of the course, the curriculum design, and/or the faculty, which impart the same. Amongst the notable Public Policy Schools is the Indian School of Public Policy, launched in October, 2018 by N. K. Singh, Rajiv Mehrishi, Gurcharan Das and others, and driven by Parth Shah and Luis Miranda of the Centre for Civil Society. The School is an outcome of academics, policymakers and professionals conceptualizing together a world-class policy education institute in India, and offering a one-year postgraduate programme in Policy, Design & Management.

The School’s Academic Advisory Council includes Vijay Kelkar, Shekhar Shah, Jessica Seddon, Arvind Panagariya, Shamika Ravi, Ajay Shah and many others; it enjoys the guidance and support of patrons such as Nandan Nilekani, Vallabh Bhanshali and Jerry Rao. The faculty includes Shubhashis Gangopadhyay, Amitabh Mattoo, Dipankar Gupta, Sanjaya Baru, Pronab Sen, Sudipto Mundle, and other accomplished faculty members.

Industry partners like GMR, PwC, Deloitte, Uber, Ministry of Skill Development & Entrepreneurship, EY, Manipal Education, APCO Worldwide, and many others (https://www.ispp.org.in/industry-partners/) have already come forward endorsing the programme.

The ISPP commences its first batch in August, 2019.

More details on: www.ispp.org.in

### The EVM – VVPAT saga

by Abhay Bhatt and Rajeeva Karandikar.

There has been lot of grumbling in the Indian media about Electronic voting machines (EVM), esspecially over last 3 months, with opposition parties accusing the government of manipulating EVMs. This reached a crescendo recently when all the opposition parties joined hands and filed a petition in the Supreme Court seeking a directive to the Election Commission to cross verify the vote count reported by the EVM, with that of paper slips produced by the VVPAT (Voter verifiable paper audit trail), in 50% of the booths.

The Election Commission had consulted us about the sampling plan to be put in place to instill confidence about the sanctity of the election process. Here is our take on how to think about these questions.

### The background

1. EVMs were first used in the Paravur Assembly constituency in Ernakulam district in the 1982 Assembly poll in about 50 booths. The CPI candidate, Mr. Pillai, defeated the Congress candidate, Mr. Jose, by a thin margin of 123 votes. Mr. Jose went to the High Court with the contention that the Representation of the People Act, 1951, and the Conduct of Elections Rules, 1961, did not empower the Election Commission to use EVMs. While the High Court dismissed the petition, subsequently the Supreme Court upheld the contention and ordered a repoll in these 50 booths. Mr. Jose won the election after this repoll.
2. Subsequently, the Representation of the People Act was amended, and S.61A was inserted, in December 1988, empowering the use of EVMs.
3. These EVMs were designed by public sector undertakings, BEL (a Defence Ministry PSU) and ECIL (an Atomic Energy Ministry PSU), who also manufacture the same. They are subjected to a thorough testing process.
4. The design and production are overseen by a technical expert committee consisting of senior professors in electronics and computer science from leading Indian Institutions.
5. The EVMs do not have any networking hardware. There is no ethernet port, no wifi or bluetooth capability, and thus it is not possible to alter or tamper the memory remotely.
6. The names of candidates on the EVM appear in an order that is determined by the same convention that had been followed since the sixties for the order on the ballot paper. First, the candidates of the national parties appear, in an alphabetical order of their names, then candidates of state parties (again in an alphabetical order of their names) and lastly the rest, again in an alphabetical order of their names. Thus, the order gets determined only after the last date of withdrawal of nomination.
7. The EVMs are distributed across the constituencies via randomisation. The EVMs used in India consist of two units: BU, the balloting unit and CU, the control unit. The BUs and CUs are distributed independently and on the day of polling, the two are connected. If one of them has been tampered or replaced, the handshake between the two units will fail and the pair will have to be replaced.
8. After the voting, the machine is locked by the presiding officer and then the EVM are physically sealed using the same techniques that were used to seal the ballot boxes of old.
9. In the parliamentary elections in 1999, EVMs were used in some constituencies and from 2004, all the elections to the parliament and to the state legislatures are conducted using the EVMs. It should be noted that in several instances, including in 2 out of three parliamentary elections – 2004, 2009 and 2014, the ruling party has lost the election.
10. The fact that there was no way to do a recount, even if a court ordered the same, was a reason for the Supreme Court to ask the EC to find a way of generating a paper trail. The Election Commission appointed a committee of experts, who came up with a design of the VVPAT machine and the Supreme Court ordered that the same be introduced as soon as possible for all elections.
11. The Supreme Court had not ordered the Election Commission to routinely carry out any cross verification of EVM and VVPAT count. The purpose of VVPAT was to have the possibility of a recount if a court so ordered, as a result of an election petition.
12. Since so many doubts had been raised about EVMs, the Election Commission decided that in every assembly constituency, one booth will be picked at random by a draw of lots and for the chosen booth, the VVPAT slips will be counted and cross checked with the count on the EVM. This done in order to increase public confidence in elections.

