The Supreme Court of India: A People’s Court?

ABSTRACT

The consensus among scholars is that the Indian Supreme Court no longer acts as a court for the common person. Contrary to this view, we argue that the Court is still a court of the people in certain ways. We analyse by party and case type all cases decided by the Supreme Court from 2010 to 2014 under its special leave petition jurisdiction, and published in the case reporter Supreme Court Cases. The Court exercises significant discretion when deciding whether or not to admit cases pursuant to this appellate jurisdiction. However, we do not have information about the parties and types of cases that seek and are granted admission. To overcome this gap, we use the hit rate analysis method for deducing whether or not the Court grants easier access at the admissions stage to certain parties and case types. We find that the Court favours for admission individuals over the government in civil cases and defendants over the prosecution in criminal cases. It also favours cases that involve constitutional challenges over cases that don’t. Taken together, we argue that in these contexts, the court tends to favour weaker parties over stronger ones for access to the Court.

KEYWORDS:

• Supreme Court of India
• win rates
• judicialdecision-making
• public interest litigation
• special leave petitions
• Article 136
• empirical methods

Acknowledgements

This project was conceived and initiated by the late Ted Eisenberg, known as the “grandfather” of empirical legal studies. We dedicate this article to him. We would like to thank the University of Chicago Center in Delhi for their grant. We are grateful to a large number of individuals who graciously shared their time with us in interviews that informed every aspect of this research project. We thank the sitting and retired Supreme Court and high court judges, Supreme Court registry staff, Senior Advocates and academicians who met with us in Delhi in January 2016. We thank the participants at the Supreme Court of India and Progressive Social Change conference at Columbia Law School, the Faculty Work-in-Progress Lunch at the University of Chicago Law School, the Cornell Summer Workshop Series Lunch, the consultation held in January 2016 at the University of Chicago Center in Delhi, the Twelfth Conference on Empirical Legal Studies, Nick Robinson and Vik Khanna for their feedback. We are indebted to Omri-Ben Shahar for the vivid phrase “case-by-case rescue”. William Hubbard thanks the Paul H. Leffman Fund and the Coase-Sandor Institute for Law & Economics for research support. This project has been made possible by the dedication and hard work of a large group of students at the National Law University, Delhi, who spent considerable time and took painstaking effort to read and hand-code hundreds of cases each. The student team was brilliantly led over the years by Hemangini Kalra, Sucheta Roy, Sanya Kumar, Shweta Kabra and Kudrat Agrawal. Along with the team leaders, Anwesha Choudhary, Aarushi Mahajan, Shreya Raman and Malavika Parthasarathy hand-coded the bulk of the cases and reviewed the entire data set for errors and consistency. Devanshi Saxena, Akanksha Gautam, Anurag Goswami and Vanya Chhabra also contributed to the project.

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

¹ Constitution of India, Article 32.

² Upendra Baxi, ‘Taking Suffering Seriously: Social Action Litigation in the Supreme Court of India’ (1985) 4(1) Third World Legal Studies 107.

³ Bihar Legal Support Society v Chief Justice of India, AIR 1987 SC 38.

⁴ See, e.g. Mayur Suresh and Siddharth Narrain, The Shifting Scales of Justice: The Supreme Court in Neo-Liberal India (Orient BlackSwan 2014); Prashant Bhushan, ‘Supreme Court and PIL: Changing Perspectives under Liberalization’ (2004) 39(18) Economic and Political Weekly 1770.

⁵ Usha Ramanathan, ‘In the Name of the People: The Expansion of Judicial Power’ in Mayur Suresh and Siddharth Narrain (eds), The Shifting Scales of Justice: The Supreme Court in Neo-liberal India (Orient BlackSwan 2014).

⁶ Prashant Bhushan, ‘Supreme Court and PIL: Changing Perspectives under Liberalization’ (2004) 39(18) Economic and Political Weekly 1770.

⁷ Varun Gauri, ‘Public Interest Litigation in India: Overreaching or Underachieving?’ (2009) World Bank, Policy Research Working Paper No 5109, 13. <http://documents.worldbank.org/curated/en/675001468042007347/pdf/WPS5109.pdf> accessed 18 November 2017.

