Faculty & Research

Research Productive

Show result

Search Query :
Area :
Search Query :
2738 items in total found

Working Papers | 2023

Towards explainable AI for addressing class overlap problem in supervised learning

Samrat Gupta, Payel Sadhukhan, Kausik Sengupta

The correctness of predictions rendered by an AI/ML model is key to its acceptability. Moreover, in order to foster researchers’ and practitioners’ confidence in the model, it is necessary to render an intuitive understanding of the working of a model. In this work, we attempt to explain a model’s working by providing some insights into the training data. While doing this, it is important to consider that revealing the training data to the users is not feasible due to logistical and security reasons. However, sharing some interpretable parameters of the training data and correlating them with the model’s performance can be helpful in this regard. To this end, we propose a new measure based on Euclidean Minimum Spanning Tree (EMST) for quantification of the intrinsic separation (or overlaps) between the classes of the data. We also validate our approach using Davies Bouldin Index (DB). We use datasets from diverse domains such as finance - credit card fraud using PaySim dataset and bank fraud using BankSim dataset, medical - pregnancy risk prediction using Maternal Health dataset, and marketing expenditure on different media using Advertisement dataset. The results of this study establish the viability of the proposed approach in explaining the working and efficiency of a classifier. Firstly, the proposed measure of class-overlap quantification has shown a better correlation with the classification performance. Secondly, the results on multi-class datasets demonstrate that the proposed measure can be used to determine the feature importance so as to learn a better model.

Read More

Working Papers | 2023

Tests to Determine Employer-Employee Relationships in India: Looking towards the Future?

M. P. Ram Mohan and Sai Muralidhar K.

The underpinning of an employee’s social security benefits is an employer-employee relationship. Workers are traditionally classified as employees (contract of service) or independent contractors (contract for service). Over the years, Indian courts relied on the control, integration and multifactor tests to determine the correct nature of employment contracts. The paper explores the evolution of these tests and examines whether the standards of the burden of proof in classification disputes require modification. The authors then dissect the efficacy of the current multifactor tests in emerging platforms and gig economies by looking at standard form contracts signed by a popular food delivery platform in India. Finally, the ability of newly enacted labour codes, particularly the Code on Social Security 2020, to address the classification conundrum and its consistency with precedents is explored.

Read More

Working Papers | 2023

Unlocking the Power of Accelerators: The Crucial Role of Institutions in Boosting New Venture Performance

Pramendra Singh Tank

Accelerators are becoming popular in helping new ventures succeed. However, current research on their impact on new venture performance does not consider the role of institutional environment. This paper examines the importance of institutions and assesses the impact of accelerators on new venture performance using a generalized difference-in-differences technique on a worldwide accelerator database. The study finds that while accelerators positively impact revenue and equity funding, however the impact is higher in countries with strong institutions. The study highlights the significance of institutions in assessing the impact of accelerators on new ventures, contributing to nascent research in this area.

Read More

Working Papers | 2023

Green Bonds and the Inflation Reduction Act (IRA)

Sanjay Kumar Jain

I study the impact of the Inflation Reduction Act (2022) on the cost of green bonds in the USA. In this study, I employ a difference-in-difference methodology to compare the cost of green bonds issued by US firms in USD after the implementation of IRA by taking green bonds issued by Non-US firms as a counterfactual. I find that after the implementation of IRA, the cost of green bonds in the USA decreased compared to that in other countries. Findings indicate that Governments can mitigate constraints on the cost and supply of climate financing through fiscal policy interventions that improve business environments for firms and spur the demand for cleaner fuels and products from individuals

Read More

Working Papers | 2023

Environmental Claims under Indian Insolvency Law: Concepts and Challenges

M. P. Ram Mohan and Sriram Prasad

The intersection between environmental liability claims and insolvency of the entity concerned have grown increasingly complex. Over the years, India has seen enactment of several laws and proactive judicial decisions to ensure liability from environmental harm are addressed through application of no-fault and absolute liability principles. A consequence of these principles is, if an entity harms the environment, they must bear the cost of clean-up. If the entity defaults on the compensation payment or is unable to pay, then, under the Insolvency and Bankruptcy Code 2016 (IBC) they may be able to declare themselves as insolvent. When admitted under insolvency, a moratorium on all claims is imposed. Once resolution has taken place, the corporate debtor is provided with a “fresh start”, relieving the debtor from all its previous debts and liabilities. If the debtor goes into liquidation, through the waterfall mechanism, financial creditors are given priority over environmental claimants who would mostly be categorised either as contingent claimants or decree holders. In these scenarios, insolvency law supersedes environmental law/policy by design, creating a visible human rights implication. While the IBC seems to be agnostic to social causes, there are other avenues to deal with social causes, such as the Public Liability Insurance Act which deals with hazardous environmental accidents in a limited way. In the paper, we argue, insurance provides a better framework to resolve environmental liabilities and that the insurance schemes should remain intact regardless of a fresh start.