### Our analysis of the recent debate

Over the last year, demands started coming up for verification of much larger number of EVMs. Various petitions were filed. The election commission engaged the two of us along with Mr. Onkar Prasad Ghosh to advise on an appropriate sample size, so if that many EVMs are randomly chosen and the machine count is verified with the VVPAT count and if no mismatches are found, then we can be confident that defective EVMs, if any, are in insignificant proportion.

We felt that given that the EVM's are assigned randomly to constituencies and that the order of names on the EVMs is determined at a late stage, only after the last date of withdrawal, it is not possible to manipulate or tamper the EVMs centrally.

Given that there is no networking component in EVM, tampering is possible only by getting physical access. If someone can get physical access to an EVM and tamper with the EVM, then the VVPAT slips can also be tampered and so validating EVM count by VVPAT count does not give any guarantee that EVM has not been tampered with. For this we must rely on the elaborate process that the Election Commission has about sealing the EVM in a bag, closing the same and storing in such a way that tampering can be detected. This was also the case with the paper ballot and ballot box method of earlier years. We are no worse with EVMs as compared with the old ways, in this regard.

Also, if some smart mind does figure out a way of changing the memory of EVMs remotely, he/she will not stop with doing so in one or two booths. Certainly, the effort will be to tamper with a larger number of constituencies so as to make an impact on the national level.

That is the reason that we took our objective to be to conclude with high degree of confidence that defective EVMs, if any, are less than a certain percentage of all EVMs in use nationwide. In our report we had taken this as 2%, but it can be 1% or 0.5%.

It is true that our suggested method does not give a guarantee at constituency level. But in our view, such a guarantee is not needed. Our sampling scheme can guarantee that the national picture has not been distorted by tampering or by a manufacturing defect. To those who have been insisting upon constituency level guarantees, we would ask: Why stop at the constituency? Should not every voter be assured that his or her vote has been counted correctly? The only way to do so without compromising privacy is to use cryptography. This is the subject of present research, with Prof. Bimal Roy, former director of ISI involved with one such initiative along with a team in UK. But when such a solution would become available, it can and will be attacked as opaque!

The Election Commission, in its affidavit to the Supreme Court, stated that in the last 2 years, over 1500 EVM counts have been matched with VVPAT counts and in all cases the matching has been perfect. Statistically, this alone is sufficient to conclude that EVM-VVAPT in use currently along with all the safeguards and practices in place are good.

We believe there is a lot of misinformation in the media. Various cases of EVM malfunction are being reported in the media and social media. These relate to local body elections or even college student union elections etc., which are outside the purview of the Election Commission. Other reports of EVM malfunction on the day of election are mostly about EVMs which fail the test before voting starts and are replaced.

While our recommendation was to draw a random sample from the population of all the EVMs in use, the EC decided to continue with their policy (for operational simplicity) to draw one booth per assembly segment, which translates to verifying 4125 EVMs by cross checking VVPAT counts. Based on the data about the number of booths across 4125 assembly segments, we were able to assert that if no defective is found in these 4125 chosen booths, we can say with 99.99999% confidence confidence that the proportion defective is less than 1%. This bound is true irrespective of the configuration of the defective EVMs. For example, a group of miscreants could have tried to tamper with EVMs selectively across a few constituencies. However, as long as the number of defective EVMs is 1% or more, the sampling procedure will catch this with a high probability.

The Supreme Court has ordered or suggested that instead of 1 per segment, 5 EVMs per assembly segment be drawn and VVPAT count and EVM count be cross verified. Based on our calculations, we conclude that if there are no defectives found in the 20625 randomly chosen booths (5 per assembly segment), then with 99.99999% confidence, the proportion of defectives, if any, are less than 0.25%, irrespective of the configuration of the defectives.

The authors are Professor, Indian Statistical Institute, Delhi and Director, Chennai Mathematical Institute, respectively.