⁸ Shylashri Shankar, Scaling Justice, India’s Supreme Court, Anti-Terror Laws, and Social Rights (OUP 2009).

⁹ Gauri focuses on cases under Article 32 of the Constitution of India. Article 32 petitions go through an admissions process similar to SLPs. Article 32 petitions are listed along with SLPs for admission every Monday and Friday. During the admissions hearing, the judges decide whether or not to admit that case based on a variety of factors, including but not limited to, whether the petition discloses a prima facie violation of a fundamental right, whether the matter is justiciable, whether the case should be barred by laches, etc. Therefore, judges exercise discretion in deciding whether or not to admit an Article 32 petition.

¹⁰ Our comparison to the functioning of the US Supreme Court is not meant to suggest that court is a model, which the Indian Supreme Court should follow. The reference is made merely to offer a contrasting example.

¹¹ For example, Justice Bhagwati had called for the creation of a National Court of Appeal. Bihar Legal Support Society v Chief Justice of India, AIR 1987 SC 38. In the Special Leave Petition (C) No 7105 of 19 March 2010, 2010 for Mathai @ Joby v George (2010) 4 SCC 358, two Justices called for the Court to narrow Article 136 jurisdiction.

¹² Law Commission of India, 229th Report on the Need for division of the Supreme Court into a Constitution Bench at Delhi and Cassation Benches in Four Regions at Delhi, Chennai/Hyderabad, Kolkata and Mumbai (2009); TR Andhyarujina, ‘Restoring the Supreme Court’s Exclusivity’ The Hindu (28 February 2014) <http://www.thehindu.com/opinion/lead/restoring-the-supreme-courts-exclusivity/article5076293.ece> accessed 18 November 2017; Nick Robinson, ‘A Court Adrift’ Frontline (3 May 2013)  <http://www.frontline.in/cover-story/a-court-adrift/article4613892.ece> accessed 18 November 2017; KK Venugopal, ‘For Proximate and Speedy Justice’ The Hindu (2 May 2010) <http://www.thehindu.com/opinion/lead/for-proximate-and-speedy-justice/article418735.ece> accessed 18 November 2017.

¹³ See Writ Petition (Civil) 36/2016, V Vasanthakumar v HC Bhatia & Ors. (petition to create a separate national court of appeals between the high courts and the Supreme Court).

¹⁴ See ibid.

¹⁵ Granville Austin, The Indian Constitution: Cornerstone of a Nation (OUP 1966) 27.

¹⁶ Constitution of India, Art 38.

¹⁷ Some high courts have jurisdiction over more than one state. Cumulatively, 24 high courts have jurisdiction over 35 states and union territories.

¹⁸ Constitution of India, Art 141.

¹⁹ Supreme Court of India, ‘Annual Report 2014’ (2015) 76. <http://sci.nic.in/annualreport/annualreport2014-15.pdf>  accessed 18 November 2017.

²⁰ George H Gadbois, Jr., ‘The Supreme Court of India: A Preliminary Report of an Empirical Study’ (1970) 4 J Constitutional and Parliamentary Studies 34.

²¹ Constitution of India, Art 124, § 1.

²² Supreme Court of India, Annual Report 2014, 79.

²³ JUDIS, the official e-reporter of the Supreme Court of India records 900 judgements for 2014.

²⁴ Constituent Assembly Debates, Volume VII (9 December 1948).

²⁵ See generally Marc Galantar, ‘Snakes and Ladders: Suo Moto Intervention and the Indian Judiciary’ (2014) 10 FIU L Rev 69 (discussing instances of suo moto jurisdiction, chiefly by the Supreme Court of India).

²⁶ S.P. Gupta v Union of India, AIR 1982 SC 149. The Court’s own data reveals, however, that even among cases admitted for merits hearing, PILs constitute only 1% of the Court’s cases (though, of course, given the complex nature of many PILs, they may take up a significant proportion of the Court’s time and resources). Nick Robinson, ‘A Quantitative Analysis of the Indian Supreme Court’s Workload’ (2013) 10 J Empirical Legal Stud 570, 590, 598.