Read More

Working Papers | 2023

Social Token Economics for GAU Coin

Nishant Krishna, Amit Garg, Gaurav Kumar Kedia, and Aprajita Mishra

This paper discusses the Token Economics around the GAU Coin. All the aspects of Token Economics, including rewards system, incentivisation, ways to increase adoption, methods to measure social impact, strategies to bring the developer community into the ecosystem, commodity flows, and many others are discussed. Various GAU Coin Ecosystem Partners and Enablers and how they interact with the ecosystem are also identified. Later, GAU NFT for the participants of the GAU Coin network can also be thought of, along with creating the entire network of artists and infrastructure around it. GAU Coin is a Social Token and is not a Cryptocurrency. The GAU Coin represents the digital form of what a person owns as part of the GAU ecosystem and brings Social Credibility.

Read More

Working Papers | 2023

Neonatal Mortality Rate (NMR) in India: A study using one-way ANOVA and multiple linear regression (MLR)

Rohan Kar and Sourav Bikash Borah

Neonatal Mortality Rate (NMR) is of grave concern for India and other low-income and middle-income countries aspiring to meet the Sustainability Development Goals by 2030 (SDG30). As per government estimates, the NMR in India was 30 per 1000 live births in 2019. Achieving the target of 12 deaths per 1000 live births by 2030 remains a considerable challenge. This study was conducted using indicators from the State Health Index Round 4 (SHI-R4), covering 34 states and union territories (N=34). One-way ANOVA was performed to identify significant differences in mean NMR, if any, between states and union territories (UTs). Later, a model was built using multiple linear regression techniques to predict the NMR in India using indicators available in the SHI-R4. The model obtained had an R 2 value of 0.37. Among the significant predictors that most influenced the NMR were the average occupancy of a district Chief Medical Officer (CMO), the number of caesarean sections performed at First Referral Units (FRUs), and the Kayakalp score of public health facilities. The study findings add to the existing scholarship on NMR in India. The results are significant both in terms of future research and policymaking decisions.

Read More

Working Papers | 2022

In Pursuit of Balance: Vicarious Liability Doctrine in the United Kingdom and India

M. P. Ram Mohan and Sai Muralidhar K

The Doctrine of Vicarious Liability is a unique exception to the principle of fault-based liability and holds persons liable for the actions of third parties. The recent verdicts in WM Morrison Supermarkets v Various Claimants (2020) and Various Claimants v Barclays Bank (2020) by the UK Supreme Court restricting the scope of vicarious liability through its interpretation of the akin to employment test as well as the close connection test deserves scrutiny. The Supreme Court apart from reaffirming the traditional distinction between independent contractors and employees also has limited the circumstances in which claims of vicarious liability can be upheld. Given that tort law in India is deeply rooted in the common law of the UK, it is unsurprising that principally vicarious liability in India and UK has evolved similarly. The paper analyses the various principled justifications behind the doctrine and focuses on the various tests such as the akin to employment test, course of employment test & close connection test which are used to impose liability. Further, it comprehensively examines the evolution of the doctrine in the UK and India, and analyses the varying approach taken by the judiciary in both countries against the backdrop of the socio-economic conditions of the workforce. Lastly, the paper identifies the difficulties that the doctrine may face in the future.

Read More

Working Papers | 2022

Performance of quality factor in Indian Equity Market

Joshy Jacob, Pradeep K.P., and Jayanth R.Varma

We study the characteristics of Quality factor (QMJ) in India, which is the second largest emerging market. Dimensions of quality factor are impacted by the weaker enforcement of corporate governance norms in emerging markets. Diversion of revenues by promoters would result in poor profitability, while tunneling of profits would result in lower payout and lower growth. Therefore, investors are likely to attach greater significance to the quality dimensions in stock pricing. Consistent with this hypothesis, the Quality factor is even more important for asset pricing in India than in developed markets. The QMJ factor earns a four factor alpha of 0.92% per month, significantly outperforming the other widely employed factors, market, size, value and momentum factors. A long-only Quality factor earns an alpha of 0.69% per month. The alpha of quality factors is highly significant, judged by the thresholds recommended by Harvey, Liu, and Zhu (2016). The key drivers of the alpha are profitability and payout, which are both consistent with the tunnelling hypothesis. Besides the alpha, the low portfolio churn, lower risk, shorter drawdowns, and viability of long-only strategies restricted to large capitalization stocks suggest that portfolios tilted towards high-quality stocks are highly attractive to institutional and retail investors.

Read More

Working Papers | 2022

Risk information - normal markets and the COVID-19 pandemic period

Pranjal Srivastava and Joshy Jacob

The paper investigates how the market infers changes in the firm-level discount rate (risk information) in normal and turbulent times. The study focuses on two key sources of risk information, earnings announcements of firms and changes in the market risk premium. We employ a recently proposed measure that limits the impact of event risk while estimating the forward-looking risk information from option prices. We find that both earnings announcements and the changes in market risk impact firm-level discount rates, but both sources exhibit a significant time variation. The impact of market risk changes is lower in favorable conditions and higher during crisis periods. Using COVID19 as an exogenous shock, we show that the influence of earnings announcements becomes insignificant during a crisis. The results suggest lower attention to firm-specific risk factors in times of a systemic crisis, in contrast to normal times.

Read More
IIMA