²⁷ Kunhayammed v State of Kerala (2000) 6 SCC 359 (“Article 136 of the Constitution is a special jurisdiction conferred on the Supreme Court which is sweeping in its nature. It is a residuary power in the sense that it confers an appellate jurisdiction on the Supreme Court subject to the special leave being granted in such matters as may not be covered by the preceding articles. It is an overriding provision conferring a special jurisdiction providing for invoking of the appellate jurisdiction of Supreme Court not fettered by the sweep of preceding articles. Article 136 opens with a non-obstante clause and conveys a message that even in the field covered by the preceding articles, jurisdiction conferred by Article 136 is available to be exercised in an appropriate case”.).

²⁸ Jamshed Hormusji Wadia v Board of Trustees, Port of Mumbai AIR 2004 SC 1815.

²⁹ As the Court itself has put it, it has stated that it has the power to interfere “even with findings of fact … [as for example when] the acquittal is based on an irrelevant ground, or where the High Court allows itself to be deflected by red herrings drawn across the track, or where the evidence accepted by the trial court is rejected by the High Court after a perfunctory consideration, or where the baneful approach of the High Court has resulted in vital and crucial evidence being ignored, or for any such adequate reason, this Court may feel obliged to step in to secure the interests of justice, to appease the judicial conscience, as it were”. Arunachalam v P.S.R. Sadanatham (1979) 2 SCC 297.

³⁰ Subedar v State of U.P. AIR 1971 SC 125.

³¹ See Nick Robinson, ‘A Quantitative Analysis of the Indian Supreme Court’s Workload’ (2013) 10 J Empirical Legal Stud 570, 598.

³² Supreme Court of India, Annual Report 2014, 76–79 (average of cases filed in 2010–14).

³³ Bihar Legal Support Society v Chief Justice of India AIR 1987 SC 38 (stating the view that cases under Article 136 “would be exceptional by their very nature” and that the Court should not interfere in every case where “some injustice has been done”). See also Jamshed Hormusji Wadia v Board of Trustees, Port of Mumbai AIR 2004 SC 1815; Chandi Prasad Chokhani v State of Bihar AIR 1961 SC 1708; State of Bombay v Rusy Mistry AIR 1960 SC 391; Pritam Singh v State AIR 1950 SC 169.

³⁴ Mathai @ Joby v George (2016) SCC OnLine SC 410.

³⁵ Constitution of India, Art 132, 133, 134. Although the Court’s jurisdiction can be invoked through procuring a certificate of appeal from the high court, this practice is rarely used. One possible reason for the low use of the “Certificate of Appeal” jurisdiction is that while ordinarily a petitioner has 90 days to file an SLP, the limitation for filing an SLP after the high court has refused a certificate of appeal is 60 days. Some experts suggested during interviews and interactions with us that lawyers do not invoke the certificate of appeal process so as to give themselves more time to file in the Supreme Court.

³⁶ Supreme Court of India, Annual Report 2014, 59–63.

³⁷ This is not the limit of the Court’s jurisdiction. The Supreme Court has original jurisdiction with respect to inter-state disputes over certain election matters. Constitution of India, Art 132 & 711. The President may also refer any matter to the Court for its advisory (non-binding) opinion. Constitution of India, Art 143.

³⁸ Supreme Court of India, Practice and Procedure: A Handbook of Information (Supreme Court of India 2010) 35.

³⁹ Supreme Court of India, Annual Report 2014, 76–79. Data available up to November 2014.

⁴⁰ Ibid 63–64. This data is available only up to 31 October 2014.

⁴¹ See, e.g.- Nathan Rehn and others, ‘Justice Without Delay: Recommendations for Legal and Institutional Reforms in Indian Courts’ (Jindal Global Legal Research Paper No 4/2011, 2010).

⁴² Amrit Amirapu, ‘Justice Delayed is Development Denied: The Effect of Slow Courts on Economic Outcomes in India’. <http://www.ideasforindia.in/article.aspx?article=Justice-delayed-is-development-denied-The-effect-of-slow-courts-on-economic-outcomes-in-India> accessed 26 August 2016.

⁴³ Ibid 2–3. From 2005 to 2011, lower court disposals grew only by 7.8%. Ibid 16.

⁴⁴ See Table 1.

⁴⁵ The median is not much better. It is 1260 days, or 3 years and 5 months.

⁴⁶ We are currently gathering data that may shed light on this.

⁴⁷ The median case duration is 4231 days, or 11 years and 7 months.

⁴⁸ The Court’s share of the total in this subset of 123 cases is comparable to its share among all cases. See Table 1.

⁴⁹ Rakesh Kumar Srivastava, ‘A Guide to India’s Legal Research and Legal System’ (April 2014) GlobaLex. <http://www.nyulawglobal.org/globalex/india_legal_research.htm#_10._Law_Reporting> accessed 18 November 2017 (Chief Librarian of the Supreme Court, stating that this reporter is used around 60% of the time before the Supreme Court itself).

⁵⁰ Judgment Information System (JUDIS) <http://judis.nic.in/supremecourt/chejudis.asp>.

⁵¹ For most variables, such discrepancies were avoided through the use of pre-filled drop-down menus that allowed coders to choose among multiple options. Some variables, however, required coders to input unique text rather than use pre-filled drop-down menus.

⁵² Computer code documenting these corrections is available upon request.

⁵³ In the end, some variables had to be dropped from the final data set altogether because, even after review by the research team, the number of unique entries could not be simplified in a manner suitable for quantitative analysis.

⁵⁴ We excluded one-judge benches because they generally deal with procedural matters, such as certain types of minor interim applications, which do not generate merits judgements (although they occasionally generate orders that appear in SCC). We also exclude cases with missing information on which party was appellant or information civil or criminal subject matter, and we exclude PIL cases and continuing mandamus cases, which represent important but very distinct segments on the Supreme Court’s docket.

⁵⁵ Ideally, we would simply include only those cases that reached the Court through the SLP process. In processing the data, however, we discovered that this is not currently feasible. When the Court admits an SLP for a separate merits hearing, the administrative office of the Court converts the case designation from “SLP” to “appeal”, which is also used to designate cases that reach a merits hearing by appeal as of right.

⁵⁶ The consultation took place on 16 January 2016.

⁵⁷ See generally Mathai @ Joby v George (2010) 4 SCC 358.

⁵⁸ Andrew Green and Albert Yoon, ‘Triaging the Law: Developing the Common Law on the Indian Supreme Court’ (unpublished draft, 23 July 2016) <https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2816666> accessed 18 November 2017 6, 12, 13.

⁵⁹ Ibid 15.

⁶⁰ Ibid 15–16.

⁶¹ Ibid 16.

⁶² Supreme Court of the United States, ‘Frequently Asked Questions (FAQ)’. <https://www.supremecourt.gov/faq.aspx> accessed 18 November 2017 (answering “How many cases are appealed to the Court each year and how many cases does the Court hear?” with “The Court receives approximately 7000–8000 petitions for a writ of certiorari each Term. The Court grants and hears oral argument in about 80 cases”). Of course, we understand that the structure of the Indian judiciary and American judiciary differ. While each state in the US has its own supreme court that has exclusive jurisdiction over certain matters, the Indian judiciary is a unitary system where appeals from most courts would come to the Supreme Court eventually. Yet, this difference does not explain why the rate at which the Indian Supreme Court admits the cases that are appealed to it is so much higher than the rate at which the US Supreme Court admits the cases that seek admission.

⁶³ For statistics on SLPs and appeals, see Table 1. (The total number of opinions per year is greater than those listed in Table 1, due to cases within the Court’s original jurisdiction.) Note that the 10-page average length is inclusive of all opinions for a given case.

⁶⁴ See generally Rishad Chowdhury, Note, ‘Missing the Wood for the Trees: The Unseen Crisis in the Supreme Court’ (2012) 5 NUJS L Rev 251; Nick Robinson, ‘India’s Judicial Architecture’ in Sujit Choudhry and others (eds), The Oxford Handbook of Indian Constitutional Law (OUP 2016).

⁶⁵ Y.S. Jagan Mohan Reddy v CBI (2013) 7 SCC 439.

⁶⁶ See, e.g. Maruti Nivrutti Navale v State of Maharashtra (2012) 9 SCC 235; Jignesh v State of Gujarat (2011) 10 SCC 591 (both pertaining to bail where the Court decided – without reference to or discussion of any prior law – whether, on the facts of the case, bail was warranted or not).

⁶⁷ Darshan Gupta v Radhika Gupta (2013) 9 SCC 1.

⁶⁸ Constitution of India, Art 142.

⁶⁹ See, e.g. Ramchander v Ananta (2015) 11 SCC 539; K. Srinivas v K. Sunita (2014) 16 SCC 34; Alok Mishra v Garima Mishra (2009) 12 SCC 270 (deciding whether or not the lower court was right in granting divorce, approving the terms of settlement in a divorce by mutual consent and deciding whether lower courts made the correct decision on custody of minor children).

⁷⁰ Mathai @ Joby v George (2010) 4 SCC 358.

⁷¹ Jamshed Hormusji Wadia v Board of Trustees, Port of Mumbai AIR 2004 SC 1815.

⁷² Jamshed Hormusji Wadia v Board of Trustees, Port of Mumbai AIR 2004 SC 1815.

⁷³ Balakrishnan Rajagopal, ‘Pro-Human Rights but Anti-Poor? A Critical Evaluation of the Indian Supreme Court from a Social Movement Perspective’ (2007) 18 Hum Rts Rev 157, 166.

⁷⁴ Manoj Mate, ‘Globalization, Rights, and Judicial Review in the Supreme Court of India’ (2016) 25 Wash Intl’L LJ 643.

⁷⁵ Varun Gauri, ‘Public Interest Litigation in India: Overreaching or Underachieving?’ (World Bank, Policy Research Working Paper No 5109, 2009) <http://documents.worldbank.org/curated/en/675001468042007347/pdf/WPS5109.pdf> accessed 18 November 2017 13.

⁷⁶ Sudhir Krishnaswamy and Madhav Khosla, ‘Social Justice and the Supreme Court’, in Mayur Suresh and Siddharth Narrain (eds), The Shifting Scales of Justice: The Supreme Court in Neo-Liberal India (Orient BlackSwan 2014) 109, 110.

⁷⁷ Sudhir Krishnaswamy and Madhav Khosla, ‘Social Justice and the Supreme Court’, in Mayur Suresh and Siddharth Narrain (eds), The Shifting Scales of Justice: The Supreme Court in Neo-Liberal India (Orient BlackSwan 2014) 111.

⁷⁸ The best-known paper on this method is John Knowles, Nicola Persico and Petra Todd, ‘Racial Bias in Motor Vehicle Searches: Theory and Evidence’ (2001) 109 J Pol Econ 203. Notable critiques and extensions include Dhammika Dharmapala and Stephen L Ross, ‘Racial Bias in Motor Vehicle Searches: Additional Theory and Evidence’ (2004) 3 Contributions to Econ. Analysis & Pol’y 1; Shamena Anwar and Hanming Fang, ‘An Alternative Test of Racial Prejudice in Motor Vehicle Searches: Theory and Evidence’ (2006) 96 Am. Econ. Rev. 127. See Nicola Persico, ‘Racial Profiling? Detecting Bias Using Statistical Evidence’ (2009) 1 Ann. Rev. Econ. 229, for a review.

⁷⁹ This pattern also depends on an additional assumption, which is that all citizens (or cases) are “marginal”, in the sense that in equilibrium they adjust their behaviour to the behaviour of the police (or the Court). Another more intuitive but less theoretically grounded version of this assumption is that the distribution of behaviour across types (e.g. citizens of different races or petitions brought by different parties) is similar. In this article, we assume that cases that have made their way through the litigation and appellate process and to the Supreme Court have roughly similar distribution of cases across those that are highly likely to be reversed, moderately likely to be reversed, and so on. In Appendix 2, we offer evidence suggesting that this assumption may hold for our data.

⁸⁰ It bears repeating that when we refer to “errors”, we mean anything that would lead the Court to reverse the judgement below, regardless of whether there was a misapplication of law or whether any observer would agree with the Court’s decision. This approach simply accepts the Court’s definition of error in any given case.

⁸¹ We make the following technical assumption: the relative numbers of, for example, cases from Delhi that are 80% likely to be reversed versus 60% likely to be reversed are similar to the relative numbers of such cases from Mumbai. Because we cannot independently observe the quality of cases the court admits, we cannot empirically verify this assumption. But we might expect that, although cases involving different litigants or different geographic origins may be very different across different courts of first instance, the set of cases that make it to the point of appeal at the Supreme Court are highly selected and must share a large number of similar traits: they involve parties that can afford the time and expense of continuing to litigate, stakes that justify continued litigation, an unwillingness to settle, and underlying legal merits that justify the above traits.

⁸² Of course, not all criminal defendants are individuals. In our data, however, 82.5% of defendants are individuals. Further, in the criminal context at least, we believe that the government prosecutor is the more powerful party, even against an institutional defendant.

⁸³ In our data, 66.0% of cases with constitutional challenges have an individual as a party.

⁸⁴ When a case involved more than one plaintiff or defendant, party status was coded based on the first-named party.

⁸⁵ See Table 2.

⁸⁶ George H Gadbois, Jr., ‘The Supreme Court of India: A Preliminary Report of an Empirical Study’ (1970) 4 Journal of Constitutional and Parliamentary Studies 44.

⁸⁷ Of these, only 156 were appeals and petitions from high courts and thus within the scope of the data set employed in this article. See Table 2. The remaining cases were within the original jurisdiction of the Court or were appeals from tribunals. By comparison, the US Supreme Court issued only 391 total merits opinions – whether on constitutional or non-constitutional questions – in its 2010 through 2014 terms. Supreme Court of the United States, ‘Opinions’. <https://www.supremecourt.gov/opinions/opinions.aspx> accessed 18 November 2017 (the number of opinions was attained by adding the number of opinions in each of the five terms according to the SCOTUS site).

⁸⁸ The Judgment Information System <http://judis.nic.in/>.

⁸⁹ We test whether the difference in reversal rates is statistically significant using a t-test of difference-in-means on unpaired data with unequal variances. The difference in reversal rates is not even close to statistical significance (p = 0.191).

⁹⁰ See generally Theodore Eisenberg, Nick Robinson and Sital Kalantry, ‘Litigation as a Measure of Well-Being’ (2013) 62 DePaul L Rev 247 (establishing empirical relationships between litigation activity in region and economic indicators); Nick Robinson, ‘A Quantitative Analysis of the Indian Supreme Court’s Workload’ (2013) 10 J Empirical Legal Stud 570 (documenting correlations between distance to Delhi and petitions filed in the Court).

⁹¹ Regression analysis (unreported but available from authors) yields equivalent conclusions.

⁹² We have GDP data for a twenty-first high court, the High Court of Meghalaya, which split in 2013 from the High Court of Gauhati. We combine it with the High Court of Gauhati in these regressions. For high courts whose jurisdiction covers more than one Indian state, we use the population-weighted average state-level GDP in assigning GDP to high courts.

⁹³ Note that the differences in case numbers cannot be explained by population differences across high courts. In our data, per capita GDP and population are negatively correlated, which strengthens the inference that cases coming from high per capita GDP regions are different in nature than those from lower per capita GDP regions. We should note that we have determined the number of cases coming from these high courts based on the number of reported decisions of the Supreme Court dealing with such matters. We do not currently have independent data on the number of cases being filed for admission from each high court.

⁹⁴ To statistically test our claim that there is no relationship between per capita GDP and reversal rate, we run a bivariate regressions of reversal rate on per capita GDP. The coefficient on per capita GDP is extraordinarily small and not statistically significant (β = −0.000, p = 0.947). For comparison, we run a bivariate regression of number of cases on per capita GDP. Consistent with the graphical evidence, the coefficient on per capita GDP is large and highly statistically significant (β = 0.116, p = 0.007). We note, too, that analysis based on total GDP, rather than per-capita GDP, is nearly identical.

⁹⁵ To statistically test our claim that there is no relationship between distance and reversal rate, we run a bivariate regressions of reversal rate on distance. The coefficient on distance is not statistically significant (β = −0.018, p = 0.524). For comparison, we run a bivariate regression of number of cases on distance. Consistent with the graphical evidence, the coefficient of distance is large, but it is not statistically significant (β = −34.47, p = 0.430).

⁹⁶ This finding supports, but of course does not prove, our hypothesis. For example, these results are consistent with the (implausible) hypothesis that the Court admits cases randomly and also decides cases randomly by reversing 57% of the time, regardless of the facts of the case.

Additional information

Funding

We would like to thank the University of Chicago Center in Delhi for their grant. William Hubbard thanks the Paul H. Leffman Fund and the Coase-Sandor Institute for Law & Economics for research support.